Split (1)

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

identifier stringlengths 11 32	pdf_url stringlengths 17 4.62k ⌀	lang stringclasses 120 values	error stringclasses 1 value	title stringlengths 2 500 ⌀	source_name stringlengths 1 435 ⌀	publication_year float64 1.9k 2.02k	license stringclasses 3 values	word_count int64 0 1.64M	text stringlengths 1 9.75M
https://openalex.org/W2118516311	https://www.scienceopen.com/document_file/55c2616a-12e9-4ad5-96ba-bf094e77c47b/ScienceOpen/001_Day.pdf	English	null	Breaking the Campus Bubble: Informed, Engaged, Connected	Electronic workshops in computing	2,007	cc-by	3,475	Keywords Interacting with public displays, mobile phones, expressing opinions, human connectedness, voting. Breaking the Campus Bubble: Informed, Engaged, Connected Nick Day Lancaster University Computing Department LA1 4WA +44 (0) 1524 510492 nick@njday.com Motoko Toma Bosch Japan SAP +81 (0) 806567 2484 motoko.toma@jp.bosch.co Nick Day Lancaster University Computing Department LA1 4WA +44 (0) 1524 510492 nick@njday.com Alan Dix Lancaster University Computing Department LA1 4WA +44 (0) 1524 510319 alan@hcibook.com Corina Sas Lancaster University Computing Department LA1 4WA +44 (0) 1524 510318 corina@comp.lancs.ac.uk Corina Sas Lancaster University Computing Department LA1 4WA +44 (0) 1524 510318 corina@comp.lancs.ac.uk Motoko Toma Bosch Japan SAP +81 (0) 806567 2484 motoko.toma@jp.bosch.co Dave Clare 17 Millthrop Sedbergh LA10 5SP +44 (0) 7811 7044474 dave_clare_millthrop@hotmail.com Dave Clare 17 Millthrop Sedbergh LA10 5SP +44 (0) 7811 7044474 Chris Bevan University of Bath Computer Science and Psychology +44 (0) 7770 640406 bevan.chris@gmail.com +81 (0) 806567 2484 motoko.toma@jp.bosch.co dave_clare_millthrop@hotmail.com ABSTRACT This often produces insular and self-centred behaviour which can easily develop into a feeling of isolation, particularly on the Lancaster campus which is physically isolated from the city itself. This lends itself to the notion of students living in a “bubble”, with students becoming more and more detached from the world outside of the university campus. Students at the university are aware of this problem, as the following quote suggests: This paper introduces UniVote, a system supporting mobile phone-based interaction with public displays. The case study carried out at Lancaster University indicates that the campus "bubble" in which students live can lead to feelings of isolation within an insular community cut off from the outside world. UniVote makes use of a voting system to help elicit user involvement, keep users informed of campus- and world-wide events and news and create a sense of community. Findings of this preliminary study suggest that the campus "bubble" can indeed be broken, and the voting component of the system particularly fosters interaction and human connectedness. “Campus is a place cut-off from the rest of the world… Students are disinterested and often blatantly unaware of what’s going on around them.” “Campus is a place cut-off from the rest of the world… Students are disinterested and often blatantly unaware of what’s going on around them.” It is this issue of student isolation from the outside world and each other that we would like to address by developing the UniVote system, through understanding campus life and exploring ways in which technology can assist in making the campus more responsive to student needs. Two concepts are particularly relevant here: user engagement and human connectedness, which are considered throughout the entire design process. © Day, Sas, Dix, Toma, Bevan, Clare, 2007 Published by the British Computer Society Volume 2 Proceedings of the 21st BCS HCI Group Conference HCI 2007, Lancaster University, 3-7 September 2007 Devina Ramduny-Ellis & Dorothy Rachovides (Editors) 1. INTRODUCTION Across the Lancaster campus there is an ongoing deployment of public displays designed to enable pervasive interaction as well as broadcast more traditional multimedia content. This infrastructure offers an opportunity to explore ways in which technology can improve student quality of life on university campuses. People in general have a desire to be involved in meaningful social relationships, a topic which has been explored in relation to technology by Agamanolis during his development of nine human connectedness principles [1]. These principles explore how such essential relationships are built, maintained or enhanced by technology [2], and as such can assist in designing effective public displays that encourage acceptance and entice interaction through building and maintaining a relationship with the user. To build this relationship a public display must engage its users and encourage interaction. A phenomenon coined as the honey pot effect was observed by Brignull and Rogers [5], which describes the social buzz produced by an increasing number of people gathering in the proximity of a public display, attracted by its interaction potential and the social payoff of congregating. Such displays would only need to attract the critical mass of people before the social facilitation of the display would maintain a high level of users’ engagement and interaction with the display. However, there are no accounts of how such phenomena would evolve over time, particularly once the novelty of the display wears off. Our work is driven by the ever-increasing levels of stress and work experienced by university students, with students’ anxiety levels being particularly high during their first year at university [6]. Students therefore have very little time to explore new things [10] and often have a general lack of interest in anything that is not directly related to student life. © Day, Sas, Dix, Toma, Bevan, Clare, 2007 Published by the British Computer Society Volume 2 Proceedings of the 21st BCS HCI Group Conference HCI 2007, Lancaster University, 3-7 September 2007 Devina Ramduny-Ellis & Dorothy Rachovides (Editors) Through the study of ambient displays Mankoff and Dey [8] identified that the information source is a crucial factor in user adoption and acceptance of a display. Since interaction with public displays is usually short (even for users interested in its content) the display has to ensure a transition from the users periphery to the focus of their attention. 2.2 System Architecture UniVote is based on a client-server model (Figure 1) and has been designed with multiple campus installations in mind. Figure 1. System Architecture We hope to reduce the feeling of isolation on campus by keeping students better informed of campus- and world-wide events and allowing them to voice their own opinions on such matter through a voting procedure. Given that interaction with public displays often encounters resistance from a public audience [5], we feel that a voting system will help to elicit audience involvement as it has proven to do so in radio and TV. This research is particularly relevant given that there has been little work exploiting voting as an activity that can successfully promote interaction with public displays [13][12]. Opinionizer [5] includes similar voting capabilities to UniVote based on open-ended questions, however a serious limitation is that it does not provide anonymity to voters and cannot allow for simultaneous interaction by multiple users due to its use of a keyboard for input. The mobile phone interaction of UniVote will overcome these limitations and will consist of closed multiple-choice questions to ensure that the threshold to participation is perceived as low, so that the benefits of interacting outweigh the perceived costs [5]. Figure 1. System Architecture Each UniVote display will require its own Mac running the server application and front end. The components of the system are described below. Our work focuses on the use of mobile phones to act as a display and input for larger public displays to leverage on the strengths of both components: the personal control and market saturation of mobile phones (particularly in a predominantly student-centred environment); and the larger presentation space, and greater computational power and bandwidth of public displays [12]. It is hoped that the combination of the rich media potential of public displays and the communications possibilities of mobile phones, will produce a truly interactive system to entice user interaction and try to build a sense of community on Lancaster campus. 2. THE UNIVOTE SYSTEM The UniVote system and further screenshots can be downloaded from www.univote.co.uk. The system has been developed and tested on a Nokia 6230 and Mac OS 10.4. 1. INTRODUCTION It is hoped that by providing information of interest to students (broadening their awareness of the outside world) and asking them to provide their input, the display will be woven into the fabric of students’ interests and needs and will help to increase both acceptance and interaction. In addition an aesthetically pleasing front end will help to entice interaction in situations where the content on the display does not speak directly to the interests of students. have a strong interest for knowledge of these areas, despite being uninformed, and responded enthusiastically to the proposed system. Naturalistic observation found that while there were many paper-based notices and advertisements along common campus routes, people rarely stopped to read them. Focussing on observing the natural patters of peoples’ movement throughout campus will inform the decision of where the UniVote displays should be located. 2.2.1 Client Application pp The client application was designed to run on any J2ME and Bluetooth enabled mobile phone and is distributable over the air using a WAP connection. It uses the standard J2ME API to ensure it inherits the look-and-feel of the host mobiles’ operating system, and operates in a step-by-step linear manner to ensure maximum usability (Figure 2). Users must have this application running on their phone before they can cast a vote. Users are connected to the server application nearest to their location (as determined by the Bluetooth protocol) which is highly likely to be the UniVote public display they are intending to interact with. On the development phone (Nokia 6230) it takes on average 16 seconds to cast a vote (including starting the application). 2.3 Information Sources There are two sources of information used in UniVote: news (displayed on the front end) and questions (on which users vote). The BBC was chosen as the source of news data because they provide an XML and two-minute RealVideo feed of the latest news headlines. The XML feed (parsed by the web service, and presented as a vertical scrolling marquee) and the two-minute video summary are embedded into the front end. At this stage of the development it was decided that the questions would be added to UniVote by university staff and not by students directly, partly due to the proof-of-concept nature of this work. However, giving students the capability to manipulate campus displays in real-time could have negative consequences for the campus community through the posting of offensive or nonsensical questions (cf. “Error prevention and user control” [8]). Figure 2. Client Application 2.2.2 Server Application pp The server application coordinates communication between client devices and the central database. It accepts incoming connections, retrieves the questions from the database (applicable to the display the user is interacting with) and sends them to the user’s mobile phone. It also sends users’ votes to the central database where they are saved. Our needs analysis involved both questionnaire and observation of students on campus. The questionnaire was administered to 31 students and captured factual data including: access to television, radio and Internet; level of interest in campus, local, national and international news; news categories of interest; level of knowledge about current affairs; and level of interest for a campus news system. The findings suggested that campus residents used the Internet as their main source of information (partly due to inadequate TV and radio signal on campus) and as a result residents were insufficiently informed about current news and events – supported by an overall poor level of knowledge in factual questions. Off-campus students preferred national news, politics and sport, whereas campus-based students were more interested in information about social events on a campus level. Both on- and off-campus students Figure 2. Client Application can be viewed remotely by university staff as well as a timeline to show how voting is spread throughout any given 24-hour period. There is also a simulator for university staff to test a screen configuration before physically deploying it to an on- campus display. 2.2.3 Front End The front end application (Figure 3) running on the public displays shows a scrolling news headlines feed and two-minute headline summary; cycles through the questions, inviting users to cast a vote and displaying a voting outcome; displays information on how users can cast a vote; and provides immediate feedback of new incoming votes by means of an animated bar chart. The front end application was implemented using Director, a common choice for delivering content on large public displays [5][13]. This notion of control also ensures that the questions posted on UniVote are neither highly controversial nor have right or wrong answers, so people do not feel defensive or inhibited in expressing their opinions. Public expression of opinions or attitudes for which people hold strong beliefs is often problematic [3][4][9] and can lead to the “spiral of silence” effect [11] in which those who hold minority opinions will choose to remain silent because of fear of isolation from the majority [14]. Given we are trying to increase the feeling of community and diminish the sense of isolation, this is certainly something we wish to avoid – if people perceive support for their opinions from a social network, they are more willing to express them [7]. Figure 3. Front End Application 3. EVALUATION UniVote was evaluated by means of a lab-based evaluation session comprising of 23 first year undergraduate Computer Science students – 21 male, 2 female, with an average age of 20. Only 21% of participants had prior experience of public display systems such as UniVote. The front end of the system was projected at the front of the room. Figure 3. Front End Application 2.2.4 Central Database & Web Service A central MySQL database and PHP-based web service are remotely hosted at univote.co.uk to allow screens to be deployed around campus wherever an Internet connection is available. The central database contains the questions, screen configuration settings and votes, and the web service provides a compatibility layer for passing data to the Director front end. Participants were given a brief presentation as an introduction to the system and the context in which it would be used in a real-world situation. Participants were asked to complete two tasks and were given an incentive for their participation. The first task required participants to use the web administration site for posting their own questions, as a means of gaining familiarity with how the system works. Qualitative and quantitative data was collected from participants through a worksheet. The second task required participants with Java- and Bluetooth-enabled mobile phones to download the client application and cast a vote, and complete another worksheet based on their interaction experience with the voting functionality of UniVote. 4. DISCUSSION & CONCLUSION Despite its prevalence, the lab-based evaluation session has limitations that need to be acknowledged. Ethnographic studies on future versions of UniVote running “in the wild” for a long period of time will undoubtedly capture aspects of the system’s success that we cannot foresee at this stage. While the evaluation session showed a positive reaction to the system, a longer-term evaluation would be required to measure the system’s success in terms of both human connectedness, and whether the system would outlast the novelty effect and recreate the honey pot effect to ensure continued interaction. [8] Mankoff, J. and Dey, A.K. From Conception to Design: A Practical Guide to Designing Ambient Displays. In O’Hara K., Perry M., Churchill E., Russell D., Public and Situated Displays: Social and Interactional Aspects of Shared Display Technologies, pages 210-230, Kluwer Academic Publishers, 2003. [9] Miller, K. Communication theories: perspectives, processes, and contexts. (2nd ed). New York, NY: McGraw-Hill, 2005. There are two features that we feel are very important to include in the next version of UniVote (given the evaluation findings) to help encourage interaction. Firstly, the system should handle questions with more than three answer options. This would be essential for using UniVote in student elections – a key usage area identified by the study participants and during the needs analysis. Secondly, a “points system” could be used to reward frequent voters with gifts and vouchers from the student union, who would be identified by the unique Bluetooth address of their mobile phone. [10] Newton, F.B. (1998). The stressed student, About Campus. (May-June 1998), 4-10. [11] Noelle-Neumann, E. The Spiral of Silence. Chicago: University of Chicago Press, 1993. [12] O’Neill, E., D. Woodgate and V. Kostakos, 2004, “Easing the wait in the Emergency Room: designing public information systems”. ACM Designing Interactive Systems, Boston, MA. The outcomes of this study highlight that interaction with shared displays has a lot to offer through encouraging users to express opinions by casting votes on topics of interest, and our preliminary findings suggest that the campus “bubble” can indeed be broken. A series of real-world testing and evaluations would have to be conducted to confirm this with respect to human connectedness principles. Unsurprisingly we replicated the general finding that a strong well-founded rationale for developing an interactive system is the best [13] Scheible, J. and Ojala, T. 2.2.5 Web Administration Site The web administration site allows university staff to manage the UniVote system. A central bank of questions is available to which staff can add, edit or delete questions. Questions comprise of a single question statement and three fixed answers and can be assigned start and expiry dates/times. Questions from this central bank are then assigned to question groups. Each screen around campus has a unique name and is assigned specific question groups to show: only questions from the central bank which belong to those groups will be shown on this screen. Screens have three configuration options: colour scheme (red, green, blue, black); question change interval (10, 20 or 30 seconds; 1-5 minutes); and voting information interval (1-10 questions) which controls how frequently the front end shows more detailed and eye-catching instructions on how to cast a vote. Voting statistics (for any screen around campus) Initial reactions to the system after the first task were positive, with 87% of participants reporting they would use such a system if it was deployed around the university campus. All participants found it easy to use. Over two thirds thought it suitable for deployment around campus, with one student particularly liking the idea of anonymous voting, and another suggesting that their real-world usage would very much depend on the types of questions available on it. As previous discussed this is crucial to the success of the system – participants predictor for its success, particularly when this is matched with users’ greatest interests and needs. suggested topics including campus-related topics (where to build more parking spaces, new bus routes); student topics (student elections particularly); current events; or just anything fun or useful. Two usability issues were identified during this phase. A few students pointed out that the scrolling news feed would be difficult to read on smaller screens, suggesting the need for different “themes” of the front end which are optimised for certain screen resolutions. While the immediate feedback of the animated bar chart was thought advantageous, one student noted that the bar chart would be continually animating during periods with high voting levels. This suggests the need for scheduling of screen updates, for which further research would have to be conducted to find the optimal trade- off between immediacy of feedback and system capacity. 5. REFERENCES [1] Agamanolis, S. Designing Displays for Human Connectedness. In O’Hara K., Perry M., Churchill E., Russell D., Public and Situated Displays: Social and Interactional Aspects of Shared Display Technologies, pages 309-334, Kluwer Academic Publishers, 2003. [2] Agamanolis, S. New Technologies for Human Connectedness, ACM Interactions, 12, 4 (July - August 2005), 33 – 37. [3] Anderson, J.A. Communication theory: epistemological foundations. New York, NY: The Guilford Press, 1996. Despite the positive initial reaction, only 39% of participants were able to complete the second task (despite 70% having compatible phones). This was due to two factors. There were major problems downloading the client application: although 91% of participants had WAP-enabled mobile phones only 4% had experience using it. This suggests that WAP is not the ideal distribution method for such applications and a more convenient method such as SMS should be considered. Secondly, implementations of the J2ME API do vary between mobile phone manufacturers which caused runtime errors for some participants: further testing is required here. [4] Borovoy, R., Martin, F., Vemuri, S., Resnick, M., Silverman, B. and Hancock, C. Meme Tags and Community Mirrors: Moving from Conferences to Collaboration, 1998. [5] Brignull H. and Rogers Y. Enticing People to Interact with Large Public Displays in Public Spaces. In Proceedings of INTERACT’03 (Zürich, Switzerland, Sep. 2003), 17-24. [6] Cooke, R., Bewick, B.M., Barkham, M. Bradley, M. and Audin, K. Measuring, monitoring and managing the psychological well-being of first year university students. British Journal of Guidance and Counselling 34, 4, (2006), 505-517. Aside from these usability and technical problems, the general consensus of the system was highly positive and the majority of students would use such a system if deployed around campus. [7] Hayes, A.F., Shanahan, J. and.Glynn, C.J. Willingness to express one’s opinion in a realistic situation as a function of perceived support for that opinion, International Journal of Public Opinion Research, 13, 1, (2001), 45-58. 4. DISCUSSION & CONCLUSION MobiLenin combining a multi- track music video, personal mobile phones and a public display into multi-user interactive entertainment. In Proceedings of ACM Multimedia 2005, 199-208. [14] Shoemaker, P.J., Breen, M. and Stamper, M. Fear of isolation: Testing an assumption from the spiral of silence, Irish Communications Review, 8, (2000), 65-78.
https://openalex.org/W2900993236	https://europepmc.org/articles/pmc6267503?pdf=render	English	null	Size Fractionation of Fluorescent Graphene Quantum Dots Using a Cross-Flow Membrane Filtration System	Nanomaterials	2,018	cc-by	9,342	Received: 28 October 2018; Accepted: 12 November 2018; Published: 21 November 2018 Abstract: Graphene quantum dots (GQDs) have received great attention as optical agents because of their low toxicity, stable photoluminescence (PL) in moderate pH solutions, and size-dependent optical properties. Although many synthetic routes have been proposed for producing GQD solutions, the broad size distribution in GQD solutions limits its use as an efﬁcient optical agent. Here, we present a straightforward method for size fractionation of GQDs dispersed in water using a cross-ﬂow ﬁltration system and a track-etched membrane with cylindrical uniform nanopores. The GQD aqueous suspension, which primarily contained blue-emitting GQDs (B-GQDs) and green-emitting GQDs (G-GQDs), was introduced to the membrane in tangential ﬂow and was fractionated with a constant permeate ﬂow of about 800 L m−2 h−1 bar−1. After ﬁltration, we observed a clear blue PL spectrum from the permeate side, which can be attributed to selective permeation of relatively small B-GQDs. The process provided a separation factor (B-GQDs/G-GQDs) of 0.74. In the cross-ﬂow ﬁltration system, size-dependent permeation through cylindrical nanochannels was conﬁrmed by simulation. Our results demonstrate a feasible method facilitating size fractionation of two-dimensional nanostructures using a cross-ﬂow membrane ﬁltration system. Since membrane ﬁltration is simple, cost-effective, and scalable, our approach can be applied to prepare a large amount of size-controlled GQDs required for high performance opto-electronics and bio-imaging applications. Keywords: graphene quantum dots; membrane ﬁltration; tangential ﬂow ﬁltration; 2D nanomaterials nanomaterials nanomaterials nanomaterials nanomaterials nanomaterials Size Fractionation of Fluorescent Graphene Quantum Dots Using a Cross-Flow Membrane Filtration System Sang-Gu Yim 1, Yong Jin Kim 2, Ye-Eun Kang 1, Byung Kee Moon 3, Eun Sang Jung 4,* and Seung Yun Yang 1,* 1 Department of Biomaterials Science, Life and Industry Convergence Institute, Pusan National University, Miryang 50463, Korea; sg.yim0425@gmail.com (S.-G.Y.); kang.ye0525@gmail.com (Y.-E.K.) y g g y g g y g 2 Center for Multidimensional Carbon Materials, Institute of Basic Science, Ulsan National Institute of Scien and Technology, Ulsan 44919, Korea; dibykim@gmail.com 3 Department of Physics, Pukyong National University, Busan 48513, Korea; bkmoon@pknu.ac.kr 4 Department of Bio Environmental Energy, Life and Industry Convergence Institute, Pusan National University, Miryang 50463, Korea y, y g , * Correspondence: esjung@pusan.ac.kr (E.S.J.); syang@pusan.ac.kr (S.Y.Y.); Tel.: +82-55-350-5433 (E.S.J.); +82-55-350-5382 (S.Y.Y.) Received: 28 October 2018; Accepted: 12 November 2018; Published: 21 November 2018 Nanomaterials 2018, 8, 959; doi:10.3390/nano8110959 www.mdpi.com/journal/nanomaterials 1. Introduction Graphene quantum dots (GQDs) are nano-sized monolayers or few-layer graphene sheets (below 20 nm) with a two-dimensional hexagonal lattice structure [1–3]. GQDs exhibit quantum conﬁnement because the size of the system is comparable with the de Broglie wavelength of an electron, which is not present in bulk graphene [2]. As quantum conﬁnement is dependent on the size of the system [4–7], GQDs generate different photoluminescence (PL) spectra as a function of the size of the GQD; this is well-known from theoretical predictions and experimental results [8,9]. Thus, GQDs have received tremendous attention in nanoscience and nanotechnology [10,11]; they have been used as a key nanomaterial in diverse applications including sensors, energy conversion, and bioanalysis [12–14]. O i l i i h l d i l l f i li i li i ll i bi l i l di Optical imaging agents have played a crucial role for visualizing living cells in biological studies and detecting biological reactions involving bioactive agents like enzymes and antibodies [15]. Recently, Nanomaterials 2018, 8, 959; doi:10.3390/nano8110959 www.mdpi.com/journal/nanomaterials 2 of 11 Nanomaterials 2018, 8, 959 GQDs with tunable optical properties have been used in several biological applications [16,17]. GQDs smaller than 10 nanometers easily pass cell membranes and are very stable without photo-bleaching [18]. Moreover, they exhibit better biocompatibility compared with other nano-sized imaging agents [15,19]. These superior properties make GQDs very promising candidates for cell-imaging [20,21] and many other biological applications [22]. GQDs have been synthesized by diverse methods based on top-down or bottom-up approaches [23,24]. Top-down methods involve cutting process graphene sheets into GQDs by hydrothermal method, laser ablation, electrochemical oxidation, or oxygen plasma treatment [23]. In the case of the bottom-up approach, GQDs are synthesized starting from molecular precursors, such as citric acid, glucose, or aromatic hydrocarbons for high yields [24]. These molecules can be converted into GQDs using different chemical or physical routes, including microwave pyrolysis, polymerization, and dehydrogenation. The cage-opening of a fullerene also can be used [8]. However, these synthetic methods inevitably produce GQDs with a broad size distribution. Since the optical properties of two-dimensional nanomaterials including GQDs depends on their size and shape, separation of GQDs according to size is important to expand the application range of this material as an optical agent [25,26]. Additional separation processes have been used to produce uniform GQDs following synthesis. 1. Introduction High-speed centrifugation techniques using cascade sedimentation depending on mass are effective for size fractionation of particles (solutes), but there is a fractionation limitation using centrifugation due to small differences in mass between GQDs [27]. Chromatographic separation methods have been used for GQD puriﬁcation, but this approach is difﬁcult to obtain large quantities for practical uses [28]. Alternatively, a pressure-driven membrane ﬁltration system has been used for GQD separation [29]. Depending on the pore size of the membrane, solute sizes with selective cut-off can be obtained during ﬁltration process [30]. The membrane ﬁltration system can operate in two modes: Dead-end ﬁltration and cross-ﬂow ﬁltration [31]. In dead-end ﬁltration, the feed ﬂow is subjected to parallel permeation (ﬁltrated) ﬂow and the feed solution completely passes the membrane, similar to ﬁltration with a syringe ﬁlter. Filtered matter larger than the pore size accumulates on the membrane surface or remains plugged within the membrane [32,33]. However, severe fouling on the membrane surface occurs during ﬁltration of GQD solutions, thereby requiring additional treatment to ensure high purity [12,29]. In addition, surface cleaning of the membrane or membrane replacement is required to recover the permeate rate. Cross-ﬂow ﬁltration, also known as tangential ﬂow ﬁltration, is a ﬁltration technique in which the initial feed solution passes tangentially along the surface of the membrane [34]. A pressure difference across the membrane drives components that are smaller than the pores through the membrane. Components larger than membrane pores are retained and pass along the membrane surface, thus ﬂowing back to the feed reservoir. This ﬁltration mode could minimize membrane fouling and provides a stable ﬂux compared to dead-end ﬁltration [35,36]. For a given membrane ﬁltration system and feed solution, cross-ﬂow velocity is the primary parameter that determines mass transfer through the membrane [37]. Although cross-ﬂow ﬁltration has been successfully applied to sort 1-dimensional carbon nanotubes by their sizes [38], this membrane system has not been investigated for fractionating 2D nanomaterials, including GQDs. Here, we report a facile method for separating ﬂuorescent GQDs dispersed in water by size based on cross-ﬂow ﬁltration using membranes with uniform pore size. Selective permeation of relatively small GQDs through the membrane occurred for a given pair of cross-ﬂow velocity and pore size values; this result was conﬁrmed by comparing with simulation results. 2.2. Membrane Filtration Tests paper and stored at 25 °C. In dead-end ﬁltration tests, 20 µg/mL of the GQD solution was ﬁltered at 0.1 bar using a stirred cell system (Millipore Co., Burlington, MA, USA), which has an effective ﬁltration area of 4.1 cm2. Filtration experiments were conducted using 25 mm-disc track-etched membranes (No. 110603, Whatman, Inc., Maidstone, UK) with uniformly distributed cylindrical nanopores (50 nm diameter) [40]. The membrane was pre-wetted with DI water for 1 day prior to mounting in the stirred cell to allow polymer chains to swell and reach equilibrium [41,42]. Filtration experiments were conducted at a constant stirring speed and temperature (25 ◦C). The working pressure was adjusted using N2 gas and applied to the feed side. The permeate ﬂux was determined using an electronic balance. Before ﬁltering GQDs, DI water was circulated for 20 min in order to obtain a non-fouled state permeable membrane. 2.2. Membrane Filtration Tests In dead-end filtration tests, 20 μg/mL of the GQD solution was filtered at 0.1 bar using a stirred cell system (Millipore Co., Burlington, MA, USA), which has an effective filtration area of 4.1 cm2. Filtration experiments were conducted using 25 mm-disc track-etched membranes (No. 110603, Whatman, Inc., Maidstone, UK) with uniformly distributed cylindrical nanopores (50 nm diameter) [40]. The membrane was pre-wetted with DI water for 1 day prior to mounting in the stirred cell to allow polymer chains to swell and reach equilibrium [41,42]. Filtration experiments were conducted at a constant stirring speed and temperature (25 °C). The working pressure was adjusted using N2 gas and applied to the feed side. The permeate flux was determined using an electronic balance. B f filt i GQD DI t i l t d f 20 i i d t bt i f l d t t p In circulating cross-ﬂow ﬁltration experiments, 20 µg/mL of the GQD feed solution was introduced into a track-etched membrane mounted in a customized module (Figure 1a) and ﬁltered at the same conditions used in the dead-end ﬁltration experiments. The crossﬂow module had an effective ﬁltration area of 0.5 cm2 and was operated with a peristaltic pump (Cole-Parmer, IL, USA) for in order to maintain consistent solution inﬂow. The feed solution ﬂowed into the upper chamber of the module with 15 mL/min ﬂow rate. The operation pressure was 0.01 bar and was measured with a pressure gauge installed in the feed line. 1. Introduction We evaluated the feasibility of the crossﬂow membrane system for size fractionation of GQDs by characterizing the selectivity and permeability in membrane ﬁltration, and the fouling on membrane surfaces by comparing with a conventional dead-end ﬁltration system. 3 of 11 Nanomaterials 2018, 8, 959 2.1. Preparation of Fluorescent GQD Suspensions 2 Materials and Methods 2.1. Preparation of Fluorescent GQD Suspensions 2 Materials and Methods GQD aqueous suspensions primarily containing blue-emitting GQDs (B-GQDs) and green-emitting GQDs (G-GQDs) were prepared with a bottom-up method from organic compounds [39]. Brieﬂy, 20 g of ethanol, 3 g of 70% nitric acid and 1 g of acetylacetone were mixed and heated in a high-pressure reactor of 60 cm3. The reaction proceeded while stirring (120 rpm) in the reactor at 250 ◦C for 240 h. After the reaction, the blackish solution was diluted 5 times with water and heated at 100 ◦C to remove volatile components. Then, the solution was ﬁltered through an Advantec 5C ﬁlter paper and stored at 25 ◦C. 2.1. Preparation of Fluorescent GQD Suspensions GQD aqueous suspensions primarily containing blue-emitting GQDs (B-GQDs) and green- emitting GQDs (G-GQDs) were prepared with a bottom-up method from organic compounds [39]. Briefly, 20 g of ethanol, 3 g of 70% nitric acid and 1 g of acetylacetone were mixed and heated in a high-pressure reactor of 60 cm3. The reaction proceeded while stirring (120 rpm) in the reactor at 250 °C for 240 h. After the reaction, the blackish solution was diluted 5 times with water and heated at 100 °C to remo e olatile components Then the solution as filtered through an Ad antec 5C filter 2.2. Membrane Filtration Tests paper and stored at 25 °C. The permeate passing through the track-etched membrane was collected in the lower chamber while the retentate was fed back to the feed solution (Figure 1b). In the cross-ﬂow system, the majority of the feed ﬂow passed tangentially across the membrane surface. The concentrated GQD solutions (retentate) was sampled from the feed bottle and characterized after membrane ﬁltration. The membrane was also pre-wetted with DI water for 1 day prior to mounting in the module. The permeate ﬂux was determined using an electronic balance. DI water was circulated for 2 h in order to obtain a stable ﬂux before ﬁltering the GQDs. Before filtering GQDs, DI water was circulated for 20 min in order to obtain a non-fouled state permeable membrane. In circulating cross-flow filtration experiments, 20 μg/mL of the GQD feed solution was introduced into a track-etched membrane mounted in a customized module (Figure 1a) and filtered at the same conditions used in the dead-end filtration experiments. The crossflow module had an effective filtration area of 0.5 cm2 and was operated with a peristaltic pump (Cole-Parmer, IL, USA) for in order to maintain consistent solution inflow. The feed solution flowed into the upper chamber of the module with 15 mL/min flow rate. The operation pressure was 0.01 bar and was measured with a pressure gauge installed in the feed line. The permeate passing through the track-etched membrane was collected in the lower chamber while the retentate was fed back to the feed solution (Figure 1b). In the cross-flow system, the majority of the feed flow passed tangentially across the membrane surface. The concentrated GQD solutions (retentate) was sampled from the feed bottle and characterized after membrane filtration. The membrane was also pre-wetted with DI water for 1 day prior to mounting in the module. The permeate flux was determined using an electronic balance. DI water was circulated for 2 h in order to obtain a stable flux before filtering the GQDs. Figure 1. (a) Photographic image and (b) schematic drawing of the cross-flow filtration system used in this study. Figure 1. (a) Photographic image and (b) schematic drawing of the cross-ﬂow ﬁltration system used in this study. Figure 1. (a) Photographic image and (b) schematic drawing of the cross-flow filtration system used in this study. Figure 1. (a) Photographic image and (b) schematic drawing of the cross-ﬂow ﬁltration system used in this study. Figure 1. 2.2. Membrane Filtration Tests paper and stored at 25 °C. (a) Photographic image and (b) schematic drawing of the cross-flow filtration system used in this study. Figure 1. (a) Photographic image and (b) schematic drawing of the cross-ﬂow ﬁltration system used in this study The permeability (flux) of the GQD solutions filtered using the dead-end and cross-flow modes was determined from the filtrated volume per unit area and time (L m−2 h−1 bar−1) as follows (Equation (1)): The permeability (ﬂux) of the GQD solutions ﬁltered using the dead-end and cross-ﬂow modes was determined from the ﬁltrated volume per unit area and time (L m−2 h−1 bar−1) as follows (Equation (1)): J(L m−2 h−1 bar−1) = V A·t × 1 P = m A·t·ρ × 1 P (1) (1) 4 of 11 Nanomaterials 2018, 8, 959 where J is the ﬂux, V is the ﬁltrated volume, A is the effective area of the membrane, t is the operation time, P is the operation pressure, m is the mass, and ρ is the ﬂuid density. To investigate the surface morphology of membranes after GQDs ﬁltration, the membrane surfaces were examined using a ﬁeld-emission scanning electron microscope (FE-SEM, S-4700, Hitachi, Tokyo, Japan) at 10 kV working voltage. The membranes were coated with platinum using an ion sputter coater (E-1010, Hitachi, Tokyo, Japan) for 40 s. 2.3. Simulation of GQDs Transport through Nanochannels To determine the effective ﬁltration conditions, a transport model for nanoparticles through nanochannels was simulated using computational ﬂuid dynamics (CFD, COMSOL MultiphysicsTM, COMSOL INC, Burlington, MA, USA). The Navier–Stokes equation, continuity equation, Brownian force, and drag force were solved in the simulation. The Navier–Stokes equation (Equation (2)) and continuity equation (Equation (3)) for predicting laminar ﬂow describe the motion of a viscous ﬂuid under the assumption that mass is conserved [43]. These equations can be written as 0 = ∇·[−pI + µ(∇u + (∇u)T) −2 3µ(∇·u)I] + F (2) (2) and ∇·(ρu) = 0 ∇·(ρu) = 0 (3) (3) where p is the ﬂuid pressure, I is the unity tensor, u is the ﬂuid velocity, µ is the ﬂuid dynamic viscosity, and T indicates a transpose operation. The Brownian force (Equation (4)) and drag force (Equation (5)) for predicting particle motion describe the random motion of particles suspended in a ﬂuid and a force acting opposite to the relative motion of any object moving relative to the surrounding ﬂuid [44,45]. These equations can be written as F = ζ r 12πkBµTrp ∆t (4) F = 1 τp mp(u −v) (5) ese eq o s c e e s F = ζ r 12πkBµTrp ∆t (4) d F = 1 τp mp(u −v) (5) F = ζ r 12πkBµTrp ∆t (4) (4) and F = 1 τp mp(u −v) (5) (5) where ζ is a random number with zero mean, kB is Boltzman’s constant, rp is the particle radius, ∆t is the time step taken by the solver, τp is the particle velocity response time, mp is the particle mass, and v is the particle velocity. where ζ is a random number with zero mean, kB is Boltzman’s constant, rp is the particle radius, ∆t is the time step taken by the solver, τp is the particle velocity response time, mp is the particle mass, and v is the particle velocity. The designed space of the ﬁltration chamber in the CFD model was subdivided using the ﬁnite element method, which is a numerical technique for ﬁnding approximate solutions to boundary value problems for partial differential equations (Figure S1). Many contact points were set in and around the membrane for observing ﬂuid and particle motion in detail. 3. Results and Discussion where ܥ௙ଵ,௣ଵ are the c wavelength, and ܥ௙ଶ,௣ଶ ௙ଵ,௣ଵ wavelength, and ܥ௙ଶ,௣ଶ i i l h P Figure 2 shows the ﬂux of DI water and 20 µg/mL GQDs in aqueous solution through track-etched membranes with uniform cylindrical nanopores using dead-end and cross-ﬂow ﬁltration modes. Dead-end ﬁltration was operated at a constant operation pressure of 0.1 bar, which was controlled with nitrogen gas. The operating pressure in cross-ﬂow ﬁltration was controlled using the rotation speed of a peristaltic pump. In dead-end ﬁltration, the 10 mL of feed solution was completely forced through the membrane for 1 h. The two ﬁltration modes proceeded at a constant permeate ﬂux. The ﬂux was relatively stable in cross-ﬂow ﬁltration for a long period of time (12 h). The permeability of the GQDs solution by cross-ﬂow ﬁltration was two times higher than that operated by dead-end ﬁltration. This ﬂux decrease during dead-end ﬁltration might be due to partial blocking of pores during the initial ﬁltration process. emission wavelength. Peak deconvolutions were performed using Gaussian components. 3. Results and Discussion Figure 2 shows the flux of DI water and 20 μg/mL GQDs in aqueous solution through track- etched membranes with uniform cylindrical nanopores using dead-end and cross-flow filtration modes. Dead-end filtration was operated at a constant operation pressure of 0.1 bar, which was controlled with nitrogen gas. The operating pressure in cross-flow filtration was controlled using the rotation speed of a peristaltic pump. In dead-end filtration, the 10 mL of feed solution was completely forced through the membrane for 1 h. The two filtration modes proceeded at a constant permeate flux. The flux was relatively stable in cross-flow filtration for a long period of time (12 h). The permeability of the GQDs solution by cross-flow filtration was two times higher than that operated by dead-end filtration. This flux decrease during dead-end filtration might be due to partial blocking of pores during the initial filtration process. emission wavelength. Peak deconvolutions were performed using Gaussian components. 3. Results and Discussion Figure 2 shows the flux of DI water and 20 μg/mL GQDs in aqueous solution through track- etched membranes with uniform cylindrical nanopores using dead-end and cross-flow filtration modes. Dead-end filtration was operated at a constant operation pressure of 0.1 bar, which was controlled with nitrogen gas. The operating pressure in cross-flow filtration was controlled using the rotation speed of a peristaltic pump. In dead-end filtration, the 10 mL of feed solution was completely forced through the membrane for 1 h. 3. Results and Discussion where ܥ௙ଵ,௣ଵ are the c wavelength, and ܥ௙ଶ,௣ଶ ௙ଵ,௣ଵ wavelength, and ܥ௙ଶ,௣ଶ i i l h P The two filtration modes proceeded at a constant permeate flux. The flux was relatively stable in cross-flow filtration for a long period of time (12 h). The permeability of the GQDs solution by cross-flow filtration was two times higher than that operated by dead-end filtration. This flux decrease during dead-end filtration might be due to partial blocking Figure 2. Permeability tests of graphene quantum dot (GQD) solutions with track-etched membranes operated in the two different modes. (a) Dead-end filtration and (b) cross-flow filtration. Figure 2. Permeability tests of graphene quantum dot (GQD) solutions with track-etched membranes operated in the two different modes. (a) Dead-end ﬁltration and (b) cross-ﬂow ﬁltration. of pores during the initial filtration process. Figure 2. Permeability tests of graphene quantum dot (GQD) solutions with track-etched membranes p Figure 2. Permeability tests of graphene quantum dot (GQD) solutions with track-etched membranes operated in the two different modes. (a) Dead-end filtration and (b) cross-flow filtration. Figure 2. Permeability tests of graphene quantum dot (GQD) solutions with track-etched membranes operated in the two different modes. (a) Dead-end ﬁltration and (b) cross-ﬂow ﬁltration. Figure 2. Permeability tests of graphene quantum dot (GQD) solutions with track-etched membranes Figure 3 shows SEM images of the membrane surface at the feed side before and after filtration, respectively. Figure 3 shows SEM images of the membrane surface at the feed side before and afte ﬁltration, respectively. operated in the two different modes. (a) Dead end filtration and (b) cross flow filtration. Figure 3 shows SEM images of the membrane surface at the feed side before and after filtration, respectively. Figure 3. Scanning electron microscope (SEM) images of the top surface of track-etched membranes before and after filtration of GQD solutions in both modes. (a) Top surface image of the membrane before filtration. (b) Membrane surface at the feed side after dead-end filtration (left) and zoomed-in Figure 3. Scanning electron microscope (SEM) images of the top surface of track-etched membranes before and after filtration of GQD solutions in both modes. (a) Top surface image of the membrane before filtration. (b) Membrane surface at the feed side after dead-end filtration (left) and zoomed-in Figure 3. Scanning electron microscope (SEM) images of the top surface of track-etched membranes before and after ﬁltration of GQD solutions in both modes. (a) Top surface image of the membrane before ﬁltration. 2.4. Characterization of GQDs UV-Vis absorbance and photoluminescence of GQDs were measured to compare the separation efﬁciency between the two ﬁltration modes. UV-Vis absorbance in the GQDs dispersed in DI water was measured using a UV/VIS spectrophotometer (Optizen 2120UV, Mecasys, Daejeon, Korea). Absorbance was scanned from 200 nm to 600 nm in 1 nm increments. PL spectra from the GQD solutions were obtained using a ﬂuorescence spectrophotometer (FP-6300, JASCO, Tokyo, Japan). Excitation wavelengths were determined from the UV-Vis measurement results. To determine the ﬁltration selectivity between B-GQDs and G-GQDs, the peaks in the PL spectra were deconvoluted using Origin 5 of 11 Nanomaterials 2018, 8, 959 8.0. Based on the concentration of GQDs calculated from the ﬂuorescence intensity, the selectivity of the membrane was obtained in terms of the separation factor (α) as expressed in Equation (6): αp/ f = Cp2/Cp1 Cf2/Cf1 (6) 5 of 11 t i d i t f th ti f t ( ) d i 5 of 11 ained in terms of the separation factor (ߙ) as expressed in (6) where Cf1,p1 are the concentrate of solutes in the feed and permeate solutions at a ﬁrst emission wavelength, and Cf2,p2 are the concentrate of solutes in the feed and permeate solution at a second emission wavelength. Peak deconvolutions were performed using Gaussian components. selectivity of the membrane was obtained in terms of the separation factor (ߙ) as expressed in Equation (6): ߙ௣௙ ⁄ = ܥ௣ଶܥ௣ଵ ⁄ ܥ௙ଶܥ௙ଵ ⁄ (6) Equation (6): ߙ௣௙ ⁄ = ܥ௣ଶܥ௣ଵ ⁄ ܥ௙ଶܥ௙ଵ ⁄ (6) here ܥ are the concentrate of solutes in the feed and permeate solutions at a first emission 3. Results and Discussion where ܥ௙ଵ,௣ଵ are the c wavelength, and ܥ௙ଶ,௣ଶ ௙ଵ,௣ଵ wavelength, and ܥ௙ଶ,௣ଶ i i l h P This simulation result suggests that smaller particles have a higher permeate flux in the cross-flow filtration system, thus achieving a selective permeation depending on particle size. Figure 4. Pressure (left) and particle (right) trajectories simulated using computational fluid dynamics (CFD) simulations in the cross-flow filtration system. Figure 4. Pressure (left) and particle (right) trajectories simulated using computational ﬂuid dynamics (CFD) simulations in the cross-ﬂow ﬁltration system. Figure 4. Pressure (left) and particle (right) trajectories simulated using computational fluid dynamics (CFD) simulations in the cross-flow filtration system. Figure 4. Pressure (left) and particle (right) trajectories simulated using computational ﬂuid dynamics (CFD) simulations in the cross-ﬂow ﬁltration system. After filtration of the B-GQD/G-GQD mixture, we measured UV-Vis absorbance of permeates obtained from the two filtration modes in order to analyze the separation efficiency. Absorption peaks obtained from the feed and permeate in dead-end filtration were observed at 245, 268, and 368 nm with similar absorption intensities (Figure 5a). The solutions before (feed) and after (permeate) dead-end filtration exhibited a similar blue-green color under UV illumination at 365 nm, indicating filtration did not change the mole ratio of B-GQDs to G-GQDs. In contrast, cross-flow filtration produced a significant difference in the absorption spectra. While the peak position was nearly the same as in the feed and permeate, the absorption peak at 368 nm disappeared in the permeate solution (Figure 5b). A photograph taken of the permeate under UV illumination at 365 nm showed a distinct blue color, highlighting the increased mole fraction of B-GQDs in the B-GQD/G-GQD feed mixture. In the literature, GQDs with diameters ranging from 5 to 10 nm exhibited blue fluorescence under UV illumination at 365 nm [46]. From transmission electron microscopy (TEM) measurements, the average size of GQDs in retentate and permeate sides was estimated to be 5.4 ± 1.1 nm and 2.5 ± 0.3 nm, respectively (Figure S3). This result indicates relatively small GQDs selectively permeate through the nanoporous membrane during cross-flow filtration. After ﬁltration of the B-GQD/G-GQD mixture, we measured UV-Vis absorbance of permeates obtained from the two ﬁltration modes in order to analyze the separation efﬁciency. Absorption peaks obtained from the feed and permeate in dead-end ﬁltration were observed at 245, 268, and 368 nm with similar absorption intensities (Figure 5a). 3. Results and Discussion where ܥ௙ଵ,௣ଵ are the c wavelength, and ܥ௙ଶ,௣ଶ ௙ଵ,௣ଵ wavelength, and ܥ௙ଶ,௣ଶ i i l h P (b) Membrane surface at the feed side after dead-end ﬁltration (left) and zoomed-in image of the blocked region highlighted by dashed lines (right). (c) Membrane surface at the feed side after cross-ﬂow ﬁltration. before filtration. (b) Membrane surface at the feed side after dead-end filtration (left) and zoomed-in Figure 3. Scanning electron microscope (SEM) images of the top surface of track-etched membranes before and after filtration of GQD solutions in both modes. (a) Top surface image of the membrane before filtration. (b) Membrane surface at the feed side after dead-end filtration (left) and zoomed-in Figure 3. Scanning electron microscope (SEM) images of the top surface of track-etched membranes before and after ﬁltration of GQD solutions in both modes. (a) Top surface image of the membrane before ﬁltration. (b) Membrane surface at the feed side after dead-end ﬁltration (left) and zoomed-in image of the blocked region highlighted by dashed lines (right). (c) Membrane surface at the feed side after cross-ﬂow ﬁltration. 6 of 11 11 Nanomaterials 2018, 8, 959 Nanomaterials 2018, 8, x The track-etched membrane contains uniform cylindrical pores with ~ 50 nm diameter, 7 um thickness, and surface porosity of ~ 2% (Figure 3a and Figure S2). After dead-end ﬁltration of the GQD solution (10 mL for 1 h), the pores in the membrane were partially blocked and a cake layer was observed in the speciﬁc regions indicated by arrows in Figure 3b. In contrast, cross-ﬂow ﬁltration did not cause signiﬁcant fouling on the membrane surface, even at large ﬁltration quantity (30 mL) for 12 h (Figure 3c). As anticipated, we conﬁrmed the cross-ﬂow ﬁltration system is more suitable for ﬁltering large amounts while minimizing membrane fouling. image of the blocked region highlighted by dashed lines (right). (c) Membrane surface at the feed side after cross-flow filtration. The track-etched membrane contains uniform cylindrical pores with ~50 nm diameter, 7 um thickness, and surface porosity of ~2% (Figures 3a and Figure S2). After dead-end filtration of the GQD solution (10 mL for 1 h), the pores in the membrane were partially blocked and a cake layer was observed in the specific regions indicated by arrows in Figure 3b. 3. Results and Discussion where ܥ௙ଵ,௣ଵ are the c wavelength, and ܥ௙ଶ,௣ଶ ௙ଵ,௣ଵ wavelength, and ܥ௙ଶ,௣ଶ i i l h P In contrast, cross-flow filtration did not cause significant fouling on the membrane surface, even at large filtration quantity (30 mL) for 12 h (Figure 3c) As anticipated we confirmed the cross flow filtration system is more suitable for A simulation was conducted to determine the ﬂuid pressure and particle trajectories between the upper and lower chambers in the cross-ﬂow ﬁltration system. The diameter of model particles was set to 10 nm and 20 nm to estimate the permeation rate of GQDs through the nanochannel. Given the membrane ﬁltration conditions, the cylindrical pore size in the membrane and ﬂow rate were set to 50 nm and 15 mL/min, respectively. The pressure difference between the upper and lower chambers was estimated to be approximately 5 mbar. Considering calculating the simulation time, the number of particles was set to 5000 at a ratio of 1:1. The ﬂow of GQDs was simulated in a laminar condition with a mean velocity of 1.51 × 10−3 m/s (Figure 4). As the particles pass through the nanopores, the ratio between 10 nm and 20 nm particles is changed from 0.97 to 1.22. This simulation result suggests that smaller particles have a higher permeate ﬂux in the cross-ﬂow ﬁltration system, thus achieving a selective permeation depending on particle size. h (Figure 3c). As anticipated, we confirmed the cross-flow filtration system is more suitable for filtering large amounts while minimizing membrane fouling. A simulation was conducted to determine the fluid pressure and particle trajectories between the upper and lower chambers in the cross-flow filtration system. The diameter of model particles was set to 10 nm and 20 nm to estimate the permeation rate of GQDs through the nanochannel. Given the membrane filtration conditions, the cylindrical pore size in the membrane and flow rate were set to 50 nm and 15 mL/min, respectively. The pressure difference between the upper and lower chambers was estimated to be approximately 5 mbar. Considering calculating the simulation time, the number of particles was set to 5000 at a ratio of 1:1. The flow of GQDs was simulated in a laminar condition with a mean velocity of 1.51 × 10−3 m/s (Figure 4). As the particles pass through the nanopores, the ratio between 10 nm and 20 nm particles is changed from 0.97 to 1.22. 3. Results and Discussion where ܥ௙ଵ,௣ଵ are the c wavelength, and ܥ௙ଶ,௣ଶ ௙ଵ,௣ଵ wavelength, and ܥ௙ଶ,௣ଶ i i l h P (a) Dead-end filtration and (b) cross-flow filtration (Inset: Photographs of the GQD solutions taken We also measured PL spectra to confirm the separation efficiency. A detailed PL study was conducted using excitation wavelengths of 245, 268 and 368 nm based on the absorbance results to further explore the optical properties of the GQD solutions [47]. Figure 6 shows the PL spectra of GQDs obtained after dead-end and cross-flow filtration, respectively. The GQDs used in this study were strongly excited at 368 nm, resulting in generation of two emission peaks at 420 nm and 520 nm. The feed and permeate solutions obtained from dead-end filtration exhibited similar PL spectra. However, the peak intensity at 520 nm from the permeate solution produced by cross-flow filtration was dramatically reduced with increasing concentration of small B-GQDs in the permeate solution. We also measured PL spectra to conﬁrm the separation efﬁciency. A detailed PL study was conducted using excitation wavelengths of 245, 268 and 368 nm based on the absorbance results to further explore the optical properties of the GQD solutions [47]. Figure 6 shows the PL spectra of GQDs obtained after dead-end and cross-ﬂow ﬁltration, respectively. The GQDs used in this study were strongly excited at 368 nm, resulting in generation of two emission peaks at 420 nm and 520 nm. The feed and permeate solutions obtained from dead-end ﬁltration exhibited similar PL spectra. However, the peak intensity at 520 nm from the permeate solution produced by cross-ﬂow ﬁltration was dramatically reduced with increasing concentration of small B-GQDs in the permeate solution. under illumination with 365 nm light). We also measured PL spectra to confirm the separation efficiency. A detailed PL study was conducted using excitation wavelengths of 245, 268 and 368 nm based on the absorbance results to further explore the optical properties of the GQD solutions [47]. Figure 6 shows the PL spectra of GQDs obtained after dead-end and cross-flow filtration, respectively. The GQDs used in this study were strongly excited at 368 nm, resulting in generation of two emission peaks at 420 nm and 520 nm. The feed and permeate solutions obtained from dead-end filtration exhibited similar PL spectra. However, the peak intensity at 520 nm from the permeate solution produced by cross-flow filtration was dramatically reduced with increasing concentration of small B-GQDs in the permeate solution. Figure 6. Photoluminescence (PL) spectra from GQD solutions obtained after two filtration modes. 3. Results and Discussion where ܥ௙ଵ,௣ଵ are the c wavelength, and ܥ௙ଶ,௣ଶ ௙ଵ,௣ଵ wavelength, and ܥ௙ଶ,௣ଶ i i l h P The solutions before (feed) and after (permeate) dead-end ﬁltration exhibited a similar blue-green color under UV illumination at 365 nm, indicating ﬁltration did not change the mole ratio of B-GQDs to G-GQDs. In contrast, cross-ﬂow ﬁltration produced a signiﬁcant difference in the absorption spectra. While the peak position was nearly the same as in the feed and permeate, the absorption peak at 368 nm disappeared in the permeate solution (Figure 5b). A photograph taken of the permeate under UV illumination at 365 nm showed a distinct blue color, highlighting the increased mole fraction of B-GQDs in the B-GQD/G-GQD feed mixture. In the literature, GQDs with diameters ranging from 5 to 10 nm exhibited blue ﬂuorescence under UV illumination at 365 nm [46]. From transmission electron microscopy (TEM) measurements, the average size of GQDs in retentate and permeate sides was estimated to be 5.4 ± 1.1 nm and 2.5 ± 0.3 nm, respectively (Figure S3). This result indicates relatively small GQDs selectively permeate through the nanoporous membrane during cross-ﬂow ﬁltration. 7 of 11 7 f 11 7 of 11 7 f 11 Nanomaterials 2018, 8, 959 7 of 11 Figure 5. Absorbance spectra of GQD solutions before and after applying the two filtration modes. (a) Dead-end filtration and (b) cross-flow filtration (Inset: Photographs of the GQD solutions taken under illumination with 365 nm light). Figure 5. Absorbance spectra of GQD solutions before and after applying the two ﬁltration modes. (a) Dead-end ﬁltration and (b) cross-ﬂow ﬁltration (Inset: Photographs of the GQD solutions taken under illumination with 365 nm light). Nanomaterials 2018, 8, x FOR PEER REVIEW 7 of 11 Figure 5. Absorbance spectra of GQD solutions before and after applying the two filtration modes. (a) Dead-end filtration and (b) cross-flow filtration (Inset: Photographs of the GQD solutions taken Nanomaterials 2018, 8, x FOR PEER REVIEW Figure 5. Absorbance spectra of GQD solutions before and after applying the two filtration modes. (a) Dead-end filtration and (b) cross-flow filtration (Inset: Photographs of the GQD solutions taken under illumination with 365 nm light). Figure 5. Absorbance spectra of GQD solutions before and after applying the two ﬁltration modes. (a) Dead-end ﬁltration and (b) cross-ﬂow ﬁltration (Inset: Photographs of the GQD solutions taken under illumination with 365 nm light). Figure 5. Absorbance spectra of GQD solutions before and after applying the two filtration modes. 3. Results and Discussion where ܥ௙ଵ,௣ଵ are the c wavelength, and ܥ௙ଶ,௣ଶ ௙ଵ,௣ଵ wavelength, and ܥ௙ଶ,௣ଶ i i l h P (a,b) PL spectra from (a) feed and (b) permeate solutions in dead-end filtration. (c–e) PL spectra from Figure 6. Photoluminescence (PL) spectra from GQD solutions obtained after two filtration modes. (a,b) PL spectra from (a) feed and (b) permeate solutions in dead-end filtration. (c–e) PL spectra from (c) feed, (d) retentate, and (e) permeate solutions in cross-flow filtration. The feed solutions used in two filtration modes was prepared by the same synthetic methods of GQDs. Figure 6. Photoluminescence (PL) spectra from GQD solutions obtained after two ﬁltration modes. (a,b) PL spectra from (a) feed and (b) permeate solutions in dead-end ﬁltration. (c–e) PL spectra from (c) feed, (d) retentate, and (e) permeate solutions in cross-ﬂow ﬁltration. The feed solutions used in two ﬁltration modes was prepared by the same synthetic methods of GQDs. Figure 6. Photoluminescence (PL) spectra from GQD solutions obtained after two filtration modes. (a,b) PL spectra from (a) feed and (b) permeate solutions in dead-end filtration. (c–e) PL spectra from Figure 6. Photoluminescence (PL) spectra from GQD solutions obtained after two filtration modes. (a,b) PL spectra from (a) feed and (b) permeate solutions in dead-end filtration. (c–e) PL spectra from (c) feed, (d) retentate, and (e) permeate solutions in cross-flow filtration. The feed solutions used in two filtration modes was prepared by the same synthetic methods of GQDs. Figure 6. Photoluminescence (PL) spectra from GQD solutions obtained after two ﬁltration modes. (a,b) PL spectra from (a) feed and (b) permeate solutions in dead-end ﬁltration. (c–e) PL spectra from (c) feed, (d) retentate, and (e) permeate solutions in cross-ﬂow ﬁltration. The feed solutions used in two ﬁltration modes was prepared by the same synthetic methods of GQDs. Figure 6. Photoluminescence (PL) spectra from GQD solutions obtained after two filtration modes. (a,b) PL spectra from (a) feed and (b) permeate solutions in dead-end filtration. (c–e) PL spectra from Figure 6. Photoluminescence (PL) spectra from GQD solutions obtained after two filtration modes. (a,b) PL spectra from (a) feed and (b) permeate solutions in dead-end filtration. (c–e) PL spectra from (c) feed, (d) retentate, and (e) permeate solutions in cross-flow filtration. The feed solutions used in two filtration modes was prepared by the same synthetic methods of GQDs. Figure 6. Photoluminescence (PL) spectra from GQD solutions obtained after two ﬁltration modes. (a,b) PL spectra from (a) feed and (b) permeate solutions in dead-end ﬁltration. 3. Results and Discussion where ܥ௙ଵ,௣ଵ are the c wavelength, and ܥ௙ଶ,௣ଶ ௙ଵ,௣ଵ wavelength, and ܥ௙ଶ,௣ଶ i i l h P (c–e) PL spectra from (c) feed, (d) retentate, and (e) permeate solutions in cross-ﬂow ﬁltration. The feed solutions used in two ﬁltration modes was prepared by the same synthetic methods of GQDs. (c) feed, (d) retentate, and (e) permeate solutions in cross-flow filtration. The feed solutions used in two filtration modes was prepared by the same synthetic methods of GQDs. After cross-flow filtration, we observed the absorption of GQDs in the interior of circulating b h f l d h b d l h h h d d h After cross-flow filtration, we observed the absorption of GQDs in the interior of circulating tubes in the filtration system and at the membrane module into which the water was introduced. This physical adsorption might be occurred by relatively large GQD with high surface area, resulting in After cross-ﬂow ﬁltration, we observed the absorption of GQDs in the interior of circulating tubes in the ﬁltration system and at the membrane module into which the water was introduced. This physical adsorption might be occurred by relatively large GQD with high surface area, resulting Nanomaterials 2018, 8, 959 Nanomaterials 2018, 8, x FO 8 of 11 8 of 11 Nanomaterials 2018, 8, 959 Nanomaterials 2018, 8, x FO in the decrease of PL intensity at 520 nm in the retentate solution (Figure 6d). Since the optical properties of GQDs can be inﬂuenced by multiple factors such as chemical functionality and edge-state of GQDs as well as size, more detailed investigation is needed to characterize the size-dependent optical properties of GQDs. the decrease of PL intensity at 520 nm in the retentate solution (Figure 6d). Since the optical properties of GQDs can be influenced by multiple factors such as chemical functionality and edge-state of GQDs as well as size, more detailed investigation is needed to characterize the size-dependent optical properties of GQDs. For more detailed analysis, we performed deconvolution of the PL spectra at 368 nm excitation p p p For more detailed analysis, we performed deconvolution of the PL spectra at 368 nm excitation to separate the contribution of the two emission peaks (420 nm and 520 nm) and calculate the filtration selectivity (Figure 7). Clear blue PL in the permeate side was attributed to selective permeation of relatively small B-GQDs. 3. Results and Discussion where ܥ௙ଵ,௣ଵ are the c wavelength, and ܥ௙ଶ,௣ଶ ௙ଵ,௣ଵ wavelength, and ܥ௙ଶ,௣ଶ i i l h P We estimate that the selectivity in terms of separation factor (B-GQDs/G-GQDs) between B-GQDs and G-GQDs was 0.74, whereas the separation factor from dead-end filtration was 0.04, meaning B-GQDs were enriched in the permeate side. From these results, we confirmed that cross-flow filtration is a more efficient method than dead-end filtration for separating GQDs by size. y , p p to separate the contribution of the two emission peaks (420 nm and 520 nm) and calculate the filtration selectivity (Figure 7). Clear blue PL in the permeate side was attributed to selective permeation of relatively small B-GQDs. We estimate that the selectivity in terms of separation factor (B-GQDs/G- GQDs) between B-GQDs and G-GQDs was 0.74, whereas the separation factor from dead-end filtration was 0.04, meaning B-GQDs were enriched in the permeate side. From these results, we confirmed that cross-flow filtration is a more efficient method than dead-end filtration for separating GQDs by size. Figure 7. Peak deconvolution in photoluminescence spectra excited at 368 nm. (a,b) PL spectra from (a) feed and (b) permeate solutions in dead-end filtration. (c,d) PL spectra from (c) feed and (d) permeate solutions in cross-flow filtration. Figure 7. Peak deconvolution in photoluminescence spectra excited at 368 nm. (a,b) PL spectra from (a) feed and (b) permeate solutions in dead-end ﬁltration. (c,d) PL spectra from (c) feed and (d) permeate solutions in cross-ﬂow ﬁltration. Figure 7. Peak deconvolution in photoluminescence spectra excited at 368 nm. (a,b) PL spectra from (a) feed and (b) permeate solutions in dead-end filtration. (c,d) PL spectra from (c) feed and (d) permeate solutions in cross-flow filtration. Figure 7. Peak deconvolution in photoluminescence spectra excited at 368 nm. (a,b) PL spectra from (a) feed and (b) permeate solutions in dead-end ﬁltration. (c,d) PL spectra from (c) feed and (d) permeate solutions in cross-ﬂow ﬁltration. References 1. Baker, S.N.; Baker, G.A. Luminescent carbon nanodots: Emergent nanolights. Angew. Chem. Int. Ed. 2010, 49, 6726–6744. [CrossRef] [PubMed] 2. Dutta, M.; Sarkar, S.; Ghosh, T.; Basak, D. ZnO/Graphene quantum dot solid-state solar cell. J. Phys. Chem. C 2012, 116, 20127–20131. [CrossRef] 3. Zheng, X.T.; Ananthanarayanan, A.; Luo, K.Q.; Chen, P. Glowing graphene quantum dots and carbon Dots: properties, syntheses, and biological applications. Small 2015, 11, 1620–1636. [CrossRef] [PubMed] 3. Zheng, X.T.; Ananthanarayanan, A.; Luo, K.Q.; Chen, P. Glowing graphene quantum dots and carbon Dots: properties, syntheses, and biological applications. Small 2015, 11, 1620–1636. [CrossRef] [PubMed] 4. Zhu, S.; Tang, S.; Zhang, J.; Yang, B. Control the size and surface chemistry of graphene for the rising ﬂuorescent materials. Chem. Commun. 2012, 48, 4527–4539. [CrossRef] [PubMed] 4. Zhu, S.; Tang, S.; Zhang, J.; Yang, B. Control the size and surface chemistry of graphene for the rising ﬂuorescent materials. Chem. Commun. 2012, 48, 4527–4539. [CrossRef] [PubMed] 5. Sk, M.A.; Ananthanarayanan, A.; Huang, L.; Lim, K.H.; Chen, P. Revealing the tunable photolumin properties of graphene quantum dots. J. Mater. Chem. C 2014, 2, 6954–6960. [CrossRef] 6. Kim, S.; Shin, D.H.; Kim, C.O.; Kang, S.S.; Kim, J.M.; Choi, S.H.; Jin, L.-H.; Cho, Y.-H.; Hwang, S.W.; Sone, C. Size-dependent radiative decay processes in graphene quantum dots. Appl. Phys. Lett. 2012, 101, 163103. [CrossRef] 6. Kim, S.; Shin, D.H.; Kim, C.O.; Kang, S.S.; Kim, J.M.; Choi, S.H.; Jin, L.-H.; Cho, Y.-H.; Hwang, S.W.; Sone, C. Size-dependent radiative decay processes in graphene quantum dots. Appl. Phys. Lett. 2012, 101, 163103. [CrossRef] 7. Eda, G.; Lin, Y.-Y.; Mattevi, C.; Yamaguchi, H.; Chen, H.-A.; Chen, I.S.; Chen, C.-W.; Chhowalla, M. Blue photoluminescence from chemically derived graphene oxide. Adv. Mater. 2010, 22, 505–509. [CrossRef] [PubMed] 8. Shen, J.; Zhu, Y.; Yang, X.; Li, C. Graphene quantum dots: Emergent nanolights for bioimaging, sensors, catalysis and photovoltaic devices. Chem. Commun. 2012, 48, 3686–3699. [CrossRef] [PubMed] 9. Jin, S.H.; Kim, D.H.; Jun, G.H.; Hong, S.H.; Jeon, S. Tuning the photoluminescence of graphene quantu through the charge transfer effect of functional groups. ACS Nano 2013, 7, 1239–1245. [CrossRef] [Pu 9. Jin, S.H.; Kim, D.H.; Jun, G.H.; Hong, S.H.; Jeon, S. Tuning the photoluminescence of graphene quantum dots through the charge transfer effect of functional groups. ACS Nano 2013, 7, 1239–1245. [CrossRef] [PubMed] 10. Zhang, Z.; Zhang, J.; Chen, N.; Qu, L. Graphene quantum dots: An emerging material for energy-related applications and beyond. Energ. Environ. 4. Conclusions 4. Conclusions and Y.-E.K.; Data curation, Y.J.K., B.K.M. and E.S.J.; Writing—original draft, S.-G.Y.; Writing—review & editing, S.-G.Y., Y.J.K., E.S.J. and S.Y.Y.; Supervision, S.Y.Y. Funding: This research was supported by a grant from the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health & Welfare, Republic of Korea (HI16C0313). Funding: This research was supported by a grant from the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health & Welfare, Republic of Korea (HI16C0313). nowledgments: The work was supported by the Ministry of Health & Welfare, Republic of Korea (HI16C0313) Acknowledgments: The work was supported by the Ministry of Health & Welfare, Republic of Korea (HI16C0313). Conﬂicts of Interest: The authors declare no conﬂict of interest. Conﬂicts of Interest: The authors declare no conﬂict of interest. Conﬂicts of Interest: The authors declare no conﬂict of interest. 4. Conclusions 4. Conclusions We first presented a method for separating two-dimensional GQDs using cross-flow membrane filtration with uniform transport nanochannels. A cross-flow membrane filtration system was found to be effective for fractionation of fluorescent GQDs by selective permeation through a membrane, exhibiting 18-fold higher selectivity and 2-fold higher permeability compared to a dead-end filtration system. In addition, cross-flow filtration provides stable flux without significant fouling or flux decline during filtration. Since the membrane filtration process is simple, cost-effective, and scalable, our approach can be used to obtain a large amount of uniform-sized GQDs required for high performance opto-electronic and bio-imaging applications. We ﬁrst presented a method for separating two-dimensional GQDs using cross-ﬂow membrane ﬁltration with uniform transport nanochannels. A cross-ﬂow membrane ﬁltration system was found to be effective for fractionation of ﬂuorescent GQDs by selective permeation through a membrane, exhibiting 18-fold higher selectivity and 2-fold higher permeability compared to a dead-end ﬁltration system. In addition, cross-ﬂow ﬁltration provides stable ﬂux without signiﬁcant fouling or ﬂux decline during ﬁltration. Since the membrane ﬁltration process is simple, cost-effective, and scalable, our approach can be used to obtain a large amount of uniform-sized GQDs required for high performance opto-electronic and bio-imaging applications. Supplementary Materials: The following are available online at www.mdpi.com/xxx/s1, Figure S1: Space subdivided by the membrane with nanochannels used in finite element method simulations; Figure S2: SEM images of the cross sectional image of the membrane before filtration; Figure S3: TEM images of GQDs in (a) Supplementary Materials: The following are available online at http://www.mdpi.com/2079-4991/8/11/959/s1, Figure S1: Space subdivided by the membrane with nanochannels used in ﬁnite element method simulations; Figure S2: SEM images of the cross-sectional image of the membrane before ﬁltration; Figure S3: TEM images of 9 of 11 Nanomaterials 2018, 8, 959 GQDs in (a) retentate and (b) permeate sides obtained after cross-ﬂow ﬁltration. The average size is deﬁned as half of (width (W) + length (L)) of a GQD as shown in the inset of (a). GQDs in (a) retentate and (b) permeate sides obtained after cross-ﬂow ﬁltration. The average size is deﬁned as half of (width (W) + length (L)) of a GQD as shown in the inset of (a). Author Contributions: Conceptualization, S.-G.Y. and S.Y.Y.; Methodology, S.-G.Y. and Y.-E.K.; Data curation, Y.J.K., B.K.M. and E.S.J.; Writing—original draft, S.-G.Y.; Writing—review & editing, S.-G.Y., Y.J.K., E.S.J. and S.Y.Y.; Supervision, S.Y.Y. Author Contributions: Conceptualization, S.-G.Y. and S.Y.Y.; Methodology, S.-G.Y. References Sci. 2012, 5, 8869–8890. [CrossRef] 9. Jin, S.H.; Kim, D.H.; Jun, G.H.; Hong, S.H.; Jeon, S. Tuning the photoluminescence of graphene quantum dots through the charge transfer effect of functional groups. ACS Nano 2013, 7, 1239–1245. [CrossRef] [PubMed] 10 Zhang Z ; Zhang J ; Chen N ; Qu L Graphene quantum dots: An emerging material for energy-related 10. Zhang, Z.; Zhang, J.; Chen, N.; Qu, L. Graphene quantum dots: An emerging material for energy-related applications and beyond. Energ. Environ. Sci. 2012, 5, 8869–8890. [CrossRef] 11. Roy, P.; Ravindranath, R.; Periasamy, A.P.; Lien, C.-W.; Liang, C.-T.; Chang, H.-T. Green synthesis of Si–GQD nanocomposites as cost-effective catalysts for oxygen reduction reaction. RSC Adv. 2016, 6, 108941–108947. [CrossRef] 12. Shin, Y.; Park, J.; Hyun, D.; Yang, J.; Lee, H. Generation of graphene quantum dots by the oxidative cleavage of graphene oxide using the oxone oxidant. New J. Chem. 2015, 39, 2425–2428. [CrossRef] 13. Bacon, M.; Bradley, S.J.; Nann, T. Graphene quantum dots. Part. Part. Syst. Char. 2014, 31, 415–428. [Cross 14. Liu, G.; Zhang, K.; Ma, K.; Care, A.; Hutchinsond, M.R.; Goldysa, E.M. Graphene quantum dot based “switch-on” nanosensors for intracellular cytokine monitoring. Nanoscale 2017, 9, 4934–4943. [CrossRef] [PubMed] 15. Nurunnabi, M.; Khatun, Z.; Naﬁujjaman, M.; Lee, D.-g.; Lee, Y.-K. Surface coating of graphene quantum dots using mussel-inspired polydopamine for biomedical optical imaging. ACS Appl. Mater. interfaces 2013, 5, 8246–8253. [CrossRef] [PubMed] 16. Smith, A.M.; Duan, H.; Mohs, A.M.; Nie, S. Bioconjugated quantum dots for in vivo molecular and cellular imaging. Adv. Drug Deliv. Rev. 2008, 60, 1226–1240. [CrossRef] [PubMed] 17. Hwang, E.; Hwang, H.M.; Shin, Y.; Yoon, Y.; Lee, H.; Yang, J.; Bak, S.; Lee, H. Chemically modulated graphene quantum dot for tuning the photoluminescence as novel sensory probe. Sci. Rep. 2016, 6, 39448. [CrossRef] [PubMed] 10 of 11 Nanomaterials 2018, 8, 959 18. Bae, P.K.; Kim, K.N.; Lee, S.J.; Chang, H.J.; Lee, C.K.; Park, J.K. The modiﬁcation of quantum dot probes used for the targeted imaging of his-tagged fusion proteins. Biomaterials 2009, 30, 836–842. [CrossRef] [PubMed] 19. Tiwari, D.K.; Jin, T.; Behari, J. Bio-distribution and toxicity assessment of intravenously injected anti-HER2 antibody conjugated CdSe/ZnS quantum dots in Wistar rats. Int. J. Nanomedicine 2011, 6, 463–475. [CrossRef] [PubMed] 20. Zheng, X.T.; Than, A.; Ananthanaraya, A.; Kim, D.-H.; Chen, P. Graphene quantum dots as universal ﬂuorophores and their use in revealing regulated trafﬁcking of insulin receptors in adipocytes. References ACS Nano 2013, 7, 6278–6286. [CrossRef] [PubMed] 21. Dong, Y.; Pang, H.; Yang, H.B.; Guo, C.; Shao, J.; Chi, Y.; Li, C.M.; Yu, T. Carbon-based dots Co-doped with nitrogen and sulfur for high quantum yield and excitation-independent emission. Angew. Chem. Int. Ed. 2013, 52, 7800–7804. [CrossRef] [PubMed] 22. Nurunnabi, M.; Khatun, Z.; Huh, K.M.; Park, S.Y.; Lee, D.Y.; Cho, K.J.; Lee, Y.-K. In vivo biodistribution and toxicology of carboxylated graphene quantum dots. ACS Nano 2013, 7, 6858–6867. [CrossRef] [PubMed] 23. Iannazzo, D.; Ziccarelli, I.; Pistone, A. Graphene quantum dots: Multifunctional nanoplatforms for anticancer therapy. J. Mater. Chem. B 2017, 5, 6471–6489. [CrossRef] 24. Zhang, J.; Yu, S.-H. Carbon dots: Large-scale synthesis, sensing and bioimaging. Mater. Today 2015, 19, 382–393. [CrossRef] 25. Coleman, J.N.; Lotya, M.; O’Neill, A.; Bergin, S.D.; King, P.J.; Khan, U.; Young, K.; Gaucher, A.; De, S.; Smith, R.J.; et al. Two-dimensional nanosheets produced by liquid exfoliation of layered materials. Science 2011, 331, 568–571. [CrossRef] [PubMed] 26. Nicolosi, V.; Chhowalla, M.; Kanatzidis, M.G.; Strano, M.S.; Coleman, J.N. Liquid exfoliation of layered materials. Science 2013, 340, 1420. [CrossRef] 27. Xue, Q.; Huang, H.; Wang, L.; Chen, Z.; Wu, M.; Li, Z.; Pan, D. Nearly monodisperse graphene quantum dots fabricated by amine-assisted cutting and ultraﬁltration. Nanoscale 2013, 5, 12098–12103. [CrossRef] [PubMed] 28. Koutsioukis, A.; Akouros, A.; Zboril, R.; Georgakilas, V. Solid phase extraction for the puriﬁcation of violet, blue, green and yellow emitting carbon dots. Nanoscale 2018, 10, 11293–11296. [CrossRef] [PubMed] 29. Sun, Y.; Wang, S.; Li, C.; Luo, P.; Tao, L.; Wei, Y.; Shi, G. Large scale preparation of graphene quantum dots from graphite with tunable ﬂuorescence properties. Phys. Chem. Chem. Phys. 2013, 15, 9907–9913. [CrossRef] [PubMed] 30. Bucher, T.; Filiz, V.; Abetz, C.; Abetz, V. Formation of thin, isoporous block copolymer membranes by an upscalable proﬁle roller coating process—A promising way to save block copolymer. Membranes 2018, 8, 57. [CrossRef] [PubMed] 31. Belfort, G.; Davis, R.H.; Zydney, A.L. The behavior of suspensions and macromolecular solutions in crossﬂow microﬁltration. J. Membr. Sci. 1994, 96, 1–58. [CrossRef] 32. Decarolis, J.; Hong, S.; Taylor, J. Fouling behavior of a pilot scale inside-out hollow ﬁber UF membrane during dead-end ﬁltration of tertiary wastewater. J. Membr. Sci. 2001, 191, 165–178. [CrossRef] 33. Van Zwieten, R.; van de Laar, T.; Sprakel, J.; Schroën, K. From cooperative to uncorrelated clogging in cross-ﬂow microﬂuidic membranes. Sci. Rep. 2018, 8, 5687. [CrossRef] [PubMed] 34. References Yeh, C.-H.; Hung, C.-W.; Wu, C.-H.; Lin, Y.-C. Using the developed cross-ﬂow ﬁltration chip for collecting blood plasma under high ﬂow rate condition and applying the immunoglobulin E detection. J. Micromech. Microeng. 2014, 24, 095013. [CrossRef] g 35. Pradel, K.C.; Sohn, K. Huang, cross-ﬂow puriﬁcation of nanowires. Angew. Chem. Int. Ed. 2011, 50, 3412–3416. [CrossRef] [PubMed] 36. Dorney, K.M.; Baker, J.D.; Edwards, M.L.; Kanel, S.R.; O’Malley, M.; Sizemore, I.E.P. Tangential ﬂow ﬁltration of colloidal silver nanoparticles: A “Green” laboratory Experiment for chemistry and engineering students. J. Chem. Educ. 2014, 91, 1044–1049. [CrossRef] 37. Koo, C.H.; Mohammad, A.W.; Suja, F. Effect of cross-ﬂow velocity on membrane ﬁltration performance in relation to membrane properties. Desalin. Water Treat. 2014, 55, 678–692. [CrossRef] 38. Ohmori, S.; Saito, T.; Shukla, B.; Yumura, M.; Iijima, S. Fractionation of single wall carbon nanotubes by length using cross ﬂow ﬁltration method. ACS Nano 2010, 4, 3606–3610. [CrossRef] [PubMed] 39. Dong, Y.; Shao, J.; Chen, C.; Li, H.; Wang, R.; Chi, Y.; Lin, X.; Chen, G. Blue luminescent graphene quantum dots and graphene oxide prepared by tuning the carbonization degree of citric acid. Carbon 2012, 50, 4738–4743. [CrossRef] 11 of 11 Nanomaterials 2018, 8, 959 40. Nguyen, Q.H.; Ali1, M.; Nasir, S.; Ensinger, W. Transport properties of track-etched membranes having variable effective pore-lengths. Nanotechnology 2015, 26, 485502. [CrossRef] [PubMed] 41. Hester, J.; Mayes, A. Design and performance of foul-resistant poly(vinylidene ﬂuoride) membranes prepared in a single-step by surface segregation. J. Membr. Sci. 2002, 202, 119–135. [CrossRef] 42. Yang, S.Y.; Ryu, I.; Kim, H.Y.; Kim, J.K.; Jang, S.K.; Russell, T.P. Nanoporous membranes with ultrahigh selectivity and ﬂux for the ﬁltration of viruses. Adv. Mater. 2006, 18, 709–712. [CrossRef] 43. Kim, J.; Moin, P. Application of a fractional-step method to incompressible Navier-Stokes equations. J. Comput. Phys. 1985, 59, 308–323. [CrossRef] p y 44. Cook, H. Brownian motion in spinodal decomposition. Acta Metall. 1970, 18, 297–306. [CrossRef] 45. Bishop, R.E.D.; Hassan, A.Y. The lift and drag forces on a circular cylinder oscillating in a ﬂowing ﬂuid. Proc. R. Soc. A 1964, 277, 51–75. [CrossRef] 46. Kim, S.; Hwang, S.W.; Kim, M.-K.; Shin, D.Y.; Shin, D.H.; Kim, C.O.; Yang, S.B.; Park, J.H.; Hwang, E.; Choi, S.-H. Anomalous behaviors of visible luminescence from graphene quantum dots: Interplay between size and shape. ACS Nano 2012, 6, 8203–8208. [CrossRef] [PubMed] 47. © 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/). References Liu, J.; Rojas-Andrade, M.D.; Chata, G.; Peng, Y.; Roseman, G.; Lu, J.-E.; Millhauser, G.L.; Saltikov, C.; Chen, S. Photo-enhanced antibacterial activity of ZnO/graphene quantum dot nanocomposites. Nanoscale 2018, 10, 158–166. [CrossRef] [PubMed] © 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/). © 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
https://openalex.org/W4383864724	https://zenodo.org/records/8133634/files/ilm-fan%201707.pdf	Tagalog	null	TASVIRIY SAN'AT SIRLARI! CHIZMACHILIK FANINI O'QITISHDA YANGILIKLAR!	Zenodo (CERN European Organization for Nuclear Research)	2,023	cc-by	1,461	Uning asosiy g'oyasi shundan iboratki, amaliyot nazariy bilimlar singari muhimdir. Ya'ni, o'rganish paytida biz nafaqat miyyamiz bilan, balki qo'limiz bilan ham ishlashimiz kerak. Faqat sinf devorlarida o'rganish tez o'zgaruvchan dunyo bilan hamqadam emas. STEAM yondashuvining asosiy farqi shundaki, bolalar turli xil mavzularni muvaffaqiyatli o'rganish uchun ham miyani, ham qo'llarini ishlatadilar. Ular olgan bilimlarni o'zlari «uqib oladilar». STEAM ta'lim muhitida bolalar bilimga ega bo'ladilar va darhol undan foydalanishni o'rganadilar. Shuning uchun, ular o'sib ulg'ayganlarida va hayotiy muammolarga duch kelganda, atrof-muhitning ifloslanishi yoki global iqlim o'zgarishi bo'ladimi, bunday murakkab masalalarni faqat turli sohalardagi bilimlarga tayanib va birgalikda ishlash orqali hal qilish mumkinligini tushunadilar. Bu yerda faqat bitta mavzu bo'yicha bilimga tayanish yetarli emas. STEAM yondashuvi bizning ta'lim va ta'limga bo'lgan qarashimizni o'zgartirmoqda. Amaliy qobiliyatga e'tibor berib, talabalar o'zlarining irodasini, ijodkorligini, moslashuvchanligini rivojlantiradi va boshqalar bilan hamkorlik qilishni o'rganadi. Ushbu ko'nikmalar va bilimlar asosiy ta'lim vazifasini tashkil etadi, ya'ni, bu butun ta'lim tizimi nimaga intilishini belgilaydi. Ta'limga ushbu yangi yondashuv qanday paydo bo'ldi? Bu nazariya va amaliyotni birlashtirishning mantiqiy natijasidir. STEAM Amerikada ishlab chiqilgan. Ba'zi maktablar bitiruvchilarning martabalarini e'tiborga olishdi va fan, texnologiya, muhandislik va matematika kabi fanlarni birlashtirishga qaror qilishdi va STEM tizimi shu tarzda shakllandi. (Fan, texnika, muhandislik va matematika). Keyinchalik bu yerda Art qo'shildi va endi STEAM oxirigacha shakllandi. O'qituvchilar ushbu mavzular, aniqrog'i ushbu fanlardan bilimlar kelajakda talabalarning yuqori malakali mutaxassis bo'lib yetishishiga yordam beradi, deb hisoblashadi. Oxir oqibat, bolalar yaxshi bilim olishga intilishadi va uni darhol amalda qo'llashadi. D ' ib b d h 'li bi j d h So'nggi o'n yilliklardagi o'zgarishlar yoqimli, ammo shu bilan birga bizni havotirlantiradi. Ushbu yangi narsalarning ixtiro qilinishi bilan odamlar ilgari duch kelmagan ko'plab yangi muammolar mavjud. Har kuni yangi ish turlari va hattoki butun kasbiy sohalar paydo bo'ladi, shuning uchun zamonaviy o'qituvchilar o'qitadigan bilimlari va mahoratlari vaqt talablariga javob beradimi yoki yo'qmi deb o'ylashlari kerak. O'zingizning g'oyangizni topishga bilim yordam beradi, ammo haqiqiy ish bu g' oyani haqiqatga aylantiradi. Agar biz an'anaviy ta'limning asosiy maqsadi bilimlarni o'rgatish va bu bilimlardan fikrlash hamda ijod qilish uchun foydalanish deb aytsak, STEAM yondashuvi bizni olgan bilimlarni haqiqiy ko'nikmalar bilan birlashtirishga o'rgatadi. Bu maktab o'quvchilariga nafaqat ba'zi bir g'oyalarga ega bo'lish, balki ularni amalda qo'llash va amalga oshirish imkoniyatini beradi. O'sha haqiqatda ishlatilishi mumkin bo'lgan bilimgina haqiqatan ham qadrlidir. STEAM yondashuvining eng mashhur namunasi - Massachusets Texnologiya Instituti (MIT). Ushbu dunyo universitetining shiori «Mens et Manus» (Aql va qo'l). ILM-FAN VA INNOVATSIYA ILMIY-AMALIY KONFERENSIYASI in-academy.uz/index.php/si TASVIRIY SAN'AT SIRLARI! CHIZMACHILIK FANINI O‘QITISHDA YANGILIKLAR! Annotatsiya: Maqolada, rassomlikning asosiy prinsiplari, tuzilishi, ranglarni qo'llash, chizmachilik vositalari va yangi texnologiyalardan foydalanish kabi ko'plab mavzular keltirilgan. Maqola, tasviriy san'at va chizmachilik fanlari bo'yicha o'quvchilarning rassomlikni o'rganish va o'zlashtirishiga yordam beruvchi ko'plab usullarni taqdim etadi. Maqolada, tasviriy san'at sirlari keltirilgan va ular o'quvchilarning rassomlikni o'rganish va o'zlashtirish uchun katta ahamiyatga ega. Tasviriy san'atning asosiy prinsiplari, masalan, proporsiya, perspektiva, tsvet i kompozitsiya haqida tushunchalar berilgan. Tasviriy san'atning tuzilishi, masalan, rassomlikda chizmachilik vositalaridan foydalanish, rassomlik materiallari va chizmachilik usullari haqida ham ma'lumotlar berilgan. Maqolada, ranglarni qo'llash, chizmachilik vositalari va yangi texnologiyalardan foydalanish kabi ko'plab mavzular ham keltirilgan. Maqolada, chizmachilik fanini o'qitishda yangiliklar ham keltirilgan. Bu, o'quvchilarning chizmachilikda keng yaxlitlashi, chizmachilik vositalaridan foydalanish, chizmachilik texnikalari va yangi texnologiyalardan foydalanish kabi ko'plab mavzularni o'z ichiga oladi. Kalit so'zlar: Milliy san'at, o'quv materiallari, san'at sirlari, amaliy mashg' ulotlar, takomillashtirish. KIRISH: Tasviriy san'at fani bo'yicha yuqori malakali mutaxassislar tayyorlashda «Tasviriy san'at metodikasi» maktablarda o'rganiladigan maxsus o'quv fanlardan biri bo'lib, u bo'lajak boshlang'ich ta'lim o'qituvchilarini tasviriy san'atdan amaliy mashg'ulotlarda metodik jihatdan tayyorgarligini takomillashtirishning muhim omilidir. Bu nazariy va amaliy mashg'ulotlar o'quvchilarni badiiy-kasbiy bilimini, malaka va mahoratini shakllantirishga hamda ularni maktab o'quvchilari va pedagogika ta'lim-tarbiya berish jarayonida qo'llashga va o' rgatishga qaratilgan. ASOSIY QISM: Tasviriy va amaliy san'at o'qituvchilarini shakllantirish uchun bu sohadagi barcha amaliy hamda nazariy bilimlar ta'lim sifatida zarur. Bular, asosan, qalamtasvir, rangtasvir, kompozitsiya, amaliy bezak san'ati, haykaltaroshlik, san'at tarixi va eng asosiysi, bu bilimlarini o'rgatishning metodik asoslarini egallash hisoblanadi. Ta'lim mazmunidagi bu bilimlarni egallash vositasida bo'lajak o'qituvchi o'z sohasining ustasi bo'libgina qolmasdan, balki ana shu san'at sirlarini o'quvchi va yoshlarga ulasha olish malakasiga ega bo'ladi. Agar ushbu qisqartmani yoysak, quyidagilarni olamiz: STEAM bu - S - science, T - technology, E -engineering, A - art va M - math. Ingliz tilida bu shunday bo'ladi: tabiiy fanlar, texnologiya, muhandislik, san'at va matematika. Ushbu yo'nalishlar zamonaviy dunyoda eng mashhur bo'lib kelayotganini unutmang. Shuning uchun bugungi kunda STEAM tizimi asosiy tendentsiyalardan biri sifatida rivojlanmoqda. STEAM ta'limi yo'nalishi va amaliy yondashuvni qo'llash, shuningdek, barcha beshta sohani yagona ta'lim tizimiga integratsiyalashuviga asoslangan. STEAM yondashuvi o'quv samaradorligiga qanday ta'sir qiladi? uslubiyatiga tayanish lozim. Yosh pedagog tayyorlashda ta'lim mazmunining muhim yo'nalishlari, umum badiiy va milliy an'analariing ustuvorligiga amal qilishi darkor. O'qitishning asosiy talablari uchun zaruriy shartlar esa tasviriy va amaliy san'at sirlarini o'zlashtirish bilan birga o'quvchi va yoshlarga har jihatdan o'rgata olish usullarini egallashdir. Metodika ta'lim-tarbiya jarayonida o'qituvchining o'quvchilar bilan ishlash usullarining, mazmunining xususiyatlarini qarab chiqadi. Ma'lumki, o'qitish jarayonida xilma-xil metodlar o zlashtirish bilan birga o quvchi va yoshlarga har jihatdan o rgata olish usullarini egallashdir. Metodika ta'lim-tarbiya jarayonida o'qituvchining o'quvchilar bilan ishlash usullarining, mazmunining xususiyatlarini qarab chiqadi. Ma'lumki, o'qitish jarayonida xilma-xil metodlar qo'llaniladi. «Metodika» so'zining ma'nosi, avvalo, ta'lim va tarbiyadagi samarali usullarining majmui hisoblanadi. Tasviriy san'at metodikasi oldida pedagogika fanining tarmog'i sifatida quyidagi vazifalar turadi: 1. Tasviriy san'at metodikasining aniq maqsadlari va uning fan sifatida bilim berish hamda tarbiyaviy ahamiyatini aniqlash. 1. Tasviriy san'at metodikasining aniq maqsadlari va uning fan sifatida bilim berish hamda tarbiyaviy ahamiyatini aniqlash. 2. Tasviriy san'at metodikasining mazmuni va tuzilishini aniqlash. 2. Tasviriy san'at metodikasining mazmuni va tuzilishini aniqlash. 2. Tasviriy san'at metodikasining mazmuni va tuzilishini aniqlash. 3. O'quvchilarning chuqur bilim, ko'nikma va malakalarini egallashda eng samarali uslub, vosita va shakllarini ishlab chiqishi. 3. O'quvchilarning chuqur bilim, ko'nikma va malakalarini egallashda eng samarali uslub, vosita va shakllarini ishlab chiqishi. 4. O'quvchilarning bilim olish jarayonini kuzatish va tadqiqot ishlarini olib borish. 4. O quvchilarning bilim olish jarayonini kuzatish va tadqiqot ishlarini olib borish. Tasviriy san'at metodlari deganda, o'qituvchining o'quvchilar bilan ishlash usullari natijasida o'quv materiallarini yaxshi bilib olish darajasiga erishish va o'zlashtirish darajasining oshirilishi nazarda tutiladi. Har bir o'qitish metodlari o'quvchilar egallashi lozim bo'lgan bilimlar mazmuni (hajmi)ga monand tarzda ta'lim maqsadiga, o'quvchilarning yosh va fiziologik xususiyatlariga bog'liq. O'qitish usuli - bu ta'lim-tarbiyaning alohida qismlari hisoblanib, ularning yig'indisi o'qitish metodlariga kiradi. Tasviriy san'at metodlari deganda, o'qituvchining o'quvchilar bilan ishlash usullari natijasida o'quv materiallarini yaxshi bilib olish darajasiga erishish va o'zlashtirish darajasining oshirilishi nazarda tutiladi. Har bir o'qitish metodlari o'quvchilar egallashi lozim bo'lgan bilimlar mazmuni (hajmi)ga monand tarzda ta'lim maqsadiga, o'quvchilarning yosh va fiziologik xususiyatlariga bog'liq. O'qitish usuli - bu ta'lim-tarbiyaning alohida qismlari hisoblanib, ularning yig'indisi o'qitish metodlariga kiradi. Xulosa qilish mumkin, "Tasviriy san'at sirlari! Chizmachilik fanini o‘qitishda yangiliklar!" nomli maqola, tasviriy san'at va chizmachilik fanlari bo'yicha o'quvchilarning rassomlikni o'rganish va o'zlashtirishiga yordam beruvchi ko'plab ma'lumotlarni taqdim etadi. Maqolada, tasviriy san'atning asosiy prinsiplari, tuzilishi, ranglarni qo'llash, chizmachilik vositalari va yangi texnologiyalardan foydalanish kabi ko'plab mavzularni ko'rsatilgan. Bu ma'lumotlar o'quvchilarning rassomlikka qiziqishini oshirish uchun yaxshi yordam beradi. References: 1. N. Abdullayev. San'at tarixi. 2 jildli, 1-jild, T ., 1986. 2. B.Z. Azimova. Natyurmort tuzish va tasvirlash. T ., « O'qituvchi», 1984. 3. B. Boymetov. Qalamtasvir. «Musiqa» nashriyoti Toshkent, 2006. 4. X.E. Sultanov, N.A. Mansurov. Chizmatasvir (Laboratoriya ishlari). Toshkent-2014. 5. G'.I. Muxamedov, U.N. Xodjamqulov. Pedagogik ta'lim innovatsion klasteri: ta'rif, tavsif, tasnif. Chirchiq «Universitet», 2019. Massachusets Texnologiya Instituti bolalarga STEAM tushunchasini oldindan o'rganish va tanishish imkoniyatini berish uchun STEAM kurslarini ishlab chiqdi va hattoki ba'zi ta'lim muassasalarida STEAM o'quv markazlarini yaratdi. Tasviriy, amaliy san'at ta'limi va tarbiyasi mazmunida milliy san'at merosimiz bilan birgalikda umum badiiy ta'limshunosligi va 28 O'quvchilarga o'rgatish, ko'nikma va malakalarini shakllantirish amaliy mashg'ulotlar mobaynida uzoq ijodiy mehnat orqali egallanadi. Demak, tasviriy san'at metodikasi - pedagogika ilmida o'quvchilarni tasviriy san'atga o'rgatish mazmuni, vazifa va metodlarini aniqlovchi, ijodiy ishlaridagi oqilona usullarni o'rganuvchi, ta'lim-tarbiyaning maqsad va vazifalariga tayangan holda samarali o'quv jarayonini tashkil etuvchi, shakl va yo'llarini tadqiq etuvchi sohasidir. Tasviriy san'at metodikasi o'qitish metodikasining boshqa sohalari kabi umumiy va xususiylarga ajratiladi. 29 ILM-FAN VA INNOVATSIYA ILMIY-AMALIY KONFERENSIYASI in-academy.uz/index.php/si 6. X.E. Sultanov. Rangtasvir (Akvarelda natyurmort ishlash) o'quv qo'llanma. Toshkent- 2019. 7 X X Muratov Qalamtasvir Toshkent Ijod-print 2020 ILM-FAN VA INNOVATSIYA ILMIY-AMALIY KONFERENSIYASI in-academy.uz/index.php/si 6. X.E. Sultanov. Rangtasvir (Akvarelda natyurmort ishlash) o'quv qo'llanma. Toshkent- 2019. 7. X.X. Muratov, Qalamtasvir. Toshkent Ijod-print, 2020. 8. X.E. Sultanov, S.T. Sobirov. Tasviriy san'at o'qitish metodikasi. Toshkent «Ishonchli hamkor» 2021 ILM-FAN VA INNOVATSIYA ILMIY-AMALIY KONFERENSIYASI in-academy.uz/index.php/si 6. X.E. Sultanov. Rangtasvir (Akvarelda natyurmort ishlash) o'quv qo'llanma. Toshkent- 2019. 7. X.X. Muratov, Qalamtasvir. Toshkent Ijod-print, 2020. 8. X.E. Sultanov, S.T. Sobirov. Tasviriy san'at o'qitish metodikasi. Toshkent «Ishonchli hamkor» 2021 7. X.X. Muratov, Qalamtasvir. Toshkent Ijod-print, 2020. 7. X.X. Muratov, Qalamtasvir. Toshkent Ijod-print, 2020. 8. X.E. Sultanov, S.T. Sobirov. Tasviriy san'at o'qitish metodikasi. Toshkent «Ishonchli hamkor» 2021 8. X.E. Sultanov, S.T. Sobirov. Tasviriy san'at o'qitish metodikasi. Toshkent «Ishonchli hamkor» 2021 8. X.E. Sultanov, S.T. Sobirov. Tasviriy san'at o'qitish metodikasi. Toshkent «Ishonchli hamkor» 2021 30
https://openalex.org/W2806198418	https://europepmc.org/articles/pmc5861037?pdf=render	English	null	Lysine acetylation of DosR regulates the hypoxia response of <i>Mycobacterium tuberculosis</i>	Emerging microbes & infections	2,018	cc-by	11,802	The Author(s) 2018 OpenAccessThisarticleislicensedunderaCreativeCommonsAttribution4.0InternationalLicense,whichpermitsuse,sharing,adaptation,distributionandreproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. Abstract Tuberculosis caused by Mycobacterium tuberculosis (Mtb) infection remains a large global public health problem. One striking characteristic of Mtb is its ability to adapt to hypoxia and trigger the ensuing transition to a dormant state for persistent infection, but how the hypoxia response of Mtb is regulated remains largely unknown. Here we performed a quantitative acetylome analysis to compare the acetylation proﬁle of Mtb under aeration and hypoxia, and showed that 377 acetylation sites in 269 Mtb proteins were signiﬁcantly changed under hypoxia. In particular, deacetylation of dormancy survival regulator (DosR) at K182 promoted the hypoxia response in Mtb and enhanced the transcription of DosR-targeted genes. Mechanistically, recombinant DosRK182R protein demonstrated enhanced DNA-binding activity in comparison with DosRK182Q protein. Moreover, Rv0998 was identiﬁed as an acetyltransferase that mediates the acetylation of DosR at K182. Deletion of Rv0998 also promoted the adaptation of Mtb to hypoxia and the transcription of DosR-targeted genes. Mice infected with an Mtb strain containing acetylation-defective DosRK182R had much lower bacterial counts and less severe histopathological impairments compared with those infected with the wild-type strain. Our ﬁndings suggest that hypoxia induces the deacetylation of DosR, which in turn increases its DNA-binding ability to promote the transcription of target genes, allowing Mtb to shift to dormancy under hypoxia. A R T I C L E O p e n A c c e s s Lysine acetylation of DosR regulates the hypoxia response of Mycobacterium tuberculosis Hua Yang1, Wei Sha1, Zhonghua Liu1, Tianqi Tang2, Haipeng Liu1, Lianhua Qin1, Zhenling Cui1, Jianxia Chen1, Feng Liu1, Ruijuan Zheng1, Xiaochen Huang1, Jie Wang1, Yonghong Feng1 and Baoxue Ge1,2 Yang et al. Emerging Microbes & Infections (2018) 7:34 DOI 10.1038/s41426-018-0032-2 Yang et al. Emerging Microbes & Infections (2018) 7:34 DOI 10.1038/s41426-018-0032-2 Emerging Microbes & Infections Introduction prevent Mtb infection or even cure individuals harboring Mtb in the dormant stage3. Adaptations to hypoxia have been implicated to play a prominent role in the dormancy of Mtb4. Accumulating evidence has demonstrated that oxygen tension correlates tightly with Mtb growth, the formation of hard, ﬁbrous, hypoxic granulomas, and, eventually, the development of infection5. Wayne et al.6, 7 have pioneered the development of culture conditions that gradually deprive bacteria of oxygen to generate nonreplicating persistent bacilli in vitro as a model for latency. With this, Schubert et al.8 proﬁled the proteome of dormant Mtb and observed that the stress-induced dormancy survival regulator (DosR) regulon consisted of ~20% of the cellular protein content during dormancy. However, whether posttranslational modiﬁcations (PTMs) are involved in this process remains elusive. In 2015, ~10.4 million new infections and 1.8 million deaths globally were caused by Mycobacterium tubercu- losis (Mtb)1. The remarkable success of Mtb as a pathogen is due to its marked phenotypic drug resistance and its ability to evade the host immune system. Mtb can also enter into a nonreplicating state and persist in humans for extended periods without causing disease in what is known as a latent tuberculosis (TB) infection (LTBI)2. Clariﬁcation of the mechanisms underlying the LTBI state may lead to the development of useful pharmaceuticals to Correspondence: Baoxue Ge (baoxue_ge@tongji.edu.cn) 1Shanghai Key Laboratory of Tuberculosis, Clinic and Research Center of Tuberculosis, Shanghai Pulmonary Hospital, Tongji University School of Medicine, Shanghai 200433, PR China 2Department of Microbiology and Immunology, Tongji University School of Medicine, Shanghai 200049, PR China © The Author(s) 2018 OpenAccessThisarticleislicensedunderaCreativeCommonsAttribution4.0InternationalLicense,whichpermitsuse,sharing,adaptation,distributionandreproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. Page 2 of 14 Page 2 of 14 Page 2 of 14 Yang et al. Emerging Microbes & Infections (2018) 7:34 Nε-lysine acetylation is an abundant and evolutionarily conserved PTM found in prokaryotes and eukaryotes9. As a dynamic and reversible process, lysine acetylation affects protein conformation and/or charge, thus altering DNA- binding afﬁnity, enzymatic activity, protein–protein interactions, and protein or mRNA stability10, 11. Increased global analyses of lysine acetylation in bacteria have been reported, including in Escherichia coli12, 13, Erwinia amylovora14, Bacillus subtilis15, Salmonella enteric16, and Mtb17, 18, as well as in others19. Protein acetylation is implicated in nearly all cellular processes, such as central metabolism, protein translation, and pathogen virulence20, 21. Increasing evidence has shown that protein acetylation plays an important regulatory role in mycobacteria22, 23. The universal stress protein (USP) was the ﬁrst-characterized acetylated protein in myco- bacterium24. Recently, Bi et al.25 identiﬁed the modulation of central carbon metabolism by the acetylation of iso- citrate lyase in Mtb. Given the widespread regulatory role of lysine acetylation, it is tempting to speculate that lysine acetylation is involved in the hypoxia adaption of Mtb. Hence, we examined alterations in the protein acetylation proﬁle of Mtb in response to hypoxia by quantitative proteomics and subsequently identiﬁed 377 acetylation sites in 269 proteins that were signiﬁcantly changed. Furthermore, we conﬁrmed the lysine acetylation of DosR and its role in the response of Mtb to hypoxia. observation that most lysine-acetylated proteins are involved in metabolism. Furthermore, we performed Gene Ontology (GO) enrichment analysis on the acetylation data. As revealed previously, hypoxia induces comprehensive changes in the bacterium, with Mtb shifting to exploit lipids as the pri- mary nutrient source27. © The Author(s) 2018 Consistently, our data demon- strated that upregulated acetylated proteins under hypoxia were enriched for biological processes related to lipid metabolism (Fig. 1d). From the GO enrichment analysis, we also found that upregulated acetylated pro- teins under hypoxia were signiﬁcantly enriched for some speciﬁc molecular functions, such as ﬂavin adenine dinucleotide binding, oxidoreductase activity, and coen- zyme binding (Fig. 1d). For the 54 acetylated proteins downregulated under hypoxia, biological processes asso- ciated with the growth of a symbiont involved in its interaction with a host were found to be enriched (data not shown). Additionally, proteins with ligase activity that form carbon–nitrogen bonds also had a higher tendency to be acetylated under normal aerobic conditions. For the Kyoto Encyclopedia of Genes and Genomes (KEGG) metabolic pathways analysis, upregulated acetylated pro- teins under hypoxia were highly enriched for carbon metabolism, RNA degradation, and propanoate metabo- lism, among others (Fig. 1e), while acetylated proteins involved in the ribosome were highly enriched under aerobic conditions. Overall, these data indicate that lysine acetylation might play a comprehensive regulatory role in the adaption of Mtb to hypoxia. Hypoxia induces deacetylation of DosR To proﬁle the lysine acetylation of Mtb during exponential growth and hypoxia-induced dormancy, we cultured Mtb H37Rv under normal aerobic conditions and gradually decreasing oxygen levels using the Wayne model (Supplementary Figure S1)26. We har- vested cultures at day 12 during exponential growth and hypoxia. Acetylated peptides were enriched using a speciﬁc acetyl lysine antibody and analyzed by liquid chromatography–tandem mass spectrometry (LC–MS/ MS) (Fig. 1a). We identiﬁed 1215 acetylation sites in 679 proteins, of which 852 acetylation sites in 497 proteins were quantiﬁed (Supplementary Table S1). Relative quantitation was divided into two categories: under hypoxia, 321 acetylation sites in 215 proteins were upregulated; and 56 acetylation sites in 54 pro- teins were downregulated (Supplementary Table S2). Classiﬁcation results for biological processes and molecular functions both showed that the largest group of upregulated acetylated proteins under hypoxia con- sisted of enzymatic proteins associated with metabo- lism, which accounted for 31% and 56% of the upregulated acetylated proteins, respectively (Fig. 1b, c). These ﬁndings are consistent with the previous Numerous studies have demonstrated the important role of DosR during the physiological adaptation of Mtb to hypoxia28, 29. Three amino acids in DosR, namely, Lys179, Lys182, and Asn183, directly bind to nucleotide bases in the DNA motif and induce the expression of target genes30. Mutation of Lys182 directly reduces the DNA-binding afﬁnity of DosR and abrogates the induc- tion of its regulon genes, thus highlighting the important role of K182 in DosR31; however, the relevant mechanism is still unclear. From our quantitative acetylome proﬁling data, DosR was shown to be acetylated at K182 (Fig. 1f). Considering the location and positive charge of this lysine residue, we hypothesized that acetylation of K182 may be involved in regulating DosR activity, speciﬁcally the DNA- binding ability of DosR. Sequence alignment analysis demonstrated that K182, the acetylated lysine residue, is evolutionarily conserved among different mycobacteria species (Fig. 1g). Interestingly, quantitative proteomics analysis revealed that the acetylation of DosR K182 was reduced under hypoxia. To further conﬁrm whether the acetylation levels of DosR K182 respond to hypoxic stress, we produced anti-DosR- and anti-DosR Ac-K182-speciﬁc Yang et al. Emerging Microbes & Infections (2018) 7:34 Page 3 of 14 polyclonal antibodies (characterized as shown in Supple- mentary Figure S2) and applied these to examine the acetylation levels of DosR K182 in Mtb H37Rv cultured under normal aerobic or hypoxic conditions. Hypoxia induces deacetylation of DosR Consistently, a signiﬁcant reduction of acetylation at K182 of DosR under hypoxia was observed (Fig. 1h), indicating that Fig. 1 (See legend on next page.) Fig. 1 (See legend on next page.) under normal aerobic or hypoxic conditions. Consistently, a signiﬁcant reduction of acetylation at K182 of DosR under hypoxia was observed (Fig. 1h), indicating that polyclonal antibodies (characterized as shown in Supple- mentary Figure S2) and applied these to examine the acetylation levels of DosR K182 in Mtb H37Rv cultured Page 4 of 14 Yang et al. Emerging Microbes & Infections (2018) 7:34 (see ﬁgure on previous page) Fig. 1 Hypoxia induces deacetylation of DosR. a Workﬂow of quantitative proteomics analysis of lysine acetylation in H37Rv under hypoxia. Mtb H37Rv was cultured under aeration or hypoxia in conical screw-capped Nephelo ﬂasks with 20 mm side arms according to the Wayne model. Samples for aeration (Aera) and hypoxia (Hypo) cultures were harvested after 12 days. Proteins were extracted and digested by trypsin. After separate labeling with a tandem mass tag (TMT) 6-plex kit, peptides from Aera or Hypo were mixed 1:1, enriched with anti-acetyl lysine antibody beads and loaded for LC–MS/MS analysis. The resulting MS/MS data were processed for bioinformatics analysis. b, c Gene ontology functional classiﬁcation of biological processes (b) and molecular functions (c) of 215 upregulated acetylated proteins under hypoxia. d Gene ontology enrichment analysis of biological processes and molecular functions of 215 upregulated acetylated proteins under hypoxia. The differently colored bars indicate the corrected p values for the enrichment of the annotations. e KEGG pathway enrichment analysis of 215 upregulated acetylated proteins under hypoxia. f Mass spectroscopy (MS) analysis identiﬁed K182 acetylation of DosR. g K182 in DosR is conserved among various mycobacteria species. The asterisk denotes the conserved lysine residue, and this result was analyzed by DNASTAR software. h Acetylation of DosRK182 in H37Rv WCL under hypoxia was downregulated. H37Rv were grown under aeration or hypoxia in 7H9 broth supplemented with 10% ADC for 3 days. DosR proteins were immunoprecipitated (IP) by the anti-DosR antibody. The acetylation level of DosR was determined by immunoblotting (IB) using an anti-DosR Ac- K182-speciﬁc polyclonal antibody. Western blots are representative of at least three independent experiments DosR may undergo deacetylation at K182 in response to hypoxia. hypothesized that acetylation of DosR K182 regulated the transcription of dormancy survival-related genes. Hypoxia induces deacetylation of DosR Several known DosR-regulated genes, including acr, dosR, fdxA, Rv1738, Rv2030, Rv2623, Rv2626c, and Rv3130c, were selected to evaluate whether K182 acetylation affects the activity of DosR as a transcription factor. A DosR deletion mutant of Mtb, WT, and a deacetylation complementation mutant were cultured under hypoxic conditions as described previously. As expected, the transcription levels of eight candidate genes were all signiﬁcantly upregulated for the K182R mutant under hypoxia compared with normal aeration conditions and the WT strain (Fig. 2e–l). Furthermore, we wanted to determine whether acetylation of K182 directly hindered the DNA-binding activity of DosR. We constructed expression plasmids of WT DosR and its derivatives, K182R and K182Q (substituted with glutamine to mimic the constitutively acetylated form), using a 6× His tag sequence at their C terminals. All proteins were overexpressed, puriﬁed, and adjusted to a similar concentration. electrophoretic mobility shift assay (EMSA) was performed by incubating the above puriﬁed DosR derivatives with the acr promoter DNA region, fol- lowed by non-denaturing polyacrylamide gel electrophor- esis and silver staining analysis. EMSA showed that K182Q had almost no DNA-binding afﬁnity compared to WT DosR, while K182R had a similar DNA-binding ability to WT DosR (Fig. 2m). Therefore, acetylation of the K182 residue is critical for the DNA-binding ability of DosR, the transcription of DosR downstream target genes, and the growth of Mtb under hypoxic conditions. Deacetylation of DosR promotes the adaption of Mtb to hypoxia yp Based on the quantitative acetylome data generated in this study, six of the key regulated proteins from the hypoxia regulatory network27 were acetylated at various levels in response to hypoxic stress, namely, DosR, PhoP, SigA, SigB, SigF, and Lsr2 (Table 1). To examine whether acetylation affects the bacterium’s adaption to hypoxia, we generated recombinant Mycobacterium smegmatis expressing corresponding wild-type (WT) genes, as well as counterparts with a lysine-to-arginine substitution of the identiﬁed acetylated lysine to mimic the non- acetylated form. The effect of deacetylation on the growth of 13 recombinant M. smegmatis strains under hypoxia was then analyzed. We found that DosR K182R and SigA K334R mutants have signiﬁcant effects on growth under hypoxia in M. smegmatis (Fig. 2a, b and Supplementary Figure S3), which suggests that the adap- tion to hypoxia by mycobacteria may be regulated by the lysine acetylation of DosR or SigA. Furthermore, we constructed Mtb H37Rv mutants deﬁcient in DosR (H37Rv:ΔdosR) (characterized as shown in Supplementary Figure S4A–S4C), complemented with WT DosR (H37Rv: ΔdosR::dosR) and deacetylation mutants of K182 (mimicked by an arginine substitution, K182R; H37Rv: ΔdosR::dosRK182R). In line with the observations of recombinant M. smegmatis, the deacetylation mutant did not have a signiﬁcant effect on normal aerobic growth but signiﬁcantly promoted the response to hypoxia by Mtb (Fig. 2c, d), suggesting that acetylation of DosR K182 directly regulates the adaption of Mtb to hypoxia. Rv0998 acetylates DosR and negatively regulates the adaption of Mtb to hypoxia Lysine acetylation is reversible and is performed in bac- teria by members of the GCN5-related N-acetyltransferase (GNAT) superfamily, as well as by NAD+-dependent dea- cetylases33. In Mtb, 20 GNATs have been predicted. The cAMP-dependent protein lysine acetyltransferase Rv0998 was recently shown to catalyze the acetylation of many important proteins34. To examine whether Rv0998 catalyzes Deacetylation of K182 regulates target gene transcription and the DNA-binding activity of DosR DosR controls the dormancy survival regulon that con- sists of ~50 genes32. Since deacetylation of DosR K182 directly promotes Mtb adaption to hypoxia, we Page 5 of 14 Yang et al. Emerging Microbes & Infections (2018) 7:34 Table 1 Acetylation of six regulated proteins from Mtb hypoxia regulatory network Protein Protein description Position Hypo-vs- Aera P71814 OmpR family two-component system response regulator 197 3.47 P9WGI3 RNA polymerase sigma factor SigF 121 2.45 P9WGI5 RNA polymerase sigma factor SigB 308 2.21 P9WGI1 RNA polymerase sigma factor SigA 334 1.57 P9WMF9 Transcriptional regulatory protein DevR (DosR) 179 1.53 182 0.65 P9WIP7 Nucleoid-associated protein Lsr2 43 0.53 Table 1 Acetylation of six regulated proteins from Mtb hypoxia regulatory network studies of Parish35. The K182R mutant exhibited lower bacterial proliferation than the WT complement mutants (Fig. 4a). We also used the deletion mutant of Rv0998 to detect the intracellular survival of Mtb in peritoneal macrophages. As expected, we observed signiﬁcantly retarded growth of the Rv0998 deletion mutant, with multiplicities of infection of 1 or 2 at 24 h (Fig. 4b). To further understand the functional role of DosR K182 acetylation in the pathogenesis of TB, we performed intraperitoneal (i.p.) infections with Mtb in mice. We found that even though the average injected dose for the four groups was similar (Fig. 4c), the numbers of bacteria in the lungs and spleen were both elevated in mice infected with the DosR deletion mutant for 28 days (Fig. 4d, e). The number of DosR complement mutant bacilli in the lungs and spleen was signiﬁcantly decreased compared with the deletion mutant, while that of the K182R complement mutant was even lower (Fig. 4d, e). From lung histopathological analysis, larger lesions and the pathological involvement of a greater proportion of the lung were observed for mice infected with the DosR deletion mutant relative to mice infected with DosR and K182R complement mutant bacilli, although damage by the K182R mutant was milder (Fig. 4f). In conclusion, the deacetylation of K182 caused signiﬁcantly lower Mtb bacterial loads as well as attenuated pulmonary inﬂam- mation in a mouse model. the acetylation of DosR K182, we constructed an Rv0998 deletion mutant of Mtb H37Rv (characterized as shown in Supplementary Figures S4D and S4E). Western blotting revealed that the acetylation level of DosR at K182 in the Rv0998 deletion mutant was signiﬁcantly reduced (Fig. 3a). Deacetylation of K182 regulates target gene transcription and the DNA-binding activity of DosR To further conﬁrm whether DosR is a substrate of Rv0998, we overexpressed and puriﬁed 6× His-tagged DosR pro- teins from E. coli, as well as Rv0998. DosR was incubated with Rv0998 in the presence of AcCoA and/or cAMP. Western blotting revealed that DosR K182 could be acety- lated by Rv0998 in vitro without cAMP, although cAMP further enhanced the function of Rv0998 (Fig. 3b). We also synthesized a short peptide of DosR, from 175 to 188 amino acids in length, termed K182-pep. In vitro acetylation reactions further conﬁrmed that Rv0998 could acetylate DosR K182-pep (Fig. 3c). Next, we used the Rv0998 dele- tion mutant to conﬁrm the effect of K182 acetylation on growth under hypoxia and on target gene transcription. The Rv0998 deletion mutant was found to signiﬁcantly promote the Mtb response to hypoxia (Fig. 3d) and to dramatically activate the transcription of eight DosR target genes (Fig. 3e–l). Discussion As is well known, hypoxia stress limits the in vivo growth of Mtb. However, Mtb can adapt to limited oxygen by entering into a metabolically altered state while awaiting the opportunity to reactivate. Deﬁning the stra- tegies employed by Mtb to shape its adaption to hypoxia is of intense interest. Many “omics” analyses have been used to comprehensively evaluate the response of Mtb to hypoxia8, 27, 37. The utilization of speciﬁc antibodies to acetylated peptides in immunoprecipitation has sig- niﬁcantly improved the enrichment and identiﬁcation of acetylated lysine residues with proteomic proﬁling38. Although an analysis of the acetylome of Mtb has recently been reported17, 18, 25, the impact of lysine acetylation on the adaption of Mtb to hypoxia has not yet been addres- sed. Understanding how Mtb modulates lysine acetylation in response to oxygen limitation may lead to new strate- gies for its control. Protein acetylation contributes to the pathogenesis of Mtb Compared with the previous study, there are many proteins with the same acetylation site in the quantitative experiment (data not shown), which also includes the K182 site of DosR. This reﬂects the reliability of our quantitative results. The acetylation of transcription factors has recently been demonstrated to be a conserved regulatory means to modulate gene expression in both bacteria and eukar- yotes42, 43. Similarly, our results suggest that deacetylation of K182 is critical for the DNA-binding activity of DosR (Fig. 2m), which promoted the transcription of DosR downstream target genes (Fig. 2e–l), leading to the rapid response of Mtb to hypoxia. We subsequently attempted to identify the key acetyltransferases that regulated the reversible acetylation of DosR K182. We measured the K182 acetylation level in recombinant M. smegmatis overexpressing ﬁve acetyltransferases from Mtb and found that one lysine acetyltransferase, Rv0998, signiﬁcantly promoted the acetylation of DosR (data not shown). From experiments using the deletion mutant and an in vitro acetylation assay (Fig. 3a–c), we postulated that DosR was a substrate of Rv0998. The residual acetylation of DosR in the Rv0998 deletion mutant indicated that another as yet to be identiﬁed acetyltransferase(s) is also involved in the regulation of the acetylation of DosR. The phenotype obtained from the Rv0998 deletion mutant further supported the notion that deacetylation of DosR promoted target gene transcription and had a profound effect on the growth of Mtb under hypoxia (Fig. 3d–l). DosR is a key transcription factor that facilitates Mtb adaptation to hypoxia, and K182 plays an important role in the DNA-binding function of DosR41. Based on the altered K182 acetylation under hypoxia as detected by quantitative proteomics, we further detected this mod- iﬁcation using a speciﬁc polyclonal antibody to lysine acetylation. This showed that oxygen limitation stress led to the deacetylation of DosR at K182. Because of the important role of DosR in the response of Mtb to hypoxia, we focused on the functional and regulatory mechanism of K182 acetylation in DosR. We found that K182 Of note, earlier studies revealed a contradictory role for DosR in Mtb virulence. Parish et al.35 reported that the DosR mutant of Mtb is hypervirulent in both SCID and C57Bl/6 mice, while studies in guinea pigs, rabbits, and even macaques revealed that mutants of DosR exhibited reduced infectivity36, 44, 45. Protein acetylation contributes to the pathogenesis of Mtb Protein acetylation contributes to the pathogenesis of Mtb Earlier studies revealed that DosR plays a critical reg- ulatory role in the adaptation and survival of Mtb within cells and tissues35, 36. We hypothesized that DosR K182 acetylation was critical for bacterial survival and pro- liferation in macrophages. To answer this question, pri- mary peritoneal macrophages were infected with Mtb H37Rv, the deletion mutant of DosR, and WT or K182R complement mutants with a multiplicity of infection (MOI) of 1. After 24 h, cells were lysed, and the number of intracellular bacteria was calculated by colony-forming unit (CFU) counting. We found that the deletion mutant multiplied more efﬁciently than Mtb H37Rv, WT, or K182R complement mutants, in agreement with the We implemented a universally applicable, tandem mass tag (TMT)-labeled approach to estimate variations in lysine acetylation in Mtb during exponential growth and hypoxia-induced dormancy39. We identiﬁed 1215 acet- ylation sites on 679 proteins, many more than found by Bi et al.25. The 852 acetylation sites on 497 proteins were identiﬁed under both aerated and hypoxic conditions and were quantiﬁed. In addition, 363 residual unquantiﬁable Yang et al. Emerging Microbes & Infections (2018) 7:34 Page 6 of 14 Fig. 2 (See legend on next page.) 2 (See legend on next page ) Fig. 2 (See legend on next page.) Fig. 2 (See legend on next page.) Page 7 of 14 Yang et al. Emerging Microbes & Infections (2018) 7:34 (see ﬁgure on previous page) Fig. 2 Deacetylation of DosR promotes the hypoxia adaption of Mtb. a, b Growth in aeration (a) or hypoxia (b) of M. smegmatis overexpressing DosR or a K-R mutant. Recombinant M. smegmatis overexpressing DosR or the K-R mutant were grown to mid-log phase; the growth curve was measured using a Bioscreen Growth Curve Instrument. M. smegmatis harboring pVV16 was used as a control. Hypoxic conditions were established by covering bacterial suspensions with parafﬁn oil. The optical density was measured at an absorbance of 590 nm every 2 h. Cultures were grown at 37 ° C for 7 days. The results were combined (mean ± standard deviation (SD)) from two independent experiments, with each experiment performed in triplicate. A statistically signiﬁcant difference was found between Ms::dosR and Ms::dosRK182R under hypoxic conditions, as analyzed by Student’s t-test (P < 0.05). Protein acetylation contributes to the pathogenesis of Mtb c, d Growth under aeration (c) or hypoxia (d) of H37Rv, the deletion mutant of dosR, and WT or K182R complement mutants. Growth curves were measured as described above, and cultures were grown at 37 °C for 14 days. A signiﬁcant difference was also found between the WT complement strain and K182R complement mutants under hypoxia. P < 0.05. Student’s t-test. e–l The K182R mutant promoted the transcription of the target genes acr (e), dosR (f), fdxA (g), Rv1738 (h), Rv2030 (i), Rv2623 (j), Rv2626c (k), and Rv3130c (l). Bacteria were cultured and harvested at day 3 to isolate total RNA. Relative transcriptional levels were determined using the 2−ΔΔCt method. The reference gene used was 16S rRNA. Values represent the mean ± SD from three independent experiments. P < 0.05. * P<0.01. Student’s t-test. m Deacetylation of K182 strengthened the DNA- binding activity of DosR. An electrophoretic mobility shift assay (EMSA) was used to test the binding of DosR and the derivative mutant proteins K182R and K182Q to the promoter region of acr. The EMSA result is representative of at least three independent experiments deacetylation directly regulated Mtb growth under hypoxia with knockout and complement strains of Mtb H37Rv (Fig. 2d), in accordance with the results from recombinant M. smegmatis. acetylation sites may have also been derived from dra- matic variations in this modiﬁcation under different growth conditions, or they could have been due to the poor replicability of the high-throughput analyses40. Among the quantiﬁed sites, we found 377 acetylation sites on 269 proteins that were signiﬁcantly changed. These ﬁndings strongly indicate that protein acetylation is important for the regulation of protein function and the subsequent adaption to hypoxia. Unfortunately, the quantiﬁcation experiment was performed only once (two samples and one biological replicate for each experi- mental condition), so there was neither a description of reproducibility nor any statistical signiﬁcance. Although we lacked biological replicates, we produced many func- tional experiments to verify the reliability of the pro- teomics data. After screening by proteomics data, we selected DosR as our research object. Taking into account the reliability of the data, we produced an anti-DosR Ac- K182-speciﬁc polyclonal antibody to determine whether the acetylation of the target protein was consistent with the proteomics data. On the other hand, we produced a qualitative experiment of acetylation in the same strain of H37Rv. Protein acetylation contributes to the pathogenesis of Mtb DosR (2 μg) was incubated with or without Rv0998 (2 μg) or cAMP (200 μM). Acetylation levels were determined by immunoblotting. The western blot shown is representative of at least three independent experiments. c Rv0998 acetylates K182 peptides of DosR in vitro. K182 peptide (2 μg) was incubated, with or without Rv0998 (2 μg) or cAMP (200 μM). Acetylation levels were determined by dot blot. The dot blot shown is representative of at least three independent experiments. d Deletion of Rv0998 promoted the hypoxia response of Mtb. A growth curve was measured for the Rv0998 deletion mutant and wild- type Mtb H37Rv under aeration and hypoxia. e–l Deletion of Rv0998 promoted the transcription of acr (e), dosR (f), fdxA (g), Rv1738 (h), Rv2030 (i), Rv2623 (j), Rv2626c (k), and Rv3130c (l). Bacteria were cultured and harvested on day 3 to isolate total RNA. Relative transcriptional levels were determined using the 2−ΔΔCt method. The expression of tested genes was normalized to that of 16S rRNA. Values represent the mean ± SD from three independent experiments. P < 0.05.* P<0.01. Student’s t-test Fig. 3 Rv0998 acetylates DosR and negatively regulates the adaption of Mtb to hypoxia. a Acetylation of DosR in the Rv0998 deletion mutant of Mtb H37Rv was downregulated. DosR from the Rv0998 deletion mutant or the wild-type Mtb H37Rv was immunoprecipitated (IP) by an anti-DosR antibody. Acetylation levels were detected by immunoblotting (IB) with an anti-DosR Ac-K182-speciﬁc polyclonal antibody. Western blots were repeated at least three times—a representative blot is shown. b Rv0998 acetylates DosR K182 in vitro. DosR (2 μg) was incubated with or without Rv0998 (2 μg) or cAMP (200 μM). Acetylation levels were determined by immunoblotting. The western blot shown is representative of at least three independent experiments. c Rv0998 acetylates K182 peptides of DosR in vitro. K182 peptide (2 μg) was incubated, with or without Rv0998 (2 μg) or cAMP (200 μM). Acetylation levels were determined by dot blot. The dot blot shown is representative of at least three independent experiments. d Deletion of Rv0998 promoted the hypoxia response of Mtb. A growth curve was measured for the Rv0998 deletion mutant and wild- type Mtb H37Rv under aeration and hypoxia. e–l Deletion of Rv0998 promoted the transcription of acr (e), dosR (f), fdxA (g), Rv1738 (h), Rv2030 (i), Rv2623 (j), Rv2626c (k), and Rv3130c (l). Protein acetylation contributes to the pathogenesis of Mtb Here we showed that the deletion mutant of DosR promoted the intracellular sur- vival of Mtb in macrophages, while a K182R mutant exhibited lower proliferation than WT complement Yang et al. Emerging Microbes & Infections (2018) 7:34 Page 8 of 14 Fig. 3 Rv0998 acetylates DosR and negatively regulates the adaption of Mtb to hypoxia. a Acetylation of DosRK182 in the Rv0998 deletion mutant of Mtb H37Rv was downregulated. DosR from the Rv0998 deletion mutant or the wild-type Mtb H37Rv was immunoprecipitated (IP) by an anti-DosR antibody. Acetylation levels were detected by immunoblotting (IB) with an anti-DosR Ac-K182-speciﬁc polyclonal antibody. Western blots were repeated at least three times—a representative blot is shown. b Rv0998 acetylates DosR K182 in vitro. DosR (2 μg) was incubated with or without Rv0998 (2 μg) or cAMP (200 μM). Acetylation levels were determined by immunoblotting. The western blot shown is representative of at least three independent experiments. c Rv0998 acetylates K182 peptides of DosR in vitro. K182 peptide (2 μg) was incubated, with or without Rv0998 (2 μg) or cAMP (200 μM). Acetylation levels were determined by dot blot. The dot blot shown is representative of at least three independent experiments. d Deletion of Rv0998 promoted the hypoxia response of Mtb. A growth curve was measured for the Rv0998 deletion mutant and wild- type Mtb H37Rv under aeration and hypoxia. e–l Deletion of Rv0998 promoted the transcription of acr (e), dosR (f), fdxA (g), Rv1738 (h), Rv2030 (i), Rv2623 (j), Rv2626c (k), and Rv3130c (l). Bacteria were cultured and harvested on day 3 to isolate total RNA. Relative transcriptional levels were determined using the 2−ΔΔCt method. The expression of tested genes was normalized to that of 16S rRNA. Values represent the mean ± SD from three independent experiments. P < 0.05.* P<0.01. Student’s t-test Fig. 3 Rv0998 acetylates DosR and negatively regulates the adaption of Mtb to hypoxia. a Acetylation of Fig. 3 Rv0998 acetylates DosR and negatively regulates the adaption of Mtb to hypoxia. a Acetylation of DosR in the Rv0998 deletion mutant of Mtb H37Rv was downregulated. DosR from the Rv0998 deletion mutant or the wild-type Mtb H37Rv was immunoprecipitated (IP) by an anti-DosR antibody. Acetylation levels were detected by immunoblotting (IB) with an anti-DosR Ac-K182-speciﬁc polyclonal antibody. Western blots were repeated at least three times—a representative blot is shown. b Rv0998 acetylates DosR K182 in vitro. Protein acetylation contributes to the pathogenesis of Mtb Bacteria were cultured and harvested on day 3 to isolate total RNA. Relative transcriptional levels were determined using the 2−ΔΔCt method. The expression of tested genes was normalized to that of 16S rRNA. Values represent the mean ± SD from three independent experiments. P < 0.05.* P<0.01. Student’s t-test mutants (Fig. 4a) as well as an Rv0998 deletion mutant (Fig. 4b). In infected mice, deacetylation of K182 also caused signiﬁcantly lower bacterial loads as well as attenuated lung damage (Fig. 4c–f). Although isolated macrophages and a C57Bl/6 mouse infection model may not represent the real hypoxic situation in vivo, these Yang et al. Emerging Microbes & Infections (2018) 7:34 Page 9 of 14 Fig. 4 Deacetylation of DosR contributes to the intracellular survival and pathogenesis of Mtb. a Deacetylation of K182 inhibited the intracellular survival of Mtb. Primary peritoneal macrophages were harvested as described in Materials and Methods and infected at a multiplicity of infection (MOI) of 1 for 24 h. Infected cells were lysed, and the number of viable intracellular bacteria was determined by serial dilutions and plating on 7H10 agar plates. Values represent the mean ± SD of three independent experiments. b The deletion of Rv0998 inhibited the intracellular survival of Mtb. Peritoneal macrophages were infected at an MOI of 1 or 2 for 24 h. Bacterial colonies were counted to calculate intracellular survival efﬁciency. Values represent the mean ± SD of three independent experiments. c Doses of Mtb H37Rv, the deletion mutant of DosR, and WT or K182R complement mutants were injected into mice. The number of bacteria in the injected doses for the four bacterial groups was determined by serial dilutions and plating on 7H10 agar plates and was found to be similar for all groups. d, e Lung (d) and spleen (e) colony-forming units (CFUs) of C57Bl/6 mice infected for 28 days with Mtb H37Rv, the deletion mutant of dosR, and WT or K182R complement mutants. Female C57Bl/6 mice (6–8 weeks old) were challenged by intraperitoneal (i.p.) injection with 5 × 106 CFU in 100 μL PBS. The left side of the lung tissues from infected mice and the whole spleen were homogenized, diluted, and plated on 7H10 agar plates. Colonies were counted after 4 weeks. f Lung histopathology of C57Bl/6 mice infected with Mtb H37Rv, the deletion mutant of dosR, and WT or K182R complement mutants for 28 days. Protein acetylation contributes to the pathogenesis of Mtb 4 Deacetylation of DosR contributes to the intracellular survival and pathogenesis of Mtb. a Deacetylation of K182 inhibited the Protein acetylation contributes to the pathogenesis of Mtb Lung halves of mice infected for 28 days were ﬁxed and embedded in parafﬁn, and then 5 μm-thick sections were stained with hematoxylin and eosin (HE) by standard methods. The pathology was evaluated by pathologists in a blinded manner. Images are pseudocolored representations at ×200 magniﬁcation. Values represent the mean ± SD from two independent experiments. P < 0.05, P < 0.01, *P < 0.001. Student’s t-test Fig. 4 Deacetylation of DosR contributes to the intracellular survival and pathogenesis of Mtb. a Deacetylation of K182 inhibited the intracellular survival of Mtb. Primary peritoneal macrophages were harvested as described in Materials and Methods and infected at a multiplicity of infection (MOI) of 1 for 24 h. Infected cells were lysed, and the number of viable intracellular bacteria was determined by serial dilutions and plating on 7H10 agar plates. Values represent the mean ± SD of three independent experiments. b The deletion of Rv0998 inhibited the intracellular survival of Mtb. Peritoneal macrophages were infected at an MOI of 1 or 2 for 24 h. Bacterial colonies were counted to calculate intracellular survival efﬁciency. Values represent the mean ± SD of three independent experiments. c Doses of Mtb H37Rv, the deletion mutant of DosR, and WT or K182R complement mutants were injected into mice. The number of bacteria in the injected doses for the four bacterial groups was determined by serial dilutions and plating on 7H10 agar plates and was found to be similar for all groups. d, e Lung (d) and spleen (e) colony-forming units (CFUs) of C57Bl/6 mice infected for 28 days with Mtb H37Rv, the deletion mutant of dosR, and WT or K182R complement mutants. Female C57Bl/6 mice (6–8 weeks old) were challenged by intraperitoneal (i.p.) injection with 5 × 106 CFU in 100 μL PBS. The left side of the lung tissues from infected mice and the whole spleen were homogenized, diluted, and plated on 7H10 agar plates. Colonies were counted after 4 weeks. f Lung histopathology of C57Bl/6 mice infected with Mtb H37Rv, the deletion mutant of dosR, and WT or K182R complement mutants for 28 days. Lung halves of mice infected for 28 days were ﬁxed and embedded in parafﬁn, and then 5 μm-thick sections were stained with hematoxylin and eosin (HE) by standard methods. The pathology was evaluated by pathologists in a blinded manner. Images are pseudocolored representations at ×200 magniﬁcation. Materials and methods infection results suggest that deacetylation of K182 can promote the regulatory effect of DosR on Mtb virulence under different conditions. Studies in more effective infection models should be performed to further conﬁrm our ﬁndings. Protein acetylation contributes to the pathogenesis of Mtb Primary peritoneal macrophages were harvested as described in Materials and Methods and infected at a multiplicity of infection (MOI) of 1 for 24 h Infected cells were lysed and the number of viable intracellular bacteria was determined by serial dilutions and plating Fig. 4 Deacetylation of DosR contributes to the intracellular survival and pathogenesis of Mtb. a Deacetylation of K182 inhibited the intracellular survival of Mtb. Primary peritoneal macrophages were harvested as described in Materials and Methods and infected at a multiplicity of infection (MOI) of 1 for 24 h. Infected cells were lysed, and the number of viable intracellular bacteria was determined by serial dilutions and plating on 7H10 agar plates. Values represent the mean ± SD of three independent experiments. b The deletion of Rv0998 inhibited the intracellular survival of Mtb. Peritoneal macrophages were infected at an MOI of 1 or 2 for 24 h. Bacterial colonies were counted to calculate intracellular survival efﬁciency. Values represent the mean ± SD of three independent experiments. c Doses of Mtb H37Rv, the deletion mutant of DosR, and WT or K182R complement mutants were injected into mice. The number of bacteria in the injected doses for the four bacterial groups was determined by serial dilutions and plating on 7H10 agar plates and was found to be similar for all groups. d, e Lung (d) and spleen (e) colony-forming units (CFUs) of C57Bl/6 mice infected for 28 days with Mtb H37Rv, the deletion mutant of dosR, and WT or K182R complement mutants. Female C57Bl/6 mice (6–8 weeks old) were challenged by intraperitoneal (i.p.) injection with 5 × 106 CFU in 100 μL PBS. The left side of the lung tissues from infected mice and the whole spleen were homogenized, diluted, and plated on 7H10 agar plates. Colonies were counted after 4 weeks. f Lung histopathology of C57Bl/6 mice infected with Mtb H37Rv, the deletion mutant of dosR, and WT or K182R complement mutants for 28 days. Lung halves of mice infected for 28 days were ﬁxed and embedded in parafﬁn, and then 5 μm-thick sections were stained with hematoxylin and eosin (HE) by standard methods. The pathology was evaluated by pathologists in a blinded manner. Images are pseudocolored representations at ×200 magniﬁcation. Values represent the mean ± SD from two independent experiments. P < 0.05, P < 0.01, P < 0.001. Student’s t-test Fig. Protein acetylation contributes to the pathogenesis of Mtb Values represent the mean ± SD from two independent experiments. P < 0.05, P < 0.01, P < 0.001. Student’s t-test Fig. 4 Deacetylation of DosR contributes to the intracellular survival and pathogenesis of Mtb. a Deacetylation of K182 inhibited the intracellular survival of Mtb. Primary peritoneal macrophages were harvested as described in Materials and Methods and infected at a multiplicity of infection (MOI) of 1 for 24 h. Infected cells were lysed, and the number of viable intracellular bacteria was determined by serial dilutions and plating on 7H10 agar plates. Values represent the mean ± SD of three independent experiments. b The deletion of Rv0998 inhibited the intracellular survival of Mtb. Peritoneal macrophages were infected at an MOI of 1 or 2 for 24 h. Bacterial colonies were counted to calculate intracellular survival efﬁciency. Values represent the mean ± SD of three independent experiments. c Doses of Mtb H37Rv, the deletion mutant of DosR, and WT or K182R complement mutants were injected into mice. The number of bacteria in the injected doses for the four bacterial groups was determined by serial dilutions and plating on 7H10 agar plates and was found to be similar for all groups. d, e Lung (d) and spleen (e) colony-forming units (CFUs) of C57Bl/6 mice infected for 28 days with Mtb H37Rv, the deletion mutant of dosR, and WT or K182R complement mutants. Female C57Bl/6 mice (6–8 weeks old) were challenged by intraperitoneal (i.p.) injection with 5 × 106 CFU in 100 μL PBS. The left side of the lung tissues from infected mice and the whole spleen were homogenized, diluted, and plated on 7H10 agar plates. Colonies were counted after 4 weeks. f Lung histopathology of C57Bl/6 mice infected with Mtb H37Rv, the deletion mutant of dosR, and WT or K182R complement mutants for 28 days. Lung halves of mice infected for 28 days were ﬁxed and embedded in parafﬁn, and then 5 μm-thick sections were stained with hematoxylin and eosin (HE) by standard methods. The pathology was evaluated by pathologists in a blinded manner. Images are pseudocolored representations at ×200 magniﬁcation. Values represent the mean ± SD from two independent experiments. P < 0.05, P < 0.01, *P < 0.001. Student’s t-test Fig. 4 Deacetylation of DosR contributes to the intracellular survival and pathogenesis of Mtb. a Deacetylation of K182 inhibited the intracellular survival of Mtb. LC–MS/MS analysis, database search, and bioinformatics analysis The tryptic peptides were dissolved in 0.1% formic acid (solvent A), directly loaded onto a homemade reversed- phase analytical column (15 cm length and 75 μm inner diameter). The gradient consisted of an increase from 6 to 23% solvent B (0.1% formic acid in 98% acetonitrile) over 26 min, 23 to 35% over 8 min, an increase to 80% over 3 min, and then holding at 80% for the last 3 min, all at a constant ﬂow rate of 400 nL/min on an EASY-nLC 1000 UPLC system. The peptides were subjected to an NSI source followed by MS/MS in Q ExactiveTM Plus (Thermo Fisher Scientiﬁc) coupled online to the UPLC. The electrospray voltage applied was 2.0 kV. The m/z scan range was 350–1800 for a full scan, and intact peptides were detected in the Orbitrap at a resolution of 70 000. Peptides were then selected for MS/MS using an NCE setting of 28, and the fragments were detected in the Orbitrap at a resolution of 17 500. A data-dependent procedure alternated between one MS scan followed by 20 MS/MS scans with a 15.0 s dynamic exclusion. Auto- matic gain control was set at 5E4. The ﬁxed ﬁrst mass was set as 100 m/z. Bacterial strains, plasmids, primers, and media All strains, plasmids, and primers used in this study have been described in Supplementary Table S3 and S4. Middlebrook 7H9 broth (Difco/Becton Dickson, Franklin Lakes, NJ) supplemented with 10% ADC (5% bovine serum albumin (BSA), 2% dextrose, and 5% catalase) was used as an enriched medium. Middlebrook 7H10 agar (Difco) plates contained 10% ADC, 1.5% (w/v) agar, and antibiotic supplements as required. The antibiotics used In conclusion, we propose that hypoxia may induce the deacetylation of DosR, which in turn increases its DNA- binding ability to promote the transcription of DosR downstream target genes, allowing Mtb to rapidly adapt to hypoxia (Supplementary Figure S5). Page 10 of 14 Page 10 of 14 Yang et al. Emerging Microbes & Infections (2018) 7:34 were 100 μg/mL ampicillin, 50 μg/mL or 75 μg/mL hygromycin B, and 50 μg/mL kanamycin. to the manufacturer’s protocol for a TMT 6-plex kit (Thermo Fisher Scientiﬁc). To enrich Kac peptides, tryptic peptides dissolved in NETN buffer (100 mM NaCl, 1 mM EDTA, 50 mM Tris-HCl, and 0.5% NP-40, pH 8.0) were incubated with pre-washed antibody beads (PTM Biolabs, Hangzhou, China) at 4 °C overnight with gentle shaking. The beads were washed four times with NETN buffer and twice with ddH2O. Bound peptides were eluted from beads with 0.1% triﬂuoroacetic acid. The eluted fractions were combined and vacuum-dried. The resulting peptides were cleaned with C18 ZipTips (Millipore, Bill- erica, MA) according to the manufacturer’s instructions. Bacterial cultures Experiments performed by Wayne26 in Mtb H37Rv have been described previously. In brief, conical screw-capped Nephelo ﬂasks with 20 mm side arms and ﬂat bases (Wheaton Scientiﬁc Products, Millville, NJ) were used to culture bacteria. Mtb H37Rv was grown in 7H9 broth supplemented with 10% ADC and 0.05% Tween 80 at 37 ° C with shaking to mid-log phase (OD590 ≈0.4, ~2.5 × 108 CFU/mL). For aerobic, fully replicating control cultures, 200 mL of medium was inoculated with 2 mL of the cul- ture (OD590 of 0.4) in the ﬂask and incubated at 37 °C on a magnetic stirrer set to rotate at 180 r.p.m. Simultaneously, 400 mL of medium was inoculated with 4 mL of identical culture in a tightly capped ﬂask, which left 200 mL of air space, corresponding to a 0.5 head space ratio. The tightly capped ﬂask culture was then placed on a tissue culture magnetic stirrer set at 70 r.p.m. and incubated at 37 °C to avoid perturbation of the surface. Cultures of the two ﬂasks were gently tipped into the side arm to read OD590 daily. Samples for the quantitative proteomics assay were taken after 12 days of vigorous aerated culture (Aera), which then entered the platform stage. Similarly, after 12 days of hypoxic culture (Hypo), samples entered anaerobic nonreplicating persistent stage 2. Using 20 mm screw-capped culture tubes, small volumes of cultures under aeration or hypoxia were collected as samples for western blot or real-time (RT)–polymerase chain reaction (PCR) detection. Rv0998 and DosR mutant constructions in Mtb For the puriﬁcation of DosR, pET28a-dosR, pET28a- dosRK182R, and pET28a-dosRK182Q were constructed and veriﬁed by sequencing. All conﬁrmed plasmids were trans- formed individually into E. coli BL21, and recombinant strains were grown in Luria Bertani medium containing 100 μg/mL ampicillin at 37 °C. Isopropyl β-D-1-thiogalactopyr- anoside was added to the culture at a ﬁnal concentration of 0.1 mM when the absorbance at 600 nm reached 0.6–0.8. The culture was continuously incubated for 16 h at 18 °C. Afterwards, cells were harvested by centrifugation and stored at −80 °C. Thawed bacteria were resuspended in lysis buffer (50 mM Tris-HCl and 0.5 M NaCl, pH 8.0) and lysed by sonication. Insoluble material was removed by centrifugation for 30 min at 19 000 × g. The soluble extract was puriﬁed by Ni2+ afﬁnity chromatography (Qiagen, Hilden Germany) according to the manu- facturer’s instructions. Protein purity was monitored by 12% sodium dodecyl sulfate (SDS)–polyacrylamide gel electrophoresis (PAGE) and Coomassie blue staining. The concentration was determined using a BCA protein assay kit (Thermo Fisher Scientiﬁc) with BSA as the standard. Rv0998 and DosR mutant constructions in Mtb Gene knockouts of Rv0998 and DosR were generated using allelic exchange and a specialized transducing phage, phAE87 (provided by Dr. LD Lyu, Shanghai Institutes for Biological Sciences)47. Constructs for allelic exchange were generated by amplifying the upstream and downstream ﬂanking regions of each of the two genes using the primer pairs listed in Supplementary Table S4. Upstream and downstream ﬂanking regions were cloned into the delivery vector pYUB854. Allelic exchange con- structs were incorporated into the phAE87 phage. Pha- gemid DNA was electroporated into electrocompetent M. smegmatis cells, and these were then plated on 7H10 agar plates to obtain plaques at the permissive temperature of 30 °C. Specialized transducing phages were picked and ampliﬁed at 30 °C to generate high titer mycobacter- iophages. The desired phage was transduced into the exponential phase of Mtb H37Rv to delete genes of interest by specialized transduction. The transductants were plated on selective medium consisting of 7H10 agar containing 10% ADC enrichment and 75 μg/mL hygro- mycin and then cultured at 37 °C. Hygromycin-resistant clones were isolated and analyzed by PCR and RT–PCR to conﬁrm the deletion of target genes. Western blotting was used to conﬁrm the deletion of DosR with anti-DosR polyclonal antibody; anti-SigA (Biolegend, San Diego, CA) monoclonal antibody was used as the reference antibody. Protein extraction, trypsin digestion, TMT labeling, and enrichment of acetylated lysine peptides Cultured cells were harvested and washed twice with cold phosphate-buffered saline (PBS); lysed in 8 M urea supplemented with 1 mM dithiothreitol (DTT), 2 mM ethylenediaminetetraacetic acid (EDTA), 1% Protease Inhibitor Cocktail Set III (Calbiochem, San Diego, CA), and 30 mM nicotinamide; and then disrupted by sonica- tion. Soluble proteins were collected after centrifugation at 20 000 × g for 10 min at 4 °C. Protein concentrations were determined with a BCA protein assay kit (Thermo Fisher Scientiﬁc, Waltham, MA) according to the manu- facturer’s instructions. For trypsin digestion, the protein sample was diluted by adding 100 mM triethylammonium bicarbonate buffer (TEAB; Sigma-Aldrich, St. Louis, MI) to a urea concentration <2 M; trypsin was then added at a 1:50 trypsin-to-protein mass ratio for the ﬁrst digestion overnight and at a 1:100 trypsin-to-protein mass ratio for a second 4 h digestion. After trypsin digestion, peptides were desalted using a Strata X C18 SPE column (Phe- nomenex, Torrance, CA) and vacuum-dried. Peptides were reconstituted in 0.5 M TEAB and labeled according The resulting MS/MS data were processed using the Maxquant search engine (v.1.5.2.8). Tandem mass spectra were searched against an Mtb database concatenated with a reverse decoy database. Trypsin/P was speciﬁed as the cleavage enzyme, allowing up to two missing cleavages. The ﬁrst search range was set to 5 p.p.m. for precursor ions; the main search range was set to 5 p.p.m. and 0.02 Da for fragment ions. Carbamidomethyl on Cys was speciﬁed as a ﬁxed modiﬁcation, and oxidation on Met was speciﬁed as a variable modiﬁcation. False discovery rate was adjusted to <1%, and the minimum score for peptides was set to >40. For the quantiﬁcation method, TMT 6-plex was selected. After checking the mass error and length of all identiﬁed peptides that ﬁt both requirements, the relative quantitation was divided into two categories: a quantitative ratio of more than 1.5 was considered to be upregulated, while a quantitative ratio of <0.67 was considered to be downregulated for the Page 11 of 14 Yang et al. Emerging Microbes & Infections (2018) 7:34 bioinformatics analysis. GO annotation of the proteome was performed using the UniProt-GOA Database (http:// www.ebi.ac.uk/GOA/). Proteins were categorized into biological processes and molecular functions according to the GO annotation. The KEGG was utilized to annotate pathways. Rv0998 and DosR mutant constructions in Mtb The complementation plasmid for the dosR and K182R mutant was generated by ligating the PCR product into the shuttle vector, pVV16. The resulting plasmid was transformed into the Δ dosR strain and plated on a 7H10 plate containing kanamycin at 50 μg/mL. Positive inte- grants carrying the required insert were screened by col- ony PCR and validated by RT–PCR. Anti-DosR- and anti-DosR Ac-K182-speciﬁc polyclonal antibody preparation The 6×His-tagged DosR was puriﬁed as previously described46 and used to immunize rabbits for the pro- duction of anti-DosR polyclonal antibody. Simulta- neously, the acetylated Ac-K182 peptide FLAEKTVK(ace) NYVSRL, a 175–188 amino acid of DosR, was coupled with keyhole limpet hemocyanin and used as an antigen to immunize rabbits. The antiserum was collected and puriﬁed with acetylated peptide, and control K182 peptide FLAEKTVKNYVSRL was used to remove nonspeciﬁc antibodies. The sensitivity and speciﬁcity of anti-DosR and anti-DosR Ac-K182 antibodies were evaluated by western or dot blots. Protein extraction, trypsin digestion, TMT labeling, and enrichment of acetylated lysine peptides For functional enrichment analysis, we used Fisher’s exact test to assess the enrichment or depletion (two-tailed test) of speciﬁc annotation terms among members of the resulting protein clusters. Any terms with adjusted p values below 0.05 in any of the clusters were treated as signiﬁcant. template and gene-speciﬁc primers with NdeI and HindIII enzyme-cutting sites (Supplementary Table S4) were cloned into the E. coli/mycobacterium shuttle plasmid, pVV16, and the K-R mutant. Recombinant plasmids were electrically transformed into M. smegmatis, and trans- formants were selected on 7H10 agar containing 10% ADC, 50 µg/mL hygromycin B, and 50 µg/mL kanamycin. Recombinant mycobacteria were further cultured in 7H9 broth supplemented with 10% ADC enrichment, 0.05% Tween 80, 50 µg/mL hygromycin B, and 50 µg/mL kana- mycin. M. smegmatis harboring the vector backbone, pVV16, was included as a control. In vitro acetylation Assays were performed in a 20 μL total reaction volume containing 50 mM Tris-Cl (pH 8.0), 100 mM NaCl, 100 μM acetyl-CoA, 2 μg Rv0998, and 2 μg of DosR or K182 peptide as the substrate in the presence or absence of 200 μM cAMP. Reactions occurred at 25 °C for 10 min and were stopped by boiling in SDS sample buffer. They were then analyzed by western blot with anti-DosR Ac-K182 antibody. Growth curve Mycobacteria were grown to mid-log phase in 7H9 broth with 10% ADC, 0.05% Tween 80, and antibiotics, as required. Growth curves for each strain were determined using a Bioscreen C Growth Curve Instrument (Labsys- tems Oy, Helsinki, Finland) and a honeycomb plate with Construction of plasmids and recombinant M. smegmatis The dosR, phop, sigA, sigB, sigF, and lsr2 genes of Mtb Construction of plasmids and recombinant M. smegmatis The dosR, phop, sigA, sigB, sigF, and lsr2 genes of Mtb were ampliﬁed using Mtb H37Rv genomic DNA as a Page 12 of 14 Page 12 of 14 Page 12 of 14 Yang et al. Emerging Microbes & Infections (2018) 7:34 beads were pelleted at 4 °C, washed ﬁve times with PBS buffer, boiled in SDS sample buffer, and subsequently analyzed by SDS–PAGE and western blot. 100 wells (Labsystems Oy)48. Brieﬂy, 200 μL of each bac- terial suspension, adjusted to a similar density, was added to each well and cultured with shaking at 37 °C. The optical density was measured at an absorbance of 590 nm every 2 h. Hypoxic conditions were established by cover- ing each culture with 50 μL of parafﬁn oil. Cultures were incubated at 37 °C for 7 days for M. smegmatis, and 14 days for Mtb. Two independent experiments were performed, each in triplicate. Western or dot blots Standard western and dot blot procedures were used. Puriﬁed recombinant protein (2 μg) or cell extracts (50 μg) were separated by 12% SDS–PAGE and then transferred to a polyvinylidene ﬂuoride membrane. The concentra- tions of the primary antibodies used in the corresponding blot were anti-DosR (1:1000), anti-DosR Ac-K182 (1:1000), and anti-SigA (1:1000). Horseradish peroxidase- conjugated goat anti-rabbit polyclonal antibody was used as the secondary antibody at a 1:5000 dilution. Quantitative RT–PCR assay Bacteria were cultured as previously described, har- vested, and disrupted by a bead homogenizer. RNA was isolated using Trizol (Invitrogen). Primers for quantitative PCR are listed in Table S4. Samples were run in triplicate and ampliﬁed using real-time SYBR Green quantitative reagents (Toyobo, Osaka, Japan). The relative transcrip- tional level was determined by the 2−ΔΔCt method. The reference gene was 16S rRNA. Intracellular CFUs in peritoneal macrophages Primary peritoneal macrophages were harvested as described previously50. Brieﬂy, C57Bl/6 mice were intra- peritoneally injected with 4% Brewer thioglycollate med- ium (Sigma-Aldrich). After 3 days, the mice were sacriﬁced by cervical dislocation, and the cells were iso- lated by ﬂushing the peritoneal cavity with 10 mL of RPMI 1640 per mouse. Cells were seeded in 24-well dishes, and non-adherent cells were removed by extensive washing with RPMI 1640. The adherent peritoneal macrophages were used for subsequent experiments. Bacteria were diluted to achieve a MOI of 1 or 2 and then incubated with cells for 3 h at 37 °C in 5% CO2. Then, uninfected cells were washed twice, and infected cells were incubated in fresh tissue culture medium containing 50 μg/mL amikacin for 2 h post infection. Infected cells were lysed at 24 h with 0.1% (v/v) Triton X-100 in PBS. The number of viable intracellular bacteria (CFU) was determined by serial dilutions and plating out. Puriﬁcation of Rv0998 Rv0998 were overexpressed and puriﬁed using similar procedure as that for DosR. A plasmid, pET28a-Rv0998/ BL21, was constructed and used to overexpress proteins. The 6× His-tagged Rv0998 was puriﬁed by Ni2+ afﬁnity chromatography, and the amount was estimated by SDS–PAGE and BCA protein assay. Electrophoretic mobility shift assays DNA fragments used for EMSAs were ampliﬁed by PCR using Mtb H37Rv genomic DNA as a template. The pro- moter region of acr was ampliﬁed using the primers acr-F and acr-R, which are listed in Supplementary Table S4. PCR ampliﬁcation yielded fragments of 148 bp. Approxi- mately 0.25 μg of DNA in a 20 μL volume was incubated at room temperature for 10 min with 5 μg of puriﬁed DosR or derivative mutant protein. The binding buffer used for protein–DNA incubations contained 25 mM Tris-HCl (pH 8.0), 20 mM KCl, 0.5 mM EDTA, 6 mM MgCl2, 1 mM DTT, and 1 μg/mL poly I:C. Samples were separated on a 6.5% non-denaturing Tris-glycine poly- acrylamide gel and detected by silver staining, as descri- bed previously49. References Statistical signiﬁcance was determined with an unpaired two-tailed Student’s t-test at a P < 0.05 level of sig- niﬁcance using GraphPad Prism 5.0 software. 1. World Health Organization. Global tuberculosis report 2016. WHO/HTM/TB/ 2016.13 (2016). 2. Kurthkoti, K. et al. The capacity of Mycobacterium tuberculosis to survive iron starvation might enable it to persist in iron-deprived microenvironments of human granulomas. MBio 8, e01092–17 (2017). Received: 6 November 2017 Revised: 10 January 2018 Accepted: 14 January 2018 Received: 6 November 2017 Revised: 10 January 2018 Accepted: 14 January 2018 Mouse infection Previously produced polyclonal antibodies to DosR were used to precipitate speciﬁc DosR proteins. For preclear- ing, cell lysates (500 μg) were incubated with normal rabbit IgG for 1 h at 4 °C. Protein A/G beads (Thermo Fisher Scientiﬁc) were added to lysates and incubated for an additional 30 min. The beads were removed by cen- trifugation, and the precleared supernatant was interacted with DosR antibody at 4 °C for 1 h. Protein A/G beads were added to the lysates and incubated overnight. The Female C57Bl/6 mice (6–8 weeks old) were challenged by i.p. injection with 5 × 106 CFUs per mouse of the indicated bacterial strains as described previously50. Mice were housed in a speciﬁc pathogen-free animal laboratory before being moved into a biosafety level (BSL)-3 laboratory for Mtb infection. Four weeks after infection, the mice were sacriﬁced, and the bacterial loads in the lungs and spleens were determined by CFU counts. Page 13 of 14 Page 13 of 14 Yang et al. Emerging Microbes & Infections (2018) 7:34 Ethical approval Animal procedures were approved by the Shanghai Pulmonary Hospital, and this study was performed in strict accordance with the National Research Council Guide for Care and Use of Laboratory Animals (SYXK2009-0083). All surgeries were performed under sodium pentobarbital anesthesia, and all efforts were made to minimize suffering. The mouse infection study was performed in an ABSL-3 laboratory of the Second Military Medical University. Animal study protocols were also reviewed and approved by the Institutional Animal Care and Use Committee of the Second Military Medical University (Permit Number: ABSL3-003). Data availability g 3. Bhusal, R. P., Bashiri, G., Kwai, B. X. C., Sperry, J. & Leung, I. K. H. Targeting isocitrate lyase for the treatment of latent tuberculosis. Drug Discov. Today 22, 1008–1016 (2017). The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identiﬁer PXD008587. Material requests for bacterial experiments should be addressed to X.C.H. (huangxiao- chen.1001@163.com) or J.W. (wjtwo@163.com). Material requests for proteomics studies should be addressed to H. Y. (yanghua97065@163.com) or W.S. (shfksw@126.com). Material requests for puriﬁed proteins or antibodies should be addressed to Z.H.L. (nllzh@126.com) or F.L. (liufeng@sibs.ac.cn). All other material requests should be addressed to H.Y. (yanghua97065@163.com) or B.X.G. (baoxue_ge@tongji.edu.cn). 4. Peddireddy, V., Doddam, S. N. & Ahmed, N. Mycobacterial dormancy systems and host responses in tuberculosis. Front Immunol. 8, 84 (2017). 4. Peddireddy, V., Doddam, S. N. & Ahmed, N. Mycobacterial dormancy systems and host responses in tuberculosis. Front Immunol. 8, 84 (2017). 5. Hudock, T. A. et al. Hypoxia sensing and persistence genes are expressed during the intragranulomatous survival of Mycobacterium tuberculosis. Am. J. Respir. Cell Mol. Biol. 56, 637–647 (2017). 6. Wayne, L. G. & Hayes, L. G. An in vitro model for sequential study of shiftdown of Mycobacterium tuberculosis through two stages of nonreplicating persis- tence. Infect. Immun. 64, 2062–2069 (1996). 7. Wayne, L. G. & Sohaskey, C. D. Nonreplicating persistence of Mycobacterium tuberculosis. Annu. Rev. Microbiol. 55, 139–163 (2001). 8. Schubert, O. T. et al. Absolute proteome composition and dynamics during dormancy and resuscitation of Mycobacterium tuberculosis. Cell Host Microbe 18, 96–108 (2015). 9. Ouidir, T., Kentache, T. & Hardouin, J. Protein lysine acetylation in bacteria: current state of the art. Proteomics 16, 301–309 (2016). 9. Ouidir, T., Kentache, T. & Hardouin, J. Protein lysine acetylation in bacteria: current state of the art. Proteomics 16, 301–309 (2016). 10. Hu, L. I., Lima, B. P. & Wolfe, A. J. Bacterial protein acetylation: the dawning of a new age. Mol. Microbiol. 77, 15–21 (2010). 10. Hu, L. I., Lima, B. P. & Wolfe, A. J. Bacterial protein acetylation: the dawning of a new age. Mol. Microbiol. 77, 15–21 (2010). The authors declare that they have no conﬂict of interest. The authors declare that they have no conﬂict of interest. The left portions of lung tissues from infected mice and the whole spleen were homogenized separately in 1 mL of PBS. Homogenates in 10-fold serial dilutions were plated on 7H10 agar supplemented with 10% ADC enrichment and incubated at 37 °C. Colonies were counted after 4 weeks. Acknowledgements 11. Zhao, S. et al. Regulation of cellular metabolism by protein lysine acetylation. Science 327, 1000–1004 (2010). 11. Zhao, S. et al. Regulation of cellular metabolism by protein lysine acetylation. Science 327, 1000–1004 (2010). This project was supported by grants from the National Natural Science Foundation of China (grant 81301391 to H.Y.; 81370108 to H.P.L.; 81330069, 91542111, and 31730025 to B.X.G.); the Chinese National Program on Key Basic Research Project (grant 2017YFA0505900 to B.X.G.), and the Chinese National Key Project for Infectious Disease (grant 2017ZX10202302 to W.S.). 12. Zhang, J. et al. Lysine acetylation is a highly abundant and evolutionarily conserved modiﬁcation in Escherichia coli. Mol. Cell Proteomics 8, 215–225 (2009). 12. Zhang, J. et al. Lysine acetylation is a highly abundant and evolutionarily conserved modiﬁcation in Escherichia coli. Mol. Cell Proteomics 8, 215–225 (2009). 13. Schilling, B. et al. Protein acetylation dynamics in response to carbon overﬂow in Escherichia coli. Mol. Microbiol. 98, 847–863 (2015). Authors’ contributions 14. Wu, X. et al. Differential lysine acetylation proﬁles of Erwinia amylovora strains revealed by proteomics. J. Proteomics 79, 60–71 (2013). B.X.G. and H.Y. designed the study. H.Y., W.S., Z.H.L., T.Q.T., L.H.Q., Z.L.C., J.X.C., X. C.H., and J.W. performed the experiments. B.X.G., H.Y., H.P.L., F.L., R.J.Z., and Y.H. F. analyzed the data. B.X.G., H.Y., W.S., Z.H.L., H.P.L., F.L., X.C.H., and J.W. contributed reagents/materials/analysis tools. B.X.G., H.Y., H.P.L., and R.J.Z. drafted the manuscript. All authors contributed to the ﬁnal version of the manuscript. revealed by proteomics. J. Proteomics 79, 60–71 (2013). 15. Kim, D. et al. The acetylproteome of Gram-positive model bacterium Bacillus subtilis. Proteomics 13, 1726–1736 (2013). 16. Wang, Q. et al. Acetylation of metabolic enzymes coordinates carbon source utilization and metabolic ﬂux. Science 327, 1004–1007 (2010). 17. Liu, F. et al. Acetylome analysis reveals diverse functions of lysine acetylation in Mycobacterium tuberculosis. Mol. Cell. Proteomics 13, 3352–3366 (2014). y 18. Xie, L. et al. Proteome-wide lysine acetylation proﬁling of the human patho- gen Mycobacterium tuberculosis. Int. J. Biochem. Cell Biol. 59, 193–202 (2015). 18. Xie, L. et al. Proteome-wide lysine acetylation proﬁling of the human patho- M b t i t b l i I t J Bi h C ll Bi l 59 193 202 (2015) Histopathology analysis Half of each lung was ﬁxed in 4% neutral-buffered paraformaldehyde solution for 24 h. Lung tissue was then embedded within parafﬁn. A series of sections with a thickness of 4–7 μm were then cut and stained with hematoxylin and eosin by standard methods. The pathology was evaluated by pathologists in a blinded manner. Supplementary Information accompanies this paper at (https://doi.org/ 10.1038/s41426-018-0032-2). Supplementary Information accompanies this paper at (https://doi.org/ 10.1038/s41426-018-0032-2). y 10.1038/s41426-018-0032-2). B.X.G. and H.Y. designed the study. H.Y., W.S., Z.H.L., T.Q.T., L.H.Q., Z.L.C., J.X.C., X. C.H., and J.W. performed the experiments. B.X.G., H.Y., H.P.L., F.L., R.J.Z., and Y.H. F. analyzed the data. B.X.G., H.Y., W.S., Z.H.L., H.P.L., F.L., X.C.H., and J.W. contributed reagents/materials/analysis tools. B.X.G., H.Y., H.P.L., and R.J.Z. drafted the manuscript. All authors contributed to the ﬁnal version of the manuscript. Author details 1 J., Kim, J. A., Moon, J. H., Ryu, S. E. & Pan, J. G. The diversity of lysine- acetylated proteins in Escherichia coli. J. Microbiol. Biotechnol. 18, 1529–1536 (2008). 26. Wayne, L. G. In vitro model of hypoxically induced nonreplicating persistence of Mycobacterium tuberculosis. Methods Mol. Med. 54, 247–269 (2001). of Mycobacterium tuberculosis. Methods Mol. Med. 54, 247–269 41. Majumdar, S. D. et al. Appropriate DevR (DosR)-mediated signaling determines transcriptional response, hypoxic viability and virulence of Mycobacterium tuberculosis. PLoS ONE 7, e35847 (2012). y 27. Galagan, J. E. et al. The Mycobacterium tuberculosis regulatory network and hypoxia. Nature 499, 178–183 (2013). 28. Leistikow, R. L. et al. The Mycobacterium tuberculosis DosR regulon assists in metabolic homeostasis and enables rapid recovery from nonrespiring dor- mancy. J. Bacteriol. 192, 1662–1670 (2010). 42. Ren, J. et al. Acetylation of lysine 201 Inhibits the DNA-binding ability of PhoP to regulate salmonella virulence. PLoS Pathog. 12, e1005458 (2016). p mancy. J. Bacteriol. 192, 1662–1670 (2010). 43. Sang, Y. et al. Acetylation regulating protein stability and DNA-binding ability of HilD, thus modulating Salmonella typhimurium virulence. J. Infect. Dis. 216, 1018–1026 (2017). 29. Schubert, O. T. et al. The Mtb proteome library: a resource of assays to quantify the complete proteome of Mycobacterium tuberculosis. Cell Host Microbe 13, 602–612 (2013). 44. Mehra, S. et al. The DosR regulon modulates adaptive immunity and is essential for Mycobacterium tuberculosis persistence. Am. J. Respir. Crit. Care Med. 191, 1185–1196 (2015). 30. Wisedchaisri, G. et al. Structures of Mycobacterium tuberculosis DosR and DosR- DNA complex involved in gene activation during adaptation to hypoxic latency. J. Mol. Biol. 354, 630–641 (2005). 45. Malhotra, V. et al. Disruption of response regulator gene, devR, leads to attenuation in virulence of Mycobacterium tuberculosis. FEMS Microbiol. Lett. 231, 237–245 (2004). 31. Gupta, R. K., Chauhan, S. & Tyagi, J. S. K182G substitution in DevR or C□G mutation in the Dev box impairs protein-DNA interaction and abrogates DevR-mediated gene induction in Mycobacterium tuberculosis. FEBS J. 278, 2131–2139 (2011). 46. Yang, H. et al. Selection and application of peptide mimotopes of MPT64 protein in Mycobacterium tuberculosis. J. Med. Microbiol. 60, 69–74 (2011). 32. Chen, T., He, L., Deng, W. & Xie, J. The mycobacterium DosR regulon structure and diversity revealed by comparative genomic analysis. J. Cell Biochem. 114, 1–6 (2013). 47. Lyu, L. D., Tang, B. K., Fan, X. Y., Ma, H. & Zhao, G. P. Author details 1 1Shanghai Key Laboratory of Tuberculosis, Clinic and Research Center of Tuberculosis, Shanghai Pulmonary Hospital, Tongji University School of Medicine, Shanghai 200433, PR China. 2Department of Microbiology and Immunology, Tongji University School of Medicine, Shanghai 200049, PR China 19. Chen, Z. et al. Lysine acetylome analysis reveals photosystem II manganese- stabilizing protein acetylation is involved in negative regulation of oxygen evolution in model cyanobacterium Synechococcus sp. PCC 7002. Mol. Cell Proteomics 16, 1297–1311 (2017). 19. Chen, Z. et al. Lysine acetylome analysis reveals photosystem II manganese- stabilizing protein acetylation is involved in negative regulation of oxygen evolution in model cyanobacterium Synechococcus sp. PCC 7002. Mol. Cell Proteomics 16, 1297–1311 (2017). Page 14 of 14 Yang et al. Emerging Microbes & Infections (2018) 7:34 20. Ren, J., Sang, Y., Lu, J. & Yao, Y. F. Protein acetylation and its role in bacterial virulence. Trends Microbiol. 25, 768–779 (2017). 36. Converse, P. J. et al. Role of the dosR-dosS two-component regulatory system in Mycobacterium tuberculosis virulence in three animal models. Infect. Immun. 77, 1230–1237 (2009). arabetta, V. J. & Cristea, I. M. Regulation, function, and detection of p 21. Carabetta, V. J. & Cristea, I. M. Regulation, function, and detection of protein acetylation in bacteria. J. Bacteriol. 199, pii: e00107–pii: e00117 (2017). acetylation in bacteria. J. Bacteriol. 199, pii: e00107–pii: e00117 (2017) 37. Du, P., Sohaskey, C. D. & Shi, L. Transcriptional and physiological changes during Mycobacterium tuberculosis reactivation from non-replicating persis- tence. Front. Microbiol. 7, 1346 (2016). 22. You, D. et al. Sirtuin-dependent reversible lysine acetylation of glutamine synthetases reveals an autofeedback loop in nitrogen metabolism. Proc. Natl Acad. Sci. USA 113, 6653–6658 (2016). 38. Chen, Y. et al. Quantitative acetylome analysis reveals the roles of SIRT1 in regulating diverse substrates and cellular pathways. Mol. Cell. Proteomics 11, 1048–1062 (2012). 23. Vergnolle, O. et al. Post-translational acetylation of MbtA modulates myco- bacterial siderophore biosynthesis. J. Biol. Chem. 291, 22315–22326 (2016). 24. Nambi, S., Basu, N. & Visweswariah, S. S. cAMP-regulated protein lysine acet- ylases in mycobacteria. J. Biol. Chem. 285, 24313–24323 (2010). 39. Zhang, Z. et al. Interconversion of peptide mass spectral libraries derivatized with iTRAQ or TMT Labels. J. Proteome Res. 15, 3180–3187 (2016). ylases in mycobacteria. J. Biol. Chem. 285, 24313–24323 (2010). 25. Bi, J. et al. Modulation of central carbon metabolism by acetylation of isocitrate lyase in Mycobacterium tuberculosis. Sci. Rep. 7, 44826 (2017). 40. Yu, B. Yang et al. Emerging Microbes & Infections (2018) 7:34 50. Wang, Y. et al. Long noncoding RNA derived from CD244 signaling epigen- etically controls CD8 + T-cell immune responses in tuberculosis infection. Proc. Natl Acad. Sci. USA 112, E3883–E3892 (2015). Author details 1 Mycobacterial MazG safeguards genetic stability via housecleaning of 5-OH-dCTP. PLoS Pathog. 9, e1003814 (2013). 33. Starai, V. J. & Escalante-Semerena, J. C. Identiﬁcation of the protein acetyl- transferase (Pat) enzyme that acetylates acetyl-CoA synthetase in Salmonella enterica. J. Mol. Biol. 340, 1005–1012 (2004). 48. Liu, Z. et al. Impact of hypoxia on drug resistance and growth characteristics of Mycobacterium tuberculosis clinical isolates. PLoS One 11, e0166052 (2016). 49. Park, H. D. et al. Rv3133c/dosR is a transcription factor that mediates the hypoxic response of Mycobacterium tuberculosis. Mol. Microbiol. 48, 833–843 (2003). 34. Lee, H. J., Lang, P. T., Fortune, S. M., Sassetti, C. M. & Alber, T. Cyclic AMP regulation of protein lysine acetylation in Mycobacterium tuberculosis. Nat. Struct. Mol. Biol. 19, 811–818 (2012). 50. Wang, Y. et al. Long noncoding RNA derived from CD244 signaling epigen- etically controls CD8 + T-cell immune responses in tuberculosis infection. Proc. Natl Acad. Sci. USA 112, E3883–E3892 (2015). 35. Parish, T. et al. Deletion of two-component regulatory systems increases the virulence of Mycobacterium tuberculosis. Infect. Immun. 71, 1134–1140 (2003). 35. Parish, T. et al. Deletion of two-component regulatory systems increases the virulence of Mycobacterium tuberculosis. Infect. Immun. 71, 1134–1140 (2003).
https://openalex.org/W3011944654	https://digital.csic.es/bitstream/10261/205625/1/Finding_Massonis%20.pdf	English	null	Finding and Breaking Lie Symmetries: Implications for Structural Identifiability and Observability in Biological Modelling	Symmetry	2,020	cc-by	10,336	Article Finding and Breaking Lie Symmetries: Implications for Structural Identiﬁability and Observability in Biological Modelling Gemma Massonis and Alejandro F. Villaverde * Gemma Massonis and Alejandro F. Villaverde * BioProcess Engineering Group, IIM-CSIC, 36208 Vigo, Galicia, Spain; gmassonis@gmail.com * Correspondence: afvillaverde@iim.csic.es BioProcess Engineering Group, IIM-CSIC, 36208 Vigo, Galicia, Spain; gmassonis@gmail.com * Correspondence: afvillaverde@iim.csic.es BioProcess Engineering Group, IIM-CSIC, 36208 Vigo, Galicia, Spain; gmassonis@gmail.com * Correspondence: afvillaverde@iim.csic.es Received: 27 February 2020; Accepted: 14 March 2020; Published: 16 March 2020 Received: 27 February 2020; Accepted: 14 March 2020; Published: 16 March 2020 Abstract: A dynamic model is structurally identiﬁable (respectively, observable) if it is theoretically possible to infer its unknown parameters (respectively, states) by observing its output over time. The two properties, structural identiﬁability and observability, are completely determined by the model equations. Their analysis is of interest for modellers because it informs about the possibility of gaining insight into a model’s unmeasured variables. Here we cast the problem of analysing structural identiﬁability and observability as that of ﬁnding Lie symmetries. We build on previous results that showed that structural unidentiﬁability amounts to the existence of Lie symmetries. We consider nonlinear models described by ordinary differential equations and restrict ourselves to rational functions. We revisit a method for ﬁnding symmetries by transforming rational expressions into linear systems. We extend the method by enabling it to provide symmetry-breaking transformations, which allows for a semi-automatic model reformulation that renders a non-observable model observable. We provide a MATLAB implementation of the methodology as part of the STRIKE-GOLDD toolbox for observability and identiﬁability analysis. We illustrate the use of the methodology in the context of biological modelling by applying it to a set of problems taken from the literature. Keywords: dynamic modelling; nonlinear systems; observability; structural identiﬁability; Lie symmetries Symmetry 2020, 12, 469; doi:10.3390/sym12030469 symmetry S S symmetry S S 1. Introduction The present work is motivated mainly by problems arising in identiﬁcation and modelling of biological systems, although its results are applicable in other ﬁelds. We consider nonlinear dynamic models deﬁned by ordinary differential equations. This framework is sufﬁciently powerful to model a wide range of biological processes, from intracellular networks to whole ecosystems, with the appropriate level of detail. When compared to other applications, biological models often pose speciﬁc challenges due to the combination of nonlinearity, uncertainty about the underlying system, and experimental limitations regarding the possibility of perturbations and of measurements [1]. These features make the identiﬁcation of biological models particularly challenging, and call for new methodological developments and computational tools. Indeed, many theoretical advances in nonlinear systems identiﬁcation have been motivated by biological problems, even though the type of problems being considered are often common to other scientiﬁc areas, which makes the resulting methodologies generally applicable [2,3]. An example is the study of structural identiﬁability, a property that was introduced in the context of physiological modelling [4] and has since then been adopted in many scientiﬁc and technological areas; since its study is of particular interest for biological models, many related theoretical developments have been motivated by biological problems [5–7]. Structural Symmetry 2020, 12, 469; doi:10.3390/sym12030469 www.mdpi.com/journal/symmetry 2 of 20 Symmetry 2020, 12, 469 identiﬁability refers to the possibility of inferring the values of the unknown (constant) parameters in the model equations from observations (measurements) of the model output. It is closely related to—and in fact can be considered a particular case of—another property, observability, which describes the possibility of determining the (time-varying) state variables of the model [8]. By considering the parameters as constant state variables, structural (local) identiﬁability amounts to observability (In accordance with the usual terminology, we speak of structural identiﬁability to distinguish it from practical identiﬁability. The latter quantiﬁes the uncertainty in parameter values taking into account the information content of the available data, which may be limited due to experimental errors or low sampling rates. We also specify that it is structural local identiﬁability to distinguish it from structural global identiﬁability. Although observability is also a structural local property, for historical reasons we do not add those adjectives to its name and refer to it simply as observability) we do not add those adjectives to its name, and refer to it simply as observability.). 1. Introduction The cause of lack of structural identiﬁability or observability can be traced back to the existence of symmetries in the model equations. In the present work we use Lie’s theory on symmetry analysis of differential equations [9–11]. It has been shown [12] that the existence of Lie symmetries is equivalent to the existence of similarity transformations, i.e., transformations of parameters and variables that leave the output(s) unchanged [13,14]. This means that the existence of symmetries amounts to lack of structural identiﬁability and/or observability. Here we revisit a method for ﬁnding symmetries by transforming rational expressions into linear systems [15]. Determining the existence of symmetries in a model is a way of analysing its structural identiﬁability and observability. Furthermore, if symmetries exist, their mathematical expressions provide information about the relationships between model variables that cause loss of identiﬁability and/or observability. One way of exploiting these insights is by ﬁxing one or more parameters involved in a symmetry, in order to render the remaining ones identiﬁable. Another way is by using the symmetry-breaking transformations to reformulate the model, applying the transformations so that the symmetries disappear and the new model is identiﬁable and observable. In this paper, we extend the method by enabling it to provide symmetry-breaking transformations, which allows for a semi-automatic model reformulation that renders a non-observable model observable. Thus, the approach can be used not only for characterizing the identiﬁability and observability of a model, but also for suggesting alternative reformulations if the original model does not possess those properties. We illustrate the usefulness and applicability of this approach in biological applications with four models of biochemical processes. Furthermore, we provide an implementation of the methodology as part of a new version of the STRIKE-GOLDD software https://github.com/afvillaverde/strike-goldd_ 2.1. STRIKE-GOLDD is a MATLAB toolbox that analyses structural identiﬁability and observability using a differential geometry approach [16–18]. The organization of this paper is as follows: we begin by providing an overview of the methodological aspects in Section 2. Then we illustrate its application to a number of modelling problems in Section 3. Finally, we discuss the implications and provide some conclusions in Section 4. . Methods 2. Methods Structural Identiﬁability and Observability 2.1. Structural Identiﬁability and Observability 2.1. Structural Identiﬁability and Observability 2.1. Structural Identiﬁability and Observability The same deﬁnition applies to the unknown entries vector but instead of observable, it is called reconstructible. A model is observable if all states are observable. The same deﬁnition applies to the unknown entries vector but instead of observable, it is called reconstructible. The term Full Input-State-Parameter Observability (FISPO) has been recently proposed [8] to refer to a model that is observable, identiﬁable and reconstructible. Its formal deﬁnition is given by: The term Full Input-State-Parameter Observability (FISPO) has been recently proposed [8] to refer to a model that is observable, identiﬁable and reconstructible. Its formal deﬁnition is given by: Deﬁnition 3. Given the model (1), the unknown quantities vector is considered z(t) = [x(t), θ, w(t)] ∈ Rm+q+mw. Denoting each component of z in time τ as zi(τ) with t0 ≤τ ≤t ≤t f , for a ﬁnite t f . The model MNL is FISPO if the following condition holds: ˆz(τ) ∈N (z∗(τ)) and y(t, ˆz(τ)) = y(t, z∗(τ)) ⇒ˆzi(τ) = z∗ i (τ) . (3) (3) Observability and identiﬁability can be studied jointly if the parameters are considered as constant states. Thus, the observability of these states is equivalent to the local structural identiﬁability. Therefore, the augmented state vector is [18]: ex(t) =   x(t) θ w(t)  , ˙ex(t) =   f (ex(t), u(t)) 0 ˙w(t)  . (4) (4) In this way, the model (1) is transformed into: In this way, the model (1) is transformed into: In this way, the model (1) is transformed into: MNL := ( ˙ex(t) = f (ex(t), u(t)), y(t) = g(ex(t), u(t)) . (5) (5) To determine the observability of a model it is necessary to calculate the nonlinear observability matrix ONL using Lie derivatives. In the case of time-dependent entries, i.e., u(t), Lie derivatives are computed as follows: Deﬁnition 4. The ﬁrst order Lie derivative of g(ex) with respect to f (ex) is: Deﬁnition 4. The ﬁrst order Lie derivative of g(ex) with respect to f (ex) is: L f g(ex) = ∂g(ex) ∂ex f (ex, u) + ∞ ∑ j=0 ∂g(ex) ∂u(j) u(j+1) , (6) (6) where u(j) is the jth derivative of u. 2.1. Structural Identiﬁability and Observability For the study of observability and identiﬁability we consider the following type of models: For the study of observability and identiﬁability we consider the following type of models: MNL :=        ˙x(t) = f (x(t), θ, u(t), w(t)) , y(t) = g(x(t), θ, u(t), w(t)) , x(t0) = x0(θ) (1) (1) with a parameter vector θ ∈Rq, known input functions u(t) ∈Rmu, unknown input functions w(t) ∈Rmw, state vector x(t) ∈Rm, output vector y(t) ∈Rn and, f and g vectors constituted by analytical functions. From this point on, the time dependency will be omitted to simplify the notation. with a parameter vector θ ∈Rq, known input functions u(t) ∈Rmu, unknown input functions w(t) ∈Rmw, state vector x(t) ∈Rm, output vector y(t) ∈Rn and, f and g vectors constituted by analytical functions. From this point on, the time dependency will be omitted to simplify the notation. Symmetry 2020, 12, 469 3 of 20 Deﬁnition 1. A parameter θi, denoting the ith component, of the MNL model is structurally locally identiﬁable (SLI) if for almost any parameter vector θ∗∈Rq there is a neighbourhood N (θ∗) in which the following relation holds: Deﬁnition 1. A parameter θi, denoting the ith component, of the MNL model is structurally locally identiﬁable (SLI) if for almost any parameter vector θ∗∈Rq there is a neighbourhood N (θ∗) in which the following relation holds: ˆθ ∈N (θ∗) and y(t, ˆθ) = y(t, θ∗) ⇒ˆθi = θ∗ i , i = 1, ..., q . (2 (2) If the above deﬁnition does not hold for any neighbourhood of θ∗, θi is structurally unidentiﬁable (SU). If all model parameters are SLI, the model is SLI too. Otherwise, the model is SU. If the above deﬁnition does not hold for any neighbourhood of θ∗, θi is structurally unidentiﬁable (SU). If all model parameters are SLI, the model is SLI too. Otherwise, the model is SU. Similarly, the observability of a state can be deﬁned as: Deﬁnition 2. Given an output vector y(t) and a known input vector u(t), the state xi(τ) is observable if it can be determined from y(t) and u(t) in the interval t0 ≤τ ≤t ≤t f , where t f is a ﬁnite time. Otherwise, it is unobservable. A model is observable if all states are observable. 2.1. Structural Identiﬁability and Observability For higher orders, i ≥2, the calculation is done through a recursive procedure: For higher orders, i ≥2, the calculation is done through a recursive procedure: For higher orders, i ≥2, the calculation is done through a recursive procedure: Li f g(ex) = ∂Li−1 f g(ex) ∂ex f (ex, u) + ∞ ∑ j=0 ∂Li−1 f g(ex) ∂u(j) u(j+1) . (7) (7) 4 of 20 2, 469 4 of 20 Symmetry 2020, 12, 469 Symmetry 2020, 12, 469 The observability-identiﬁability matrix of the previous model is deﬁned as: The observability-identiﬁability matrix of the previous model is deﬁned as: OI(ex, u) =                   ∂ ∂ex g(ex, u) ∂ ∂ex(L f g(ex, u)) ∂ ∂ex(L2 f g(ex, u)) ... ∂ ∂ex(Lnex−1 f g(ex, u))                   . (8) OI(ex, u) =                   ∂ ∂ex g(ex, u) ∂ ∂ex(L f g(ex, u)) ∂ ∂ex(L2 f g(ex, u)) ... ∂ ∂ex(Lnex−1 f g(ex, u))                   . (8) Theorem 1. Nonlinear Observability-Identiﬁability Condition (OIC). If the model (5) satisﬁes rank(OI(ex0, u)) = m + q + mw, with OI(ex0, u) given by (8), where ex0 is a point in the augmented state space, then the model is locally observable and locally structurally identiﬁable in a neighbourhood N (ex0) of ex0 [18]. (8) Theorem 1. Nonlinear Observability-Identiﬁability Condition (OIC). If the model (5) satisﬁes rank(OI(ex0, u)) = m + q + mw, with OI(ex0, u) given by (8), where ex0 is a point in the augmented state space, then the model is locally observable and locally structurally identiﬁable in a neighbourhood N (ex0) of ex0 [18]. Theorem 1. Nonlinear Observability-Identiﬁability Condition (OIC). If the model (5) satisﬁes rank(OI(ex0, u)) = m + q + mw, with OI(ex0, u) given by (8), where ex0 is a point in the augmented state space, then the model is locally observable and locally structurally identiﬁable in a neighbourhood N (ex0) of ex0 [18]. If the OIC is not fulﬁlled, there is at least one unobservable (respectively, unidentiﬁable) state (respectively, parameter). 2.1. Structural Identiﬁability and Observability In that case, full observability may sometimes be achieved by measuring more variables or functions. However, sometimes it is not possible to perform more measurements due to the characteristics of the experiments. 2.2. Lie Symmetries and Structural Unidentiﬁability For a more detailed background, we refer the reader to [9,11]. 2.2. Lie Symmetries and Structural Unidentiﬁability (15) dx∗ dτ = η(x∗) , (14) x∗ x when τ 0 (15) ﬁ dx∗ dτ = η(x∗) , (14) x∗= x when τ = 0 . (15) (14) (15) (14) (15) x∗= x when τ = 0 . The Lie group of transformations (9) is equivalent to the above IVP with the parametrization τ(ε): The Lie group of transformations (9) is equivalent to the above IVP with the parametrization τ(ε): τ(ε) = Z ε 0 Γ(ε′)dε′ , τ(ε) = Z ε 0 Γ(ε′)dε′ , where where Γ(ε) = ∂φ(a, b) ∂b \|(a,b)=(ε−1, ε) , Γ(0) = 1 . Γ(ε) = ∂φ(a, b) ∂b \|(a,b)=(ε−1, ε) , Γ(0) = 1 . φ(a, b) is the law of composition and ε−1 denotes the inverse element to ε. φ(a, b) is the law of composition and ε−1 denotes the inverse element to ε. φ(a, b) is the law of composition and ε−1 denotes the inverse element to ε. φ(a, b) is the law of composition and ε−1 denotes the inverse element to ε. The above theorem shows how inﬁnitesimal transformations contain the essential information that determines the uniparametric Lie group of transformations. From Theorem (2) and without loss of generality, it is assumed that the Lie group of transformations has as its law of composition φ(a, b) = a + b with ε−1 = −ε and Γ(ε) = 1 [9]. Therefore, the uniparametric Lie group of transformations (9) is rewritten, in terms of its inﬁnitesimals η(x∗), as: dx∗ dε = η(x∗) , x∗= x at ε = 0 . (16) (16) The above expression deﬁnes an Initial Value Problem (IVP) for the Lie group of transformations in terms of its inﬁnitesimals generators. The exponential parametrization of the Lie group around the identity is: Theorem 3 ([9]). The one-parameter Lie group of transformations (9) is equivalent to: x∗= exp[εX]x = x + εXx + 1 2ε2X2x + ... = (17) = 1 + εX + 1 2ε2X2 + ... x = ∞ ∑ k=0 εk k! Xkx , where X is given by (12) and Xk = XXk−1, k = 1, 2... with X0x = x. (17) where X is given by (12) and Xk = XXk−1, k = 1, 2... with X0x = x. For a more detailed background, we refer the reader to [9,11]. For a more detailed background, we refer the reader to [9,11]. 2.2. Lie Symmetries and Structural Unidentiﬁability Yates et al. [12] showed that the existence of Lie symmetries entails the existence of similarity transformations, and therefore denotes lack of structural identifiability [13,14]. Similarity transformations allow the existence of parameters and variables transformations that leave the output(s) unchanged. In Lie algebra, similarity transformations are one-parameter Lie group morphisms that map a solution of the differential equation onto themselves in terms of state variables. There are an inﬁnite number of ways to represent this morphism, however the representation is unique when independent variables are ﬁxed. The uniqueness of representation is a key property for the purpose of implementing the algorithm. The expression of the one-parameter Lie group of transformations is: The expression of the one-parameter Lie group of transformations is: x∗= X(x; ε) . (9) (9) x∗= X(x; ε) . (9) Expanding the above expression in some neighborhood of ε = 0 : Expanding the above expression in some neighborhood of ε = 0 : x∗= x + ε ∂X(x; ε) ∂ε \|ε=0 + 1 2ε2 ∂2X(x; ε) ∂ε2 \|ε=0 + ... = x + ε ∂X(x; ε) ∂ε \|ε=0 + O(ε2) , (10) (10) where where η(x) = ∂X(x; ε) ∂ε \|ε=0 (11) (11) is the inﬁnitesimal of (9) and x + εη(x) is the inﬁnitesimal transformation of the Lie group of transformations. Deﬁnition 5 ([9]). The inﬁnitesimal generator of the one-parameter Lie group of transformations is the differential operator: X = X(x) = η(x) · ▽= n ∑ i=1 ηi(x) ∂ ∂xi , (12) (12) Symmetry 2020, 12, 469 5 of 20 Symmetry 2020, 12, 469 5 of 20 where ▽deﬁnes the gradient ▽= ∂ ∂x1 , ∂ ∂x2 , ..., ∂ ∂xn . (13) where ▽deﬁnes the gradient ∂ ∂ ∂ (13) where ▽deﬁnes the gradient ▽= ∂ ∂x1 , ∂ ∂x2 , ..., ∂ ∂xn . (13) ▽= ∂ ∂x1 , ∂ ∂x2 , ..., ∂ ∂xn . (13) (13) Theorem 2. First Fundamental Theorem of Lie [9]. Given an initial value problem (IVP) for a system of ﬁrst-order ODEs: Theorem 2. First Fundamental Theorem of Lie [9]. Given an initial value problem (IVP) for a system of ﬁrst-order ODEs: dx∗ dτ = η(x∗) , (14) x∗= x when τ = 0 . • Partially variate: ηi(x) = \|d\|=dmax ∑ di,dm+1,...,dm+q=0 ri,dxdi i xdm+1 m+1 · · · x dm+q m+q , i = 1, ..., m , ηi(x) = \|d\|=dmax ∑ dm+1,...,dm+q=0 ri,dxdm+1 m+1 · · · x dm+q m+q , i = m + 1, ..., m + q , ηi(x) = \|d\|=dmax ∑ di,dm+1,...,dm+q=0 ri,dxdi i xdm+1 m+1 · · · x dm+q m+q , i = m + q + 1, ..., n∗. (19) ηi(x) = \|d\|=dmax ∑ di,dm+1,...,dm+q=0 ri,dxdi i xdm+1 m+1 · · · x dm+q m+q , i = 1, ..., m , ηi(x) = \|d\|=dmax ∑ dm+1,...,dm+q=0 ri,dxdm+1 m+1 · · · x dm+q m+q , i = m + 1, ..., m + q , (19) (19) ηi(x) = \|d\| dmax ∑ di,dm+1,...,dm+q=0 ri,dxdi i xdm+1 m+1 · · · x dm+q m+q , i = m + q + 1, ..., n∗. • Multivariate: • Multivariate: where n∗= m + q + mw. We will study three different inﬁnitesimals. Considering dmax, d ∈N as the maximum degree for polynomials and ri,d unknown constants to determine, the expressions of inﬁnitesimals are: • Univariate: ηi(x) = dmax ∑ d=0 ri,dxd i , i = 1, ..., n∗. (18) (18) • Partially variate: • Partially variate: 2.2.1. Computing Symmetries Let us consider the same ODE system as in (1). The state vector will be augmented with parameters and unknown inputs, as mentioned earlier: ˙xi(t) = fi(x(t), u(t)), i = 1, ..., m xi(t) = θ, i = m + 1, ..., m + q xi(t) = wi(t), i = m + q + 1, ..., n∗ xi(t) = wi(t), i = m + q + 1, ..., n∗ 6 of 20 Symmetry 2020, 12, 469 where n∗= m + q + mw. where n∗= m + q + mw. • Multivariate: ηi(x) = \|d\|=dmax ∑ d1,...,dm+q=0 ri,dxd1 1 · · · x dm+q m+q , i = 1, ..., m , ηi(x) = \|d\|=dmax ∑ dm+1,...,dm+q=0 ri,dxdm+1 m+1 · · · x dm+q m+q , i = m + 1, ..., m + q , ηi(x) = \|d\|=dmax ∑ d1,...,dn∗=0 ri,dxd1 1 · · · xdn∗ n∗, i = 1, ..., n∗. (20) ηi(x) = \|d\|=dmax ∑ d1,...,dm+q=0 ri,dxd1 1 · · · x dm+q m+q , i = 1, ..., m , (20) The derivative of inﬁnitesimal generators is also deﬁned, so that it can act on the ˙x(t): X′ = n∗ ∑ i=1 ηi(x) ∂ ∂xi + n∗ ∑ i=1 η′ i(x) ∂ ∂˙xi , (21) (21) where η′ i(x) = n∗ ∑ j=1 ˙xj ∂ηi ∂xj . Using the above formulation for inﬁnitesimals generators, the following explicit criterion for admittance of a Lie group of transformations is obtained: Theorem 4 ([9,10]). The system of ordinary differential equations admits a one-parameter Lie group of transformations deﬁned by the inﬁnitesimal generator (12) if and only if: X′ · ( ˙xk −fk(x)) = 0, k = 1, ..., m (22) X · (yl −gl(x)) = 0, l = 1, ..., n (23) X′ · ( ˙xk −fk(x)) = 0, k = 1, ..., m (22) (22) X (xk fk(x)) 0, k 1, ..., m (22) X · (yl −gl(x)) = 0, l = 1, ..., n (23) X · (yl −gl(x)) = 0, l = 1, ..., n X · (yl −gl(x)) = 0, l = 1, ..., n (23) (23) Symmetry 2020, 12, 469 7 of 20 Applying the previous theorem to the initial system (1), we obtain an explicit criterion: Applying the previous theorem to the initial system (1), we obtain an explicit criterion: n∗ ∑ j=1 ˙xj ∂ηk ∂xj (x) − n∗ ∑ i=1 ηi(x)∂fk ∂xi (x) =0, k = 1, ..., m n∗ ∑ i=1 ηi(x)∂gl ∂xi (x) =0, l = 1, ..., n (24) (24) An admitted Lie symmetry is a continuous group of transformations X such that the observation (observed datum) is unchanged: g(x∗(t), u∗(t)) = g(x(t), u(t)) The output map should not be modiﬁed. The above system deﬁnes a system of partial differential equations. • Multivariate: If we consider the rational form of fk(x) and gl(x): ˙xk = fk(x) = Pk(x) Qk(x), k = 1, ..., m , yl = gl(x) = Rl(x) Sl(x) , l = 1, ..., n , (25) (25) then, the system of PDEs (24) is converted into a system of ODEs. then, the system of PDEs (24) is converted into a system of ODEs. then, the system of PDEs (24) is converted into a system of ODEs. There are different expressions of (24) depending on the inﬁnitesimal generator: There are different expressions of (24) depending on the inﬁnitesimal generator: U i i d i l • Univariate and partial: • Univariate and partial: PkQk ∂ηk ∂xk − n∗ ∑ i=1 ηi[Pk xiQk −PkQk xi] =0, k = 1, ..., m , n∗ ∑ i=1 ηi[Rl xiSl −RlSl xi] =0, l = 1, ..., n . (26) (26) • Multivariate: m ∑ j=1 PjQk ∏ b̸=j Qb ! ∂ηk ∂xj − n∗ ∑ i=1 ηi ∏ b̸=k Qb ! [Pk xiQk −PkQk xi] =0, k = 1, ..., m , n∗ ∑ i=1 ηi[Rl xiSl −RlSl xi] =0, l = 1, ..., n . (27) (27) Each of these equations can be reordered based on the combinations among the components of x. Let r be a vector containing all ri,d: Each of these equations can be reordered based on the combinations among the components of x. Let r be a vector containing all ri,d: ∑ i1,...,in ci1,...,in(r)xi1 1 · · · xinn = 0 . (28) (28) The coefﬁcients ci1,...,in are linear in r and its matrix form expression allows to reformulate condition (26) and (27) into: C 0 The coefﬁcients ci1,...,in are linear in r and its matrix form expression allows to reformulate condition (26) and (27) into: C · r = 0 . C · r = 0 . The problem of ﬁnding symmetries is equivalent to solving a linear system of equations with numeric entries. However, C is a non-square matrix and, in order to ﬁnd all the solutions, it is necessary to compute its kernel. It is possible that the obtained solutions are not independent of each other, as a result of linear combinations or multiplication by xi; for this reason, we will only consider solutions that are independent of each other and in their greatest degree of simpliﬁcation. • Multivariate: Symmetry 2020, 12, 469 8 of 20 The next step is to build the expression of x∗with the inﬁnitesimal generators as in Theorem (3). When the inﬁnitesimal transformation is given by powers of the variable, the exact transformation is known and it is classiﬁed as “elementary”. Some examples are: x∗ i = xi + ε, X = ∂ ∂xi (translation) x∗ i = exp(ε)xi, X = xi ∂ ∂xi (scaling) x∗ i = xi 1 −εxi , X = x2 i ∂ ∂xi (Mobius) x∗ i = xi [1 −(p −1)εxp−1 i ] 1 p−1 , X = xp i ∂ ∂xi (higher order) (29) (29) The most common types of symmetries in biochemical models are translation and scaling. However, those elementary transformations cover only a part of the possible symmetries that a model can contain. The others must be calculated using Lie series or solving the IVP (16). It is possible to maximize the number of elementary transformations given by an inﬁnitesimal generator before applying Lie series. This process starts by searching for all parameter combinations that forbid an elementary transformation, and dividing the initial inﬁnitesimal generator by them. Each new inﬁnitesimal generator provides a number of elementary transformations that can be greater than, lower than or equal to the initial one. Maximizing the total amount, we will obtain an inﬁnitesimal generator that provides the largest possible number of elementary transformations. 2.2.2. Initial Conditions The deﬁnition of a dynamic model may include speciﬁc initial conditions (ICS), which can be numeric, parametric (known or unknown) or a combination of both [19,20]. Perturbations in ICS must produce changes in the output y(t) for the model to be observable. If the symmetries are studied without taking into account the ICS, it may happen that the generators do not fulﬁll them. It is important to consider only the generators that satisfy the ICS [19]; by including them as output vectors, a symmetry is admitted by the system if and only if: X · (xk(t0) −x0(θ))\|x=x0(θ) = 0, k = 1, ..., m , (30) (30) wherex0(θ) are the ICS (parametric, numerical or a combination of both). Expanding the above equation: wherex0(θ) are the ICS (parametric, numerical or a combination of both). Expanding the above equation: n∗ ∑ i=1 ηi ∂ ∂xi xk(t0)\|x=x0(θ) − n∗ ∑ i=1 ηi ∂ ∂xi x0(θ)\|x=x0(θ) = 0, k = 1, ..., m . (31) (31) Considering the rational form of x0(θ): Considering the rational form of x0(θ): x0(θ) = Vk(x) Wk(x), k = 1, ..., m . Expression (31) is reformulated as follow: Expression (31) is reformulated as follow: n∗ ∑ i=1 ηi(x0(θ)) − n∗ ∑ i=1 ηi Vk xiWk −VkWk xi (Wk)2 \|x=x0(θ) = 0, k = 1, ..., m . (32) (32) This new restriction allows us to consider only those symmetries that fulﬁll the initial conditions, and provides a tool for study the inﬂuence of them. This new restriction allows us to consider only those symmetries that fulﬁll the initial conditions, and provides a tool for study the inﬂuence of them. Symmetry 2020, 12, 469 9 of 20 9 of 20 2.3. Implementation An overview of the algorithm described in the preceding subsections is shown in Figure 1. We have made a MATLAB implementation available at https://github.com/GemmaMasFes6/Lie-Symmetries, and as part of a new version (v2.1.6) of the STRIKE-GOLDD toolbox (https://github.com/afvillaverde/ strike-goldd_2.1); it will also be included in future releases of said toolbox. Our code represents an addition to the set of existing tools for studying differential equation symmetries, which include symmetryDetection [15] in Python, MinimalOutputSets [19] in Mathematica, and SADE [21] in Maple. Our software is open source and, to the best of our knowledge, it is the ﬁrst tool of its characteristics available in MATLAB. available in MATLAB. Figure 1. Diagram of the algorithm. The code consists of a main MATLAB script, ‘Lie_Symmetry’, and ten auxiliary functions deﬁned in separate ﬁles. Each of these functions performs one of the stages outlined above: calculation of Figure 1. Diagram of the algorithm. The code consists of a main MATLAB script, ‘Lie_Symmetry’, and ten auxiliary functions deﬁned in separate ﬁles. Each of these functions performs one of the stages outlined above: calculation of Figure 1. Diagram of the algorithm. Figure 1. Diagram of the algorithm. The code consists of a main MATLAB script, ‘Lie_Symmetry’, and ten auxiliary functions deﬁned in separate ﬁles. Each of these functions performs one of the stages outlined above: calculation of The code consists of a main MATLAB script, ‘Lie_Symmetry’, and ten auxiliary functions deﬁned in separate ﬁles. Each of these functions performs one of the stages outlined above: calculation of Symmetry 2020, 12, 469 Symmetry 2020, 12, 469 10 of 20 inﬁnitesimal polynomials (univariate, partially variate and multivariate); computation of polynomials for states (depending on the type of polynomial there are two possibilities), observations, and ICS (if they are speciﬁed); and obtaining the transformations. The last step in turn incorporates two other functions, corresponding to maximizing the number of elementary transformations and, if necessary, calculating the transformations using Lie series. The algorithm has a number of options that can be speciﬁed by the user: the type of inﬁnitesima ynomial, its degree, and the number of terms in Lie series (in case it has to be used). The input of the programme must include the following vectors of symbolic variables, declared through the MATLAB sym command: parameters, states, initial conditions, ODEs, observations, and inputs (known or unknown). 2.3. Implementation For initial conditions, two vectors must be provided: a vector called known_ics with entries equal to 1 for a known initial condition and 0 otherwise, and a vector ics with the values of the known ICS, either numeric or (known or unknown) parametric value. The output of the programme ﬁrst reports whether there is any symmetry. If a symmetry exists, the programme prints the inﬁnitesimal generator and the transformations on the screen. 3. Results To illustrate the application of the method in biological and biomedical modelling, we use it to analyse a set of models taken from the literature. The models are listed in Table 1 and their schematic representations are shown in Figure 2. Table 1. List of models analysed in this paper and summary of their features. Table 1. List of models analysed in this paper and summary of their features. Model Name (and Acronym) Reference States Parameters Outputs Simple chemical reaction (CR) [15] A k, s1, s2 Aobs Pharmacokinetic model (PK) [22] x1, x2, x3, x4 k1, k2, k3, k5, xobs 2 , xobs 3 k6, k7, s2, s3 NF-κB signalling pathway (NFKB) [23] x1, x2, x3, x4, x5, k0, k1, k1p, k2, k3, k4, y1, y2, y3, y4 x6, x7, x8, x9, x10 k5, k6, k7, k8, k9, k10, k11, s1, s2, s3, s4, ρvol, I0cyt, I0nuc Glucose-insulin regulation (Bolie) [24] q1, q2 p1, p2, p3, p4, Vp h 3.1. Simple Chemical Reaction Output of STRIKE-GOLDD for the initial model (Page 1) and the model with one-parameter Figure 3. Output of STRIKE-GOLDD for the initial model (Page 1) and the model with one-parameter Lie transformations (Page 2). 3.1. Simple Chemical Reaction This model represents a bimolecular reaction described by one ODE and one observation [15]: A = −2kA2 , Aobs = s1 A 1 + s2A . (33) (33) Aobs = s1 A 1 + s2A . (33) It is used to provide a basic illustration of the methodology, due to its simplicity. Without considering initial conditions, and using an univariate polynomial of second order, the programme ﬁnds two inﬁnitesimal generators: X = A ∂ ∂A −k ∂ ∂k −s1 ∂ ∂s1 −s2 ∂ ∂s2 , X = A2 ∂ ∂A + ∂ ∂s2 . (34) (34) Symmetry 2020, 12, 469 11 of 20 Figure 2. Diagrams of the models analysed in this article. (A) Simple chemical reaction. (B) Pharmacokinetic model. (C) NF-κB signalling pathway. (D) Glucose-insulin regulation system. Figure 2. Diagrams of the models analysed in this article. (A) Simple chemical reaction. (B) Pharmacokinetic model (C) NF-κB signalling pathway (D) Glucose-insulin regulation system Figure 2. Diagrams of the models analysed in this article. (A) Simple chemical reaction. Figure 2. Diagrams of the models analysed in this article. (A) Simple chemical reaction. (B) Pharmacokinetic model. (C) NF-κB signalling pathway. (D) Glucose-insulin regulation system. All the transformations are elementary: A∗= eεA , k∗= e−εk , s∗ 1 = e−εs1 , s∗ 2 = e−εs2 . A∗= A 1 −εA , s∗ 2 = s2 + ε . (35) (35) Our results coincide with those reported in [15]. It is possible to include ICS in order to study its inﬂuence in the model. This example, because of its simplicity, allows us to check the results manually. Once the transformations are computed, it is easy to see that the second group of transformations solves the same ODE. The time derivative of A∗is: ˙A∗= ˙A (1 −εA)2 Incorporating the above expression with A∗in (33), the ODE is still fulﬁlled. Incorporating the above expression with A∗in (33), the ODE is still fulﬁlled. Below are two screenshots of the results of the observability and identiﬁability analysis obtained with STRIKE-GOLDD. In the ﬁrst panel, Figure 3 (Page 1), corresponding to the initial model, all states and parameters are unobservable; in the second one, Figure 3 (Page 2), corresponding to the model with Lie transformations given by the second generator, states and parameters are observable. 12 of 20 Symmetry 2020, 12, 469 Figure 3. 3.2. Pharmacokinetic Model A pharmacokinetic (PK) model describes the time course of the concentrations of a drug in different compartments, after entering an organism. This model includes one input, four ODEs, and two outputs [22,25]: ˙x1 = u −(k1 + k2)x1 , ˙x2 = k1x1 −(k3 + k6 + k7)x2 + k5x4 , ˙x3 = k2x1 + k3x2 −k4x3 , ˙x4 = k6x2 −k5x4 , xobs 2 = s2x2 , xobs 3 = s3x3 . ˙x1 = u −(k1 + k2)x1 , ˙x2 = k1x1 −(k3 + k6 + k7)x2 + k5x4 , ˙x3 = k2x1 + k3x2 −k4x3 , ˙x4 = k6x2 −k5x4 , xobs 2 = s2x2 , xobs 3 = s3x3 . ˙x1 = u −(k1 + k2)x1 , ˙x2 = k1x1 −(k3 + k6 + k7)x2 + k5x4 , ˙x3 = k2x1 + k3x2 −k4x3 , ˙x4 = k6x2 −k5x4 , xobs 2 = s2x2 , xobs 3 = s3x3 . We use partial variate polynomials of second order, without ICS. Maximizing the number of elementary transformations leads to four of them, and the procedure yields the following inﬁnitesimal generator: X = k1 ∂ ∂k1 −∂ ∂k2 −k3(k1 + k2) k2 ∂ ∂k3 −∂ ∂k7 −s2 ∂ ∂s2 + k1s3 k2 ∂ ∂s3 + x2 ∂ ∂x2 −k1s3 k2 ∂ ∂x3 + x4 ∂ ∂x4 . The formulation of the IVP (16) for this inﬁnitesimal generator, considering only the non-elementary transformations, is: Symmetry 2020, 12, 469 13 of 20 13 of 20 Symmetry 2020, 12, 469 ˙k2 ∗= −k∗ 1 , k∗ 2(0) = k2 , ˙k3 ∗= −k∗ 3 (k∗ 1 + k∗ 2) k∗ 2 , k∗ 3(0) = k3 , ˙k7 ∗= k∗ 3 (k∗ 1 + k∗ 2) k∗ 2 , k∗ 7(0) = k7 , ˙s3∗= k∗ 1s∗ 3 k∗ 2 , s∗ 3(0) = s3 , ˙x3∗= −k∗ 1x∗ 3 k∗ 2 , x∗ 3(0) = x3 . The solution of the ODE system, replacing k∗ 1 with its transformation, is: The solution of the ODE system, replacing k∗ 1 with its transformation, is: k∗ 2 =k1 + k2 −k1eε , (36) k∗ 3 =k3e−ε(k1 + k2 −k1eε) k2 , (37) k∗ 7 =k7 + k3(k1 + k2) k2 −k3e−ε(k1 + k2) k2 , (38) x∗ 3 = x3(k1 + k2 −k1eε) k2 , (39) s∗ 3 = k2s3 (k1 + k2 −k1eε) . 3.2. Pharmacokinetic Model (49) (49) It is only necessary to prove that (46) is the same as (49) to show that both transformations are equal. Equation (46) can be reordered in terms of the powers of k1/k2: It is only necessary to prove that (46) is the same as (49) to show that both transformations are equal. Equation (46) can be reordered in terms of the powers of k1/k2: s∗ 3 = s3 1 + k1 k2 ε + ε2 2 + ε3 6 + ε4 24 + k2 1 k2 2 ε2 + ε3 + 7ε4 12 + k3 1 k3 2 ε3 + 3ε4 2 + k4 1ε4 k4 2 ! . (50) (50) The coefﬁcient of k1/k2 is eε −1, as was proven before in (47) and (48). The coefﬁcient of the third term must be equal to The coefﬁcient of k1/k2 is eε −1, as was proven before in (47) and (48). The coefﬁcient of k1/k2 is eε −1, as was proven before in (47) and (48). The coefﬁcient of the third term must be equal to The coefﬁcient of the third term must be equal to (eε −1)2 = e2ε + 1 −2eε . Considering the Taylor series of the previous expression until the ﬁfth order, 1 + 2ε + 2ε2 + 8ε3 6 + 16ε4 24 + 1 −2 −2ε −ε2 −ε3 3 −2ε4 24 = ε2 + ε3 + 7ε4 12 , (51) (51) we obtain that the ﬁnal expression is the same as that presented in (50). we obtain that the ﬁnal expression is the same as that presented in (50). The coefﬁcient of k3 1/k3 2 is the result of considering some terms of the Taylor series of (eε −1)3: we obtain that the ﬁnal expression is the same as that presented in (50). 3.2. Pharmacokinetic Model (40) (36) (37) (38) (39) (40) These new transformations coincide with those presented in [25]. Using Lie series in the inﬁnitesimal generator, the new expressions are: These new transformations coincide with those presented in [25]. Using Lie series in the inﬁnitesimal generator, the new expressions are: x∗ 2 = x2eε , x∗ 4 = x4eε , k∗ 1 = k1eε , s∗ 2 = s2e−ε , (41) x∗ 3 = x3 −εk1x3 k2 −ε2k1x3 2k2 −ε3k1x3 6k2 −ε4k1x3 24k2 , (42) k∗ 2 = k2 −εk1 −ε2k1 2 −ε3k1 6 −ε4k1 24 , (43) k∗ 3 = k3 −k3(k1 + k2)ε k2 + ε2k3(k1 + k2) 2k2 −ε3k3(k1 + k2) 6k2 + ε4k3(k1 + k2) 24k2 , (44) k∗ 7 = k7 + k3(k1 + k2)ε k2 −ε2k3(k1 + k2) 2k2 + ε3k3(k1 + k2) 6k2 −ε4k3(k1 + k2) 24k2 , (45) s∗ 3 = s3 + εk1s3 k2 + ε2k1s3(2k1 + k2) 2k2 2 + ε3k1s3(6k2 1 + 6k1k2 + k2 2) 6k3 2 + (46) + ε4k1s3(24k3 1 + 36k2 1k2 + 14k1k2 2 + k3 2) 24k4 2 . (41) (42) (43) (46) These expressions seem to differ from the closed form, however, the result can be reformulated in order to obtain them. Taking as example x∗ 3: x∗ 3 = x3 −k1x3 k2 ε + ε2 2 + ε3 6 + ε4 24 . (47) (47) Symmetry 2020, 12, 469 14 of 20 The last part of the previous expression consists of the ﬁrst terms of the following series: The last part of the previous expression consists of the ﬁrst terms of the following series: ∞ ∑ n=0 εn n! −1 = eε −1 . (48) (48) Substituting (48) in (47), the transformation of x∗ 3 is the same as the one obtained from IVP: x∗ 3 = x3 −k1x3 k2 (eε −1) = x3k2 −k1x3(eε −1) k2 = x3(k2 −k1(eε −1)) k2 . The most complicated case appears to be s∗ 3. The expression obtained from the IVP (40) can be rearranged as: ∞ n The most complicated case appears to be s∗ 3. The expression obtained from the IVP (40) can be rearranged as: s∗ 3 = k2s3 k2 + k1(1 −eε) = s3 1 −k1 k2 (eε −1) = s3 ∞ ∑ n=0 k1 k2 (eε −1) n . 3.3. NF-κB Signalling Pathway The model studied in this example was described in [23,26]. It represents a cellular signaling pathway found in most animal cells, corresponding to the NF-κB transcription factor. It is depicted in Figure 2C, where black arrows indicate exit routes. ˙x1 = k11x10 − k1u 1 + k0u + k1p x1 , ˙x2 = k1u 1 + k0u + k1p x1 −k2x2 , ˙x3 = k2x2 −k3x3 , ˙x4 = k2x2 −k4x4 , ˙x5 = k3ρvolx3 −k5x5 , ˙x6 = k5x5 −k10x9x6 , ˙x7 = k6x6 −k7x7 , ˙x8 = k8x7 −k9x8 , ˙x9 = k9ρvolx8 −k10x9x6 , ˙x10 = k10x9x6 −k11ρvolx10 , yobs 1 = s1(x1 + x2 + x3) + I0cyt , yobs 2 = s2(x10 + x5 + x6) + I0nuc , yobs 3 = s3(x2 + x3) , yobs 4 = s4(x2 + x4) . ˙x1 = k11x10 − k1u 1 + k0u + k1p x1 , ˙x2 = k1u 1 + k0u + k1p x1 −k2x2 , ˙x3 = k2x2 −k3x3 , ˙x4 = k2x2 −k4x4 , ˙x5 = k3ρvolx3 −k5x5 , ˙x6 = k5x5 −k10x9x6 , ˙x7 = k6x6 −k7x7 , ˙x8 = k8x7 −k9x8 , ˙x9 = k9ρvolx8 −k10x9x6 , ˙x10 = k10x9x6 −k11ρvolx10 , yobs 1 = s1(x1 + x2 + x3) + I0cyt , yobs 2 = s2(x10 + x5 + x6) + I0nuc , yobs 3 = s3(x2 + x3) , yobs 4 = s4(x2 + x4) . This model is the only one studied with ICS. In this case, they are parametric and include an additional parameter, x10, that does not appear in the model equations: x1(0) = x10, x2(0) = k1px1 k2 , x3(0) = k1px1 k3 , x4(0) = k1px1 k4 , x5(0) = k3ρvolx3 k5 , x6(0) = k7x7 k6 , x7(0) = k9x8 k8 , x8(0) = k3x3 k9 , x9(0) = k5x5 x6k10 , x10(0) = k1px1 k11 . Using a second order univariate polynomial, the following inﬁnitesimal transformations were found: Using a second order univariate polynomial, the following inﬁnitesimal transformations were found: • Scaling symmetry was found for the known input function and two parameters: • Scaling symmetry was found for the known input function and two parameters: u∗= ue−ε; k∗ i = kieε i = 0, 1 . 3.2. Pharmacokinetic Model The coefﬁcient of k3 1/k3 2 is the result of considering some terms of t p p ( ) The coefﬁcient of k3 1/k3 2 is the result of considering some terms of the Taylor series of (eε −1)3: p p The coefﬁcient of k3 1/k3 2 is the result of considering some terms of the Taylor series o 3 1 + ε + ε2 2 + ε3 6 + ε4 24 −3 1 + 2ε + 2ε2 + 8ε3 6 + 16ε4 24 + 1 + 3ε + 9ε2 2 + 27ε3 6 + 81ε4 24 −1 = = ε3 + 36ε4 24 = ε3 + 3ε4 2 . It is necessary to consider more terms of the Lie series to prove that the coefﬁcient of k4 1/k4 2 is (eε −1)4. In any case, regardless of the model under consideration, the transformations obtained with Lie series—that is, the output of the programme—are the expanding form of the solution of the IVP. In order to obtain a good approximation of the closed form (36)–(40) the programme needs to consider sufﬁcient terms of the Lie series. For this case, two terms are sufﬁcient to achieve full observability (i.e., FISPO). This model had been previously analysed in [15,25]. In [15] a different inﬁnitesimal generator was obtained; due to the lack of a method to compute non-elementary transformations, it was not possible to obtain all transformations. In [25] the same inﬁnitesimal generator as reported here was obtained, and transformations were computed using Hermite-Padé polynomials. The resulting transformations in [25] are the same as those obtained from the IVP, as well as those obtained by our programme using Lie series. 15 of 20 Symmetry 2020, 12, 469 3.3. NF-κB Signalling Pathway 3.3. NF-κB Signalling Pathway Symmetry 2020, 12, 469 3.3. NF-κB Signalling Pathway • The second symmetry was also for the input function and one parameter, in this case, a Mobius and translation symmetry, respectively: • The second symmetry was also for the input function and one parameter, in this case, a Mobius and translation symmetry, respectively: u∗= − u εu −1; k∗ 0 = k0 + ε . • Another scaling symmetry involving one state and two parameters: • Another scaling symmetry involving one state and two parameters: x∗ 7 = x7e−ε; k∗ 6 = k6e−ε; k∗ 8 = k8eε . Symmetry 2020, 12, 469 16 of 20 • One scaling type symmetry is admitted using the parameter ρvol. All the nucleus states, as well as four parameters, take part in the symmetry: • One scaling type symmetry is admitted using the parameter ρvol. All the nucleus states, as well as four parameters, take part in the symmetry: • One scaling type symmetry is admitted using the parameter ρvol. All the nucleus states, as well as four parameters, take part in the symmetry: x∗ i = xieε i = 5, 6, 9, 10; s∗ 2 = s2e−ε ; k∗ i = kie−ε i = 6, 10, 11; ρ∗ vol = ρvoleε . • The last symmetry is the only one that involves the initial condition parameter, x10. All of the states have a scaling type symmetry, compensated by the scaling factor of si and k10: • The last symmetry is the only one that involves the initial condition parameter, x10. All of the states have a scaling type symmetry, compensated by the scaling factor of si and k10: • The last symmetry is the only one that involves the initial condition parameter, x10. All of the states have a scaling type symmetry, compensated by the scaling factor of si and k10: x∗ i = xieε i = 1, ..., 10; s∗ i = sie−ε i = 1, ..., 4 ; k∗ 10 = k10e−ε; x∗ 10 = x10eε . All of the symmetries are elementary transformations and it was not necessary to use Lie series. If ICS had not been considered, the symmetries would be elementary too. This model was studied in [15]. The results of the ﬁrst four transformations presented above coincide with those found in [15]; however, the last transformation includes the ICS parameter, unlike in the aforementioned article. 3.4. Glucose-Insulin Regulation The solution of the ODE system is: The solution of the ODE system is: q∗ 1 =q1e−ε , q∗ 2 =q2 + e−εp3q1 + eεu p2 , p∗ 1 =p1 + p3 −p3e−ε , p∗ 2 =p2e−ε , p∗ 3 =p3e−ε , p∗ 4 = p2 3e−ε p2 −p2 3 + p1p3 p2 + eε(p1p3 + p2p4) p2 , V∗ p =Vpe−ε . q∗ 1 =q1e−ε , q∗ 2 =q2 + e−εp3q1 + eεu p2 , p∗ 1 =p1 + p3 −p3e−ε , p∗ 2 =p2e−ε , p∗ 3 =p3e−ε , p∗ 4 = p2 3e−ε p2 −p2 3 + p1p3 p2 + eε(p1p3 + p2p4) p2 , V∗ p =Vpe−ε . It is possible, using the procedure described for the pharmacokinetic model, to verify that the IVP solutions are the closed form of the solutions obtained through Lie series provided by the programme. It is possible, using the procedure described for the pharmacokinetic model, to verify that the IVP solutions are the closed form of the solutions obtained through Lie series provided by the programme. This model was proposed by Bolie [24] and its structural identiﬁability was analysed in [27] for This model was proposed by Bolie [24] and its structural identiﬁability was analysed in [27] for the ﬁrst time. However, its symmetries had not been studied until now. 3.4. Glucose-Insulin Regulation This model describes the regulation of blood glucose and insulin [24]. It has two states (glucose and insulin), one output (a glucose measurement), and a known input (the glucose entering from the digestive system): ˙q1 = u + p1q1 −p2q2 , ˙q2 = p3q2 + p4q1 , y = q1 Vp . Here we analyse its symmetries without considering ICS. The model has two inﬁnitesimal generators, which can be found using multivariate polynomials of second order: X = q2 ∂ ∂q2 −p2 ∂ ∂p2 + p4 ∂ ∂p4 , X = −q1 ∂ ∂q1 + u + p3q1 p2 ∂ ∂q2 + p3 ∂ ∂p1 −p2 ∂ ∂p2 −p3 ∂ ∂p3 + + p1p3 −p2 3 + p2p4 p2 ∂ ∂p4 −Vp ∂ ∂Vp . The ﬁrst inﬁnitesimal generator includes only elementary transformations: The ﬁrst inﬁnitesimal generator includes only elementary transformations: q∗ 2 = q2eε , p∗ 2 = p2e−ε , p∗ 4 = p4eε . Symmetry 2020, 12, 469 17 of 20 17 of 20 The second inﬁnitesimal generator has seven transformations, four of which are elementary: q∗ 1 = q1e−ε , p∗ 2 = p2e−ε , p∗ 3 = p3e−ε , V∗ p = Vpe−ε , q∗ 2 = (u −p3q1)ε4 24p2 + (u + p3q1)ε3 6p2 + (u −p3q1)ε2 2p2 + (u + p3q1)ε p2 + q2 , p∗ 1 = −p3ε4 24 + p3ε3 6 −p3ε2 2 + p3ε + p1 , p∗ 4 = (p2 3 + p1p3 + p2p4)ε4 24p2 + (−p2 3 + p1p3 + p2p4)ε3 6p2 + + (p2 3 + p1p3 + p2p4)ε2 2p2 + (−p2 3 + p1p3 + p2p4)ε p2 + p4 . Considering two terms of Lie series, the reparameterized model is FISPO, as classiﬁed by STRIKE-GOLDD. Considering two terms of Lie series, the reparameterized model is FISPO, as classiﬁed by STRIKE-GOLDD. The formulation of the IVP for the second inﬁnitesimal generator is: ˙q1∗= −q∗ 1 , q∗ 1(0) = q1 , ˙q2∗= u + p∗ 3q∗ 1 p∗ 2 , q∗ 2(0) = q2 , ˙p1∗= p∗ 3 , k∗ 7(0) = k7 , ˙p2∗= −p∗ 2 , p∗ 2(0) = p2 , ˙p3∗= −p∗ 3 , p∗ 3(0) = p3 , ˙p4∗= p∗ 1 p∗ 3 −(p∗ 3)2 + p∗ 2 p∗ 4 p∗ 2 , p∗ 4(0) = p4 . ˙Vp ∗= −V∗p , V∗p (0) = Vp . 4. Discussion This article has addressed the relationship between non-observability and Lie algebra. Its main contribution is a computational method that searches for inﬁnitesimal transformations in models Symmetry 2020, 12, 469 18 of 20 composed of rational functions, in order to undo the symmetries that these may present. The procedure is based on expressing each transformation admitted by the ODE system according to its inﬁnitesimal generator in polynomial form. In this way, the search for symmetries is equivalent to solving a system of linear equations, whose solution yields a transformation of the parameters that makes the model observable while leaving the observations invariant. Our method builds on previous work [12,19,25], and especially on the procedure presented by Merkt et al. [15], with the addition of two features. The ﬁrst one is the a priori maximization of the number of explicit transformations that can be obtained from the inﬁnitesimal generator. The second one is the calculation of non-elementary transformations by means of Lie series. Increasing the number of explicit transformations is beneﬁcial not only because it reduces the number of terms to consider from the Lie series, but also for calculating the solutions of the IVP. The complexity of the IVP solutions is inversely proportional to the number of explicit transformations. The pharmacokinetic model (PK) analysed here illustrates this point: without the use of the maximization of explicit transformations, MATLAB’s symbolic math toolbox did not manage to solve the IVP. The algorithm allows to study the inﬂuence of the initial conditions in the model. The type of ICS (parametric, numeric or both) and the states that incorporate them may affect the number and type of symmetries of the models, varying from explicit to non-elementary transformations and reducing the number of inﬁnitesimal generators. We have implemented the method as a MATLAB programme that automates both the search for symmetries and the reconstruction of the model from the inﬁnitesimal generators found. The programme has been integrated in the STRIKE-GOLDD toolbox for observability and identiﬁability analysis. The software has been tested with four previously published biomedical models, one of which—Bolie’s glucose-insulin regulation model—had not been tested for symmetries before. In the other cases our diagnoses mostly agree with those previously reported in the literature. An exception is the NF-κB model, for which we found an inﬁnitesimal generator that includes the parameter introduced by the initial conditions and that was not found in a previous analysis [15]. References 1. DiStefano, J., III. Dynamic Systems Biology Modeling and Simulation; Academic Press: Amsterdam, The Netherlands, 2015. 1. DiStefano, J., III. Dynamic Systems Biology Modeling and Simulation; Academic Press: Amsterdam, The Netherlands, 2015. 1. DiStefano, J., III. Dynamic Systems Biology Modeling and Simulation; Academic Press: Amsterdam, The Netherlands, 2015. 2. Sontag, E.D. Some new directions in control theory inspired by systems biology. IET Syst. Biol. 2004, 1, 9–18. [CrossRef] [PubMed] 2. Sontag, E.D. Some new directions in control theory inspired by systems biology. IET Syst. Biol. 2004, 1, 9–18. [CrossRef] [PubMed] 3. Åström, K.J.; Kumar, P.R. Control: A perspective. Automatica 2014, 50, 3–43. [CrossRef] 4. Bellman, R.; Åström, K.J. On structural identiﬁability. Math. Biosci. 1970, 7, 329–339. [CrossRef] 4. Bellman, R.; Åström, K.J. On structural identiﬁability. Math. Biosci. 1970, 7, 329–339. [CrossRef] 5. Miao, H.; Xia, X.; Perelson, A.S.; Wu, H. On identiﬁability of nonlinear ODE models and applications in viral dynamics. SIAM Rev. 2011, 53, 3–39. [CrossRef] 5. Miao, H.; Xia, X.; Perelson, A.S.; Wu, H. On identiﬁability of nonlinear ODE models and applications in viral dynamics. SIAM Rev. 2011, 53, 3–39. [CrossRef] 6. Chi¸s, O.T.; Banga, J.R.; Balsa-Canto, E. Structural identiﬁability of systems biology models: A critical comparison of methods. PLoS ONE 2011, 6, e27755. [CrossRef] 6. Chi¸s, O.T.; Banga, J.R.; Balsa-Canto, E. Structural identiﬁability of systems biology models: A critical comparison of methods. PLoS ONE 2011, 6, e27755. [CrossRef] 7. Raue, A.; Karlsson, J.; Saccomani, M.P.; Jirstrand, M.; Timmer, J. Comparison of approaches for parameter identiﬁability analysis of biological systems. Bioinformatics 2014, 30, 1440–1448. [CrossRef] 7. Raue, A.; Karlsson, J.; Saccomani, M.P.; Jirstrand, M.; Timmer, J. Comparison of approaches for parameter identiﬁability analysis of biological systems. Bioinformatics 2014, 30, 1440–1448. [CrossRef] 8. Villaverde, A.F. Observability and Structural Identiﬁability of Nonlinear Biological Systems. Complexity 2019, 2019, 8497093. [CrossRef] 9. Bluman, G.; Anco, S. Symmetry and Integration Methods for Differential Equations; Springer Science & B Media: New York, NY, USA, 2008. 10. Oliveri, F. Lie symmetries of differential equations: Classical results and recent contributions. Sy 2010, 2, 658–706. [CrossRef] 11. Arrigo, D.J. Symmetry Analysis of Differential Equations: An Introduction; John Wiley & Sons: New York, NY, USA, 2015. 12. Yates, J.W.; Evans, N.D.; Chappell, M.J. Structural identiﬁability analysis via symmetries of differential equations. Automatica 2009, 45, 2585–2591. [CrossRef] 13. Vajda, S.; Godfrey, K.R.; Rabitz, H. Similarity transformation approach to identiﬁability analysis of nonlinear compartmental models. Math. Biosci. 1989, 93, 217–248. 4. Discussion and A.F.V.; writing—original draft preparation, G.M. and A.F.V.; writing—review and editing, G.M. and A.F.V.; supervision, A.F.V. All authors have read and agreed to the published version of the manuscript. Funding: This research was funded by the Spanish Ministry of Science, Innovation and Universities through the project SYNBIOCONTROL (ref. DPI2017-82896-C2-2-R). Conﬂicts of Interest: The authors declare no conﬂict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results. Conﬂicts of Interest: The authors declare no conﬂict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results. 4. Discussion We observed another difference between our software and the one provided with [15]: when analysing the chemical reaction (CR) and pharmacokinetic (PK) models, the generators obtained with our code remained the same when varying the type of polynomial and degree; in contrast, the generators obtained with the programme of [15] changed when using partially varied and multivariate polynomials of order three or higher. These discrepancies may be due to implementation issues. Our symmetry-detecting algorithm can be directly used to analyse structural identiﬁability and observability, providing an alternative to the OIC-checking algorithm already included in STRIKE-GOLDD for that purpose. More importantly however, this new code provides additional information about the relationships between model variables that cause loss of identiﬁability and/or observability. These insights can be exploited in two ways: (i) by ﬁxing one or more parameters involved in a symmetry, in order to render the remaining ones identiﬁable, and (ii) by using the symmetry-breaking transformations to reformulate the model, yielding a modiﬁed model that is identiﬁable and observable. To facilitate the application of the latter procedure, we have implemented in our programme the semi-automatic transformation of a non-observable (respectively, non-identiﬁable) model into an observable (respectively, identiﬁable) model. It should be noted that, while said transformation may render a model fully observable, it also modiﬁes the expression of the variables involved in its equations, which lose their original mechanistic meaning. Thus, while the results of the procedure can offer valuable insight about the model structure, they should be applied carefully for the purpose of model reformulation. Our programme has some known limitations. First, while it considers high order generators and it can uncover a wide range of possible symmetries, it lacks procedures for determining a priori the type and total number of symmetries present in a model. Second, it does not provide a bound on the number of terms of the Lie series needed to obtain the inﬁnitesimal transformations, when these are 19 of 20 Symmetry 2020, 12, 469 not given by explicit transformations. To the best of our knowledge, these limitations are shared with other existing methodologies. The possibility of overcoming them will be considered in future work. not given by explicit transformations. To the best of our knowledge, these limitations are shared with other existing methodologies. The possibility of overcoming them will be considered in future work. Author Contributions: Conceptualization, A.F.V.; methodology, G.M.; software, G.M.; investigation, G.M. Abbreviations The following abbreviations are used in this manuscript: The following abbreviations are used in this manuscript: CR Chemical reaction FISPO Full input, state, and parameter observability ICS Initial conditions IVP Initial value problem ODE Ordinary differential equation OIC Observability-identiﬁability condition PDE Partial differential equation PK Pharmacokinetic SLI Structurally locally identiﬁable SU Structurally unidentiﬁable The following abbreviations are used in this manuscript: CR Chemical reaction FISPO Full input, state, and parameter observability ICS Initial conditions IVP Initial value problem ODE Ordinary differential equation OIC Observability-identiﬁability condition PDE Partial differential equation PK Pharmacokinetic SLI Structurally locally identiﬁable SU Structurally unidentiﬁable References [CrossRef] 14. Evans, N.D.; Chapman, M.J.; Chappell, M.J.; Godfrey, K.R. Identiﬁability of uncontrolled nonlinear rational systems. Automatica 2002, 38, 1799–1805. [CrossRef] Symmetry 2020, 12, 469 20 of 20 15. Merkt, B.; Timmer, J.; Kaschek, D. Higher-order Lie symmetries in identiﬁability and predictability analysis of dynamic models. Phys. Rev. E 2015, 92, 012920. [CrossRef] [PubMed] 16. Villaverde, A.F.; Barreiro, A.; Papachristodoulou, A. Structural identiﬁability of dynamic systems biology models. PLoS Comput. Biol. 2016, 12, e1005153. [CrossRef] [PubMed] 17. Villaverde, A.F.; Evans, N.D.; Chappell, M.J.; Banga, J.R. Input-dependent structural identiﬁability of nonlinear systems. IEEE Control Syst. Lett. 2019, 3, 272–277. [CrossRef] 18. Villaverde, A.F.; Tsiantis, N.; Banga, J.R. Full observability and estimation of unknown inputs, states, and parameters of nonlinear biological models. J. R. Soc. Interface 2019, in review. [CrossRef] 19. Anguelova, M.; Karlsson, J.; Jirstrand, M. Minimal output sets for identiﬁability. Math. Biosci. 2012, 239, 139–153. [CrossRef] 20. Saccomani, M.P.; Audoly, S.; D’Angiò, L. Parameter identiﬁability of nonlinear systems: The role of initial conditions. Automatica 2003, 39, 619–632. [CrossRef] 21. Rocha Filho, T.M.; Figueiredo, A. [SADE] a Maple package for the symmetry analysis of differential equations. Comput. Phys. Commun. 2011, 182, 467–476. [CrossRef] 22. Raksanyi, A. Utilisation du calcul formel pour l’étude des systèmes d’équations polynomiales applications en modélisation. Ph.D. Thesis, Université de Paris-Dauphine, Paris, France, 1986. 3. Lipniacki, T.; Paszek, P.; Brasier, A.R.; Luxon, B.; Kimmel, M. Mathematical model of NF-kB regula module. J. Theor. Biol. 2004, 228, 195–215. [CrossRef] 24. Bolie, V.W. Coefﬁcients of normal blood glucose regulation. J. Appl. Physiol. 1961, 16, 783–788. [CrossRef] 24. Bolie, V.W. Coefﬁcients of normal blood glucose regulation. J. Appl. Physiol. 1961, 16, 783–788. [CrossRef] 25. Sedoglavic, A. A probabilistic algorithm to test local algebraic observability in polynomial time. J. Symb. Comput. 2002, 33, 735–755. [CrossRef] 25. Sedoglavic, A. A probabilistic algorithm to test local algebraic observability in polynomial time. J. Symb. Comput. 2002, 33, 735–755. [CrossRef] 26. Cheong, R.; Hoffmann, A.; Levchenko, A. Understanding NF-kB signaling via mathematical modeling. Mol. Syst. Biol. 2008, 4, 192. [CrossRef] [PubMed] 27. Cobelli, C.; DiStefano, J. Parameter and structural identiﬁability concepts and ambiguities: A critical review and analysis. Am. J. Physiol.-Regulat. Integr. Comp. Physiol. 1980, 239, R7–R24. [CrossRef] [PubMed] c⃝2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
https://openalex.org/W3133438047	https://pure.eur.nl/files/48436353/Repub_135256_O-A.pdf	English	null	Polygenic hazard score is associated with prostate cancer in multi-ethnic populations	Nature communications	2,021	cc-by	10,449	ARTICLE ARTICLE ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-021-21287-0 P P rostate cancer is the second most common cancer diag- nosed in men worldwide, causing substantial morbidity and mortality1. Prostate cancer screening may reduce morbidity and mortality2–5, but to avoid overdiagnosis and overtreatment of indolent disease6–9, it should be targeted and personalized. Prostate cancer age at diagnosis is important for clinical decisions regarding if/when to initiate screening for an individual10,11. Survival is another key cancer endpoint recommended for risk models12. each genetic ancestry group, men with high PHS had HRs of 5.54 [5.18–5.93], 4.49 [3.23–6.33], and 2.54 [2.08–3.10] for men of European, Asian, and African ancestry, respectively. each genetic ancestry group, men with high PHS had HRs of 5.54 [5.18–5.93], 4.49 [3.23–6.33], and 2.54 [2.08–3.10] for men of European, Asian, and African ancestry, respectively. PHS association with aggressive prostate cancer in OncoArray. PHS2 was associated with age at aggressive prostate cancer diagnosis in all three OncoArray-deﬁned genetic ancestry groups (Table 2). Comparing the 80th and 20th percentiles of genetic risk, men with high PHS had an HR of 5.88 [5.46–6.33] for aggressive prostate cancer; within each genetic ancestry group, men with high PHS had HRs of 5.62 [5.23–6.05], 5.16 [4.79–5.55], and 2.43 [2.26-2.61] for men of European, Asian, and African ancestry, respectively. Genetic risk stratiﬁcation is promising for identifying indivi- duals with a greater predisposition for developing cancer13–16, including prostate cancer17. Polygenic models use common var- iants—identiﬁed in genome-wide association studies—whose combined effects can assess the overall risk of disease develop- ment18,19. Recently, a polygenic hazard score (PHS) was devel- oped as a weighted sum of 54 single-nucleotide polymorphisms (SNPs) that models a man’s genetic predisposition for developing prostate cancer13. Validation testing was done using ProtecT trial data2 and demonstrated the PHS to be associated with age at prostate cancer diagnosis, including aggressive prostate cancer13. However, the development and validation datasets were limited to men of European ancestry. While genetic risk models might be important clinical tools for prognostication and risk stratiﬁcation, using them may worsen health disparities20–24 because most models are constructed using European data and may under- represent genetic variants important in persons of non-European ancestry20–24. Indeed, this is particularly concerning in prostate cancer, as race/ethnicity is an important prostate cancer risk factor; diagnostic, treatment, and outcomes disparities continue to exist between different races/ethnicities25,26. PHS association with fatal prostate cancer in OncoArray. Results Adaption of PHS for OncoArray. Of the 30 SNPs from PHS1 not directly genotyped on OncoArray, proxy SNPs were identiﬁed for 22 (linkage disequilibrium ≥0.94). Therefore, PHS2 included 46 SNPs, in total (Supplementary Information). PHS2 association with age at aggressive prostate cancer diagnosis in ProtecT was similar to that previously reported for PHS1 (z = 21.7, p = 3.6 × 10−104 for PHS1; z = 21.4, p = 1.3 × 10−101 for PHS2). HR98/50 was 4.68 [95% CI: 3.62–6.15] for PHS2, compared to 4.61 [3.52–5.99] for PHS1. ARTICLE PHS2 was associated with age at prostate cancer death for all men in the multi-ethnic dataset (z = 15.9, p = 6.3 × 10−57). Table 3 shows z- scores and corresponding HRs for fatal prostate cancer. Com- paring the 80th and 20th percentiles of genetic risk, men with high PHS had a HR of 5.68 [5.07–6.46] for prostate cancer death. Sensitivity analyses. Sensitivity analyses demonstrated that large changes in assumed population incidence had minimal effect on the calculated HRs for any, aggressive, or fatal prostate cancer (Supplementary Information). PHS and family history. Family history was also associated with any prostate cancer (z = 39.7, p < 10−300; Table 4), aggressive prostate cancer (z = 32.4, p = 2.7 × 10−230), and fatal prostate cancer (z = 8.76, p = 1.4 × 10−18) in the multi-ethnic dataset. Among those with known family history, the combination of family history and PHS performed better than family history alone (log-likelihood p < 10−300). This pattern held true when analyses were repeated on each genetic ancestry. Additional family history analyses are reported in the Supplementary Information. Here, we assessed PHS performance in a multi-ethnic dataset that includes individuals of European, African, and Asian genetic ancestry. This dataset also includes long-term follow-up infor- mation, affording an opportunity to evaluate PHS for association with fatal prostate cancer. Polygenic hazard score is associate cancer in multi-ethnic populations Genetic models for cancer have been evaluated using almost exclusively European data, which could exacerbate health disparities. A polygenic hazard score (PHS1) is associated with age at prostate cancer diagnosis and improves screening accuracy in Europeans. Here, we evaluate performance of PHS2 (PHS1, adapted for OncoArray) in a multi-ethnic dataset of 80,491 men (49,916 cases, 30,575 controls). PHS2 is associated with age at diagnosis of any and aggressive (Gleason score ≥7, stage T3-T4, PSA ≥10 ng/mL, or nodal/distant metas- tasis) cancer and prostate-cancer-speciﬁc death. Associations with cancer are signiﬁcant within European (n = 71,856), Asian (n = 2,382), and African (n = 6,253) genetic ancestries (p < 10−180). Comparing the 80th/20th PHS2 percentiles, hazard ratios for prostate cancer, aggressive cancer, and prostate-cancer-speciﬁc death are 5.32, 5.88, and 5.68, respectively. Within European, Asian, and African ancestries, hazard ratios for prostate cancer are: 5.54, 4.49, and 2.54, respectively. PHS2 risk-stratiﬁes men for any, aggressive, and fatal prostate cancer in a multi-ethnic dataset. A full list of authors and their afﬁliations appears at the end of the paper. A full list of authors and their afﬁliations appears at the end of the paper. NATURE COMMUNICATIONS \| (2021) 12:1236 \| https://doi.org/10.1038/s41467-021-21287-0 \| www.nature.com/naturecommunications 1 NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-021-21287-0 NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-021-21287-0 Table 2 Association of PHS with aggressive prostate cancer. OncoArray genetic ancestry z (p Value) Hazard ratios [95% CI] comparing percentiles of PHS2 HR20/50: ≤20th vs. 30–70th HR80/50: ≥80th vs. 30–70th HR98/50: ≥98th vs. 30–70th HR80/20: ≥80th vs. ≤20th All (n = 58,600) 47.6 (p < 10−300) 0.43 [0.41–0.44] 2.50 [2.42–2.60] 4.61 [4.33–4.90] 5.88 [5.48–6.34] European (n = 53,608) 46.4 (p < 10−300) 0.44 [0.42–0.45] 2.45 [2.36–2.55] 4.40 [4.15–4.70] 5.62 [5.25–6.05] Asian (n = 1806) 43.8 (p < 10−300) 0.45 [0.37–0.55] 2.32 [1.88–2.89] 4.14 [2.92–6.03] 5.16 [3.45–7.78] African (n = 3186) 23.6 (p = 7.2 × 10−123) 0.64 [0.49–0.81] 1.55 [1.23–2.00] 2.18 [1.44–3.43] 2.43 [1.51–4.05] Hazard ratios (HRs) derived from Cox proportional hazards models are shown comparing men in the highest 2% of genetic risk (≥98th percentile of PHS), highest 20% of genetic risk (≥80th percentile), average risk (30–70th percentile), and lowest 20% of genetic risk (≤20th percentile) across genetic ancestry. p Values reported are two-tailed from the Cox models. Table 2 Association of PHS with aggressive prostate cancer. Table 3 Association of PHS with death from prostate cancer. Ancestry z (p Value) Hazard ratios [95% CI] comparing percentiles of PHS2 HR20/50: ≤20th vs. 30–70th HR80/50: ≥80th vs. 30-70th HR98/50: ≥98th vs. 30–70th HR80/20: ≥80th vs. ≤20th All (n = 78,221) 15.9 (p = 6.3 × 10−57) 0.43 [0.41–0.56] 2.47 [2.33–2.64] 4.46 [4.04–4.98] 5.68 [5.07–6.46] Hazard ratios (HRs) from Cox proportional hazards models are shown comparing men in the highest 2% of genetic risk (≥98th percentile of PHS), highest 20% of genetic risk (≥80th percentile), average risk (30–70th percentile), and lowest 20% of genetic risk (≤20th percentile). p Values reported are two-tailed from the Cox models. Table 3 Association of PHS with death from prostate cancer. Table 4 Multivariable models with both PHS and family history of prostate cancer (≥1 ﬁrst-degree relative affected) for association with any prostate cancer in the multi-ethnic dataset, and by genetic ancestry. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-021-21287-0 OncoArray genetic ancestry Variable beta z-score p Value HR All (n = 46,030) PHS 1.98 53.3 <10−300 4.48 Family history 0.94 38.6 <10−300 2.55 European (n = 39,445) PHS 2.06 56.2 <10−300 4.80 Family history 0.92 38.1 <10−300 2.50 Asian (n = 1028) PHS 1.89 50.7 <10−300 4.17 Family history 0.72 21.2 9.5 × 10−100 2.05 African (n = 5557) PHS 1.11 26.2 2.6 × 10−151 2.22 Family history 1.14 46.7 <10−300 3.11 This analysis is limited to individuals with known family history. Both family history and PHS were signiﬁcantly associated with any prostate cancer in the combined models. Hazard ratios (HRs) for family history were calculated as the exponent of the beta from the multivariable Cox proportional hazards regression56. The HR for PHS in the multivariable models was estimated as the HR80/20 (men in the highest 20% vs. those in the lowest 20% of genetic risk by PHS2) in each cohort. p Values reported are two-tailed from the Cox models. The model with PHS performed better than family history alone (log-likelihood p < 10−300). Table 4 Multivariable models with both PHS and family history of prostate cancer (≥1 ﬁrst-degree relative affected) for association with any prostate cancer in the multi-ethnic dataset, and by genetic ancestry. Table 4 Multivariable models with both PHS and family history of prostate cancer (≥1 ﬁrst-degree association with any prostate cancer in the multi-ethnic dataset, and by genetic ancestry. This analysis is limited to individuals with known family history. Both family history and PHS were signiﬁcantly associated with any prostate cancer in the combined models. Hazard ratios (HRs) for family history were calculated as the exponent of the beta from the multivariable Cox proportional hazards regression56. The HR for PHS in the multivariable models was estimated as the HR80/20 (men in the highest 20% vs. those in the lowest 20% of genetic risk by PHS2) in each cohort. p Values reported are two-tailed from the Cox models. The model with PHS performed better than family history alone (log-likelihood p < 10−300). PHS associations with aggressive prostate cancer using alter- native ancestry groupings y g p g Agnostic genetic ancestry groupings with fastSTRUCTURE. With fastSTRUCTURE, the optimal model was the one with K = 2 clusters: cluster 1 had mainly men of European OncoArray- deﬁned genetic ancestry and self-reported race/ethnicity, cluster 2 had only men of African OncoArray-deﬁned genetic ancestry and mostly Black/African American self-reported race/ancestry, while the Admixed cluster included men of all Oncotype-deﬁned genetic ancestries. Table 5 demonstrates the HR80/20 for aggres- sive prostate cancer for these K = 2 fastSTRUCTURE-deﬁned clusters. Comparing the 80th and 20th percentiles of genetic risk, men with high PHS had HRs for aggressive prostate cancer of 5.60 [5.55, 5.64], 2.06 [2.03, 2.09], and 5.05 [4.89, 5.21] for PHS association with any prostate cancer in OncoArray. PHS2 was associated with age at prostate cancer diagnosis in all three OncoArray-deﬁned genetic ancestry groups (Table 1). Comparing the 80th and 20th percentiles of genetic risk, men with high PHS had an HR of 5.32 [4.99–5.70] for any prostate cancer. Within Table 1 Association of PHS with prostate cancer. OncoArray genetic ancestry z (p Value) Hazard ratios [95% CI] comparing percentiles of PHS2 HR20/50: ≤20th vs. 30–70th HR80/50: ≥80th vs. 30–70th HR98/50: ≥98th vs. 30–70th HR80/20: ≥80th vs. ≤20th All (n = 80,491) 54.3 (p < 10−300) 0.45 [0.43–0.46] 2.39 [2.31–2.47] 4.21 [3.99–4.47] 5.32 [4.99–5.70] European (n = 71,856) 55.8 (p < 10−300) 0.44 [0.43–0.45] 2.44 [2.35–2.53] 4.34 [4.09–4.60] 5.54 [5.18–5.93] Asian (n = 2382) 46.7 (p < 10−300) 0.48 [0.40–0.56] 2.15 [1.81–2.57] 3.77 [2.80–5.13] 4.49 [3.23–6.33] African (n = 6253) 28.7 (p = 3.8 × 10−181) 0.63 [0.57–0.69] 1.59 [1.44–1.76] 2.27 [1.91–2.71] 2.54 [2.08–3.10] Hazard ratios (HRs) are shown comparing men in the highest 2% of genetic risk (≥98th percentile of PHS), highest 20% of genetic risk (≥80th percentile), average risk (30–70th percentile), and lowest 20% of genetic risk (≤20th percentile) across genetic ancestry. p Values reported are two-tailed from the Cox models. Table 1 Association of PHS with prostate cancer. NATURE COMMUNICATIONS \| (2021) 12:1236 \| https://doi.org/10.1038/s41467-021-21287-0 \| www.nature.com/naturecommunications 2 ARTICLE ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-021-21287-0 The ability of these models with more K clusters to risk-stratify men well (while possibly being less representative of the available data) emphasizes the dire need for more complex and deeper studies evaluating the intersection of genetics, the granularity of ancestry, and prostate cancer risk. In addition, the PHS may not include all SNPs associated with prostate cancer; in fact, over 60 additional SNPs have been reported since the development of the original PHS18. Some of these SNPs are ethnicity-speciﬁc, including within non-European populations44–46, and will be included in further model optimization to improve prostate cancer risk stratiﬁcation. Future work could also evaluate the PHS performance in relation to epidemiological risk factors associated with prostate cancer risk beyond those currently used in clinical practice (i.e., family history and race/ethnicity). Finally, various circumstances and disease-modifying treatments may have inﬂuenced post-diagnosis survival to an unknown degree. Despite this possible source of variability in survival among men with fatal prostate cancer, PHS was still associated with age at death, an objective, and meaningful endpoint. Future development and optimization hold promise for improving upon the encouraging risk stratiﬁcation achieved here in men of different genetic ances- tries, particularly African. I PHS i d i h d i The ﬁrst PHS validation study used data from ProtecT, a large prostate cancer trial2,13. ProtecT’s screening design yielded biopsy results from both controls and cases with PSA ≥3 ng/mL, making it possible to demonstrate improved accuracy and efﬁciency of prostate cancer screening with PSA testing. Limitations of the ProtecT analysis, though, include few recorded prostate cancer deaths in the available data, and the exclusion of advanced cancer from that trial2. The present study includes long-term observa- tion, with both early and advanced disease18, allowing for eva- luation of PHS association with any, aggressive, and fatal prostate cancer; we found PHS to be associated with all outcomes. Age is critical in clinical decisions of whether men should be offered prostate cancer screening31–34 and in how to treat men diagnosed with prostate cancer31,32. Age may also inform prog- nosis32,35. Age at diagnosis or death is therefore of clinical interest in inferring how likely a man is to develop cancer at an age when he may beneﬁt from treatment. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-021-21287-0 NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-021-21287-0 reasoning applies to men of all genetic ancestries. Risk-stratiﬁed screening should be prospectively evaluated. Current guidelines suggest considering a man’s individual cancer risk factors, overall life expectancy, and medical comor- bidities when deciding whether to screen6. The most prominent clinical risk factors used in practice are family history and race/ ethnicity6,36,37. Combined PHS and family history performed better than either alone in this multi-ethnic dataset. This ﬁnding is consistent with a prior report that PHS adds considerable information over family history alone. The prior study did not ﬁnd an association of family history with age at prostate cancer diagnosis, perhaps because the universal screening approach of the ProtecT trial diluted the inﬂuence of family history on who is screened in typical practice13. In the present study, family history and PHS appear complementary in assessing prostate cancer genetic risk. Moreover, the HRs for PHS suggest clinical relevance similar or greater to predictive tools routinely used for cancer screening (e.g., breast cancer) and for other diseases (e.g., diabetes and cardiovascular disease). HRs reported for those tools are around 1–3 for disease development or other adverse outcome38–42; HRs reported here for PHS (for any, aggressive, or fatal prostate cancer) are similar or greater. g p p y PHS performance was better in those with OncoArray-deﬁned European and Asian genetic ancestry than in those with African ancestry. For example, comparing the highest and lowest quintiles of genetic risk, men with OncoArray-deﬁned European and Asian genetic ancestry with high PHS had HRs for any prostate cancer of 5.54 and 4.49 times, respectively, while the analogous HR for men of African genetic ancestry was 2.54. This trend was also observed for aggressive prostate cancer. Moreover, the optimal fastSTRUCTURE clustering of our dataset (K = 2) yielded one cluster that consisted of almost only men of African ancestry (by both self-report and OncoArray-deﬁned genetic ancestry) and had inferior risk stratiﬁcation with PHS2 (HR 2.06), compared to the performance observed in the other cluster (nearly all Eur- opean) and an admixed cluster (HRs 5.60 and 5.05, respectively). Overall, these results suggest PHS can differentiate men of higher and lower risk in each ancestral group, but the range of risk levels may be narrower in those of African ancestry. Possible reasons for relatively diminished performance include increased genetic diversity with less linkage disequilibrium in those of African genetic ancestry27–29. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-021-21287-0 Known health disparities may also con- tribute25, as the availability—and timing—of PSA results may depend on healthcare access. Alarmingly, there has historically been a poor representation of African populations in clinical or genomic research studies20,21. This pattern is reﬂected in the present study, where most men of African genetic ancestry were missing clinical diagnosis information used to determine disease aggressiveness. That such clinical information is less available for men of African ancestry also leaves open the possibility of sys- tematic differences in the diagnostic workup—and therefore the age of diagnosis—across different ancestry populations. These are critical health disparities that will need to be addressed (and ultimately eliminated) to ensure equitable and accurate genomic prostate cancer stratiﬁcation for all men. Notwithstanding these caveats, the present PHS is associated with age at prostate cancer diagnosis in men of African ancestry, possibly paving the way for more personalized screening decisions for men of African des- cent. Promising efforts are also underway to further improve PHS performance in men of African ancestry30. Limitations to this work include that the dataset comes from multiple, heterogeneous studies, from various populations with variable screening rates. This allowed for a large, multi-ethnic dataset that includes clinical and survival data, but comes with uncertainties avoided in the ProtecT dataset used for original vali- dation. However, the heterogeneity would likely reduce the PHS performance, not systematically inﬂate the results. Second, we note that no germline SNP tool, including this PHS, has been shown to discriminate men at risk of aggressive prostate cancer from those at risk of only indolent prostate cancer. Third, while the OncoArray- deﬁned and fastSTRUCTURE genetic ancestry classiﬁcations used here may be more accurate than self-reported race/ethnicity alone43 and allowed for evaluation of admixed genetic ancestry, detailed analysis of local ancestry was not assessed. As noted above, clinical data availability was not uniform across contributing studies and was lower in men of OncoArray-deﬁned African genetic ancestry. Efforts to improve genetic risk prediction should focus on con- sistent data collection patterns and elimination of data disparities so that models are widely applicable for all men. We also found that while the optimal fastSTRUCTURE model had K = 2 clusters for risk stratiﬁcation men for aggressive prostate cancer, models with more K clusters also produced comparable (or larger ranges) of hazard ratios for risk stratiﬁcation. Discussion Table 5 Association of PHS with aggressive prostate cancer, by two clusters using fastSTRUCTURE. fastSTRUCTURE K Cluster HR80/20: ≥80th vs. ≤20th K = 2 1 5.60 [5.55–5.64] 2 2.06 [2.03–2.09] Admixed 5.05 [4.89–5.21] Hazard ratios (HRs) from Cox proportional hazards models are shown comparing men in the highest 20% of genetic risk (≥80th percentile) vs. the lowest 20% of genetic risk (≤20th percentile). Table 5 Association of PHS with aggressive prostate cancer, by two clusters using fastSTRUCTURE. These results conﬁrm the previously reported association of PHS with age at prostate cancer diagnosis in Europeans and show that this ﬁnding generalizes to a multi-ethnic dataset, including men of Eur- opean, Asian, and African ancestry. PHS is also associated with age at aggressive prostate cancer diagnosis and at prostate cancer death. Comparing the highest and lowest quintiles of genetic risk, men with high PHS had HRs of 5.32, 5.88, and 5.68 for any prostate cancer, aggressive prostate cancer, and prostate cancer death, respectively. gg p p p y We found that PHS is associated with prostate cancer in men of European, Asian, and African genetic ancestry (and a wider range of self-reported race/ethnicities). Current prostate cancer screening guidelines suggest possible initiation at earlier ages for men of African ancestry, given higher incidence rates and worse survival when compared to men of European ancestry26. Using the PHS to risk- stratify men might help with decisions regarding when to initiate prostate cancer screening: perhaps a man with African genetic ancestry in the lowest percentiles of genetic risk by PHS could safely delay or forgo screening to decrease the possible harms associated with overdetection and overtreatment9, while a man in the highest risk percentiles might consider screening at an earlier age. Similar cluster 1, cluster 2, and admixed cluster, respectively. Corre- sponding results for the K = 3–6 clustering approaches are shown in the Supplementary Information. cluster 1, cluster 2, and admixed cluster, respectively. Corre- sponding results for the K = 3–6 clustering approaches are shown in the Supplementary Information. Self-reported race/ethnicity. HRs for aggressive prostate cancer comparing the 80th and 20th percentiles of genetic risk when participants are stratiﬁed by their self-reported race/ethnicity are shown in the Supplementary Information. TURE COMMUNICATIONS \| (2021) 12:1236 \| https://doi.org/10.1038/s41467-021-21287-0 \| www.nature.com/naturecommunications 3 ARTICLE NATURE COMMUNICATIONS \| (2021) 12:1236 \| https://doi.org/10.1038/s41467-021-21287-0 \| www.nature.com/naturecommunications Methods P ti i t Recognizing that not all prostate cancer is clinically signiﬁcant, we also tested PHS2 for association with age at aggressive prostate cancer diagnosis in the multi-ethnic dataset. For these analyses, we included cases that had known tumor stage, Gleason score, and PSA at diagnosis (n = 60,617 cases, Table 6). Aggressive prostate cancer cases were those that met any of the following criteria6,13: Gleason score ≥7, PSA ≥10 ng/mL, T3–T4 stage, nodal metastases, or distant metastases. As before, Cox proportional hazards models and sensitivity analysis were used to assess the association. Aggressive prostate cancer. Recognizing that not all prostate cancer is clinically signiﬁcant, we also tested PHS2 for association with age at aggressive prostate cancer diagnosis in the multi-ethnic dataset. For these analyses, we included cases that had known tumor stage, Gleason score, and PSA at diagnosis (n = 60,617 cases, Table 6). Aggressive prostate cancer cases were those that met any of the following criteria6,13: Gleason score ≥7, PSA ≥10 ng/mL, T3–T4 stage, nodal metastases, or distant metastases. As before, Cox proportional hazards models and sensitivity analysis were used to assess the association. Fatal prostate cancer. Using an even stricter deﬁnition of clinical signiﬁcance, we evaluated the association of PHS2 with age at prostate cancer death in the multi- ethnic dataset. All cases (regardless of staging completeness) and controls were included, and the endpoint was the age at death due to prostate cancer. This analysis was not stratiﬁed by genetic ancestry due to low numbers of recorded prostate cancer deaths in the non-European datasets. The cause of death was OncoArray-deﬁned genetic ancestry. Self-reported race/ethnicities47,52, included European, Black, or African American (includes Black African, Black Caribbean), East Asian, South Asian, Hawaiian, Hispanic American, and Other/Unknown. Table 6 Participant characteristics, n = 80,491. Methods P ti i t Participants. We obtained data from the OncoArray project47 that had undergone quality control steps18. This dataset includes 91,480 men with genotype and phe- notype data from 64 studies (Supplementary Information). Individuals whose data were used in the prior development or validation of the original PHS model (PHS1) were excluded (n = 10,989)13, leaving 80,491 in the independent dataset used here. Table 6 describes available data. Individuals not meeting the endpoint for each analysis were censored at age of last follow-up. Any prostate cancer. We tested PHS2 for association with age at diagnosis of any prostate cancer in the multi-ethnic dataset (n = 80,491, Table 6). PHS2 was calculated for all patients in the multi-ethnic dataset and used as the sole independent variable in Cox proportional hazards regressions for the endpoint of age at prostate cancer diagnosis. Due to the potential for Cox proportional hazards results to be biased by a higher number of cases in our dataset than in the general population, sample-weight corrections were applied to all Cox models using population data from Sweden13,53 (additional details are in Supplementary Information). Signiﬁcance was set at α = 0.0113. y g All contributing studies were approved by the relevant ethics committees; written informed consent was acquired from the study participants48. The present analyses used de-identiﬁed data from the PRACTICAL consortium. written informed consent was acquired from the study participants48. The present analyses used de-identiﬁed data from the PRACTICAL consortium. Polygenic hazard score. The original PHS1 was validated for association with age at prostate cancer diagnosis in men of European ancestry using a survival analysis13. To ensure the score was not simply identifying men at risk of indolent disease, PHS1 was also validated for association with age at aggressive prostate cancer (deﬁned as an intermediate-risk disease, or above6) diagnosis13. PHS1 was calculated as the vector product of a patient’s genotype (Xi) for n selected SNPs and the corresponding para- meter estimates (βi) from a Cox proportional hazards regression: These Cox proportional hazards regressions (with PHS2 as the sole independent variable and age at prostate cancer diagnosis as the outcome) were then repeated for subsets of data, stratiﬁed by OncoArray-deﬁned genetic ancestry: European, Asian, and African. Percentiles of genetic risk were calculated using data from the 9,728 men in the original (iCOGS) development set who were less than 70 years old and without prostate cancer13,54. ARTICLE ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-021-21287-0 dataset. PHS performance was relatively diminished in men of African genetic ancestry, compared to performance in men of European or Asian genetic ancestry. PHS risk-stratiﬁes men of various genetic ancestries for prostate cancer and should be pro- spectively studied as a means to individualize screening strategies seeking to reduce prostate cancer morbidity and mortality. Genetic ancestry for each individual from the OncoArray project47 was provided with the PRACTICAL consortium data. Brieﬂy, genotypes from 2318 ancestry informative markers were mapped into a two-dimensional space representing the ﬁrst two principal components, which has been shown to yield results very similar to those obtained with the STRUCTURE approach52. The distance from the individual’s mapping to the three reference clusters (European, African, and Asian) was then used to estimate the individual’s genetic ancestry47,52. Individuals were classiﬁed into one of three OncoArray-deﬁned labels; European: greater than 80% European ancestry, Asian: greater than 40% Asian ancestry, and African: greater than 20% African ancestry. Individuals not meeting any of the aforementioned three labels were classiﬁed as “other,” but all of the individuals in the present prostate cancer dataset met the criteria for one of the three OncoArray- deﬁned genetic ancestries. Methods P ti i t HRs and 95% CIs for each genetic ancestry group were calculated to make the following comparisons: HR98/50, men in the highest 2% of genetic risk vs. those with average risk (30–70th percentile); HR80/50, men in the highest 20% vs. those with average risk, HR20/50, men in the lowest 20% vs. those with average risk; and HR80/20, men in the highest 20% vs. lowest 20%. CIs were determined by bootstrapping 1000 random samples from each genetic ancestry group50,51 while maintaining the same number of cases and controls. HRs and CIs were calculated for age at prostate cancer diagnosis separately for each genetic ancestry group. PHS ¼ X n i Xiβi ð1Þ ð1Þ The 54 SNPs in PHS1 were selected using PRACTICAL consortium data (n = 31,747 men) genotyped with a custom array (iCOGS, Illumina, San Diego, CA)13. Adapting the PHS to OncoArray. Genotyping for the present study was per- formed using a commercially available, cancer-speciﬁc array (OncoArray, Illumina, San Diego, CA)18. Twenty-four of the 54 SNPs in PHS1 were directly genotyped on OncoArray. We identiﬁed proxy SNPs for those not directly genotyped and re- calculated the SNP weights in the same dataset used for the original development of PHS113 (Supplementary Methods). Given that the overall incidence of prostate cancer in different populations varies, we performed a sensitivity analysis of the population case/control numbers, allowing the population incidence to vary from 25 to 400% of that reported in Sweden (chosen as an example population; Supplementary Information). The performance of the adapted PHS (PHS2), was compared to that of PHS1 in the ProtecT dataset originally used to validate PHS1 (n = 6411). PHS2 was calculated for all patients in the ProtecT validation set and was tested as the sole predictive variable in a Cox proportional hazards regression model (R v.3.5.1, “survival” package49) for age at aggressive prostate cancer diagnosis, the primary endpoint of that study. The performance was assessed by the metrics reported during the PHS1 development:13 z-score and hazard ratio (HR98/50) for aggressive prostate cancer between men in the highest 2% of genetic risk (≥98th percentile) vs. those with average risk (30–70th percentile). HR 95% conﬁdence intervals (CIs) were determined by bootstrapping 1000 random samples from the ProtecT dataset50,51 while maintaining the same number of cases and controls. PHS2 percentile thresholds are shown in the Supplementary Information. Aggressive prostate cancer. NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-021-21287-0 One important advantage of the survival analysis used here is that it permits men without cancer at the time of the last follow-up to be censored while allowing for the possibility of them developing prostate cancer (including aggressive or fatal prostate cancer) later on. prostate cancer death is a hard endpoint with less uncertainty than clinical diagnosis (which may vary with screening practices and delayed medical attention). PHS may help identify men with a high (or low) genetic predisposition to develop lethal prostate cancer and could assist physicians in deciding when to initiate screening. In summary, PHS was associated with age at any and aggressive prostate cancer, and at death from prostate cancer in a multi-ethnic NATURE COMMUNICATIONS \| (2021) 12:1236 \| https://doi.org/10.1038/s41467-021-21287-0 \| www.nature.com/naturecommunicatio 4 ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-021-21287-0 determined by the investigators of each contributing study using cancer registries and/or medical records (Supplementary Information). At last follow-up, 3983 men had died from prostate cancer, 5806 had died from non-prostate cancer causes, and 70,702 were still alive. The median age at the last follow-up was 70 years (IQR: 63–76). As before, Cox proportional hazards models and sensitivity analysis were used to assess the association. 5. Jones, C. U. et al. Radiotherapy and short-term androgen deprivation for localized prostate cancer. N. Engl. J. Med. 365, 107–118 (2011). determined by the investigators of each contributing study using cancer registries and/or medical records (Supplementary Information). At last follow-up, 3983 men had died from prostate cancer, 5806 had died from non-prostate cancer causes, and 70,702 were still alive. The median age at the last follow-up was 70 years (IQR: 63–76). As before, Cox proportional hazards models and sensitivity analysis were used to assess the association. p g 6. NCCN Clinical Practice Guidelines in Oncology. Prostate Cancer. Version 1.2019. , g p y ( Q 63–76). As before, Cox proportional hazards models and sensitivity analysis were used to assess the association. 7. Grossman, D. C. et al. Screening for prostate cancer: US Preventive Services Task Force recommendation statement. J. Am. Med. Assoc. 319, 1901–1913 (2018). 8. Wolf, A. M. D. et al. American Cancer Society Guideline for the early detection of prostate cancer: update 2010 CA Cancer J Clin 60 70 98 (2 8. Wolf, A. M. D. et al. American Cancer Society Guideline for the early detection of prostate cancer: update 2010. CA Cancer J. Clin. 60, 70–98 (2010) 8. Wolf, A. M. D. et al. American Cancer Society Guideline for the early detection of prostate cancer: update 2010. CA Cancer J. Clin. 60, 70–98 (2010). PHS and family history. Prostate cancer family history was also tested for asso- ciation with any, aggressive, or fatal prostate cancer. Information on family history was standardized across studies included in PRACTICAL consortium data. A family history of prostate cancer was deﬁned as the presence or absence of a ﬁrst- degree relative with a prostate cancer diagnosis. There were 46,030 men with available prostate cancer family history data. 9. Ilic, D., Neuberger, M. M., Djulbegovic, M. & Dahm, P. Screening for prostate cancer. Cochrane Database Syst. Rev. 2013, CD004720 (2013). 10. Stangelberger A., Waldert M., Djavan B. Prostate cancer in elderly men. Rev. Data availability PRACTICAL consortium data are available upon request to the Data Access Committee (http://practical.icr.ac.uk/blog/?page_id=135). Questions and requests for further information may be directed to PRACTICAL@icr.ac.uk. All other data are available within the Article, Supplementary information, or upon request to the authors. PRACTICAL consortium data are available upon request to the Data Access Committee (http://practical.icr.ac.uk/blog/?page_id=135). Questions and requests for further 27. Rotimi, C. N. et al. The genomic landscape of African populations in health and disease. Hum. Mol. Genet. 26, R225–R236 (2017). information may be directed to PRACTICAL@icr.ac.uk. All other data are available within the Article, Supplementary information, or upon request to the authors. 28. Campbell, M. C. & Tishkoff, S. A. African genetic diversity: implications for human demographic history, modern human origins, and complex disease mapping. Annu. Rev. Genomics Hum. Genet. 9, 403–433 (2008). Explorations of alternative ancestry groupings OncoArray-deﬁned genetic ancestry was mostly concordant with self-reported race/ethnicity. Participants with other/unknown self-reported race/ethnicity were mostly grouped into OncoArray’s European genetic ancestry. Additional details are shown in the Supplementary Information. 21. Duncan L., et al. Analysis of polygenic risk score usage and performance in diverse human populations. Nat Commun. https://doi.org/10.1038/s41467- 019-11112-0 (2019). Participants with other/unknown self-reported race/ethnicity were mostly grouped into OncoArray’s European genetic ancestry. Additional details are shown in the Supplementary Information. Participants with other/unknown self-reported race/ethnicity were mostly grouped into OncoArray’s European genetic ancestry. Additional details are shown in the Supplementary Information. 22. Petrovski S., Goldstein D. B. Unequal representation of genetic variation across ancestry groups creates healthcare inequality in the application of precision medicine. Genome Biol. https://doi.org/10.1186/s13059-016-1016-y (2016). 23. Grinde, K. E. et al. Generalizing polygenic risk scores from Europeans to Hispanics/Latinos. Genet. Epidemiol. 43, 50–62 (2019). Self-reported race/ethnicity. Finally, we also evaluated PHS performance for association with aggressive prostate cancer using participants’ self-reported race/ethnicity. 24. Popejoy, A. B. & Fullerton, S. M. Genomics is failing on diversity. Nature 538, 161–164 (2016). 25. DeSantis, C. E. et al. Cancer statistics for African Americans, 2016: progress and opportunities in reducing racial disparities. CA Cancer J. Clin. 66, 290–308 (2016). Reporting summary. Further information on research design is available in the Nature Research Reporting Summary linked to this article. 26. Tsodikov, A. et al. Is prostate cancer different in black men? Answers from three natural history models. Cancer 123, 2312 (2017). ARTICLE Urol. http://www.ncbi.nlm.nih.gov/pubmed/18660852 (2008). 11. Leitzmann M. F., Rohrmann S. Risk factors for the onset of prostatic cancer: age, location, and behavioral correlates. Clin. Epidemiol. https://doi.org/ 10.2147/CLEP.S16747 (2012). Cox proportional hazards models were used to assess family history for association with any, aggressive, or fatal prostate cancer. To evaluate the relative importance of each, a multivariable model using both family history and PHS was compared to using family history alone (log-likelihood test; α = 0.01). HRs were calculated for each variable. 12. Kattan M. W., et al. American Joint Committee on Cancer acceptance criteria for inclusion of risk models for individualized prognosis in the practice of precision medicine. CA Cancer J Clin. https://doi.org/10.3322/caac.21339 (2016). 13. Seibert, T. M. et al. Polygenic hazard score to guide screening for aggressive prostate cancer: development and validation in large scale cohorts. Br. Med. J. 360, 1–7 (2018). Code availability 29. Gomez F., Hirbo J., Tishkoff S. A. Genetic variation and adaptation in Africa: Implications for human evolution and disease. Cold Spring Harb. Perspect. Biol. https://doi.org/10.1101/cshperspect.a008524 (2014). y Code used for this work has been made available along with this paper (Supplementary Software 1). Code used for this work has been made available along with this paper (Supplementary Software 1). 30. Karunamuni R., et al. African-speciﬁc improvement of a polygenic hazard score for age at diagnosis of prostate cancer. Int. J Cancer. https://doi.org/ 10.1101/2020.04.20.20072926 (2020). Received: 15 May 2020; Accepted: 12 January 2021; Received: 15 May 2020; Accepted: 12 January 2021; Received: 15 May 2020; Accepted: 12 January 2021; 31. NCCN Guidelines Version 1.2019 Older Adult Oncology. (2019). 32. Bechis S. K., Carroll P. R., Cooperberg M. R. Impact of age at diagnosis on prostate cancer treatment and survival. J. Clin. Oncol. https://doi.org/10.1200/ JCO.2010.30.2075 (2011). 33. Huynh-Le, M. P. et al. Age dependence of modern clinical risk groups for localized prostate cancer—a population-based study. Cancer 126, 1691–1699 (2020). Explorations of alternative ancestry groupings Agnostic genetic ancestry groupings with FastSTRUCTURE. The primary analyses, above, used OncoArray-deﬁned genetic ancestries, as prior reports have shown genetic ancestry may be more informative than self-reported race/ethnicities43. However, for the purpose of this study, the OncoArray-deﬁned categories may underestimate the impact of the inherent complexity of human genetic ancestry. Therefore, we further explored the impact of an array of alternative genetic ancestry subgroup deﬁnitions on PHS2 performance using fastSTRUCTURE55, which infers global admixture/ancestry via a Bayesian approach. We ran fas- tSTRUCTURE v1.0 on all individuals in the multi-ethnic dataset using approxi- mately 2300 ancestry informative markers and multiple (K) levels of population complexity to agnostically cluster the data into K = 2–6 populations. For each iteration of K populations, participants were placed into the cluster for which their maximum admixture proportion was ≥0.8. Those participants without a cluster for which their maximum admixture proportion was ≥0.8 were placed into a separate group termed “admixed.” The optimal number of clusters (K) for fastSTRUCTURE was chosen as that which maximized the marginal likelihood of the data55. PHS2 was evaluated for association with aggressive prostate cancer (HR80/20) after stra- tiﬁcation by each K population subgroup. 14. Witte, J. S. Personalized prostate cancer screening: improving PSA tests with genomic information. Sci. Transl. Med. 2, 62ps55 (2010). 15. Chen, H. et al. Adding genetic risk score to family history identiﬁes twice as many high-risk men for prostate cancer: results from the prostate cancer prevention trial. Prostate 76, 1120–1129 (2016). 16. Michailidou, K. et al. Large-scale genotyping identiﬁes 41 new loci associated with breast cancer risk. Nat. Genet. 45, 353–361 (2013). 17. Fantus, R. J. & Helfand, B. T. Germline genetics of prostate cancer: time to incorporate genetics into early detection tools. Clin. Chem. 65, 74–79 (2019). 18. Schumacher, F. R. et al. Association analyses of more than 140,000 men identify 63 new prostate cancer susceptibility loci. Nat. Genet. 50, 928–936 (2018). 19. Benaﬁf S., Kote-Jarai Z., Eeles R. A. A review of prostate cancer Genome-Wide Association Studies (GWAS). Cancer Epidemiol. Biomarkers Prev. https://doi. org/10.1158/1055-9965.EPI-16-1046 (2018). 20. Martin, A. R. et al. Clinical use of current polygenic risk scores may exacerbate health disparities. Nat. Genet. 51, 584–591 (2019). A comparison of fastSTRUCTURE clustering, OncoArray-determined genetic ancestry, and self-reported race/ethnicity was compiled. OncoArray-deﬁned genetic ancestry was mostly concordant with self-reported race/ethnicity. A comparison of fastSTRUCTURE clustering, OncoArray-determined genetic ancestry, and self-reported race/ethnicity was compiled. Methods P ti i t OncoArray-deﬁned genetic ancestry All European Asian African Participants Controls 30,575 26,377 1185 3013 Prostate cancer cases 49,916 45,479 1197 3240 Aggressive prostate cancer casesa 26,419 24,279 716 1424 Fatal prostate cancer cases 3983 3908 57 18 Number of participants with known ﬁrst-degree family history information Family history of prostate cancer available (prostate cancer cases; controls) 46,030 (28,204; 17,826) 39,445 (24,921; 14,524) 1,028 (519; 509) 5,557 (2,764; 2,793) Age demographics Median age, at diagnosis (IQR) 65 [60–71] 66 [60–71] 68 [62–74] 62 [56–68] Median age, at last follow up (IQR) 70 [63–76] 70 [64–77] 70 [63–76] 62 [56–68] aAggressive prostate cancer deﬁned as: Gleason scores ≥7, PSA ≥10 ng/mL, T3–T4 stage, nodal metastases, or distant metastases. IQR interquartile range. Table 6 Participant characteristics, n = 80,491. NATURE COMMUNICATIONS \| (2021) 12:1236 \| https://doi.org/10.1038/s41467-021-21287-0 \| www.nature.com/naturecommunicatio 5 Reprints and permission information is available at http://www.nature.com/reprints Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional afﬁliations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/ licenses/by/4.0/. ARTICLE ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-021-21287-0 B.G.N., F.W., C.M.T., G.G.G., A.W., D.A., R.C.T., W.J.B., W.Z., M.S., J.L.S., L.A.M., C.M.L.W., A.S.K., O.C., S.I.B., S.K., K.D.S., C.C., E.M.G., F.M., K.-T.K., J.Y.P., S.A.I., C.M., R.J.M., S.N.T., B.S.R., T.-J.L., S.W., A.V., M.K., K.L.P., C.H., M.R.T., L.M., R.J.L., L. C.-A., H.B., E.M.J., R.K., C.J.L., S.L.N., K.D.R., H.P., A.R., L.F.N., J.H.F., M.G., N.U., F.C., M.G.-D., P.A.T., W.S.B., M.J.R., M.E.P., J.J.H., and T.M.S. acquired the data. M.-P.H.-L. and T.M.S. wrote the original drafts of the paper and Supplementary Information. All authors approved the ﬁnal version of the paper and Supplementary Information. B.G.N., F.W., C.M.T., G.G.G., A.W., D.A., R.C.T., W.J.B., W.Z., M.S., J.L.S., L.A.M., C.M.L.W., A.S.K., O.C., S.I.B., S.K., K.D.S., C.C., E.M.G., F.M., K.-T.K., J.Y.P., S.A.I., C.M., R.J.M., S.N.T., B.S.R., T.-J.L., S.W., A.V., M.K., K.L.P., C.H., M.R.T., L.M., R.J.L., L. C.-A., H.B., E.M.J., R.K., C.J.L., S.L.N., K.D.R., H.P., A.R., L.F.N., J.H.F., M.G., N.U., F.C., M.G.-D., P.A.T., W.S.B., M.J.R., M.E.P., J.J.H., and T.M.S. acquired the data. M.-P.H.-L. and T.M.S. wrote the original drafts of the paper and Supplementary Information. All authors approved the ﬁnal version of the paper and Supplementary Information. 37. Ankerst, D. P. et al. Prostate cancer prevention trial risk calculator 2.0 for the prediction of low- vs high-grade prostate cancer. Urology 83, 1362–1367 (2014). 37. Ankerst, D. P. et al. Prostate cancer prevention trial risk calculator 2.0 for the prediction of low- vs high-grade prostate cancer. Urology 83, 1362–1367 (2014). p g g p gy 38. Brentnall, A. R., Cuzick, J., Buist, D. S. M. & Bowles, E. J. A. Long-term accuracy of breast cancer risk assessment combining classic risk factors and breast density. JAMA Oncol. 4, e180174 (2018). M.G.-D., P.A.T., W.S.B., M.J.R., M.E.P., J.J.H., and T.M.S. acquired the data. M.-P.H.-L. and T.M.S. wrote the original drafts of the paper and Supplementary Information. All authors approved the ﬁnal version of the paper and Supplementary Information. y 39. Yeh, H. C., Duncan, B. B., Schmidt, M. I., Wang, N. Y. & Brancati, F. L. Smoking, smoking cessation, and risk for type 2 diabetes mellitus: a cohort study. Ann. Intern. Med. 152, 10–17 (2010). 40. Wang, T. J. et al. Plasma natriuretic peptide levels and the risk of cardiovascular events and death. N. Engl. J. Med. 350, 655–663 (2004). Additional information 55. Raj, A., Stephens, M. & Pritchard, J. K. FastSTRUCTURE: variational inference of population structure in large SNP data sets. Genetics 197, 573–589 (2014). Supplementary information The online version contains supplementary material available at https://doi.org/10.1038/s41467-021-21287-0. g 56. Klein J. P., Houwelingen H. C., Ibrahim J. G. S. T., ed. Handbook of Survival Analysis. (Chapman and Hall/CRC, London, 2013). Correspondence and requests for materials should be addressed to T.M.S. Correspondence and requests for materials should be addressed to T.M.S. Peer review information Nature Communications thanks Ewan Birney and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available. Competing interests A.M. Dale and T.M. Seibert report a research grant from the US Department of Defense. O.A. Andreassen reports research grants from K.G. Jebsen Stiftelsen, Research Council of Norway, and South East Norway Health Authority. N. Usmani reports grants from Astra Zeneca and Astellas, research collaboration, and ﬁnancial in-kind support from Best Medical Canada and Concure Oncology. R.M. Martin reports grants from Cancer Research UK, during the conduct of the study. K.D. Sørensen reports grants from Danish Cancer Society, grants from Velux Foundation, during the conduct of the study. T.M. Seibert reports honoraria from Multimodal Imaging Services Corporation for imaging segmen- tation and honoraria from Varian Medical Systems and WebMD, Inc. for educational content. A.S. Kibel reports advisory board memberships for Sanoﬁ-Aventis, Dendreon, and Profound. R.A. Eeles reports honoraria from GU-ASCO, honoraria/speaker fees from Janssen, honoraria from an invited talk to the University of Chicago, and educational honoraria from Bayer&Ipsen. K.D. Sørensen reports personal fees from AstraZeneca, personal fees from Sanoﬁ, outside the submitted work. N. Usmani reports honoraria from Janssen Canada and Bayer, outside the submitted work. M. Gamulin reports speaker/ advisor board/travel fees for BMS, Pﬁzer, Novartis, Astellas, Sanoﬁ, Janssen, Roche, Sandoz, Amgen, Bayer, PharmaSwiss, MSD, Alvogen. M. Gamuli also reports non-ﬁnancial report for drugs from BMS, Roche, Janssen. A.M. Dale has additional disclosures outside the present work: founder, equity holder, and advisory board member for CorTechs Labs, Inc.; advisory board member of Human Longevity, Inc.; recipient of nonﬁnancial research support from General Electric Healthcare. K.D. Sørensen is co-inventor on an issued patent (“Biomarkers for prostate cancer”/# US10106854B2, # AU2013275761B2, # JP6242388B2) licensed to Qiagen, on an issued patent (“A microRNA-based method for early detection of prostate cancer in urine samples”/# US10400288B2, # EP3256602B1, # ES2749651T3) licensed to Qiagen, and on an issued patent (“A microRNA-based method for assessing the prognosis of a prostate cancer patient”/# US10358681B2, # EP3262186B1, # ES2724404T3, #JP6769979B2), licensed to Qiagen. N. Usmani has a patent (US Provisional Patent Application No. 62/688,481: “Theranostic radiophotodynamic therapy nanoparticles”) pending, and a patent (US Patent Application No. 15/978,996: “Hand-held device and computer-implemented system and method for assisted steering of a percutaneously inserted needle”) pending. The remaining authors declare no competing interests. Addi- tional acknowledgments for the PRACTICAL consortium and contributing studies are described in the Supplemental Material. 41. Yang, X. et al. Evaluation of polygenic risk scores for ovarian cancer risk prediction in a prospective cohort study. J. Competing interests Med. Genet. 55, 546–554 (2018). p p p y 42. Torkamani, A., Wineinger, N. E. & Topol, E. J. The personal and clinical utility of polygenic risk scores. Nat. Rev. Genet. 19, 581–590 (2018). 43. Marini S., et al. Comparison of genetic and self-identiﬁed ancestry in modeling intracerebral hemorrhage risk. Front. Neurol. https://doi.org/ 10.3389/fneur.2018.00514 (2018). 44. Haiman C. A., et al. Characterizing genetic risk at known prostate cancer susceptibility loci in African Americans. PLoS Genet. https://doi.org/10.1371/ journal.pgen.1001387 (2011). j pg 45. Han, Y. et al. Generalizability of established prostate cancer risk variants in men of African ancestry. Int. J. Cancer 136, 1210–1217 (2015). 46. Cheng, I. et al. Evaluating genetic risk for prostate cancer among Japanese and Latinos. Cancer Epidemiol. Biomark. Prev. 21, 2048–2058 (2012). 47. Amos, C. I. et al. The OncoArray consortium: a network for understanding the genetic architecture of common cancers. Cancer Epidemiol. Biomark. Prev. 26, 126–135 (2017). 48. Kote-Jarai, Z. et al. Multiple novel prostate cancer predisposition loci conﬁrmed by an international study: the PRACTICAL consortium. Cancer Epidemiol. Biomark. Prev. 17, 2052–2061 (2008). p 49. R Core Team. R: A Language and Environment for Statistical Computing. (R Foundation for Statistical Computing, Vienna, Austria, 2015). 50. Efron, B. Bootstrap methods: another look at the jackknife. Ann. Stat. 7, 1–26 (1979). 51. Efron B., Tibshirani R. Bootstrap Methods for Standard Errors, Conﬁdence Intervals, and Other Measures of Statistical Accuracy. https://about.jstor.org/ terms (1986). 52. Li Y., et al. FastPop: A rapid principal component derived method to infer intercontinental ancestry using genetic data. BMC Bioinform. https://doi.org/ 10.1186/s12859-016-0965-1 (2016). 53. Therneau, T. M. & Li, H. Computing the Cox Model for Case Cohort Designs. Lifetime Data Anal. 5, 99–112 (1999). 54. Karunamuni R. A., et al. The effect of sample size on polygenic hazard models for prostate cancer. Eur. J. Hum. Genet. https://doi.org/10.1038/s41431-020- 0664-2 (2020). Acknowledgements Acknowledgments for the PRACTICAL consortium and contributing studies are described in the Supplementary Material. This study was funded in part by a grant from the United States National Institute of Health/National Institute of Biomedical Imaging and Bioengineering (#K08EB026503), United States Department of Defense (#W81XWH-13-1-0391), University of California CRCC C21CR2060, the Research Council of Norway (#223273), K.G. Jebsen Stiftelsen, and South East Norway Health Authority. RM Martin is supported in part by the National Institute for Health Research Bristol Biomedical Research Centre. The CAP trial is funded by Cancer Research UK and the UK Department of Health (C11043/A4286, C18281/A8145, C18281/A11326, C18281/A15064, and C18281/A24432). R.M. Martin was supported by a Cancer Research UK (C18281/A19169) program grant (the Integrative Cancer Epidemiology Programme). The views and opinions expressed by authors in this publication are those of the authors and do not necessarily reﬂect those of the UK National Institute for Health Research (NIHR) or the Department of Health and Social Care. The content is solely the responsibility of the authors and does not necessarily represent the ofﬁcial views of any of the funding agencies, who had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. Funding for the PRACTICAL consortium member studies is detailed in the Supplementary Information. Reprints and permission information is available at http://www.nature.com/reprints Author contributions M.-P.H.-L., C.C.F., R.K., W.K.T., I.G.M., O.A.A., A.M.D., and T.M.S. designed the study concept, created the methodology, and analyzed/interpreted the data. M.-P.H.-L., M.E. M., R.A.E., Z.K.-J., K.M., J.S., N.P., J.B., H.G., D.E.N., J.L.D., F.C.H., R.M.M., S.F.N., © The Author(s) 2021 References Re e e ces 1. Torre, L. A. et al. Global cancer statistics, 2012. CA Cancer J. Clin. 65, 87–108 (2015). 2. Hamdy, F. C. et al. 10-Year outcomes after monitoring, surgery, or radiotherapy for localized prostate cancer. N. Engl. J. Med. 375, 1415–1424 (2016). 3. Bill-Axelson, A. et al. Radical prostatectomy or watchful waiting in prostate cancer—29-year follow-up. N. Engl. J. Med. 379, 2319–2329 (2018). 4. Bolla, M. et al. Duration of androgen suppression in the treatment of prostate cancer. N. Engl. J. Med. 360, 2516–2527 (2009). 1. Torre, L. A. et al. Global cancer statistics, 2012. CA Cancer J. Clin. 65, 87–108 (2015). 1. Torre, L. A. et al. Global cancer statistics, 2012. CA Cancer J. Clin. 65, 87–108 (2015). 34. Huynh-Le, M.-P. et al. A genetic risk score to personalize prostate cancer screening, applied to population data. Cancer Epidemiol. Biomark. Prev. 29, 1731–1738 (2020). ( ) 2. Hamdy, F. C. et al. 10-Year outcomes after monitoring, surgery, or radiotherapy for localized prostate cancer. N. Engl. J. Med. 375, 1415–1424 (2016). 2. Hamdy, F. C. et al. 10-Year outcomes after monitoring, surgery, or radiotherapy for localized prostate cancer. N. Engl. J. Med. 375, 1415–1424 (2016). 35. Pettersson A., Robinson D., Garmo H., Holmberg L., Stattin P. Age at diagnosis and prostate cancer treatment and prognosis: a population-based cohort study. Ann. Oncol. https://doi.org/10.1093/annonc/mdx742 (2018). 3. Bill-Axelson, A. et al. Radical prostatectomy or watchful waiting in prostate cancer—29-year follow-up. N. Engl. J. Med. 379, 2319–2329 (2018). 36. Giri, V. N. & Beebe-Dimmer, J. L. Familial prostate cancer. Semin Oncol. 43, 560–565 (2016). y g 4. Bolla, M. et al. Duration of androgen suppression in the treatment of prostate cancer. N. Engl. J. Med. 360, 2516–2527 (2009). NATURE COMMUNICATIONS \| (2021) 12:1236 \| https://doi.org/10.1038/s41467-021-21287-0 \| www.nature.com/naturecommunicatio 6 © The Author(s) 2021 7 7 NATURE COMMUNICATIONS \| (2021) 12:1236 \| https://doi.org/10.1038/s41467-021-21287-0 \| www.nature.com/naturecommunications 1Department of Radiation Medicine and Applied Sciences, University of California San Diego, La Jolla, CA, USA. 2Center for Multimodal Imaging and Genetics, University of California San Diego, La Jolla, CA, USA. 3Division of Biostatistics and Halicioğlu Data Science Institute, University of California San Diego, La Jolla, CA, USA. 4Department of Family Medicine and Public Health, University of California San Diego, La Jolla, CA, USA 5Moores Cancer Center, Department of Family Medicine and Public Health, University of California San Diego, La Jolla, CA, USA. 6The Institute of Cancer Research, London, UK. 7Royal Marsden NHS Foundation Trust, London, UK. 8Division of Population Health, Health Services Research and Primary Care, University of Manchester, Oxford Road, Manchester, UK. 9Warwick Medical School, University of Warwick, Coventry, UK. 10Institute of Biomedicine, Kiinamyllynkatu 10, FI-20014 University of Turku, Turku, Finland. 11Department of Medical Genetics, Genomics, Laboratory Division Turku University Hospital, Turku, Finland. 12University College London, Department of Applied Health Research, London, UK. 13Centre for Cancer Genetic Epidemiology, Department of Oncology, University of Cambridge, Strangeways Laboratory, Worts Causeway, Cambridge, UK. 14Department of Applied Health Research, University College London, London, UK. 15Australian Prostate Cancer Research Centre-Qld, Institute of Health and Biomedical Innovation and School of Biomedical Sciences, Queensland University of Technology, Brisbane, QLD, Australia. 16Translational Research Institute, Brisbane, QLD, Australia. 17Department of Medical Epidemiology and Biostatistics, Karolinska Institute, Stockholm, Sweden. 18Nufﬁeld Department of Surgical Sciences, University of Oxford, John Radcliffe Hospital, Headington, Oxford, UK. 19Department of Oncology, University of Cambridge, Addenbrooke’s Hospital, Cambridge, UK. 20Cancer Research UK, Cambridge Research Institute, Li Ka Shing Centre, Cambridge, UK. 21Population Health Sciences, Bristol Medical School, University of Bristol, Bristol, UK. 22Nufﬁeld Department of Surgical Sciences, University of Oxford, Oxford, UK. 23Faculty of Medical Science, University of Oxford, John Radcliffe Hospital, Oxford, UK. 24National Institute for Health Research (NIHR) Biomedical Research Centre, University of Bristol, Bristol, UK. 25Medical Research Council (MRC) Integrative Epidemiology Unit, University of Bristol, Bristol, UK. 26Faculty of Health and Medical Sciences, University of Copenhagen Copenhagen, Denmark. 27Department of Clinical Biochemistry, Herlev and Gentofte Hospital, Copenhagen University Hospital, Herlev, Copenhagen, Denmark. 28Faculty of Health and Medical Sciences, University of Copenhagen, Copenhagen, Denmark. 29Department of Clinical Biochemistry, Herlev and Gentofte Hospital, Copenhagen University Hospital, Herlev, Copenhagen, Denmark. 30Department of Medical Epidemiology and Biostatistics, Karolinska Institute, Stockholm, Sweden. 31SWOG Statistical Center, Fred Hutchinson Cancer Research Center, Seattle, WA, USA. 32Cancer Epidemiology Division, Cancer Council Victoria, Melbourne, VIC, Australia. 33Centre for Epidemiology and Biostatistics Melbourne School of Population and Global Health, The University of Melbourne, Parkville, VIC, Australia. 34Precision Medicine, School of Clinica Sciences at Monash Health, Monash University, Clayton, VIC, Australia. 35Division of Nutritional Epidemiology, Institute of Environmental Medicine Karolinska Institutet, Stockholm, Sweden. 36Department of Surgical Sciences, Uppsala University, Uppsala, Sweden. 37Division of Cancer Epidemiology and Genetics, National Cancer Institute, NIH, Bethesda, MD, USA. 38Cancer Epidemiology Unit, Nufﬁeld Department of Population Health, University of Oxford, Oxford, UK. 39Division of Epidemiology, Department of Medicine, Vanderbilt University Medical Center, Nashville, TN USA. 40International Epidemiology Institute, Rockville, MD, USA. 41Division of Epidemiology, Department of Medicine, Vanderbilt University Medical Center, Nashville, TN, USA. 42Department of Family and Community Medicine, Meharry Medical College, Nashville, TN, USA. 43Division of Public Health Sciences, Fred Hutchinson Cancer Research Center, Seattle, WA, USA. 44Department of Epidemiology, School of Public Health, University of Washington, Seattle, WA, USA. 45Department of Epidemiology, Harvard T. H. Chan School of Public Health, Boston, MA, USA. 46Division of Cancer Sciences, University of Manchester, Manchester Academic Health Science Centre, Radiotherapy Related Research, The Christie Hospital NHS Foundation Trust, Manchester, UK. 47Division of Urologic Surgery, Brigham and Womens Hospital, Boston, MA, USA. 48Sorbonne Universite, GRC n°5, AP-HP, Tenon Hospital, 4 Rue de la Chine, Paris, France. 49CeRePP, Tenon Hospital, Paris, France. 50Division of Cancer Epidemiology and Genetics, National Cancer Institute, NIH, Bethesda, MD, USA. 51Department of Molecular Medicine, Aarhus University Hospital, Aarhus, Denmark. 52Department of Clinical Medicine, Aarhus University, Aarhus, Denmark. 53International Hereditary Cancer Center, ARTICLE Minh-Phuong Huynh-Le 1,2, Chun Chieh Fan 2, Roshan Karunamuni1,2, Wesley K. Thompson3,4, Maria Elena Martinez5, Rosalind A. Eeles 6,7, Zsoﬁa Kote-Jarai6, Kenneth Muir 8,9, Johanna Schleutker 10,11, Nora Pashayan 12,13,14, Jyotsna Batra 15,16, Henrik Grönberg 17, David E. Neal18,19,20, Jenny L. Donovan21, Freddie C. Hamdy22,23, Richard M. Martin 21,24,25, Sune F. Nielsen26,27, Børge G. Nordestgaard 28,29, Fredrik Wiklund 30, Catherine M. Tangen31, Graham G. Giles 32,33,34, Alicja Wolk 35,36, Demetrius Albanes37, Ruth C. Travis 38, William J. Blot39,40, Wei Zheng 41, Maureen Sanderson42, Janet L. Stanford43,44, Lorelei A. Mucci45, Catharine M. L. West 46, Adam S. Kibel47, Olivier Cussenot48,49, Sonja I. Berndt50, Stella Koutros50, Karina Dalsgaard Sørensen 51,52, Cezary Cybulski53, Eli Marie Grindedal54, Florence Menegaux55,56, Kay-Tee Khaw57, Jong Y. Park 58, Sue A. Ingles59, Christiane Maier60, Robert J. Hamilton61,62, Stephen N. Thibodeau63, Barry S. Rosenstein64,65, Yong-Jie Lu66, Stephen Watya67, Ana Vega 68,69,70, Manolis Kogevinas71,72,73, Kathryn L. Penney74, Chad Huff75, Manuel R. Teixeira 76,77, Luc Multigner 78, Robin J. Leach79, Lisa Cannon-Albright 80,81, Hermann Brenner82,83,84, Esther M. John85, Radka Kaneva86, Christopher J. Logothetis87, Susan L. Neuhausen88, Kim De Ruyck89, Hardev Pandha90, Azad Razack91, Lisa F. Newcomb43,92, Jay H. Fowke93,94, Marija Gamulin 95, Nawaid Usmani96,97, Frank Claessens98, Manuela Gago-Dominguez99,100, Paul A. Townsend101, William S. Bush 102, Monique J. Roobol 103, Marie-Élise Parent104,105, Jennifer J. Hu106, Ian G. Mills107, Ole A. Andreassen 108, Anders M. Dale2,109, Tyler M. Seibert 1,2,109,110✉, UKGPCS collaborators, APCB (Australian Prostate Cancer BioResource), NC-LA PCaP Investigators, The IMPACT Study Steering Committee and Collaborators, Canary PASS Investigators, The Proﬁle Study Steering Committee & The PRACTICAL Consortium Minh-Phuong Huynh-Le 1,2, Chun Chieh Fan 2, Roshan Karunamuni1,2, Wesley K. Thompson3,4, Maria Elena Martinez5, Rosalind A. Eeles 6,7, Zsoﬁa Kote-Jarai6, Kenneth Muir 8,9, Johanna Schleutker 10,11, Nora Pashayan 12,13,14, Jyotsna Batra 15,16, Henrik Grönberg 17, David E. Neal18,19,20, Jenny L. Donovan21, Freddie C. Hamdy22,23, Richard M. Martin 21,24,25, Sune F. Nielsen26,27, Børge G. Nordestgaard 28,29, Fredrik Wiklund 30, Catherine M. Tangen31, Graham G. Giles 32,33,34, Alicja Wolk 35,36, Demetrius Albanes37, Ruth C. Travis 38, William J. Blot39,40, Wei Zheng 41, Maureen Sanderson42, Janet L. Stanford43,44, Lorelei A. Mucci45, Catharine M. L. West 46, Adam S. Kibel47, Olivier Cussenot48,49, Sonja I. Berndt50, Stella Koutros50, Karina Dalsgaard Sørensen 51,52, Cezary Cybulski53, Eli Marie Grindedal54, Florence Menegaux55,56, Kay-Tee Khaw57, Jong Y. Park 58, Sue A. Ingles59, Christiane Maier60, Robert J. Hamilton61,62, Stephen N. Thibodeau63, Barry S. NATURE COMMUNICATIONS \| (2021) 12:1236 \| https://doi.org/10.1038/s41467-021-21287-0 \| www.nature.com/naturecommunications NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-021-21287-0 ARTICLE Department of Genetics and Pathology, Pomeranian Medical University, Szczecin, Poland. 54Department of Medical Genetics, Oslo University Hospital, Oslo, Norway. 55Cancer & Environment Group, Center for Research in Epidemiology and Population Health (CESP), INSERM, University Paris-Sud, University Paris-Saclay, Villejuif Cédex, France. 56Paris-Sud University, UMRS 1018, Villejuif Cedex, France. 57Clinical Gerontology Unit, University of Cambridge, Cambridge, UK. 58Department of Cancer Epidemiology, Mofﬁtt Cancer Center, Tampa, FL, USA. 59Department of Preventive Medicine, Keck School of Medicine, University of Southern California/Norris Comprehensive Cancer Center, Los Angeles, CA, USA. 60Humangenetik Tuebingen, Tuebingen, Germany. 61Department of Surgical Oncology, Princess Margaret Cancer Centre, Toronto, ON, Canada. 62Department of Surgery (Urology), University of Toronto, Toronto, ON, Canada. 63Department of Laboratory Medicine and Pathology, Mayo Clinic, Rochester, MN, USA. 64Department of Radiation Oncology and Department of Genetics and Genomic Sciences, Icahn School of Medicine at Mount Sinai, One Gustave L. Levy Place, New York, NY, USA. 65Department of Genetics and Genomic Sciences, Icahn School of Medicine at Mount Sinai, New York, NY, USA. 66Centre for Molecular Oncology, Barts Cancer Institute, Queen Mary University of London, John Vane Science Centre, Charterhouse Square, London, UK. 67Uro Care, Kampala, Uganda. 68Fundación Pública Galega Medicina Xenómica, Santiago De Compostela, Spain. 69Instituto de Investigación Sanitaria de Santiago de Compostela, Santiago De Compostela, Spain. 70Centro de Investigación en Red de Enfermedades Raras (CIBERER), Santiago De Compostela, Spain. 71ISGlobal, Barcelona, Spain. 72IMIM (Hospital del Mar Medical Research Institute), Barcelona, Spain. 73Universitat Pompeu Fabra (UPF), Barcelona, Spain. 74Channing Division of Network Medicine, Department of Medicine, Brigham and Women’s Hospital/Harvard Medical School, Boston, MA, USA. 75The University of Texas M. D. Anderson Cancer Center, Houston, TX, USA. 76Department of Genetics, Portuguese Oncology Institute of Porto (IPO-Porto), Porto, Portugal. 77Biomedical Sciences Institute (ICBAS), University of Porto, Porto, Portugal. 78Univ Rennes, Inserm, EHESP, Irset (Institut de Recherche en Santé, Environnement et Travail)— UMR_S 1085, Rennes, France. 79Department of Urology, Mays Cancer Center, University of Texas Health Science Center at San Antonio, San Antonio, TX, USA. 80Division of Epidemiology, Department of Internal Medicine, University of Utah School of Medicine, Salt Lake City, UT, USA. 81George E. Wahlen Department of Veterans Affairs Medical Center, Salt Lake City, UT, USA. 82Division of Clinical Epidemiology and Aging Research, German Cancer Research Center (DKFZ), Heidelberg, Germany. 83German Cancer Consortium (DKTK), German Cancer Research Center (DKFZ), Heidelberg, Germany. 84Division of Preventive Oncology, German Cancer Research Center (DKFZ) and National Center for Tumor Diseases (NCT), Im Neuenheimer Feld 460, Heidelberg, Germany. 85Department of Medicine, Division of Oncology, Stanford Cancer Institute, Stanford University School of Medicine, Stanford, CA, USA. 86Molecular Medicine Center, Department of Medical Chemistry and Biochemistry, Medical University of Soﬁa, Soﬁa, Bulgaria. 87The University of Texas M. D. Anderson Cancer Center, Department of Genitourinary Medical Oncology, Houston, TX, USA. 88Department of Population Sciences, Beckman Research Institute of the City of Hope, Duarte, CA, USA. 89Ghent University, Faculty of Medicine and Health Sciences, Basic Medical Sciences, Gent, Belgium. 90The University of Surrey, Guildford, Surrey, UK. 91Department of Surgery, Faculty of Medicine, University of Malaya, Kuala Lumpur, Malaysia. 92Department of Urology, University of Washington, Seattle, WA, USA. 93Department of Medicine and Urologic Surgery, Vanderbilt University Medical Center, Nashville, TN, USA. 94Division of Epidemiology, Department of Preventive Medicine, The University of Tennessee Health Science Center, Memphis, TN, USA. 95Department of Oncology, University Hospital Centre Zagreb, University of Zagreb, School of Medicine, Zagreb, Croatia. 96Department of Oncology, Cross Cancer Institute, University of Alberta, Edmonton, Alberta, Canada. 97Division of Radiation Oncology, Cross Cancer Institute, Edmonton, Alberta, Canada. 98Department of Cellular and Molecular Medicine, Molecular Endocrinology Laboratory, KU Leuven, Leuven, Belgium. 99Genomic Medicine Group, Galician Foundation of Genomic Medicine, Instituto de Investigacion Sanitaria de Santiago de Compostela (IDIS), Complejo Hospitalario Universitario de Santiago, Servicio Galego de Saúde, SERGAS, Santiago de Compostela, Spain. 100University of California San Diego, Moores Cancer Center, La Jolla, CA, USA. 101Division of Cancer Sciences, Manchester Cancer Research Centre, Faculty of Biology, Medicine and Health, Manchester Academic Health Science Centre, NIHR Manchester Biomedical Research Centre, Health Innovation Manchester, University of Manchester, Manchester, UK. 102Case Western Reserve University, Department of Population and Quantitative Health Sciences, Cleveland Institute for Computational Biology, Cleveland, OH, USA. 103Department of Clinical Chemistry, Erasmus University Medical Center, Rotterdam, The Netherlands. 104Epidemiology and Biostatistics Unit, Centre Armand-Frappier Santé Biotechnologie, Institut National de la Recherche Scientiﬁque, Laval, QC, Canada. 105Department of Social and Preventive Medicine, School of Public Health, University of Montreal, Montreal, QC, Canada. 106The University of Miami School of Medicine, Sylvester Comprehensive Cancer Center, Miami, FL, USA. 107Nufﬁeld Department of Surgical Sciences, University of Oxford, Oxford, UK. 108NORMENT, KG Jebsen Centre, Oslo University Hospital and University of Oslo, Oslo, Norway. 109Department of Radiology, University of California San Diego, La Jolla, CA, USA. 110Department of Bioengineering, University of California San Diego, La Jolla, CA, USA. Lists of members and their afﬁliations appear in the Supplementary Information. ✉email: tseibert@ucsd.edu ARTICLE Rosenstein64,65, Yong-Jie Lu66, Stephen Watya67, Ana Vega 68,69,70, Manolis Kogevinas71,72,73, Kathryn L. Penney74, Chad Huff75, Manuel R. Teixeira 76,77, Luc Multigner 78, Robin J. Leach79, Lisa Cannon-Albright 80,81, Hermann Brenner82,83,84, Esther M. John85, Radka Kaneva86, Christopher J. Logothetis87, Susan L. Neuhausen88, Kim De Ruyck89, Hardev Pandha90, Azad Razack91, Lisa F. Newcomb43,92, Jay H. Fowke93,94, Marija Gamulin 95, Nawaid Usmani96,97, Frank Claessens98, Manuela Gago-Dominguez99,100, Paul A. Townsend101, William S. Bush 102, Monique J. Roobol 103, Marie-Élise Parent104,105, Jennifer J. Hu106, Ian G. Mills107, Ole A. Andreassen 108, Anders M. Dale2,109, Tyler M. Seibert 1,2,109,110✉, UKGPCS collaborators, APCB (Australian Prostate Cancer BioResource), NC-LA PCaP Investigators, The IMPACT Study Steering Committee and Collaborators, Canary PASS Investigators, The Proﬁle Study Steering Committee & The PRACTICAL Consortium 8 NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-021-21287-0 ARTICLE NATURE COMMUNICATIONS \| https://doi.org/10.1038/s41467-021-21287-0 NATURE COMMUNICATIONS \| (2021) 12:1236 \| https://doi.org/10.1038/s41467-021-21287-0 \| www.nature.com/naturecommunicatio 9
https://openalex.org/W2049065363	https://escholarship.org/content/qt7p94v9gv/qt7p94v9gv.pdf?t=m82qqp	English	null	Behavioral states may be associated with distinct spatial patterns in electrocorticogram	Cognitive neurodynamics	2,010	cc-by	9,441	UC Berkeley UC Berkeley Previously Published Works Title Behavioral states may be associated with distinct spatial patterns in electrocorticogram. Permalink https://escholarship.org/uc/item/7p94v9gv Journal Cognitive neurodynamics, 5(1) ISSN 1871-4099 Authors Panagiotides, Heracles Freeman, Walter J J, III Holmes, Mark D et al. Publication Date 2011-11-13 Copyright Information This work is made available under the terms of a Creative Commons Attribution License, availalbe at https://creativecommons.org/licenses/by/3.0/ UC Berkeley UC Berkeley Previously Published Works Title Behavioral states may be associated with distinct spatial patterns in electrocorticogram. Permalink https://escholarship.org/uc/item/7p94v9gv Journal Cognitive neurodynamics, 5(1) ISSN 1871-4099 Authors Panagiotides, Heracles Freeman, Walter J J, III Holmes, Mark D et al. Publication Date 2011-11-13 Copyright Information This work is made available under the terms of a Creative Commons Attribution License, availalbe at https://creativecommons.org/licenses/by/3.0/ UC Berkeley UC Berkeley Previously Published Works Title Behavioral states may be associated with distinct spatial patterns in electrocorticogram. Permalink https://escholarship.org/uc/item/7p94v9gv Journal Cognitive neurodynamics, 5(1) ISSN 1871-4099 Authors Panagiotides, Heracles Freeman, Walter J J, III Holmes, Mark D et al. Publication Date 2011-11-13 Copyright Information This work is made available under the terms of a Creative Commons Attribution License, availalbe at https://creativecommons.org/licenses/by/3.0/ P i d Peer reviewed Peer reviewed UC Berkeley y UC Berkeley Previously Published Works Powered by the California Digital Library University of California eScholarship.org Cogn Neurodyn (2011) 5:55–66 DOI 10.1007/s11571-010-9139-4 RESEARCH ARTICLE Behavioral states may be associated with distinct spatial patterns in electrocorticogram Heracles Panagiotides • Walter J. Freeman • Mark D. Holmes • Dimitrios Pantazis Received: 30 December 2009 / Revised: 13 October 2010 / Accepted: 25 October 2010 / Published online: 13 November 2010 The Author(s) 2010. This article is published with open access at Springerlink.com Received: 30 December 2009 / Revised: 13 October 2010 / Accepted: 25 October 2010 / Published online: 13 November 2010 The Author(s) 2010. This article is published with open access at Springerlink.com Abstract To determine if behavioral states are associated with unique spatial electrocorticographic (ECoG) patterns, we obtained recordings with a microgrid electrode array applied to the cortical surface of a human subject. The array was constructed with the intent of extracting maximal spa- tial information by optimizing interelectrode distances. A 34-year-old patient with intractable epilepsy underwent intracranial ECoG monitoring after standard methods failed to reveal localization of seizures. During the 8-day period of invasive recording, in addition to standard clinical elec- trodes a square 1 9 1 cm microgrid array with 64 electrodes (1.25 mm separation) was placed on the right inferior tem- poral gyrus. Careful review of video recordings identiﬁed four extended naturalistic behaviors: reading, conversing on the telephone, looking at photographs, and face-to-face interactions. ECoG activity recorded with the microgrid that corresponded to these behaviors was collected and ECoG spatial patterns were analyzed. During periods of ECoG selected for analysis, no electrographic seizures or epileptiform patterns were present. Moments of maximal spatial variance are shown to cluster by behavior. Com- parisons between conditions using a permutation test reveal signiﬁcantly different spatial patterns for each behavior. We conclude that ECoG recordings obtained on the cortical surface with optimal high spatial frequency resolution reveal distinct local spatial patterns that reﬂect different behavioral states, and we predict that similar patterns will be found in many if not most cortical areas on which a microgrid is placed. Keywords Neural circuits Behavioral states Perception EEG Epilepsy Spatial EEG patterns D. Pantazis Signal & Image Processing Institute, Department of Electrical Engineering, University of Southern California, 3740 McClintock Ave, Los Angeles, CA 90089-2564, USA Introduction The aim of our project was to seek and measure spatial patterns of cortical electrical activity that we could correlate with intentional cognitive behaviors. Our choice of the electrocorticogram (ECoG) was based on prior success in animal studies. Spatial patterns were extracted from multi- channel ECoGs recorded from the sensory and limbic cortices of rabbits (Barrie et al. 1996), cats (Freeman et al. 2003c) and gerbils (Ohl et al. 2001) with high-density planar arrays of electrodes ﬁxed on the pial surfaces. Reinforce- ment training shaped the patterns, so that they could be classiﬁed with respect to the conditioned stimuli (CS) elic- iting them. The choice of ECoG rather than local ﬁeld potentials (LFP) and units from arrays of depth electrodes (Ghovanloo et al. 2003; Pang et al. 2005) was based on ease of surgical placement, immunity from tissue damage by avoid- ing insertions, and long-term stability of signals. An addi- tional advantage was the fact that the ECoG contained only contributions from neuron populations generating open H. Panagiotides (&) Department of Neurological Surgery, University of Washington, 1959 Paciﬁc Ave, Seattle, WA, USA e-mail: hercp@u.washington.edu W. J. Freeman Department of Molecular & Cell, Division of Neurobiology, University of California, Donner 101, Berkeley, CA, USA M. D. Holmes Regional Epilepsy Center, Department of Neurology, University of Washington, Harborview Medical Center, 325 Ninth Ave, Seattle, WA, USA M. D. Holmes Regional Epilepsy Center, Department of Neurology, University of Washington, Harborview Medical Center, 325 Ninth Ave, Seattle, WA, USA D. Pantazis In particular, the radial phase gradients (‘phase cones’) were readily documented, showing that the spatial resolution provided by the grid would be adequate for classiﬁcation of amplitude patterns. The principal difﬁculty encountered was the deﬁnition of an appropriate set of behaviors. The animal studies were designed to study the ECoGs of sen- sory cortices, with the exception of a study in cats, in which three 4 9 4 microgrids were ﬁxed on the visual, auditory and somatic cortices. Two electrodes were placed in the olfactory bulb, and a 2 9 8 array was ﬁxed on the en- torhinal cortex, taking advantage of the ﬂat surface of the bony tentorium that underlay the limbic cortex (Freeman et al. 2003c). That study revealed global amplitude patterns with an intermittently shared waveform. The goodness of the classiﬁcation was maximal when all 64 channels were used. The goodness declined signiﬁcantly when any com- ponent was removed, showing that widely separated areas of the forebrain synchronized repeatedly at frame rates in the theta range (Freeman and Burke 2003) and sustained classiﬁable spatial patterns. This result encouraged us to search for classiﬁable amplitude patterns in the temporal lobe. The lack of dependence of classiﬁcation on ECoG frequencies led us to adopt the magnitude of the spatial variance of amplitude as our criterion for selection of potentially classiﬁable ECoG segments. The performance of the microgrid in extracting spatial ECoG patterns was tested thoroughly in recordings from sensory cortices of animals that were trained to discriminate olfactory (Freeman and Grajski 1987), visual, auditory, and somatic CS (Freeman and Barrie 2000; Freeman and Rogers 2002). Spatial patterns of amplitude were shown to be classiﬁable with respect to conditioned stimuli in both classical and operant conditioning (Freeman and Viana Di Prisco 1986). The spatial patterns were found to occur in brief segments lasting 80–120 ms, which had high ampli- tude and high spatial variance of a common waveform but minimal temporal and spatial variance of the frequency of the waveform (Freeman 2004a, b). The classiﬁcation was by measuring the 64 amplitudes of the common waveform in each frame, constructing a 64 9 1 feature vector for each frame, expressing the vector as a point in 64-space, and forming clusters of points from multiple trials with each discriminated stimulus. D. Pantazis 12 3 3 Cogn Neurodyn (2011) 5:55–66 56 ﬁelds, whereas the LFP contained far more complex mix- tures of open and closed ﬁelds (Chap. 4 in: Freeman 1975). frequencies. Every frame had the same waveform with the same carrier frequency but with spatial amplitude modula- tions of the amplitudes and phases. The effectiveness of the microgrid for extracting the spatial differences of both the amplitude and phase of the high-frequency components of the ECoG was measured by calculating the point spread function (PSF). This was the bell-shaped surface distribu- tion of dendritic potential above an active point in cortex) that was predicted for activation of a single dendrite or a cortical column (Fig. 2 in: Freeman 2006). The results simulated the full range of observed spatial differences in amplitude modulation of the shared waveform. The most important experimental parameter to set in the search for classiﬁable spatial patterns in the ECoG was the distance between the electrodes in the array. The search for spatial patterns differed from the search for temporal sig- nals. The main requirement for identifying a temporal signal was that the digitizing rate had to be high enough to capture the highest temporal frequency components, thereby avoiding undersampling and aliasing. The requirement for measuring the spatial patterns was sampling at intervals sufﬁciently small to avoid aliasing but large enough to avoid oversampling. The criterion for an optimal interval was based on calculation of the spatial power spectral density (PSDX) of human ECoG (Freeman et al. 2000) from 64 ECoG signals recorded with a 64 9 1 linear array on the superior temporal gyrus (0.5 mm spacing, 32 mm length, giving a spectral range from 0.016 to 1 cycle/mm). The PSDX showed a power-law (linear) decrease in log power with increasing log spatial frequency up to a spatial fre- quency of 0.4 cycles/mm, above which the spectrum was ﬂat. Activity above that inﬂection was treated as noise. By the Nyquist criterion the optimal sampling frequency was 0.8 cycles/mm, giving a spacing of 1.25 mm. An 8 9 8 array was 10 9 10 mm in width, which ﬁt onto many single gyri in the human brain without bridging a sulcus. p The properties of the phase of the ECoG that were derived from animals were replicated in the analysis of ECoG from a human subject (Freeman et al. 2006a, b). 123 Subject The subject, a 34-year-old woman with a history of intrac- table partial complex seizures, had subdural strip electrodes placed over lateral and basal-temporal regions for pre- operative evaluation prior to surgical intervention. In addi- tion to the clinical electrodes, a high-density electrode microgrid through a right temporal burr hole was placed onto the anterior surface of the right inferior temporal lobe gyrus. The patient underwent continuous ECoG and behavior monitoring for 8 days; the length of recording time was determined solely by clinical consideration. The microgrid remained in place during the entire 8-day period, as did the clinical electrodes. No complications were observed during the procedure. The data were collected in the EEG and Clinical Neurophysiology laboratory of Har- borview Hospital, University of Washington in Seattle. Data collection and management were in agreement with the protocol approved by the University’s Human Subjects Review Committee. The patient gave informed consent for the placement and use of the clinical electrodes and the microgrid. D. Pantazis resulted in inappropriate vocal and facial social signals, inability to recognize social signals of other monkeys, lack of interest in social peer monkeys lack of aggression and neglect of their infants (Bucher et al. 1970; Kling and Steklis 1976; Kling et al. 1993; Olson et al. 2007). In humans, socioemotional deﬁcits have been associated with temporal pole damage, such as the Kluver-Bucy syndrome that can result from herpes encephalitis affects the medial and anterior temporal lobes (Lilly et al. 1983). Similar symptoms can be observed following anterior temporal lobe lesions that result from epilepsy (Anson and Kuhlman 1993). In patients with temporal variant of frontal temporal dementia and only right temporal pole atrophy personality and social behavior deﬁcits are observed, such as loss of social dominance, lack of empathy and neurotic behavior(Mummery et al. 2000). Temporal pole damage can lead to unpredictable mood states, rapidly cycling bipolar disorder (Murai and Fujimoto 2003) and cycling levels of depression, anxiety and irritation (Glosser et al. 2000). Methods The temporal poles have also been implicated in face processing. Although the fusiform gyrus (Grill-Spector et al. 2004; Kanwisher et al. 1997, 1999; Kanwisher and Yovel 2006) and the superior temporal sulcus (Gross 1992; Perrett et al. 1985) are the structures mostly known for their involvement in the perception of face identity and expres- sion, there evidence suggesting that the temporal poles are also involved in face processing. Bilateral and even unilat- eral lesions of the temporal poles can produce a form of prosopagnosia that is not accompanied by generalized memory failure (Damasio et al. 1990). Cognitive deﬁcits of this ‘‘amnesic associative prosopagnosia’’ also include voice and gait mediated recognition (Olson et al. 2007). Left temporal pole lesions cause impairments in naming famous faces (Glosser et al. 2000), while atrophy of this region produces progressive prosopagnosia (Thompson et al. 2003). With electrodes implanted in the right ventral tem- poral pole ERP components with latencies around 350 ms were recorded (Allison et al. 1994, 1999; McCarthy et al. 1999; Puce et al. 1999). Finally, imaging studies have revealed that the temporal pole is involved in person (not face) recognition and the appraisal of facial emotional expressions (Damasio et al. 1990). D. Pantazis Classiﬁcation was by dividing data sets into a training set to locate centers of gravity of clusters and a test set to classify feature vectors by the Euclidean distance to the closest center of gravity. The placement of the human microgrid onto the inferior temporal gyrus, far from any sensory or motor area, was dictated entirely by neurosurgical considerations for the welfare of the patient, who gave informed consent for the recordings. This region has rich interconnections with both the amygdala and the orbital frontal cortex through the uncinate fascicule. It also sends projections to the hypo- thalamus and receives inputs from the three sensory sys- tems in the temporal lobes, the visual, the auditory and the olfactory systems as well as from the insula (gestation and awareness of internal physiological states. The temporal pole has been implicated in social behavior. Lesions of the temporal pole that spared the amygdala in female monkeys The goodness of classiﬁcation was reduced or abolished by randomization of the data: spatial scrambling of chan- nels, temporal scrambling of phase values, and most effec- tively by random shifts of ECoG segments in the 64 samples from each frame (Freeman 2005). The classiﬁcatory infor- mation was solely in the amplitudes of the signals, not in the 123 Cogn Neurodyn (2011) 5:55–66 57 Given the broad involvement of the temporal lobe in cognitive and social behaviors, and in consideration of the comfort and convenience of the patient, we did not proceed to a behavioral design with double blind controls. Rather, we sought the amplitude correlates of the ECoG in four behaviors that the patient engaged in repeatedly: (face-to- face interaction, looking at pictures, reading, speaking on the telephone). Pattern localization was based on local maxima in the spatial variance of the microgrid electrical activity; pattern classiﬁcation was based on the feature vectors at the maxima. Principal components analysis and a permutation test revealed signiﬁcant pattern differences between behavioral conditions. Our aim was to demon- strate the existence of classiﬁcatory information relating to cognition in the human ECoG at the high spatial frequen- cies that have heretofore been neglected. resulted in inappropriate vocal and facial social signals, inability to recognize social signals of other monkeys, lack of interest in social peer monkeys lack of aggression and neglect of their infants (Bucher et al. 1970; Kling and Steklis 1976; Kling et al. 1993; Olson et al. 2007). Electrode placement The temporal pole has also been implicated in tasks involving theory of mind (inferences about the intensions, desires and beliefs of others). Functional neuroimaging studies have revealed activation during theory of mind tasks such as thinking of other people’s emotions, making moral decisions or judging intentionality in viewing geo- metric shapes that move around. Evidence comes from neuroimaging studies and patients with temporal pole lesions who exhibit lower levels of empathy (Olson et al. 2007). Fifty-two intracranial subdural clinical electrodes were placed over lateral and basal temporal regions of both hemispheres through temporal burr holes. Additionally, a microgrid array of 8 9 8 0.5 mm stainless steel wires spaced 1.25 mm apart was placed over a 1 cm2 cortical patch on the right anterior inferior temporal gyrus (Fig. 1). The interelectrode dimensions of the microgrid were established on the basis of knowledge of the spatial 12 3 58 Cogn Neurodyn (2011) 5:55–66 Fig. 1 Left Location of microgrid in a radiograph. Right Drawing of microgrid and its electrode arrangement Fig. 1 Left Location of microgrid in a radiograph. Right Drawing of microgrid and its electrode arrangement was placed. As reviewed in the introduction, the temporal pole is implicated in both visual and auditory social behavior, and particularly with face processing. This region is believed to be active during ‘‘theory of mind’’ tasks. The inferior temporal lobe in associated with the integration of higher processing of complex stimuli. frequency spectrum of human EEG from the cortical sur- face (Freeman et al. 2000). More details of the electrode placement and microgrid design are available in Freeman et al. 2006a, which uses the same experimental data but a different set of behavioral moments, in order to investigate the power spectral density and phase of brain signals. Data collection Only the microgrid physiological record during the afore- mentioned behaviors was considered for analysis. The record during those time periods was seizure-free. Unlike conventional EEG recordings, corticographic recordings are less susceptible to muscle artifacts, such as eye blinks, body, head and tongue movement, than conventional EEG recordings. The main artifact that may be seen with intra- cranial EEG is pulse signals, if the electrode is near a vessel. Other artifacts include the electrode wire being grossly pulled from the outside, which is a rare occurrence because the subdural electrodes are securely anchored (Spencer et al. 2008). In this study, the physiological record for each condition was visually inspected and motion and other artifacts were identiﬁed and demarcated. Artifacts were characterized by their amplitude and overall mor- phology. The most common artifact was signals sharply deviating from the overall norm in one or more electrodes. Figure 2 shows examples of such artifact segments. Overall, the percentage of the data that were excluded from analysis was: 24.65% for the face condition, 24.8% for the pictures condition, 0.3% for the reading condition, and 14.7% for the phone condition. All artifact-free segments of each speciﬁc behavior were entered into subsequent analysis. The EEG was ampliﬁed with a Nicolet BMSI 5000 system having a ﬁxed gain of 1628 and analog ﬁlters set at 0.5 Hz high pass and 120 Hz low pass. The ADC gave 12 bits with the least signiﬁcant bit of 0.9 lV and maximal range of ±2048 9 0.9 lV. The reference and ground placements for all 64 electrodes were, respectively, the Cz and Pz scalp locations [midline frontal and parietal sites in the standard 10–20 clinical system (Reilly 1999). The Nicolet system digitized data at 420 Hz and down-sampled to 200 Hz. Off-line low pass ﬁltering at or below 55 Hz was per- formed to minimize 60 Hz artifact. Fig. 1 Left Location of microgrid in a radiograph. Right Drawing of microgrid and its electrode arrangement 1 The bar indicates averaging over the dotted subscript. V ¼ SCORE PCT ð3Þ where SCORE is a TM 9 N matrix, and PC is a N 9 N matrix with each column j populated by the prin- cipal component pcj. Principal ECoG components of conditions Selection of behavior moments The entire video record of the patient during the 8 days of hospitalization was reviewed. Sustained behaviors involv- ing visual and auditory social interactions and reading were sought and identiﬁed. Four behaviors were selected: (a) face to face interaction with another person (5 min), (b) looking at pictures (27 min), (c) reading a book or a magazine (10 min), and (d) speaking on the telephone (14 min). The selection of these behaviors was motivated by the function of the location over which the microgrid 123 123 59 Cogn Neurodyn (2011) 5:55–66 Spatial variance order to visualize the main components of the data For this Fig. 3 Left Selection of time points for data analysis. Spatial standard deviation r(t) is computed for every time point (t [ S) and a threshold of 35 mV is used to identify time regions of high spatial variance. Only time points corresponding to local maxima (t [ SM) within these regions are considered for further analysis (points marked with circles). Right Histograms of spatial standard deviation for each condition Fig. 2 Sample segments excluded from analysis due to artifacts Fig. 2 Sample segments excluded from analysis due to artifacts Fig. 2 Sample segments excluded from analysis due to artifacts Fig. 3 Left Selection of time points for data analysis. Spatial standard deviation r(t) is computed for every time point (t [ S) and a threshold of 35 mV is used to identify time regions of high spatial variance. Only time points corresponding to local maxima (t [ SM) within these regions are considered for further analysis (points marked with circles). Right Histograms of spatial standard deviation for each condition Spatial variance Spatial variance order to visualize the main components of the data. For this purpose, the 8 9 8 measurement matrices were reshaped into N 9 1 vectors v(t) and concatenated (after being transposed) into a voltage matrix V with dimension TM 9 N, where TM is the number of time points with maximum variance for each condition (cardinality of SM) and N = 64. We applied PCA on matrix V. PCA ﬁrst centers the data by making each variable (electrode in our case) zero mean. We denote the centered data as voðtÞ ¼ vðtÞ vð:Þ and Vo. PCA produced N principal component vectors pcj, j = 1…N, as well as the projections of the centered voltage vectors vo(t) on the principal components, called scores: The spatial variance r2(t) based on the voltage values at the microgrid electrodes was computed for each time point1: r2ðtÞ ¼ PN i¼1 ðviðtÞ v:ðtÞÞ2 N 1 ; t 2 S ð1Þ ð1Þ where vi(t) is the voltage at electrode i at time t, N = 64 is the total number of electrodes, and S is the artifact-free recording time period for each condition. Time segments with large spatial standard deviation (r(t) [ 35 mV, a threshold that allows enough trials per condition) were selected (Fig. 3). Among these time segments, only the time points with the local maximum variance were kept for subsequent analysis (t [ SM, where SM is the set of time points with local maximum variance), leading to a total of TM = 1138 time points for speaking on the phone, 1347 for reading, 199 for face-to-face interaction, and 254 for looking at pictures. Each one of these time points consisted of an 8 9 8 matrix of voltage values, therefore each con- dition yielded a set of 8 9 8 matrices. where vi(t) is the voltage at electrode i at time t, N = 64 is the total number of electrodes, and S is the artifact-free recording time period for each condition. Time segments with large spatial standard deviation (r(t) [ 35 mV, a threshold that allows enough trials per condition) were selected (Fig. 3). Fig. 3 Left Selection of time points for data analysis. Spatial standard deviation r(t) is computed for every time point (t [ S) and a threshold of 35 mV is used to identify time regions of high spatial variance. Only time points corresponding to local maxima (t [ SM) within these regions are considered for further analysis (points marked with circles). Right Histograms of spatial standard deviation for each condition Spatial variance Among these time segments, only the time points with the local maximum variance were kept for subsequent analysis (t [ SM, where SM is the set of time points with local maximum variance), leading to a total of TM = 1138 time points for speaking on the phone, 1347 for reading, 199 for face-to-face interaction, and 254 for looking at pictures. Each one of these time points consisted of an 8 9 8 matrix of voltage values, therefore each con- dition yielded a set of 8 9 8 matrices. voðtÞ ¼ X N j¼1 scoretjpcj; t 2 SM ð2Þ ð2Þ or in matrix form: or in matrix form: V ¼ SCORE PCT ð3Þ V ¼ SCORE PCT Fig. 2 Sample segments excluded from analysis due to artifacts Principal ECoG components of conditions Principal ECoG components of conditions Assume a total of TA voltage measurements vA i ðtÞ and TB voltage measurements vB i ðtÞ for two conditions A and B, respectively. We calculate a t-statistic map ti, i = 1…N, by taking the difference of the means across time between the two conditions and dividing by the standard deviation: Microgrid measurements with the highest spatial variance were grouped by condition and separately analyzed with principal component analysis as described in the ‘‘Method’’ section (Eq. 2). Figure 5 shows scatter plots of the scores for the ﬁrst two principal components (scoret,1 and scoret,2, t [ SM) for each condition. In other words, it displays the projections (inner product) of each microgrid measurement v(t) on the ﬁrst two principal components pc1 and pc2. ti ¼ vA i ð:Þ vB i ð:Þ rAB i ð5Þ rAB i ¼ ðTA 1ÞrA i þ ðTB 1ÞrB i TA TB þ 2 ð6Þ ð5Þ ð6Þ For all conditions, the scatter plots form two clusters approximately symmetric around the center (0,0). The bottom right clusters have 559 points for speaking on the phone, 597 for reading, 93 for face-to-face interaction, and 120 for looking at pictures. The top left clusters have 579 points for speaking on the phone, 750 for reading, 106 for face-to-face interaction, and 134 for looking at pictures. where rA i and rB i are the standard deviations of the voltage measurements at electrode i across time. Equation 6 allows for different number of samples and variances for each group. To detect electrodes that demonstrate statistical signiﬁcant difference across conditions, we need to threshold the statistical map ti, while controlling for mul- tiple comparisons. To focus on topographic patterns rather than voltage level differences, we normalized the voltage measurements into zero mean and unit variance across electrodes, as in Eq. 4. Then, we averaged the microgrid measurements corresponding to each cluster separately. Figure 6 shows the average topographic voltage maps for the bottom right clusters (ﬁrst row), top left clusters (second row), and their summation (third row). The average topographic voltage maps revealed that the two distributions are symmetrically We assume the voltage measurements were exchange- able, i.e. we can permute the matrices between the two conditions, under the null hypothesis of no condition effect. Multiple discriminant analysis Multiple discriminant analysis Each group of 8 9 8 matrices of voltage values was sep- arately subjected to principle component analysis (PCA), in To investigate whether the recorded signals on the micro- grid vary depending on the condition, we applied multiple 12 3 60 Cogn Neurodyn (2011) 5:55–66 that variables are produced by permutation. We then esti- mate the permutation statistic: discriminant analysis (MDA), which is a generalization of the Fisher Linear Discriminant analysis for multiple classes (Duda et al. 2000, p. 121). Our data comprises c = 4 classes, which correspond to 4 conditions: Phone, Reading, Face, Pictures. While PCA seeks directions that are efﬁ- cient for representation of the data, MDA seeks directions that are efﬁcient for discrimination. In particular, it ﬁnds weight vectors wk, k = 1…c - 1, which linearly transform the data into new measurements ykt ¼ wT k vðtÞ that are maximally separable among the classes. t i ¼ vA i ð:Þ vB i ð:Þ rAB i ð7Þ ð7Þ By repeating this process multiple times, we create M = 100000 permutation samples (in one of them we keep the original set), where each time we randomly permute the matrices between the two conditions. To control the family-wise error across all electrodes, we use the maximum statistic approach, which computes the maximum absolute value of the t-statistic for each permutation sample: t max ¼ max i ð t i Þ ð8Þ To explore whether differences across conditions are attributed to topographical changes in voltage rather than changes of the overall voltage level, we normalized the voltage measurements vi(t) into zero mean unit variance across electrodes before being subjected to MDA: ð8Þ We use the M permutation samples to build the empirical distribution of t max and then deﬁne a threshold that leaves a = 5% of the distribution on the right side. By applying this threshold to the t-statistic matrix of the original data, we identify the electrodes that have statistical signiﬁcant effects while controlling the family-wise error among all spatial locations. Figure 4 schematically outlines the methodology of this portion of the analysis. Multiple discriminant analysis vN i ðtÞ ¼ viðtÞ v:ðtÞ rðtÞ ð4Þ Permutation test vN i ðtÞ ¼ viðtÞ v:ðtÞ rðtÞ ð4Þ vN i ðtÞ ¼ viðtÞ v:ðtÞ rðtÞ ð4Þ Permutation test Permutation test To identify the spatial proﬁle of the electrodes that differ among conditions, we used a permutation test (Pantazis et al. 2005). The permutation method does not assume any parametric distribution (for example normally distributed samples), and automatically corrects for multiple compar- isons. We used a t-statistic contrast between conditions, which allows for different number of samples and different variance for each condition. 123 Fig. 5 Scatter plots of the scores (scoret,1 and scoret,2) of each voltage vector v(t) for the ﬁrst two principal components. Only time points of maximum local variance were analyzed (t [ SM), separately for each condition Principal ECoG components of conditions For every permutation sample, we exchange voltage mea- surements between the two conditions to create vA i ðtÞ and vB i ðtÞ, while keeping the number of observations for each group the same as the original data. The star (*) indicates 123 Cogn Neurodyn (2011) 5:55–66 61 Fig. 4 Illustration of the permutation method: the original voltage matrices produce many permutation samples by exchanging Condi- tion A and Condition B labels. The matrices are then averaged and normalized with the standard deviation to produce t-statistic matrices (maps t i from Eq. 7) separately for each permutation sample. We compute the maximum absolute t-statistic over all electrodes for each permutation sample (t max from Eq. 8) and use it to estimate the empirical distribution of the maximum statistic. The value that leaves a = 5% of the distribution on the right side is the threshold that is then applied to the t-statistic matrix of the original data (map ti from Eq. 5). Values above the threshold are signiﬁcantly different between the two conditions. The procedure guarantees that the family-wise error rate is exactly controlled at level a llustration of the permutation method: the original voltage empirical distribution of the maximum statistic The value that leaves Fig. 4 Illustration of the permutation method: the original voltage matrices produce many permutation samples by exchanging Condi- tion A and Condition B labels. The matrices are then averaged and normalized with the standard deviation to produce t-statistic matrices (maps t i from Eq. 7) separately for each permutation sample. We compute the maximum absolute t-statistic over all electrodes for each permutation sample (t max from Eq. 8) and use it to estimate the empirical distribution of the maximum statistic. The value that leaves a = 5% of the distribution on the right side is the threshold that is then applied to the t-statistic matrix of the original data (map ti from Eq. 5). Values above the threshold are signiﬁcantly different between the two conditions. The procedure guarantees that the family-wise error rate is exactly controlled at level a of the nts. m d 123 3 62 Cogn Neurodyn (2011) 5:55–66 sets of electrodes were signiﬁcant for each pair of condi- tions. The t-statistical maps ti of the original data, estimated using Eq. 5, are displayed at the bottom of Fig. 8. opposite to each other in each condition. Principal ECoG components of conditions Also, despite the great resemblance of topographies across conditions, dif- ferences are still visible. Our MDA and permutation test analysis explored whether these differences are signiﬁcant. Multiple discriminant analysis Discussion We have detected and analyzed local spatial ECoG patterns registered by a high-density array during naturalistic behaviors. These topographic patterns appear very similar across conditions, as shown in Fig. 6 (ﬁrst and second row). However, further investigation with multiple discriminant analysis demonstrated that the topographic patterns are different between conditions. A permutation analysis further localized these differences on various sets of electrodes across the microgrid. Depending on the conditions being contrasted, different sets of electrodes were signiﬁcantly different. Therefore, the local maxima of spatially variant activity revealed topographic patterns that signiﬁcantly differed between conditions of naturalistic behaviors. We performed MDA analysis only on topographies corre- sponding to the bottom right cluster of Fig. 5. Joint anal- ysis of both clusters would hinder statistical analysis, since clusters are far apart and located in similar areas for all conditions. Furthermore, voltages were normalized as in Eq. 4. Figure 7 shows scatter plots of MDA coefﬁcients ykt, k [ {1, 2, 3}, t [ SM, bottom right cluster. All four conditions are clearly separable, indicating that microgrid voltage measurements recording above 1 cm2 of cortex have dis- tinct topographies per condition. Inferential statistical comparisons 7 Scatter plots of the ﬁrst 3 linear discriminant coefﬁcients ykt, for voltage measurements corresponding to the bottom right clusters of Fig. 3 (t [ SM, bottom right cluster). All four conditions are clearly separable The usual artifacts apparent on scalp recordings, such as eye movements, body/head/tongue movements and EMG, are virtually absent with invasive recordings. The main artifacts that may be seen with intracranial EEG include pulse artifacts, if the electrode is on or near a vessel, or if the electrode wire is grossly pulled from the outside (a rare occurrence since the subdural electrodes are securely anchored). Other artifacts, such as 60 Hz, can rarely be seen if the electrode makes poor contact with the cortical surface (e.g. from underlying ﬂuid collection). As we stated earlier, during preprocessing ECoG recordings were visu- ally inspected for artifact identiﬁcation and rejection. Furthermore, other artifacts, such as systematic contami- nation from electromagnetic interference, would most probably cause smooth ﬁelds in the recordings, since the interfering source would be distant from the recordings. Therefore, such recordings would be excluded from our analysis because of our high spatial variance selection criterion. Mayhew 1989; Tramo et al. 1995). The remarkable ﬁnding of behavioral correlates in a sample size of 1 cm2 (1% of the temporal lobe, 0.05% of the cortex) lead us to predict that such information will be found in many, if not most, cortical locations, and could have an effect on applications, such as the design of brain-computer interfaces. Our pre- diction is supported by the global phase-locked pattern correlates of CS found in cats (Freeman and Burke 2003), and by the demonstration of classiﬁable spatial patterns of beta activity in the scalp EEG (Ruiz et al. 2009) recurring in phase-locked segments recurring at rates in the theta range (Freeman et al. 2003a, b; Ruiz et al. 2007; Pockett et al. 2009). We have observed two distinct voltage topographies, opposite to each other (Fig. 6). Such brain responses are often reported in literature. For example, in EEG/MEG brain responses are denoted as positive or negative, with the letters P and N, followed by the time delay in milliseconds, such as P20, N30 etc. An example is a study of median nerve stimulation, which yielded a number of these components with inverted topographies (Allison et al. 1991). Dipolar current source-sink pairs that inverted potentials from den- dritic current ﬂow can cause such responses. Inferential statistical comparisons 2009) recurring egments recurring at rates in the theta t al. 2003a, b; Ruiz et al. 2007; Pockett rved two distinct voltage topographies, other (Fig. 6). Such brain responses are terature. For example, in EEG/MEG brain noted as positive or negative, with the llowed by the time delay in milliseconds, tc. An example is a study of median nerve h yielded a number of these components ographies (Allison et al. 1991). Dipolar k pairs that inverted potentials from den- can cause such responses. We postulate nomenon occurs close to the microgrid. n the vicinity of the microgrid produce with alternating phases, therefore causing terns we observe in Fig. 6. The usual artifacts apparent on scalp recordings, su eye movements, body/head/tongue movements and E are virtually absent with invasive recordings. The artifacts that may be seen with intracranial EEG in pulse artifacts, if the electrode is on or near a vessel, the electrode wire is grossly pulled from the outside (a occurrence since the subdural electrodes are sec anchored). Other artifacts, such as 60 Hz, can rare seen if the electrode makes poor contact with the co surface (e.g. from underlying ﬂuid collection). As we s earlier, during preprocessing ECoG recordings were ally inspected for artifact identiﬁcation and rejec Furthermore, other artifacts, such as systematic con nation from electromagnetic interference, would probably cause smooth ﬁelds in the recordings, sinc interfering source would be distant from the record Therefore, such recordings would be excluded from analysis because of our high spatial variance sele criterion. The possibility still exists that some other reason related to artifacts, such as fatigue, alertness, etc. mig the main cause for the difference between conditions not the behavioral engagement we labeled. We under that some other internal reason might be the cause. indeed difﬁcult in natural settings where isolated beha are studied to separate all factors. We would like, how to point out that given the placement of the micro of the ﬁrst oltage ponding to ters of right cluster). re clearly 1) 5:55–66 Cogn Neurodyn (2011) 5:55–66 63 Mayhew 1989; Tramo et al. 1995). The remarkable ﬁnding of behavioral correlates in a sample size of 1 cm2 (1% of the temporal lobe, 0.05% of the cortex) lead us to predict that such information will be found in many, if not most, cortical locations, and could have an effect on applications, such as the design of brain-computer interfaces. Inferential statistical comparisons Our pre- diction is supported by the global phase-locked pattern correlates of CS found in cats (Freeman and Burke 2003), and by the demonstration of classiﬁable spatial patterns of beta activity in the scalp EEG (Ruiz et al. 2009) recurring in phase-locked segments recurring at rates in the theta range (Freeman et al. 2003a, b; Ruiz et al. 2007; Pockett et al. 2009). We have observed two distinct voltage topographies, opposite to each other (Fig. 6). Such brain responses are often reported in literature. For example, in EEG/MEG brain responses are denoted as positive or negative, with the letters P and N, followed by the time delay in milliseconds, such as P20, N30 etc. An example is a study of median nerve stimulation, which yielded a number of these components with inverted topographies (Allison et al. 1991). Dipolar current source-sink pairs that inverted potentials from den- d iti t ﬂ h W t l t The usual artifacts apparent on scalp recordings, such as eye movements, body/head/tongue movements and EMG, are virtually absent with invasive recordings. The main artifacts that may be seen with intracranial EEG include pulse artifacts, if the electrode is on or near a vessel, or if the electrode wire is grossly pulled from the outside (a rare occurrence since the subdural electrodes are securely anchored). Other artifacts, such as 60 Hz, can rarely be seen if the electrode makes poor contact with the cortical surface (e.g. from underlying ﬂuid collection). As we stated earlier, during preprocessing ECoG recordings were visu- ally inspected for artifact identiﬁcation and rejection. Furthermore, other artifacts, such as systematic contami- nation from electromagnetic interference, would most probably cause smooth ﬁelds in the recordings, since the interfering source would be distant from the recordings. Therefore, such recordings would be excluded from our analysis because of our high spatial variance selection criterion. The possibility still exists that some other reason not related to artifacts, such as fatigue, alertness, etc. might be the main cause for the difference between conditions, and t th b h i l t l b l d W d t d Fig. 7 Scatter plots of the ﬁrst 3 linear discriminant coefﬁcients ykt, for voltage measurements corresponding to the bottom right clusters of Fig. 3 (t [ SM, bottom right cluster). All four conditions are clearly separable Cogn Neurodyn (2011) 5:55 66 63 Fig. Inferential statistical comparisons The present experiment was designed to test the pre- diction, which was based on animal studies of ECoG, that textured high spatial frequency information would be found in human ECoG that was correlated with cognitive behaviors. The location of the microgrid was based on proximity to an expected pathological focus, but it was random with respect to an area engaged in a certain cog- nitive function. Estimates of the surface area of human cerebral cortex range from 1640 to 2500 cm2, with 22% in the temporal lobe (Nieuwenhuys et al. 1998; Henery and Between-condition analysis of the matrices at the local maxima of the spatial variance was done using a permu- tation test, as described in the ‘‘Method’’ section. Similarly to MDA analysis, only the bottom right clusters of Fig. 5 were analyzed, and voltage measurements were normalized into zero mean unit variance as in Eq. 4. The results for each condition comparison are shown in Fig. 8. Signiﬁcant electrodes (corrected for multiple com- parisons) are displayed on top with white color. Different Fig. 6 Average topographic distributions of normalized voltages vN i ðtÞ across the microgrid. First row average of topographies for time points corresponding to the bottom right clusters of Fig. 5, separately for each condition. Second row average of top left clusters of Fig. 5. Third row summation of above. Topographies on ﬁrst and second rows are nearly opposite. Also, there is great similarity of topographies between conditions, but also visible differences 123 Fig. 6 Average topographic distributions of normalized voltages vN i ðtÞ across the microgrid. First row average of topographies for time points corresponding to the bottom right clusters of Fig. 5, separately for each condition. Second row average of top left clusters of Fig. 5. Third row summation of above. Topographies on ﬁrst and second rows are nearly opposite. Also, there is great similarity of topographies between conditions, but also visible differences 123 123 amo et al. 1995). The remarkable ﬁnding elates in a sample size of 1 cm2 (1% of 0.05% of the cortex) lead us to predict tion will be found in many, if not most, and could have an effect on applications, n of brain-computer interfaces. Our pre- ted by the global phase-locked pattern ound in cats (Freeman and Burke 2003), stration of classiﬁable spatial patterns of e scalp EEG (Ruiz et al. Inferential statistical comparisons We postulate that a similar phenomenon occurs close to the microgrid. Current dipoles in the vicinity of the microgrid produce electric currents with alternating phases, therefore causing the two spatial patterns we observe in Fig. 6. The possibility still exists that some other reason not related to artifacts, such as fatigue, alertness, etc. might be the main cause for the difference between conditions, and not the behavioral engagement we labeled. We understand that some other internal reason might be the cause. It is indeed difﬁcult in natural settings where isolated behaviors are studied to separate all factors. We would like, however, to point out that given the placement of the microgrid, 123 3 123 Cogn Neurodyn (2011) 5:55–66 64 namely the right tip of the temporal lobes, such a separa- tion by behaviors as the ones we selected would not be impossible. Furthermore, our goal is to demonstrate the existence of spatially variable signals of brain origin, even if they are not introduced by strictly deﬁned conditions. Our method was designed to extract high spatial fre- quency information from recordings of cortical activity channels, there must always be a trade-off between sam- pling high spatial frequencies versus measuring signals from extended cortical areas. Microgrids sample high spatial frequencies but only in a small cortical area, whereas con- ventional ECoGs sample extended cortical areas but at the expense of limited spatial resolution and therefore aliasing. Evidence was required that spatial textures could be Fig. 8 Permutation test statistical analysis. Top Signiﬁcant voxels when comparing different pairs of conditions, corrected for multiple comparisons at a 0.05 family-wise error level. Bottom Statistical maps of original data, estimated using Eq. 5 64 Cogn Neurodyn (2011) 5:55–66 channels, there must always be a trade-off between sam- pling high spatial frequencies versus measuring signals from extended cortical areas. Microgrids sample high spatial frequencies but only in a small cortical area, whereas con- ventional ECoGs sample extended cortical areas but at the expense of limited spatial resolution and therefore aliasing. namely the right tip of the temporal lobes, such a separa- tion by behaviors as the ones we selected would not be impossible. Furthermore, our goal is to demonstrate the existence of spatially variable signals of brain origin, even if they are not introduced by strictly deﬁned conditions. Fig. 8 Permutation test statistical analysis. Top Signiﬁcant voxels when comparing different pairs of conditions, corrected for multiple comparisons at a 0.05 family-wise error level. Bottom Statistical maps of original data, estimated using Eq. 5 Inferential statistical comparisons Our method was designed to extract high spatial fre- quency information from recordings of cortical activity, particularly when the microgrid was positioned on cortical areas for which behavioral correlates had already been established by conventional ECoG, MEG and fMRI meth- ods, which by wide spacing gave low spatial frequency information. Given a limited number of electrodes and Evidence was required that spatial textures could be detected with such close spacing, despite the smoothing of ECoG signals by volume conduction on referential record- ing. The evidence in animal studies came from classiﬁcation of spatial patterns of amplitude modulation in the beta and gamma ranges with respect to conditioned stimuli (Freeman 123 123 Cogn Neurodyn (2011) 5:55–66 65 2005, 2006). This method demonstrated that, despite the shared frequencies of the 64 waveforms of the ECoG, the local differences in amplitude gave signiﬁcant classiﬁcatory information. Even more signiﬁcant were the spatial patterns of phase modulation, which were determined by the relative amplitudes of the Fourier components of sine and cosine. The phase gradients provided unequivocal proof of infor- mation at high spatial frequencies in the referentially recorded ECoG from animals (Freeman 2004a, b; Freeman and Barrie (2000). Very similar phase gradients from the microgrid array provided comparable proof for the preser- vation of amplitude information in ECoG signals in human referential recording (Freeman 2005). Allison T, McCarthy G, Nobre A, Puce A, Belger A (1994) Human extrastriate visual cortex and the perception of faces, words, numbers, and colors. Cereb Cortex 4:544–554 Allison T, Puce A, Spencer DD, McCarthy G (1999) Electrophys- iological studies of human face perception. I: potentials gener- ated in occipitotemporal cortex by face and non-face stimuli. Cereb Cortex 9:415–430 Anson JA, Kuhlman DT (1993) Post-ictal Kluver-Bucy syndrome after temporal lobectomy. J Neurol Neurosurg Psychiatr 56: 311–313 Barrie JM, Freeman WJ, Lenhart M (1996) Modulation by discrim- inative training of spatial patterns of gamma EEG amplitude and phase in neocortex of rabbits. J Neurophysiol 76:520–539 Bucher K, Myers RE, Southwick C (1970) Anterior temporal cortex and maternal behavior in monkey. Neurology 20:415 Damasio AR, Tranel D, Damasio H (1990) Face agnosia and the neural substrates of memory. Annu Rev Neurosci 13:89–109 The microgrid recordings utilized in this study, which provided access to ‘‘ﬁne-grained’’ spatial information on human cortical EEG, resulted in ﬁndings that may also be consistent with one concept of the cortical dynamics of cognition. Allison T, McCarthy G, Wood CC, Jones SJ (1991) Potentials evoked in human and monkey cerebral cortex by stimulation of the median nerve. Brain 114:2465–2503 Inferential statistical comparisons This theory proposes that neural activity related to perception is modulated by ‘‘cinematographic’’ spatio- temporal EEG patterns (Henery and Mayhew 1989). Shared cortical oscillations occurring at the same instan- taneous frequency across the cortical surface may serve as a ‘‘carrier wave’’ for perception by means of spatial pat- terns (of ‘‘wave packets’’) of amplitude modulation (AM). The information content is broadly distributed over the cortex, not localized. Endogenous state transitions, which occur several times per second, and result in the formation a wave packet with speciﬁc AM patterns may be the essence of act of perception (Henery and Mayhew 1989; Tramo et al. 1995). We speculate that the distinct spatial patterns that we ﬁnd are associated with speciﬁc behaviors in this study may in fact correspond to AM patterns of perception and intention. Future research that employs microgrid technology in other subjects will be required to better understand the dynamics of human cognition. We predict that the ECoG correlates at high spatial frequencies will be found in most if not all areas of cortex studied in subjects engaged in normal behaviors. Duda RO, Hart PE, Stork DG (2000) Pattern classiﬁcation, 2nd edn. Wiley-Interscience, New York Freeman WJ (1975) Mass action in the nervous system. Academic Press, New York Freeman WJ (2004a) Origin, structure, and role of background EEG Freeman WJ (2004a) Origin, structure, and role of background EEG activity. Part 1. Analytic amplitude. Clin Neurophysiol 115: 2077–2088. http://repositories.cdlib.org/postprints/1006 activity. Part 1. Analytic amplitude. Clin Neurophysiol 115: 2077–2088. http://repositories.cdlib.org/postprints/1006 Freeman WJ (2004b) Origin, structure, and role of background EEG activity. Part 2. Analytic phase. Clin Neurophysiol 115: 2089–2107. http://repositories.cdlib.org/postprints/1486 Freeman WJ (2005) Origin, structure, and role of background EEG activity. Part 3. Neural frame classiﬁcation. Clin Neurophysiol 116 (5): 1118–1129. http://authors.elsevier.com/sd/article/S1388 245705000064; http://repositories.cdlib.org/postprints/2134/ Freeman WJ (2006) Origin, structure, and role of background EEG activity. Part 4. Neural frame simulation. Clin Neurophysiol 117/ 3: 572–589. http://repositories.cdlib.org/postprints/1480/; http:// dx.doi.org/10.1016/j.clinph.2005.10.025 Freeman WJ, Barrie JM (2000) Analysis of spatial patterns of phase in neocortical gamma EEGs in rabbit. J Neurophysiol 84: 1266–1278 Freeman WJ, Burke BC (2003) A neurobiological theory of meaning in perception. Part 4. Multicortical patterns of amplitude modulation in gamma EEG. Int J Bifurc Chaos 13: 2857–2866. http://repositories.cdlib.org/postprints/3345 Freeman WJ, Grajski KA (1987) Relation of olfactory EEG to behavior: factor analysis. Inferential statistical comparisons Behav Neurosci 101:766–777 Freeman WJ, Rogers LJ (2002) Fine temporal resolution of analytic phase reveals episodic synchronization by state transitions in gamma EEG. J Neurophysiol 87:937–945 Freeman WJ, Viana Di Prisco G (1986) Relation of olfactory EEG to behavior: time series analysis. Behav Neurosci 100:753–763 Acknowledgments The Authors wish to acknowledge Mr. Nickolas Noah for his assistance with editorial corrections. A preliminary report on this work has been presented (Panagiotides et al. 2008). Freeman WJ, Rogers LJ, Holmes MD, Silbergeld DL (2000) Spatial spectral analysis of human electrocorticograms including the alpha and gamma bands. J Neurosci Methods 95:111–121 Freeman WJ, Burke BC, Holmes MD, Vanhatalo S (2003a) Spatial spectra of scalp EEG and EMG from awake humans. Clin Neurophysiol 114:1053–1068 Open Access This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which per- mits any noncommercial use, distribution, and reproduction in any medium, provided the original author(s) and source are credited. Freeman WJ, Burke BC, Holmes MD (2003b) Aperiodic phase re-setting in scalp EEG of beta-gamma oscillations by state transitions at alpha-theta rates. Hum Brain Mapp 19:248–272. http://repositories.cdlib.org/postprints/3347 Freeman WJ, Gaa´l G, Jornten R (2003c) A neurobiological theory of meaning in perception. Part 3. Multiple cortical areas synchro- nize without loss of local autonomy. Int J Bifurc Chaos 13: 2845–2856. http://repositories.cdlib.org/postprints/3344 References Allison T, McCarthy G, Wood CC, Jones SJ (1991) Potentials evoked in human and monkey cerebral cortex by stimulation of the median nerve. Brain 114:2465–2503 Freeman WJ, Holmes MD, West GA, Vanhatalo S (2006a) Fine spatiotemporal structure of phase in human intracranial EEG. Clin Neurophysiol 117:1228–1243 12 3 3 66 Cogn Neurodyn (2011) 5:55–66 Olson IR, Plotzker A, Ezzyat Y (2007) The Enigmatic temporal pole: a review of ﬁndings on social and emotional processing. Brain 130:8–31 Freeman WJ, Holmes MD, West GA, Vanhatalo S (2006b) Dynamics of human neocortex that optimizes its stability and ﬂexibility. Intern J Intell Syst 21: 1–21. http://repositories.cdlib.org/ postprints/2385 Olson IR, Plotzker A, Ezzyat Y (2007) The Enigmatic temporal pole: a review of ﬁndings on social and emotional processing. Brain 130:8–31 Panagiotides H, Freeman WJ, Holmes MD, Pantazis D (2008) Behavioral states exhibit distinct spatial EEG patterns. Abstr #1.051, 62nd Ann Mtg Amer Epilepsy Soc, Seattle WA Ghovanloo MW, Wise KD, NajaﬁK (2003) Towards a button-sized 1024-site wireless cortical microstimulating array. Neural engi- neering conference proceedings, ﬁrst international IEEE EMBS conference on neural engineering, pp 138–141 Pang CC et al (2005) A new multi-site probe array with monolith- ically integrated parylene ﬂexible cable for neural prostheses. The 27th annual international conference of the IEEE engineer- ing in medicine and biology society (EMBS), Shanghai, China Glosser G, Zwil AS, Glosser DS, O’Connor MJ, Sperling MR (2000) Psychiatric aspects of temporal lobe epilepsy before and after anterior temporal lobectomy. J Neurol Neurosurg Psychiatr 68:53–58 Pantazis D, Nichols TE, Baillet S, Leahy RM (2005) A comparison of random ﬁeld theory and permutation methods for the statistical analysis of MEG data. Neuroimage 25:383–394 Grill-Spector K, Knouf N, Kanwisher N (2004) The fusiform face area subserves face perception, not generic within-category identiﬁcation. Nat Neurosci 7:555–562 Perrett DI, Smith PA, Potter DD, Mistlin AJ, Head AS, Milner AD et al (1985) Visual cells in the temporal cortex sensitive to face view and gaze direction. Proc R Soc B Biol Sci 223:293–317 Gross CG (1992) Representation of visual stimuli in inferior temporal cortex. Philos Trans R Soc Lond B 335:3–10 cortex. Philos Trans R Soc Lond B 335:3–10 Henery CC, Mayhew TM (1989) The cerebrum and cerebellum of the ﬁxed human brain: efﬁcient and unbiased estimates of volumes and cortical surface areas. References J Anat 167:167–180 Pockett S, Bold GEJ, Freeman WJ (2009) EEG synchrony during a perceptual-cognitive task: widespread phase synchrony at all frequencies. Clin Neurophysiol 120:695–708. doi:10.1016/ j.clinph.2008.12.044 Kanwisher N, Yovel G (2006) The fusiform face area: a cortical region specialized for the perception of faces. Philos Trans R Soc Lond B Biol Sci 361:2109–2128 Puce A, Allison T, McCarthy G (1999) Electrophysiological studies of human face perception. III: effects of top-down processing on face-speciﬁc potentials. Cereb Cortex 9:445–458 Kanwisher N, McDermott J, Chun MM (1997) The fusiform face area: a module in human extrastriate cortex specialized for face perception. J Neurosci 17:4302–4311 Reilly EL (1999) EEG recording and operation of the apparatus. In: Niedermeyer E, Lopes da Silva F (eds) Electroencephalography: basic principles, clinical applications, and related ﬁelds. Williams & Wilkins, Baltimore, pp 122–142 Kanwisher N, Stanley D, Harris A (1999) The fusiform face area is selective for faces not animals. Neuroreport 10:183–187 Kling A, Steklis HD (1976) A neural substrate for afﬁliative behavior in nonhuman primates. Brain Behav Evol 13:216–238 Ruiz YG, Li G, Freeman WJ, Moreira EG (2007) A new approach to detect stable phase structure in high density EEG signals. 1st international conference on Cognitive Neurodynamics 2007. Shanghai, P.R. China: Springer, Chap. 128, pp 741–745 Kling AS, Tachiki K, Lloyd R (1993) Neurochemical correlates of the Kluver-Bucy syndrome by in vivo microdialysis in monkey. Behav Brain Res 56:161–170 Ruiz Y, Li G, Freeman WF, Gonzalez E (2009) Detecting stable phase structures on EEG signals to classify brain activity amplitude patterns. J Zhejiang Univ 10:1483–1491 Lilly R, Cummings JL, Benson F, Frankel M (1983) The human Kluver-Bucy syndrome. Neurology 33:1141–1145 McCarthy G, Puce A, Belger A, Allison T (1999) Electrophysiolog- ical studies of human face perception. II: response properties of face-speciﬁc potentials generated in occipitotemporal cortex. Cereb Cortex 9:431–444 Spencer S, Sperling M, Shewmon A, Kahane P (2008) Intracranial electrodes. In: Engel J Jr, Pedley T (eds) Epilepsy: a compre- hensive textbook, vol. 171. Wolters Kluwer/Lippincott, Phila- delphia, pp 1791–1815 Mummery CJ, Patterson K, Prices CJ, Ashburner J, Frackowiak RSJ, Hodges JR (2000) A voxel-based morphometry study of semantic dementia: relationship between temporal lobe atrophy and semantic memory. Ann Neurol 47:36–45 Thompson SA, Patterson K, Hodges JR (2003) Left/right asymmetry of atrophy in semantic dementia: behavioral-cognitive implica- tions. References Neurology 61:1196–1203 Tramo MJ, Loftus WC, Thomas CE, Green RL, Mott LA, Gazzaniga MS (1995) Surface area of human cerebral cortex and its gross morphological subdivisions: in vivo measurements in monozy- gotic twins suggest differential hemisphere effects of genetic factors. J Cogn Neurosci 7:292–302 Murai T, Fujimoto S (2003) Rapid cycling bipolar disorder after left temporal polar damage. Brain Inj 17:355–358 Nieuwenhuys R, Ten Donkelaar HJ, Nicholson C (1998) The central nervous system of vertebrates, vol. 3. Springer, Berlin Ohl FW, Scheich H, Freeman WJ (2001) Change in pattern of ongoing cortical activity with auditory category learning. Nature 412:733–736 123 123
https://openalex.org/W2909198863	https://link.springer.com/content/pdf/10.1007/s13659-019-0197-y.pdf	English	null	Anti-oral Microbial Flavanes from Broussonetia papyrifera Under the Guidance of Bioassay	Natural products and bioprospecting	2,019	cc-by	4,153	Natural Products and Bioprospecting (2019) 9:139–144 https://doi.org/10.1007/s13659-019-0197-y Natural Products and Bioprospecting (2019) 9:139–144 https://doi.org/10.1007/s13659-019-0197-y ORIGINAL ARTICLE ORIGINAL ARTICLE Abstractl A new flavane, bropapyriferol (1), and eleven known ones were isolated from the EtOAc part of Broussonetia papyrifera under the guidance of bioassay. The structure of compound 1 was determined by extensive 1D and 2D NMR, [α]D spec- troscopic data and quantum computation. Daphnegiravan F (2) and 5,7,3′,4′-tetrahydroxy-3-methoxy-8,5′-diprenylflavone (3) showed significantly anti-oral microbial activity against five Gram-positive strains and three Gram-negative strains in vitro. Especially, compound 3 was more potent in suppressing Actinomyces naeslundii and Porphyromonas gingivalis (MIC = 1.95 ppm) than the positive control, triclosan. Graphical Abstract y = 1.0278x - 6.8615 R² = 0.9997 0 20 40 60 80 100 120 140 160 0.0 20.0 40.0 60.0 80.0 100.0 120.0 140.0 160.0 Exp-1b bropapyriferol (1) MIC = 1.95 ppm A. naeslundii P. gingivalis Compound 3 y = 1.0278x - 6.8615 R² = 0.9997 0 20 40 60 80 100 120 140 160 0.0 20.0 40.0 60.0 80.0 100.0 120.0 140.0 160.0 Exp-1b bropapyriferol (1) MIC = 1.95 ppm A. naeslundii P. gingivalis Compound 3 Keywords Bropapyriferol · Broussonetia papyrifera · Anti-oral microbial activity Keywords Bropapyriferol · Broussonetia papyrifera · Anti-oral microbial activity Anti‑oral Microbial Flavanes from Broussonetia papyrifera Under the Guidance of Bioassay Chang‑An Geng1,2 · Meng‑Hong Yan1,2 · Xue‑Mei Zhang1,2 · Ji‑Jun Chen1,2,3 Chang‑An Geng1,2 · Meng‑Hong Yan1,2 · Xue‑Mei Zhang1,2 · Ji‑Jun Chen1,2,3 Received: 26 November 2018 / Accepted: 6 January 2019 / Published online: 16 January 2019 © The Author(s) 2019 * Ji‑Jun Chen chenjj@mail.kib.ac.cn 1 Introduction Electronic supplementary material The online version of this article (doi:https://doi.org/10.1007/s13659-019-0197-y) contains supplementary material, which is available to authorized users. * Ji‑Jun Chen chenjj@mail.kib.ac.cn 1 State Key Laboratory of Phytochemistry and Plant Resources in West China, Kunming Institute of Botany, Chinese Academy of Sciences, 132# Lanhei Road, Kunming 650201, Yunnan, People’s Republic of China 2 Yunnan Key Laboratory of Natural Medicinal Chemistry, Kunming 650201, People’s Republic of China 3 University of Chinese Academy of Sciences, Beijing 100049, People’s Republic of China Electronic supplementary material The online version of this article (doi:https://doi.org/10.1007/s13659-019-0197-y) contains supplementary material, which is available to authorized users. Broussonetia papyrifera, a deciduous broadleaf plant in the family Moraceae, is native to Asia, and mainly distributed in China, Japan, Vietnam and India. B. papyrifera also known as paper mulberry is a highly valued plant in China, whose leaves, roots, barks and fruits are used in traditional Chinese medicines (TCMs) for diuretic, anti-rheumatic, anti-bacte- rial, anti-inflammatory purposes. Its bark is a good source for papermaking and leaves are high quality feed. Now, this plant widely grows in Europe, the United States and Africa as introduced a species [1, 2]. The main constituents of B. papyrifera are flavones, lignans, diterpenes and triterpenes * Ji‑Jun Chen chenjj@mail.kib.ac.cn 1 State Key Laboratory of Phytochemistry and Plant Resources in West China, Kunming Institute of Botany, Chinese Academy of Sciences, 132# Lanhei Road, Kunming 650201, Yunnan, People’s Republic of China 2 Yunnan Key Laboratory of Natural Medicinal Chemistry, Kunming 650201, People’s Republic of China 3 University of Chinese Academy of Sciences, Beijing 100049, People’s Republic of China (0121 140 C.-A. Geng et al. Table 1 1H and 13C NMR data of compound 1 in pyridine-d5 (δ in ppm, J in Hz) No. 2.1 Structure Elucidation Compound 1 was assigned the molecular formula of C25H30O6 from the positive HRESIMS at m/z 449.1954 ([M + Na]+, calcd. for 449.1940), with 11 degrees of unsatu- ration. The IR absorptions at 3419, 1623, 1600, 1509 and 1463 cm−1 were indicative for the presence of hydroxy, double bond and phenyl groups. In the 1H NMR spectrum, three ABX-coupling protons at δH 7.10 (d, J = 8.2 Hz, H-5), 6.87 (dd, J = 8.2, 2.4 Hz, H-6) and 6.93 (d, J = 2.4 Hz, H-8) suggested a 1,2,4-trisubstituted phenyl ring. In addition, one aromatic proton at δH 7.20 and four methyls at δH 1.75, 1.63, 1.49 and 1.41, all in singlet, were obviously recog- nized. In the 13C NMR spectrum, one methine (δC 75.7) and two methylenes (δC 30.7 and 25.8) were characteristic for a flavane skeleton when taking 12 aromatic carbons into con- sideration. One prenyl group was deduced from the methyl- ene (δH 3.83, d, J = 6.6 Hz; δC 25.4), tri-substituted double bond (δH 5.49, t, J = 6.6 Hz; δC 124.8 and 130.5) and two tertiary methyls (δH 1.75 and 1.63, both s; δC 18.0 and 25.8). From the above analyses, this compound was proposed to be a prenylated flavane [6]. The stereochemistry at C-12′ and C-13′ was deduced as trans-form by the coupling constant (JH-12′,H-13′ = 4.4 Hz) and comparing with that of 16-hydroxycudratrixanthone Q [11]. In order to determine the stereochemistry of 1, the two possible configurations, 2S,12′S,13′S-1 (1a) and 2S,12′R,13′R-1 (1b), were applied to quantum compu- tation at the B3LYP/6-311 + g(d,2p) level for the 13C NMR data. As shown in Fig. 2, the calculated NMR data for 1b matched well with the experimented data, with the correla- tion coefficient (R2) of 0.9979, when comparing with that of the calculated 1a. In addition, the correlation coefficient (R2) was up to 0.9997 when only the carbons (C-4′, 5′, 6′, 12′, 13′ and 14′) near C-12′ and C-13′ takes into consid- eration. The absolute stereochemistry was determined to be 2R, 12′S, 13′S by the positive [α]D value which is consist- ent with the previous report and quantum calculation [12]. Thus, the structure of 1 was determined and named to be bropapyriferol. 1 Introduction 1H NMR (400 MHz) 13C NMR (100 MHz) 2 5.42, brd, 9.2 75.7, d 3 2.13, m 2.02, m 30.7, t 4 2.94, m 2.71, dd, 15.7, 2.9 25.8, t 4a ‒ 113.1, s 5 7.10, d, 8.2 130.8, d 6 6.87, dd, 8.2, 2.4 109.3, d 7 ‒ 158.4, s 8 6.93, d, 2.4 104.3, d 8a ‒ 157.4, s 1′ ‒ 133.8, s 2′ ‒ 128.1, s 3′ ‒ 140.1, s 4′ ‒ 148.8, s 5′ ‒ 129.8, s 6′ 7.20, overlapped 114.2, d 7′ 3.83, d, 6.6 25.4, t 8′ 5.49, t, 6.6 124.8, d 9′ ‒ 130.5, s 10′ 1.75, s 18.0, q 11′ 1.63, s 25.8, q 12′ 6.07, d, 4.4 74.3, d 13′ 4.94, d, 4.4 99.1, d 14′ ‒ 70.8, s 15′ 1.49, s 26.4, q 16′ 1.41, s 25.8, q Table 1 1H and 13C NMR data of compound 1 in pyridine-d5 (δ in ppm, J in Hz) [3–8]. In the previous investigation, three flavanes, kazinol B, 7,4′-dihydroxy-3′-prenylflavan and 3′-(3-methylbut-2- enyl)-3′,4′,7-trihydroxyflavane, and one triterpene, oleanolic acid, from the cortex were revealed with anti-oral microbial activity [9]. In order to clarify the anti-oral microbial effects of B. papyrifera and the active constituents, the preliminary bioassay revealed that the EtOAc part showed increased activity comparing to the total extraction. Thus, detailed investigation was performed on the active EtOAc part and yielded one new flavane, bropapyriferol, and eleven known ones. Herein, we report their isolation, structural elucidation and anti-oral microbial activities. 2.1 Structure Elucidation l The NMR data of 1 showed similarity with those of daph- negiravan F [10], and the main difference was that com- pound 1 had an addition 2-hydroxypropan-2-yl group (two methyls: δH 1.41 and 1.49, δC 26.4 and 25.8; one oxygenated quaternary carbon: δC 70.8) and two methines (δH 6.07 and 4.94, δC 74.3 and 99.1) instead of a double bond (Table 1). Detailed analysis of the 2D NMR spectra well established its planar structure: the 1H-1H COSY correlation of H-5 with H-6, and HMBC correlations from H-6 and H-8 to C-4a, and from H-5 to C-7 verified a hydroxyl group at C-7; the HMBC correlations from H-8′ to C-2′, and from H-7′ to C-1′ and C-3′ suggested the prenyl group at C-2′; the 1H-1H COSY correlation of H-12′ with H-13′, and HMBC correlations from H-15′ and H-16′ to C-13′ indicated a hydroxyl group at C-12′ and 2-hydroxypropan-2-yl group at C-13′ (Fig. 1). T h e k n o w n c o m p o u n d s w e r e d e t e r- m i n e d a s d a p h n e g i r av a n F ( 2 ) [ 1 0 ] , 1 3 3 Anti‑oral Microbial Flavanes from Broussonetia papyrifera Under the Guidance of Bioassay 141 141 Anti oral Microbial Flavanes from Broussonetia papyrifera Under the Guidance of Bioassay Fig. 1 Structure of 1 and the key 2D NMR correlations (→ HMBC, ▬ 1H-1H COSY) y = 0.9657x - 0.2619 R² = 0.9977 0 20 40 60 80 100 120 140 160 180 0.0 20.0 40.0 60.0 80.0 100.0 120.0 140.0 160.0 180.0 Exp-1a y = 0.9656x - 0.3006 R² = 0.9979 0 20 40 60 80 100 120 140 160 180 0.0 20.0 40.0 60.0 80.0 100.0 120.0 140.0 160.0 180.0 Exp-1b y = 1.0283x - 6.8081 R² = 0.9994 0 20 40 60 80 100 120 140 160 0.0 20.0 40.0 60.0 80.0 100.0 120.0 140.0 160.0 Exp-1a y = 1.0278x - 6.8615 R² = 0.9997 0 20 40 60 80 100 120 140 160 0.0 20.0 40.0 60.0 80.0 100.0 120.0 140.0 160.0 Exp-1b B A D C Fig. 2 The correlations between calculated and experimental 13C NMR data for two possible configurations of 1 (1a and 1b). 2.1 Structure Elucidation gingivalis Total extraction 39 39 39 19.5 PE part > 1250 > 1250 > 1250 39 EtOAc part 19.5 19.5 19.5 9.8 n-Butanol part 78.1 39 78.1 39 Water part > 1250 > 1250 > 1250 > 1250 Triclosanc 3.9 3.9 7.8 7.8 Table 2 Anti-oral microbial activity of fractions from B. papyrifera expressed as MIC (ppm) 2.1 Structure Elucidation All carbons (a and b) and only C-4′, 5′, 6′, 12′, 13′, 14′ (c and d) were considered Fig. 1 Structure of 1 and the key 2D NMR correlations (→ HMBC, ▬ 1H-1H COSY) Fig. 1 Structure of 1 and the key 2D NMR correlations (→ HMBC, ▬ 1H-1H COSY) Fig. 1 Structure of 1 and the key 2D NMR correlations (→ HMBC, ▬ 1H-1H COSY) y = 0.9657x - 0.2619 R² = 0.9977 0 20 40 60 80 100 120 140 160 180 0.0 20.0 40.0 60.0 80.0 100.0 120.0 140.0 160.0 180.0 Exp-1a Exp-1a A C y = 0.9656x - 0.3006 R² = 0.9979 0 20 40 60 80 100 120 140 160 180 0.0 20.0 40.0 60.0 80.0 100.0 120.0 140.0 160.0 180.0 Exp-1b Exp-1b B D B A Exp-1b C D y = 1.0278x - 6.8615 R² = 0.9997 0 20 40 60 80 100 120 140 160 0.0 20.0 40.0 60.0 80.0 100.0 120.0 140.0 160.0 Exp-1b D y = 1.0283x - 6.8081 R² = 0.9994 0 20 40 60 80 100 120 140 160 0.0 20.0 40.0 60.0 80.0 100.0 120.0 140.0 160.0 Fig. 2 The correlations between calculated and experimental 13C NMR data for two possible configurations of 1 (1a and 1b). All carbons (a and b) and only C-4′, 5′, 6′, 12′, 13′, 14′ (c and d) were considered Fig. 2 The correlations between calculated and experimental 13C NMR data for two possible configurations of 1 b) and only C-4′, 5′, 6′, 12′, 13′, 14′ (c and d) were considered n calculated and experimental 13C NMR data for two possible configurations of 1 (1a and 1b). All carbons (a and 14′ (c and d) were considered 5,7,3′,4′-tetrahydroxy-3-methoxy-8,5′-diprenylflavone (3) [13], broussoflavonol B (4) [14], broussonin A (5) [15], isoliquiritigenin (6) [16], uralenol (7) [17], broussoflavan A (8) [18], uralene (9) [19], uralenol-3-methylether (10) [19], abyssinone VII (11) [20] and 3,5,7,3′,5′-pentahydroxyfla- vanone (12) [21] by comparing with the previous reports. 1 3 Table 2 Anti-oral microbial activity of fractions from B. papyrifera expressed as MIC (ppm) Fractions A. viscosus S. mutans F. nucleatum P. 3.2 Plant Materials The plant materials of Broussonetia papyrifera (Linn.) L’Hér. ex Vent. were collected from Chongqing, China, in August 2009, and authenticated by Dr. Li-Gong Lei (Kun- ming Institute of Botany, CAS). A voucher specimen (No. 200908001) was deposited in the Laboratory of Antivirus and Natural Medicinal Chemistry, Kunming Institute of Botany, CAS. 3.1 General Procedures Optical rotations were collected on a Jasco model 1020 pola- rimeter (Horiba, Tokyo, Japan). UV spectra were determined on a Shimadzu UV-2401A spectrophotometer (Shimadzu, Kyoto, Japan). IR (KBr) spectra were recorded on a Bio- Rad FTS-135 spectrometer (Bio-Rad, Hercules, California, USA). 1D and 2D NMR spectra were recorded on a Bruker AM-400 NMR or DRX-500 spectrometer with TMS as the internal standard (Bruker, Bremerhaven, Germany). MS data were collected on a VG Auto Spec-3000 spectrometer (VG, Manchester, UK) and an API Qstar Pulsar hybrid Q-TOF mass spectrometer (AB-Sciex, Framingham, MA, USA). Compounds 1–12 were assayed for their anti-oral microbial activity against five Gram-positive strains (A. naeslundii, A. viscosus, S. mutans, S. sanguinis and S. sorbrinus) and three Gram-negative strains (A. actinomy- cetemcomitans, F. nucleatum, P. gingivalis). As shown in Table 3, compound 2 exhibited significant activity against all the strains except for S. sorbrinus, with MIC values of 7.8, 7.8, 3.9, 15.6, 62.5, 3.9, 3.9 and 7.8 ppm, which is comparable to the positive control, triclosan. Compound 3 could obviously inhibit five Gram-positive strains and one Gram-negative strain (P. gingivalis), with MIC values between 1.95 and 15.6 ppm. For A. naeslundii and P. gin- givalis, compound 3 showed an MIC value of 1.95 ppm, even higher than triclosan (MIC = 3.9 and 7.8 ppm, rep- ectively). In addition, compound 10 also showed com- parable inhibition with triclosan on two Gram-negative strains, F. nucleatum (MIC = 7.8 ppm) and P. gingivalis (MIC = 7.8 ppm). 3 Experimental strains, A. viscosus, S. mutans, F. nucleatum and P. gin- givalis with MIC values of 39, 39, 39 and 19.5 ppm, respectively. After extraction by different solvents, the EtOAc part showed the highest activity with MIC values of 19.5, 19.5, 19.5 and 9.8 ppm; the n-butanol part exhib- ited slightly decreased activity with MIC values of 78.1, 39, 78.1 and 39 ppm; and the petroleum ether (PE) and water parts are inactive except for the PE part showing inhibition on P. gingivalis (MIC = 39 ppm). 2.2 Anti‑oral Microbial Activity The total extraction and each fraction were assayed for the anti-oral microbial activity against two Gram-positive and two Gram-positive strains. As shown in Table 2, the total extraction showed activity against the four assayed 142 C.-A. Geng et al. Table 3 Anti-oral microbial activity of compounds 1–12 expressed as MIC (ppm) A. na, Actinomyces naeslundii ATCC 12104; A. vi, Actinomyces viscosus ATCC 27044; S. mu, Strepto- coccus mutans ATCC 25175; S. sa, Streptococcus sanguinis ATCC 10556; S. so, Streptococcus sorbrinus ATCC6715; A. ac, Aggregatibacter actinomycetemcomitans ATCC 43717; F. nu, Fusobacterium nuclea- tum ATCC 10953; P. gi: Porphyromonas gingivalis ATCC 33277 3.3 Extraction and Isolation D-3-2 was purified by repeated silica gel and Sephadex LH-20 CC to give compounds 5 (8 mg), 6 (11 mg) and 11 (18 mg). Fr. D-3-3 was purified by HPLC to yield compounds 7 (5 mg), 9 (7 mg), 10 (3 mg) and 12 (4 mg). Fr. D-4 was loaded on Si CC and eluted with CHCl3–MeOH–H2O system (85:15:1.5) to provide three fractions, Frs. D-4-1 to D-4-3. Compounds 2 (5 mg), 4 (3 mg) and 8 (9 mg) was obtained from Fr. D-4-1 after HPLC purification. Fr. D-4-2 was separated by HPLC to generated compounds 1 (5 mg) and 3 (6 mg). for two times. The combined extract was solved in water and partitioned with petroleum ether (PE), EtOAc and n-butanol, successively. Based on the preliminary bioassay, the EtOAc part showed the highest anti-oral microbial activity, and was thus applied for the following investigation. The EtOAc part was separated by silica gel CC using CHCl3–MeOH gradi- ent (from 9:1 to 5:5) to yield six fractions, Frs. A-F. Fr. D was fractionated by MPLC system on a MCI gel CHP20P column with MeOH-H2O system (5:95, 10:90, 30:70, 50:50 and 0:100) as the mobile phase to provide five fractions, Frs. D-1 to D-5. Fr. D-3 was subjected to Si CC and eluted with CHCl3–MeOH–H2O system (9:1:0.1) to generate three frac- tions, Fr. D-3-1 to D-3-3. Fr. D-3-2 was purified by repeated silica gel and Sephadex LH-20 CC to give compounds 5 (8 mg), 6 (11 mg) and 11 (18 mg). Fr. D-3-3 was purified by HPLC to yield compounds 7 (5 mg), 9 (7 mg), 10 (3 mg) and 12 (4 mg). Fr. D-4 was loaded on Si CC and eluted with CHCl3–MeOH–H2O system (85:15:1.5) to provide three fractions, Frs. D-4-1 to D-4-3. Compounds 2 (5 mg), 4 (3 mg) and 8 (9 mg) was obtained from Fr. D-4-1 after HPLC purification. Fr. D-4-2 was separated by HPLC to generated compounds 1 (5 mg) and 3 (6 mg). aerobically or anaerobically at 37 °C for about 20 h, the plate was determined by a BioTek EL × 808 Microplate Reader. The bacteria growth was indicated by the turbidity deter- mined at OD630, which was performed in duplicate. The minimum inhibitory concentration (MIC) is defined as the lowest concentration of the sample at which the microorgan- ism tested does not demonstrate visible growth. 3.3 Extraction and Isolation Triclosan (Purity > 99%, Ciba Specialty Chemicals) was used as the positive control. 3.5 Quantum Computation The 13C NMR and [α]D calculations for compound 1 were carried out using Gaussian 09 program. Conformational search was achieved by Spartan ‘14 in MMFF94 s force field, and the lowest conformer was further optimized with the hf/3-21 g and DFT b3lyp/6-311 + g(d,p) methods in Gaussian 09 program package. 13C NMR shielding con- stants were calculated with the GIAO method at b3lyp/6- 311 + g(d,2p) level in pyridine with PCM, which were converted into chemical shifts by referencing to TMS. [α]D values were calculated at b3lyp/6-31 g(d,p) level based on the above DFT optimized geometries. 3.3.1 Spectroscopic Data of Bropapyriferol (1) Acknowledgements This work was financially supported by the Pro- gram of Yunling Scholarship, the CAS “Light of West China” Program (Western Youth Scholars “A”), the Youth Innovation Promotion Asso- ciation CAS (2013252), and the Applied Basic Research Programs of Yunnan Province (2017FB137). Colorless gum, C25H30O6, [α]22D +27.6º (c 0.16, CHCl3–MeOH, 1:1); UV (CHCl3–MeOH) λmax (log ε) 236 (4.01) and 284 (3.69) nm; IR (KBr) νmax 3419, 2928, 1623, 1600, 1509, 1463, 1377, 1307, 1226, 1155, 1114, 1055, 999 cm−1; 1H and 13C NMR data see Table 1; EIMS m/z 426 [M]+ (1), 390 (6), 350 (20), 294 (9), 227 (14), 214 (14), 185 (29), 171 (11), 123 (100); (+) HRESIMS m/z 449.1954 (calcd. for C25H30O6Na, 449.1940). Compliance with Ethical Standards Conflict of interest These authors have no conflict of interest to de- clare. 3.4 Anti‑oral Microbial Test Open Access This article is distributed under the terms of the Crea- tive Commons Attribution 4.0 International License (http://creativeco mmons.org/licenses/by/4.0/), which permits unrestricted use, distribu- tion, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. Microorganisms including Gram-positive [Actinomyces naeslundii (ATCC 12104), Actinomyces viscosus (ATCC 27044), Streptococcus mutans (ATCC 25175), Streptococ- cus sanguinis (ATCC 10556) and Streptococcus sorbrinus (ATCC6715)] and Gram-negative [Aggregatibacter actino- mycetemcomitans (ATCC 43717), Fusobacterium nuclea- tum (ATCC 10953) and Porphyromonas gingivalis (ATCC 33277)] bacteria were used for anti-oral microbial test. All bacterial strains were obtained from the American Type Culture Collection (ATCC). A broth microdilution method was used to assess the anti-bacterial activity of the frac- tions and isolates by determining the minimum inhibitory concentration (MIC) [22]. In brief, a serial doubling dilu- tion of each sample in the range 0.12–250 ppm was pre- pared in a 96-well plate. 100 μL of bacterial suspension was added into the 96-well plate, and the final concentration in each well was adjusted to 5 × 107 CFU/mL. After incubated 3.3 Extraction and Isolation The air-dried aerial part (5 kg) of B. papyrifera was pow- dered and extracted with EtOH (90%) at room temperature A. na, Actinomyces naeslundii ATCC 12104; A. vi, Actinomyces viscosus ATCC 27044; S. mu, Strepto- coccus mutans ATCC 25175; S. sa, Streptococcus sanguinis ATCC 10556; S. so, Streptococcus sorbrinus ATCC6715; A. ac, Aggregatibacter actinomycetemcomitans ATCC 43717; F. nu, Fusobacterium nuclea- tum ATCC 10953; P. gi: Porphyromonas gingivalis ATCC 33277 Compd. Gram-positive Gram-negative A. na A. vi S. mu S. sa S. so A. ac F. nu P. gi 1 250 125 125 125 250 62.5 31.25 125 2 7.8 7.8 3.9 15.6 62.5 3.9 3.9 7.8 3 1.95 3.9 15.6 7.8 7.8 62.5 > 250 1.95 4 31.25 7.8 62.5 31.25 62.5 15.6 7.8 3.9 5 62.5 62.5 62.5 62.5 125 62.5 31.25 31.25 6 125 125 125 125 125 31.25 15.6 62.5 7 250 250 250 250 250 15.6 31.25 31.25 8 62.5 31.25 250 62.5 125 250 31.25 31.25 9 > 250 250 > 250 62.5 125 31.25 15.6 7.8 10 125 31.25 62.5 62.5 125 15.6 7.8 7.8 11 62.5 31.25 62.5 62.5 62.5 62.5 15.6 15.6 12 125 125 62.5 62.5 125 > 250 250 250 Triclosan 3.9 3.9 3.9 7.8 3.9 1.95 7.8 7.8 A. na, Actinomyces naeslundii ATCC 12104; A. vi, Actinomyces viscosus ATCC 27044; S. mu, Strepto- coccus mutans ATCC 25175; S. sa, Streptococcus sanguinis ATCC 10556; S. so, Streptococcus sorbrinus ATCC6715; A. ac, Aggregatibacter actinomycetemcomitans ATCC 43717; F. nu, Fusobacterium nuclea- tum ATCC 10953; P. gi: Porphyromonas gingivalis ATCC 33277 1 3 3 Anti‑oral Microbial Flavanes from Broussonetia papyrifera Under the Guidance of Bioassay 143 for two times. The combined extract was solved in water and partitioned with petroleum ether (PE), EtOAc and n-butanol, successively. Based on the preliminary bioassay, the EtOAc part showed the highest anti-oral microbial activity, and was thus applied for the following investigation. The EtOAc part was separated by silica gel CC using CHCl3–MeOH gradi- ent (from 9:1 to 5:5) to yield six fractions, Frs. A-F. Fr. D was fractionated by MPLC system on a MCI gel CHP20P column with MeOH-H2O system (5:95, 10:90, 30:70, 50:50 and 0:100) as the mobile phase to provide five fractions, Frs. D-1 to D-5. Fr. D-3 was subjected to Si CC and eluted with CHCl3–MeOH–H2O system (9:1:0.1) to generate three frac- tions, Fr. D-3-1 to D-3-3. Fr. References 1. J. González-Lorca, A. Rivera-Hutinel, X. Moncada, S. Lobos, D. Seelenfreund, A. Seelenfreund, N. Z. J. Bot. 53, 75–89 (2015) 2. R.H. Ji, X.F. Lin, W.B. Zhang, S.L. Bai, China For. Prod. Ind. 44, 3–6 (2017) 3. Y. Ding, J. Qiu, X.Y. Liu, Nat. Prod. Res. Dev. 26(1327–1331), 1337 (2014) 4. H.W. Ryu, B.W. Lee, M.J. Curtis-Long, S. Jung, Y.B. Ryu, W.S. Lee, K.H. Park, J. Agric. Food Chem. 58, 202–208 (2010) 5. H.H. Ko, W.L. Chang, T.M. Lu, J. Nat. Prod. 71, 1930–1933 (2008) 1 3 144 C.-A. Geng et al. 15. T. Mitsuo, A. Masaki, M. Tadashi, S. Akira, T. Kokichi, Chem. Lett. 3, 339–340 (1980) 16. Z.Q. Yin, J.F. Chao, L.H. Zhang, W.C. Ye, W.B. Shen, S.X. Zhao, Nat. Prod. Res. Dev. 18(420–422), 425 (2006) 7. D. Lee, K.P.L. Bhat, H.H.S. Fong, N.R. Farnsworth, J.M. Pezzuto, A.D. Kinghorn, J. Nat. Prod. 64, 1286–1293 (2001) 7. D. Lee, K.P.L. Bhat, H.H.S. Fong, N.R. Farnsworth, J.M. Pezzuto, A.D. Kinghorn, J. Nat. Prod. 64, 1286–1293 (2001) g 8. H.W. Ryu, M.J. Curtis-Long, S. Jung, Y. Jeong, D.S. Kim, K.Y. Kang, K.H. Park, Food Chem. 132, 1244–1250 (2012) 17. S.S. Jia, C.M. Ma, J.M. Wang, Acta Pharm. Sin. 25, 758–762 (1990) g Z.M. Li, J.J. Chen, M.H. Yan, Med. Plant 1, 33–35 (2010 18. S.C. Fang, B.J. Shieh, C.N. Lin, Phytochemistry 37, 851–853 (1994) 10. Q. Sun, F.F. Li, D. Wang, J. Wu, G.D. Yao, X. Li, L.Z. Li, Q.B. Liu, X.X. Huang, S.J. Song, RSC Adv. 6, 55919–55929 (2016) 19. S.S. Jia, D. Liu, X.P. Zheng, Y. Zhang, Y.K. Li, Acta Pharm. Sin. 28, 28–31 (1993) g g 11. J. Kwon, N.T. Hiep, D.W. Kim, S. Hong, Y. Guo, B.Y. Hwang, H.J. Lee, W. Mar, D. Lee, J. Nat. Prod. 79, 1938–1951 (2016) 20. F. Machumi, G. Bojase-Moleta, R. Mapitse, I. Masesane, R.R.T. Majinda, Nat. Prod. Comm. 1, 287–292 (2006) 12. L. Cui, H.S. Lee, D.T. Ndinteh, J.T. Mbafor, Y.H. Kim, T.V.T. Le, P.H. Nguyen, W.K. Oh, Planta Med. 76, 713–718 (2010) j 21. L.S. Ding, Q.L. Liang, Y.F. Teng, Acta Pharm. Sin. 32, 600–602 (1997) 13. F.J. Guo, L. Feng, C. Huang, H.X. Ding, X.T. Zhang, Z.Y. Wang, Y.M. Li, Phytochemistry Lett. 6, 331–336 (2013) 22. J.F. Liu, Z.Y. Jiang, C.A. Geng, X.B. Zou, Y. Shi, Y.B. Ma, X.M. Zhang, J.J. Chen, Planta Med. 76, 1464–1467 (2010) y y 14. J. Matsumoto, T. Fujimoto, C. Takino, M. Saitoh, Y. 6. S.C. Fang, B.J. Shieh, R.R. Wu, C.N. Lin, Phytochemistry 38, 535–537 (1995) 22. J.F. Liu, Z.Y. Jiang, C.A. Geng, X.B. Zou, Y. Shi, Y.B. Ma, X.M. Zhang, J.J. Chen, Planta Med. 76, 1464–1467 (2010) References Hano, T. Fukai, T. Nomura, Chem. Pharm. Bull. 33, 3250–3256 (2008) 1 3 3
https://openalex.org/W4324149189	https://bmccardiovascdisord.biomedcentral.com/counter/pdf/10.1186/s12872-023-03164-4	English	null	LncRNA HIF1A-AS2: a potential biomarker for early diagnosis of acute myocardial infarction and predictor of left ventricular dysfunction	BMC cardiovascular disorders	2,023	cc-by	6,584	Abstract Background Rapid diagnosis of acute myocardial infarction (AMI) is the subject of many clinical studies as it ena- bles an effective therapy, preventing adverse progression of AMI and increasing survival rates. Recent studies have revealed that specific blood-based long non-coding RNAs (lncRNAs) are deregulated in patients with AMI and serve as promising diagnostic and prognostic tools. The current study aimed to determine the potential role of a hypoxia- responsive lncRNA, hypoxia-inducible factor 1A antisense RNA 2 (HIF1A-AS2), as a biomarker for early diagnosis and predictor of left ventricular dysfunction (LVD). Methods This study was carried out on 48 patients with AMI and 50 age-and sex-matched controls. The relative quantification of HIF1A-AS2 expression was done using reverse transcription real‐time polymerase chain reaction. Results Compared to the control group, HIF1A-AS2 were significantly higher in MI patients (P < 0.001). Interestingly, patients presenting within 3 h of chest pain onset had elevated levels of HIF1A‐AS2 as compared to patients with late presentation. The ROC curve was constructed to assess HIF1A-AS2 as an early marker. It demonstrated higher sensi- tivity (94%) and specificity (86%). Moreover, the multivariate regression analysis revealed that HIF1A‐AS2 was signifi- cantly associated with LVD in the patient group after 6 months follow up (p = 0.018). Conclusion Our study suggests that HIF1A‐AS2 may be a potential early diagnostic biomarker of AMI with high sen- sitivity. In addition, it might have a promising role as a predictor of left ventricular dysfunction. Keywords Long non-coding RNA, Hypoxia-inducible factor 1A, Myocardial infarction, Antisense RNA Keywords Long non-coding RNA, Hypoxia-inducible factor 1A, Myocardial infarction, Antisense RNA, Early biomarker © The Author(s) 2023. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativeco mmons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. LncRNA HIF1A‑AS2: a potential biomarker for early diagnosis of acute myocardial infarction and predictor of left ventricular dysfunction Eman Tayae1* , Eman Amr1, Amr Zaki2 and Dalal Elkaffash1 Eman Tayae1* , Eman Amr1, Amr Zaki2 and Dalal Elkaffash1 Tayae et al. BMC Cardiovascular Disorders (2023) 23:135 https://doi.org/10.1186/s12872-023-03164-4 Tayae et al. BMC Cardiovascular Disorders (2023) 23:135 https://doi.org/10.1186/s12872-023-03164-4 BMC Cardiovascular Disorders Open Access Introduction According to the latest WHO Global Health Estimates, cardiovascular diseases, specifically ischemic heart dis- ease (IHD), is still the leading cause of mortality world- wide over the past two decades. However, it currently results in higher increased mortality rates than ever before [1]. Myocardial infarction (MI) is a severe mani- festation of IHD. It results from rupture or erosion of a vulnerable atherosclerotic coronary plaque where a totally occluding thrombus leads to Acute ST-segment– elevation MI (STEMI) or from reduced coronary artery Correspondence: Eman Tayae Eman.Tayae@alexmed.edu.eg 1 Clinical Pathology Department, Faculty of Medicine, Alexandria University, Alexandria, Egypt 2 Cardiology Department, Faculty of Medicine, Alexandria University, Alexandria, Egypt © The Author(s) 2023. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativeco mmons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Subjects and methods Subjects After receiving the approval of the ethics commit- tee, the current study was carried out on 98 subjects. Forty-eight patients were admitted to the Cardiology Department at Alexandria main University Hospital, suffering from AMI diagnosed based on WHO criteria, excluding patients with a history of old infarction, pre- vious myocardial intervention, and malignant tumor. Fifty healthy volunteers of matching ages and sex served as a control group. All participants were sub- jected to full history taking, physical examination, and laboratory investigations (hs Troponin I, CBC, hs CRP, Lipid profile). Six months follow up was done to assass the prognostic value of HIF1A-AS2. Relative quantifi- cation of HIF1A-AS2 expression was done using real‐ time reverse transcription–polymerase chain reaction. More than 80% of ncRNAs represent a heterogeneous group of long non-coding RNA (lncRNAs) (200 nt–1 kb in length). LncRNAs are significantly more tissue-specific than protein-coding genes and are predominantly local- ized in the nucleus, whereas mRNA is abundant in the cytoplasm [8]. LncRNAs exert their effects through a variety of mechanisms that include chromatin and chro- mosome condensation through histone modifications, transcriptional regulation, posttranscriptional regulation, and regulating miRNA expression [9]. Relative quantification of HIF 1A ANTISENSE LONG NON‑CODING RNA 2 (HIF1A‐AS2) expression RNA extraction In the past few years, many studies have demonstrated that a group of lncRNA expressions is modulated by hypoxia [10]. As MI results from acute and prolonged deficits in oxygen supply, which trigger a series of severe biochemical and metabolic disturbances in the cardiomy- ocytes [11], we attempted to determine the potential role of one of hypoxia-induced lncRNA, hypoxia-inducible factor 1A antisense RNA 2 (HIF1A-AS2) as a potential biomarker for AMI. Genomic RNA was extracted from peripheral blood cells using RNeasy mini kit according to manufacturer instructions (QIAGEN, Germany). The concentration and purity of RNA were measured at 260& 280 &230 nm using NanoDrop 2000c Spectrophotometer (Thermo Sci- entific, USA). Total RNA was kept at -80 °C until further analysis. Open Access T Page 2 of 9 Tayae et al. BMC Cardiovascular Disorders (2023) 23:135 Tayae et al. BMC Cardiovascular Disorders (2023) 23:135 blood flow without complete coronary occlusion, coro- nary artery spasm, coronary embolism results in Non- STEMI (NSTEMI) or unstable angina [2, 3]. Restricted blood supply would lead to irreversible damage of func- tional cardiomyocytes, ultimately resulting in ventricular failure, with significant adverse effects on life quality and increased mortality [4]. both physiologically and when the tissues had become cancerous [13]. These observations prompted further investigation into HIF1A-AS2. These studies revealed that HIF1A-AS2 was upregulated in various tumors. Therefore, HIF1A-AS2 was identified as a potential bio- marker in oncology [14]. HIF1A-AS2 was also found to be hypoxia-inducible and negatively regulates HIF-1a mRNA expression [15]. HIF-1a is a crucial transcription factor that controls oxy- gen delivery and utilization as a key modulator of oxygen homeostasis. It controls the expression of genes involved in vascular remodeling and angiogenesis, an essential component of the heart’s response to ischemia. Conse- quently, it was suggested that HIF-1a play a critical role in the pathophysiology of ischemic heart disease and heart failure [16, 17]. In addition to clinical symptoms and electrocardio- graphic (ECG) findings, biomarkers levels are the most crucial factor in diagnosing AMI. Cardiac troponin is the diagnostic gold standard for AMI, and its use is rec- ommended by current guidelines [5]. However, explor- ing new diagnostic and prognostic biomarkers with high sensitivity and specificity, especially for early AMI, could pave the way for the continuous discovery of new thera- peutic targets.i Epigenetic processes, defined as heritable changes in gene expression that occur without changes to the DNA sequence, have emerged as a promising area in cardiovascular disease research. Epigenetic biomarkers include DNA methylation, histone modifications, and non-coding RNA mechanisms [6]. Over the past decade, whole transcriptome screening has revealed that nearly the entire human genome is transcribed. However, only a minor fraction (2%) of these transcripts are translated into proteins, while the remaining plethora of transcripts represent ncRNAs [7]. Complementary DNA (cDNA) synthesis HIF1A-AS2 is the antisense transcript of hypoxia- inducible factor 1α (HIF1α). It is localized at chromo- some 14q23.2 and was first discovered in 1990 to be abnormally expressed in clear cell renal carcinoma. It binds in a complementary manner to the 3’ untranslated region (3’UTR) of HIF1α mRNA [12]. In 2002, HIF1A- AS2 was found to be expressed in several human tissues, First, cDNA was synthesized from purified RNA using High-Capacity cDNA Reverse Transcription Kits (Thermo Scientific, USA) according to the manufacturer’s protocol. In a volume of 20 µL, 1 µg of RNA was added to the reverse transcription (RT) reaction mix. The prepared reaction mix Tayae et al. BMC Cardiovascular Disorders (2023) 23:135 Page 3 of 9 was incubated in Arktik thermocycler at 25 °C for 10 min, 37 °C for 120 min, and then 85 °C for 5 min. was incubated in Arktik thermocycler at 25 °C for 10 min, 37 °C for 120 min, and then 85 °C for 5 min. diabetics, and 27 patients had hypertension. According to ECG changes, 27 patients were diagnosed with Acute ST- segment–elevation MI (STEMI) and 21 with Non-STEMI (NSTEMI). After 6 months of follow up, echocardiogra- phy revealed that ejection fraction (EF) in 18 patients was (EF) ≤ 40%, while EF was > 40% in 30 patients. Age and gen- der were not significantly different between AMI patients and controls (P = 0.07 and P = 0.14, respectively). Demo- graphic, laboratory, and clinical data for the study groups are illustrated in Table 2. Statistical analysis Assessment of HIF1A‑AS2 value as an early AMI biomarker Interestingly, patients presenting within three hours showed significantly higher HIF1A‐AS2 expression levels than patients presenting later (P = 0.002) (Table 4). Data were analyzed using Statistical Package of Social Science (SPSS/version 23) software. Qualitative data were described using numbers and percentages and were compared using the Chi-square test. The Kolmog- orov–Smirnov test was used to verify the normality of the distribution of variables. The student’s t-test was used to compare normally distributed quantitative data, expressed in mean ± standard deviation (SD). Mann– Whitney test was used to compare abnormally distrib- uted quantitative variables expressed in the median, minimum, and maximum. Correlations between two quantitative variables were assessed using the Pearson coefficient. The HIF1A-AS2 Receiver operating charac- teristic curve (ROC) was used to detect diagnostic per- formance. The significance of the obtained results was determined at the 5% level. In order to determine the role of HIF1A-AS2 as an early marker for AMI, another ROC curve for HIF1A- AS2 expression was constructed to detect AMI in the patient group with pain onset within three hours vs. the control group. We found that HIF1A-AS2 can detect early MI with 94% sensitivity and 86% specificity at a cut- off > 1.8 (Fig. 2). Expression of HIF1A‑AS2 in acute myocardial infarction Expression of HIF1A‑AS2 in acute myocardial infarction The HIF1A‐AS2 expression level was significantly higher in AMI patients than in controls (P ≤ 0.001) (Table 3). A ROC curve analysis was done to assess the diagnostic value of HIF1A-AS2 expression in AMI. It was found that HIF1A-AS2 could detect AMI in the patient group with 85.4% sensitivity and 86% specificity at cut-off 1.8. The corresponding area under the curve (AUC) was found to be 0.931 (Fig. 1). Associations between HIF1A‐AS2 expression and clinical and laboratory parameters in acute myocardial infarction patients (Table 5) There was no correlation between the expression of the HIF1A-AS2 gene and age (P = 0.639). In addition, no significant association was found between HIF1A-AS2 and gender or any AMI risk factors such as smoking, hypertension, and diabetes mellitus. In addition, no significant difference in HIF1A-AS2 levels between STEMI and NSTEMI patients was detected. Quantitative real‑time (qPCR)i Q (q ) Relative quantification of HIF1A-AS2 was performed using Maxima SYBR Green/ROX qPCR Master Mix (Thermo Scientific, USA), custom made primers (10 pmol/µL), and cDNA in a total volume of 25 µL according to the manu- facturer’s instructions. Thermal cycling was done using Rotor gene Q real time PCR system (Qiagen, Germany) and included initial denaturation at 95 °C for 10 min fol- lowed by 40 cycles of denaturation at 95 °C for 15 s and final annealing and extension at 56 °C for 1 min. Acquisi- tion of SYBR Green I dye was done during the annealing- extension step. A negative control (No template control) was included in each run. A normalizer target, SF3a1, was used as a housekeeping gene, and relative quantifica- tion of HIF1A-AS2 was measured using the Comparative CT method (2−ΔΔCt). Primers’ sequences are illustrated in Table 1 Characteristics of the study participantsh The current case–control study was conducted on 48 patients diagnosed with AMI according to WHO criteria, with ages ranging from 30.0—75.0 years. Patients included 32 males (66.7%) and 16 females (33.3%). Concerning risk factors, 23 patients were smokers, 16 patients were HIF1A-AS2 demonstrated positive correlation with high sensitivity cTnI level (r = 0.437, p = 0.002). With Table 1 PCR primers NCBI Accession Number Forward Primer Reverse Primer HIF1A-AS2 NR_045406 GGTCTGCCATCTATTACTT TCTCAGCATTATAGTCACAA SF3a1 NM_005877 GATTGGCCCCAGCAAGCC TGCGGAGACAACTGTAGTACG NCBI Accession Number Tayae et al. BMC Cardiovascular Disorders (2023) 23:135 Page 4 of 9 Table 2 Demographic, clinical and laboratory data for the studied groups SD Standard deviation, STEMI ST-segment–elevation MI, NSTEMI Non-STEMI, χ2 Chi square test, t Student t-test Acute myocardial infarction Cases (n = 48) Controls (n = 50) Test of Sig P Age (year) Mean ± SD 56.23 ± 10.34 52.70 ± 8.74 t = 1.828 0.071 Min. – Max 30.0—75.0 30.0—69.0 Sex Male 32 (66.7%) 26 (52.0%) χ2 = 2.181 0.140 Female 16 (33.3%) 24 (48.0%) Type of AMI NSTEMI 21 (43.8%) - - - STEMI 27 (56.3%) - Smoking Nonsmoker 22 (45.8%) 30 (60.0%) χ2 = 3.351 MCp = 0.187 Smoker 23 (47.9%) 15 (30.0%) Ex-smoker 3 (6.3%) 5 (10.0%) Diabetes Mellitus Yes 16 (33.3%) - NO 32 (66.7%) Hypertension Yes 27 (56.3%) - No 21 (43.7%) hs- cTnI (ng/ml) Mean ± SD 64.43 ± 109.63 - - - Median (Min. – Max.) 30.04 (1.03–450.0) WBCs (× 103/ul) Mean ± SD 12.45 ± 5.62 - - - Median (Min. – Max.) 11.18 (3.65–27.80) CRP (mg/dl) Mean ± SD 90.38 ± 84.99 - - - Median (Min. – Max.) 63.0 (1.36—325.0) Ejection fraction ≤ 40 18(37.5%) > 40 30(62.5%) Table 2 Demographic, clinical and laboratory data for the studied groups respect to markers of inflammation, positive cor- relation was found between HIF1A-AS2 and CRP level (r = 0.42, P = 0.002) and WBC count (r = 318, P = 0.028). HIF1A‐AS2 expression as a predictor of left ventricular dysfunction In the patient group, a significant association was found between HIF1A-AS2 and EF% after 6 months of follow up Table 3 HIF1A-AS2 expression in acute myocardial infarction U Mann Whitney test Statistically significant at p ≤ 0.05 HIF1A‐AS2 expression Acute myocardial infarction Cases (n = 48) Controls (n = 50) Test of Sig P Median (Min. Characteristics of the study participantsh – Max.) 5.75 (0.99—310.0) 1.07 (0.50—2.60) U = 164.50* < 0.001* Table 3 HIF1A-AS2 expression in acute myocardial infarction U Mann Whitney test * Statistically significant at p ≤ 0.05 HIF1A‐AS2 expression Acute myocardial infarction Cases (n = 48) Controls (n = 50) Test of Sig P Median (Min. – Max.) 5.75 (0.99—310.0) 1.07 (0.50—2.60) U = 164.50* < 0.001* Table 3 HIF1A-AS2 expression in acute myocardial infarction HIF1A‐AS2 expression as a predictor of left ventricular dysfunction In the patient group, a significant association was found between HIF1A-AS2 and EF% after 6 months of follow up HIF1A‐AS2 expression as a predictor of left ventricular dysfunctioni In the patient group, a significant association was found between HIF1A-AS2 and EF% after 6 months of follow up Tayae et al. BMC Cardiovascular Disorders (2023) 23:135 Page 5 of 9 Fig. 1 ROC analysis to assess the diagnostic performance of HIF1A-AS2 expression in acute myocardial infarction Fig. 1 ROC analysis to assess the diagnostic performance of HIF1A-AS2 expression in acute myocardial infarction Discussion MI is one of the most prevalent cardiovascular diseases characterized by a disruption in the blood and oxygen supply to the heart, leading to gradual myocardial necro- sis and heart failure [18]. Therefore, the identification of an ideal biomarker for early detection of MI can contrib- ute to improving MI diagnosis and prognosis. Table 4 Association between HIF1A‐AS2 expression levels and timing of presentation Table 4 Association between HIF1A‐AS2 expression levels and timing of presentation p p values for Mann Whitney test for comparing between the two groups * Statistically significant at p ≤ 0.05 Onset of chest pain (P-value) ≤ 3 Hour (N = 17) > 3 h (N = 31) HIF1A-AS2 expression Median 28.2 3.1 0.002* Min.—Max 1.6 – 265.3 0.99 – 310.0 Table 4 Association between HIF1A‐AS2 expression levels and timing of presentation In the last two decades, epigenetics has emerged as a regulatory interface between the environmental factors and genome with remarkable advances in high through- put’ omics’ technologies. Due to the limited proliferative ability of human cardiomyocytes, epigenetic regulation appears to play a particularly crucial role in the heart. It permits the heart to respond quickly and adequately to environmental challenges and cardiac stress [19]. Although lncRNAs that control gene expression through diverse mechanisms have been primarily asso- ciated with cancers [14], it is now evident that various lncRNAs play a vital role in cardiac development and function. Their dysregulation is involved in the onset and progression of various cardiovascular diseases [20]. It has been found that some lncRNAs are modu- lators in hypoxia, acting as stimulators or inhibitors of the HIF-pathway and thus called hypoxia responsive lncRNAs [21]. (p = 0.001). The overexpression of HIF1A-AS2 using multi- variate regression analysis was found to be still significantly associated with low EF (P = 0.018, OR 1.133). The ROC curve of HIF1A-AS2 expression was constructed to pre- dict post-MI left ventricular (LV) dysfunction (EF% < 40) in the patient group. The corresponding area under the curve (AUC) was found to be 0.874. The best cut-off value for HIF1A-AS2 when predicting post-MI LV dysfunction in this patient group was fold change > 6.7, where sensitiv- ity = 77.8% and specificity = 80.0% (Fig. 3). (p = 0.001). The overexpression of HIF1A-AS2 using multi- variate regression analysis was found to be still significantly associated with low EF (P = 0.018, OR 1.133). Discussion The ROC curve of HIF1A-AS2 expression was constructed to pre- dict post-MI left ventricular (LV) dysfunction (EF% < 40) in the patient group. The corresponding area under the curve (AUC) was found to be 0.874. The best cut-off value for HIF1A-AS2 when predicting post-MI LV dysfunction in this patient group was fold change > 6.7, where sensitiv- ity = 77.8% and specificity = 80.0% (Fig. 3). HIF1A-AS2, the antisense lncRNA transcribed from HIF-1a mRNA, has been identified in cancer processes Tayae et al. BMC Cardiovascular Disorders (2023) 23:135 Page 6 of 9 Fig. 2 ROC analysis to assess the diagnostic performance of HIF1A-AS2 expression as an early marker for acute myocardial infarction o assess the diagnostic performance of HIF1A-AS2 expression as an early marker for acute myocardial infarction Fig. 2 ROC analysis to assess the diagnostic performance of HIF1A-AS2 expression as an early marker for acute myocardia Table 5 Associations between HIF1A‐AS2 expression and clinical and laboratory parameters in acute myocardial infarction patients (n = 48) U Mann Whitney test, H H for Kruskal Wallis test * Statistically significant at p ≤ 0.05 HIF1A‐AS2 expression Test of Sig P Median (Min. – Max.) Sex Male 5.70 (1.27- 310.0) U = 252.0 0.930 Female 6.01 (0.99 – 132.0) Type NSTEMI 6.70 (0.99–265.32) U = 255.0 0.554 STEMI 4.70 (1.14–310.0) Smoking Nonsmoker 7.9 (1.0 – 265.3) H = 2.322 0.313 Smoker 4.6 (1.3 – 114.7) Ex-smoker 3.1 (1.6 – 310.0) Diabetes Mellitus No 6.02 (1.27—180.0) U = 252.5 0.939 Yes 4.7 (1.0 – 310.0) Hypertension No 3.1 (1.3 – 114.7) U = 212.50 0.140 Yes 6.7 (1.0 – 310.0) Ejection fraction ≤ 40 3.1 (0.99 – 28.2) U = 149.5* 0.001* > 40 63.8 (2.6 – 310.0) Table 5 Associations between HIF1A‐AS2 expression and clinical and laboratory parameters in acute myocardial infarction patients (n = 48) associated with hypoxia. Consequently, it is proposed to be involved in physiological and pathological processes in ischemic and hypoxic conditions [22]. Two previous studies have suggested that HIF1A‐AS2 is overexpressed in MI and coronary heart disease [23, 24], highlighting its potential as a MI biomarker and therapeutic target. How- ever, its role as an early diagnostic and prognostic marker remains unclear. The current study was designed to determine the clini- cal significance of HIF1A-AS2 in AMI as a biomarker for early diagnosis (in the first 3 h from pain onset) and prognosis. Discussion It was found that there was a highly significant increase in the levels of HIF1A-AS2 expression in MI patients when compared to the control group with 85.4% sensitivity and 86% specificity at a cut-off value > 1.8. Interestingly, we found that the HIF1A-AS2 level is sub- stantially higher in the first 3 h from the onset of chest pain than later, and when compared with its level in con- trols, we found that it has a higher sensitivity (94%) and the same specificity as an early biomarker. i According to previous research, the HIF1A-AS2 pro- moter shows several putative hypoxia response elements (HRE), which can explain the overexpression of HIF1A- AS2 under hypoxic conditions present in AMI. In addi- tion, it was found that HIF1A-AS2 could expose AU-rich elements present in the 3’ untranslated region of HIF-1a mRNA, thereby destabilizing HIF-1A mRNA and thus Tayae et al. BMC Cardiovascular Disorders (2023) 23:135 Page 7 of 9 Fig. 3 ROC analysis to assess the performance of HIF1A-AS2 expression as a predictor of left ventricular dysfunction in acute myocardial infarction ig. 3 ROC analysis to assess the performance of HIF1A-AS2 expression as a predictor of left ventricular dysfunction in acute m o assess the performance of HIF1A-AS2 expression as a predictor of left ventricular dysfunction in acute myocardial infarction Fig. 3 ROC analysis to assess the performance of HIF1A-AS2 expression as a predictor of left ventricular dysfunction in acute myocardial infarction possibly increasing the degradation of HIF-1a mRNA, inducing a negative loop of regulation of HIF-1a expres- sion [10]. may explain why HIF1A‐AS2 was significantly higher in patients who presented within 3 h of chest pain onset than patients who later presented a phase where mainly neutrophils become activated. In another study, HIF1A-AS2 was found to be over- expressed in human umbilical vein endothelial cells (HUVECs) in hypoxia. However, Luciferase reporter assay revealed that HIF1A-AS2 upregulates HIF-1α by sponging miR-153-3p, which is responsible for posttran- scriptional silencing of HIF-1α. HIF1A-AS2 was sug- gested to promote angiogenesis in HUVECs in hypoxia by increasing HIF-1 α expression [22]. The association between HIF1A-AS2 and inflamma- tion was also proposed in another study. They found that HIF1A-AS2 elevates the expression of activating tran- scription factor 2 (ATF2). ATF2 is expressed in a vari- ety of tissues and involved in inflammatory responses. HIF1A-AS2 also was found to be highly expressed in atherosclerotic mice. Discussion Downregulation of HIF1A-AS2 inhibited inflammation by suppressing the levels of pro- inflammatory factors and adhesion molecules. Therefore, HIF1A-AS2 was proposed as a promising biomarker for coronary artery disease (CAD) and a therapeutic target for atherosclerosis [25]. In the current study, HIF1A‐AS2 was positively corre- lated with inflammatory markers such as CRP and WBCs count in the patient group. Vausort et al. also found positive associations between HIF1A‐AS2 and several inflammatroy markers measured at admission, including CRP, white blood cell count, and percentage of neutro- phils. They suggested that the increased expression of lncRNA is not due to hypoxia, but rather the inflamma- tory response to MI. They measured the expression of HIF1A‐AS2 in different blood cells from healthy donors and found that monocytes and neutrophils are the pri- mary sources of HIF1A‐AS2. In addition, they revealed a strong positive association between HIF1A‐AS2 and the percentage of circulating neutrophils. Therefore, they suggested that the upregulation of HIF1A‐AS2 in patients with MI is mostly because of neutrophils [23], which We found a positive correlation between HIF1A-AS2 and known specific cardiac markers in the patient group, specifically with troponin. As reported, troponin is corre- lated with infarct size and severity [26]. Therefore, there may be a potential association between HIF1A‐AS2 and infarct severity. In the current study, no significant correlation was found between HIF1A‐AS2 and age. In addition, no sig- nificant association was detected between HIF1A‐AS2 and gender, type of MI or any risk factors for MI as with smoking, diabetes mellitus, hypertension, among patient Tayae et al. BMC Cardiovascular Disorders (2023) 23:135 Page 8 of 9 groups. This finding agrees with Vausort. et al. and Zhang. et al. concluded that HIF1A‐AS2 is not affected by age, gender, smoking, diabetes mellitus, or hyperten- sion [23, 24]. comorbidities using larger number of patients. Extended multicentric studies is also recommended to assess the prognostic value of HIF1A-AS2 as a predictor of late ventricular dysfunction in the initial evaluation of the patients. This may be useful in choosing the revascu- larization strategy to be initially adopted in a STEMI patient, or in deciding about the urgency of an invasive evaluation in the face of a patient without ST-segment elevation MI. Subsequently, we investigated whether HIF1A-AS1 could be used as a predictor for LV dysfunction after 6 months follow up. Funding O Open access funding provided by The Science, Technology & Innovation Funding Authority (STDF) in cooperation with The Egyptian Knowledge Bank (EKB). The authors received no financial support for the research, authorship, and/or publication of this article. Availability of data and materials The data supporting the conclusions are included within the article. Acknowledgements Acknowledgements Authors would like to thank all participants in this study. Authors would like to thank all participants in this study. Authors’ contributions ET and DE were responsible for the study design. EA and AZ were responsible for the recruitment of the study subjects, collection of clinical information. ET, DE and EA were responsible for analysis of the overall data, and drafted the manuscript. EA conducted relative quantification of HIF1A-AS2 by quantitative real-time PCR. ET, DE, and EA reviewed and edited the manuscript. All the authors read and approved the final form of the manuscript. Early diagnosis as well as rapid treatment with tar- geted drugs are both crucial in the management of myocardial infarction patients. Example for targeted druges in myocardial infarction is alirocumab, which is proprotein convertase subtilisin kexin type 9 (PCSK9) inhibitors. PSCK9 was discovered to be associated with elevated low-denisity lipoprotein (LDL) and adverse cardiovascular outcomes. It was observed that among acute myocardial infarction patients, the early addition of alirocumab, to statin therapy resulted in greater coronary atherosclerotic plaque regression in non–infarct-related arteries with subsequent improve- ment of patients’ management, treatment adherence, and quality of life [28–30].i Ethics approval and consent to participate Based on the previous findings, we may conclude that HIF1A‐AS2 might be used as a potential diagnostic bio- marker for early MI as well as a predictor for post-MI heart failure. Furthermore, it paves the way for further research to investigate its value as a therapeutic target for MI and heart failure. The study was approved by the Ethics Committee of Alexandria University, Egypt. Informed consent was obtained from all individual participants included in the study. All methods were performed in accordance with the relevant guidelines and regulations. Discussion We found that HIF1A-AS1 levels can be used as a predictor for LV dysfunction after using multivariate logistic regression analysis with 77.8% sensi- tivity and 80.0% specificity. i The only available data about the role of HIF1A-AS1 as a predictor for post-MI heart failure is from the Vau- sort V. et al. study. However, they found no association between HIF1A-AS1 and LV dysfunction at a 4-month follow-up, as demonstrated by EF ≤ 40%. Abbreviations AMI Acute myocardial infarction lncRNAs Long non-coding RNAs HIF1A-AS2 Hypoxia-inducible factor 1A antisense RNA 2 LVD Left ventricular dysfunction IHD Ischemic heart disease STEMI ST-segment–elevation MI ECG Electrocardiographic HIF1α Hypoxia-inducible factor 1α qPCR Quantitative real-time PCR SD Standard deviation ROC Receiver operating characteristic curve EF Ejection fraction AUC Area under the curve HUVECs Human umbilical vein endothelial cells WBC White blood cells count CRP C- reactive protein ATF2 Activating transcription factor 2 HMCs Hypoxia-treated human cardiomyocytes The effect of HIF1A-AS1 overexpression on myo- cardial apoptosis has been experimentally studied to reveal a novel molecular-targeted therapy in MI. Luo et al. found that the HIF1A-AS2 level is upregulated in hypoxia-treated human cardiomyocytes (HMCs) com- pared with normal cardiomyocytes. Through increased TRIM44 expression, HIF1A‐AS2 inhibits the Akt path- way via sponging miR‐623 to promote apoptosis of car- diomyocytes. In contrast, HIF1A‐AS2 silencing led to the repression of cardiomyocyte apoptosis and an increase in cardiomyocyte viability, migration, and invasion. These findings propose the promoting effect of HIF1A‐AS2 in MI injury induced by hypoxia and underscore the poten- tial role of HIF1A‐AS2 in targeted treatment for MI patients [27]. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Competing interests The authors declare no competing interests. We recommend to study the diagnostic perfor- mance of HIF1A-AS2 in different clinical settings and Publisher’s Note 6. Herman AB, Occean JR, Sen P. Epigenetic dysregulation in cardiovascular aging and disease. J Cardiovasc Aging. 2021;1:10. 6. Herman AB, Occean JR, Sen P. Epigenetic dysregulation in cardiovascular aging and disease. J Cardiovasc Aging. 2021;1:10. Springer Nature remains neutral with regard to jurisdictional claims in pub- lished maps and institutional affiliations. Springer Nature remains neutral with regard to jurisdictional claims in pub- lished maps and institutional affiliations. 7. Li J, Liu C. Coding or Noncoding, the Converging Concepts of RNAs. Front Genet. 2019;10:496. 7. Li J, Liu C. Coding or Noncoding, the Converging Concepts of RNAs. Front Genet. 2019;10:496. 8. Zhang X, Wang W, Zhu W, et al. Mechanisms and Functions of Long Non-Coding RNAs at Multiple Regulatory Levels. Int J Mol Sci. 2019;20(22):5573. 8. Zhang X, Wang W, Zhu W, et al. Mechanisms and Functions of Long Non-Coding RNAs at Multiple Regulatory Levels. Int J Mol Sci. 2019;20(22):5573. 9. Statello L, Guo CJ, Chen LL, Huarte M. Gene regulation by long non-coding RNAs and its biological functions. Nat Rev Mol Cell Biol. 2021;22(2):96–118. 9. Statello L, Guo CJ, Chen LL, Huarte M. Gene regulation by long non-coding RNAs and its biological functions. Nat Rev Mol Cell Biol. 2021;22(2):96–118. 10. Kuo TC, Kung HJ, Shih JW. Signaling in and out: long-noncoding RNAs in tumor hypoxia. J Biomed Sci. 2020;27(1):59. 11. Xie L, Zhang Q, Mao J, Zhang J, Li L. The Roles of lncRNA in Myocardial Infarction: Molecular Mechanisms, Diagnosis Biomarkers, and Therapeutic Perspectives. Front Cell Dev Biol. 2021;9:680713. 11. Xie L, Zhang Q, Mao J, Zhang J, Li L. The Roles of lncRNA in Myocardial Infarction: Molecular Mechanisms, Diagnosis Biomarkers, and Therapeutic Perspectives. Front Cell Dev Biol. 2021;9:680713. 12. Thrash-Bingham CA, Tartof KD. aHIF: a natural antisense transcript over- expressed in human renal cancer and during hypoxia. J Natl Cancer Inst. 1999;91:143–51. 12. Thrash-Bingham CA, Tartof KD. aHIF: a natural antisense transcript over- expressed in human renal cancer and during hypoxia. J Natl Cancer Inst. 1999;91:143–51. 13. Rossignol F, Vache C, Clottes E. Natural antisense transcripts of hypoxia- inducible factor 1alpha are detected in different normal and tumour human tissues. Gene. 2002;299:135–40. 13. Rossignol F, Vache C, Clottes E. Natural antisense transcripts of hypoxia- inducible factor 1alpha are detected in different normal and tumour human tissues. Gene. 2002;299:135–40. 14. Shih JW, Kung HJ. Long non-coding RNA and tumor hypoxia: new players ushered toward an old arena. References 27. Luo F, Wu Y, Zhu L, et al. Knockdown of HIF1A-AS2 suppresses TRIM44 to protect cardiomyocytes against hypoxia-induced injury. Cell Biol Int. 2020;44:1523–34. 1. Roth GA, Mensah GA, Johnson CO, Addolorato G, Ammirati E, Baddour LM, et al. Global Burden of Cardiovascular Diseases and Risk Factors, 1990–2019: Update From the GBD 2019 Study. J Am Coll Cardiol. 2020;76(25):2982–3021. 28. Cesaro A, Gragnano F, Fimiani F, Moscarella E, Diana V, Pariggiano I, et al. Impact of PCSK9 inhibitors on the quality of life of patients at high cardio- vascular risk. Eur J Prev Cardiol. 2020;27:556–8. 2. Kingma J. Myocardial Infarction: An Overview of STEMI and NSTEMI Physi- opathology and Treatment. World J Cardiovasc Dis. 2018;8:498–517. 2. Kingma J. Myocardial Infarction: An Overview of STEMI and NSTEMI Physi- opathology and Treatment. World J Cardiovasc Dis. 2018;8:498–517. 29. Gragnano F, Natale F, Concilio C, Fimiani F, Cesaro A, Sperlongano S, et al. Adherence to proprotein convertase subtilisin/kexin 9 inhibitors in high cardiovascular risk patients: an Italian single-center experience. J Cardio- vasc Med (Hagerstown). 2018;19:75–7. 3. Amsterdam EA, Wenger NK, Brindis RG, Casey DE, Ganiats TG, Holmes DR, et al. AHA/ACC guideline for the management of patients with non-ST-elevation acute coronary syndromes: a report of the American College of Cardiology/ American Heart Association Task Force on Practice Guidelines. J Am Coll Cardiol. 2014;64(24):e139-228. 3. Amsterdam EA, Wenger NK, Brindis RG, Casey DE, Ganiats TG, Holmes DR, et al. AHA/ACC guideline for the management of patients with non-ST-elevation acute coronary syndromes: a report of the American College of Cardiology/ American Heart Association Task Force on Practice Guidelines. J Am Coll Cardiol. 2014;64(24):e139-228. 30. Räber L, Ueki Y, Otsuka T, Losdat S, Häner JD, Lonborg J, et al. Effect of Alirocumab Added to High-Intensity Statin Therapy on Coronary Athero- sclerosis in Patients With Acute Myocardial Infarction: The PACMAN-AMI Randomized Clinical Trial. JAMA. 2022;327:1771–81. 30. Räber L, Ueki Y, Otsuka T, Losdat S, Häner JD, Lonborg J, et al. Effect of Alirocumab Added to High-Intensity Statin Therapy on Coronary Athero- sclerosis in Patients With Acute Myocardial Infarction: The PACMAN-AMI Randomized Clinical Trial. JAMA. 2022;327:1771–81. 4. Reed GW, Rossi JE, Cannon CP. Acute myocardial infarction. Lancet. 2017;389(10065):197–210. 5. Thygesen K, Alpert JS, Jaffe AS, Chaitman BR, Bax JJ, Morrow DA, et al. Fourth universal definition of myocardial infarction. J Am Coll Cardiol. 2018;72(18):2231–64. Received: 14 December 2022 Accepted: 2 March 2023 25. Li P, Xing J, Zhang J, et al. Inhibition of long non-coding RNA HIF1A-AS2 confers protection against atherosclerosis via ATF2 downregulation. J Adv Res. 2020;26:123–35. 26. Tiller C, Reindl M, Holzknecht M, Klapfer M, Beck A, Henninger B, et al. Biomarker assessment for early infarct size estimation in ST-elevation myocardial infarction. Eur J Intern Med. 2019;64:57–62. Competing interests The authors declare no competing interests. Page 9 of 9 Tayae et al. BMC Cardiovascular Disorders (2023) 23:135 Tayae et al. BMC Cardiovascular Disorders (2023) 23:135 Received: 14 December 2022 Accepted: 2 March 2023 Publisher’s Note J Biomed Sci. 2017;24(1):53. 14. Shih JW, Kung HJ. Long non-coding RNA and tumor hypoxia: new players ushered toward an old arena. J Biomed Sci. 2017;24(1):53. 14. Shih JW, Kung HJ. Long non coding RNA and tumor hypox ushered toward an old arena. J Biomed Sci. 2017;24(1):53. 15. Zheng F, Chen J, Zhang X, et al. The HIF-1α antisense long non-coding RNA drives a positive feedback loop of HIF-1α mediated transactivation and glycolysis. Nat Commun. 2021;12(1):1341. 15. Zheng F, Chen J, Zhang X, et al. The HIF-1α antisense long non-coding RNA drives a positive feedback loop of HIF-1α mediated transactivation and glycolysis. Nat Commun. 2021;12(1):1341. g y y 16. Corrado C, Fontana S. Hypoxia and HIF Signaling: One Axis with Diver- gent Effects. Int J Mol Sci. 2020;21(16):5611. y y 16. Corrado C, Fontana S. Hypoxia and HIF Signaling: One Axis with Diver- gent Effects. Int J Mol Sci. 2020;21(16):5611. f 17. Semenza GL. Hypoxia-inducible factor 1 and cardiovascular disease. Annu Rev Physiol. 2014;76:39–56. 17. Semenza GL. Hypoxia-inducible factor 1 and cardiovascular disease. Annu Rev Physiol. 2014;76:39–56. 18. Li M, Chen F, Zhang Y, Xiong Y, Li Q, Huang H. Identification of Post- myocardial Infarction Blood Expression Signatures Using Multiple Feature Selection Strategies. Front Physiol. 2020;11:483. 18. Li M, Chen F, Zhang Y, Xiong Y, Li Q, Huang H. Identification of Post- myocardial Infarction Blood Expression Signatures Using Multiple Feature Selection Strategies. Front Physiol. 2020;11:483. 19. Ameer SS, Hossain MB, Knöll R. Epigenetics and Heart Failure. Int J Mol Sci. 2020;21(23):9010. 19. Ameer SS, Hossain MB, Knöll R. Epigenetics and Heart Failure. Int J Mol Sci. 2020;21(23):9010. • fast, convenient online submission • thorough peer review by experienced researchers in your ﬁeld • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year • At BMC, research is always in progress. Learn more biomedcentral.com/submissions Ready to submit your research Ready to submit your research ? Choose BMC and benefit from: ? Publisher’s Note Choose BMC and benefit from: • fast, convenient online submission • thorough peer review by experienced researchers in your ﬁeld • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year • At BMC, research is always in progress. Learn more biomedcentral.com/submissions Ready to submit your research Ready to submit your research ? Choose BMC and benefit from: ? Choose BMC and benefit from: • fast, convenient online submission • thorough peer review by experienced researchers in your ﬁeld • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year • At BMC, research is always in progress. Learn more biomedcentral.com/submissions Ready to submit your research Ready to submit your research ? Choose BMC and benefit from: ? Choose BMC and benefit from: 20. García-Padilla C, Aránega A, Franco D. The role of long non-coding RNAs in cardiac development and disease. AIMS Genet. 2018;5(2):124–40. 20. García-Padilla C, Aránega A, Franco D. The role of long non-coding RNAs in cardiac development and disease. AIMS Genet. 2018;5(2):124–40. 21. Barth DA, Prinz F, Teppan J, Jonas K, Klec C, Pichler M. Long-Noncoding RNA (lncRNA) in the Regulation of Hypoxia-Inducible Factor (HIF) in Cancer. Noncoding RNA. 2020;6(3):27. g 22. Li L, Wang M, Mei Z, et al. lncRNAs HIF1A-AS2 facilitates the up-regulation of HIF-1α by sponging to miR-153-3p, whereby promoting angiogenesis in HUVECs in hypoxia. Biomed Pharmacother. 2017;96:165–72. 23. Vausort M, Wagner DR, Devaux Y. Long non-coding RNAs in patients with acute myocardial infarction. Circ Res. 2014;115:668–77. 24. Zhang Y, Zhang L, Wang Y, Ding H, Xue S, Yu H, Hu L, Qi H, Wang Y, Zhu W, Liu D, Li P. KCNQ1OT1, HIF1A-AS2 and APOA1-AS are promising novel biomarkers for diagnosis of coronary artery disease. Clin Exp Pharmacol Physiol. 2019;46:635–42. 24. Zhang Y, Zhang L, Wang Y, Ding H, Xue S, Yu H, Hu L, Qi H, Wang Y, Zhu W, Liu D, Li P. KCNQ1OT1, HIF1A-AS2 and APOA1-AS are promising novel biomarkers for diagnosis of coronary artery disease. Clin Exp Pharmacol Physiol. 2019;46:635–42.
https://openalex.org/W4312642705	https://repositorio.iscte-iul.pt/bitstream/10071/26533/1/conferenceobject_89442.pdf	English	null	Impact of Traffic Load and Spectral Occupancy on Gaussian Noise Models Performance for Multiband Networks	2022 13th International Symposium on Communication Systems, Networks and Digital Signal Processing (CSNDSP)	2,022	cc-by	6,340	Citation for published item: p Venda, P., Rebola, J. & Cancela, L. (2022). Impact of traffic load and spectral occupancy on Gaussian noise models performance for multiband networks. In 2022 13th International Symposium on Communication Systems, Networks and Digital Signal Processing (CSNDSP) . (pp. 240-245). Porto: IEEE. Peer-review status of attached file: Peer-reviewed Serviços de Informação e Documentação, Instituto Universitário de Lisboa (ISCTE-IUL) Av. das Forças Armadas, Edifício II, 1649-026 Lisboa Portugal Phone: +(351) 217 903 024 \| e-mail: administrador.repositorio@iscte-iul.pt https://repositorio.iscte-iul.pt Creative Commons CC BY 4.0 The full-text may be used and/or reproduced, and given to third parties in any format or medium, without prior permission or charge, for personal research or study, educational, or not-for-profit purposes provided that: Publisher's copyright statement: This is the peer reviewed version of the following article: Venda, P., Rebola, J. & Cancela, L. (2022). Impact of traffic load and spectral occupancy on Gaussian noise models performance for multiband networks. In 2022 13th International Symposium on Communication Systems, Networks and Digital Signal Processing (CSNDSP) . (pp. 240-245). Porto: IEEE., which has been published in final form at https://dx.doi.org/10.1109/CSNDSP54353.2022.9907947. This article may be used for non- commercial purposes in accordance with the Publisher's Terms and Conditions for self-archiving. Creative Commons CC BY 4.0 Use policy Use policy Creative Commons CC BY 4.0 The full-text may be used and/or reproduced, and given to third parties in any format or medium, without prior permission or charge, for personal research or study, educational, or not-for-profit purposes provided that: Creative Commons CC BY 4.0 Creative Commons CC BY 4.0 The full-text may be used and/or reproduced, and given to third parties in any format or medium, without prior permission or charge, for personal research or study, educational, or not-for-profit purposes provided that: • a full bibliographic reference is made to the original source • a full bibliographic reference is made to the original source • a link is made to the metadata record in the Repository • a link is made to the metadata record in the Repository • the full-text is not changed in any way • the full-text is not changed in any way The full-text must not be sold in any format or medium without the formal permission of the copyright holders. Serviços de Informação e Documentação, Instituto Universitário de Lisboa (ISCTE-IUL) Av. das Forças Armadas, Edifício II, 1649-026 Lisboa Portugal Phone: +(351) 217 903 024 \| e-mail: administrador.repositorio@iscte-iul.pt https://repositorio.iscte-iul.pt I. INTRODUCTION The Gaussian-noise (GN) model is an efﬁcient and widely adopted tool to estimate the nonlinear interference (NLI) due to Kerr nonlinearity, simplifying current wavelength-division multiplexing (WDM) systems design and analysis [1], [2]. Re- cently, the generalized Gaussian-noise (GGN) model has been proposed for assessing the performance of multiband C+L transmission in WDM optical systems, where the interaction between NLI and stimulated Raman scattering (SRS) must be accurately characterized [3]–[5]. This paper is organized as follows. In section II, the main deﬁnitions and assumptions of the optical network simulator developed to account for dynamic trafﬁc load and variable WDM channels spectral occupation are presented and ex- plained. In section III, the performance of the two closed-form GN models is compared and the impact of the network trafﬁc load, channels add/drop and spectral occupancy on the OSNR estimation is quantiﬁed. Section IV presents the conclusions. For C-band transmission, various closed-form GN-model approximations have been proposed over the years [1], [6]– [9], while, recently, GGN-model closed formulas have been derived for transmission networks operating beyond the C- band [5], [10], the so-called multiband networks. For optical networking, both GN and GGN models closed forms allow fast and reliable analysis of the network physical layer im- pairments, enabling the development and implementation of more rigorous real time network optimization tools [1], [5], [11]. Impact of Trafﬁc Load and Spectral Occupancy on Gaussian Noise Models Performance for Multiband Networks Pedro Venda⇤, Jo˜ao Rebola⇤†, Lu´ıs Cancela⇤† ⇤Instituto Universit´ario de Lisboa (ISCTE-IUL), Lisboa, Portugal †Optical Communications and Photonics Group, Instituto de Telecomunicac¸˜oes, Lisboa, Portugal E-mail: pdsva@iscte-iul.pt, joao.rebola@iscte-iul.pt, luis.cancela@iscte-iul.pt the full network capacity. Network performance studies using the GGN-model have been presented in [13], [14]. In [13], the capacity gain of reducing the link margin is studied in a dynamic scenario considering the C+L band and a blocking probability of 10%, which can be considered too high for network normal operation [15], [16]. In [14], a comparison between a worst-case network with a full C+L band occu- pancy and the same unlikely dynamic trafﬁc case with the 10% blocking probability is performed. In [5], a preliminary study of the effect of the dynamic trafﬁc variations on NLI predictions using a closed-form GGN-model was carried out for validation purposes, without assessing its impact on the optical signal-to-noise ratio (OSNR). Abstract—In a network scenario, wavelength division- multiplexing channels are added and dropped leading to ﬂuc- tuations on the network trafﬁc loads along the optical path. In this work, a comparison between the optical signal-to-noise ratio (OSNR) predictions of the recently proposed closed-form generalized Gaussian noise (GGN) model and a closed-form Gaussian noise (GN) model that does not take into account the stimulated Raman scattering (SRS) is performed, for different network trafﬁc loads and spectral occupancy over the entire C+L band. In all results obtained, the maximum difference between the OSNR predictions of GN (without SRS) and GGN models closed forms is below 0.7 dB at optimum OSNR and maximum C+L band occupancy, indicating that the GN-model can also be used in C+L band transmission. For channel launch powers higher than the optimum, the OSNR differences increase up to 3 dB, being the GN-model (without SRS) unsuitable to assess the network performance in such situations. In this work, the closed-form GGN-model introduced in [5] is investigated against the closed-form GN-model [9] proposed for C-band network transmission and their accuracy for per- formance estimation in C+L mulitband systems is discussed, which to the best of our knowledge has not been done yet. The accuracy of the OSNR predictions is assessed using both models, considering ﬂuctuations on the trafﬁc loads along the lightpath of the network, increasingly ﬁlling the WDM spectral occupancy up to the full range of C+L band and taking into account the effect of the add/drop operations on the OSNR. II. NETWORK SIMULATION The launch power of the added channels has a random offset of ±1 dB relative to the launch power of the CUTs and, for a more realistic approach, the added channels always maintain the same power until they are dropped, i.e., for a given lightpath, the power of a channel only changes if it is added to the optical network more than one time. This last assumption differs from the optical transmission scenario considered in [5], where the add/drop channels that have not been dropped may not maintain the power they had in the previous span. The launch powers of the CUTs remain the same along the complete lightpath. 92 km 105 km 82 km 81 km B 73 km ROADM ROADM ROADM ROADM Rx OSNR measurement (b) (b) (b) Fig. 1. Two lightpaths examples taken from the British Telecommunications topology of the United Kingdom core network [5]. Throughout this work, the lightpaths in (a) and (b) are referred to as green and red lightpaths, respectively. -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 0 0.5 1 1.5 2 Fig. 2. Randomly generated normalized PSD of a WDM signal transmitted in a network scenario, at the input (z = 0) and at the end (z = Ls) of a span, for ✏network = 80% and ✏occupancy ⇡35%. Blue lines: WDM signal PSD at z = 0. Orange lines: WDM signal PSD at z = Ls. -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 0 0.5 1 1.5 2 into account in this work. The post-ampliﬁers are designed to perfectly compensate the ROADM losses. The inline and pre-ampliﬁers compensate perfectly the previous ﬁber losses. All optical ampliﬁers considered in the lightpaths are erbium doped ﬁber ampliﬁers (EDFAs) with dynamic gain equaliza- tion [3]. Thus, the EDFAs also compensate the power transfer in each span due to SRS. For the analysis of the lightpaths represented in Fig. 1, the WDM channels are classiﬁed in two categories: channels under test (CUTs) and add/drop channels. The CUTs correspond to the WDM channels that are transmitted along the complete lightpath, i.e., from the ﬁrst node to the last node without any add/drop occurring in those wavelengths. The add/drop channels are the wavelengths that can be added or dropped in any ROADM of the considered lightpath. We also introduce the two following deﬁnitions: network utilization and C+L band occupancy. II. NETWORK SIMULATION In this section, the simulator developed for assessing the performance of a multiband optical network and related as- sumptions are presented and explained. In reconﬁgurable optical add/drop multiplexer (ROADM) based networks, the WDM channels can be added, dropped or expressed [12], leading to dynamic trafﬁc variations in the ﬁber spans. Thus, the number and the power of WDM channels transmitted in a given span of a given lightpath vary constantly, meaning that the WDM signal bandwidth does not always fully covers the span maximum capacity, and consequently, The two lightpaths examples studied in this work, taken from the British Telecommunications network topology [5], are illustrated in Fig. 1. The ROADMs architecture considered for the optical node is the route-and-select with maximum losses of 18 dB [17]. The ﬁltering effects and crosstalk due to ROADM components imperfections [18] are not taken 99 km 98 km 101 km 102 km 27 km 80 km 80 km 75 km A B ROADM Rx OSNR measurement ROADM ROADM ROADM ROADM ROADM (a) 113 km 113 km 57 km 58 km 91 km 92 km 105 km 82 km 81 km A B 73 km ROADM ROADM ROADM ROADM ROADM ROADM ROADM Rx OSNR measurement (b) Fig. 1. Two lightpaths examples taken from the British Telecommunications topology of the United Kingdom core network [5]. Throughout this work, the lightpaths in (a) and (b) are referred to as green and red lightpaths, respectively. where Nch,j is the number of channels transmitted in the j- th span, Ns is the number of spans in the lightpath and Nch is the total number of WDM channels considered. The C+L band occupancy ✏occupancy is deﬁned as ✏occupancy = ∆fNch BC+L-band (2) (2) where ∆f is the channel spacing and BC+L-band is the total C+L optical transmission bandwidth, assumed as 11.5 THz [19], which corresponds to 100% C+L band occupancy. p p y In order to emulate the behavior of dynamic trafﬁc load variations occurring in each ROADM node, we consider that the add/drop channels are added or dropped randomly following a uniform distribution. The number of add/drop channels depends on the required ✏network. II. NETWORK SIMULATION For a given span of a lightpath, the ratio between the number of channels transmitted in the j-th span and the total number of channels is deﬁned as span utilization, which is denoted as ✏span,j. For the entire lightpath, the network utilization ✏network is deﬁned as the average of the network utilizations of all spans, written as Fig. 2. Randomly generated normalized PSD of a WDM signal transmitted in a network scenario, at the input (z = 0) and at the end (z = Ls) of a span, for ✏network = 80% and ✏occupancy ⇡35%. Blue lines: WDM signal PSD at z = 0. Orange lines: WDM signal PSD at z = Ls. A randomly generated normalized power spectral density (PSD) of a transmitted WDM signal GWDM(f) is shown in Fig. 2, at the input z = 0 and at the end z = Ls of a ﬁber span, for a network utilization ✏network = 80% and a C+L band occupancy ✏occupancy ⇡35%. The WDM signal PSDs shown are normalized to the PSD value of the CUTs. Notice that, as a consequence of the considered scenario with ✏network = 80%, there is the presence of unused channel frequency slots and oscillations in the PSD amplitudes. For ✏network = 1 Ns Ns X j=1 ✏span,j = 1 Ns Ns X j=1 Nch,j Nch (1) 0 2 4 6 8 10 12 14 5 10 15 -5 THz -2.5 THz 0 THz 2.5 THz 5 THz 0 2 4 6 8 10 12 14 5 10 15 -5 THz -2.5 THz 0 THz 2.5 THz 5 THz Lightpath in green Lightpath in red Fig. 3. OSNR as a function of the power transfer between the outer channels ∆⇢(L), for ✏occupancy = 87%, ✏network = 90% , NCUT = 20% · Nch and the red and green lightpaths. z = 0, these oscillations are only a consequence of the add/drop channels having a launch power with a random offset of ±1 dB relative to CUTs power. For z = Ls, i.e., after the WDM signal has been transmitted through the optical ﬁber, the PSD variation is tilted due to the impact of the SRS, which causes a power transfer ∆⇢(L) ⇡4.8 dB from higher to lower frequency components. II. NETWORK SIMULATION 0 2 4 6 8 10 12 14 5 10 15 -5 THz -2.5 THz 0 THz 2.5 THz 5 THz Lightpath in red All NLI contributions along an optical link composed by several ﬁber spans sum coherently or incoherently along the signal propagation until the receiver [6]. The coherent and incoherent variants of the GGN-model will be studied and referred as the coherent GGN-model and incoherent GGN- model, respectively. Due to the asymptotic expansion of the dilog function performed in its derivation, we will refer to the GN-model proposed for the C-band as the asymptotic GN- model [9]. It is important to notice that this model considers that the NLI accumulates incoherently along the optical path. It should be noted that among the GN models studied, the coherent GGN-model is the one that provides more accurate estimates of the normalized NLI power ⌘NLI in comparison with the split-step Fourier method [5], and therefore, it is the more accurate model to estimate the OSNR. The network simulator used in this work has been successfully validated by obtaining the results shown in Fig. 8 of [5]. 0 2 4 6 8 10 12 14 5 10 15 -5 THz -2.5 THz 0 THz 2.5 THz 5 THz Lightpath in green Fig. 3. OSNR as a function of the power transfer between the outer channels ∆⇢(L), for ✏occupancy = 87%, ✏network = 90% , NCUT = 20% · Nch and the red and green lightpaths. TABLE I SYSTEM PARAMETERS FOR SECTION III System parameters Maximum number of WDM of channels (Nch) 201 Number of channels under test (NCUT ) 41 Symbol rate (Rs) [GBaud] 32 Channel bandwidth (Bm) [GHz] 32 Loss coefﬁcient (↵) [dB/km] 0.22 Dispersion (β2) [ps/nm/km] 16.7 Dispersion slope (Sr) [ps/nm2/km] 0.067 NLI coefﬁcient (γ) [W−1.km−1] 1.3 Raman gain slope (Cr) [W−1km−1THz−1] 0.028 EDFA noise ﬁgure (Fn) [dB] 5 III. GN MODELS COMPARISON In the following, the impact of the network utilization and C+L band occupancy on the OSNR estimation using the GGN- model [5, Eq. (5)] and the asymptotic GN-model [9, Eq. (16)] is evaluated. The system parameters considered for the OSNR calculation are presented in Table I. In this work, we assume that the impact of the SRS on the ASE noise power is negligible. In order to assess the optimum optical power, in Fig. 3, the OSNR is shown as a function of the SRS power transfer between the WDM outer channels ∆⇢(L), for ✏occupancy = 87%, ✏network = 90% and for the channels at the following frequencies: -5, -2.5, 0, 2.5 and 5 THz. The power transfer is increased by increasing the CUTs launch power [3]. uniform distribution, the remaining CUTs are chosen randomly until 20% of the available WDM channels are occupied, i.e., NCUT = 20%Nch. Two channels spacings are considered: 50 GHz (blue lines) and 100 GHz (yellow lines). To increase the C+L band occupancy, the channels are added sequentially in the total available bandwidth, i.e., ﬁrst, the C-band is ﬁlled and then the L-band. The division that marks the end of the C-band and the beginning of the L-band is highlighted by dashed black vertical lines. For the center channel, the CUTs launch power of approxi- mately 0 dBm leads to the power transfer of around 4 dB that allows obtaining the maximum OSNR, for both lightpaths. For this optimum power, i.e., when the OSNR maximum is reached in Fig. 3, the maximum OSNR variation between the ﬁve WDM channels is only about 0.7 dB. In Fig. 4(b), for the center channel and the two channel spac- ings, the OSNR predictions using the asymptotic GN-model show a very good agreement with the ones obtained with the incoherent GGN-model for all the values of the C+L band occupancy. Due to higher NLI predictions, the coherent GGN- model provides lower OSNRs in the center channel, with a 0.2 dB maximum difference relative to the other two GN models. When comparing to the incoherent GGN-model results, it can be seen that, for ∆f = 50 GHz and a ✏occupancy above 70%, the asymptotic GN-model overestimates and underestimates the OSNR for the lowest and highest WDM frequencies, A. OSNR over the C+L band at optimum launch power The OSNR and power transfer ∆⇢(L) as a function of the C+L band occupancy are shown in Fig. 4(a) for the WDM lowest frequency channel, in Fig. 4(b) for the center channel and in Fig. 4(c) for the highest frequency channel. The network utilization considered is about 95% and the CUTs power is set to PCUT = 0 dBm, which, for the C+L band system studied, leads approximately to the maximum OSNR for all WDM channels, as shown in Fig. 3. The lowest, center and highest frequency channels are always assumed to be CUTs. Using an respectively. When the C+L band is completely ﬁlled and the average power transfer is approximately 6 dB, the difference in OSNR between these models reaches its maximum with 0.3 dB. The maximum OSNR discrepancy between the asymptotic GN-model and the coherent GGN-model is about 0.5 dB and is reached for ∆f = 50 GHz and ✏occupancy = 100%. These higher OSNR discrepancies encountered for ∆f = 50 GHz are due to the higher SRS power transfer relative to the one obtained for the 100 GHz channel spacing. The OSNR differences between the GN models are approximately the same for both lightpaths considered. 0 20 40 60 80 100 15 15.5 16 16.5 0 2 4 6 0 20 40 60 80 100 16.5 17 17.5 18 0 2 4 6 Red lightpath Green lightpath (a) L-band C-band L-band C-band 0 20 40 60 80 100 15 15.6 16.2 0 2 4 6 0 20 40 60 80 100 16 16.5 17 17.5 0 2 4 6 C-band C-band L-band L-band Red lightpath Green lightpath (b) 0 20 40 60 80 100 15 15.5 16 16.5 0 2 4 6 0 20 40 60 80 100 16.5 17 17.5 18 0 2 4 6 L-band C-band L-band C-band Red lightpath Green lightpath (c) Fig. 4. OSNR for the red and green lightpaths as a function of the C+L band occupancy, for the center channel in (b) and the lowest and highest WDM frequency channels in (a) and (c), respectively. The network utilization is 95% and NCUT = 20% · Nch. Circles: closed form GGN-model (incoherent). Dashed lines: closed form GGN-model (coherent). Asterisks: asymptotic GN- model. Points: ∆⇢(L). B. OSNR considering different CUTs launch powers In Fig. 5, the OSNR for only the red lightpath is depicted as a function of the CUTs frequencies, for (a) PCUT = −3 dBm, (b) PCUT = 0 dBm and (c) PCUT = 3 dBm, for ✏occupancy = 100% and ✏network = 95%. In contrast to the closed-form GGN-model, the asymptotic GN-model does not take into account the inﬂuence of the dispersion slope on the NLI prediction. For a fair comparison, in these results, a null dispersion slope is considered in the OSNR estimation by the GGN models. The remaining system parameters are presented in Table I. 0 20 40 60 80 100 15 15.6 16.2 0 2 4 6 C-band L-band Red lightpath In Fig. 5(a), since there is no considerable SRS power trans- fer for launch powers below 0 dBm, the difference between the OSNRs is below 0.3 dB for all the GN-models considered. In Fig. 5(b), it can be seen that, due to the enhancement of the SRS effect caused by increasing the optical launch power to 0 dBm, the maximum discrepancy between the asymptotic GN- model and incoherent and coherent GGN models increases to about 0.4 dB and 0.7 dB, respectively. In Fig. 5(c), the considered CUTs power surpasses the optimal power of 0 dBm, and due to SRS, a much sharper tilt in the OSNR can be observed in the GGN-model results. The maximum average power transfer rises, respectively, from 6 dB to 11.9 dB, for the 50 GHz channel spacing, and from 3 dB to 5.9 dB for the 100 GHz spacing, in relation to the 0 dBm launch power. Consequently, the maximum OSNR difference between the asymptotic estimation and coherent and incoherent GGN models increases, respectively, to about 3 dB and 2.4 dB. Notice that the OSNR variation along the CUTs frequency given by the asymptotic GN-model does not predict the OSNR tilt, leading to higher OSNR discrepancies for more pronounced OSNR tilts. B. OSNR considering different CUTs launch powers 0 20 40 60 80 100 16 16.5 17 17.5 0 2 4 6 C-band L-band Green lightpath (b) 0 20 40 60 80 100 15 15.5 16 16.5 0 2 4 6 0 20 40 60 80 100 16.5 17 17.5 18 0 2 4 6 L-band C-band L-band C-band Red lightpath Green lightpath (c) As a main conclusion, the results presented in this sub- section show that the asymptotic GN-model can provide reasonably accurate OSNR predictions in C+L band optical networks at optimum launch power, but can lead to differences in the OSNR of up to 3 dB, when the launch power leads to a signiﬁcant SRS effect. Fig. 4. OSNR for the red and green lightpaths as a function of the C+L band occupancy, for the center channel in (b) and the lowest and highest WDM frequency channels in (a) and (c), respectively. The network utilization is 95% and NCUT = 20% · Nch. Circles: closed form GGN-model (incoherent). Dashed lines: closed form GGN-model (coherent). Asterisks: asymptotic GN- model. Points: ∆⇢(L). A. OSNR over the C+L band at optimum launch power 0 20 40 60 80 100 15 15.5 16 16.5 0 2 4 6 0 20 40 60 80 100 16.5 17 17.5 18 0 2 4 6 Red lightpath Green lightpath (a) L-band C-band L-band C-band C. OSNR for different network utilizations Lastly, it is important to analyze how the network utilization impacts the performance of the GN models when the WDM signal covers the full C+L band, i.e., ✏occupancy = 100%. The OSNR as a function of ✏network is represented in Fig. 6(a) for -5 -2.5 0 2.5 5 10 11 12 13 14 15 16 -5 -2.5 0 2.5 5 13 14 15 16 17 18 19 (a) -5 -2.5 0 2.5 5 12 13 14 15 16 17 18 19 -5 -2.5 0 2.5 5 12 13 14 15 16 17 18 19 (b) -5 0 5 10 11 12 13 14 15 16 17 -5 0 5 10 11 12 13 14 15 16 17 (c) Fig. 5. OSNR for the red lightpath as a function of the CUTs frequencies, for ✏occupancy = 100%, ✏network = 95%, NCUT = 20% · Nch, Sr = 0 and the launch powers of (a) -3 dBm, (b) 0 dBm and (c) 3 dBm. Channel spacings ∆f = 50 GHz and ∆f = 100 GHz are considered. Circles: closed-form GGN-model (incoherent). Dashed lines: closed-form GGN-model (coherent). Asterisks: asymptotic GN-model. Points: ∆⇢(L). -5 -2.5 0 2.5 5 10 11 12 13 14 15 16 -5 -2.5 0 2.5 5 13 14 15 16 17 18 19 (a) -5 -2.5 0 2.5 5 12 13 14 15 16 17 18 19 -5 -2.5 0 2.5 5 12 13 14 15 16 17 18 19 (b) -5 0 5 10 11 12 13 14 15 16 17 -5 0 5 10 11 12 13 14 15 16 17 (c) Fig. 5. OSNR for the red lightpath as a function of the CUTs frequencies, for ✏occupancy = 100%, ✏network = 95%, NCUT = 20% · Nch, Sr = 0 and the launch powers of (a) -3 dBm, (b) 0 dBm and (c) 3 dBm. Channel spacings ∆f = 50 GHz and ∆f = 100 GHz are considered. Circles: closed-form GGN-model (incoherent). Dashed lines: closed-form GGN-model (coherent). Asterisks: asymptotic GN-model. Points: ∆⇢(L). C. OSNR for different network utilizations -5 -2.5 0 2.5 5 10 11 12 13 14 15 16 -5 -2.5 0 2.5 5 13 14 15 16 17 18 19 (a) 40 60 80 100 12 14 16 40 60 80 100 14 16 40 60 80 100 13 13.5 14 40 60 80 100 14 16 OSNR [dB] 40 60 80 100 12 14 16 40 60 80 100 14 16 OSNR [dB] First WDM channel First WDM channel Center WDM channel Center WDM channel Last WDM channel Last WDM channel (a) 40 60 80 100 12 14 16 40 60 80 100 14 16 First WDM channel First WDM channel 17 14 40 60 80 100 13 13.5 14 40 60 80 100 14 16 OSNR [dB] Center WDM channel Center WDM channel 40 60 80 100 12 14 16 40 60 80 100 14 16 OSNR [dB] Last WDM channel Last WDM channel (a) (a) 17 18 19 17 18 19 40 60 80 100 13 15 17 40 60 80 100 13 15 17 First WDM channel First WDM channel 17 17 -5 -2.5 0 2.5 5 12 13 14 15 16 17 -5 -2.5 0 2.5 5 12 13 14 15 16 17 (b) 40 60 80 100 13 15 17 40 60 80 100 13 15 17 Center WDM channel Center WDM channel 40 60 80 100 13 15 17 40 60 80 100 13 15 17 Last WDM channel Last WDM channel (b) (b) 40 60 80 100 12 14 16 40 60 80 100 13 15 17 40 60 80 100 12 14 16 40 60 80 100 13 15 17 40 60 80 100 13 15 17 40 60 80 100 13 15 17 (c) First WDM channel First WDM channel Center WDM channel Center WDM channel Last WDM channel Last WDM channel -5 0 5 10 11 12 13 14 15 16 17 -5 0 5 10 11 12 13 14 15 16 17 (c) 40 60 80 100 13 15 17 40 60 80 100 13 15 17 (c) Last WDM channel Last WDM channel Fig. 5. OSNR for the red lightpath as a function of the CUTs frequencies, for ✏occupancy = 100%, ✏network = 95%, NCUT = 20% · Nch, Sr = 0 and the launch powers of (a) -3 dBm, (b) 0 dBm and (c) 3 dBm. C. OSNR for different network utilizations 40 60 80 100 12 14 16 40 60 80 100 14 16 40 60 80 100 13 13.5 14 40 60 80 100 14 16 OSNR [dB] 40 60 80 100 12 14 16 40 60 80 100 14 16 OSNR [dB] First WDM channel First WDM channel Center WDM channel Center WDM channel Last WDM channel Last WDM channel (a) 40 60 80 100 13 15 17 40 60 80 100 13 15 17 40 60 80 100 13 15 17 40 60 80 100 13 15 17 40 60 80 100 13 15 17 40 60 80 100 13 15 17 First WDM channel Center WDM channel Last WDM channel Last WDM channel Center WDM channel First WDM channel (b) 40 60 80 100 12 14 16 40 60 80 100 13 15 17 40 60 80 100 12 14 16 40 60 80 100 13 15 17 40 60 80 100 13 15 17 40 60 80 100 13 15 17 (c) First WDM channel First WDM channel Center WDM channel Center WDM channel Last WDM channel Last WDM channel ig. 6. OSNR for the red lightpath as a function of the network utilization, for a) PCUT = −3 dBm, (b) PCUT = 0 dBm and (c) PCUT = 3 dBm. Only esults for the ﬁrst (-5.7 THz), center and last (5.7 THz) WDM channels are resented. The channel spacings ∆f = 50 GHz and 100 GHz are used. Lines with circles: closed-form GGN-model (incoherent). Dashed lines: closed-form GGN-model (coherent). Lines with asterisks: asymptotic GN-model. C. OSNR for different network utilizations 40 60 80 100 12 14 16 40 60 80 100 14 16 40 60 80 100 13 13.5 14 40 60 80 100 14 16 OSNR [dB] 40 60 80 100 12 14 16 40 60 80 100 14 16 OSNR [dB] First WDM channel First WDM channel Center WDM channel Center WDM channel Last WDM channel Last WDM channel (a) 40 60 80 100 13 15 17 40 60 80 100 13 15 17 40 60 80 100 13 15 17 40 60 80 100 13 15 17 40 60 80 100 13 15 17 40 60 80 100 13 15 17 First WDM channel Center WDM channel Last WDM channel Last WDM channel Center WDM channel First WDM channel (b) 40 60 80 100 12 14 16 40 60 80 100 13 15 17 40 60 80 100 12 14 16 40 60 80 100 13 15 17 40 60 80 100 13 15 17 40 60 80 100 13 15 17 (c) First WDM channel First WDM channel Center WDM channel Center WDM channel Last WDM channel Last WDM channel Fig. 6. OSNR for the red lightpath as a function of the network utilization, for (a) PCUT = −3 dBm, (b) PCUT = 0 dBm and (c) PCUT = 3 dBm. Only results for the ﬁrst (-5.7 THz), center and last (5.7 THz) WDM channels are presented. The channel spacings ∆f = 50 GHz and 100 GHz are used. Lines with circles: closed-form GGN-model (incoherent). Dashed lines: closed-form GGN-model (coherent). Lines with asterisks: asymptotic GN-model. REFERENCES [1] P. Poggiolini and Y. Jiang, “Recent advances in the modeling of the impact of nonlinear ﬁber propagation effects on uncompensated coherent transmission systems,” J. Lightw. Technol., vol. 35, no. 3, pp. 458–480, Feb. 2017. The good agreement between the models predictions for PCUT = −3 dBm is due to the lower power transfer of only 3.1 dB, which occurs when ✏network = 100% and ∆f = 50 GHz. For PCUT = 0 dBm, the overall power transfer increases, leading to higher differences between the OSNR es- timates as the network utilization increases, as can be observed in Fig. 6 (b). The OSNR differences using the incoherent GGN-model and the asymptotic GN-model reach 0.3 dB in the edge channels, when ✏network = 100% and ∆⇢(L) ⇡6.2 dB. For the coherent GGN-model and for ∆f = 50 GHz, the OSNR differences in the ﬁrst channel relative to the asymptotic GN-model are about 0.3 dB and 0.5 dB for ✏network = 50% and ✏network = 100%, respectively. For ∆f = 100 GHz, this discrepancy is always below about 0.3 dB. For the last channel, the OSNR estimates are very similar to the estimates of the asymptotic GN-model for all ✏network considered. For PCUT = 3 dBm, the OSNR predictions from the GN models follow a similar behavior as with PCUT = 0 dBm, but due to the higher power transfer, the OSNR prediction deviations between the models for the edge channels become larger, for all network utilizations considered. For instance, for the ﬁrst channel with ∆f = 50 GHz, the asymptotic GN-model overestimates the OSNR at least by about 1 dB and 2 dB for ✏network = 50% and ✏network = 100%, respectively. For ∆f = 100 GHz, a maximum OSNR difference of about 1.6 dB occurs for the ﬁrst WDM channel and ✏network = 100%. pp [2] A. Ferrari et al., “GNPy: an open source application for physical layer aware open optical networks,” J. Opt. Commun. Netw., vol. 12, no. 6, pp. C31–C40, Jun. 2020. [3] D. Semrau, R. I. Killey, and P. Bayvel, “The Gaussian noise model in the presence of inter-channel stimulated Raman scattering,” J. Lightw. Technol., vol. 36, no. 14, pp. 3046–3055, Jul. 2018. arXiv: 1801.02460. [4] M. Cantono, J. L. Auge, and V. Curri, “Modelling the impact of SRS on NLI generation in commercial equipment: an experimental investigation,” in Proc. Opt. Fiber Commun. Conf. REFERENCES Expo., San Diego, CA, USA, Mar. 2018, paper M1D.2. [5] D. Semrau, R. I. Killey, and P. Bayvel, “A closed-form approximation of the gaussian noise model in the presence of inter-channel stimulated raman scattering,” J. Lightw. Technol., vol. 37, no. 9, pp. 1924–1936, May 2019. [6] P. Poggiolini, “The GN model of non-linear propagation in uncompen- sated coherent optical systems,” J. Lightw. Technol., vol. 30, no. 24, pp. 3857–3879, Dec. 2012. pp [7] X. Chen and W. Shieh, “Closed-form expressions for nonlinear transmission performance of densely spaced coherent optical OFDM systems,” Opt. Express, vol. 18, no. 18, pp. 19 039–19 054, Aug. 2010. [8] S. J. Savory, “Approximations for the nonlinear self-channel interfer- ence of channels with rectangular spectra,” IEEE Photon. Technol. Lett., vol. 25, no. 10, pp. 961–964, May 2013. [9] P. Johannisson and E. Agrell, “Modeling of nonlinear signal distortion in ﬁber-optic networks,” J. Lightw. Technol., vol. 32, no. 23, pp. 4544– 4552, Dec. 2014. [10] D. Semrau, R. Killey, and P. Bayvel, “Achievable rate degradation of ultra-wideband coherent ﬁber communication systems due to stimu- lated raman scattering,” Opt. Express, vol. 25, no. 12, pp. 13 024– 13 034, Jun. 2017. [11] V. Anagnostopoulos, C. T. Politi, C. Matrakidis, and A. Stavdas, “Physical layer impairment aware wavelength routing algorithms based on analytically calculated constraints,” Opt. Commun., vol. 270, no. 2, pp. 247–254, Feb. 2007. pp [12] S. Gringeri, B. Basch, V. Shukla, R. Egorov, and T. J. Xia, “Flexible architectures for optical transport nodes and networks,” IEEE Commun. Mag., vol. 48, no. 7, pp. 40–50, Jul. 2010. IV. CONCLUSION In this work, we studied the impact of the network utiliza- tion and C+L band occupancy on the OSNR estimation using closed-form GN models. Comparing the performances of the asymptotic and GGN models in optimum OSNR conditions and for the full C+L band occupancy, the maximum OSNR difference using the asymptotic GN-model is only 0.7 dB com- pared to the optimum OSNR obtained with the GGN-model. Hence, we have shown that, at optimum launch power and for applications that do not have high accuracy requirements, the asymptotic GN-model can represent a viable alternative to estimate the NLI in C+L band transmissions systems for power transfers below 6 dB. Furthermore, is is expected that most optical networks are working near this low nonlinear effects regime [3]. However, for higher launch powers that lead to higher SRS power transfer, the OSNR discrepancies can increase up to 3 dB. Therefore, the use of the asymptotic GN-model is not recommended in this case. Moreover, in this work, we also showed that higher network utilizations lead to an increasing impact of the SRS, which increases the NLI impact and decreases the overall network performance. g pp [13] A. Mitra et al., “Effect of reduced link margins on C+L band elastic optical networks,” J. Opt. Commun. Netw., vol. 11, no. 10, pp. C86– C93, Oct. 2019. [14] A. Mitra et al., “Effect of channel launch power on ﬁll margin in C+L band elastic optical networks,” J. Lightw. Technol., vol. 38, no. 5, pp. 1032–1040, Mar. 2020. [15] N. Sambo et al., “Provisioning in multi-band optical networks,” J. Lightw. Technol., vol. 38, no. 9, pp. 2598–2605, May 2020. [16] B. Correia, R. Sadeghi, E. Virgillito, A. Napoli, N. Costa, J. Pedro, and V. Curri, “Power control strategies and network performance assessment for C+ L+ S multiband optical transport,” J. Opt. Commun. Netw., vol. 13, no. 7, pp. 147–157, Jul. 2021. [17] J. Pedro, “Designing transparent ﬂexible-grid optical networks for maximum spectral efﬁciency [invited],” IEEE/OSA J. Opt. Commun. Netw., vol. 9, no. 4, pp. C35–C44, Apr. 2017. [18] D. Sequeira, L. Cancela, and J. Rebola, “CDC ROADM design tradeoffs due to physical layer impairments in optical networks,” Opt. Fiber Technol., vol. 62, p. 102 461, Mar. 2021. [19] A. Ferrari et al., “Assessment on the achievable throughput of multi- band ITU-T G.652.D ﬁber transmission systems,” J. Lightw. Technol., vol. 38, no. 16, pp. 4279–4291, Aug. 2020. C. OSNR for different network utilizations Channel spacings ∆f = 50 GHz and ∆f = 100 GHz are considered. Circles: closed-form GGN-model (incoherent). Dashed lines: closed-form GGN-model (coherent). Asterisks: asymptotic GN-model. Points: ∆⇢(L). Fig. 6. OSNR for the red lightpath as a function of the network utilization, for (a) PCUT = −3 dBm, (b) PCUT = 0 dBm and (c) PCUT = 3 dBm. Only results for the ﬁrst (-5.7 THz), center and last (5.7 THz) WDM channels are presented. The channel spacings ∆f = 50 GHz and 100 GHz are used. Lines with circles: closed-form GGN-model (incoherent). Dashed lines: closed-form GGN-model (coherent). Lines with asterisks: asymptotic GN-model. PCUT = −3 dBm, in Fig. 6(b) for PCUT = 0 dBm and in Fig. 6(c) for PCUT = 3 dBm. For convenience, it is assumed that the lowest and highest frequency channels correspond to the ﬁrst and last channels of the WDM signal, respectively. PCUT = −3 dBm, in Fig. 6(b) for PCUT = 0 dBm and in Fig. 6(c) for PCUT = 3 dBm. For convenience, it is assumed that the lowest and highest frequency channels correspond to the ﬁrst and last channels of the WDM signal, respectively. REFERENCES ACKNOWLEDGMENT This work was supported under the project of Instituto de Telecomunicac¸˜oes UIDB/50008/2020.
https://openalex.org/W4226376166	https://dl.acm.org/doi/pdf/10.1145/3522676	English	null	Smart mobility	Crossroads	2,022	cc-by	4,073	tremendous benefits, it also poses several challenges. Some of the chal lenges are obvious such as pollution (22% of CO2 emissions in 2019) and safety (40,732 deaths in 2020 [5]). Oth er challenges are not so obvious such as equitable access, where equity may be limited due to income, race, or dis abilities. These challenges provide op portunities for academics, the public feature feature Smart Mobility: Challenges and opportunities for the next generation of transportation Transportation is an essential component of living in smart cities, but what would mobility in smart cities look like? This article is an overview of the opportunities and challenges presented by smart mobility. By Arnav Choudhry DOI: 10.1145/3522676 OPEN ACCESS By Arnav Choudhry DOI: 10.1145/3522676 OPEN ACCESS I n 2020, American households spent $1.2 trillion on transportation, making it the fourth- largest spending category after healthcare, housing, and food [1]. An average American is estimated to have spent about 286.7 hours traveling in 2020, or about 438 hours in 2019 [2]. People use transportation systems to access better employment or education opportunities, get medical care, for recreational purposes, to obtain daily goods, or to meet friends and family. Another major aspect of transportation networks is moving freight. Freight transportation affords many benefits by moving raw materials, intermediate goods, I Examples of indirect impacts are the economic activity spurred through the availability of transportation ($19 trillion worth goods moved using U.S. transportation networks in 2017 [4]) and societal benefits such as access to better jobs. Hence, the importance of transportation in everybody’s lives seems self-evident. and finished products from where they are in abundance/produced to places where there is demand. The vital role played by transportation can be seen both directly and indirectly. In the United States alone, direct impacts in clude economic value associated with the jobs (14.8 million jobs in 2019) and fixed assets ($8.3 trillion in 2018) asso ciated directly with transportation [3]. Where transportation provides XRDS • SPRING 2022 • VOL .28 • NO.3 14 15 XRDS • SPRING 2022 • VOL .28 • NO.3 15 XRDS • SPRING 2022 • VOL .28 • NO.3 feature Transportation has its roots in Latin words transportare, which liter ally translates to carry across. Con temporary use of transportation is related primarily to moving people and goods. Currently, this movement is accomplished through modes pow ered by humans or animals, air trans port, land transport, water transport, pipelines, or even space transport to name a few. Each mode of transpor tation has its own requirements for fixed infrastructure, operation, and maintenance and by using a different technological solution it provides dif ferent and very specific benefits. For example, a remote mountainous area may be inaccessible to land or water transport and may require a helicop ter or drone to meet local needs. While we want to develop technology that accomplishes our goal of improving transportation, it is worthwhile to re flect on what improved transportation might look like. Consider that due to the differences in required resources for each mode of transportation and constraints supplied by the society served by a transportation system, dif ferent modes of transportation may be optimal for different communities. This leads to differences in cost, ac cessibility (due to availability), safety, and reliability of use. This then leads to the idea that any future transporta tion solution must not only enable the movement of goods and people but also be able to do so while being cost- effective, equitable, safe, reliable/ resilient, and sustainable. Although these guiding principles are enough if we consider transportation in the tra ditional context, we may want to con sider some other principles as tech nology becomes more ubiquitous. Privacy is an example of such an addi tional consideration as city transport systems become smarter. Another ex ample is making resilient systems not just safe for one in 50-year floods but secure from cyberattacks. To reflect the complexity of what an advanced (and smart) transporta i h ld d h teractions happen, and their effects, are something that changes with the availability of technology. But specifi cally in the context of cities, what it means to be smart has changed over time. Earlier, smart cities were just communities that were using a nifty gadget in an existing ecosystem. How ever, more recently, the idea of a smart city is about developing and using technology with a specific outcome in mind. The key here is the develop ment of technology. 15 XRDS • SPRING 2022 • VOL .28 • NO.3 In earlier cycles, the technology was developed inde pendently, and then applications for them were identified in the wild. Now there is a push to identify end goals first, and then develop technologies to accomplish those goals. Better trans portation is one of those goals that we hope to achieve through technology. sector, and the private sector to create solutions and direct change that will impact billions of lives. There is a lot of interest in smart transportation from self-driving cars to drones performing Amazon deliv eries. More recently the passing of the “Infrastructure Investment and Jobs Act” last November provides ap proximately $1.2 trillion of funding to improve electric, internet, and transportation infrastructure with a significant portion of the spending allocated toward making safe policy and constructing infrastructure for the next generation of transportation enabling smart cities in the U.S. UNLOCKING OPPORTUNITIES Since smart mobility exists in a symbi otic relationship with physical, and in creasingly digital, infrastructure, we can group the opportunities provided by smart mobility based on the stages of life within an infrastructure sys tem. A typical infrastructure life cycle broadly consists of four stages: plan ning, building, operation, and end of life. Let us explore some of these op portunities through examples. g g g Use stage. This stage of an infra structure system is the one most peo ple interact with. There are opportu nities afforded by smart technologies in this stage that don’t just help users use the system better, but also help with maintaining the infrastructure. An example of this is using optical fiber cables embedded in highways to detect potholes and traffic charac teristics like congestion in real-time. This is accomplished using distrib uted acoustic sensing and has been in use to monitor sensitive installations for a long time [8]. Combined with ad vanced roadway material technology, such as self-healing pavements [9, 10], the infrastructure repairs proactively rather than reactively. Smart technol ogy can also similarly enable proac tive monitoring for other pieces of infrastructure, such as bridges, using autonomous drones [11]. Planning stage. This stage in the infrastructure life cycle is roughly associated with assessing needs, fea sibility, impact assessment, and solu tion identification. For example, if it were possible to pinpoint the routes traveled by vehicles traversing the ex isting network, it would be possible to figure out the exact demand for pas senger travel. We can then plan for links to add to the current transporta tion network to accommodate trans portation demands. This can replace current traffic travel pattern studies that make several assumptions that may not hold true over time. This can be especially useful in freight infra structure planning. By understanding the freight weight, volume, and value demand, we can locate the best facili ties for collaborative use. This could be especially important if we wanted to include newer methods of urban freight delivery such as drones, self- pickup hubs, crowd workers, or side Communication between differ ent participants in the transporta tion system can be enabled by utiliz ing smart technologies. The ability to communicate brings about a host of benefits. Communication between infrastructure and users can enable more flexible use of infrastructure by better informing users. SMART MOBILITY Smart mobility is the portmanteau used to refer to transportation in the context of smart cities. To have a good understanding of what smart mobility is, we need to take a closer look at what smart and mobility mean. “Smart” is an adjective that has gained increas ing use to describe complex systems of the future. Broadly, a smart system is a technology-driven system that in teracts with end-users and/or other smart systems. It uses some com bination of hardware (e.g., sensors) and software (e.g., signal processing, machine learning) to accomplish its intelligent behavior. How those in We must ensure everybody has access to safe, convenient, cheap, and reliable transportation. This is all enabled by technology. To reflect the complexity of what an advanced (and smart) transporta tion system should do, the term mo bility is often used over transporta tion. Although the two terms sound similar, there is a slight difference in what they refer to. Transportation it self is simple in that it is the ability to XRDS • SPRING 2022 • VOL .28 • NO.3 16 Devices that are used for providing a seamless or otherwise better service could also be used for surveillance, persecution, and harassment. due to pedestrian use of roads during the day and possibly increasing them at night. Devices that are used for providing a seamless or otherwise better service could also be used for surveillance, persecution, and harassment. move goods or people, and mobility is the ability to access multiple modes of transportation while having a rea sonable quality of score in the guiding principles outlined earlier. So while good transportation is a binary ques tion of whether we can move goods or people, good mobility is a continuous scale of how well we can move goods or people. And therefore, the future of advanced, technology-enabled trans portation is smart mobility. Smart mobility is when we use technology to not just improve the modes of trans portation themselves, but also use technology to improve the factors affecting their quality and availabil ity. Pursuing smart mobility, as is the case with most things, has trade-offs. Both opportunities and challenges manifest in the pursuit of this goal. Construction stage. SMART MOBILITY Beyond the ob vious benefits of using robots, like au tonomous bulldozers to perform the construction tasks to finish construc tion faster [7], there are other benefits that smart technologies can provide during this stage of the infrastructure life cycle. A recurring example of con struction impacting everyday users are the closures that are associated with construction. Road, sidewalk, or parking closures are not impromptu and are typically planned-out weeks or even months in advance. In a smart city, it would be possible to encode these closures during the permitting process itself. Hence when motor ists, pedestrians, and bikers (or even delivery robots) are using the built in frastructure, they can plan routes that avoid going through shut-off areas. walk robots. These types of methods are hyper-local in terms of coverage areas and would need optimally lo cated facilities to provide reasonable coverage and economic advantage. Let us look at another example for using smart technologies in the plan ning phase of transportation systems. Current highway design follows a one size fits all approach by using the “Manual on Uniform Traffic Control Devices for Streets and Highways,” which provides guidance for every aspect of roadway design, including setting speed limits [6]. On the one hand, where this provides certainty to local officials and planners in terms of highway design, not having enough flexibility for local context makes the roads more dangerous. In this case, the speed limit is set by performing a speed survey of motorists using the road and setting the speed limit to the 85th percentile of the results. This results in very high-speed limits for arterial roads leading to most of the fatalities involving car crashes. It also biases the system toward fast-moving motorists who can revise the speed upward through the years by going faster [6]. XRDS • SPRING 2022 • VOL .28 • NO.3 A SMART MOBILITY SYSTEM Smart mobility has a lot of potential to improve the quality of transportation available to users, however like most things it is not a panacea. There exist trade-offs and difficulties in the pur suit of smart mobility. While a particu lar smart mobility solution may help alleviate one shortcoming of the trans portation system, it may introduce challenges based on multiple guiding principles outlined earlier (economic, equitable, safe, reliable/resilient, sus tainable, and privacy preserving). Communication between infra structure and users could also inform users about future road conditions so that there is more time to plan. Future road conditions may also be commu nicated through user-to-user interac tion. In self-driving technology, this can lead to a network effect in the efficacy of the autonomous systems, where you have current road infor mation before your camera or LiDAR system even sees anything. Through infrastructure-to-infrastructure com munication, we may be able to set up green light corridors or open alter nate paths to combat congestion in real-time. ◊◆◊◆◊ Request a media kit and pricing: Request a media kit and pricing: Ilia Rodriguez +1 212-626-0686 acmmediasales@acm.org Communication between infrastructure and users can enable more flexible use of infrastructure by better informing users. are blind to better interact with a city’s transportation network through voice- based navigation. As a bonus, by modi fying the system to also give directions through a screen rather than voice, others can also better navigate the city. by only displaying what is active at a given point of time directly to the user through proximity. It would also make it easier to set up notification alerts such as parking time limits. This would avoid scenarios where cars get towed, which could cost an average of $557 in San Francisco [12] where these types of costs disproportionately af fect lower-income people [13]. It also enables congestion and usage control policies such as dynamic availability and pricing parking spots. UNLOCKING OPPORTUNITIES A case in point is static curb signage, which can get complicated in bigger cities. Confusing signage can be simplified Smart technologies can help in this case through two means: incor porating local contexts while setting the speed limit and having dynamic speed limits. Even preserving the ex isting rule-based nature of designing roads, computer systems can enable storing and applying a higher number of more contextually aware rules while also making the design process faster. In addition, speed limits can be made to be dynamic such as lowering them 17 XRDS • SPRING 2022 • VOL .28 • NO.3 feature Advertise with ACM! Reach the innovators and thought leaders working at the cutting edge of computing and information technology through ACM’s magazines, websites and newsletters. An example of a broader challenge is the requirement for sensors every where. This requirement creates chal lenges on multiple fronts. The first challenge is the increased amount of electricity that sensors would require. To manufacture sensors, there would be a higher number of raw materials needed. Both electricity and raw ma terial requirements may lead to high er economic and/or environmental costs. From a networking perspective, creating and maintaining a massive network of sensors is a big engineer ing challenge. Other engineering challenges include processing and storing the volume of data created. There are also requirements for pri vacy and security. Highly networked systems present additional challeng es from the perspective of cybersecu rity. Privacy becomes an issue since devices that are used for providing a seamless or otherwise better service could also be used for surveillance, persecution, and harassment. Smart technology can also help in crease mobility itself by allowing peo ple to use multiple modes of transport more conveniently. Consider Mobility as a Service (MaaS), which posits that we can plan, pay for, and use intercity or intra-city travel with one applica tion. A MaaS app can be a true door- to-door travel application. Instead of needing to check multiple apps [14], everything happens through one app. Another example is technology en abling precise turn-by-turn naviga tion, which would benefit people who Communication between Another challenge that could hin der mobility gains is access to the in ternet and cashless payment methods such as credit cards. Several current proposals for advancing smart mobili ty, such as MaaS, would require people to own a smartphone with an internet connection. Programs such as tap-to- ride require a linked credit card, but 5% of the U.S. adult population do not use a bank instead they use alternative infrastructure and users can enable more flexible use of infrastructure by better informing users. 18 XRDS • SPRING 2022 • VOL .28 • NO.3 financial services [15]. Most house holds that were unbanked had a fam ily income of less than $25,000, while Black and Hispanic households with out banking services reported higher numbers as well. This is important to note because these are groups who may be more at risk of having low mo bility in the first place [16]. This is an issue of equitable access. [5] Bureau of Transportation Statistics. Transportation Statistics Annual Report 2021. U.S. Department of Transportation, Washington D.C., 2021 DOI: https:// doi.org/10.21949/1524191 Smart mobility has a lot of potential to improve the quality of transportation available to users, however like most things it is not a panacea. [6] Bronin, S. C. and Shill, G. H. Rewriting our nation’s deadly traffic manual. Harv. L. Rev. F.135 (2021), 1; https://harvardlawreview.org/2021/10/rewriting- our-nations-deadly-traffic-manual [7] Phillips, Z. Komatsu’s bulldozer technology eases skilled labor needs. Construction Drive. February 5, 2020; https://www.constructiondive.com/ news/komatsu-to-unveil-semi-autonomous- bulldozers/571751 [8] Owen, A., Duckworth, G., and Worsley, J. OptaSense: Fibre optic distributed acoustic sensing for border monitoring. In Proceedings of the 2012 European Intelligence and Security Informatics Conference. IEEE, 2012, 362–364. The development of smart mobility systems should be done with the diver sity of users in mind. This is especially important since the development of these systems are currently underway. There exist examples of leaving large proportions of the population out while developing the next generation of transportation. For example, as a result of excluding women while de veloping car safety standards, it puts women at a 47% higher risk of severe injuries compared to men, if they’re involved in similar crashes [17]. [9] Tabaković, A. and Schlangen, E. In Self-healing Materials. Advances in Polymer Science, vol 273. Eds. M. Hager, S. van der Zwaag, and U. Schubert. Springer, Cham., 2015, 285–306; https://doi. CONCLUSION [12] Sernoffsky, E. SF gives low-income people a break in city’s steep vehicle towing fees. San Francisco Chronicle. May 15, 2018; https://www.sfchronicle. com/news/article/SF-gives-low-income-people-a- break-in-city-s-12917281.php Smart mobility is the future of tech- enabled transportation. In an ideal future, regardless of the mode of transport, self-driving cars, or tele portation devices, we must ensure everybody has access to safe, con venient, cheap, and reliable trans portation. This is all enabled by technology. We’ve explored some examples showcasing how smart mo bility expands opportunities. We’ve then seen some challenges that pres ent themselves for the adoption of smart mobility as well as challenges that could arise from adopting smart mobility. Transportation is an es sential service and really should be treated as a utility. Although there are challenges, there are many inge nious solutions that already exist and are in further development. Issues of access are also exempli fied by reducing access to fixed infra structure. Consider autonomous cars, for example, it has been shown that under certain conditions, autonomous vehicles may prefer to cruise on roads rather than park [18]. This reduces ac cess to roads for users wanting to trav el. Or consider autonomous sidewalk robots for deliveries, if robots are using sidewalks, it may reduce access for pe destrians including wheelchair users. Although smart mobility may increase access, there might be challenges in making sure accessibility is improved for everybody. [13] The Times Editorial Board. Editorial: Towing can devastate a poor family. But the solution isn’t to throw out the rules. Los Angeles Times. June15, 2019; https://www.latimes.com/opinion/editorials/ la-ed-poverty-towing-bill-20190615-story.html [13] The Times Editorial Board. Editorial: Towing can devastate a poor family. But the solution isn’t to throw out the rules. Los Angeles Times. June15, 2019; https://www.latimes.com/opinion/editorials/ la-ed-poverty-towing-bill-20190615-story.html [14] Hertzfeld, E. Survey: Nearly two-thirds of travelers rely on mobile apps during trips. Hotel Management. Nov. 20, 2018; https://www.hotelmanagement.net/ tech/nearly-two-thirds-travelers-rely-mobile- apps-during-trip [14] Hertzfeld, E. Survey: Nearly two-thirds of travelers rely on mobile apps during trips. Hotel Management. Nov. 20, 2018; https://www.hotelmanagement.net/ tech/nearly-two-thirds-travelers-rely-mobile- apps-during-trip [15] Division of Consumer and Community Affairs (DCCA). Economic Well-Being of U.S. Households in 2020. Board of Governors of the Federal Reserve System. May 2021; https://www.federalreserve.gov/ publications/files/2020-report-economic-well- being-us-households-202105.pdf [15] Division of Consumer and Community Affairs (DCCA). Economic Well-Being of U.S. Households in 2020. Board of Governors of the Federal Reserve System. May 2021; https://www.federalreserve.gov/ publications/files/2020-report-economic-well- being-us-households-202105.pdf [16] Sanchez, T. W., Stolz, R., and Ma, J. S. Communication between org/10.1007/12_2015_335 a people problem and working with multiple people with competing goals who may or may not be wary of the fu ture is a challenge in itself. [10] Norambuena-Contreras, J. and Garcia, A. Self-healing of asphalt mixture by microwave and induction heating. Materials & Design 106 (2016), 404–414. [10] Norambuena-Contreras, J. and Garcia, A. Self-healing of asphalt mixture by microwave and induction heating. Materials & Design 106 (2016), 404–414. [11] Azari, H., O’Shea, D., and Campbell, J. Application of unmanned aerial systems for bridge inspection. Transportation Research Record 2676, 1 (2022), 401–407. This work is licensed under a http:// creativecommons.org/licenses/by/4.0/ Funding agency: U.S. Department of Energy. Grant no: DE-EE0008463, DE-EE0008466 [2] U.S. Bureau of Labor Statistics. American time use survey summary. United States Department of Labor. July 22, 2021; https://www.bls.gov/news. release/atus.nr0.htm [3] Bureau of Transportation Statistics. Transportation Statistics Annual Report 2020. U.S. Department of Transportation, Washington D.C., 2020 DOI: https:// doi.org/10.21949/1520449 CONCLUSION Moving to equity: Addressing inequitable effects of transportation policies on minorities. The Civil Rights Project at Harvard University, Cambridge, 2003. [17] Bose, D., Segui-Gomez, M., and Crandall, J. R. Vulnerability of female drivers involved in motor vehicle crashes: an analysis of US population at risk. American Journal of Public Health 101, 12 (2011), 2368–2373. y y Drones have been touted as the next generation of last-mile movers of people and goods for a while. But one of the reasons they have not been ad opted yet is noise pollution. This is an example of an externality not covered under one of the guiding principles. Although drones are getting quieter, we’re not there yet. Ultimately, inno vators in this space need to realize that technology exists to support a community and improve the qual ity of life. And technological advance ments must go toward that improve ment. Public policy can go a long way in ensuring that technology works for the people. However, creating a policy for something that does not yet exist is a hard problem. Policy is essentially [18] Millard-Ball, A. The autonomous vehicle parking problem. Transport Policy 75 (2019), 99–108. [18] Millard-Ball, A. The autonomous vehicle parking problem. Transport Policy 75 (2019), 99–108. References References [1] Bureau of Transportation Statistics. Transportation economic trends. U.S. Department of Transportation; www.bts.gov/product/ transportation-economic-trends [2] U.S. Bureau of Labor Statistics. American time use survey summary. United States Department of Labor. July 22, 2021; https://www.bls.gov/news. release/atus.nr0.htm [3] Bureau of Transportation Statistics. Transportation Statistics Annual Report 2020. U.S. Department of Transportation, Washington D.C., 2020 DOI: https:// doi.org/10.21949/1520449 [4] Bureau of Transportation Statistics. New freight benchmark shows value, weight of goods moved on America’s transportation network rose over a five- year period. February 26, 2021. U.S. Department of Transportation; https://www.bts.dot.gov/newsroom/ new-freight-benchmark-shows-value-weight-goods- moved-americas-transportation-network-rose [1] Bureau of Transportation Statistics. Transportation economic trends. U.S. Department of Transportation; www.bts.gov/product/ transportation-economic-trends Biography Arnav Choudhry is a fourth-year Ph.D. student in the Advanced Infrastructure Systems program at Carnegie Mellon University. He is a member of the Mobility Data Analytics Center (MAC lab), where he researches ways to improve the quality of the built environment using data and algorithms. [2] U.S. Bureau of Labor Statistics. American time use survey summary. United States Department of Labor. July 22, 2021; https://www.bls.gov/news. release/atus.nr0.htm [3] Bureau of Transportation Statistics. Transportation Statistics Annual Report 2020. U.S. Department of Transportation, Washington D.C., 2020 DOI: https:// doi.org/10.21949/1520449 [3] Bureau of Transportation Statistics. Transportation Statistics Annual Report 2020. U.S. Department of Transportation, Washington D.C., 2020 DOI: https:// doi.org/10.21949/1520449 Funding agency: U.S. Department of Energy. Grant no: DE-EE0008463, DE-EE0008466 [4] Bureau of Transportation Statistics. New freight benchmark shows value, weight of goods moved on America’s transportation network rose over a five- year period. February 26, 2021. U.S. Department of Transportation; https://www.bts.dot.gov/newsroom/ new-freight-benchmark-shows-value-weight-goods- moved-americas-transportation-network-rose Copyright 2022 held by Owner/Author 1528-4972/22/03 $15.00 Copyright 2022 held by Owner/Author 1528-4972/22/03 $15.00 Copyright 2022 held by Owner/Author 1528-4972/22/03 $15.00 XRDS • SPRING 2022 • VOL .28 • NO.3 19
https://openalex.org/W1863687648	https://link.springer.com/content/pdf/10.1007/JHEP04(2015)078.pdf	English	null	Observable effects of general new scalar particles	The Journal of high energy physics/The journal of high energy physics	2,015	cc-by	20,705	Published for SISSA by Springer Received: January 26, 2015 Accepted: February 24, 2015 Published: April 15, 2015 Received: January 26, 2015 Accepted: February 24, 2015 Published: April 15, 2015 Open Access, c⃝The Authors. Article funded by SCOAP3. Observable eﬀects of general new scalar particles JHEP04(2015)078 J. de Blas,a M. Chala,b M. P´erez-Victoriac and J. Santiagoc,d aINFN, Sezione di Roma, Piazzale A. Moro 2, Rome, I-00185 Italy bDESY, Notkestrasse 85, Hamburg, 22607 Germany cDepartamento de F´ısica Te´orica y del Cosmos and CAFPE, Universidad de Granada, Campus de Fuentenueva, Granada, E-18071 Spain dTheory Division, CERN, Geneva 23, CH1211 Switzerland E-mail: Jorge.DeBlasMateo@roma1.infn.it, mikael.chala@desy.de, mpv@ugr.es, jsantiago@ugr.es Notkestrasse 85, Hamburg, 22607 Germany cDepartamento de F´ısica Te´orica y del Cosmos and CAFPE, Universidad de Granada, Campus de Fuentenueva, Granada, E-18071 Spain dTheory Division, CERN, Geneva 23, CH1211 Switzerland E-mail: Jorge.DeBlasMateo@roma1.infn.it, mikael.chala@desy.de, @ j i @ E-mail: Jorge.DeBlasMateo@roma1.infn.it, mikael.chala@desy.de, mpv@ugr.es, jsantiago@ugr.es Abstract: We classify all possible new scalar particles that can have renormalizable linear couplings to Standard Model ﬁelds and therefore be singly produced at colliders. We show that this classiﬁcation exhausts the list of heavy scalar particles that contribute at the tree level to the Standard Model eﬀective Lagrangian to dimension six. We compute this eﬀective Lagrangian for a general scenario with an arbitrary number of new scalar particles and obtain ﬂavor-preserving constraints on their couplings and masses. This completes the tree-level matching of the coeﬃcients of dimension ﬁve and six operators in the eﬀective Lagrangian to arbitrary extensions of the Standard Model. Keywords: Beyond Standard Model, Phenomenological Models ArXiv ePrint: 1412.8480 Open Access, c⃝The Authors. Article funded by SCOAP3. doi:10.1007/JHEP04(2015)078 Contents 1 Introduction 1 2 Standard model extensions with extra scalar ﬁelds 3 3 The eﬀective Lagrangian for heavy new scalar particles 5 4 Observable eﬀects of new scalars 8 5 Precision constraints on new scalars 11 6 Scalar extensions with other new particles 15 7 Conclusions 18 A Basis of dimension-six operators 20 B Operator coeﬃcients in the eﬀective Lagrangian 23 Contents 1 Introduction 1 2 Standard model extensions with extra scalar ﬁelds 3 3 The eﬀective Lagrangian for heavy new scalar particles 5 4 Observable eﬀects of new scalars 8 5 Precision constraints on new scalars 11 6 Scalar extensions with other new particles 15 7 Conclusions 18 A Basis of dimension-six operators 20 B Operator coeﬃcients in the eﬀective Lagrangian 23 JHEP04(2015)078 A Basis of dimension-six operators B Operator coeﬃcients in the eﬀective Lagrangian 1 Introduction The discovery of the Higgs boson at the Large Hadron Collider (LHC) [1, 2] has opened a new era in particle physics in which we have, for the ﬁrst time, direct access to the electroweak symmetry breaking sector. The naturalness problem associated to this scalar sector is still one of the main reasons to expect new physics beyond the Standard Model (SM) at the TeV scale, and thus accessible to the LHC. However, the lack of signiﬁcant deviations from the SM predictions after Run 1 suggests that, even if really present, the new particles may be too heavy to be produced on-shell, so that only their indirect eﬀects can be observed at the LHC —although it is certainly possible that they are just above the current reach and can still be directly produced at the higher energies of Run 2. In such a case, the natural language to parameterize the expected eﬀects of new physics is that of eﬀective theories. Eﬀective Lagrangians provide a model-independent description of the eﬀects of new particles at energies much smaller than their masses. Hence, they are the perfect tool to study any new physics that lies beyond the reach of our current experiments. A single higher-dimensional gauge-invariant operator contributes in general to several diﬀerent cou- plings between SM particles after electroweak symmetry breaking, leading to non-trivial correlations between diﬀerent observables [3]. These correlations can be tested experimen- tally or, alternatively, be used to predict the size of the expected deviations with respect to the SM predictions [4]. In addition, higher-dimensional operators are often generated – 1 – by a smaller number of couplings in speciﬁc ultraviolet completions and therefore the co- eﬃcients of diﬀerent operators are correlated as well. It could be argued that these latter correlations are model-dependent, defeating the very purpose of the eﬀective Lagrangian. However, model independence can be recovered if a complete dictionary between ultra- violet completions and eﬀective operators is built. Such a dictionary would provide a comprehensive classiﬁcation of new physics with potentially observable eﬀects at the LHC. This classiﬁcation can hence guide experimental searches to ensure that no viable option is missed at the LHC, and help to identify the origin of possible deviations from the SM predictions. Encouraged by the recent experimental observation of a Higgs sector, in this article we focus on new scalar particles. 1 Introduction We ﬁrst classify all the possible new scalars that can cou- ple linearly, with renormalizable interactions, to SM ﬁelds, and write their most general phenomenologically-relevant interactions. These particles can be singly produced at col- liders with sizable couplings and have therefore the most promising discovery potential at the LHC. The corresponding general interactions of new quarks, leptons and vector bosons have already been given in refs. [5, 6] and [7], respectively. Our results here thus complete the description of arbitrary new particles with linear gauge-invariant renormalizable cou- plings to the SM ﬁelds. This provides an extremely useful scheme for a model-independent interpretation of LHC searches. Particular models correspond to speciﬁc choices of the general couplings and masses in this set-up. Therefore, once the experimental results are written in terms of the general parameters it is straightforward to derive consequences for any model of choice. JHEP04(2015)078 It turns out that the same classiﬁcation also covers all possible new scalar particles that contribute, once integrated out at the tree level, to the SM eﬀective Lagrangian of dimension ﬁve and six. We perform this integration explicitly. In predictive models loop contributions are suppressed, so it is expected that the leading observable consequences of new heavy scalars are those generated at tree level. Except for this assumption, our results are completely general and can be used for an arbitrary extension of the SM with heavy scalars, independently of their amount and quantum numbers. Furthermore, extra heavy particles with diﬀerent spins do not mix with the heavy scalars in their contribution to the dimension-six eﬀective Lagrangian at tree level. Therefore, together with the eﬀective Lagrangians generated by the most general extension of the SM with new quarks [5], leptons [6] and vector bosons [7], our results complete the tree-level dictionary between any model of new physics and the dimension-six SM eﬀective Lagrangian. This dictionary can be used to trivially obtain the observable implications of an arbitrary model of new physics at energies much smaller than the masses of the new particles involved. It is also a powerful tool to investigate correlations or cancellations predicted among observables in speciﬁc extensions of the SM. Finally, it provides a rationale for the calculation of the constraints on the coeﬃcients of the SM eﬀective Lagrangian, as one can put constraints on the sources of the eﬀective operators in a correlated way rather than on arbitrary combinations of them. 1 Introduction The article is organized as follows. We classify in section 2 all possible new scalars that can have linear interactions with SM ﬁelds. We describe in section 3 how to compute – 2 – the eﬀective Lagrangian that results from the tree-level integration of an arbitrary number and type of new heavy scalars, and show that the previous classiﬁcation exhausts the list of new spin-0 particles that can contribute to the dimension-six eﬀective Lagrangian at tree level. In section 4 we discuss what eﬀects induced by the heavy scalars are observable at this order. Some applications of the eﬀective Lagrangian are presented in section 5. Finally, we comment on the interplay between new scalars and particles with diﬀerent spin in section 6, and conclude in section 7. The basis of dimension-six operators and all the relevant scalar interactions with the resulting eﬀective Lagrangians, are given in appendix A and B, respectively. JHEP04(2015)078 1We use capital indices A, B, C as color indices, whereas lower case indices a, b = 1, 2, 3 tag ﬁelds in the adjoint of SU(2)L. Latin indices i, j, k are used to label diﬀerent generations. 2 Standard model extensions with extra scalar ﬁelds We consider a general renormalizable theory for extra scalars and SM ﬁelds, invariant under Lorentz transformations and under the complete SU(3)c ⊗SU(2)L ⊗U(1)Y gauge group. The new scalar ﬁelds will come in complete representations of this group, which can be decomposed into their irreducible components σ. Non-renormalizable interactions are also possible in principle, but in a predictive theory they will be suppressed by a scale larger than the mass of the extra scalars. Here we concentrate on the leading eﬀects, which are generically described by operators of dimension four at most. The most general Lagrangian for such an extension of the SM can be written as L = LSM + Lσ + Lint, (2.1) (2.1) where LSM is the SM Lagrangian, Lσ contains the kinetic (with covariant derivatives) and mass terms for the new scalars and Lint describes the non-gauge interactions of the extra scalars. where LSM is the SM Lagrangian, Lσ contains the kinetic (with covariant derivatives) and mass terms for the new scalars and Lint describes the non-gauge interactions of the extra scalars. The SM Lagrangian reads, in standard notation1 The SM Lagrangian reads, in standard notation1 LSM = −1 4GA µνGA µν −1 4W a µνW a µν −1 4BµνBµν + li L i D li L + qi L i D qi L + ei R i D ei R + ui R i D ui R + di R i D di R + (Dµφ)† Dµφ −U (φ) − ye ii li Lφei R + yd ii qi Lφdi R + V † ijyu jj qi L ˜φuj R + h.c. . (2.2) (2.2) In order to ﬁx the meaning of ﬂavor indices, we have chosen a basis in which the Yukawa interactions for the charged-leptons and down-type quarks are diagonal. As usual, ˜φ = iσ2φ∗denotes the iso-doublet of hypercharge −1/2, constructed with the Higgs doublet φ. The Higgs scalar potential is In order to ﬁx the meaning of ﬂavor indices, we have chosen a basis in which the Yukawa interactions for the charged-leptons and down-type quarks are diagonal. As usual, ˜φ = iσ2φ∗denotes the iso-doublet of hypercharge −1/2, constructed with the Higgs doublet φ. The Higgs scalar potential is U (φ) = −µ2 φ \|φ\|2 + λφ \|φ\|4 . 2 Standard model extensions with extra scalar ﬁelds (2.3) (2.3) The Lagrangian Lσ contains the gauge-invariant kinetic and mass terms for the new scalars: The Lagrangian Lσ contains the gauge-invariant kinetic and mass terms for the new scalars: Lσ = X σ ησ h (Dµσ)† Dµσ −M2 σσ†σ i , (2.4) (2.4) 1We use capital indices A, B, C as color indices, whereas lower case indices a, b = 1, 2, 3 tag ﬁelds in the adjoint of SU(2)L. Latin indices i, j, k are used to label diﬀerent generations. – 3 – where ησ = 1, 1 2, for complex and real scalars, respectively. Note that we are working in a basis with canonical kinetic terms and diagonal mass matrices for all scalar ﬁelds, including the Higgs doublet. To match models written in a diﬀerent basis, the diagonalization must be performed prior to using our formulas.2 Finally, Lint contains the renormalizable interactions of the extra scalars (among them- selves and with the SM ﬁelds), except for the gauge interactions, which are already included in Lσ. We can distinguish between interactions with fermions and purely scalar interac- tions: Lint = −V ({σ}, φ) − X σ ησ σ†Jσ + h.c. . (2.5) (2.5) JHEP04(2015)078 The chirality-ﬂipping fermionic currents Jσ ∼ψL ⊗ξR or Jσ ∼ξR ⊗ψL couple to one scalar ﬁeld with a dimensionless coupling. The potential V contains scalar interactions between the new particles and, possibly, the SM Higgs ﬁelds. Together with the mass terms and U, it forms the total scalar potential. As explained above, V does not include mass mixing terms, since we work in a basis with diagonal quadratic terms. Thus, each term in V contains either three or four scalars, with couplings of dimension one or zero, respectively. Furthermore, all the terms in V have at least one σ ﬁeld. The new scalars with linear interactions in Lint can be singly produced at tree level in colliders.3 By “linear interactions” we mean that Lint contains some non-vanishing term that is the product of SM ﬁelds and a single power of the given extra scalar ﬁeld, with no other extra scalars. Gauge invariance and the particle content of the SM strongly constraints the quantum numbers of new scalars that can have such linear interactions. 2Furthermore, we assume that there are no tadpole operators in the electroweak symmetric phase. This entails no loss of generality since these tadpoles, which are only possible for new singlet scalars, can always be eliminated by a shift of the singlet ﬁeld(s). The only eﬀect of this shift is a redeﬁnition of the parameters that we write explicitly. y 3Other scalar ﬁelds can also be singly produced if they mix with these after electroweak symmetry breaking, but if they are heavy the production rate will be suppressed by the square of small mixings. 2 Standard model extensions with extra scalar ﬁelds As we show in the next section, only the scalars in representations that allow for these linear couplings can contribute at tree level, to order 1/M 2 σ in the heavy mass limit, to observable processes with SM particles in the initial and ﬁnal states. We list in table 1 all the irreducible representations of scalars with linear interactions of this kind. The scalar representations S, Ξ0, Θ1 and Θ3 do not couple to the SM fermions. In appendix B we write the interactions in Lint explicitly, including only those that have an impact in the SM eﬀective Lagrangian at dimension six. These include the mentioned linear interactions of the scalar ﬁelds in the irreducible representations of table 1, as well as terms involving two or three scalar ﬁelds in the same set of representations. Note that, in particular, new scalars can always couple with the SM through a Higgs-portal type of coupling σ†σ φ†φ [8]. However, only the ones in our list with the right quantum numbers to allow for linear couplings to the SM will induce dimension-six operators when integrated at tree level. Moreover, as we explain in the next section, interactions involving two or more new scalars are relevant for the calculation of the eﬀective Lagrangian to dimension six only in those – 4 – Colorless S S1 S2 ϕ Ξ0 Ξ1 Θ1 Θ3 Scalars Irrep (1, 1)0 (1, 1)1 (1, 1)2 (1, 2) 1 2 (1, 3)0 (1, 3)1 (1, 4) 1 2 (1, 4) 3 2 Colored ω1 ω2 ω4 Π1 Π7 ζ Scalars Irrep (3, 1)−1 3 (3, 1) 2 3 (3, 1)−4 3 (3, 2) 1 6 (3, 2) 7 6 (3, 3)−1 3 Colored Ω1 Ω2 Ω4 Υ Φ Scalars Irrep (6, 1) 1 3 (6, 1)−2 3 (6, 1) 4 3 (6, 3) 1 3 (8, 2) 1 2 Table 1. Scalar bosons with linear renormalizable interactions with the SM ﬁelds. The quantum numbers (Rc, RL)Y denote the irreducible representation (Irrep) Rc under SU(3)c, RL under SU(2)L and the hypercharge Y , respectively. The hypercharge is normalized such that the electric charge is Q = Y + T3. Looking only at the quantum numbers, some readers might miss in this list a scalar particle transforming as a (1, 2) 3 2 and coupling linearly to three Higgs doublets. 2 Standard model extensions with extra scalar ﬁelds However, the corresponding operator actually vanishes, since it involves an antisymmetric combination of the Higgs ﬁelds. Colorless S S1 S2 ϕ Ξ0 Ξ1 Θ1 Θ3 Scalars Irrep (1, 1)0 (1, 1)1 (1, 1)2 (1, 2) 1 2 (1, 3)0 (1, 3)1 (1, 4) 1 2 (1, 4) 3 2 Colored ω1 ω2 ω4 Π1 Π7 ζ Scalars Irrep (3, 1)−1 3 (3, 1) 2 3 (3, 1)−4 3 (3, 2) 1 6 (3, 2) 7 6 (3, 3)−1 3 Colored Ω1 Ω2 Ω4 Υ Φ Scalars Irrep (6, 1) 1 3 (6, 1)−2 3 (6, 1) 4 3 (6, 3) 1 3 (8, 2) 1 2 Table 1. Scalar bosons with linear renormalizable interactions with the SM ﬁelds. The quantum numbers (Rc, RL)Y denote the irreducible representation (Irrep) Rc under SU(3)c, RL under SU(2)L and the hypercharge Y , respectively. The hypercharge is normalized such that the electric charge JHEP04(2015)078 Table 1. Scalar bosons with linear renormalizable interactions with the SM ﬁelds. The quantum numbers (Rc, RL)Y denote the irreducible representation (Irrep) Rc under SU(3)c, RL under SU(2)L and the hypercharge Y , respectively. The hypercharge is normalized such that the electric charge is Q = Y + T3. Looking only at the quantum numbers, some readers might miss in this list a scalar particle transforming as a (1, 2) 3 2 and coupling linearly to three Higgs doublets. However, the corresponding operator actually vanishes, since it involves an antisymmetric combination of the Higgs ﬁelds. cases where the scalar potential contains trilinear couplings of new scalars to two Higgs ﬁelds (see tables 9, 13, 14 and 28 in appendix B). Extensions of the SM with new scalar ﬁelds have been extensively considered in the past and partial classiﬁcations have been presented, with special emphasis on their eﬀect on baryon and lepton number violation [9–11], collider physics [12, 13], top physics [14] or ﬂavor physics [15]. However the complete classiﬁcation of the scalar ﬁelds that can couple linearly to the SM and the calculation of the tree-level dimension-six eﬀective Lagrangian for the most general extension of the SM with an arbitrary number of new scalar ﬁelds has, to the best of our knowledge, never been presented before. 3 The eﬀective Lagrangian for heavy new scalar particles In this paper we are mainly interested in the eﬀects of heavy new scalars, with masses Mσ that are large in comparison to the Higgs vacuum expectation value (vev) and to the energies probed by the available experimental data. An eﬃcient way to describe these eﬀects at leading order is to integrate the heavy scalars out and expand in 1/Mσ to obtain an eﬀective Lagrangian with gauge-invariant local operators of dimension up to six. This allows for a direct comparison with model-independent analyses and also for a simple combination with other extensions of the SM. In our theory with general extra scalars, the coeﬃcients of these operators will be simple functions of the couplings and masses of the heavy scalars. Once again, we work at the tree level for simplicity and to avoid further suppressions (see [13, 16–18] for examples of one-loop integration of scalar multiplets). In this section we describe the integration of the heavy scalar ﬁelds. In particular, we show that among the inﬁnite scalar representations that can appear in Lint, only the – 5 – ones in table 1 contribute to the eﬀective Lagrangian to dimension six. For clarity, our discussion will be slightly schematic. The complete explicit results, i.e. the dimension-six operators and the values of their coeﬃcients, are collected in the appendices. In the following we assume that, before electroweak symmetry breaking, the scalar mass matrix has only one negative eigenvalue, −µ2 φ. The eigenvector is a (1, 2)1/2 scalar ﬁeld φ, which we identify with the SM scalar doublet. The other eigenvalues, M2 σ, are assumed to be large in comparison with the Higgs vev and with the relevant energies. Finally, we assume that the dimensionful couplings κσ that multiply the dimension-three operators in V are at most of the size of the smaller heavy-scalar mass, \|κσ\| ≲M, with M = Min {\|Mσ\|}. These assumptions are well motivated by the agreement, within the available precision, of experimental results and SM predictions for Higgs observables. They lead to a decoupling scenario and allow us to perform the integration in the electroweak symmetric phase, which is extremely convenient. The occurrence of electroweak symmetry breakdown and all its eﬀects are captured to order 1/M 2 by the eﬀective Lagrangian. This includes the case in which the extra scalars acquire (suppressed) vevs in the Higgs phase [16, 17]. 3 The eﬀective Lagrangian for heavy new scalar particles Wi W j ∆(i) ∆(j) W i k Wi W j W k j ∆(j) ∆(i) ∆(k) W i j W i j W i ∆(i) Wi Wi W ijk Wj Wk ∆(i) ∆(j) ∆(k) W k j Wj W j JHEP04(2015)078 Wk Figure 1. Feynman diagrams contributing to the eﬀective Lagrangian to order 1/M 2. Non- equivalent permutations of the arrow directions shown here should be considered as well. Note that the right-hand sides depend on the ﬁelds σ, so these are not explicit solutions. But the equations can be solved iteratively. The crucial point is that ∆i is O(1/M 2), whereas the couplings in Lint are either dimensionless or O(M), at most. Therefore, the iterative solution of (3.4) starts at O(1/M), and each correction (step of the iteration) is suppressed by at least another 1/M factor. Using (3.3) in the right-hand side of (3.4), we can perform the iteration explicitly. Then we plug the iterative solution in (3.2) and (3.3), neglecting O(1/M 3) terms, to obtain Leﬀ= LSM −∆(i)WiW i −∆(i)∆(j) WijW iW j + W ijWiWj + W i jWiW j −∆(i)∆(j)∆(k) W ijkWiWjWk + WijkW iW jW k + W ij k WiWjW k + W i jkWiW jW k +2W ikW j kWiWj + 2WikW k j W iW j + 4W ikWkjWiW j + W i kW k j WiW j + O(1/M 3). (3.5) (3.5) In each term, ∆(i) acts on the operator with an upper index i. The covariant propagators ∆i are to be expanded in 1/M 2 i , as in (3.1). This result matches precisely the one obtained from the possible Feynman diagrams with heavy-scalar propagators that contribute to or- der 1/M 2, shown in ﬁgure 1. The blobs in this ﬁgure represent the SM operators W i1...im j1...jn with m incoming and n outgoing lines, and the arrowed lines represent the covariant prop- agators ∆i. In each term, ∆(i) acts on the operator with an upper index i. The covariant propagators ∆i are to be expanded in 1/M 2 i , as in (3.1). This result matches precisely the one obtained from the possible Feynman diagrams with heavy-scalar propagators that contribute to or- der 1/M 2, shown in ﬁgure 1. 3 The eﬀective Lagrangian for heavy new scalar particles JHEP04(2015)078 The integration at the tree level can be performed solving the classical equations of motion for the heavy ﬁelds and inserting the solutions into the original Lagrangian. This procedure manifestly preserves the gauge invariance of the original theory. Let σi be each of the scalar ﬁelds. Diﬀerent values of the index i label diﬀerent scalars, in the same or in diﬀerent representations. We will use upper and lower indices for the ﬁelds and their complex conjugates, respectively. The covariant propagator is ∆i = − D2 i + M2 i −1 = −1 M2 i 1 −D2 i M2 i + O(1/M 6), (3.1) (3.1) with Di the covariant derivative acting on σi. The part of the Lagrangian that contains the extra scalars reads (3.2) η(i)σ† i ∆−1 (i) σi + Lint. (3.2) We are using the convention of repeated indices, contracting upper and lower indices. A parenthesis indicates an index that can run but does not count as repeated to induce the running (indices in parenthesis actually refer to the diagonal elements of a diagonal matrix, the propagator). The interaction Lagrangian is a polynomial in σ of degree 4 with no constant term. Hence, it will be of the form Lint = − 4 X m+n=1 σ† j1 · · · σ† jnW j1...jn i1...im σi1 · · · σim, (3.3) (3.3) where m and n vary independently subject to the indicated constraint. The operators W j1...jn i1...im are formed with SM ﬁelds only and obey the obvious hermiticity conditions inher- ited from Lint. In general, they carry reducible representations of the gauge group, but the operators with one index i belong to the irreducible representation of the associated σi or σ† i . The equations of motion for the new scalars read σi = −∆(i) ∂Lint ∂σ† i . (3.4) (3.4) – 6 – – 6 – W i ∆(i) Wi W j ∆(i) ∆(j) Wi Wi W ijk Wj Wk ∆(i) ∆(j) ∆(k) W i k Wi W j W k j ∆(j) ∆(i) ∆(k) W i j Figure 1. Feynman diagrams contributing to the eﬀective Lagrangian to order 1/M 2. Non- equivalent permutations of the arrow directions shown here should be considered as well. 3 The eﬀective Lagrangian for heavy new scalar particles The blobs in this ﬁgure represent the SM operators W i1...im j1...jn with m incoming and n outgoing lines, and the arrowed lines represent the covariant prop- agators ∆i. Observe that in all the terms in (3.5) except the ones in the last line, there is one Wi or W i operator for each propagator index i. Wi and W i arise from terms in Lint with only one heavy scalar (σi or σ† i ), which are what we have called linear interactions. In the – 7 – last line (corresponding to the last Feynman diagram), on the other hand, all terms have a propagator with index k that is not attached to any Wk or W k, but only to operators with two indices. However, these terms are actually the contraction of the two one-index operators f Wk = [Wkj ∆(j)W j] and f W k = [W kj ∆(j)W j] (or variations in the position of the indices). The operators f Wk and f W k are in the same gauge representation as Wk and W k, respectively. Moreover, to allow for an O(1/M 2) contribution, the operators f Wk and f W k must have a dimensionless coeﬃcient and, hence, scaling dimension four. Therefore, the scalars σk (σ† k) associated to f Wk (f W k) must also belong to a representation that allows for renormalizable linear interactions. We conclude that only the scalar ﬁelds in the irreducible representations of table 1 contribute at the tree level to the eﬀective Lagrangian to dimension six. JHEP04(2015)078 Note also that the last two topologies in ﬁgure 1 only contribute to this order when the four blobs contain O(M) dimensionful couplings, which requires that all of them arise from trilinear terms in the scalar potential. In particular, the SM fermions only appear through the diagrams with the ﬁrst and second topologies. Gauge bosons only arise at dimension six from the covariant propagator of the ﬁrst topology, when both blobs represent trilinear interactions in the potential. A complete basis of gauge-invariant operators to dimension six, including the ones generated by heavy scalars, is given in appendix A. The relevant interactions of arbitrary scalars and the detailed results of the integration are collected in appendix B.4 4In some cases, we employ algebraic identities and/or ﬁeld redeﬁnitions to transform the induced oper- ators into the ones in our basis. 4 Observable eﬀects of new scalars The leading indirect eﬀects of new heavy scalar particles on physical observables are de- scribed by the dimension ﬁve and six eﬀective operators, with coeﬃcients given in tables 9– 28 in appendix B. In this section we give an overview of these eﬀects. We discuss colored and colorless scalars in turn. The scalar ﬁelds with SU(3)c quantum numbers manifest themselves, at dimension six, only through four-fermion interactions. Therefore they can be tested in two-to-two fermion processes or in particle decays. The ﬂavor structure of the scalar interactions makes them particularly sensitive to constraints from ﬂavor-violating processes, although it is always possible to go into an alignment limit in which each new scalar couples exclusively to certain fermion generations, up to factors of the Cabibbo-Kobayashi-Maskawa (CKM) matrix. In such a case, the bounds from ﬂavor-preserving processes can become dominant. We will only consider this scenario in the following. The scalar ﬁelds ω1, ω4, Π1, Π7 and ζ carry quantum numbers that allow for lepton- quark Yukawa interactions. These scalar leptoquarks can be tested in the dilepton processes e+e−→had at LEP2 and pp →ℓ+ℓ−at the LHC, where the new particles are exchanged in the t channel, and also in low-energy experiments (e.g. parity violation in atoms). The multiplets ω1, ω4 and ζ admit purely hadronic interactions too. The simultaneous presence of all these interactions introduces a violation of lepton (L) and baryon (B) number, as is – 8 – manifest by the generation of the operators labeled as “B & L” in tables 19, 17, and 22. These contributions are proportional to the product of one lepton-quark and one quark- quark Yukawa coupling, so the strong constraints set by the non-observation of proton decay forces one of these two couplings to be very small (see for instance [19]). The remaining colored representations, ω2, Ω1, Ω2, Ω4, Υ and Φ, only admit quark Yukawa interactions and therefore generate only four-quark contact interactions. Aside from ﬂavor observables, these can be tested in dijet production at hadron colliders. If coupled to the ﬁrst and third family, top pair production (with opposite or same sign) is also possible. The phenomenology of colorless scalars is signiﬁcantly richer. Again, the exchange of some colorless heavy scalars generates four-fermion operators, which in this case include operators with four leptons. In particular, all the four-fermion interactions induced by the multiplets S1,2 and Ξ1 are purely leptonic. 5Four-lepton operators not involving electrons could in principle be accessible through four-lepton pro- duction at the LHC or ILC. The sensitivity is however only marginal if there is no resonant production [20]. 4 Observable eﬀects of new scalars Nevertheless, the constraints on the coeﬃ- cient of OφD from EWPD are signiﬁcantly stronger than the ones from Higgs physics [26, 27], and completely dominate in the global ﬁt. The operator Oφ introduces corrections to the Higgs vev and mass parameters, which can always be absorbed in the physical values. It also gives a direct contribution to the self-coupling of the physical Higgs, JHEP04(2015)078 ∆LH3 = 5 6αφ v3 Λ2 H3, (4.3) (4.3) which is in principle observable in Higgs pair production. However, the LHC data at 8 TeV are not sensitive enough to probe the Higgs self-coupling. Given the small cross sections for Higgs pair production in the SM, a measurement of the Higgs self-coupling seems to be challenging even with the results of Run 2, and may require of the high-luminosity upgrade of the LHC [28]. On the other hand, a relatively large enhancement in diHiggs production due to the eﬀect of Oφ could unveil the presence of new physics eﬀects before sensitivity to the SM coupling is attained. With the exception of S1 and S2, all the colorless multiplets contribute to Higgs ob- servables (the singlet S and the two quadruplets Θ1 and Θ2 only contribute to Higgs observables). The singlet S and the iso-triplets Ξ0 and Ξ1 contribute to Oφ□and their trilinear couplings can therefore be constrained by Higgs measurements. The latter two, however, contribute to OφD via the same trilinear coeﬃcients, which are therefore con- strained mainly by EWPD. The quadruplets Θ1 and Θ2 only contribute to Oφ, so, to dimension six, their eﬀects are not observable in current data. Finally, the scalar-fermion operators Oeφ = φ†φ lLφeR , Odφ = φ†φ (qLφdR), and Ouφ = φ†φ qL ˜φuR correct the SM Yukawa interactions: ∆LYukawa = 1 √ 2H (αeφ)ij ei Lej R + (V αuφ)ij ui Luj R + (αdφ)ij di Ldj R + h.c. v2 Λ2 . (4.4) (4.4) In this equation we have already reabsorbed the corrections to the masses in the deﬁnition of the SM Yukawa matrices. These operators are generated by the colorless iso-triplets and iso-doublet. For Ξ0 and Ξ1, the coeﬃcients are proportional to the (squared) trilinear couplings, which as pointed out above contribute to the T parameter and are strongly constrained by EWPD. 4 Observable eﬀects of new scalars The eﬀects of these operators could show up in the e+e−→ℓ+ℓ−data taken at LEP2 or in low-energy experiments such as measurements of parity violation in Møller scattering.5 In general, these four-lepton operators also con- tribute to very sensitive lepton ﬂavor violating processes, such as µ−→e+e−e−(note that this cannot be mediated by S1) or τ −→µ−e+e−. But similarly to the quark case, these dangerous eﬀects are absent when the scalar couplings are properly aligned with the SM lepton ﬂavors. JHEP04(2015)078 Colorless scalars can also have visible eﬀects in other types of observables. To start with, the hypercharge-one iso-triplet Ξ1 is the only scalar multiplet that can produce lepton- number violation. Indeed, it contributes to the dimension ﬁve Weinberg operator [21] O5 = lc L ˜φ∗˜φ†lL, which generates Majorana masses for the SM neutrinos. This is nothing but the well-known seesaw mechanism of type II. Unless the scalar iso-triplet is very heavy (and thus does not contribute to other observables), the smallness of neutrino masses requires that either the lepton Yukawa couplings of the scalar or its linear interactions with the Higgs be tiny [22–24]. Colorless scalars contribute to the following three purely bosonic dimension-six opera- tors: Oφ□= φ†φ □ φ†φ , OφD = φ†Dµφ Dµφ†φ and Oφ = φ†φ 3 /3. The ﬁrst one, OφD, is in one to one correspondence with the Peskin-Takeuchi oblique T parameter [25], T = −αφD 2αem v2 Λ2 , (4.1) (4.1) which is very strongly constrained by electroweak precision data (EWPD), especially now that the value of the Higgs mass is known. Only the colorless iso-triplets Ξ0 and Ξ1 contribute to OφD at the tree level. Hence, these are the only scalars whose eﬀects break custodial symmetry at tree level, to dimension six in the eﬀective Lagrangian expansion. In this regard, note that, unlike the weak iso-singlets and iso-doublets, quadruplets do also break custodial isospin at the tree level if they acquire a vev. However, because of the absence of trilinear interactions with two Higgs ﬁelds, this vev is suppressed by a factor – 9 – of O(1/M 2). Therefore, quadruplet custodial isospin breaking eﬀects appear starting at dimension eight. Both OφD and Oφ□renormalize the wavefunction of the physical Higgs ﬁeld, H → 1 + αφ□ v2 Λ2 −1 4αφD v2 Λ2 H, (4.2) (4.2) and therefore enter in most Higgs observables. 5 Precision constraints on new scalars 5 As explained in the previous section, the eﬀects of the scalar couplings in tables 9–27 can potentially be observed in several diﬀerent physical processes. The good agreement of the SM predictions with most of the current observations implies bounds on the diﬀerent interactions. In this section, we use the eﬀective Lagrangian results obtained in the previous sections to derive ﬂavor-conserving limits on some of the scalar representations. For the sake of simplicity, in the ﬁts presented here we consider only one scalar multiplet at a time and always assume that only one of the possible couplings of each scalar is non- vanishing.6 In most cases, this assumption gives rise to conservative limits. More general scenarios are certainly interesting and can be studied with the tools provided in this paper. At any rate, it is important to observe that there are strong phenomenological reasons for not considering certain couplings simultaneously, as we explain next. JHEP04(2015)078 As stressed above, the new scalar fermionic interactions do not conserve ﬂavor in general. Thus, they are subject to the constraints imposed by observables measured in ﬂavor-violating processes, which are usually much stronger than the ones from ﬂavor- conserving observables. For pure hadronic interactions in the form of four-quark operators, for instance, the observables with ∆F = 2 transitions (e.g. ϵK or ∆mK, measured in K0 −K0 mixing) impose bounds on the new physics scale typically around 102–104 TeV, assuming order-one couplings [30]. Lepton ﬂavor violating processes also impose strong bounds, especially from rare decays such as µ−→e+e−e−or µ →eγ. Flavor-preserving results are meaningful in scenarios in which ﬂavor constraints are subdominant or do not apply. For instance, ﬂavor constraints can be avoided in a natural manner by enforcing an appropriate symmetry on these SM extensions, which requires extending each scalar gauge multiplet to a full multiplet under the corresponding ﬂavor group [15]. From the point of view of our model-independent description of new scalars, each of these ﬂavor multiplets corresponds to several copies of one of our gauge-covariant multiplets, with correlated couplings.7 The presence of tree-level ﬂavor changing neutral currents can also be softened if the new Yukawa interactions are adequately aligned with the SM ﬂavors. In particular, any ﬂavor violation can be removed — up to terms suppressed by the corresponding CKM matrix elements — if, for each scalar multiplet, only one entry of the new Yukawa matrices is non-zero. 6Besides the obvious simpliﬁcations of reducing the number of free parameters and allowing for a simple one-dimensional presentation of the results, this assumption allows us to use the pp →jj results of [29]. Indeed, the four-quark operators O(1) qud and O(8) qud, not considered in that reference, are generated for some scalar representations, but their coeﬃcients always involve the product of two diﬀerent Yukawa couplings. 7Since ref. [15] concentrates on quark processes, its classiﬁcation does not include the scalar representa- tions that do not have purely quark interactions. 4 Observable eﬀects of new scalars In the case of the doublet ϕ, the coeﬃcients are proportional to the product of the scalar coupling λϕ ϕ†φ φ†φ + h.c. and the corresponding fermionic coupling. The Higgs observables only constrain this product but not the individual cou- plings. (The scalar coupling enters quadratically in the coeﬃcient of Oφ but, as indicated above, there is no signiﬁcant bound on this coeﬃcient at present.) In the next section we give some numerical results for bounds that can be obtained on the couplings and masses of the heavy scalar particles from the available measurements. – 10 – p , y p p g 7Since ref. [15] concentrates on quark processes, its classiﬁcation does not include the scalar represent tions that do not have purely quark interactions. 5 Precision constraints on new scalars We will restrict ourselves to this case in the present section. This tuned choice provides conservative bounds. It also helps to establish in which places certain new physics eﬀects might be hidden and to determine their maximum size allowed by current data. We also noted in the discussion of the previous section that in several cases the new scalars can contribute to other extremely sensitive physical observables, such as proton decay or neutrino masses. Since one can always assign deﬁnite B and L numbers to the new scalars, such contributions can only appear as the product of two interactions selecting – 11 – diﬀerent assignments of these quantum numbers. Therefore, they are always avoided when only one of these couplings is non-vanishing. Finally, certain contributions to four fermion operators that would give rise to charged- current interactions mediating rare decays are also absent when we only consider one non- zero coupling. For instance, the observable Rπ = Γ(π+ →νe+)/Γ(π+ →νµ+) set bounds on the operators Oqde, Oledq and Oluqe, which are signiﬁcantly stronger than the ones from the LHC and EWPD considered here [31, 32]. The same holds for same-sign top pair production [33]. In the scenario we are considering, with no contribution to any of these sensitive observables, the most relevant constraints on the couplings and masses of the new scalars come from ﬂavor-, B- and L-blind observables. Our ﬁts combine the bounds on dimension six interactions from EWPD [34] (see also [35]),8 LHC dilepton [32] and dijet searches [29], and measurements of Higgs observables [26].9 In all the analyses we ﬁx the SM inputs to their best-ﬁt values in the absence of extra scalars, JHEP04(2015)078 mH = 125.1 ± 0.2 GeV, mt = 173.8 ± 0.8 GeV, MZ = 91.1880 ± 0.0020 GeV, αs(M2 Z) = 0.1186 ± 0.0006, ∆α(5) had(M2 Z) = 0.02754 ± 0.00010, (5.1) mH = 125.1 ± 0.2 GeV, mt = 173.8 ± 0.8 GeV, MZ = 91.1880 ± 0.0020 GeV, αs(M2 Z) = 0.1186 ± 0.0006, ∆α(5) had(M2 Z) = 0.02754 ± 0.00010, (5.1) (5.1) and vary only the new-physics parameters. This is a good approximation, since large eﬀects are not allowed. The limits we obtain in this way are presented in table 2 for the colorless multiplets, and in table 3 for the ones charged under SU(3)c. 9We do not include here limits from (opposite sign) top pair production on couplings mixing the ﬁrst and third generation of quarks. These can be obtained from LHC data and the results in [45] and [14] and will be considered elsewhere. 8This includes the usual Z-pole data [36], ∆α(5) had(M 2 Z) [37], αs(M 2 Z) [38], the top [39] and Higgs [40, 41] masses, the W mass and width [42, 43], the ﬁnal LEP2 results of e+e−→¯ff [44], unitarity constraints on the the CKM matrix [38], as well as several low-energy measurements [38]. 5 Precision constraints on new scalars Bound [TeV−1] S \|κS\| M2 S 1.55 S1 yl S1 MS1     − 0.08 − 0.08 − − − − −     S2 ye S2 MS2     0.36 0.19 0.28 0.19 − − 0.28 − −     ϕ \|ye ϕ\| Mϕ     0.26 0.56 0.79 0.56 − − 0.79 − −     \|(yd ϕ)11\| Mϕ 0.61 \|(yu ϕ)11\| Mϕ 0.44 Ξ0 \|κΞ0\| M2 Ξ0 0.11 Ξ1 \|κΞ1\| M2 Ξ1 0.04 yl Ξ1 MΞ1     0.33 0.09 0.18 0.09 − − 0.18 − −     JHEP04(2015)078 Table 2. Bounds on the colorless new scalars from ﬂavor-preserving observables. The results for the Yukawa matrices are obtained from a ﬁt to each one of the entries of the coupling matrices at a time. The limit on κS is determined exclusively by the Higgs data, while the ones on the κΞi couplings are dominated by the EWPD limits on the T parameter. Leptonic couplings are constrained by the LEP2 (e+e−→ℓ+ℓ−) and low energy measurements (e.g. Møller and ν-electron scattering), while the hadronic ones are bounded by the LHC dijet angular distributions. Table 2. Bounds on the colorless new scalars from ﬂavor-preserving observables. The results for the Yukawa matrices are obtained from a ﬁt to each one of the entries of the coupling matrices at a time. The limit on κS is determined exclusively by the Higgs data, while the ones on the κΞi couplings are dominated by the EWPD limits on the T parameter. Leptonic couplings are constrained by the LEP2 (e+e−→ℓ+ℓ−) and low energy measurements (e.g. Møller and ν-electron scattering), while the hadronic ones are bounded by the LHC dijet angular distributions. Table 2. Bounds on the colorless new scalars from ﬂavor-preserving observables. The results for the Yukawa matrices are obtained from a ﬁt to each one of the entries of the coupling matrices at a time. The limit on κS is determined exclusively by the Higgs data, while the ones on the κΞi couplings are dominated by the EWPD limits on the T parameter. Leptonic couplings are constrained by the LEP2 (e+e−→ℓ+ℓ−) and low energy measurements (e.g. Møller and ν-electron scattering), while the hadronic ones are bounded by the LHC dijet angular distributions. 5 Precision constraints on new scalars In all cases the limits apply to ratios of couplings and masses, which are the quantities that appear in the coeﬃcients of the eﬀective operators. (In some cases tailored searches can give better bounds when the new scalars can be directly produced [46].) Let us comment on the few absences in those tables. In the colorless case, we cannot put meaningful bounds on the quadruplet couplings. As explained in the previous section, they only modify the Higgs self-coupling, which is not signiﬁcantly constrained by the LHC data at 8 TeV. In the colored sector, we have not presented any bounds for ω2, nor for the hadronic couplings of ω4 and ζ. These could be in principle constrained by the LHC dijet data. However, the hadronic couplings of these three multiplets are antisymmetric and necessarily involve more than one family. Hence, they go beyond the ﬁrst-family approximation used in [29]. Putting bounds on them would require an extended analysis. Finally, let us discuss the range of validity of the eﬀective Lagrangian. In this approach, the results are given on the ratios yσ/Mσ, where yσ < 4π to allow for a loop expansion, or κσ/M 2 σ, where we assumed \|κσ\| ≲Mσ. To guarantee the validity of our bounds, we need to assume that the scalar masses are suﬃciently larger than the relevant energies and momenta of the processes we consider. This condition depends on each observable and coupling and is always satisﬁed by large enough values of the masses. But for large masses the upper 8This includes the usual Z-pole data [36], ∆α(5) had(M 2 Z) [37], αs(M 2 Z) [38], the top [39] and Higgs [40, 41] masses, the W mass and width [42, 43], the ﬁnal LEP2 results of e+e−→¯ff [44], unitarity constraints on the the CKM matrix [38], as well as several low-energy measurements [38]. 9We do not include here limits from (opposite sign) top pair production on couplings mixing the ﬁrst and third generation of quarks. These can be obtained from LHC data and the results in [45] and [14] and will be considered elsewhere. – 12 – Scalar Parameter 95% C.L. 5 Precision constraints on new scalars Bound [TeV−1] S \|κS\| M2 S 1.55 S1 yl S1 MS1     − 0.08 − 0.08 − − − − −     S2 ye S2 MS2     0.36 0.19 0.28 0.19 − − 0.28 − −     ϕ \|ye ϕ\| Mϕ     0.26 0.56 0.79 0.56 − − 0.79 − −     \|(yd ϕ)11\| Mϕ 0.61 \|(yu ϕ)11\| Mϕ 0.44 Ξ0 \|κΞ0\| M2 Ξ0 0.11 Ξ1 \|κΞ1\| M2 Ξ1 0.04 yl Ξ1 MΞ1     0.33 0.09 0.18 0.09 − − 0.18 − −     Table 2. Bounds on the colorless new scalars from ﬂavor-preserving observables. The results for the Yukawa matrices are obtained from a ﬁt to each one of the entries of the coupling matrices at a time. The limit on κS is determined exclusively by the Higgs data, while the ones on the κΞi couplings are dominated by the EWPD limits on the T parameter. Leptonic couplings are constrained by the LEP2 (e+e−→ℓ+ℓ−) and low energy measurements (e.g. Møller and ν-electron scattering), while the hadronic ones are bounded by the LHC dijet angular distributions. Scalar Parameter 95% C.L. Bound [TeV−1] S \|κS\| M2 S 1.55 S1 yl S1 MS1     − 0.08 − 0.08 − − − − −     S2 ye S2 MS2     0.36 0.19 0.28 0.19 − − 0.28 − −     ϕ \|ye ϕ\| Mϕ     0.26 0.56 0.79 0.56 − − 0.79 − −     \|(yd ϕ)11\| Mϕ 0.61 \|(yu ϕ)11\| Mϕ 0.44 Ξ0 \|κΞ0\| M2 Ξ0 0.11 Ξ1 \|κΞ1\| M2 Ξ1 0.04 yl Ξ1 MΞ1     0.33 0.09 0.18 0.09 − − 0.18 − −     Scalar Parameter 95% C.L. 5 Precision constraints on new scalars Bounds on the colored new scalars from ﬂavor-preserving observables. The results for the Yukawa matrices are obtained from a ﬁt to each one of the entries of the coupling matrices at a time. All these interactions are constrained by two to two fermion processes. Leptoquark interactions are bounded by LEP 2 e+e−→had data, low energy measurements (e.g. Atomic parity violation, ν-nucleon scattering), CKM unitarity, and dilepton searches at the LHC. Purely hadronic bounds are again obtained only from the LHC pp →jj angular distributions. See text for more details. Scalar Parameter 95% C.L. Bounds [TeV−1] Π7 ylu Π7 MΠ7     0.27 1.04 − 0.33 − − − − −     yeq Π7 MΠ7     0.29 0.93 1.06 0.32 − − − − −     Ω1 yud Ω1 11 MΩ1 0.78 Ω2 yd Ω2 11 MΩ2 0.68 Ω4 yu Ω4 11 MΩ4 0.47 ζ yql ζ Mζ     0.21 0.30 − 0.66 − − 0.47 − −     Φ \|(yqu Φ )11\| MΦ 0.88 ydq Φ 11 MΦ 1.12 Υ \|yq Υ\| MΥ 0.32 Scalar Parameter 95% C.L. Bound [TeV−1] ω1 yql ω1 Mω1     0.19 0.53 − 0.40 − − − − −     \|(yqq ω1)11\| Mω1 0.24 \|yeu ω1\| Mω1     0.27 0.49 − 0.48 − − − − −     (ydu ω1)11 Mω1 0.47 ω4 \|yed ω4\| Mω4     0.28 0.98 0.98 0.42 − − − − −     Π1 yld Π1 MΠ1     0.27 1.80 1.80 0.48 − − − − −     JHEP04(2015)078 Table 3. Bounds on the colored new scalars from ﬂavor-preserving observables. The results for the Yukawa matrices are obtained from a ﬁt to each one of the entries of the coupling matrices at a time. All these interactions are constrained by two to two fermion processes. Leptoquark interactions are bounded by LEP 2 e+e−→had data, low energy measurements (e.g. Atomic parity violation, ν-nucleon scattering), CKM unitarity, and dilepton searches at the LHC. Purely hadronic bounds are again obtained only from the LHC pp →jj angular distributions. See text for more details. accessible by the LHC at √s = 8 TeV. 5 Precision constraints on new scalars region of the allowed parameters may involve strong couplings that threaten perturbativity and thus the validity of the tree-level approximation. This can happen when the limits on coupling/mass ratios are weak. One example is the scalar singlet S, which is only observable through its contributions to the Higgs boson wave function via the operator Oφ□, with mild limits [26, 27]. For \|κS\| ∼MS, the bound in table 2 implies MS ≳700 GeV, which is close to the scale probed at the LHC, and the validity of this bound might be questioned. In some cases, entries involving electrons coupling with the second and third family of quarks are also relatively weak ∼O(1–2) TeV−1. For weakly coupled scenarios (yσ < 1) this implies the new scalar masses can be around 500-1000 GeV. However, these entries are only constrained by the LEP 2 data, which involves lower energies √s ≤209 GeV. Those entries that can modify dilepton production at the LHC can be as large as ∼0.5 TeV−1 for the case of Π1. This translates into a mass scale of ∼2 TeV, which is in principle – 13 – Scalar Parameter 95% C.L. Bound [TeV−1] ω1 yql ω1 Mω1     0.19 0.53 − 0.40 − − − − −     \|(yqq ω1)11\| Mω1 0.24 \|yeu ω1\| Mω1     0.27 0.49 − 0.48 − − − − −     (ydu ω1)11 Mω1 0.47 ω4 \|yed ω4\| Mω4     0.28 0.98 0.98 0.42 − − − − −     Π1 yld Π1 MΠ1     0.27 1.80 1.80 0.48 − − − − −     Scalar Parameter 95% C.L. Bounds [TeV−1] Π7 ylu Π7 MΠ7     0.27 1.04 − 0.33 − − − − −     yeq Π7 MΠ7     0.29 0.93 1.06 0.32 − − − − −     Ω1 yud Ω1 11 MΩ1 0.78 Ω2 yd Ω2 11 MΩ2 0.68 Ω4 yu Ω4 11 MΩ4 0.47 ζ yql ζ Mζ     0.21 0.30 − 0.66 − − 0.47 − −     Φ \|(yqu Φ )11\| MΦ 0.88 ydq Φ 11 MΦ 1.12 Υ \|yq Υ\| MΥ 0.32 Table 3. 10A direct comparison of the eﬀective Lagrangian results presented here with those in refs. [5–7] requires to perform certain ﬁeld redeﬁnitions and Fierz reorderings, since the basis employed in those works has some redundant interactions, and use diﬀerent deﬁnitions for some operators. All the transformations needed to relate both bases are provided in appendix A, where we also introduce the full basis of dimension six operators we use, which can be compared to the one in table 7 in ref. [7]. 5 Precision constraints on new scalars However, dilepton processes at the LHC can only be mediated by leptoquarks in t-channel. Thus, even for such values of masses and weak couplings, the eﬀective theory remains valid in a large region of the phase space and can give a good approximation to the integrated observables (see the quantitative discussion of an analogous process in [14]). Finally, the results for the pure hadronic interactions should be taken with care. Again, for order one couplings the corresponding mass scales can be relatively small, while the dijet angular distributions used in [29] to set bounds correspond to dijet masses Mjj > 3 TeV. Unlike the dilepton case, these scalars can be produced in s-channel. For instance, for the case with the weakest bound, the color-octet iso-doublet Φ, demanding MΦ > 3.9 TeV ( the highest Mjj value observed in the CMS analysis [47] used in [29]), the hadronic couplings needed to saturate the bounds must be (ydq Φ )11 ≳4, and one may start worrying about the precision of the perturbative (asymptotic) series. Note, nevertheless, that for s-channel processes the limits obtained with the eﬀective Lagrangian give a conservative estimate of the actual limit. In summary, with the current constraints – 14 – the eﬀective Lagrangian approach provides in general a good approximation for heavy scalars and a large range of values of their couplings. 6 Scalar extensions with other new particles In the previous section we have considered in detail SM extensions with only one scalar multiplet. In this case, there are strong correlations between diﬀerent observables. In this section we discuss the interplay between the eﬀects of diﬀerent scalars, and also between particles with diﬀerent spin. The eﬀective Lagrangian formalism used in this article allows an easy comparison of the eﬀects of diﬀerent sources of new physics. In particular, it helps to identify at the Lagrangian level those places where a (partial) cancellation between the virtual eﬀects of diﬀerent new particles in physical observables can take place [48, 49]. This is useful for model builders to construct scenarios with not too heavy (or not too weakly coupled) particles that are consistent with the existing phenomenological constraints. Such cancellations require a large correlation between the eﬀects of diﬀerent new particles. Al- though they correspond to small regions in the parameter space of generic models, in some cases they can be made natural by imposing extra symmetries. JHEP04(2015)078 Let us ﬁrst point out that, at the tree level, the dimension-six eﬀective Lagrangian in extensions of the SM with arbitrary new particles of spin 0, 1/2 and 1 is simply the sum of the eﬀective Lagrangian obtained here and the ones in refs. [5–7].10 Indeed, a simple extension of the argument in section 3 shows that mixed contributions from particles of diﬀerent spin only appear at higher dimension. Therefore, the eﬀective Lagrangians in those references and in this paper completely characterize the largest eﬀects of arbitrary extensions of the SM with new heavy particles. Note in this regard that particles of spin higher than 1 only interact via non-renormalizable couplings, which are naturally suppressed. In what follows we study, for each type of interaction induced by the extra scalars to dimension six, the diﬀerent sources of new physics (new scalars, fermions or vectors) that can cancel at the tree-level the eﬀects from the virtual exchange of scalar bosons. We discuss the cancellations at the operator level, which is a suﬃcient (and often necessary) condition to guarantee the cancellation in physical observables. • Dimension-ﬁve operators: the Weinberg operator only arises when we integrate the hypercharge-one iso-triplet Ξ1. Having no deﬁnite sign, contributions from diﬀer- ent triplets could cancel each other, or the ones coming from the other two possible seesaw messengers, i.e. new lepton singlets and triplets with hypercharge zero. 6 Scalar extensions with other new particles • Dimension-ﬁve operators: the Weinberg operator only arises when we integrate the hypercharge-one iso-triplet Ξ1. Having no deﬁnite sign, contributions from diﬀer- ent triplets could cancel each other, or the ones coming from the other two possible seesaw messengers, i.e. new lepton singlets and triplets with hypercharge zero. • Oblique operators: cancellation between custodial isospin breaking contributions can occur between the two triplets, as both have deﬁnite, opposite sign. This is also • Oblique operators: cancellation between custodial isospin breaking contributions can occur between the two triplets, as both have deﬁnite, opposite sign. This is also – 15 – possible with new vector singlets with hypercharge zero (hypercharge one), which yield negative (positive) deﬁnite contributions to αφD. New vector triplets with hypercharge one also yield negative contributions to this operator [7]. possible with new vector singlets with hypercharge zero (hypercharge one), which yield negative (positive) deﬁnite contributions to αφD. New vector triplets with hypercharge one also yield negative contributions to this operator [7]. possible with new vector singlets with hypercharge zero (hypercharge one), which yield negative (positive) deﬁnite contributions to αφD. New vector triplets with hypercharge one also yield negative contributions to this operator [7]. • Scalar operators: these include Oφ and Oφ□. A look at the second row of table 28 in appendix B shows all the possible contributions to the operator Oφ, including collective contributions that appear when two diﬀerent species of new scalars are present at the same time. Contributions to Oφ□, on the other hand, are much simpler, and are negative (positive) deﬁnite for S (Ξ0,1). This allows for cancellations between colorless iso-singlets and iso-triplets. JHEP04(2015)078 • Scalar-Fermion operators: in extensions with new scalars only, these operators only appear in the case of colorless iso-doublets, ϕ or iso-triplets, Ξ0,1. However, for the latter they always arise through a ﬁeld redeﬁnition, necessary to bring all the con- tributions in the dimension-six Lagrangian into the chosen basis. As a consequence, the ﬂavor structure of the coeﬃcients of the operators Oeφ and Odφ, and Ouφ coming from triplets is always SM-like (proportional to the SM Yukawa couplings), while the one from doublets ϕ can be completely generic. Moreover, while all the contributions of arbitrary triplets have the same sign, the sign of the genuine contributions from ϕ is indeﬁnite. Therefore, cancellations between scalars are always possible. 6 Scalar extensions with other new particles Contributions from new vectors [7], also appear when the SM equations of mo- tion are used. They only come from colorless hypercharge zero or one iso-singlets, or iso-triplets (B, B1, W and W1 in the notation of [7]), and are also SM-like. Finally, heavy fermions can generate these contributions either after applying the SM equation of motion, if only one fermion species is present, or as a result of the combined eﬀect of extra fermionic iso-doublets and new fermion iso-singlets or iso-triplets [5, 6]. Even in the case of only one fermion, the contributions to Ofφ al- ways involve the ﬂavor structure of the new fermionic interactions, and therefore are general a priori. Hence, some interplay with the contributions from scalar doublets is possible, although an eventual cancellation of all the scalar eﬀects may require several diﬀerent new fermion multiplets. • Four-fermion operators: upon inspection of the new scalar contributions to four- fermion operators, it can be seen that, for a ﬁxed set of ﬂavor indices with i = j, k = l, the operator coeﬃcients of all four-fermion interactions involving at most two diﬀerent types of SM fermion multiplets have a deﬁnite sign. Although the contributions to four-fermion interactions with three or four diﬀerent multiplets have no deﬁnite sign, they are always correlated with operators involving only one or two kinds of multiplets. Moreover, for a particular operator, contributions from scalars of diﬀerent types have either the same sign, or are proportional to the contribution to another operator with the same ﬁeld content where both scalars contribute additively. For instance, ϕ and Φ contributions to α(8) qu have opposite sign and can balance each other. However, each individual contribution is proportional to the corresponding one to α(1) qu , where both have the same sign. Therefore, a complete cancellation – 16 – of the eﬀects from such couplings to two-to-two fermion processes is not possible in extensions with extra scalars only. This is quite similar to the situation for extra vectors [7]. However, as illustrated in ref. [49] for the case of pure leptonic interactions, a cancellation between the four fermion eﬀects coming from new scalars and the ones from new vector particles is possible in many cases, although it comes at the price of a signiﬁcant ﬁne tuning. 6 Scalar extensions with other new particles New vectors of hypercharge Y ̸= 0 contribute with a deﬁnite sign to four-fermion operators involving at most two types of SM multiplets if i = j, k = l, exactly as in the scalar case. For vectors of zero hypercharge, only the operators where all fermions belong to the same SM representation can have a deﬁnite sign. This is always the case for i = j = k = l and, in certain cases, for i = l, j = k.11 Instead of going over each operator/scalar/vector and listing all the possible cancellations, we show in table 4 those interactions that are common for each scalar-vector pair, indicating the relative sign between the diﬀerent contributions for both cases, i = l, j = k (i ̸= j), and i = j = k = l (where some restrictions appear in the case of hypercharge-zero vector ﬁelds). In general, for a given operator with four multiplets of the same kind, one can always choose a scalar and a hypercharge-zero vector ﬁeld such that, tuning the corresponding scalar/vector couplings, contributions with opposite sign are obtained. Table 5 contains the same information for the case of four-fermion operators built from at most two types of multiplets, with i = j, k = l (i ̸= k). JHEP04(2015)078 Note that a relative minus sign between the contributions from two particles to a given operator does not always guarantee that a complete cancellation of the new physics eﬀects is possible. The reason is that the contributions to some operators with the same ﬁeld content are in many cases correlated, and a cancellation in all those operator coeﬃcients does not usually take place at the same time. However, for each conﬁguration of four fermionic ﬁelds, there are at most two independent operators in the dimension six basis. And, as can be seen from the tables, for each scalar (vector) ﬁeld and pair of such operators, one can always ﬁnd a pair of vectors (scalars) that contribute to both operators with the adequate signs to cancel the total contribution. Therefore, we conclude that for any given four-fermion process receiving contributions from one new particle through an arbitrary set of four-fermion operators, it is always possible to ﬁnd a combination of new ﬁelds that, after the adequate tuning in their couplings, cancels out all the new eﬀects. Correlations with other types of operators can be easily avoided. 11Note that in this case, the corresponding scalar contributions always have a deﬁnite sign. 6 Scalar extensions with other new particles Summarizing, we see that the existing experimental limits are compatible with many combinations of new particles with sizable couplings and masses at the LHC reach. Essen- tially, by including many new multiplets we are breaking the correlations in the coeﬃcients of the eﬀective operators. However, in most cases the corresponding models are too con- trived and ﬁne tuned. In the simplest cases, symmetries may exist which make these models more natural and appealing. The discussion in this section may be useful in the search of such symmetries. – 17 – S1 S2 Ξ1 ω1 ω2 ω4 ζ Ω1 Ω2 Ω4 Υ Bµ +α(1) ll ±αee −α(1) ll −α(1) qq +α(1) dd +α(1) uu +α(1) qq +α(1) qq −α(1) dd −α(1) uu −α(1) qq (—) (−αee) (−α(1) ll ) (−α(1) qq ) (—) (—) (—) (—) (−α(1) dd ) (−α(1) uu) (−α(1) qq ) Wµ ±α(1) ll ±α(1) ll ±α(1) qq ±α(1) qq ±α(1) qq ±α(1) qq ±α(8) qq ±α(8) qq ±α(8) qq ±α(8) qq (—) (−α(1) ll ) (+α(1) qq ) (—) (—) (+α(1) qq ) (+α(8) qq ) (—) (—) (−α(8) qq ) Gµ +α(8) qq ±α(1) dd ±α(1) uu −α(8) qq +α(8) qq ±α(1) dd ±α(1) uu −α(8) qq (+α(8) qq ) (—) (—) (—) (—) (−α(1) dd ) (−α(1) uu) (−α(8) qq ) Hµ −α(1) qq +α(1) qq +α(1) qq −α(1) qq ±α(8) qq ) ±α(8) qq ±α(8) qq ±α(8) qq (−α(1) qq ) (—) (—) (−α(1) qq ) (−α(8) qq ) (—) (—) (+α(8) qq ) Table 4. Contributions to four-fermion interactions with i = l, j = k (i ̸= j), common to new scalar and (hypercharge-zero) vector ﬁelds. Only operators involving one type of multiplet, and the particles that contribute to them, are shown. The symbols “+” (“−”) indicate that the contributions from scalars and vectors have the same (opposite) sign, while “±” indicates the absence of a deﬁnite sign in any of the sources. The same information for the case i = j = k = l is provided in parenthesis. In this case, a dash (“—”) indicates the absence of contribution from the corresponding scalar particle. JHEP04(2015)078 (+α(8) qq ) Table 4. Contributions to four-fermion interactions with i = l, j = k (i ̸= j), common to new scalar and (hypercharge-zero) vector ﬁelds. Only operators involving one type of multiplet, and the particles that contribute to them, are shown. 12Indirect evidence of gauge-Higgs interactions was available before, in EWPD. 6 Scalar extensions with other new particles The symbols “+” (“−”) indicate that the contributions from scalars and vectors have the same (opposite) sign, while “±” indicates the absence of a deﬁnite sign in any of the sources. The same information for the case i = j = k = l is provided in parenthesis. In this case, a dash (“—”) indicates the absence of contribution from the corresponding scalar particle. 7 Conclusions We use “±” to indicate the absence of a deﬁnite sign in one of the sources. For the case of antisymmetric scalar couplings the sign is compared with the exact same index conﬁguration as in tables 9–27. operators common to new scalar and vector ﬁelds. We only show operators that receive deﬁnite sign contributions from i = j, k = l (i ̸= k). The symbols “+” (“−”) indicate that the contributions from scalars and vectors have the same “±” to indicate the absence of a deﬁnite sign in one of the sources. For the case of antisymmetric scalar couplings the JHEP04(2015)078 5. Four-fermion operators common to new scalar and vector ﬁelds. We only show operators that receive deﬁnite sign contributions from one particle, for i = j, k = l (i ̸= k). The symbols “+” (“−”) indicate that the contributions from scalars and vectors have the same te) sign. We use “±” to indicate the absence of a deﬁnite sign in one of the sources. For the case of antisymmetric scalar couplings the sign is compared with the exact same index conﬁguration as in tables 9–27. 7 Conclusions The discovery at the LHC of a new particle of spin 0 has come hand in hand with the direct observation of new interactions mediated by scalar ﬁelds.12 Among these, the Yukawa interactions are quite unique in that they are not ruled by gauge invariance under the SM gauge group, although of course they are compatible with it. The exploration of this scalar sector is an important part of the LHC physics program. The results at the LHC Run 1 have already constrained signiﬁcantly its structure. So far, all the measurements are consistent with the minimal scalar sector of the SM: a Higgs iso-doublet with a non-vanishing vev for its neutral component. But the present uncertainties still allow for signiﬁcant deviations in the couplings of this doublet and for the presence of additional scalar ﬁelds, related or not to electroweak symmetry breaking. To comply with all the available data, such extra scalars must either have small couplings or be signiﬁcantly heavier than the Higgs boson. The latter is the scenario we have considered in the second part of this paper. We have followed a largely model-independent and unbiased approach, with the min- imal theoretical input of gauge-invariance. First, we have classiﬁed into 19 irreducible representations all the possible scalar ﬁelds that can interact linearly with the SM ﬁelds with gauge-invariant renormalizable couplings. Their components with a deﬁnite electric charge are the only scalar particles that can be produced at colliders with unsuppressed – 18 – JHEP04(2015)078 Table 5. Four-fermion operators common to new scalar and vector ﬁelds. We only show operators that receive deﬁnite sign contributions from at least one particle, for i = j, k = l (i ̸= k). The symbols “+” (“−”) indicate that the contributions from scalars and vectors have the same (opposite) sign. We use “±” to indicate the absence of a deﬁnite sign in one of the sources. For the case of antisymmetric scalar couplings the relative sign is compared with the exact same index conﬁguration as in tables 9–27. 5. Four-fermion operators common to new scalar and vector ﬁelds. We only show operators that receive deﬁnite sign contributions from one particle, for i = j, k = l (i ̸= k). The symbols “+” (“−”) indicate that the contributions from scalars and vectors have the same te) sign. 7 Conclusions JHEP04(2015)078 S1 S2 ϕ Ξ1 ω1 ω2 ω4 Π1 Π7 ζ Ω1 Ω2 Ω4 Υ Φ Bµ ±α(1) ll ±αee ±αle ±α(1) ll ±α(1) lq,ud ±α(1) dd ±αed ±αld ±αlu,qe ±α(1) lq ±α(1) ud ±α(1) dd ±α(1) uu ±α(1) qq ±α(1) qu,qd ±α(1) qu,qd ±αeu ±α(1) uu ±α(1) qq ±α(1) qq ±α(1) qq Wµ ±α(1) ll ±α(1) ll ±α(3) lq ±α(3) lq ±α(1) qq ±α(1) qq ±α(1) qq ±α(1) qq −α(8) qq −α(8) qq +α(8) qq +α(8) qq Gµ ±α(8) qu,qd ±α(8) ud ±α(1) dd ±α(1) uu ±α(8) qq ±α(8) qq ±α(1) dd ±α(1) uu ±α(8) qq ±α(8) qu,qd ±α(8) qq ±α(8) ud Hµ ±α(1) qq ±α(1) qq ±α(1) qq ±α(1) qq ±α(8) qq ±α(1) qq ±α(8) qq ±α(8) qq B1 µ −α(1) ud −α(1) ud +α(8) ud −α(8) ud G1 µ −α(1) ud −α(1) ud −α(8) ud +α(8) ud Lµ −αle U2 µ −α(1) lq −αed −α(1) lq +α(3) lq −α(3) lq U5 µ −αeu Q1 µ −α(1) qd −αlu −α(1) qd +α(8) qd −α(8) qd Q5 µ −α(1) qu −αld −αqe −α(1) qu +α(8) qu −α(8) qu Xµ −α(1) lq −α(1) lq −α(3) lq +α(3) lq Y1 µ −α(1) qd −α(1) qd −α(8) qd +α(8) qd Y5 µ −α(1) qu −α(1) qu −α(8) qu +α(8) qu ur-fermion operators common to new scalar and vector ﬁelds. We only show operators that receive deﬁnite sign con particle, for i = j, k = l (i ̸= k). The symbols “+” (“−”) indicate that the contributions from scalars and vectors n. We use “±” to indicate the absence of a deﬁnite sign in one of the sources. For the case of antisymmetric scala s compared with the exact same index conﬁguration as in tables 9–27. – 19 – couplings. We have written the most general renormalizable interactions of the scalar mul- tiplets (except for parts of the scalar potential that cannot be tested in the near future). Up to this point, all our results apply to either light or heavy extra scalars. In a second step, we have assumed a hierarchy of scales and have derived the dimension-six eﬀective Lagrangian that describes all the tree-level eﬀects of the heavy scalars in experiments where the probed energies are smaller than their masses. We have shown that only the 19 scalar multiplets with allowed linear interactions contribute to operators of dimension ﬁve and six. 7 Conclusions Non-linear interactions of these ﬁelds also appear in the eﬀective Lagrangian to this order. The results are collected in appendix B. Finally, we have used this eﬀective Lagrangian to discuss the observable eﬀects of the new scalars and to derive bounds on their couplings and masses. The strongest bounds come from ﬂavor observables. In order to avoid ﬂavor constraints, here we have simply assumed that, in the ﬂavor basis deﬁned by eq. (2.2), there is only one non-vanishing entry in the Yukawa couplings with the new scalar. We have then studied the limits from a range of ﬂavor-conserving observables: EWPD, LHC dilepton and dijet searches and Higgs-mediated cross sections. JHEP04(2015)078 Together with refs. [5–7], this paper provides a complete classiﬁcation of all the particles with up to dimension-four linear couplings (in the electroweak symmetric phase) to the SM ﬁelds. Even if our emphasis in this paper has been on indirect eﬀects, let us stress that this classiﬁcation and the general interactions that are explicitly written in these references provide a useful basis for model-independent direct searches at large colliders (see e.g. [12, 50–56] for applications of this gauge-invariant formalism to direct searches). On top of this, our results here complete the tree-level dictionary between particles with general couplings of dimension ≤4 and the eﬀective operators that describe their low- energy eﬀects. The dictionary entries for quarks, leptons and vector bosons can be found in refs. [5–7], respectively. We believe this correspondence can prove useful in combining the information from LHC searches of new particles with the existing precision constraints on their masses and couplings. Acknowledgments We thank L. Silvestrini and A. Weiler for useful discussions. The work of JB has been sup- ported by the European Research Council under the European Union’s Seventh Framework Programme (FP/2007-2013)/ERC Grant Agreement n. 279972. The work of MC, MPV and JS has been partially supported by the European Commission through the contract PITN- GA-2012-316704 (HIGGSTOOLS), by MINECO, under grant numbers FPA2010-17915 and FPA2013-47836-C3-2-P, and by the Junta de Andaluc´ıa grants FQM 101 and FQM 6552. A Basis of dimension-six operators In this appendix, we introduce a complete set of gauge-invariant operators Oi, which enter the general SM eﬀective Lagrangian to dimension six: L(6) Eﬀ= LSM + 1 ΛL5 + 1 Λ2 L6, with Ld = X i αiOi. – 20 – Operator Notation Operator Notation Dim 4 φ†φ 2 Oφ4 lL φ eR Oye qL ˜φ uR Oyu qL φ dR Oyd Dim 5 lc L ˜φ∗˜φ†lL O5 Table 6. Operators of dimension four and ﬁve. Table 6. Operators of dimension four and ﬁve. JHEP04(2015)078 We employ the basis in tables 6, 7 and 8. Table 6 deﬁnes our notation for the eﬀec- tive operators renormalizing the SM interactions, and presents the unique dimension-ﬁve interaction: the Weinberg operator, which gives Majorana masses to the SM neutrinos. Tables 7 and 8 gather the dimension-six operators. In these tables, TA = 1 2λA and fABC, A, B, C = 1, . . . , 8, are the SU(3)c generators and structure constants, with λA the Gell- Mann matrices; ϵABC (εabc) , A, B, C = 1, 2, 3 (a, b, c = 1, 2, 3) is the totally antisym- metric tensor in color (weak isospin) indices; σa, a = 1, 2, 3 are the Pauli matrices; and eAµν = 1 2εµνρσAρσ is the Hodge-dual of the ﬁeld strength Aµν. Finally, the superscript symbol “T” denotes transposition of the SU(2)L indices exclusively. We employ the basis in tables 6, 7 and 8. Table 6 deﬁnes our notation for the eﬀec- tive operators renormalizing the SM interactions, and presents the unique dimension-ﬁve interaction: the Weinberg operator, which gives Majorana masses to the SM neutrinos. Tables 7 and 8 gather the dimension-six operators. In these tables, TA = 1 2λA and fABC, A, B, C = 1, . . . , 8, are the SU(3)c generators and structure constants, with λA the Gell- Mann matrices; ϵABC (εabc) , A, B, C = 1, 2, 3 (a, b, c = 1, 2, 3) is the totally antisym- metric tensor in color (weak isospin) indices; σa, a = 1, 2, 3 are the Pauli matrices; and eAµν = 1 2εµνρσAρσ is the Hodge-dual of the ﬁeld strength Aµν. Finally, the superscript symbol “T” denotes transposition of the SU(2)L indices exclusively. We use essentially the same basis as the one introduced in ref. [57]. A Basis of dimension-six operators (See [58–62] for a related discussion of dimension-six physics in diﬀerent operator bases.) The only diﬀerences (apart from changes in the names) are the use of diﬀerent normalization factors in several operators, and the trade of their operators Q(3) qq = (qLγµσaqL) (qLγµσaqL) and Q(3) lequ = lLσµνeR iσ2 (qLσµνuR)T by O(8) qq and Oluqe, respectively, in our tables. Also, for consis- tency with previous works we write here the operators O(1) φψ = φ†iDµφ ψγµψ and O(3) φψ = φ†iσaDµφ ψLγµσaψL , instead of the hermitian interactions Q(1) φψ = φ†i ↔ Dµφ ψγµψ and Q(1) φψ= φ†i ↔ D a µ φ ψLγµσaψL of ref. [57]. Note that these last interactions are not gen- erated in the integration of the new scalars, and are introduced here only for completeness. Finally, for the purpose of comparing the results of the integration of new scalars with those obtained for new fermions and vector bosons in refs. [5–7], we provide below the necessary relations to translate the results in those references, which use the original basis of [58, 59] and therefore contains redundant interactions, into our basis. Again, we use the notation Qi to refer to the operator basis in other references, while we keep Oi for the operators presented in tables 7 and 8. A Basis of dimension-six operators In the sector of four-fermion interactions the following identities follow from the cor- responding Fierz reorderings: Q(3) ll ijkl = 1 2(li Lγµσalj L)(lk Lγµσall L) = 2 O(1) ll ilkj − O(1) ll ijkl , Q(1,3) qq ijkl = 1 2(qi Lγµσaqj L)(qk Lγµσaql L) = − O(1) qq ijkl + 2 3 O(1) qq ilkj + 4 O(8) qq ilkj , – 21 – Operator Notation Operator Notation LLLL 1 2 lLγµlL lLγµlL O(1) ll 1 2 (qLγµqL) (qLγµqL) O(1) qq 1 2 (qLγµTAqL) (qLγµTAqL) O(8) qq lLγµlL (qLγµqL) O(1) lq lLγµσalL (qLγµσaqL) O(3) lq RRRR 1 2 (eRγµeR) (eRγµeR) Oee 1 2 (uRγµuR) (uRγµuR) O(1) uu 1 2 dRγµdR dRγµdR O(1) dd (uRγµuR) dRγµdR O(1) ud (uRγµTAuR) dRγµTAdR O(8) ud (eRγµeR) (uRγµuR) Oeu (eRγµeR) dRγµdR Oed LLRR & LRRL lLγµlL (eRγµeR) Ole (qLγµqL) (eRγµeR) Oqe lLγµlL (uRγµuR) Olu lLγµlL dRγµdR Old (qLγµqL) (uRγµuR) O(1) qu (qLγµTAqL) (uRγµTAuR) O(8) qu (qLγµqL) dRγµdR O(1) qd (qLγµTAqL) dRγµTAdR O(8) qd lLeR dRqL Oledq LRLR (qLuR) iσ2 (qLdR)T O(1) qud (qLTAuR) iσ2 (qLTAdR)T O(8) qud lLeR iσ2 (qLuR)T Olequ lLuR iσ2 (qLeR)T Oluqe B & L ϵABC(lLiσ2qc A L )(dB Ruc C R ) Olqdu ϵABC(qA Liσ2qc B L )(eRuc C R ) Oqqeu ϵABC(lLiσ2qc A L )(qB L iσ2qc C L ) O(1) lqqq ϵABC(uA Rdc B R )(eRuc C R ) Oudeu ϵABC(lLσaiσ2qc A L )(qB L σaiσ2qc C L ) O(3) lqqq Table 7. Basis of dimension-six operators used in our analysis: four-fermion contact interactions. Flavor indices are omitted. JHEP04(2015)078 Table 7. Basis of dimension-six operators used in our analysis: four-fermion contact interactions. Flavor indices are omitted. A Basis of dimension-six operators Secondly, using – 22 – Operator Notation Operator Notation S φ†φ □ φ†φ Oφ□ 1 3 φ†φ 3 Oφ SVF φ†iDµφ lLγµlL O(1) φl φ†σaiDµφ lLγµσalL O(3) φl φ†iDµφ (eRγµeR) O(1) φe φ†iDµφ (qLγµqL) O(1) φq φ†σaiDµφ (qLγµσaqL) O(3) φq φ†iDµφ (uRγµuR) O(1) φu φ†iDµφ dRγµdR O(1) φd φT iσ2iDµφ (uRγµdR) Oφud STF lLσµνeR φ Bµν OeB lLσµνeR σaφ W a µν OeW (qLσµνuR) ˜φ Bµν OuB (qLσµνuR) σa ˜φ W a µν OuW (qLσµνdR) φ Bµν OdB (qLσµνdR) σaφ W a µν OdW (qLσµνTAuR) ˜φ GA µν OuG (qLσµνTAdR) φ GA µν OdG SF φ†φ lL φ eR Oeφ φ†φ qL ˜φ uR Ouφ φ†φ (qL φ dR) Odφ Oblique φ†Dµφ ((Dµφ)† φ) OφD φ†φ BµνBµν OφB φ†φ eBµνBµν Oφ e B φ†φ W a µνW a µν OφW φ†φ f W a µνW a µν Oφf W φ†σaφ W a µνBµν OWB φ†σaφ f W a µνBµν Of WB φ†φ GA µνGA µν OφG φ†φ eGA µνGA µν Oφ eG Gauge εabc W a ν µ W b ρ ν W c µ ρ OW εabc f W a ν µ W b ρ ν W c µ ρ Of W fABC GA ν µ GB ρ ν GC µ ρ OG fABC eGA ν µ GB ρ ν GC µ ρ O eG ble 8. Basis of dimension-six operators used in our analysis: operators other than four-fermi ntact interactions. Flavor indices are omitted. Operators in grey color do not arise in t f JHEP04(2015)078 Table 8. Basis of dimension-six operators used in our analysis: operators other than four-fermion contact interactions. Flavor indices are omitted. Operators in grey color do not arise in the integration of heavy scalars at the tree level. Table 8. Basis of dimension-six operators used in our analysis: operators other than four-fermion contact interactions. Flavor indices are omitted. Operators in grey color do not arise in the integration of heavy scalars at the tree level. a perturbative ﬁeld redeﬁnition: a perturbative ﬁeld redeﬁnition: Q(1) φ = φ†φ (Dµφ)† Dµφ = 1 2Oφ□−µ2 φOφ4 + 3λφOφ+ + 1 2 ye ii (Oeφ)ii + yd ii (Odφ)ii + V † ijyu jj (Ouφ)ij + h.c. . A Basis of dimension-six operators Q(8,1) qq ijkl = 1 2(qi LγµλAqj L)(qk LγµλAql L) = 4 O(8) qq ijkl , Q(8,3) qq ijkl = 1 2(qi LγµλAσaqj L)(qk LγµλAσaql L)= 32 9 O(1) qq ilkj−4 O(8) qq ijkl−8 3 O(8) qq ilkj , Q(8) ff ijkl = 1 2(fi RγµλAfj R)(fk RγµλAfl R) = 2 O(1) ff ilkj −2 3 O(1) ff ijkl , (f = u, d) (QFf)ijkl = (F i Lfj R)(fk RF l L) = −1 2 (OFf)ilkj , (Ff = le, lu, ld, qe) Q(1) qf ijkl = (qi Lfj R)(fk Rql L) = −1 6 O(1) qf ilkj − O(8) qf ilkj , (f = u, d) Q(8) qf ijkl = (qi LλAfj R)(fk RλAql L) = −8 9 O(1) qf ilkj + 2 3 O(8) qf ilkj . Q(8,1) qq ijkl = 1 2(qi LγµλAqj L)(qk LγµλAql L) = 4 O(8) qq ijkl , Q(8,3) qq ijkl = 1 2(qi LγµλAσaqj L)(qk LγµλAσaql L)= 32 9 O(1) qq ilkj−4 O(8) qq ijkl−8 3 O(8) qq ilkj , Q(8) ff ijkl = 1 2(fi RγµλAfj R)(fk RγµλAfl R) = 2 O(1) ff ilkj −2 3 O(1) ff ijkl , (f = u, d) (QFf)ijkl = (F i Lfj R)(fk RF l L) = −1 2 (OFf)ilkj , (Ff = le, lu, ld, qe) Q(1) qf ijkl = (qi Lfj R)(fk Rql L) = −1 6 O(1) qf ilkj − O(8) qf ilkj , (f = u, d) Q(8) qf ijkl = (qi LλAfj R)(fk RλAql L) = −8 9 O(1) qf ilkj + 2 3 O(8) qf ilkj . Finally, the operator Qqde is labeled as Oledq in table 7. There are also some diﬀerences in the case of the bosonic operators. Firstly, the operators Qφ6 and Q(3) φ correspond exactly to OφD and Oφ, respectively. B Operator coeﬃcients in the eﬀective Lagrangian In tables 9–27 we present, for each new type of scalar, the contributions to the coeﬃcients of the diﬀerent dimension-six operators that result upon integration of one scalar multiplet at the tree level. Those contributions that arise only in the case where the theory contains several scalars at the same time (in the same or diﬀerent representations) are given in – 23 – table 28. All these results are given in the basis in appendix A. In some cases, this requires performing algebraic manipulations and ﬁeld redeﬁnitions on the operators that result directly from the integration. Tables 9–28 also contain the interactions in the high-energy Lagrangian, using the notation of eq. (2.5). We only write here those interactions that are relevant for the computation of L(6) Eﬀ. When gauge indices are explicitly shown, we use the following labeling for SU(2)L indices in the diﬀerent representations: α, β = 1 2, −1 2 for SU(2)L doublets; a, b, c = 1, 2, 3 for the components of SU(2)L triplets in Cartesian coordinates; and I, J, K = 3 2, 1 2, −1 2, −3 2 for the components of the SU(2)L quadruplets. The matrices used to construct the diﬀerent invariants are the following: JHEP04(2015)078 • In constructing the triplets from two doublets we use the Pauli matrices σ1 = 0 1 1 0 ! ; σ2 = 0 −i i 0 ! ; σ3 = 1 0 0 −1 ! . • The isospin-1 product of two triplets is obtained through: • The isospin-1 product of two triplets is obtained through: fabc = i √ 2εabc. • Quadruplets are obtained from the product of an isospin-1 ﬁeld and a doublet by means of • Quadruplets are obtained from the product of an isospin-1 ﬁeld and a doublet by means of C3/2 aβ = 1 √ 2    1 0 −i 0 0 0   ; C1/2 aβ = 1 √ 6    0 1 0 −i −2 0   ; C−1/2 aβ = −1 √ 6    1 0 i 0 0 2   ; C−3/2 aβ = −1 √ 2    0 1 0 i 0 0   . B Operator coeﬃcients in the eﬀective Lagrangian C3/2 aβ = 1 √ 2    1 0 −i 0 0 0   ; C1/2 aβ = 1 √ 6    0 1 0 −i −2 0   ; C−1/2 aβ = −1 √ 6    1 0 i 0 0 2   ; C−3/2 aβ = −1 √ 2    0 1 0 i 0 0   . • The singlet product of two quadruplets is obtained through the SU(2) product • The singlet product of two quadruplets is obtained through the SU(2) product • The singlet product of two quadruplets is obtained through the SU(2) product ϵIJ = 1 2      0 0 0 1 0 0 −1 0 0 1 0 0 −1 0 0 0     . Finally, for SU(3)c indices, we use the following notation for the symmetric product of colored ﬁelds: 1 Finally, for SU(3)c indices, we use the following notation for the symmetric product of colored ﬁelds: ( \| \| ) 1 ψ(A\| 1 . . . ψ\|B) 2 ≡1 2 ψA 1 . . . ψB 2 + ψB 1 . . . ψA 2 . ψ(A\| 1 . . . ψ\|B) 2 ≡1 2 ψA 1 . . . ψB 2 + ψB 1 . . . ψA 2 . – 24 – S ∼(1, 1)0 VS = κS Sφ†φ + λS S2φ†φ + κS3 S3 Dimension-Four Operators αφ4 = κ2 S 2M2 S Scalar Operators αφ Λ2 = 3 κ2 S M2 S −λS M2 S + κS3κS M4 S αφ□ Λ2 = −κ2 S 2M4 S Table 9. Operator coeﬃcients arising from the integration of a S scalar ﬁeld. See table 28 for collective contributions of several multiplets. JHEP04(2015)078 Table 9. Operator coeﬃcients arising from the integration of a S scalar ﬁeld. See table 28 for collective contributions of several multiplets. S1 ∼(1, 1)1 JS1 = (yl S1)ij li Liσ2lc j L (yl S1)ij = −(yl S1)ji Four-Fermion Operators: LLLL α(1) ll ijkl Λ2 = 2 (yl S1)ik(yl † S1)lj M2 S1 Table 10. Operator coeﬃcients arising from the integration of a S1 scalar ﬁeld. Table 10. Operator coeﬃcients arising from the integration of a S1 scalar ﬁeld. B Operator coeﬃcients in the eﬀective Lagrangian S2 ∼(1, 1)2 JS2 = (ye S2)ij ei Rec j R (ye S2)ij = (ye S2)ji Four-Fermion Operators: RRRR (αee)ijkl Λ2 = (ye S2)ki(ye † S2 )jl M2 S2 Table 11. Operator coeﬃcients arising from the integration of a S2 scalar ﬁeld. Table 11. Operator coeﬃcients arising from the integration of a S2 scalar ﬁeld. – 25 – ϕ ∼(1, 2) 1 2 Jϕ = (ye ϕ)ij ei Rlj L + (yd ϕ)ij di Rqj L + (yu ϕ)ij iσ2qi L T uj R Vϕ = λϕ (ϕ†φ)(φ†φ) + h.c. Four-Fermion Operators: • LLRR (αle)ijkl Λ2 = − (ye ϕ)kj(ye † ϕ )il 2M2ϕ α(1) qd ijkl Λ2 = − (yd ϕ)kj(yd † ϕ )il 6M2ϕ α(8) qd ijkl Λ2 = 6 α(1) qd ijkl Λ2 α(1) qu ijkl Λ2 = − (yu ϕ)il(yu † ϕ )kj 6M2ϕ α(8) qu ijkl Λ2 = 6 α(1) qu ijkl Λ2 • LRRL • LRLR (αledq)ijkl Λ2 = (yd ϕ)kl(ye † ϕ )ij M2ϕ (αlequ)ijkl Λ2 = (yu ϕ)kl(ye † ϕ )ij M2ϕ α(1) qud ijkl Λ2 = − (yu ϕ)ij(yd † ϕ )kl M2ϕ Scalar-Fermion Operators (αeφ)ij Λ2 = λϕ(ye † ϕ )ij M2ϕ (αuφ)ij Λ2 = − λ∗ ϕ(yu ϕ)ij M2ϕ (αdφ)ij Λ2 = λϕ(yd † ϕ )ij M2ϕ Scalar Operators αφ Λ2 = 3 \|λϕ\|2 M2ϕ Table 12. Operator coeﬃcients arising from the integration of a ϕ scalar ﬁeld. See table 28 for collective contributions of several multiplets. JHEP04(2015)078 Scalar-Fermion Operators Table 12. Operator coeﬃcients arising from the integration of a ϕ scalar ﬁeld. See table 28 for collective contributions of several multiplets. Table 12. Operator coeﬃcients arising from the integration of a ϕ scalar ﬁeld. See table 28 for collective contributions of several multiplets. Ξ0 ∼(1, 3)0 VΞ0 = κΞ0 φ†Ξa 0σaφ + λΞ0 (Ξa 0Ξa 0) φ†φ Dimension-Four Operators αφ4 = κ2 Ξ0 2M2 Ξ0 1 −4 µ2 φ M2 Ξ0 Scalar-Fermion Operators (αeφ)ij Λ2 = κ2 Ξ0ye ii M4 Ξ0 δij (αuφ)ij Λ2 = κ2 Ξ0V † ijyu jj M4 Ξ0 (αdφ)ij Λ2 = κ2 Ξ0yd ii M4 Ξ0 δij Oblique Operators Scalar Operators αφD Λ2 = −2 κ2 Ξ0 M4 Ξ0 αφ□ Λ2 = κ2 Ξ0 2M4 Ξ0 αφ Λ2 = −3 κ2 Ξ0 M4 Ξ0 (λΞ0 −2λφ) Table 13. B Operator coeﬃcients in the eﬀective Lagrangian Operator coeﬃcients arising from the integration of a Ξ0 scalar ﬁeld. See table 28 for collective contributions of several multiplets. Oblique Operators Table 13. Operator coeﬃcients arising from the integration of a Ξ0 scalar ﬁeld. See table 28 for collective contributions of several multiplets. – 26 – Ξ1 ∼(1, 3)1 JΞ1 = (yl Ξ1)ij li Lσaiσ2lc j L (yl Ξ1)ij = (yl Ξ1)ji VΞ1 = κΞ1 Ξa † 1 ˜φ†σaφ + h.c. + λΞ1 Ξa † 1 Ξa 1 φ†φ + ˜λΞ1 fabc Ξa † 1 Ξb 1 φ†σcφ Dimension Four and Five Operators αφ4 = 2\|κΞ1\| 2 M2 Ξ1 1 −2 µ2 φ M2 Ξ1 (α5)ij Λ = −2 κΞ1 yl † Ξ1 ij M2 Ξ1 Four-Fermion Operators: LLLL α(1) ll ijkl Λ2 = 2 (yl Ξ1)ki(yl † Ξ1)jl M2 Ξ1 Scalar-Fermion Operators (αeφ)ij Λ2 = 2\|κΞ1\| 2ye ii M4 Ξ1 δij (αuφ)ij Λ2 = 2\|κΞ1\| 2V † ijyu jj M4 Ξ1 (αdφ)ij Λ2 = 2\|κΞ1\| 2yd ii M4 Ξ1 δij Oblique Operators Scalar Operators αφD Λ2 = 4\|κΞ1\| 2 M4 Ξ1 αφ□ Λ2 = 2\|κΞ1\| 2 M4 Ξ1 αφ Λ2 = −3\|κΞ1\| 2 M4 Ξ1 2λΞ1 − √ 2˜λΞ1 −4λφ Table 14. Operator coeﬃcients arising from the integration of a Ξ1 scalar ﬁeld. See table 28 for collective contributions of several multiplets. Dimension Four and Five Operators JHEP04(2015)078 Scalar-Fermion Operators Scalar Operators 2 Oblique Operators Table 14. Operator coeﬃcients arising from the integration of a Ξ1 scalar ﬁeld. See table 28 for collective contributions of several multiplets. Table 14. Operator coeﬃcients arising from the integration of a Ξ1 scalar ﬁeld. See table 28 for collective contributions of several multiplets. Θ1 ∼(1, 4) 1 2 VΘ1 = λΘ1 φ†σaφ CI aβ ˜φβϵIJΘJ 1 + h.c. Scalar Operators αφ Λ2 = 1 2 \|λΘ1\| 2 M2 Θ1 Table 15. Operator coeﬃcients arising from the integration of a Θ1 scalar ﬁeld. See table 28 for collective contributions of several multiplets. Table 15. Operator coeﬃcients arising from the integration of a Θ1 scalar ﬁeld. See table 28 f collective contributions of several multiplets. Θ3 ∼(1, 4) 3 2 VΘ3 = λΘ3 φ†σa ˜φ CI aβ ˜φβϵIJΘJ 3 + h.c. Scalar Operators αφ Λ2 = 3 2 \|λΘ3\| 2 M2 Θ3 Table 16. B Operator coeﬃcients in the eﬀective Lagrangian Operator coeﬃcients arising from the integration of a Θ3 scalar ﬁeld. See table 28 for collective contributions of several multiplets. Θ3 ∼(1, 4) 3 2 VΘ3 = λΘ3 φ†σa ˜φ CI aβ ˜φβϵIJΘJ 3 + h.c. Scalar Operators αφ Λ2 = 3 2 \|λΘ3\| 2 M2 Θ3 Table 16. Operator coeﬃcients arising from the integration of a Θ3 scalar ﬁeld. See table 28 for collective contributions of several multiplets. Table 16. Operator coeﬃcients arising from the integration of a Θ3 scalar ﬁeld. See table 28 for collective contributions of several multiplets. – 27 – ω1 ∼(3, 1)−1 3 Jω1 = (yql ω1)ij qc i L iσ2lj L + (yqq ω1)ij εABC qi B L iσ2qc j C L + (yeu ω1)ij ec i R uj R + (ydu ω1)ij εABC di B R uc j C R ((yqq ω1)ij = (yqq ω1)ji) ω1 ∼(3, 1)−1 3 Jω1 = (yql ω1)ij qc i L iσ2lj L + (yqq ω1)ij εABC qi B L iσ2qc j C L + (yeu ω1)ij ec i R uj R + (ydu ω1)ij εABC di B R uc j C R qq ) ( qq ) ) Four-Fermion Operators: Four-Fermion Operators: Four-Fermion Operators: • LLLL • RRRR α(1) lq ijkl Λ2 = 1 4 (yql ω1)lj(yql † ω1 )ik M2ω1 α(1) ud ijkl Λ2 = 1 3 (ydu ω1)ki(ydu † ω1 )jl M2ω1 α(3) lq ijkl Λ2 = − α(1) lq ijkl Λ2 α(8) ud ijkl Λ2 = −3 α(1) ud ijkl Λ2 α(1) qq ijkl Λ2 = 4 3 (yqq ω1)ki(yqq † ω1 )jl M2ω1 (αeu)ijkl Λ2 = 1 2 (yeu ω1)jl(yeu † ω1 )ki M2ω1 α(8) qq ijkl Λ2 = −3 α(1) qq ijkl Λ2 • LRLR • B & L α(1) qud ijkl Λ2 = 4 3 (yqq ω1)ki(ydu † ω1 )jl M2ω1 (αlqdu)ijkl Λ2 = − (ydu ω1)kl(yql † ω1 )ij M2ω1 α(8) qud ijkl Λ2 = −3 α(1) qud ijkl Λ2 (αqqeu)ijkl Λ2 = (yqq ω1)ji(yeu † ω1 )lk M2ω1 (αluqe)ijkl Λ2 = (yeu ω1)lj(yql † ω1 )ik M2ω1 α(1) lqqq ijkl Λ2 = − (yql † ω1 )ij(yqq ω1)kl M2ω1 (αlequ)ijkl Λ2 = (αluqe)ilkj Λ2 (αudeu)ijkl Λ2 = − (yeu † ω1 )lk(ydu ω1)ji M2ω1 Table 17. • LLLL α(1) lq ijkl Λ2 = 1 4 (yql ω1)lj(yql † ω1 )ik M2ω1 α(3) lq ijkl Λ2 = − α(1) lq ijkl Λ2 α(1) qq ijkl Λ2 = 4 3 (yqq ω1)ki(yqq † ω1 )jl M2ω1 α(8) qq ijkl Λ2 = −3 α(1) qq ijkl Λ2 LRLR B Operator coeﬃcients in the eﬀective Lagrangian Operator coeﬃcients arising from the integration of a ω1 scalar ﬁeld. • LLLL • LLLL α(1) lq ijkl Λ2 = 1 4 (yql ω1)lj(yql † ω1 )ik M2ω1 α(3) lq ijkl Λ2 = − α(1) lq ijkl Λ2 α(1) qq ijkl Λ2 = 4 3 (yqq ω1)ki(yqq † ω1 )jl M2ω1 α(8) qq ijkl Λ2 = −3 α(1) qq ijkl Λ2 LRLR α(1) ud ijkl Λ2 = 1 3 (ydu ω1)ki(ydu † ω1 )jl M2ω1 α(8) ud ijkl Λ2 = −3 α(1) ud ijkl Λ2 (αeu)ijkl Λ2 = 1 2 (yeu ω1)jl(yeu † ω1 )ki M2ω1 JHEP04(2015)078 Table 17. Operator coeﬃcients arising from the integration of a ω1 scalar ﬁeld. Table 17. Operator coeﬃcients arising from the integration of a ω1 scalar ﬁeld. ω2 ∼(3, 1) 2 3 Jω2 = (yd ω2)ij εABC di B R dc j C R (yd ω2)ij = −(yd ω2)ji Four-Fermion Operators: RRRR α(1) dd ijkl Λ2 = 2 (yd ω2)ki(yd † ω2 )jl M2ω2 Table 18. Operator coeﬃcients arising from the integration of a ω2 scalar ﬁe Table 18. Operator coeﬃcients arising from the integration of a ω2 scalar ﬁeld. ω4 ∼(3, 1)−4 3 Jω4 = (yed ω4)ij ec i R dj R + (yuu ω4 )ij εABC ui B R uc j C R (yuu ω4 )ij = −(yuu ω4 )ji Four-Fermion Operators: • RRRR (αed)ijkl Λ2 = (yed ω4)jl(yed † ω4 )ki 2M2ω4 α(1) uu ijkl Λ2 = 2 (yuu ω4 )ki(yuu † ω4 )jl M2ω4 • B & L (αudeu)ijkl Λ2 = 2 (yuu ω4 )il(yed † ω4 )jk M2ω4 Table 19 Operator coeﬃcients arising from the integration of a ω4 scalar ﬁeld Four-Fermion Operators: Table 19. Operator coeﬃcients arising from the integration of a ω4 scalar ﬁeld. – 28 – Π1 ∼(3, 2) 1 6 JΠ1 = (yld Π1)ij iσ2li L T dj R Four-Fermion Operators: LLRR (αld)ijkl Λ2 = − (yld Π1)il(yld † Π1 )kj 2M2 Π1 Table 20. Operator coeﬃcients arising from the integration of a Π1 scalar ﬁeld. Table 20. Operator coeﬃcients arising from the integration of a Π1 scalar ﬁeld. • LLLL JHEP04(2015)078 Π7 ∼(3, 2) 7 6 JΠ7 = (ylu Π7)ij iσ2li L T uj R + (yeq Π7)ij ei Rqj L Four-Fermion Operators: • LLRR • LRLR (αlu)ijkl Λ2 = − (ylu Π7)il(ylu † Π7 )kj 2M2 Π7 (αluqe)ijkl Λ2 = − (ylu Π7)ij(yeq † Π7 )kl M2 Π7 (αqe)ijkl Λ2 = − (yeq Π7)kj(yeq † Π7 )il 2M2 Π7 Table 21. Operator coeﬃcients arising from the integration of a Π7 scalar ﬁeld. ζ ∼(3, 3)−1 3 Jζ = (yql ζ )ij qc i L iσ2σalj L + (yqq ζ )ij εABC qi B L σaiσ2qc j C L (yqq ζ )ij = −(yqq ζ )ji Four-Fermion Operators: • LLLL α(1) lq ijkl Λ2 = 3 4 (yql ζ )lj(yql † ζ )ik M2 ζ α(3) lq ijkl Λ2 = 1 3 α(1) lq ijkl Λ2 α(1) qq ijkl Λ2 = 4 3 (yqq ζ )ki(yqq † ζ )jl M2 ζ α(8) qq ijkl Λ2 = −3 α(1) qq ijkl Λ2 • B & L α(3) lqqq ijkl Λ2 = − (yqq ζ )kl(yql † ζ )ij M2 ζ ζ ∼(3, 3)−1 3 Jζ = (yql ζ )ij qc i L iσ2σalj L + (yqq ζ )ij εABC qi B L σaiσ2qc j C L (yqq ζ )ij = −(yqq ζ )ji Four-Fermion Operators: • LLLL α(1) lq ijkl Λ2 = 3 4 (yql ζ )lj(yql † ζ )ik M2 ζ α(3) lq ijkl Λ2 = 1 3 α(1) lq ijkl Λ2 α(1) qq ijkl Λ2 = 4 3 (yqq ζ )ki(yqq † ζ )jl M2 ζ α(8) qq ijkl Λ2 = −3 α(1) qq ijkl Λ2 • B & L α(3) lqqq ijkl Λ2 = − (yqq ζ )kl(yql † ζ )ij M2 ζ Table 22. Operator coeﬃcients arising from the integration of a ζ scalar ﬁeld. • LLLL ζ ∼(3, 3)−1 3 Jζ = (yql ζ )ij qc i L iσ2σalj L + (yqq ζ )ij εABC qi B L σaiσ2qc j C L (yqq ζ )ij = −(yqq ζ )ji Four-Fermion Operators: Four-Fermion Operators: • LLLL • LLLL α(1) lq ijkl Λ2 = 3 4 (yql ζ )lj(yql † ζ )ik M2 ζ α(3) lq ijkl Λ2 = 1 3 α(1) lq ijkl Λ2 α(1) qq ijkl Λ2 = 4 3 (yqq ζ )ki(yqq † ζ )jl M2 ζ α(8) qq ijkl Λ2 = −3 α(1) qq ijkl Λ2 • B & L α(3) lqqq ijkl Λ2 = − (yqq ζ )kl(yql † ζ )ij M2 ζ Table 22. Operator coeﬃcients arising from the integration of a ζ scalar ﬁeld. – 29 – Ω1 ∼(6, 1) 1 3 JΩ1 = (yud Ω1)ij uc i (A\| R dj \|B) R + (yqq Ω1)ij qc i (A\| L iσ2qj \|B) L (yqq Ω1)ij = −(yqq Ω1)ji Four-Fermion Operators: • LLLL • RRRR α(1) qq ijkl Λ2 = 2 3 (yqq Ω1)jl(yqq † Ω1 )ki M2 Ω1 α(1) ud ijkl Λ2 = 1 3 (yud Ω1)jl(yud † Ω1 )ki M2 Ω1 α(8) qq ijkl Λ2 = 3 2 α(1) qq ijkl Λ2 α(8) ud ijkl Λ2 = 3 2 α(1) ud ijkl Λ2 • LRLR α(1) qud ijkl Λ2 = 4 3 (yud Ω1)jl(yqq † Ω1 )ik M2 Ω1 α(8) qud ijkl Λ2 = 3 2 α(1) qud ijkl Λ2 Table 23. Operator coeﬃcients arising from the integration of a Ω1 scalar ﬁeld. Four-Fermion Operators: JHEP04(2015)078 JHEP04(2015)078 Table 23. Operator coeﬃcients arising from the integration of a Ω1 scalar ﬁeld. Table 23. Operator coeﬃcients arising from the integration of a Ω1 scalar ﬁeld. Ω2 ∼(6, 1)−2 3 JΩ2 = (yd Ω2)ij dc i (A\| R dj \|B) R (yd Ω2)ij = (yd Ω2)ji Four-Fermion Operators: RRRR α(1) dd ijkl Λ2 = (yd Ω2)jl(yd † Ω2 )ki M2 Ω2 Table 24. Operator coeﬃcients arising from the integration of a Ω2 scalar ﬁeld. Table 24. Operator coeﬃcients arising from the integration of a Ω2 scalar ﬁeld. • LLLL −λij S SiSj φ†φ − λi ϕ ϕ† iφ φ†φ + h.c. −λij Ξ0Ξa 0iΞa 0j φ†φ −˜λij Ξ0Ξa 0iΞb 0jfabc φ†σcφ −λij Ξ1Ξa † 1i Ξa 1j φ†φ −˜λij Ξ1fabcΞa † 1i Ξb 1j φ†σcφ −λij SΞ0SiΞa 0j φ†σaφ − λij SΞ1SiΞa † 1j ˜φ†σaφ + λij Ξ1Ξ0fabcΞa † 1i Ξb 0j ˜φ†σcφ + h.c. − λi Θ1 φ†σaφ CI aβ ˜φβϵIJΘJ 1i + λi Θ3 φ†σa ˜φ CI aβ ˜φβϵIJΘJ 3i + h.c. Mixed contributions from {S, ϕ, Ξ0, Ξ1, Θ1, Θ3} JHEP04(2015)078 Scalar Operators p αφ Λ2 = 3 M2ϕj λj ϕ − κij Sϕκi S M2 Si − κij Ξ0ϕκi Ξ0 M2 Ξ0i −2 (κij Ξ1ϕ)∗κi Ξ1 M2 Ξ1i 2 + 1 2M2 Θ1i λi Θ1 − κj Ξ0κji Ξ0Θ1 M2 Ξ0j − (κj Ξ1)∗κji Ξ1Θ1 M2 Ξ1j 2 + 3 2M2 Θ3i λi Θ3 − (κj Ξ1)∗κji Ξ1Θ3 M2 Ξ1j 2 −3κi S M2 Si λij S κj S M2 Sj + λij SΞ0κj Ξ0 M2 Ξ0j + 4 Re n λij SΞ1(κj Ξ1)∗o M2 Ξ1j ! −3 λij Ξ0κi Ξ0κj Ξ0 M2 Ξ0iM2 Ξ0j −3 (κi Ξ1)∗κj Ξ1 M2 Ξ1iM2 Ξ1j 2λij Ξ1 − √ 2˜λij Ξ1 −6 √ 2 Re n λij Ξ1Ξ0(κi Ξ1)∗κj Ξ0 o M2 Ξ1iM2 Ξ0j −3 √ 2 κijk Ξ0Ξ1κi Ξ0(κj Ξ1)∗κk Ξ1 M2 Ξ0iM2 Ξ1j M2 Ξ1k +3 κi S M2 S κijk S κj Sκk S M2 S M2 S + κijk SΞ0κj Ξ0κk Ξ0 M2 Ξ M2 Ξ + 2 κijk SΞ1(κj Ξ1)∗κk Ξ1 M2 Ξ M2 Ξ Scalar-Fermion Operators (αeφ)ij Λ2 = 1 M2ϕj λj ϕ − κi Sκij Sϕ M2 Si − κi Ξ0κij Ξ0ϕ M2 Ξ0i −2 κi Ξ1 κij Ξ1ϕ ∗ M2 Ξ1i ! ye † ϕ ij (αdφ)ij Λ2 = 1 M2ϕj λj ϕ − κi Sκij Sϕ M2 Si − κi Ξ0κij Ξ0ϕ M2 Ξ0i −2 κi Ξ1 κij Ξ1ϕ ∗ M2 Ξ1i ! yd † ϕ ij (αuφ)ij Λ2 = − 1 M2ϕj (λj ϕ)∗− κi S(κij Sϕ) ∗ M2 Si − κi Ξ0 κij Ξ0ϕ ∗ M2 Ξ0i −2 κi Ξ1 ∗ κij Ξ1ϕ M2 Ξ1i ! yu ϕ ij Table 28. • LLLL Ω4 ∼(6, 1) 4 3 JΩ4 = (yu Ω4)ij uc i (A\| R uj \|B) R (yu Ω4)ij = (yu Ω4)ji Four-Fermion Operators: RRRR α(1) uu ijkl Λ2 = (yu Ω4)jl(yu † Ω4 )ki M2 Ω4 Table 25. Operator coeﬃcients arising from the integration of a Ω4 scalar ﬁeld. Table 25. Operator coeﬃcients arising from the integration of a Ω4 scalar ﬁeld. – 30 – Υ ∼(6, 3) 1 3 JΥ = (yq Υ)ij qc i (A\| L iσ2σaqj \|B) L (yq Υ)ij = (yq Υ)ji Four-Fermion Operators: LLLL α(1) qq ijkl Λ2 = 4 3 (yq Υ)lj(yq † Υ )ik M2 Υ α(8) qq ijkl Λ2 = 3 2 α(1) qq ijkl Λ2 Table 26. Operator coeﬃcients arising from the integration of a Υ scalar ﬁeld. JHEP04(2015)078 Table 26. Operator coeﬃcients arising from the integration of a Υ scalar ﬁeld. Φ ∼(8, 2) 1 2 JΦ = (yqu Φ )ij iσ2qi L T TAuj R + (ydq Φ )ij di RTAqj L Four-Fermion Operators: • LLRR • LRLR α(1) qu ijkl Λ2 = −2 9 (yqu Φ )il(yqu † Φ )kj M2 Φ α(8) qud ijkl Λ2 = −(yqu Φ )ij(ydq † Φ )kl M2 Φ α(8) qu ijkl Λ2 = −3 4 α(1) qu ijkl Λ2 α(1) qd ijkl Λ2 = −2 9 (ydq Φ )kj(ydq † Φ )il M2 Φ α(8) qd ijkl Λ2 = −3 4 α(1) qd ijkl Λ2 Table 27. Operator coeﬃcients arising from the integration of a Φ scalar ﬁeld. Table 27. Operator coeﬃcients arising from the integration of a Φ scalar ﬁeld. Table 27. Operator coeﬃcients arising from the integration of a Φ scalar ﬁeld. – 31 – Mixed contributions from {S, ϕ, Ξ0, Ξ1, Θ1, Θ3} ∆Lint = − ϕ† iJϕi + Ξa † 1i Ja Ξ1i + h.c. −κi SSiφ†φ −κijk S3 SiSjSk −κi Ξ0Ξa 0iφ†σaφ − κi Ξ1Ξa † 1i ˜φ†σaφ + κij SϕSiϕ† jφ + h.c. −κijk SΞ0SiΞa 0jΞa 0k −κijk SΞ1SiΞa † 1j Ξa 1k −κijk Ξ3 0 fabcΞa 0iΞb 0jΞc 0k −κijk Ξ0Ξ1fabcΞa 0iΞb † 1j Ξc 1k − κij Ξ0ϕΞa 0i ϕ† jσaφ + κij Ξ1ϕΞa † 1i ˜ϕ† jσaφ + h.c. • LLLL − κij Ξ0Θ1Ξa 0iCI aβ ˜φβϵIJΘJ 1j + κij Ξ1Θ1Ξa † 1i CI aβφβϵIJΘJ 1j + κij Ξ1Θ3Ξa † 1i CI aβ ˜φβϵIJΘJ 3j + h.c. −λij S SiSj φ†φ − λi ϕ ϕ† iφ φ†φ + h.c. −λij Ξ0Ξa 0iΞa 0j φ†φ −˜λij Ξ0Ξa 0iΞb 0jfabc φ†σcφ −λij Ξ1Ξa † 1i Ξa 1j φ†φ −˜λij Ξ1fabcΞa † 1i Ξb 1j φ†σcφ −λij SΞ0SiΞa 0j φ†σaφ − λij SΞ1SiΞa † 1j ˜φ†σaφ + λij Ξ1Ξ0fabcΞa † 1i Ξb 0j ˜φ†σcφ + h.c. − λi Θ1 φ†σaφ CI aβ ˜φβϵIJΘJ 1i + λi Θ3 φ†σa ˜φ CI aβ ˜φβϵIJΘJ 3i + h.c. Scalar Operators αφ Λ2 = 3 M2ϕj λj ϕ − κij Sϕκi S M2 Si − κij Ξ0ϕκi Ξ0 M2 Ξ0i −2 (κij Ξ1ϕ)∗κi Ξ1 M2 Ξ1i 2 + 1 2M2 Θ1i λi Θ1 − κj Ξ0κji Ξ0Θ1 M2 Ξ0j − (κj Ξ1)∗κji Ξ1Θ1 M2 Ξ1j 2 + 3 2M2 Θ3i λi Θ3 − (κj Ξ1)∗κji Ξ1Θ3 M2 Ξ1j 2 −3κi S M2 Si λij S κj S M2 Sj + λij SΞ0κj Ξ0 M2 Ξ0j + 4 Re n λij SΞ1(κj Ξ1)∗o M2 Ξ1j ! −3 λij Ξ0κi Ξ0κj Ξ0 M2 Ξ0iM2 Ξ0j −3 (κi Ξ1)∗κj Ξ1 M2 Ξ1iM2 Ξ1j 2λij Ξ1 − √ 2˜λij Ξ1 −6 √ 2 Re n λij Ξ1Ξ0(κi Ξ1)∗κj Ξ0 o M2 Ξ1iM2 Ξ0j −3 √ 2 κijk Ξ0Ξ1κi Ξ0(κj Ξ1)∗κk Ξ1 M2 Ξ0iM2 Ξ1j M2 Ξ1k +3 κi S M2 Si κijk S κj Sκk S M2 Sj M2 Sk + κijk SΞ0κj Ξ0κk Ξ0 M2 Ξ0j M2 Ξ0k + 2 κijk SΞ1(κj Ξ1)∗κk Ξ1 M2 Ξ1j M2 Ξ1k Scalar-Fermion Operators (αeφ)ij Λ2 = 1 M2ϕj λj ϕ − κi Sκij Sϕ M2 Si − κi Ξ0κij Ξ0ϕ M2 Ξ0i −2 κi Ξ1 κij Ξ1ϕ ∗ M2 Ξ1i ! ye † ϕ ij (αdφ)ij Λ2 = 1 M2ϕj λj ϕ − κi Sκij Sϕ M2 Si − κi Ξ0κij Ξ0ϕ M2 Ξ0i −2 κi Ξ1 κij Ξ1ϕ ∗ M2 Ξ1i ! • LLLL yd † ϕ ij (αuφ)ij Λ2 = − 1 M2ϕj (λj ϕ)∗− κi S(κij Sϕ) ∗ M2 Si − κi Ξ0 κij Ξ0ϕ ∗ M2 Ξ0i −2 κi Ξ1 ∗ κij Ξ1ϕ M2 Ξ1i ! yu ϕ ij Table 28. Collective contributions from several scalars to the dimension six eﬀective Lagrangian. Mixed contributions from {S, ϕ, Ξ0, Ξ1, Θ1, Θ3} ∆Lint = − ϕ† iJϕi + Ξa † 1i Ja Ξ1i + h.c. −κi SSiφ†φ −κijk S3 SiSjSk −κi Ξ0Ξa 0iφ†σaφ − κi Ξ1Ξa † 1i ˜φ†σaφ + κij SϕSiϕ† jφ + h.c. −κijk SΞ0SiΞa 0jΞa 0k −κijk SΞ1SiΞa † 1j Ξa 1k −κijk Ξ3 0 fabcΞa 0iΞb 0jΞc 0k −κijk Ξ0Ξ1fabcΞa 0iΞb † 1j Ξc 1k − κij Ξ0ϕΞa 0i ϕ† jσaφ + κij Ξ1ϕΞa † 1i ˜ϕ† jσaφ + h.c. − κij Ξ0Θ1Ξa 0iCI aβ ˜φβϵIJΘJ 1j + κij Ξ1Θ1Ξa † 1i CI aβφβϵIJΘJ 1j + κij Ξ1Θ3Ξa † 1i CI aβ ˜φβϵIJΘJ 3j + h.c. −λij S SiSj φ†φ − λi ϕ ϕ† iφ φ†φ + h.c. −λij Ξ0Ξa 0iΞa 0j φ†φ −˜λij Ξ0Ξa 0iΞb 0jfabc φ†σcφ −λij Ξ1Ξa † 1i Ξa 1j φ†φ −˜λij Ξ1fabcΞa † 1i Ξb 1j φ†σcφ −λij SΞ0SiΞa 0j φ†σaφ − λij SΞ1SiΞa † 1j ˜φ†σaφ + λij Ξ1Ξ0fabcΞa † 1i Ξb 0j ˜φ†σcφ + h.c. − λi Θ1 φ†σaφ CI aβ ˜φβϵIJΘJ 1i + λi Θ3 φ†σa ˜φ CI aβ ˜φβϵIJΘJ 3i + h.c. Mixed contributions from {S, ϕ, Ξ0, Ξ1, Θ1, Θ3} ∆Lint = − ϕ† iJϕi + Ξa † 1i Ja Ξ1i + h.c. −κi SSiφ†φ −κijk S3 SiSjSk −κi Ξ0Ξa 0iφ†σaφ − κi Ξ1Ξa † 1i ˜φ†σaφ + κij SϕSiϕ† jφ + h.c. −κijk SΞ0SiΞa 0jΞa 0k −κijk SΞ1SiΞa † 1j Ξa 1k −κijk Ξ3 0 fabcΞa 0iΞb 0jΞc 0k −κijk Ξ0Ξ1fabcΞa 0iΞb † 1j Ξc 1k − κij Ξ0ϕΞa 0i ϕ† jσaφ + κij Ξ1ϕΞa † 1i ˜ϕ† jσaφ + h.c. − κij Ξ0Θ1Ξa 0iCI aβ ˜φβϵIJΘJ 1j + κij Ξ1Θ1Ξa † 1i CI aβφβϵIJΘJ 1j + κij Ξ1Θ3Ξa † 1i CI aβ ˜φβϵIJΘJ 3j + h.c. • LLLL Collective contributions from several scalars to the dimension six eﬀective Lagrangian. Only those operators that receive contributions from interactions involving more than one scalar at a time are shown. The full contribution to the operator coeﬃcient is presented in that case. For the other operators the contributions can be read by adding the corresponding pieces from the tables obtained integrating one scalar at a time. Note that, due to the antisymmetric properties of the interaction, the couplings κijk Ξ3 0 and ˜λij Ξ0 do not contribute to L6. – 32 – Open Access. This article is distributed under the terms of the Creative Commons Attribution License (CC-BY 4.0), which permits any use, distribution and reproduction in any medium, provided the original author(s) and source are credited. References [1] ATLAS collaboration, Observation of a new particle in the search for the Standard Model Higgs boson with the ATLAS detector at the LHC, Phys. Lett. B 716 (2012) 1 [arXiv:1207.7214] [INSPIRE]. [2] CMS collaboration, Observation of a new boson at a mass of 125 GeV with the CMS experiment at the LHC, Phys. Lett. B 716 (2012) 30 [arXiv:1207.7235] [INSPIRE]. [2] CMS collaboration, Observation of a new boson at a mass of 125 GeV with the CMS experiment at the LHC, Phys. Lett. B 716 (2012) 30 [arXiv:1207.7235] [INSPIRE]. JHEP04(2015)078 JHEP04(2015)078 [3] F. del Aguila, M. P´erez-Victoria and J. Santiago, Eﬀective description of quark mixing, Phy Lett. B 492 (2000) 98 [hep-ph/0007160] [INSPIRE]. [3] F. del Aguila, M. P´erez-Victoria and J. Santiago, Eﬀective description of quark mixing, Phys. Lett. B 492 (2000) 98 [hep-ph/0007160] [INSPIRE]. [4] A. Pomarol and F. Riva, Towards the Ultimate SM Fit to Close in on Higgs Physics, JHEP 01 (2014) 151 [arXiv:1308.2803] [INSPIRE]. [4] A. Pomarol and F. Riva, Towards the Ultimate SM Fit to Close in on Higgs Physics, JHEP 01 (2014) 151 [arXiv:1308.2803] [INSPIRE]. [5] F. del Aguila, M. P´erez-Victoria and J. Santiago, Observable contributions of new exotic quarks to quark mixing, JHEP 09 (2000) 011 [hep-ph/0007316] [INSPIRE]. [5] F. del Aguila, M. P´erez-Victoria and J. Santiago, Observable contributions of new exotic quarks to quark mixing, JHEP 09 (2000) 011 [hep-ph/0007316] [INSPIRE]. [6] F. del Aguila, J. de Blas and M. P´erez-Victoria, Eﬀects of new leptons in Electroweak Precision Data, Phys. Rev. D 78 (2008) 013010 [arXiv:0803.4008] [INSPIRE]. [6] F. del Aguila, J. de Blas and M. P´erez-Victoria, Eﬀects of new leptons in Electroweak Precision Data, Phys. Rev. D 78 (2008) 013010 [arXiv:0803.4008] [INSPIRE]. [7] F. del Aguila, J. de Blas and M. P´erez-Victoria, Electroweak Limits on General New Vector Bosons, JHEP 09 (2010) 033 [arXiv:1005.3998] [INSPIRE]. [8] V. Silveira and A. Zee, Scalar Phantoms, Phys. Lett. B 161 (1985) 136 [INSPIRE]. [9] J.F. Nieves, Baryon and Lepton Number Nonconserving Processes and Intermediate Mass Scales, Nucl. Phys. B 189 (1981) 182 [INSPIRE]. [10] W. Buchm¨uller, R. Ruckl and D. Wyler, Leptoquarks in Lepton-Quark Collisions, Phys. Lett. B 191 (1987) 442 [Erratum ibid. B 448 (1999) 320] [INSPIRE]. [11] E. Ma, M. Raidal and U. Sarkar, Probing the exotic particle content beyond the standard model, Eur. Phys. J. C 8 (1999) 301 [hep-ph/9808484] [INSPIRE]. [12] E. Del Nobile, R. Franceschini, D. References Pappadopulo and A. Strumia, Minimal Matter at the Large Hadron Collider, Nucl. Phys. B 826 (2010) 217 [arXiv:0908.1567] [INSPIRE]. [13] T. Han, I. Lewis and Z. Liu, Colored Resonant Signals at the LHC: Largest Rate and Simplest Topology, JHEP 12 (2010) 085 [arXiv:1010.4309] [INSPIRE]. [14] J.A. Aguilar-Saavedra and M. P´erez-Victoria, Probing the Tevatron t¯t asymmetry at LHC, JHEP 05 (2011) 034 [arXiv:1103.2765] [INSPIRE]. [15] B. Grinstein, A.L. Kagan, J. Zupan and M. Trott, Flavor Symmetric Sectors and Collider Physics, JHEP 10 (2011) 072 [arXiv:1108.4027] [INSPIRE]. [16] W. Skiba, TASI Lectures on Eﬀective Field Theory and Precision Electroweak Measurements, arXiv:1006.2142 [INSPIRE]. [17] Z.U. Khandker, D. Li and W. Skiba, Electroweak Corrections from Triplet Scalars, Phys. Rev. D 86 (2012) 015006 [arXiv:1201.4383] [INSPIRE]. [18] B. Henning, X. Lu and H. Murayama, How to use the Standard Model eﬀective ﬁeld theory, arXiv:1412.1837 [INSPIRE]. – 33 – [19] I. Dorsner, S. Fajfer and N. Kosnik, Heavy and light scalar leptoquarks in proton decay, Phys. Rev. D 86 (2012) 015013 [arXiv:1204.0674] [INSPIRE]. [20] F. del Aguila, M. Chala, J. Santiago and Y. Yamamoto, Collider limits on leptophilic interactions, arXiv:1411.7394 [INSPIRE]. [21] S. Weinberg, Baryon and Lepton Nonconserving Processes, Phys. Rev. Lett. 43 (1979) 1566 [INSPIRE]. [22] F. del Aguila, J.A. Aguilar-Saavedra, J. de Blas and M. Zralek, Looking for signals beyond the neutrino Standard Model, Acta Phys. Polon. B 38 (2007) 3339 [arXiv:0710.2923] [INSPIRE]. [23] F. del Aguila, J.A. Aguilar-Saavedra, J. de Blas and M. P´erez-Victoria, Electroweak constraints on see-saw messengers and their implications for LHC, arXiv:0806.1023 [INSPIRE]. JHEP04(2015)078 JHEP04(2015)078 [24] F. del Aguila, J.A. Aguilar-Saavedra and J. de Blas, New neutrino interactions at large colliders, PoS(ICHEP 2010)296 [arXiv:1012.1327] [INSPIRE]. [25] M.E. Peskin and T. Takeuchi, Estimation of oblique electroweak corrections, Phys. Rev. D 46 (1992) 381 [INSPIRE]. [26] J. de Blas et al., Global Bayesian Analysis of the Higgs-boson Couplings, arXiv:1410.4204 [INSPIRE]. [27] J.R. Ellis, V. Sanz and T. You, The Eﬀective Standard Model after LHC Run I arXiv:1410.7703 [INSPIRE]. [28] J. Baglio, A. Djouadi, R. Gr¨ober, M.M. M¨uhlleitner, J. Quevillon and M. Spira, The measurement of the Higgs self-coupling at the LHC: theoretical status, JHEP 04 (2013) 151 [arXiv:1212.5581] [INSPIRE]. [29] O. Domenech, A. Pomarol and J. Serra, Probing the SM with Dijets at the LHC, Phys. Rev. D 85 (2012) 074030 [arXiv:1201.6510] [INSPIRE]. [30] G. Isidori, Y. Nir and G. Perez, Flavor Physics Constraints for Physics Beyond the Standard Model, Ann. Rev. References Nucl. Part. Sci. 60 (2010) 355 [arXiv:1002.0900] [INSPIRE]. [31] M. Carpentier and S. Davidson, Constraints on two-lepton, two quark operators, Eur. Phys J. C 70 (2010) 1071 [arXiv:1008.0280] [INSPIRE]. [32] J. de Blas, M. Chala and J. Santiago, Global Constraints on Lepton-Quark Contact Interactions, Phys. Rev. D 88 (2013) 095011 [arXiv:1307.5068] [INSPIRE]. [33] J.A. Aguilar-Saavedra and M. P´erez-Victoria, No like-sign tops at Tevatron: Constraints on extended models and implications for the t¯t asymmetry, Phys. Lett. B 701 (2011) 93 [arXiv:1104.1385] [INSPIRE]. [34] J. de Blas, Electroweak limits on physics beyond the Standard Model, EPJ Web Conf. 60 (2013) 19008 [arXiv:1307.6173] [INSPIRE]. [35] M. Ciuchini, E. Franco, S. Mishima and L. Silvestrini, Electroweak Precision Observables, New Physics and the Nature of a 126 GeV Higgs Boson, JHEP 08 (2013) 106 [arXiv:1306.4644] [INSPIRE]. [36] ALEPH, DELPHI, L3, OPAL and SLD collaborations, LEP Electroweak Working Group, SLD Electroweak Group and SLD Heavy Flavour Group, S. Schael et al., Precision electroweak measurements on the Z resonance, Phys. Rept. 427 (2006) 257 [hep-ex/0509008] [INSPIRE]. – 34 – [37] M. Davier, A. Hoecker, B. Malaescu and Z. Zhang, Reevaluation of the Hadronic Contributions to the Muon g −2 and to α(M 2 Z), Eur. Phys. J. C 71 (2011) 1515 [Erratum ibid. C 72 (2012) 1874] [arXiv:1010.4180] [INSPIRE]. [38] Particle Data Group collaboration, K.A. Olive et al., Review of Particle Physics, Chin. Phys. C 38 (2014) 090001 [INSPIRE]. [39] ATLAS, CDF, CMS and D0 collaborations, First combination of Tevatron and LHC measurements of the top-quark mass, arXiv:1403.4427 [INSPIRE]. [40] ATLAS collaboration, Measurement of the Higgs boson mass from the H →γγ and H →ZZ∗→4ℓchannels with the ATLAS detector using 25 fb−1 of pp collision data, Phys. Rev. D 90 (2014) 052004 [arXiv:1406.3827] [INSPIRE]. JHEP04(2015)078 JHEP04(2015)078 [41] CMS collaboration, Precise determination of the mass of the Higgs boson and tests of compatibility of its couplings with the standard model predictions using proton collisions at 7 and 8 TeV, arXiv:1412.8662 [INSPIRE]. [42] CDF and D0 collaborations, T.E.W. Group, 2012 Update of the Combination of CDF and D0 Results for the Mass of the W Boson, arXiv:1204.0042 [INSPIRE]. [43] ALEPH, CDF, D0, DELPHI, L3, OPAL and SLD collaborations, LEP and Tevatron Electroweak Working Group, SLD Electroweak and Heavy Flavour Groups, Precision Electroweak Measurements and Constraints on the Standard Model, arXiv:1012.2367 [INSPIRE]. [44] ALEPH, DELPHI, L3 and OPAL collaborations, LEP Electroweak Working Group, S. References Schael et al., Electroweak Measurements in Electron-Positron Collisions at W-Boson-Pair Energies at LEP, Phys. Rept. 532 (2013) 119 [arXiv:1302.3415] [INSPIRE]. [45] J.A. Aguilar-Saavedra, Eﬀective four-fermion operators in top physics: A Roadmap, Nucl. Phys. B 843 (2011) 638 [Erratum ibid. B 851 (2011) 443] [arXiv:1008.3562] [INSPIRE]. [46] I. Dorsner, S. Fajfer and A. Greljo, Cornering Scalar Leptoquarks at LHC, JHEP 10 (2014) 154 [arXiv:1406.4831] [INSPIRE]. [47] CMS collaboration, Search for quark compositeness in dijet angular distributions from pp collisions at √s = 7 TeV, JHEP 05 (2012) 055 [arXiv:1202.5535] [INSPIRE]. [48] F. del Aguila and J. de Blas, Electroweak constraints on new physics, Fortsch. Phys. 59 (2011) 1036 [arXiv:1105.6103] [INSPIRE]. [49] F. del Aguila, J. de Blas, P. Langacker and M. P´erez-Victoria, Impact of extra particles on indirect Z′ limits, Phys. Rev. D 84 (2011) 015015 [arXiv:1104.5512] [INSPIRE]. [50] F. del Aguila and J.A. Aguilar-Saavedra, Distinguishing seesaw models at LHC with multi-lepton signals, Nucl. Phys. B 813 (2009) 22 [arXiv:0808.2468] [INSPIRE]. [51] A. Atre, M. Carena, T. Han and J. Santiago, Heavy Quarks Above the Top at the Tevatron Phys. Rev. D 79 (2009) 054018 [arXiv:0806.3966] [INSPIRE]. [52] A. Atre et al., Model-Independent Searches for New Quarks at the LHC, JHEP 08 (2011) 080 [arXiv:1102.1987] [INSPIRE]. [53] A. Atre, M. Chala and J. Santiago, Searches for New Vector Like Quarks: Higgs Channels, JHEP 05 (2013) 099 [arXiv:1302.0270] [INSPIRE]. [54] C. Grojean, E. Salvioni and R. Torre, A weakly constrained W ′ at the early LHC, JHEP 07 (2011) 002 [arXiv:1103.2761] [INSPIRE]. – 35 – [55] J. de Blas, J.M. Lizana and M. P´erez-Victoria, Combining searches of Z′ and W ′ bosons, JHEP 01 (2013) 166 [arXiv:1211.2229] [INSPIRE]. [56] J.A. Aguilar-Saavedra, R. Benbrik, S. Heinemeyer and M. P´erez-Victoria, Handbook of vectorlike quarks: Mixing and single production, Phys. Rev. D 88 (2013) 094010 [arXiv:1306.0572] [INSPIRE]. [57] B. Grzadkowski, M. Iskrzynski, M. Misiak and J. Rosiek, Dimension-Six Terms in the Standard Model Lagrangian, JHEP 10 (2010) 085 [arXiv:1008.4884] [INSPIRE]. [58] W. Buchm¨uller and D. Wyler, Eﬀective Lagrangian Analysis of New Interactions and Flavor Conservation, Nucl. Phys. B 268 (1986) 621 [INSPIRE]. [59] C. Arzt, M.B. Einhorn and J. Wudka, Patterns of deviation from the standard model, Nucl. Phys. B 433 (1995) 41 [hep-ph/9405214] [INSPIRE]. [59] C. Arzt, M.B. Einhorn and J. Wudka, Patterns of deviation from the standard model, Nucl. Phys. B 433 (1995) 41 [hep-ph/9405214] [INSPIRE]. JHEP04(2015)078 [60] R. Contino, M. Ghezzi, C. References Grojean, M. Muhlleitner and M. Spira, Eﬀective Lagrangian for a light Higgs-like scalar, JHEP 07 (2013) 035 [arXiv:1303.3876] [INSPIRE]. [60] R. Contino, M. Ghezzi, C. Grojean, M. Muhlleitner and M. Spira, Eﬀective Lagrangian for a light Higgs-like scalar, JHEP 07 (2013) 035 [arXiv:1303.3876] [INSPIRE]. [61] J. Elias-Miro, J.R. Espinosa, E. Masso and A. Pomarol, Higgs windows to new physics through D = 6 operators: constraints and one-loop anomalous dimensions, JHEP 11 (2013) 066 [arXiv:1308.1879] [INSPIRE]. [61] J. Elias-Miro, J.R. Espinosa, E. Masso and A. Pomarol, Higgs windows to new physics through D = 6 operators: constraints and one-loop anomalous dimensions, JHEP 11 (2013) 066 [arXiv:1308.1879] [INSPIRE]. [62] R.S. Gupta, A. Pomarol and F. Riva, BSM Primary Eﬀects, Phys. Rev. D 91 (2015) 035001 [arXiv:1405.0181] [INSPIRE]. – 36 –
https://openalex.org/W2237318463	https://europepmc.org/articles/pmc4711018?pdf=render	English	null	County-level heat vulnerability of urban and rural residents in Tibet, China	Environmental health	2,016	cc-by	7,388	© 2016 Bai et al. Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. * Correspondence: liuqiyong@icdc.cn 1State Key Laboratory of Infectious Disease Prevention and Control, Collaborative Innovation Center for Diagnosis and Treatment of Infectious Diseases, National Institute for Communicable Disease Control and Prevention, Chinese Center for Disease Control and Prevention, 155 Changbai RoadChangping District, Beijing 102206, P. R. China 4Shandong University Climate Change and Health Center, 44 WenHua Road, Jinan 250012Shangdong, P. R. China Full list of author information is available at the end of the article County-level heat vulnerability of urban and rural residents in Tibet, China Li Bai1, Alistair Woodward2, Cirendunzhu3 and Qiyong Liu1,4* Abstract Background: Tibet is especially vulnerable to climate change due to the relatively rapid rise of temperature over past decades. The effects on mortality and morbidity of extreme heat in Tibet have been examined in previous studies; no heat adaptation initiatives have yet been implemented. We estimated heat vulnerability of urban and rural populations in 73 Tibetan counties and identified potential areas for public health intervention and further research. Methods: According to data availability and vulnerability factors identified previously in Tibet and elsewhere, we selected 10 variables related to advanced age, low income, illiteracy, physical and mental disability, small living spaces and living alone. We separately created and mapped county-level cumulative heat vulnerability indices for urban and rural residents by summing up factor scores produced by a principal components analysis (PCA). Results: For both study populations, PCA yielded four factors with similar structure. The components for rural and urban residents explained 76.5 % and 77.7 % respectively of the variability in the original vulnerability variables. We found spatial variability of heat vulnerability across counties, with generally higher vulnerability in high-altitude counties. Although we observed similar median values and ranges of the cumulative heat vulnerability index values among urban and rural residents overall, the pattern varied strongly from one county to another. Conclusions: We have developed a measure of population vulnerability to high temperatures in Tibet. These are preliminary findings, but they may assist targeted adaptation plans in response to future rapid warming in Tibet. Keywords: Tibet, Climate change, Heat, Vulnerability, Urban, Rural, Adaption Open Access Open Access Bai et al. Environmental Health (2016) 15:3 DOI 10.1186/s12940-015-0081-0 Bai et al. Environmental Health (2016) 15:3 DOI 10.1186/s12940-015-0081-0 Background increase of warm winters after 2000 [4]. Warming ap- pears to be positively associated in Tibet with elevation [1]. As a result of these changes, heat has become a new climatic threat in Tibet, although Tibetans still experi- ence periods of extreme cold [4]. Tibet of China has been identified as one of areas most vulnerable to climate variability and change in the world [1, 2]. Temperatures on the Tibetan plateau have been increasing by as much as 0.50 °C per decade during the past 30 years [3], a much faster rate of change than has been observed in China, or in Asia generally. In the 2009 summer, maximum temperatures in Lhasa, the capital city of Tibet reached 30.4 °C. Before this, the highest record was 29.9 °C in 1971. Stronger warming trends were observed in winters, with a significant The adverse impacts of short-term exposure to ambi- ent high temperatures on human health have been re- ported in China and worldwide [5–11]. In Tibet, we carried out the first studies on heat-related mortality and morbidity using time-series analyses. For mortality effects, we found those at greater risk of dying on very hot days included the elderly and men, but the effects of temperature on all-cause mortality were not consistent across all study counties [12]. For morbidity effects [13], we found that hot temperatures were more strongly as- sociated with hospital admissions in Lhasa than were cold temperatures. Heat effects were associated with in- creases of total emergency room visits, hospital Bai et al. Environmental Health (2016) 15:3 Page 2 of 10 Page 2 of 10 admissions for non-accidental diseases, renal diseases and respiratory disease. Again, those ≥65 years of age and males were more likely to be affected during high temperature days compared with young people and fe- males . In a cross-sectional survey of 619 urban residents in Lhasa, we found widespread awareness in the local population of rising temperatures and their effects on health, and nearly 40 % reported that they had experi- enced heat-related symptoms during hot summers [14]. populations may exhibit patterns of vulnerability different from those of urban populations [19]. In China, social, cultural and economic features often vary strongly be- tween city dwellers and those in rural areas. Background For example, urban areas tend to be relatively modernized with higher socioeconomic level and better infrastructure, while many rural areas still rely heavily on agriculture and have poorly developed health facilities. These factors may also contrib- ute to the different patterns of population vulnerability. Tibet includes commercial and densely populated towns as well as dispersed rural populations in remote villages. Therefore, our study aimed to separately develop and map county-level heat vulnerability indices for urban and rural populations, using variables that have been reported previ- ously to increase vulnerability to heat. These results, together with climate models that pro- ject further warming in the coming decades [3] indicate that action must be taken to minimize the downside health effects of heat, especially among those who are most vulnerable, in the unique high-elevation setting of Tibet. So far, no heat adaptation initiatives have been de- veloped in Tibet. Some of the special features of Tibet that limit its ability to address heat-related health risks include: extremely harsh geographic and environmental conditions, poorly developed public health infrastruc- ture, low coping capacity of the health sector, lack of ef- fective mechanisms for intersectoral collaboration, the distinctive effects of high altitude on health conditions, and the large fraction of the population with low educa- tional attainment. As a first step in responding to these challenges, we aimed to provide information on the vul- nerability of populations to heat, in a spatially explicit form. Study population and data sources Environmental Health (2016) 15:3 Page 3 of 10 Page 3 of 10 maximum, minimum and mean temperatures in seven re- gions in Tibet during 1970 to 2013. Nyingtri has the low- est elevation and highest average temperatures, while the lowest temperatures are observed in Nakchu, with a high- est average altitude of more than 4,500 meters. no robust comprehensive data on the prevalence of other chronic conditions in the United States. In Tibet also, such variables are not currently available at a county level. We used, instead, the percentage of the population who have partly or totally lost their labour ability due to mental or physical diseases. We selected 10 county-level demographic and socioeco- nomic variables which have been associated with mortality and morbidity during extremely hot days. The variables were chosen according to 1) previous literature, 2) our re- cent studies in Tibet and 3) data availability. They in- cluded measures of advanced age, loss of labor ability, education, living alone, poverty and small dwellings. Table 1 shows the source, definitions and units of these variables. All selected data are available for all counties, except for Shuanghu County. Shuanghu has very few resi- dents due to extreme altitude of more than 5,000 meters, and was recognized as a county only until 2013. p y We included illiteracy rather than low education as a vulnerability variable. Others have reported that those with low education attainments have a greater vulner- ability to weather-related mortality [25, 26]. Nearly all previous assessments considered low education (e.g. less than high school diploma [16, 17]) as an important vul- nerability factor. In Tibet, we found that illiterate people tended to be more affected by heat than those who were literate [12]. It should be noted that literacy rates in Tibet have increased dramatically over recent decades. However, the proportion of the population that has never attended school is still high, especially among rural residents. Higher mortality and morbidity among the elderly, es- pecially those older than 65 years, during extreme heat have been reported elsewhere [20] and observed also in our recent epidemiological studies in Tibet [12, 13]. However, in this assessment, we defined the elderly as those ≥60 years of age rather than ≥65 years which is the cut-point commonly applied in other developed countries [16, 17], because overall life expectancy in Tibet (68.17 years) is still relatively low [21]. Study population and data sources y p p Tibet Autonomous Region, with an average altitude of more than 4,000 meters, is located in the southwest of China, with a land area of 1.22 million square kilometers. There are 74 county-level administrative units in Tibet. In 2010 the population of Tibet numbered 3,002,165 (77.3 % rural residents; 22.7 % urban residents). The region has a complex climate with distinct dry and wet seasons, char- acterized by cold, dry conditions in the northwest and warm, wet weather in southeast areas. Due to its unique geographic location, Tibet experiences relatively small yearly variations in air temperature, but there are large di- urnal swings in temperature. Figure 1 shows average Elsewhere, studies have mapped heat vulnerability using variables drawn from the literature, but mainly in devel- oped countries. Principal Component Analysis (PCA), a common technique for finding patterns in data of high di- mension, has often been used in these studies to construct a composite index. For example, Wolf and McGregor [15] developed a heat vulnerability index in the City of London, UK using 11 indicators from census data. The study found that areas with high sensitivity scores coincided frequently with areas that experienced especially high temperatures. Reid et al. [16] applied a factor analysis with 10 pre- selected vulnerability variables and found cities on the United States west coast and in New England, and gener- ally the downtown areas of large urban centers, to be the places most vulnerable to heat. A similar approach was followed by Harlan et al. [17] for a heat vulnerability as- sessment in Maricopa County, Arizona, USA. Three fac- tors including socioeconomic vulnerability, elderly/ isolation, and unvegetated area explained 79.8 % of the variance in the original dataset in this study. Fig. 1 Temperatures in seven regions of Tibet during 1970–2013 (Purple, green and blue represent maximum, mean and minimum temperatures, respectively) Most of these studies limited their analysis to urban areas, because those living in urban areas are generally found to be at higher risks of death due in part, perhaps, to the higher temperatures which result from the urban heat island effect [18]. However, some report the relative increase in mortality during heat events was greater in suburban and rural counties, indicating that rural Fig. 1 Temperatures in seven regions of Tibet during 1970–2013 (Purple, green and blue represent maximum, mean and minimum temperatures, respectively) Bai et al. Study population and data sources Social isolation or living alone has been identified as a vulnerability factor in heat events [22, 27]. Iso- lated seniors who live alone tend to be at even higher risk of dying during heat waves [27]. Unlike other more developed parts of China with a rapidly increas- ing number of one-senior households, few elderly per- sons live alone in Tibet (Table 2). However, in line with previous vulnerability studies [16, 17], we in- cluded both the percentage of the total population living alone and the fraction of elderly living alone. Physical disabilities and pre-existing diseases such as cardiovascular [22], cerebrovascular [23] diabetes [24] and respiratory diseases [13] may increase human susceptibil- ity to heat-related illnesses and death. However, informa- tion on the prevalence of these conditions is often lacking. For example, Reid et al. [16] included only diabetes preva- lence in their heat vulnerability assessment, as there are People receiving low income not only suffer from poorer health in general, but also experience higher mortality during hot days [28, 29]. Poverty levels used in previous vulnerability studies [16, 17, 30] are usually country- Table 1 Sources and definitions of ten selected heat vulnerability variables Source (Year) Variable name Definition a Tibet Census (2010) Age ≥60 Percent population ≥60 years of age Loss of labor ability Percent population losing labor ability because of physical or mental diseases Illiterate Percent urban population ≥15 years of age knowing less than 2000 words Percent rural population ≥15 years of age knowing less than 1500 words Living alone Percent population living alone Age ≥60 living alone Percent population ≥60 of age living alone Households with only one room Percent households having only one room which is not a toilet, kitchen or living room Households ≤8 m2 living spaces Percent households with less than 8 m2 gross floor area Minimum Living Allowances, Ministry of Civil Affairs (2010) Low income Percent population receiving the minimum living allowances Low income among seniors Percent elderly population receiving the minimum living allowances Low income households Percent households receiving the minimum living allowances a Definitions of all variables are same for both urban and rural population in Tibet expect for illiterate persons Table 1 Sources and definitions of ten selected heat vulnerability variables Source (Year) Variable name Definitio Bai et al. Environmental Health (2016) 15:3 Page 4 of 10 specific. The 2010 Tibet Census did not include questions on income. Study population and data sources Instead, we collected county-level data on Minimum Living Allowances in 2010 from the Ministry of Civil Affairs, China. We extracted numbers of tota lation, the elderly and all households that receive th mum living allowances in urban and rural areas in Table 2 Means, standard deviations (SDs), and Spearman’s correlations for vulnerability variables for urban and rural resident counties of Tibet Age ≥ 60 Loss of labor ability Illiterate Living alone Age ≥60 living alone Low income Low income among seniors Low income house-holds House-holds with only one room House-ho living spa Urban Residents Mean 5.44 0.23 19.90 7.35 0.47 7.55 1.20 14.20 21.86 5.28 SD 1.69 0.19 10.54 4.72 0.41 7.09 1.59 16.90 12.73 4.46 Minimum 1.25 0.00 2.50 0.90 0.00 0.35 0.01 0.16 0.83 0.17 Maximum 10.14 1.04 63.12 21.30 2.65 31.71 6.82 92.79 61.07 19.89 Age ≥60 1.00 Loss of labor ability 0.54 1.00 Illiterate 0.35 0.21 1.00 Living alone −0.04 0.04 −0.19 1.00 Age ≥60 living alone 0.40 0.403 0.06 0.45 1.00 Low income 0.02 −0.12 0.21 −0.15 −0.11 1.00 Low income among seniors 0.13 0.02 0.19 −0.10 −0.11 0.68 1.00 Low income households 0.11 −0.05 0.27* −0.31 −0.15 0.89 0.74** 1.00 Households with only one room −0.19 −0.24* −0.15 0.27* −0.04 −0.16 −0.35 −0.35 1.00 Households ≤8 m2 living spaces −0.08 −0.13 −0.12 −0.04 0.02 −0.07 −0.20 −0.22 0.63 1.00 Rural Residents Mean 8.29 0.45 36.38 2.04 0.40 10.17 3.09 13.29 9.93 4.84 SD 1.32 0.16 12.36 1.09 0.24 3.03 2.85 5.91 8.99 5.40 Minimum 5.39 0.05 14.04 0.46 0.05 1.45 0.07 1.50 0.49 0.13 Maximum 11.37 1.08 70.48 5.87 1.42 17.47 11.19 30.05 51.52 28.48 Age ≥60 1.00 Loss of labor ability 0.04 1.00 Illiterate −0.30** −0.28* 1.00 Living alone 0.02 0.22 −0.24* 1.00 Age ≥60 living alone 0.13 0.19 0.06 0.75 1.00 Low income a 0.16 −0.16 0.13 −0.15 0.05 1.00 Low income among seniors −0.03 −0.10 −0.06 −0.08 −0.02 0.69 1.00 Low income households 0.23* 0.03 0.07 −0.16 0.02 0.59 0.39 1.00 Households with only one room −0.36 −0.21 0.38 0.35 0.34 −0.33 -.30 −0.27* 1.00 Households ≤8 m2 living spaces −0.18 −0.23* 0.33** 0.24* 0.26* −0.25* −0.25* −0.10 0.74** 1.00 p ≤0.05. Study population and data sources *p ≤0.01 a Low income is defined as individuals, seniors and households receiving the minimum living allowances Table 2 Means, standard deviations (SDs), and Spearman’s correlations for vulnerability variables for urban and rural residents in 73 counties of Tibet Table 2 Means, standard deviations (SDs), and Spearman’s correlations for vulnerability variables for urban and rural residents in 73 counties of Tibet Civil Affairs, China. We extracted numbers of total popu- lation, the elderly and all households that receive the mini- mum living allowances in urban and rural areas in Tibet. Civil Affairs, China. We extracted numbers of total popu- lation, the elderly and all households that receive the mini- mum living allowances in urban and rural areas in Tibet. specific. The 2010 Tibet Census did not include questions on income. Instead, we collected county-level data on Minimum Living Allowances in 2010 from the Ministry of specific. The 2010 Tibet Census did not include questions on income. Instead, we collected county-level data on Minimum Living Allowances in 2010 from the Ministry of Page 5 of 10 Bai et al. Environmental Health (2016) 15:3 Page 5 of 10 Small living spaces and poor dwelling conditions are risk factors of heat-related health conditions. One study in Chicago included dwelling-related variables (e.g. me- dian room number and proportion of housing units with only one room) in a social vulnerability index for heat [31]. Similarly, we used in this analysis the percentages of households with only one room and households with less than 8 m2 gross floor area. county. Each factor score was divided into six cat- egories based on standard deviations. Each category was assigned scores: 1 (≥2 SD below mean), 2 (1–2 SD below mean), 3 (<1 SD below mean), 4 (<1 SD above mean), 5 (1–2 SD above mean), 6 (≥2 SD above mean). We then created a heat vulnerability index that results from summing the integer scores for all four factors for each county. The cumulative vulnerability index values were then mapped using ArcMap 10. Consistent with previous similar studies [15–17, 30–32], we did not include sex as a variable, although we have observed previously that Tibetan males are more sensitive than females to extreme heat. The reason is the proportion of males varies little across counties and therefore cannot contribute materially to differences in vulnerability at the population level. Study population and data sources Similarly, we did not consider ethnic groups, as 90.48 % of the overall population in Tibet is ethnic Tibetan and only a few counties include small proportions of other ethnic groups. The analysis was carried out for urban and rural resi- dents separately. The Chinese government announced official definitions of urban and rural area for the pur- pose of statistics in 2008. The census-defined urban area includes city districts and towns, and rural are defined as all areas rather than city districts and towns. Popula- tions in a county may include urban residents living in a town and rural residents living in villages. All statistical analyses were performed using the R software (version 3.0.1). PCA was performed through “prcomp” package. Absolute values > 0.5 are the most significant loadings on that factor Bai et al. Environmental Health (2016) 15:3 Page 6 of 10 We also calculated each area’s mean cumulative heat vulnerability index value (Additional file 1). Lhasa had the second-highest vulnerability of urban people, but the low- est value for rural residents. This is mainly driven by the fact that Lhasa has the highest score of factor 2 among urban population, but very low value of the same factor among rural population. In contrast, Ngari had the highest score for rural residents, but the lowest vulnerability among those in urban areas. In Nyingtri Area, the differ- ence between heat vulnerability of the two populations is the smallest. social isolation (individuals and seniors living alone) were independently represented as factor 3 in both urban and rural residents’ vulnerability index. The variable ≥60 years of age loaded with loss of labor ability to make up factor 4 for the rural population, whereas for those in urban set- tings, the two variables along with illiteracy were identified as factor 2. We observed the same median values and similar ranges of the cumulative heat vulnerability index values between the two study populations (Additional file 1). Numbers of counties above the mean value for all four factors are 28 and 24 respectively for urban and rural residents (Additional file 1). Ten counties have the same vulnerability index scores for those living in urban and rural areas. In absolute terms the vulnerability of the two populations varies by county: Gyantse County of Shigatse Area had the highest cumulative vulnerability index value for urban residents, and this was greater than the value for the rural populations in the same county. Tsochen County of Ngari Area had the highest vulnerability scores for rural residents, while the vulner- ability of the urban population in this country was relatively low. Figure 2 and 3 show the geographic distribution of the cumulative vulnerability index for urban and rural resi- dents respectively. In urban areas, higher heat vulnerabil- ity was seen in the middle and Northwest and east of Tibet. For rural areas, clustering of high vulnerability was found in Western, central and Southeastern Tibet. Clus- tering of lower vulnerability was seen in Nyingtri Area in Southeastern Tibet for both the two populations. Results bl The 10 variables selected for analysis are described in Table 2. We carried out Spearman’s correlation ana- lysis to examine the relations between variables. Fol- lowing a method by Reid et al. [16], we carried out a principal components analysis (PCA) using varimax rotation. PCA is a statistical method to reduce the original dataset to smaller sets of new, independent variables (components). Four principal factors were identified (eigenvalues > 1) for urban and rural resi- dents. Standardized scores (mean = 0; SD = 1) were calculated for each factor and assigned to each Table 2 shows high correlations between some variables. For both study populations, PCA yielded four factors with similar structure, as shown in Table 3. The components for urban and rural residents explained 77.7 % and 76.5 % respectively of the variability in the original vulnerability variables. For rural residents, two variables related to small dwellings loaded with illiteracy to form factor 1, but they were an independent factor for urban residents. Three as- pects of low income were identified as factor 1 for urban residents, but as factor 2 for rural ones. Indicators of Table 3 Principal components analysis of heat vulnerability variables for urban and rural residents in 73 counties of Tibet Factor loading Urban Rural Factor 1: Poverty Factor 2: Elderly/ Fragile health/ Illiterate Factor 3: Social isolation Factor 4: Small dwelling Factor 1: Illiterate /Small dwelling Factor 2: Poverty Factor 3: Social isolation Factor 4: Elderly/Fragile health Age ≥60 0.05 0.87 0.10 −0.09 −0.23 0.12 0.00 0.80 Loss of labor ability −0.16 0.78 0.10 −0.25 −0.18 −0.05 0.20 0.74 Illiterate 0.47 0.55 −0.25 0.18 0.79 0.12 −0.33 −0.09 Living alone −0.12 −0.13 0.90 0.01 −0.04 −0.17 0.91 0.03 Age ≥60 living alone −0.08 0.39 0.80 0.12 0.40 0.15 0.70 0.34 Low income 0.95 0.00 −0.08 0.10 −0.04 0.94 −0.11 0.08 Low income among seniors 0.76 −0.08 0.05 −0.28 −0.31 0.78 0.13 −0.34 Low income households 0.93 0.06 −0.21 −0.07 0.22 0.68 −0.14 0.34 Households with only one room −0.13 −0.28 0.20 0.78 0.76 −0.22 0.34 −0.30 Households ≤8 m2 living spaces −0.03 0.00 −0.05 0.88 0.76 −0.05 0.28 −0.22 Absolute values > 0.5 are the most significant loadings on that factor nents analysis of heat vulnerability variables for urban and rural residents in 73 counties of Tibet Bai et al. Environmental Health (2016) 15:3 Discussion This is the first study assessing and mapping heat vul- nerability of urban and rural populations, in the Tibet Fig. 2 Map of cumulative heat vulnerability by county for urban residents in Tibet Fig. 2 Map of cumulative heat vulnerability by county for urban residents in Tibet Bai et al. Environmental Health (2016) 15:3 Page 7 of 10 Fig. 3 Map of cumulative heat vulnerability by county for rural residents in Tibet Fig. 3 Map of cumulative heat vulnerability by county for rural residents in Tibet previously from a high-elevation setting. The Tibetan plat- eau may not experience such extreme heat events as other inland provinces of China, but the temperature in Tibet is increasing more quickly than other areas of China [1, 4]. Moreover, we have already noted that some sub- populations including the elderly, men and illiterate per- sons are at higher risks of dying or being sick during high temperature days [12, 13]. The vulnerability approach could be a useful tool for a rapid developing Tibet to quickly identify potential areas for heat-health action plans. This kind of literature-based assessment does not require us to wait until all determinants of human vulner- ability are described comprehensively in regions which lack reliable long-term health data. Autonomous Region of China. Our analysis was based on 10 factors which may increase human vulnerability to heat according to previous epidemiologic research, and we applied similar statistical methods to those used by Reid et al. [16]. Overall, our study shows that the relative vulnerabilities of the two study populations varies across counties, with generally higher (and more adverse) scores for vulnerability of rural populations in the cen- tral Tibet, and for urban residents in the West and Southeast. Particular attention should be paid to those high-elevation areas in South Tibet. They not only dis- play relatively higher population vulnerability (due to higher proportions of relevant demographic risk factors), but also are warming more quickly than parts of Tibet at low altitude. The factor structure of derived heat vulnerabilities in- dices is generally the same for the two study populations in Tibet. One exception which should be noted is the configuration of factor loadings for the illiteracy variable. It loaded with small living spaces to form an independ- ent factor in rural populations, but with variables of advanced age and those who lost labor abilities in the index of urban residents. Discussion For those living in rural villages There is a growing body of work on spatial heat vulner- ability indices. However, most of the studies carried out in the past were conducted in cities of developed countries such as the UK [15], the United States [16, 31], Australia [32] and Canada [30]. Little is known about spatial vari- ability of heat vulnerability in areas of developing coun- tries like China and there have been no relevant reports Bai et al. Environmental Health (2016) 15:3 Page 8 of 10 Page 8 of 10 assessment of health risks included in the report. An- other example is the development and implementation of the “Programme to address climate change in Tibet Autonomous Region” by the Development and Reform Commission and Meteorological Administration of Tibet. The Plan did not consider the public health chal- lenges of a changing climate, although it identified sys- tematic strategies to deal with climate change effects on agriculture, livestock, forest, water resources, industry and natural ecosystems. We suggest that better know- ledge of the range and magnitude of temperature- sensitive health outcomes is the first important step to increase awareness at the local government level of health implications of climate change. in Tibet, higher proportions of elderly residents and those living in small dwellings tended to occur in the same counties. This may be explained by a tendency for the oldest sections of the population to live in old and crowed dwellings in poor and remote neighborhoods in China. In contrast, most of the urban residents in small living spaces are likely to be younger people who have been drawn to the cities or towns and are sharing rental dwellings. In the urban residents’ index, variables for illiteracy, the elderly and loss of labor ability loaded on the same factor, suggesting that illiterate residents living in cities and towns tend to be older and experience mental or physical disabilities. In this study we created cumulative heat vulnerability indices for urban and rural populations separately. We suggest that this approach may enable better targeted as- signments of adaptive interventions and resources, be- cause conditions for urban and rural residents in the same county in Tibet may be quite different. For ex- ample, in Ngari Area, we observed the highest vulner- ability score of rural populations, but the lowest for urban residents (Additional file 1). Discussion Ngari Area, with an average altitude of 4500 m, has one of the lowest popu- lation densities in the world due to its extreme altitude and very harsh natural environment. Rural residents in this area tend to live in remote villages, and are further disadvantaged by limited access to education, low in- come, poor housing quality and living conditions. The area with the lowest vulnerability values of urban resi- dents is also found to be Ngari Area, possibly explained by the fact that it has the lowest illiteracy rate among urban people in all Tibet. Educational facilities and con- ditions have been improved dramatically over recent de- cades in Ngari by the Chinese government’s substantial investment. However, most of the new schools are lo- cated in towns, and greater investments in education in rural and remote areas in Tibet are needed to minimize educational inequality and vulnerability to poor health. p g The limited capacity of public health departments in Tibet creates further challenges. The local Center of Disease Control and Prevention (CDC) is the major in- stitution working in the field of diseases control and public health managements, and is supposed to be the lead agency when it comes to developing public health components of heat preparedness measures in Tibet. There are a number of barriers that make it difficult for Tibet CDC to manage the health risks related to rising temperatures. Chronic staff shortage, insufficient expert- ise and limited technical skills of current staff are the major constraints. According to Tibetan Statistics in 2008, despite dramatic increases in income levels, sav- ings, educational attainment and dwelling conditions over past two decades, the number of public health tech- nical personnel per 1000 persons has fallen from 3.39 in 1990 to 3.02 in 2008. Options for building capacity in the health sector include increasing the number of pub- lic health personnel, and improving education and train- ing of the current work force. The capacity to develop adaptation policies and measures in health sections is re- stricted also by limited information on health impacts of climate variability due to an absence of reliable health data in Tibet. At present, there are only five counties that have been selected to carry out long-term surveil- lance of death and prevalence of chronic diseases in Tibet. Discussion Long-term hospital-based data are also limited due to very uneven spatial distributions of hospitals and lack of well-established electronic medical record sys- tems. More surveillance locations are required to collect more valid and comprehensive health data for future re- search and intervention planning. Also, sharing access to existing relevant data sets should be improved to meet needs of sustainable intersectoral collaboration and actions. Chinese governments have designed and implemented a variety of research projects, policies and plans to better build capacity to cope with meteorological disasters and long-term climate change in Tibet, while the health im- plication of climate change are relatively neglected. An Assessment Report of Climate Change Impacts in Tibet Autonomous Region has been published by the National and Tibet Climate Centers to better understand current and future adverse effects of climate variability and change. This report recognizes the adverse impacts of a rapidly warming climate in Tibet including snow line rising, glacial recession, changes in river levels, frozen soil layer movements towards the north, grassland deg- radation, increasing plant diseases and pests, decreasing biological diversity and more meteorological disasters. However, there are no human health data nor is an The heat vulnerability indices we developed have a number of limitations that users need to keep in mind when using the results for decision-making. Our analysis was limited by the lack of data for other important de- terminants of heat-health vulnerability. We did not Bai et al. Environmental Health (2016) 15:3 Page 9 of 10 Page 9 of 10 include exposure variables which are frequently used in other assessments of this kind, such as land cover and surface temperature. Extreme urban heat is still not very common in Tibet except in a few locations such as Chengguan District of Lhasa and uneven spatial reso- lution of surface temperature datasets permits only intra-urban heat vulnerability assessments [33]. Further- more, surface temperature and green spaces are not the only factors which determine air temperature and indoor temperatures. Unlike other studies, we did not include a measure of home air-conditioning (AC) although it is known access to air-conditioning may protect against heat-related illnesses and deaths [22, 34, 35]. The rea- sons are that, first, these data are not available at county-level and, second, air-conditioning at home is uncommon in Tibet (in 2008 it was estimated that only 3 % of households had AC) [36]. Conclusions In this analysis, we separately developed and mapped county-level heat vulnerability indices for urban and rural populations in Tibet. Heat vulnerability varies spatially on county scales. For both rural and urban pop- ulations, we observed higher heat vulnerability in higher-altitude counties compared with counties with relative low altitude, due to higher proportions of vul- nerable groups in remote high-elevation areas. However, we found the relative vulnerability of rural and urban residents may differ between counties. At present, no adaptive initiatives have been taken to reduce the heat- related health risks in Tibet. Our findings provide the preliminary information to public health sections to pre- pare more targeted adaptation strategies for urban and rural residents in Tibet, while future studies are required to validate the indices with local health outcome data. Acknowledgments h d This study was supported by the National Basic Research Program of China (973 Program) (Grant No. 2012CB955504). Competing interests Competing interests Competing interests The authors declare that they have no competing interests. Competing interests The authors declare that they have no competing interests. Authors’ contributions B d Q d d h LB and QL designed the study. LB analyzed the data and wrote the paper. AW and C revised the manuscript and contributed to interpret the results. QL had the overall supervision of the project. All authors read and approved the final manuscript. Discussion Similarly, we did not include a measure of use of cooling fans, as this infor- mation is not available. Besides, although many studies have indicated that individuals in occupations that entail exposure to high temperatures outdoors are more likely to develop heat-induced diseases, we were unable to in- clude a measure of occupation in this study. Housing factors (e.g. dwelling age, the number of floors, construc- tion types) may also increase the likelihood of exposure to severe heat and influence heat-health risks. However, these data were not included in our analysis, as there are too many missing values for most of the remote coun- ties. Another important limitation related to data un- availability in this study is lack of sensitivity analysis using alternative variables. not been validated with local health data since this infor- mation is not currently available. Validation is certainly an important component in vulnerability assessment studies [37], but the next step, in the absence of compre- hensive local data sets, could be to consult local Tibetan stakeholders and relevant professionals, obtain feedback on the heat vulnerability maps, and modify as appropri- ate variable selection and other aspects of the analysis. Additional file Additional file 1: Proportions and vulnerability scores of urban and rural residents in each county. (DOC 157 kb) Additional file 1: Proportions and vulnerability scores of urban and rural residents in each county. (DOC 157 kb) Given the lack of information on other domains of vulnerability, our indices were mainly based on demo- graphic variables. We observed different demographic patterns associated with heat vulnerability between urban and rural populations. For instance, illiteracy is strongly associated with household size in rural settings but matches closely with age and disability in urban set- tings. These findings indicate that the population vari- ables which can best represent heat vulnerability are different in urban and rural areas. Nevertheless, the 10 variables in this study were based on our epidemiological studies in Tibet as well as validated vulnerability indices in previous literature. Despite their imperfections we suggest the results in this study provide important and timely information to policy makers and can be used along with local meteorological records in developing heat adaptation plans. References 26. O’Neill MS, Zanobetti A, Schwartz J. Modifiers of the temperature and mortality association in seven US cities. Am J Epidemiol. 2003;157:1074–82. 1. Liu X, Chen B. Climatic warming in the Tibetan Plateay during recent decades. Int J Climatol. 2000;20:1729–42. 27. Klinenberg E. Review of heat wave: a social autopsy of disaster in Chicago. N Engl J Med. 2003;348:666–7. 2. Yu C, Zhang Y, Claus H, Zeng R, Zhang X, Wang J. Ecological and environmental issues faced by a developing Tibet. Environ Sci Technol. 2012;46:1979–80. 28. Curriero FC, Heiner KS, Samet JM, Zeger SL, Strug L, Patz JA. Temperature and mortality in 11 cities of the eastern United States. Am J Epidemiol. 2002;155:80–7. 3. China Meteorological Administration: Assessement Report of Climate Change in the Southwest of China 2012. China Meteorological Press 2013 29. Kim Y, Joh S. A vulnerability study of the low-income elderly in the context of high temperature and mortality in Seoul, Korea. Sci Total Environ. 2006; 371:82–8. 4. Du J, Yang ZG, Shi L, Ma PF. Climate change characteristics of cold and warm winters in Tibet in 50 years. Acta Geographica Sinica. 2011;66:885–94 [In Chinese]. 30. Rinner C, Patychuk D, Jakubek D. Development of a Toronto-specific, spatially explicit heat vulnerability assessment: Phase I - Final Report. Toronto:2009. 5. Chen R, Wang C, Meng X, Chen H, Thach TQ, Wong CM, et al. Both low and high temperature may increase the risk of stroke mortality. Neurology. 2013; 81:1064–70. 31. Sister C, Boone C, Golden J, Hartz D, Chuang WC. Mapping social vulnerability to heat wave in Chicago. Presented at the American Meteorological Society 2009 Meeting. Available at: https://ams.confex.com/ams/89annual/ techprogram/paper_150857.htm. 6. Guo Y, Barnett AG, Pan X, Yu W, Tong S. The impact of temperature on mortality in Tianjin, China: a case-crossover design with a distributed lag nonlinear model. Environ Health Perspect. 2011;119:1719–25. 32. Loughnan M, Tapper N, Phan T, Lynch K, McInnes J. A spatial vulnerability analysis of urban populations during extreme heat events in Australian capital cities. National Climate Change Adaptation Research Facility; 2013. Available at https://www.nccarf.edu.au/business/sites/www.nccarf.edu.au. business/files/attached_files_publications/Loughnan- ExtremeHeatEventsinAustralianCapitalCities-HighRes.pdf. p 7. Kan HD, Jia J, Chen BH. Temperature and daily mortality in Shanghai: a time-series study. Biomed Environ Sci. 2003;16:133–9. time-series study. Biomed Environ Sci. 2003;16:133–9. 8. Lindstrom SJ, Nagalingam V, Newnham HH. Impact of the 2009 Melbourne heat wave on a major public hospital. Intern Med J. 2013;43:1246–50. 9. Lowe D, Ebi KL, Forsberg B. References Heat wave early warning systems and adaptation advice to reduce human health consequences of heat waves. Int J Environ Res Public Health. 2011;8:4623–48. 33. Kestens Y, Brand A, Fournier M, Goudreau S, Kosatsky T, Maloley M, et al. Modelling the variation of land surface temperature as determinant of risk of heat-related health events. Int J Health Geogr. 2011;10:7. 10. Rocklöv J. Forsberg: The effect of temperature on mortality in Stockholm 1998–2003:a study of lag structures and heat wave effects. Scand J Public Health. 2008;36:516–23. 34. Braga AL, Zanobetti A, Schwartz J. The time course of weather-related deaths. Epidemiology. 2001;12:662–7. 11. Toloo G, FitzGerald G, Aitken P, Verrall K, Tong S. Evaluating the effectiveness of heat warning systems: systematic review of epidemiological evidence. Int J Public Health. 2013;58:667–81. 35. Kaiser R, Rubin CH, Henderson AK, Wolfe MI, Kieszak S, Parrott CL, et al. Heat-related death and mental illness during the 1999 Cincinnati heat wave. Am J Forensic Med Pathol. 2001;22:303–7. 12. Bai L, Cirendunzhu, Woodward A, Dawa, Xiraoruodeng, Liu Q. Temperature and mortality on the roof of the world: a time-series analysis in three Tibetan counties. China Sci Total Environ. 2014;485–486:41–8. 36. Tibet Burea of Statistics. Tibetan Statistical Yearbook. Beijing, China: China Statistics Press; 2008. 37. Reid CE, Mann JK, Alfasso R, English PB, King GC, Lincoln RA, et al. Evaluation of a heat vulnerability index on abnormally hot days: an environmental public health tracking study. Environ Health Perspect. 2012; 120:715–20. 13. Bai L, Cirendunzhu, Woodward A, Dawa, Zhaxisangmu, Chen B, et al. Temperature, hospital admissions and emergency room visits in Lhasa, Tibet: a time-series analysis. Sci Total Environ. 2014;490:838–48. 14. Bai L, Dunzhu C, Cuociren P, Dawa, Woodward A, Liu X, et al. Rapid warming in Tibet, China: public perception, response and coping resources in urban Lhasa. Environ Health. 2013;12:71. 15. Wolf T, McGregor G, Analitis A. Assessing vulnerability to heat stress in urban areas. The example of greater London. Epidemiology 20, S24 [Abstracts]. In: ISEE 21st Annual Conference, Dublin, Ireland, August 25–29, 2009. 16. Reid CE, O’Neill MS, Gronlund C, Brines SJ, Brown DG, Diez-Roux AV, et al. Mapping community determinants of heat vulnerability. Environ Health Perspect. 2009;117:1730–6. 17. Harlan SL, Declet-Barreto JH, Stefanov WL, Petitti DB. Neighborhood effects on heat deaths: social and environmental predictors of vulnerability in Maricopa County, Arizona. Environ Health Perspect. 2013;121:197–204. 18. Hajat S, Kovats RS, Lachowycz K. Author details 1 b 1State Key Laboratory of Infectious Disease Prevention and Control, Collaborative Innovation Center for Diagnosis and Treatment of Infectious Diseases, National Institute for Communicable Disease Control and Prevention, Chinese Center for Disease Control and Prevention, 155 Changbai RoadChangping District, Beijing 102206, P. R. China. 2School of Population Health, University of Auckland, Private Bag 92019, Auckland 1142, New Zealand. 3Tibet Center for Disease Control and Prevention, 21 Linkuo North Road, Lhasa, Tibet 850000, P. R. China. 4Shandong University Climate Change and Health Center, 44 WenHua Road, Jinan 250012Shangdong, P. R. China. Following Reid et al. [16], we created heat vulnerability indices without weighting factors. The use of a compos- ite index without weights is not ideal, and future studies may explore differential weighting of the variables based on closer understanding of heat-health vulnerability pat- tern in Tibet. The composite indices in this analysis have Received: 8 March 2015 Accepted: 7 December 2015 Received: 8 March 2015 Accepted: 7 December 2015 Page 10 of 10 Page 10 of 10 Page 10 of 10 Bai et al. Environmental Health (2016) 15:3 References Heat-related and cold-related deaths in England and Wales: who is at risk? Occup Environ Med. 2007;64:93–100. 19. Sheridan SD, Dolney TJ. Heat, mortality, and level of urbanization: measuring vulnerability across Ohio, USA. Climate Res. 2003;24:255–65. 20. Yu W, Mengersen K, Wang X, Ye X, Guo Y, Pan X, et al. Daily average temperature and mortality among the elderly: a meta-analysis and systematic review of epidemiological evidence. Int J Biometeorol. 2012;56: 569–81. Submit your next manuscript to BioMed Central and we will help you at every step: Submit your next manuscript to BioMed Central and we will help you at every step: 21. Information Office of the State Council of the People’s Republic of China, 2013. Development and Progress of Tibet. • We accept pre-submission inquiries • Our selector tool helps you to ﬁnd the most relevant journal • We provide round the clock customer support • Convenient online submission • Thorough peer review • Inclusion in PubMed and all major indexing services • Maximum visibility for your research Submit your manuscript at www.biomedcentral.com/submit and we will help you at every step: 22. Naughton MP, Henderson A, Mirabelli MC, Kaiser R, Wilhelm JL, Kieszak SM, et al. Heat-related mortality during a 1999 heat wave in Chicago. Am J Prev Med. 2002;22:221–7. 23. Stafoggia M, Forastiere F, Agostini D, Caranci N, De’Donato F, Demaria M, et al. Factors affecting in-hospital heat-related mortality: a multi-city case- crossover analysis. J Epidemiol Community Health. 2008;62:209–15. 24. Schwartz J. Who is sensitive to extremes of temperature?: A case-only analysis. Epidemiology. 2005;16:67–72. 25. Medina-Ramon M, Zanobetti A, Cavanagh DP, Schwartz J. Extreme temperatures and mortality: assessing effect modification by personal characteristics and specific cause of death in a multi-city case-only analysis. Environ Health Perspect. 2006;114:1331–6. 25. Medina-Ramon M, Zanobetti A, Cavanagh DP, Schwartz J. Extreme temperatures and mortality: assessing effect modification by personal characteristics and specific cause of death in a multi-city case-only analysis. Environ Health Perspect. 2006;114:1331–6.
https://openalex.org/W4322755322	https://www.aanda.org/articles/aa/pdf/2023/04/aa45788-22.pdf	English	null	An overview of HMI off-disk flare observations	Astronomy & astrophysics	2,023	cc-by	7,297	1 We name the events following the IAU target convention: SOLyyyy- mm-ddThh:mm:ss. ABSTRACT Context. White-light continuum observations of solar ﬂares often have coronal counterparts, including the classical ‘white-light prominence’ (WLP) phenomenon. p ( ) p Aims. Coronal emissions by ﬂares, seen in the white-light continuum, have only rarely been reported. We seek to use modern data to understand the morphology of WLP events. Methods. We have identiﬁed a set of 14 examples of WLPs detected by the Heliospheric and Magnetic Imager (HMI) experiment on board the Solar Dynamics Observatory (SDO) satellite using a new online catalogue that covers 2011–2017. These WLPs invariably accompanied white-light ﬂare (WLF) emission from the lower atmosphere from ﬂares near the limb, as identiﬁed by hard X-ray images from the Reuven Ramaty High Energy Spectroscopic Imager (RHESSI). HMI provides full Stokes information, and we have used the linear polarisations (Q and U) to distinguish Thomson scattering from cool material. esults. The event morphologies ﬁt roughly into three categories – ejection, loop, and spike – but many events sh henomena. p Conclusions. The coronal white-light continuum, observed by HMI analogously to the observations made by a coronagraph, reveals many examples of coronal emission and dynamics. Using the Stokes linear polarisation, we estimate the masses of hot coronal plasma in 11 of the 14 events and ﬁnd them to be similar to typical coronal-mass-ejection masses, but without exceeding 1015 g. We note that the HMI observations do not occult the bright solar disk and were not designed for coronal observations, resulting in relatively low signal-to-noise ratios. We therefore believe that future such observations with better optimisation will be even more fruitful. Key words. Sun: ﬂares – Sun: ﬁlaments, prominences – Sun: corona – Sun: coronal mass ejections (CMEs) Key words. Sun: ﬂares – Sun: ﬁlaments, prominences – Sun: corona – Sun: coronal mass ejections (CMEs) instrument has thus provided the ﬁrst systematic database for studying ‘white-light prominences’ (WLPs), observations of which had been extremely rare. Using HMI’s full Stokes capability, Saint-Hilaire et al. (2014) found linear polarisation in some of these ejecta, making it possible to use the familiar tool of Thomson scattering to explore the ﬂare corona right down to the limb of the Sun. HMI can thus detect two classes of ﬂare events, which we refer to as on-disk (analogous to white-light ﬂares) and oﬀ-disk (the WLPs). These often go together, especially for events near the limb. ABSTRACT The oﬀ-limb component appears in the annular zone (the ‘annulus’), some tens of arcseconds wide, just above the visible limb. HMI detects both types of ﬂare events in the ‘pseudo-continuum’ (Švanda et al. 2018). This paper introduces a catalogue of HMI ﬂare observations, the Reuven Ramaty High Energy Spectro- scopic Imager (RHESSI) HMI White Light Flare Catalogue, which is available online2. We also review the morphology of the events, which have some novel features. To our knowledge, there is no other systematic catalogue of HMI oﬀ-disk events, but we note that Namekata et al. (2017) reported a large survey of on-disk events as part of research on analogous stellar ﬂares. D. Fremstad1 , J. C. Guevara Gómez1,2 , H. Hudson3,4, and J. C. Martínez Oliveros4 1 Institute of Theoretical Astrophysics, University of Oslo, Postboks 1029 Blindern, 0315 Oslo, Norway e-mail: j.c.g.gomez@astro.uio.no j g g 2 Rosseland Centre for Solar Physics, University of Oslo, Postboks 1029 Blindern, 0315 Oslo, Norway 3 SUPA School of Physics and Astronomy University of Glasgow Glasgow G12 8QQ UK 2 Rosseland Centre for Solar Physics, University of Oslo, Postboks 1029 Blindern, 0315 Oslo, Norway 3 SUPA School of Physics and Astronomy, University of Glasgow, Glasgow G12 8QQ, UK 4 Space Sciences Laboratory, University of California Berkeley, Berkeley, CA 92037, USA 2 Rosseland Centre for Solar Physics, University of Oslo, Postboks 1029 Blindern, 0315 Oslo, Norway 3 SUPA School of Physics and Astronomy, University of Glasgow, Glasgow G12 8QQ, UK 4 Space Sciences Laboratory, University of California Berkeley, Berkeley, CA 92037, USA Received 23 December 2022 / Accepted 23 February 2023 Open Access article, published by EDP Sciences, under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. This article is published in open access under the Subscribe to Open model. Subscribe to A&A to support open access publication. 2 http://sprg.ssl.berkeley.edu/~oliveros/wlf_ catalogue/catalog.html Astronomy & Astrophysics Astronomy & Astrophysics Astronomy & Astrophysics A&A 672, A32 (2023) https://doi.org/10.1051/0004-6361/202245788 c⃝The Authors 2023 A&A 672, A32 (2023) https://doi.org/10.1051/0004-6361/202245788 c⃝The Authors 2023 Open Access article, published by EDP Sciences, under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. This article is published in open access under the Subscribe to Open model. Subscribe to A&A to support open access publication. A32, page 1 of 6 3 Event No. 14 at https://sites.google.com/berkeley.edu/ hsi-hmi-catalog/2013 2. The catalogue improvement can be seen by comparing the examples in Fig. 1; it roughly doubled the width of the annulus to about 40′′. Most white-light ﬂare (WLF) observations have come from broadband sensors, or (rarely) spectroscopic data. The HMI database we use diﬀers substantially from either of these approaches: the ‘Ic’ pseudo-continuum data product measures the continuum background for a photospheric absorption line over a narrow spectral band, using six wavelength samples over a tuneable range of 678 mÅ at the 6173.3 Å line (Hoeksema et al. 2014). This limited sampling has some disadvantages; it is a nar- row slice of the continuum adjacent to a spectral line that wob- bles in wavelength owing to the spacecraft Doppler motion, for example. In sunspot regions, where WLFs usually happen, the continuum level may be much less than that of the quiet pho- tosphere. The line itself may actually go into emission during the ﬂare. Despite these caveats, the ﬂare observations do corre- late well with independent observations as regards image mor- phology and time variations, with some uncertainties as regards absolute photometry (e.g., Švanda et al. 2018). 1. Introduction The Heliospheric and Magnetic Imager (HMI) has provided a large volume of solar observations at visible wavelengths in a narrow band, including the photospheric Fe i line at 6173.34 Å (Scherrer et al. 2012; Hoeksema et al. 2014). These observations include a glimpse of the continuum near this wavelength and have suﬃcient time resolution to have captured many solar ﬂares. The observations began on 21 April 2010 and continue to the present, thus including the full sunspot maximum of Solar Cycle 24. HMI observes solar ﬂares only because it patiently records the full Sun, with a very high duty cycle thanks to the geosynchronous orbit of the Solar Dynamics Observatory (SDO; Pesnell et al. 2012). Because HMI’s principal scientiﬁc objec- tives are related to helioseismology and solar magnetism, its ﬂare observations are serendipitous and not optimal; neverthe- less, HMI has produced a unique record of the many events that comprise the survey reported here. p y p Another serendipitous bonus is that HMI images extend into an annular region outside the solar disk, and Martínez Oliveros et al. (2014) note that two X-class solar ﬂares (SOL2013-05-13T021 and SOL2013-05-13T16) were detectable both on the disk (akin to classical white-light ﬂares; see e.g., Neidig 1989; Hudson 2016) and above the limb. The 1 We name the events following the IAU target convention: SOLyyyy- mm-ddThh:mm:ss. A32, page 1 of 6 Fremstad, D., et al.: A&A 672, A32 (2023) Fig. 1. Example of the spike (left; SOL2013-05-13T01:50), ejection (middle; SOL2017-09-10T), and loop (right; SOL2015-10-02T00:07) coronal ﬂare morphologies. The date and time of each snapshot is listed above each ﬁgure. Note the increase in annulus width between 2013 and 2015. These are running diﬀerence images, from the 45-second cadence of the original data, with black showing newly appearing emission. Fig. 1. Example of the spike (left; SOL2013-05-13T01:50), ejection (middle; SOL2017-09-10T), and loop (right; SOL2015-10-02T00:07) coronal ﬂare morphologies. The date and time of each snapshot is listed above each ﬁgure. Note the increase in annulus width between 2013 and 2015. These are running diﬀerence images, from the 45-second cadence of the original data, with black showing newly appearing emission. 3. Morphologies of off-limb white-light ﬂares Diﬀerence imaging is the essential tool underlying most HMI ﬂare observations. Occasionally, WLF events on disk are bright enough to be seen without this advantage (famously Carrington 1859; Hodgson 1859). In the annular region outside the disk, though, recognition is a little bit harder. This region is dominated by scattered light originating in the HMI optics, and the noise ﬂuctuations of this unwanted signal dictate the sensitivity limit for ﬂare observations. The annulus initially was limited to about 20′′ in width, owing to SDO telemetry limita- tions. It was increased substantially after the 2013 observations, revealing a number of interesting phenomena in the region. This p We made rough morphological classiﬁcations of the ﬂares by looking at the diﬀerence in intensity over time for each ﬂare and observing the dynamics of the ﬂare via movies and diﬀer- ence images, as contained in the catalogue. We established three tentative morphologies (see Sect. 3.5 and Fig. 3): spikes, which appear as quick ﬂashes of light close to the limb; ejections – associated with loop prominence systems (LPSs) in the classi- cal Hα sense of Bruzek (1964) – which appear as mass mov- ing rapidly through the annulus and are observed to take place after spikes; and loops, which are arc-shaped features appear- ing to move slowly outwards from the limb. Some events exhibit multiple morphological properties, and typically the spike com- ponent precedes the loop component in time, mirroring the impulsive and gradual phases of ﬂare development. Examples of these morphologies are shown in Fig. 1; the images are shown as they come by default in HMI data (i.e. they are not rotated). Diﬀerence imaging is the essential tool underlying most HMI ﬂare observations. Occasionally, WLF events on disk are bright enough to be seen without this advantage (famously Carrington 1859; Hodgson 1859). In the annular region outside the disk, though, recognition is a little bit harder. This region is dominated by scattered light originating in the HMI optics, and the noise ﬂuctuations of this unwanted signal dictate the sensitivity limit for ﬂare observations. The annulus initially was limited to about 20′′ in width, owing to SDO telemetry limita- tions. It was increased substantially after the 2013 observations, revealing a number of interesting phenomena in the region. This 3. Morphologies of off-limb white-light ﬂares Using the catalogue, we were able to explore the coronal mor- phology of WLFs systematically. This was done by looking for ﬂares with visible mass ejection in the HMI annulus region, where the detection sensitivity is much higher than on the disk because it has no direct photospheric signal. In this paper we have manually picked out 13 events from the catalogue in which we can clearly observe activity above the limb. We have also included the SOL2017-09-10T ﬂare, for a total of 14 events. The continuum in general has several possible contributions (Hiei et al. 1992; Jejˇciˇc et al. 2018), and, because of the nov- elty of this database, we make no assumptions about the phys- ical nature of the phenomena, instead proceeding initially just from morphological appearances. Candidate physical phenom- ena would include surges, sprays, jets, ﬁlament eruptions, coro- nal mass ejections, and ﬂare-induced coronal ‘evaporation’, for example. Our catalogue (at the time of writing) lists 461 entries for M- and X-class ﬂares in Cycle 24 and is based on the RHESSI ﬂare list for the years 2011 through 2016. Each entry contains inten- sity, running-diﬀerence, and power-intensity movies for ﬂares within 70 degrees of disk centre and intensity, plus running- diﬀerence and saturated running-diﬀerence movies for limb and oﬀ-limb events. A linked script then returns a cutout HMI data cube for any selected event (see the catalogue entry3 for the ﬁrst oﬀ-limb detection, of SOL2013-05-13T16:00 (X2.8), as reported by Martínez Oliveros et al. 2014). This remarkable event is one of the best detections in the HMI annulus, but the catalogue contains many more examples, as surveyed in Sect. 3. Our catalogue (at the time of writing) lists 461 entries for M- and X-class ﬂares in Cycle 24 and is based on the RHESSI ﬂare list for the years 2011 through 2016. Each entry contains inten- sity, running-diﬀerence, and power-intensity movies for ﬂares within 70 degrees of disk centre and intensity, plus running- diﬀerence and saturated running-diﬀerence movies for limb and oﬀ-limb events. A linked script then returns a cutout HMI data cube for any selected event (see the catalogue entry3 for the ﬁrst oﬀ-limb detection, of SOL2013-05-13T16:00 (X2.8), as reported by Martínez Oliveros et al. 2014). This remarkable event is one of the best detections in the HMI annulus, but the catalogue contains many more examples, as surveyed in Sect. 3. 3.3. Emission mechanisms Hiei et al. (1992) described the mechanisms likely to produce the emissions detected with HMI: Thomson scattering, recom- bination radiation (free-free and free-bound), and line emission. These three mechanisms (atomic processes) need not be opti- cally thin if the source densities are large enough (Jejˇciˇc et al. 2018). HMI’s capability for linear polarisation makes detec- tion of the Thomson-scattering component unambiguous, within signal-to-noise limitations. To disentangle the atomic processes requires consideration of radiative transfer and source geometry. Martínez Oliveros et al. (2014) included detection of the HMI target line (6173 Å), normally an absorption line in the quiet Sun, which appeared in emission in the legs of the ejection event SOL2013-05-13T16:01 (their Fig. 1). Generally, we can imagine such sources to be optically thick extensions of the photosphere itself, and thus to have the potential to be much brighter than the Thomson-scattered sources. I = I0e−a(x−x0)2+2b(x−x0)(y−y0)+c(y−y0)2 , (1) where a = cos2 θ 2σ2x + sin2 θ 2σ2y b = −sin 2θ 4σ2x + sin 2θ 4σ2y c = sin2 θ 2σ2x + cos2 θ 2σ2y . (2) I = I0e−a(x−x0)2+2b(x−x0)(y−y0)+c(y−y0)2 , (1) where a = cos2 θ 2σ2x + sin2 θ 2σ2y b = −sin 2θ 4σ2x + sin 2θ 4σ2y c = sin2 θ 2σ2x + cos2 θ 2σ2y . (2) (1) where where a = cos2 θ 2σ2x + sin2 θ 2σ2y b = −sin 2θ 4σ2x + sin 2θ 4σ2y c = sin2 θ 2σ2x + cos2 θ 2σ2y . (2) There are thus six free parameters: the helio-projective position ([x, y]), the length and width ([σx, σy]), the peak intensity (I0), and the rotation angle (θ). This process was done manually; an example appears in Fig. 2. In this case, the proximity to the limb of the on-disk emission results in a highly elongated 2D Gaussian ﬁt. Table 1 summarises the kernel parameters for the on-disk events in the catalogue. There are thus six free parameters: the helio-projective position ([x, y]), the length and width ([σx, σy]), the peak intensity (I0), and the rotation angle (θ). This process was done manually; an example appears in Fig. 2. In this case, the proximity to the limb of the on-disk emission results in a highly elongated 2D Gaussian ﬁt. Table 1 summarises the kernel parameters for the on-disk events in the catalogue. 3.4. Masses We were also able to estimate the event masses and column densities by assuming Thomson scattering as the dominant mechanism. This is the same as the usual analysis of the K corona (Billings 1966), with the advantage that we can assume a localised source, rather than a spherically extended 3D corona. We discuss this assumption further below. The method for cal- culating these quantities is described in Martínez Oliveros et al. (2022), where the mass and column density of the SOL2017- 09-10 oﬀ-limb ﬂare were measured to be of the order of 8 × 1014 grams and 1021 cm−2, respectively. Fig. 2. Flare site and kernel ﬁt for SOL2012-11-20T12:27, an ejec- tive event. This is again a running-diﬀerence frame, with new emission shown in white. We calculated speeds for each ejection event by ﬁtting a Gaussian to the signal at each height. The mean of the Gaussian thus traces the ﬂare in time. Figure 4 shows an example of such a trace. The mean at each height can then be ﬁtted with a linear function whose slope gives us the velocity projected in the plane of the sky. The ﬂare emission has substantial noise ﬂuctuations, resulting in gaps where the Gaussian ﬁts do not converge. As an example, in Fig. 4 we use the column of points at the bottom to ﬁnd the velocity of the ﬁrst part of the ﬂare (i.e. the spike) and the curve at the top for the second part (the loop), but the points in between them are not relevant and should be ignored. We thus made two separate linear ﬁts for the two parts. Fig. 2. Flare site and kernel ﬁt for SOL2012-11-20T12:27, an ejec- tive event. This is again a running-diﬀerence frame, with new emission shown in white. as a quick ﬂash of light on the disk, and, upon closer inspec- tion, they often display a two-ribbon or kernel structure (e.g., Hudson et al. 2006), often with diﬀuse brightening. Disk events not suﬃciently close to the limb cannot be linked to the morphol- ogy of the oﬀ-disk event, unfortunately. The catalogue includes the helio-projective position of the ﬂare, its size, and the time, using the RHESSI onset time for the time ﬁeld. It includes ani- mations of the full data and of diﬀerence images, selecting the image frame in which the ﬂare is clearly visible. From there we restricted the region and what intensities we included. For the classiﬁcation, we ﬁtted an asymmetric 2D Gaussian kernel to the image frame, which gives us the position, length, and width of the ﬂare brightening: 3.1. Analysis of on-disk ﬂare properties With modern observations it is fairly common to observe some continuum emission from a ﬂare (the WLF phenomenon), espe- cially for the more energetic cases but also sometimes for minor events (Jess et al. 2008). In HMI movies, such a ﬂare appears A32, page 2 of 6 Fremstad, D., et al.: A&A 672, A32 (2023) each height in the frame and left us with a strip of mean inten- sities for diﬀerent heights. We subtracted the continuum inten- sity from this at each height, which was found by calculating the mean intensity at each height for some time before or after the mass ejection takes place and then taking the mean of these intensities. We were thus left with a slice of corrected mean con- tinuum excess intensity. Doing this for each frame left us with results such as that shown in the bottom panel of Fig. 4. We can clearly see how the height of the ejected mass evolves over time, which provides a clear view of the time evolution of the coro- nal aspects of the ﬂare in the manner that a coronagraphic J-map does (Sheeley et al. 1999). 3.2. Analysis of off-limb ﬂare properties The dynamics seen in the oﬀ-limb observations is a key part of the morphological description. We quantiﬁed this by creat- ing time-distance plots, which give visual time lines of the ﬂare dynamics and can be further used to estimate the plane-of-the- sky velocity of the ejected mass. These plots show the evolution of ejected mass in height over time and allow us to more easily see the dynamics of the ejected mass. They thus give us a better understanding of the ﬂares’ morphology. There are several steps that go into making a plot that displays the dynamics of the ﬂare clearly. We began by cropping out the solar disk and rotating the frame such that the solar limb is horizontal. We selected a central pixel, which preferably is centred on the main body of the ejected mass. Next we selected some distance, d, from the central pixel and calculated the mean intensity from pixels with a distance less than d from the central pixel. This was done for This method requires estimates of the linearly polarised ﬂux, which HMI provides at a 12 min cadence. The Q and U polari- sation components are rotated into the frame of the solar limb such that the rotated Q′ component contains all of the lin- ear polarisation for a simple photospheric radiation ﬁeld (see Saint-Hilaire et al. 2021). We deﬁned a region of interest around the Stokes I source and integrated the Q′ signal in a pre-ﬂare- subtracted diﬀerence image. We obtained an estimate for each A32, page 3 of 6 Fremstad, D., et al.: A&A 672, A32 (2023) Table 1. On-disk ﬂare kernel parameters. Table 1. On-disk ﬂare kernel parameters. 3.2. Analysis of off-limb ﬂare properties Gaussian ﬁt for ﬂare SOL2013-05-13T15 (middle panel in Fig. 1). The top panel shows the ﬁt compared to the intensity for a hori- zontal slice in the middle of the time-distance plot, and the bottom panel shows the trace given by the mean of the Gaussian ﬁt at each height. We can see traces of the spike morphology as an illumination at the bottom of the ﬁgure, in addition to the ejection (LPS) morphology, which again is seen as a cloud appearing after the spike. Fig. 3. Illustration of the three morphologies. For the LPS or ejection, we see that it is the magnetic ﬁeld lines themselves that form an inde- pendent loop and travel away from the limb. For the spike and loop morphologies, the ejected mass follows the magnetic ﬁeld lines. In the case of the loop, the ejected mass follows the magnetic ﬁeld lines all the way around the loop, while for the spike we only observe them to follow it partially. closely match the amount of material injected at high tempera- tures into the ﬂare loops (e.g., Hudson & Ohki 1972). The mass estimation for the spike morphology in particular, or any cool material, would depend upon detailed radiative-transfer mod- elling, which we have not attempted in this article. frame to create a time series. For some of the events, especially those with a spike morphology, the mass could not be calculated precisely, owing to the complex and noisy image conﬁguration and its proximity to the limb. The mass estimates based on the polarisation reﬂect only the Thomson scattering and, therefore, may underestimate the total masses of the events. This is because any cool component may not be fully ionised. We can safely assume that the Thom- son component is optically thin, so unless a part of the source is occulted by foreground material, the mass estimate should 3.2. Analysis of off-limb ﬂare properties The ﬁfth and sixth columns describe the standard deviations of the Gaussian kernel along its major (σ2) and minor (σ1) axes. The seventh column shows the centre-averaged raw intensity of the ﬂare with respect to the average intensity of the solar disk centre. The last column shows the rotation angle of the Gaussian kernel with respect to the vertical axis, i.e. the clockwise rotation of the kernel with respect to north-south in degrees. Fig. 3. Illustration of the three morphologies. For the LPS or ejection, we see that it is the magnetic ﬁeld lines themselves that form an inde- pendent loop and travel away from the limb. For the spike and loop morphologies, the ejected mass follows the magnetic ﬁeld lines. In the case of the loop, the ejected mass follows the magnetic ﬁeld lines all the way around the loop, while for the spike we only observe them to follow it partially. 14:5015:0715:2415:4115:5816:1516:3216:4917:06 Time 0 2 5 7 9 12 14 16 Height [Mm] 0 101 102 Intensity [DN/s] 0 2000 4000 6000 8000 Time [s] 0 5 10 Intensity [DN/s] Fig. 4. Gaussian ﬁt for ﬂare SOL2013-05-13T15 (middle panel in Fig. 1). The top panel shows the ﬁt compared to the intensity for a hori- zontal slice in the middle of the time-distance plot, and the bottom panel shows the trace given by the mean of the Gaussian ﬁt at each height. We can see traces of the spike morphology as an illumination at the bottom of the ﬁgure, in addition to the ejection (LPS) morphology, which again is seen as a cloud appearing after the spike. 14:5015:0715:2415:4115:5816:1516:3216:4917:06 Time 0 2 5 7 9 12 14 16 Height [Mm] 0 101 102 Intensity [DN/s] 0 2000 4000 6000 8000 Time [s] 0 5 10 Intensity [DN/s] 1 2 Intensity [DN/s] 14:5015:0715:2415:4115:5816:1516:3216:4917:06 Time 0 0 Fig. 4. Gaussian ﬁt for ﬂare SOL2013-05-13T15 (middle panel in Fig. 1). The top panel shows the ﬁt compared to the intensity for a hori- zontal slice in the middle of the time-distance plot, and the bottom panel shows the trace given by the mean of the Gaussian ﬁt at each height. We can see traces of the spike morphology as an illumination at the bottom of the ﬁgure, in addition to the ejection (LPS) morphology, which again is seen as a cloud appearing after the spike. Fig. 4. A32, page 4 of 6 3.2. Analysis of off-limb ﬂare properties Event (RHESSI start time) tWL x0 y0 σ1 σ2 I0/I⊙ θ SOL2011-01-28T00:46 00:57 936 277 0.2 0.5 0.11 −19 SOL2011-09-22T10:53 10:52 −921 159 0.4 0.5 0.35 0 SOL2011-10-31T17:48 17:47 −937 204 0.2 0.5 0.51 −2 SOL2012-03-02T17:54 18:40 −914 305 0.5 0.9 0.41 34 SOL2012-11-08T02:06 02:19 −926 233 0.5 1.2 0.55 55 SOL2012-11-20T12:28 12:39 952 198 0.3 2.5 0.17 −14 SOL2013-05-13T01:50 02:08 −930 192 0.3 9 0.12 11 SOL2013-05-13T15:50 15:59 −931 181 1.2 6.3 0.32 16 SOL2014-10-16T12:58 13:01 −937 −221 0.4 4.5 0.19 −12 SOL2014-11-03T22:08 22:31 −936 244 0.4 9.8 0.24 14 SOL2015-03-02T15:15 15:24 899 360 0.2 8 0.19 −20 SOL2015-10-02T00:08 00:10 819 −364 1.7 3.3 0.60 −2 SOL2015-10-02T12:21 12:30 883 −330 0.3 0.9 0.61 16 SOL2017-09-10T16:05 15:58 942 −153 0.5 0.7 0.2 2 Notes. The ﬁrst column is the start time of the ﬂare as listed in the RHESSI catalogue. The second column is the time of the white-light brightening. The third and fourth columns give the helio-projective x and y coordinates in arcseconds. The ﬁfth and sixth columns describe the standard deviations of the Gaussian kernel along its major (σ2) and minor (σ1) axes. The seventh column shows the centre-averaged raw intensity of the ﬂare with respect to the average intensity of the solar disk centre. The last column shows the rotation angle of the Gaussian kernel with respect to the vertical axis, i.e. the clockwise rotation of the kernel with respect to north-south in degrees. Notes. The ﬁrst column is the start time of the ﬂare as listed in the RHESSI catalogue. The second column is the time of the white-light brightening. The third and fourth columns give the helio-projective x and y coordinates in arcseconds. The ﬁfth and sixth columns describe the standard deviations of the Gaussian kernel along its major (σ2) and minor (σ1) axes. The seventh column shows the centre-averaged raw intensity of the ﬂare with respect to the average intensity of the solar disk centre. The last column shows the rotation angle of the Gaussian kernel with respect to the vertical axis, i.e. the clockwise rotation of the kernel with respect to north-south in degrees. Notes. The ﬁrst column is the start time of the ﬂare as listed in the RHESSI catalogue. The second column is the time of the white-light brightening. The third and fourth columns give the helio-projective x and y coordinates in arcseconds. 3.5. Morphologies The second, third and fourth columns show the measured (projected) velocity for the morphology type present during that ﬂare. The ﬁfth and sixth columns show the inferred electron density and the total mass of the polarised component, based on peak polarisation signal and avoiding the spike morphology as described in the text. The seventh column shows the measured polarisation, and the eighth column shows the normalised intensity of the ejected material. The large uncertainties in the measurements basically restrict the interpretation to one signiﬁcant ﬁgure. Notes. The ﬁrst column shows the event time. The second, third and fourth columns show the measured (projected) velocity for the morphology type present during that ﬂare. The ﬁfth and sixth columns show the inferred electron density and the total mass of the polarised component, based on peak polarisation signal and avoiding the spike morphology as described in the text. The seventh column shows the measured polarisation, and the eighth column shows the normalised intensity of the ejected material. The large uncertainties in the measurements basically restrict the interpretation to one signiﬁcant ﬁgure. 23:08 23:24 23:38 23:53 00:07 00:21 00:35 00:50 01:04 01:18 01:32 Time 0 5 11 16 21 26 32 Height [Mm] Timeseries for flare 2015-10-02 0 101 102 Intensity [DN/s] Fig. 5. Time series for the SOL2015-10-02 coronal mass ejection, which has a loop morphology. Timeseries for flare 2015-10-02 events seem similar to surges and sprays, ejection events simi- lar to loop prominences, and loop events similar to ﬂows con- ﬁned to closed coronal ﬁelds. The examples we discuss below show surprises, though, and we cannot be sure that our three morphologies capture all of the possibilities. Spike morphology is deﬁned by rapid motion, somewhere between 100 km s−1 and 300 km s−1. In view of this, the low cadence of HMI image sequences makes polarisation measure- ments extremely uncertain. In general, the loss of polarisation data in this way tends to make our mass estimates lower lim- its. Spikes can appear as columns of mass ejected away from the solar disk or as large bursts that seemingly do not escape far from the disk and happen very quickly. In addition to this, they do not reach very large heights. 3.5. Morphologies We ﬁnd instances where the ejected mass during spikes moves away from the limb, but also at least one instance with rapid motions towards the limb, which we would associate with coronal rain. The natural expla- nation for outward-moving spike events would be as the HMI counterpart of the common Hα surge or spray. An illustration of spike morphology is shown in the middle drawings in Fig. 3, and an example of a spike time-distance plot can be found in the bot- tom panel of Fig. 4. Martínez Oliveros et al. (2014) also reported spike-like behaviour but with a downward motion in SOL2012- 05-13 (see below for further comments). This velocity, plus the appearance of bright material from the corona, strongly suggests an association with the late phase of coronal rain. Fig. 5. Time series for the SOL2015-10-02 coronal mass ejection, which has a loop morphology. form into an LPS, the ejected mass follows an isolated, station- ary arch, ascending in one branch and descending in the other (Bruzek 1964). Such behaviour suggests the presence of cool material. The bottom drawing of Fig. 3 provides an illustration of the loop morphology. In the time-distance plot, the loop mor- phology appears as two long columns connected at the top. We can measure the upward and downward velocity of the loop by restricting which part of the time-distance plot we want to ﬁt a Gaussian to. If we cut the plot along the middle of the loop, we can ﬁt a Gaussian and then a linear function to the left side of the loop. This gives us the upward-moving velocity, while the right side gives us the velocity moving downwards. The loops diﬀer from the LPSs because the former show an apparently parallel ﬂow, do not travel to larger heights, and have higher velocities. Ejection morphology appears to only happen after a spike event. It is apparent that the spikes initiate the LPSs and might give rise to the ejected mass that goes into making an LPS. The ejection morphology in HMI appears as long clumps of gas that are spread out over a larger period of time in the time-distance plot. The low velocity of this morphology causes it to have a very distinct ﬁngerprint on the time-distance plots. 3.5. Morphologies As previously mentioned, there are three morphologies found in this study; spike, ejection, and loop. With the methods described above we have made time-distance plots and estimated the pro- jected velocities for each of these. Roughly speaking, spike Fremstad, D., et al.: A&A 672, A32 (2023) Table 2. Projected velocities of the respective morphologies. j p p g Event Spike velocity [km s−1] LPS velocity [km s−1] Loop velocity [km s−1] Ne[cm−2] m [g] ∆Q′/∆I I0/I⊙ SOL2011-01-28T00:46 −47 ± 20 – – 3 × 1020 1014 0.07 8.1 × 10−3 SOL2011-09-22T10:53 61 ± 35 7 ± 8 – 5 × 1020 3 × 1014 0.11 8.6 × 10−3 SOL2011-10-31T17:47 – 30 ± 19 – 4 × 1020 2 × 1014 0.02 9.0 × 10−3 SOL2012-03-02T17:54 124 ± 34 9 ± 5 – – – – – SOL2012-11-08T02:06 −102 ± 16 – – – – – – SOL2012-11-20T12:28 209 ± 39 – – – – – – SOL2013-05-13T01:50 – 18 ± 10 – 5 × 1020 3 × 1014 0.02 5.5 × 10−3 SOL2013-05-13T15:50 – 15 ± 5 – 6 × 1020 3 × 1014 0.02 6.8 × 10−3 SOL2014-10-16T12:58 270 ± 31 – – 5 × 1020 3 × 1014 0.02 6.8 × 10−3 −72 ± 26 SOL2014-11-03T22:08 −64 ± 11 13 ± 10 – 2 × 1020 9 × 1013 0.02 6.3 × 10−3 SOL2015-03-02T15:15 −46 ± 11 – – 3 × 1020 2 × 1014 0.10 6.9 × 10−3 SOL2015-10-02T00:08 217 ± 37 – −87 ± 25 2 × 1020 1014 0.06 7.8 × 10−3 SOL2015-10-02T12:21 – 69 ± 22 – 2 × 1020 1014 0.09 8.2 × 10−3 SOL2017-09-10T16:05 33 ± 12 11 ± 4 – 2 × 1021 9 × 1014 0.02 7.3 × 10−3 Notes. The ﬁrst column shows the event time. The second, third and fourth columns show the measured (projected) velocity for the morphology type present during that ﬂare. The ﬁfth and sixth columns show the inferred electron density and the total mass of the polarised component, based on peak polarisation signal and avoiding the spike morphology as described in the text. The seventh column shows the measured polarisation, and the eighth column shows the normalised intensity of the ejected material. The large uncertainties in the measurements basically restrict the interpretation to one signiﬁcant ﬁgure. Notes. The ﬁrst column shows the event time. 3.5. Morphologies In Table 2 we can see that the velocities of LPSs are usually of the order of magnitude ∼10 km s−1, which is consistent with what is found in Hα by Bruzek (1964). In this case, the magnetic ﬁeld forms independent loops that appear to carry the mass away from the limb, as illustrated by the top drawing in Fig. 3. The complete evolution in height over time for each ﬂare, as determined by tracing the ﬂares by ﬁtting a Gaussian to each height, is plotted in Fig. 6. We ﬁnd that for all cases of LPSs the ejected mass is always propagating away from the Sun. Similar results to what is shown in the left panel are also reported by Bruzek (1964). In the case of the spikes, we can also ﬁnd cases where the mass is moving away from the Sun, but we also ﬁnd evidence that they can fall down back onto the Sun. In addition, the SOL2015-10-02 event shows evidence of loop morphology, and it can be seen to move away from the Sun before reversing. Loop morphology appears to reﬂect parallel ﬂows within loops. Instead of a spike ejecting mass that may eventually Summarising the appearance and behaviour of the HMI oﬀ-limb observations, we suggest that the spike morphology A32, page 5 of 6 A32, page 5 of 6 Fremstad, D., et al.: A&A 672, A32 (2023) 0 5 10 15 20 25 30 35 40 Time (minutes) 0 1 2 3 4 Height [104 km] 2011-09-22T10 2011-10-31T17 2012-03-02T17 2012-11-20T12 2013-05-13T01 2013-05-13T15 2014-10-16T12 2014-11-03T22 0 1 2 3 4 5 6 7 Time (minutes) 0 1 2 3 4 Height [104 km] 2011-01-28T00 2012-11-08T02 2015-03-02T15 2015-10-02T00 2015-10-02T12 Fig. 6. Resulting time evolution in height for the included ﬂares deter- mined by tracing the ﬂares using a Gaussian ﬁt for each height. The ﬂares that are found to be propagating away from the Sun are plotted in the top panel, while those that appear to fall back onto the Sun are plotted in the bottom panel. include the ﬁrst-ever systematic view of WLF manifestations in the low corona from an observatory in space. These oﬀ- limb observations follow and extend a very limited historical set of data, mainly visual, of WLPs. This paper introduces an online catalogue4 with comprehensive coverage of the oﬀ-limb events observed by HMI, and further analysis of the on-disk morphology as well. 4 http://sprg.ssl.berkeley.edu/~oliveros/wlf_ catalogue/catalog.html L119 Martínez Oliveros, J.-C., Krucker, S., Hudson, H. S., et al. 2014, ApJ, 780, L28 Martínez Oliveros J C Guevara Gómez J C Saint Hilaire P Hudson H & Martínez Oliveros, J.-C., Krucker, S., Hudson, H. S., et al. 2014, ApJ, 780, L28 Martínez Oliveros, J. C., Guevara Gómez, J. C., Saint-Hilaire, P., Hudson, H., & Martínez Oliveros, J. C., Krucker, S., Hudson, H. S., et al. 2014, ApJ, 780, L28 Martínez Oliveros, J. C., Guevara Gómez, J. C., Saint-Hilaire, P., Hudson, H., & Martínez Oliveros, J. C., Guevara Gómez, J. C., Saint-Hilaire, P., Krucker, S. 2022, ApJ, 936, 56 Namekata, K., Sakaue, T., Watanabe, K., et al. 2017, ApJ, 851, 91 Neidig, D. F. 1989, Sol. Phys., 121, 261 Pesnell, W. D., Thompson, B. J., & Chamberlin, P. C. 2012, Sol. Phys., 275, 3 Saint-Hilaire, P., Schou, J., Martínez Oliveros, J.-C., et al. 2014, ApJ, 786, L19 Saint-Hilaire, P., Martínez Oliveros, J. C., & Hudson, H. S. 2021, ApJ, 923, 276 Scherrer, P. H., Schou, J., Bush, R. I., et al. 2012, Sol. Phys., 275, 207 Sheeley, N. R., Walters, J. H., Wang, Y.-M., & Howard, R. A. 1999, J Geophys. Neidig, D. F. 1989, Sol. Phys., 121, 261 Pesnell, W. D., Thompson, B. J., & Chamberlin, P. C. 2012, Sol. Phys., 275, 3 Sheeley, N. R., Walters, J. H., Wang, Y.-M., & Howard, R. A. 1999, J Geophys. Res., 104, 24739 Sheeley, N. R., Walters, J. H., Wang, Y.-M., & Howard, R. A. 1999, J Geophys. Res., 104, 24739 Sun, X., Hoeksema, J. T., Liu, Y., Kazachenko, M., & Chen, R. 2017, ApJ, 839, 67 3.5. Morphologies The catalogue reveals three distinct non- exclusive morphologies – spike, ejection, and loop – which pro- duce an eﬀect in an annulus about 40′′ in width above the visible disk. 0 5 10 15 20 25 30 35 40 Time (minutes) 0 1 2 3 4 Height [104 km] 2011-09-22T10 2011-10-31T17 2012-03-02T17 2012-11-20T12 2013-05-13T01 2013-05-13T15 2014-10-16T12 2014-11-03T22 The full Stokes capability of HMI shows that Thomson scat- tering plays a major role in WLF ejecta (Scherrer et al. 2012). This is the classic coronagraphic K-corona observational mode, extended to the base of the corona, where the incident radiation ﬁeld may have signiﬁcant anisotropy (Saint-Hilaire et al. 2021). In some cases, the oﬀ-limb structures achieve suﬃcient den- sity (above about 1012 cm−3) to allow normal collisional exci- tation to play a role (Jejˇciˇc et al. 2018). The linear polarisation, however, allows for a direct determination of the ejected mass (Martínez Oliveros et al. 2022). 0 1 2 3 4 5 6 7 Time (minutes) 0 1 2 3 4 Height [104 km] 2011-01-28T00 2012-11-08T02 2015-03-02T15 2015-10-02T00 2015-10-02T12 Acknowledgements. J.C.G.G is supported by the SolarALMA project, which received funding from the European Research Council (ERC) under the Euro- pean Union’s Horizon 2020 research and innovation programme (grant agree- ment No. 682462), and by the Research Council of Norway through its Centres of Excellence scheme, project number 262622. H.H. thanks the School of Physics and Astronomy, University of Glasgow, for hospitality. Data sources: All data used in this article can be found in public-domain archives. Fig. 6. Resulting time evolution in height for the included ﬂares deter- mined by tracing the ﬂares using a Gaussian ﬁt for each height. The ﬂares that are found to be propagating away from the Sun are plotted in the top panel, while those that appear to fall back onto the Sun are plotted in the bottom panel. Acknowledgements. J.C.G.G is supported by the SolarALMA project, which received funding from the European Research Council (ERC) under the Euro- pean Union’s Horizon 2020 research and innovation programme (grant agree- ment No. 682462), and by the Research Council of Norway through its Centres of Excellence scheme, project number 262622. H.H. thanks the School of Physics and Astronomy, University of Glasgow, for hospitality. Data sources: All data used in this article can be found in public-domain archives. References corresponds to ﬂare surge or spray activity during the impul- sive phase, and that ejection may correspond to ﬁlament erup- tions. The ejection morphology is well understood from the classic behaviour of the Hα LPSs. A possible fourth morphology, rain, may be present. This would describe the infall observed by Martínez Oliveros et al. (2014) in the event SOL2013-05- 15T16:01, in which material appeared suddenly within the annular region and moved downwards at a projected speed estimated at 134 ± 8 km s−1; the act of appearance strongly sug- gests an association with thermal instability (Field 1965) and the enhancement of HMI low-temperature opacity, rather than Thomson scattering by hot electrons. This observation required special analysis at HMI’s highest time resolution, using all of the polarisation and wavelength images, each diﬀerenced against its preceding like exposure (see Sun et al. 2017 for a discussion of high-cadence data from HMI). This kind of analysis was beyond the scope of the present paper, and so this event does not appear in the lower panel of Fig. 6. Billings, D. E. 1966, A Guide to the Solar Corona (New York: Academic Press) Bruzek, A. 1964, ApJ, 140, 746 Carrington, R. C. 1859, MNRAS, 20, 13 Field, G. B. 1965, ApJ, 142, 531 Hiei, E., Nakagomi, Y., & Takuma, H. 1992, PASJ, 44, 55 Hodgson, R. 1859, MNRAS, 20, 15 Hoeksema, J. T., Liu, Y., Hayashi, K., et al. 2014, Sol. Phys., 289, 3483 Hudson, H. S. 2016, Sol. Phys., 291, 1273 Hudson, H. S., & Ohki, K. 1972, Sol. Phys., 23, 155 6 Billings, D. E. 1966, A Guide to the Solar Corona (New York: Academic Press) Bruzek, A. 1964, ApJ, 140, 746 p Carrington, R. C. 1859, MNRAS, 20, 13 g Hoeksema, J. T., Liu, Y., Hayashi, K., et al. 2014, Sol. Phys., 289, 3483 y Hudson, H. S. 2016, Sol. Phys., 291, 1273 Hudson, H. S. 2016, Sol. Phys., 291, 1273 Hudson, H. S., & Ohki, K. 1972, Sol. Phys., 23, 155 Hudson, H. S., Wolfson, C. J., & Metcalf, T. R. 2006, Sol. Phys., 234, 79 Hudson, H. S., Wolfson, C. J., & Metcalf, T. R. 2006, S Jejˇciˇc, S., Kleint, L., & Heinzel, P. 2018, ApJ, 867, 134 Jess, D. B., Mathioudakis, M., Crockett, P. J., & Keenan, F. P. 2008, ApJ, 688, L119 Acknowledgements. J.C.G.G is supported by the SolarALMA project, which received funding from the European Research Council (ERC) under the Euro- pean Union’s Horizon 2020 research and innovation programme (grant agree- ment No. 682462), and by the Research Council of Norway through its Centres of Excellence scheme, project number 262622. H.H. thanks the School of Physics and Astronomy, University of Glasgow, for hospitality. Data sources: All data used in this article can be found in public-domain archives. 4. Conclusion Sun, X., Hoeksema, J. T., Liu, Y., Kazachenko, M., & Chen, R. 2017, ApJ, 839, 67 Š The pseudo-continuum observations of the Sun from HMI con- stitute a unique database for characterising solar WLFs, and 67 Švanda, M., Jurˇcák, J., Kašparová, J., & Kleint, L. 2018, ApJ, 860, 144 67 Švanda, M., Jurˇcák, J., Kašparová, J., & Kleint, L. 2018, ApJ, 860, 144 A32, page 6 of 6
https://openalex.org/W4375933344	http://ejurnal.iaipd-nganjuk.ac.id/index.php/es/article/download/531/356	Arabic	null	STRATEGI OPTIMALISASI KUALITAS PRODUK KOPI DALAM MEMBANGUN LOYALITAS PELANGGAN DALAM TEORI EKONOMI SYARIAH (Studi Kasus di Giri Kopi Kelurahan Ploso Kecamatan Nganjuk Kabupaten Nganjuk)	Jurnal Dinamika Ekonomi Syariah	2,023	cc-by-sa	5,563	Juni Iswanto1, Alfin Yuli Dianto2, Putri Nur Indah Sari3 E-mail: juniiswanto14@gmail.com Received: October This is an open access article under CC-BY-SA license. ABSTRACT The amount of competition in the beverage business such as coffee that has begun to emerge has forced coffee-processed beverage entrepreneurs who have become the culture of everyday people to be able to provide innovation in maintaining product quality so that their customers remain loyal and loyal to the same product and place. The purpose of this study is to analyze strategies to improve product quality, build consumer loyalty and review sharia economics towards optimizing the quality of coffee products in Giri Kopi; Research Method: A descriptive qualitative research approach and a case study using two research variables for product quality and customer loyalty. Primary data sources were interviews with the owner, two employees, and 4 "Giri Kopi" customers. The research instrument used open interview guidelines. Data analysis used triangulation of researchers and theory. Research results show: 1. To optimize the quality of coffee products in accordance with the 9 indicators that make up the quality of the product with the raw materials of coffee beans directly from farmers, modern roaster grinder production machines in producing coffee products, both packaged or drinks in "Giri Kopi" shops 2. Giri Coffee builds customer loyalty by establishing reseller partners with 15% profit sharing. 3. Review of Islamic economics on optimizing the quality of coffee products in building customer loyalty at Giri Kopi is carried out by maintaining cleanliness, purity in the production process using modern coffee roaster machines and ensuring a level of honesty in guaranteeing the origin of raw materials 100% quality coffee beans with packaging weights according to specifications. K d P d Q li C ff C L l Keywords: Product Quality, Coffee, Consumer Loyalty Keywords: Product Quality, Coffee, Consumer Loyalty g j p g y 2 M Anang Firmansyah, Pemasaran (Dasar Dan Konsep). Surabaya: CV. Penerbit Qiara Media, 2019, 183. ABSTRAK Banyaknya persaingan bisnis minuman seperti kopi yang mulai bermunculan membuat para pengusaha minuman olahan kopi yang menjadi budaya masyarakat sehari₋hari harus mampu memberikan inovasi dalam menjaga kualitas produk agar para pelanggannya tetap setia dan loyal dalam satu produk dan tempat yang sama. Tujuan dari Penelitian ini untuk menganalisis strategi meningkatkan kualitas produk, membangun loyalitas konsumen dan tinjauan ekonomi syariah terhadap optimalisasi kualitas produk kopi di Giri Kopi; Metode Penelitian: Pendekatan penelitian kualitatif desktitif dan studi kasus dengan menggunakan dua variabel peneltian kualitas produk dan loyalitas pelanggan.Sumber data primer dari hasil wawancara owner, dua karyawan, dan 4 pelangan “Giri Kopi”.Instrumen penelitian menggunakan pedoman wawancara terbuka.Analisis data menggunakan triangulasi peneliti dan teori. Hasil Penelitian menunjukkan :1. Untuk optimalisasi kualitas produk kopi di sudah sesuai dengan 9 indikator penyusun kualitas produk dengan bahan baku biji kopi langsung dari petani, mesin produksi modern roaster grinder dalam menghasilkan produk₋produk kopi baik kemasan atau minuman dikedai “Giri Kopi” 2. Giri Kopi membangun loyalitas pelanggan dengan menjalin mitra reseller dengan bagi hasil 15%. 3. Tinjauan ekonomi syariah terhadap optimalisasi kualitas produk kopi dalam membangun loyalitas pelanggan di Giri Kopi dilakukan dengan menjaga kebersihan, kesucian dalam proses produksi dengan menggunakan mesin modern roaster kopi dan menjamin tingkat kejujuran dalam menjamin asal usul bahan baku 100% biji kopi berkualitas dengan timbangan kemasan sesuai dengan spesifikasinya. Kata Kunci : Kualitas Produk, Kopi, Loyalitas Konsumen This is an open access article under CC-BY-SA license. This is an open access article under CC-BY-SA license. 63 4 Jurnal Dinamika Ekonomi Syariah, Vol 10, No. 1, Januari 2023, Hal 63 - 73 64 1Tatang Ibrahim dan A. Rusdiana, Manajemen Mutu Terpadu. Bandung: Yrama Widya, 2021, 98. PENDAHULUAN Adanya perkembangan industri 4.0 yang mengimbangi adanya kemajuan teknologi dalam dunia bisnis membuat para pelaku usaha terus termotivasi untuk berinovasi atas usahanya. Begitu juga para konsumen juga menjadi lebih selektif dalam memilih dan menggunakan produk atau jasa yang ditawarkan oleh para pelaku usaha. Konsumen dalam dunia bisnis memang memiliki kedudukan paling tinggi dalam kemajuan dan kehidupan berkelanjutan suatu bisnis. Pengusaha harus dapat mengembangkan produk-produknya agar tetap menarik minat konsumen. Ada beberapa atribut yang harus diperhatikan dalam mengembangkan suatu produk yang berkualitas. Karena kualitas produk merupakan seluruh gabungan sifat-sifat produk atau jasa pelayanan dari pemasaran, engineering, manufaktur, dan pemeliharaan di mana produk atau jasa pelayanan dalam penggunaannya akan bertemu sesuai harapan pelanggan. 1 Atribut-atribut pengembangan produk yang menjadi perhatian konsumen dalam memberikan manfaat yang harus dipenuhi perusahaan adalah kualitas, fitur, serta gaya dan Kelurahanin.2 Atribut kualitas atas produk memiliki kedudukan poin utama yang dipertimbangkan konsumen. Karena kualitas merupakan atribut yang memiliki kemampuannya memenuhi kebutuhan yang telah ditentukan dan bersifat laten. Kualitas produk dalam suatu usaha ini memiliki sifat yang abstrak antara satu konsumen dengan konsumen lainnya. Pelaku usaha harus bisa mengukur keinginan yang diharapkan konsumen satu dengan konsumen lainnya sesuai dengan komponen penyusun produk yang diinginkan konsumen dalam mememuhi permintaan konsumen. Oleh karena itu, diperlukan analisis lebih lanjut dan mendalam dari masing-masing konsumen terkait keluh kesah dan harapan yang diinginkan pelanggan atas produk yang dihasilkan produsen agar memiliki nilai kualitas dari sudut pandang konsumen. Iswanto– Strategi Optimaisasi Kualitas 65 Para pelaku usaha harus memiliki perencanaan dan strategi yang tepat dalam mempengaruhi keputusan para konsumennya untuk tetap menjadi pelanggan dan mampu memenuhi kebutuhan konsumen agar memutuskan tetap menggunakan suatu produk. Adapun pemasaran dari suatu produk mencakup kegiatan menyelidiki dan mengetahui apa yang diinginkan oleh konsumen nya, perencanaan dan pengembangan suatu produk dan jasa yang akan memenuhi keinginan tersebut, dan membuat suatu keputusan memenuhi apa yang diinginkan oleh konsumennya. 3 Perilaku konsumen yang mampu merangsang konsumen lainnya untuk ikut memilih produk atau jasa dalam satu produk/jasa yang sama dan tetap adalah konsumen yang memiliki sikap loyal.4. Adanya perencanaan dan strategi atas kualitas produk dalam membangun konsumen untuk loyal juga menjadi manajemen bagi pelaku usaha produksi kopi “Giri Kopi” yang ada di Kelurahan Ploso, Kecamatan Nganjuk, Kabupaten Nganjuk. Giri kopi ini merupakan salah satu industri kecil menengah yang memproduksi kopi olahan menjadi bubuk kopi kemasan. 3Juni Iswanto, Puji Winarko, and Zainatun Munawarah, Analisis Hukum Islam Terhadap Strategi Pemasaran Usaha Konveksi CV. Nuratina Collection Dalam Menghadapi Persaingan Bisnis Di Era Industri 4.0, Jurnal Dinamika Ekonomi Syariah 08, no. 01 (2021): 42.https://ejurnal.iaipd- nganjuk.ac.id/index.php/es/article/view/160 4 Hengki Mangiring Parulian Simarmata, Erika Revida, Iskandar Kato, Manajemen Perilaku Konsumen dan Loyalitas, Medan: Yayasan Kita Menulis, 2021, 113. 5Hardani et al., Metode Penelitian Kualitatif & Kuantitatif, ed. Husna Abadi. Yogyakarta: CV. Pustaka Ilmu, 2020, 54. 6Ibid, 372. 7A Muri Yusuf, Metode Penelitian Kuantitatif, Kualitatif & Penelitian Gabungan, ed. Suwito, Kencana, Cetakan Ke (Jakarta: Kencana, 2017,395. PENDAHULUAN Selain itu, “Giri Kopi” ini juga memiliki kedai kopi yang juga memberikan pelayanan bagi konsumen untuk bisa langsung menikmati produk-produk olahan kopi asli produksi “Giri Kopi”. Selama masa pandemi hingga saat ini banyak sekali persaingan yang juga bergerak di bidang kedai warung kopi. Tetapi di Giri Kopi ini memberikan keunggulan bagi para konsumen dengan adanya produk khas yang hanya di produksi “Giri Kopi” yakni kemasan kopi bubuk “Giri Kopi”. Karena minum kopi merupakan suatu kebiasaan dan terkadang menjadi kebutuhan yang setiap hari masyarakat hampir semua lakukan sebagai teman santai, maupun minuman pembuka aktivitas. Adanya keunggulan produk bubuk kemasan “Giri Kopi” yang tetap menjaga kualitas, diharapkan dapat menjadi wujud manajemen dalam memicu konsumen bersikap loyal. METODE PENELITIAN Penelitian yang digunakan adalah pendekatan kualitatif. Alasan peneliti menggunakan pendekatan kualiatif adalah ingin mengetahui dan memaknai bagaimana pelaksanaan strategi peningkatan kualitas yang dilakukan oleh pelaku usaha “Giri Kopi” dalam membangun loyalitas konsumen serta kajian-kajian dalam etika bisnis Islam. sifat penelitian ini menggunakan jenis penelitian deskriptif, yang ingin mengurai, menjelaskan mengambarkan suatu fenomena yang terjadi. Penelitian deskriptif ini tidak perlu mencari atau menerangkan saling hubungan dan menguji hipotesis.5 Peneliti akan mendeskripsikan hasil pengamatan strategi peningkatan kualitas produk Giri Kopi dalam membangun loyalitas konsumen. Peneliti menggunakan keabsahan data dengan cara triangulasi. Triangulasi dalam uji kredibilitas data dilakukan dari berbagai sumber dengan cara dan berbagai waktu6. Triangulasi merupakan salah satu teknik dalam pengumpulan data untuk mendapatkan temuan dan interpretasi data yang lebih akurat dan kredibel.7 Triangulasi data terdapat tiga tahapan antar peneliti, triangulasi sumber, dan triangulasi teori yang akan menjelaskan dari Jurnal Dinamika Ekonomi Syariah, Vol 10, No. 1, Januari 2023, Hal 63 - 73 66 hasil wawancara dan observasi yang dilakukan di Giri Kopi Kelurahan Ploso, Kecamatan Nganjuk, Kabupaten Nganjuk. hasil wawancara dan observasi yang dilakukan di Giri Kopi Kelurahan Ploso, Kecamatan Nganjuk, Kabupaten Nganjuk. HASIL DAN PEMBAHASAN Strategi Optimalisasi Kualitas Produk Kopi Di Giri Kopi Kelurahan Ploso, Kecamatan Nganjuk Kabupaten Nganjuk HASIL DAN PEMBAHASAN Strategi Optimalisasi Kualitas Produk Kopi Di Giri Kopi Kelurahan Ploso, Kecamatan Nganjuk Kabupaten Nganjuk Strategi Optimalisasi Kualitas Produk Kopi Di Giri Kopi Kelurahan Ploso, Kecamatan Nganjuk Kabupaten Nganjuk Kualitas pelayanan perlunya layanan yang prima sebagai standar kinerja tinggi yang secara konsisten demi tercapainya ekspektasi/harapan pelanggan agar pelanggan merasa puas. Pelanggan akan menilai layanan berdasarkan persepsinya dari bagaimana karyawan berinteraksi dengannya (kualitas interaksi). Untuk mencapai keunggulan dalam kualitas pelayanan membutuhkan tidak hanya untuk pemasaran eksternal tetapi juga untuk pemasaran internal untuk memotivasi karyawan, serta pemasaran interaktif untuk menekankan pentingnya “teknologi tinggi” dan “sentuhan tinggi”. “Giri Kopi” merupakan salah satu UMKM asli Nganjuk yang melakukan inovasi dalam mengolah biji kopi asli dengan teknologi modern mesin roaster dalam menjaga kebersihan dan kehigienisan bubuk kopi yang murni100% tanpa bahan campuran.Adanya mesin modern tidak memerlukan banyak pekerja cukup 1 karyawan, bahkan owner sendiri yang melaukan proses roaster dalam memastikan tingkat kematangan biji kopi matang sempurna. Pengembangan teknologi produksi menggunakan roaster dan mesin grinder press kopi dipenuhi owner dalamupaya meningkatkan kualitas produk kopi baik bubuk kopi yang dikemas maupun kopi murni dan olahan kopi modern dengan cita rasa dan aroma yang khas bagi penikmat kopi spesial.Inovasi bubuk kopi kemasan memiliki keunggulan yang praktis dan mudah disajikan dimanapun dengan kualitas tetap terkamin 100% kopi asli Robusta maupun arabika. “Giri kopi” melakukan optimalisasi kualitas produk baikbubukkemasan maupun menu minuman kopi kedai dengan memenuhi aspek₋aspek bentuk kemasan dari bahan yang halal, bersih, dari alumunium foil, ruangan produksi yang steril, aspek desain kemasan mudah diingkat dengan slogan “Sing Tenang Lek Mu Ngopi”, keandalan dan ketepatan mutu dengan proses produksi yang dilengkapi temperatur pengatur suhu agar biji kopi tetap memilikikeunggulan aroma dan kemurniannya.Upaya optimalisasi ini dilakukan dalamamenciptkan produk berkualitas dan bermanfaat bagi masyarkat.Karena masyarakat juga dapat menjadi reseller dalam melakukan jual beli dengan unsurkualitas yang jelas sudah terbukti karena sudah mendapatkan ijin dari BPOM dan MUI.Sehingga konsumen dapat membuktikan sendirikualitas dengan mencobanya di kedai langsung. Hal ini sejalan dengan firman Allah SWT dalam Al Baqoroh ayat 42: َوَ َلَ تَلۡبسُواْ ٱلۡحَقَّ بٱلۡبََٰطل وَ تَكۡتُمُواْ ٱلۡحَقَّ وَ أَنتُمۡ تَعۡلَمُون Wa lā talbisul haqqa bil bāthili wa taktumul haqqa wa antum ta‘lamūna. Wa lā talbisul haqqa bil bāthili wa taktumul haqqa wa antum ta‘lamūna. Artinya, “Jangan kalian mencampur kebenaran dengan kebatilan.Jangan juga kalian menyembunyikan kebenaran.Padahal kalian menyadarinya,” (Surat Al-Baqarah ayat 42). Pada ayat ini, Allah memberikan larangan kepada Bani Israil untuk tidak mencampuradukkan antara kebenaran dan kebatilan. Dan janganlah kamu, wahai Bani Israil, campuradukkan kebenaran dengan kebatilan dengan memasukkan apa yang bukan firman Allah ke dalamKitab Taurat, dan janganlah kamu sembunyikan kebenaran firman-firman Allah seperti berita akan datangnya Nabi Muhammad, sedangkan kamu mengetahuinya. Orang-orang Yahudi menyembunyikan berita tentang kedatangan Nabi Muhammad yang termaktub di dalam Taurat Iswanto– Strategi Optimaisasi Kualitas 67 dengan maksud untuk menghalangi manusia beriman kepadanya. dengan maksud untuk menghalangi manusia beriman kepadanya. dengan maksud untuk menghalangi manusia beriman kepadanya. Kualitas Produk “Giri Kopi” diilakukan dengan optimal dengan mempekerjakan barista yang sudah berpengalaman dalam memahami persepsi konsumen dan evaluasi kualitas produk karena kepuasan pelanggan dibuat berdasarkan kepercayaan.Proses pembuatan kopi dapat dibuktikan langsung dengan konsumen melihat para barista menyajikan kopi murni saat biji kopi yang masih utuh di press dengan mesin grinder yang akan menghasilkan tetes demi tetes sari kopi murninya.Untukkemasan baik botol, drip bag dan standing pouch dikemas dengan ukuran berat yang jelas tanpa adanya penguragan timbangan.Owner “Giri Kopi” mengupayakan selalu meningkatkan alat modern agar kualitas produk inovasinya tersu berkualitas dan adanya barista yang memiliki ketrampilan diharapkan terus menambah minat konsumen yang datang dengan sajian minuman kopi berkualitas.Barista selalu menyampaikan informasi seluruh macam variansi”Giri Kopi”dengan detail jelas apasaja yang dapat dipesan dan rasa yang diberikan. Upaya Barista GiriKopi yang menyajikan informasi dengan jujur dalam upaya menginformasikan kualitas dari sajian menu”Giri Kopi” sesuai dengan ajaran Islam juga diajarkan bahwasanya dalam menginformasikan segala sesuatu kepada orang lain kita harus jujur sesuai dengan kenyataan nya, karena dengan demikian kita akan merasa lega dan orang lain akan merasa puas.8 Hal ini membuktikan bahwa “Giri KOPI” mengutamakan kualitas produk dengan pengembangan teknologimodern dicirikan dalam Lima aspek, yaitu:berorientasi pada pelanggan, adanyapartisipasi aktif yang dipimpin olehmanajemen puncak, adanya pemahamandari setiap orang terhadap tanggung jawabspesifik untuk berkualitas, adanya aktivitasyang berorientasi pada tindakan pencegahan kerusakan dan adanya suatufilosofi yang menganggap bahwa kualitasmerupakan jalan hidup (way of life).Manajemen kualitas di definisikan sebagaisuatu cara meningkatkan performansi,secara terus menerus (continuousimprovement) pada setiap level operasiatau proses, dalam setiap area fungsionaldari suatu organisasi dengan menggunakansumber daya manusia dan modal yangtersedia.Beberapa hal penting yangterkandung dalam definisi tersebut adalahadanya perencanaan kualitas, pengendaliankualitas, jaminan kualitas, dan peningkatankualitas.9 9 Ahmad Syaickhu, Bhaswarendra Guntur Hendratri, dan Binti Lu’lu’il Maknuun, Analisis Peningkatan Kualitas Produksi Abon Dalam Rangka Meningkatkan Pendapatan Pada UD. Tunas Mulya. Jurnal Dinamika Ekonomi Syariah Vol 9, No.1, 2022, 50.https://ejurnal.iaipd- nganjuk.ac.id/index.php/es/article/view/331 8Juni Iswanto, Winarko, and Munawarah, Analisis Hukum Islam Terhadap Strategi Pemasaran Usaha Konveksi CV. Nuratina Collection Dalam Menghadapi Persaingan Bisnis Di Era Industri 4.0. Jurnal Dinamika Ekonomi Syariah, 8(1), 48-62.https://ejurnal.iaipd- nganjuk.ac.id/index.php/es/article/view/160 Strategi Membangun Loyalitas Pelanggan Di Giri Kopi Kelurahan Ploso, Kecamatan Nganjuk Kabupaten Nganjuk Loyalitas pelanggan merupakan gambaran keberhasilan para entrepreneur dalam menjalankan usahanya.Loyalitas dengan memberikan kualitas dari hasil penelitian ini artinya semakin tinggi kualitas layanan yang diberikan oleh “Giri Kopi” , maka pelanggan akan semakin loyal. “Giri Kopi”secara normal sudah melayani pelanggan dengan baik seperti penggunaan staf yang berpengalaman dan waktu tunggu yang rendah akan mendorong pelanggan untuk menjalin hubungan baik dengan perusahaan dalam jangka panjang untuk berkunjung kembali ke kedai kopi. “Giri Kopi”hanya perlu mempertahankan yang sudah dilakukan. Namun “Giri Kopi”belum membuat program pelayanan untuk lebih mempertahankan Jurnal Dinamika Ekonomi Syariah, Vol 10, No. 1, Januari 2023, Hal 63 - 73 68 loyalitas pelanggan mereka berupa membership card.Sebagai contoh misalnya seperti Starbucks yang lebih mengikat pelanggan dengan Starbucks Reward dengan integrasi aplikasi dan kartu yang dimana pelanggan yang loyal bisa mendapat bonus-bonus potongan harga dan penawaran khusus tertentu.Maknaya adalah pelanggan “Giri Kopi”belum tentu loyal meskipun mendapat produk yang berkualitas. Pelanggan “Giri Kopi”belum tentu bersedia untuk datang lagi di lain waktu, merekomendasikan “Giri Kopi”kepada orang lain, maupun lebih percaya “Giri Kopi”dibanding produk kedai kopi lain. pelanggan belum tentu loyal terhadap “Giri Kopi”meskipun merasa puas terhadap produk. Walaupun kopi beserta variasi minuman lainya di “Giri Kopi”rasanya enak dan membuat mereka puas, namun hal tersebut tidak menjadi alasan mereka untuk tetap menjadi pelanggan setia di “Giri Kopi” , karena bisa jadi mereka juga ingin berganti kedai kopi lain untuk mencoba-coba. Kepuasan pelanggan pastinya akan mempengaruhi loyalitas pelanggan. Hasil dari penelitian ini menunjukan bahwa semakin tinggi kepuasan maka pelanggan “Giri Kopi”akan semakin loyal terhadap “Giri Kopi”. Hal ini menunjukkan bahwa semakin tinggi kualitas layanan dan efisiensi yang diberikan kepada pelanggan, maka secara langsung akan tercipta kepuasan pelanggan yang pada akhirnya mempengaruhi tingkat loyalitas pelanggan. Hal ini sejalan dengan penelitian terdahulu yang juga menunjukkan bahwa dengan adanya peningkatkan kepuasan pelanggan dengan melalui produk berkualitas harus dioptimalkan dengan baik.Konsumen yang merasa puas akan membantu peningkatkan penjualan dengan semakin banyaknya masyarakat yang mengetahui kualitas suatu produk dari adanya pemasaran yang dilakukan. Sehingga tidak langsung kepuasan atas kualitas produk akan berdampak pada Tindakan loyalitas konsumen.10 g j p p 11 Mashuri, Analisis Dimensi Loyalitas Pelangan Berdasarkan Perspektif Islam. IQTISHADUNA: Jurnal Ilmiah Ekonomi Kita. Juni 2020, Vol.9, No.1: 54-64.https://ejournal.stiesyariahbengkalis.ac.id/index.php/iqtishaduna/article/view/212 10 Juni Iswanto, Winarko, and Munawarah, Analisis Hukum Islam Terhadap Strategi Pemasaran Usaha Konveksi CV. Nuratina Collection Dalam Menghadapi Persaingan Bisnis Di Era Industri 4.0. Jurnal Dinamika Ekonomi Syariah, 8(1), 48-62.https://ejurnal.iaipd- nganjuk.ac.id/index.php/es/article/view/160 Dalam Menghadapi Persaingan Bisnis Di Era Industri 4.0. Jurnal Dinamika Ekonomi Syariah, 8(1), 48-62.https://ejurnal.iaipd- nganjuk.ac.id/index.php/es/article/view/160 11 Mashuri, Analisis Dimensi Loyalitas Pelangan Berdasarkan Perspektif Islam. IQTISHADUNA: Jurnal Ilmiah Ekonomi Kita. Juni 2020, Vol.9, No.1: 54-64.https://ejournal.stiesyariahbengkalis.ac.id/index.php/iqtishaduna/article/view/212 g p Q j g 13Kementrian Agama Republik Indonesia. Al-Qur’an dan terjemahan. Jakarta: Kemenag, 2019, 89. Teori Ekonomi Syariah Terhadap Optimalisasi Kualitas Produk Kopi Dalam Membangun Loyalitas Pelanggan di Giri Kopi Kelurahan Ploso, Kecamatan Nganjuk Kabupaten Nganjuk Perhatikan firman Allah dalam surat al-Maidah ayat 54-55 yang sebagai berikut: ََٰي ٓ أَيُّهَا ٱلَّذِينَ ءَامَنُواْ مَن يَرۡ تَدَّ مِنكُمۡ عَن دِينِهِۦ فَسَوۡ فَ يَأۡتِي ٱ َِّللَّ ُ ب ُقَوۡ مٖ يُحِ بُّه َمۡ وَ يُحِ بُّونَهُۥ أَذِلَّةٍ عَل ٍى ٱلۡمُؤۡ مِنِينَ أَعِزَّة َعَلَى ٱلۡكََٰفِرِ ينَ يُجََٰ هِدُونَ فِي سَبِيلِ ٱَّللَّ ِ وَ َلَ يَخَافُونَ لَوۡ مَةَ َل ٖئِم َذََٰلِك ِ َّفَضۡ لُ ٱَّلل ُي ُ َّؤۡ تِيهِ مَن يَشَا ءُ وَ ٱَّلل ََٰو سِعٌ عَلِيمٌ إِنَّمَا َوَ لِيُّكُمُ ٱَّللَّ ُ وَ رَسُولُهُۥ وَ ٱلَّذِينَ ءَامَنُواْ ٱلَّذِينَ يُقِيمُونَ ٱلصَّلَوَٰ ة ُ وَ يُؤۡ ت َونَ ٱلزَّكَوَٰ ة َوَ هُمۡ رََٰ كِعُون g y y g g ََٰي ٓ أَيُّهَا ٱلَّذِينَ ءَامَنُواْ مَن يَرۡ تَدَّ مِنكُمۡ عَن دِينِهِۦ فَسَوۡ فَ يَأۡتِي ٱ َِّللَّ ُ ب ُقَوۡ مٖ يُحِ بُّه َمۡ وَ يُحِ بُّونَهُۥ أَذِلَّةٍ عَل ٍى ٱلۡمُؤۡ مِنِينَ أَعِزَّة َعَلَى ٱلۡكََٰفِرِ ينَ يُجََٰ هِدُونَ فِي سَبِيلِ ٱَّللَّ ِ وَ َلَ يَخَافُونَ لَوۡ مَةَ َل ٖئِم َذََٰلِك ِ َّفَضۡ لُ ٱَّلل ُي ُ َّؤۡ تِيهِ مَن يَشَا ءُ وَ ٱَّلل ََٰو سِعٌ عَلِيمٌ إِنَّمَا َوَ لِيُّكُمُ ٱَّللَّ ُ وَ رَسُولُهُۥ وَ ٱلَّذِينَ ءَامَنُواْ ٱلَّذِينَ يُقِيمُونَ ٱلصَّلَوَٰ ة ُ وَ يُؤۡ ت َونَ ٱلزَّكَوَٰ ة َوَ هُمۡ رََٰ كِعُون Artinya: 54. Wahai orang-orang yang beriman! Barangsiapa di antara kamu yang murtad (keluar) dari agamanya, maka kelak Allah akan mendatangkan suatu kaum, Dia mencintai mereka dan mereka pun mencintai-Nya, dan bersikap lemah lembut terhadap orang-orang yang beriman, tetapi bersikap keras terhadap orang-orang kafir, yang berjihad di jalan Allah, dan yang tidak takut kepada celaan orang yang suka mencela. Itulah karunia Allah yang diberikan-Nya kepada siapa yang Dia kehendaki.Dan Allah Mahaluas (pemberian-Nya), Maha Mengetahui.55. Sesungguhnya penolongmu hanyalah Allah, Rasul-Nya, dan orang-orang yang beriman, yang melaksanakan salat dan menunaikan zakat, seraya tunduk (kepada Allah). (Q.S al-Maidah ayat 54-55)12 Di sisi lain, seorang Mukmin tidak boleh loyal dan cinta terhadap musuhmusuh Islam. Dalam kategori hablum minanas ini, berarti kita tidak diperbolehkan bekerjasama apalagi menjual loyalitas muslim kepada hal-hal yang berbau riba. 12Kementrian Agama Republik Indonesia. Al-Qur’an dan terjemahan. Jakarta: Kemenag, 2019, 13 Teori Ekonomi Syariah Terhadap Optimalisasi Kualitas Produk Kopi Dalam Membangun Loyalitas Pelanggan di Giri Kopi Kelurahan Ploso, Kecamatan Nganjuk Kabupaten Nganjuk Teori Ekonomi Syariah Terhadap Optimalisasi Kualitas Produk Kopi Dalam Membangun Loyalitas Pelanggan di Giri Kopi Kelurahan Ploso, Kecamatan Nganjuk Kabupaten Nganjuk Islam merupakan agama yang mengatur segala dimensi kehidupan.Al-Qur’an diturunkan Allah SWT kepada manusia untuk memberikan solusi atas segala permasalahan hidup. Oleh karena itu, setiap aktivitas hidup selalu berhubungan dengan aturan secara islam. Demikian halnya dalam penyampaian jasa, setiap aktivitas yang terkait harus didasari oleh kepatuhan terhadap Syariah yang penuh dengan nilai-nilai moral dan etika. Islam merupakan agama yang mengatur segala dimensi kehidupan.Al-Qur’an diturunkan Allah SWT kepada manusia untuk memberikan solusi atas segala permasalahan hidup. Oleh karena itu, setiap aktivitas hidup selalu berhubungan dengan aturan secara islam. Demikian halnya dalam penyampaian jasa, setiap aktivitas yang terkait harus didasari oleh kepatuhan terhadap Syariah yang penuh dengan nilai-nilai moral dan etika. Loyalitas dalam Islam disebut dengan al-wala’.Secara etimologi, alwala’ memiliki beberapa makna, antara lain mencintai, menolong, mengikuti, dan mendekat kepada sesuatu.Konsep loyalitas dalam Islam atau al-wala’ adalah ketundukan mutlak kepada Allah SWT dalam wujud menjalankan syariah Islam sepenuhnya.Loyalitas pelanggan dalam Islam terjadi apabila aktivitas muamalah itu dapat memberi manfaat yang saling menguntungkan kedua belah pihak, karena terpenuhinya kewajiban serta hak masing-masing melalui penerapan nilai-nilai Islam.11 Syaikhul Islam Ibnu Taymiyah mendefinisikan Al-Wala‟ dan Al-Baro‟ Dengan ungkapanAl-Walayah kebalikan dari Al-adawah.Asal pengertian Al-walayah adalah kecintaan dan Iswanto– Strategi Optimaisasi Kualitas 69 kedekatan.Sedangkan asal pengertian Al-adawah adalah kebencian dan kejauhan.Wali artinya orang yang dekat.Dalam Bahasa Arab “hadza yali hadza” artinya ini dekat dengan ini.Seperti dalam sabda Nabi Muhammad SAW, “Serahkan ilmu waris kepada pakarnya.Bila masih ada yang menyisa dari harta warisan, maka ia menjadi milik orang yang paling dekat dengan orang yang mati”.Berwala‟ dalam Islam ini implementasinya dilakukan dengan memberikan wala‟ kepada Allah, Rasul, dan orang-orang yang beriman dalam satu kesatuan, sebagaimana disebutkan Al Qur- an, Sesungguhnya Wali kamu hanyalah Allah, Rasul-Nya, dan orang-orang yang beriman, yang mendirikan shalat dan menunaikan zakat, seraya mereka tunduk (kepada Allah). (QS. Al Maaidah: 55). Teori Ekonomi Syariah Terhadap Optimalisasi Kualitas Produk Kopi Dalam Membangun Loyalitas Pelanggan di Giri Kopi Kelurahan Ploso, Kecamatan Nganjuk Kabupaten Nganjuk kedekatan.Sedangkan asal pengertian Al-adawah adalah kebencian dan kejauhan.Wali artinya orang yang dekat.Dalam Bahasa Arab “hadza yali hadza” artinya ini dekat dengan ini.Seperti dalam sabda Nabi Muhammad SAW, “Serahkan ilmu waris kepada pakarnya.Bila masih ada yang menyisa dari harta warisan, maka ia menjadi milik orang yang paling dekat dengan orang yang mati”.Berwala‟ dalam Islam ini implementasinya dilakukan dengan memberikan wala‟ kepada Allah, Rasul, dan orang-orang yang beriman dalam satu kesatuan, sebagaimana disebutkan Al Qur- an, Sesungguhnya Wali kamu hanyalah Allah, Rasul-Nya, dan orang-orang yang beriman, yang mendirikan shalat dan menunaikan zakat, seraya mereka tunduk (kepada Allah). (QS. Al Maaidah: 55). Loyalitas dalam muamalah ini tidak hanya memperhatikan siapa saja yang memberi keuntungan bagi kita, akan tetapi harus memperhatikan perkara-perkara syar’i yang telah dituntun oleh agama Islam. Perhatikan firman Allah dalam surat al-Maidah ayat 54-55 yang sebagai berikut: ََٰي ٓ أَيُّهَا ٱلَّذِينَ ءَامَنُواْ مَن يَرۡ تَدَّ مِنكُمۡ عَن دِينِهِۦ فَسَوۡ فَ يَأۡتِي ٱ َِّللَّ ُ ب ُقَوۡ مٖ يُحِ بُّه َمۡ وَ يُحِ بُّونَهُۥ أَذِلَّةٍ عَل ٍى ٱلۡمُؤۡ مِنِينَ أَعِزَّة َعَلَى ٱلۡكََٰفِرِ ينَ يُجََٰ هِدُونَ فِي سَبِيلِ ٱَّللَّ ِ وَ َلَ يَخَافُونَ لَوۡ مَةَ َل ٖئِم َذََٰلِك ِ َّفَضۡ لُ ٱَّلل ُي ُ َّؤۡ تِيهِ مَن يَشَا ءُ وَ ٱَّلل ََٰو سِعٌ عَلِيمٌ إِنَّمَا َوَ لِيُّكُمُ ٱَّللَّ ُ وَ رَسُولُهُۥ وَ ٱلَّذِينَ ءَامَنُواْ ٱلَّذِينَ يُقِيمُونَ ٱلصَّلَوَٰ ة ُ وَ يُؤۡ ت َونَ ٱلزَّكَوَٰ ة َوَ هُمۡ رََٰ كِعُون Loyalitas dalam muamalah ini tidak hanya memperhatikan siapa saja yang memberi keuntungan bagi kita, akan tetapi harus memperhatikan perkara-perkara syar’i yang telah dituntun oleh agama Islam. Perhatikan firman Allah dalam surat al-Maidah ayat 54-55 yang sebagai berikut: ٰ ٱ أۡ َ ُ َ َّ ُ ْ ٱ َّ أَ ُّ َّلَّ ُّ ََُ أَ َّ ُّ َّ أَ ٱۡ oleh agama Islam. Teori Ekonomi Syariah Terhadap Optimalisasi Kualitas Produk Kopi Dalam Membangun Loyalitas Pelanggan di Giri Kopi Kelurahan Ploso, Kecamatan Nganjuk Kabupaten Nganjuk Oleh karenanya, dalam beberapa firman-Nya, Allah mengingatkan orang-orang beriman tentang hal ini dalam surat Ali Imran ayat 28 sebagai berikut: َََٰۖ َََٰ ََلَّ يَتَّخِ ذِ ٱلۡمُؤۡ مِنُونَ ٱلۡكََٰفِرِ ينَ أَوۡ لِيَا ءَ مِن دُونِ ٱلۡمُؤۡ مِنِينََۖ وَ م ۡن يَفۡعَل ۡ ذََٰلِكَ فَلَي َّسَ مِنَ ٱَّللَّ ِ فِي شَيۡ ءٍ إَِل ۡ أَن تَتَّقُواْ مِنۡهُم َتُقَىَٰة ۗٗ وَ يُحَذ ِرُكُمُ ٱَّللَّ ُ نَفۡسَهُۥۗٗ و ُإِلَى ٱَّللَّ ِ ٱلۡمَصِ ير Artinya: “Janganlah orang-orang mukmin mengambil orang-orang kafir menjadi wali dengan meninggalkan orang-orang mukmin. Barangsiapa berbuat demikian, niscaya lepaslah ia dari pertolongan Allah, kecuali karena (siasat) memelihara diri dari sesuatu yang ditakuti dari mereka. Dan Allah memperingatkan kamu terhadap diri (siksa)-Nya.Dan hanya kepada Allah kembali (mu)”. (Q.S Ali Imran ayat 28)13 Artinya: “Janganlah orang-orang mukmin mengambil orang-orang kafir menjadi wali dengan meninggalkan orang-orang mukmin. Barangsiapa berbuat demikian, niscaya lepaslah ia dari pertolongan Allah, kecuali karena (siasat) memelihara diri dari sesuatu yang ditakuti dari mereka. Dan Allah memperingatkan kamu terhadap diri (siksa)-Nya.Dan hanya kepada Allah kembali (mu)”. (Q.S Ali Imran ayat 28)13 “Giri Kopi” memberikan keuntungan atas sikap loyalitas yang pelanggan lakukan promosi secara tidak langsungmounth to mounth saat berbagi minuman di rumah dengan bubuk kopi 70 Jurnal Dinamika Ekonomi Syariah, Vol 10, No. 1, Januari 2023, Hal 63 - 73 70 kemasan yang di bawa kerumah maupun kepada konsumen yang baru datang dengan terus berinovasi dan memberikan sajian menuminuman kopi murni maupun olahan kopi modern yang berkualitas dan jaminan kenikmatan tanpa adanya bahan campuran yang membahayakan kesehatan kosumen. Bagi reseller “Giri Kopi” juga memberikan sistem keuntungan bagi hasil yang jelas dan menyamakan harga jual baik pembelian langsung di kedai maupun di gerai milik reseler.Hal ini dilakukan agar tidak ada persaingan antara reseler, kedai ‘Giri Kopi” maupun mitra swalayan yang juga menjual produk “Giri Kopi”. kemasan yang di bawa kerumah maupun kepada konsumen yang baru datang dengan terus berinovasi dan memberikan sajian menuminuman kopi murni maupun olahan kopi modern yang berkualitas dan jaminan kenikmatan tanpa adanya bahan campuran yang membahayakan kesehatan kosumen. Bagi reseller “Giri Kopi” juga memberikan sistem keuntungan bagi hasil yang jelas dan menyamakan harga jual baik pembelian langsung di kedai maupun di gerai milik reseler.Hal ini dilakukan agar tidak ada persaingan antara reseler, kedai ‘Giri Kopi” maupun mitra swalayan yang juga menjual produk “Giri Kopi”. p , , 15Kementrian Agama Republik Indonesia. Al-Qur’an dan terjemahan. Jakarta: Kemenag, 2019, 127. 14Khozin Zaki, Manajemen Syariah : Viral Marketing Dalam Perspektif Pemasaran Syariah Studi Kasus Pada Perusahaan Start Up Sosial Purwokerto: Penerbit Amerta Media 2020 5 14Khozin Zaki, Manajemen Syariah : Viral Marketing Dalam Perspektif Pemasaran Syariah Studi Kasus Pada Perusahaan Start Up Sosial. Purwokerto: Penerbit Amerta Media, 2020, 5. 15Kementrian Agama Republik Indonesia Al Qur’an dan terjemahan Jakarta: Kemenag 2019 127 15Kementrian Agama Republik Indonesia. Al-Qur’an dan terjemahan. Jakarta: Kem Teori Ekonomi Syariah Terhadap Optimalisasi Kualitas Produk Kopi Dalam Membangun Loyalitas Pelanggan di Giri Kopi Kelurahan Ploso, Kecamatan Nganjuk Kabupaten Nganjuk Giri Kopi juga selalu memperhatikan etika dalam menajalankan usaha dalam selalu menjaga kualitas produk dalam meningkatkan loyalitas pelanggan dengan menjaga unsur-unsur yang dilarang oleh Rasululloh dalam menjalankan usaha harus terbebas dari unsur Maghrib, singkatan dari lima unsur yaitu maysir (judi), zhulm (aniaya), gharar (penipuan), haram, riba (bunga), iktinaz, dan Bathil. 14Hal ini dilakukan Giri Kopi dengan tidak mencari keuntungan yang besar, keuntungan sesuai dengan kualitas dari produk yang dihasilkan dengan harga yang disesuaikan dengan harga pasar sehingga transaksi jual beli yang dilakukan jelas, sesuai kesepakatan dan tidak memberatkan konsumen atau reseller yang menjual kembali produk kopi kemasan dari Giri Kopi.Karyawan Giri Kopi selalu bersikap jujur dalam memberikan informasi produk yang dijual juga sudah jelas bentuk, ukuran, rasa, komposisi dan tidak menutup-nutupinya.Karena kalau usaha dijalankan dijalan yang tidak benar maka hasilnya tidak membawa berkah.Produk yang dijual juga sudah memiliki iji halal dan tidak memberikan produk berbahaya bagi masyarakat dan menggunakan mesin moden dalam menjaga kualitas produk tetap hiegenis dan suci. Untuk ijin kehalalan dan kelayakan pangan juga sudah memiliki ijin PIRT dari BPOM. Giri Kopi memiliki struktur organisasi yang tertata dalam mengatur karyawan untuk mengawasi dan melayani dengan baik para konsumen dalam memberikan kepuasan dan menghindari adanya aktivitas yang terkait perjudian maupun yang dilarang agama.Adanya struktur organisasi baik anatar karyawan dan reseller selalu memegang teguh prinsip Rasululloh yakni shidiq, amanah, fathonah dan tablig dalam selalu menjalankan usaha terus berinovasi dan berjalan secara berkepanjangan demi meningkatkan kualitas dan loyalitas pelanggan Giri Kopi. Pengaturan struktur organisasi yang baik dalam menjalankan usaha sangatlah penting. Adanya struktur dan stratifikasi dalam Islam dijelaskan dalam surah al-An’aam ayat 165 yang dikatakan: ٰٰۗ …. ٗۗۡوَ رَ فَعَ بَعۡضَكُمۡ فَوۡ قَ بَعۡضٖ دَرَ جََٰ تٖ ل ِيَبۡلُوَ كُمۡ فِي مَا ءَاتَىَٰكُم.……. Artinya: Dia mengangkat (derajat) sebagian kamu di atas yang lain, untuk mengujimu atas (karunia) yang diberikan-Nya kepadamu.15 Hal ini menjelaskan bahwa dalam mengatur kehidupan dunia, peranan manusia tidak akan sama. 16 Destri Sonatasia, Onsardi, dan Ety Arini, Strategi Meningkatkan Loyalitas Konsumen Makanan Khas Kota Curup Kabupaten Rejang Lebong. JMIB: Jurnal Manajemen Modal Insani dan Bisnis Vol 1, No 1 (2020) Teori Ekonomi Syariah Terhadap Optimalisasi Kualitas Produk Kopi Dalam Membangun Loyalitas Pelanggan di Giri Kopi Kelurahan Ploso, Kecamatan Nganjuk Kabupaten Nganjuk Ayat tersebut mengatakan bahwa kelebihan yang diberikan itu (struktur yang berbeda-beda) bukan digunakan untuk kepentingan sendiri melainkan untuk memberikan kemudahan bagi orang lain dan memberikan kesejahteraan bagi orang lain Kejujuran diperlihatkan saat transaksi jual beli dengan reseller dan penyajian produk.Sikap Iswanto– Strategi Optimaisasi Kualitas 71 amanah ditunjukkan dengan menepati kerjasama dengan reseller yang memberikan sistem bagi hasil dan diskon kepada pelanggan yang membeli secara grosir.Sikap fathonah atau kecerdasan diperlihatkan dengan selalu berinovasi dalam mengguakan mesin produksi yang canggih dalam meningkatkan kualitas dan memberikan menu yang selalu update.sikap tablig ditunjukkan dengan jalinan komunikasi yang baik antara reseller dan para mitra yang selalu menjaga komunikasi dan memberikan pelaynan 5S kepada pelanggan. Untuk menjalankan usaha yang menajaga loyalitas pelanggan harus mampu menjaga aspek- aspek yang menjadi kekuatan dalam bisnis dengan memberikan kualitas produk yang sudah baik dengan harga yang terjangkau, pelayanan ramah, cepat tanggap terhadap kebutuhan konsumen, serta kemajuan teknologi yang memberikan kemudahan dalam mempromosikan usaha. Strategi makanan memang harus mampu membuat manajemen operasional yang mampu menciptakan loyalitas kepada pelanggan agar tetap setia, sehingga terkadang diperlukan pelatihan kepada karyawan dengan memberikan bimbingan cara memperilakukan pelanggan dengan baik.16 KESIMPULAN Berdasarkan hasil paparan dan temuan data, menunjukkan bahwa Strategi yang digunakan dalam optimalisasi kualitas produk kopi di di Giri Kopi Kelurahan Ploso, Kecamatan Nganjuk Kabupaten Nganjuk dengan memperoleh bahan baku biji kopi langsung dari petani kopi, menggunakan mesin produksi modern roaster grinder dalam menghasilkan produk₋produk kopi baik kemasan atau minuman yang dapat dinikmati di tempat “Giri Kopi”. Produk kopi bubuk dikemas dengan 3 jenis ukuran berbeda botol, drip bag, dan standing pouch dengan desain kemasan hitam elegan menjaga higienisan, aroma dan rasa tetap terjaga hingga produkhabis dipakai sampai masa kadarluwarsa selama 1 tahun sesuai dengan standard perijinan BPOM dan kehalalan dari MUI yang diberikan. Kemudian Giri Kopi membangun loyalitas pelanggan dengan menjalin mitra reseller dari masyarakat, swalayan dan minimarket yang tersebar di seluruh Indonesia dengan memberikan keuntungan penjualan 15% serta potongan harga bagi pelanggan yang juga melakukan pembelian secara grosir. “Giri Kopi” juga membangun loyalitas pelanggan dari inovasi produk kemasan dengan menjaga kualitas kopi murni 100%, harga sama di seluruh mitra, pelayanan tempat kedai kopi yang menguataman pelayanan prima, senyum, salam dan sapa dengan memberikan informasi menu maupun produk yang dapat dinikmati ditempat dan dijadikan oleh₋oleh untuk keluarga dirumah. Terbukti sebagain besar pelanggan melakukan pembelian lebih dari 2 kali datang dan membeli produk serta mengajak orang yang berbeda datang ke “Giri Kopi”. Berdasarkan Teori ekonomi syariah terhadap optimalisasi kualitas produk kopi dalam membangun loyalitas pelanggan di Giri Kopi Kelurahan Ploso, Kecamatan Nganjuk Kabupaten Nganjuk menghindari adanya unsur Maghrib (maysir (judi), zhulm (aniaya), gharar (penipuan), haram, riba (bunga), iktinaz, dan Bathil) dengan menjaga kehalalal, kebersihan dan kesucian produk, kejujuran juga diterapkan dengan memberikan jaminan produk yang 100% asli dari biji kopi, mengembangjan inovasi mesin produksi, unsur kejelasan dalam penyajian langsung dapat dilihat oleh pelanggan. Sistem kerjasama dengan mitra yang menjualkan kembali produk Giri Kopi saling menguntungkan dan tidak ada unsur keterpaksaan, maupun penipuan. 72 Jurnal Dinamika Ekonomi Syariah, Vol 10, No. 1, Januari 2023, Hal 63 - 73 REFERENSI REFERENSI Al Sa‟ud, Fahd ibn “Abd al Aziz. Al - Quran dan Terjemahannya. Madinah: Mujamma‟ Al Malik Fahd Li Thiba‟at Al Mushaf Asy Syarif. 2010. Ariani, Dorothea Wahyu. Manajemen Kualitas. Tangerang: Penerbit Universitas Terbuka Press, 2021. REFERENSI Al Sa‟ud, Fahd ibn “Abd al Aziz. Al - Quran dan Terjemahannya. Madinah: Mujamma‟ Al Malik Fahd Li Thiba‟at Al Mushaf Asy Syarif. 2010. Ariani, Dorothea Wahyu. Manajemen Kualitas. Tangerang: Penerbit Universitas Terbuka Press, 2021. Daga, Rosnaini. Citra, Kualitas Produk, dan Kepuasan Pelanggan. Gowa: Global-RCI. 2017. Dewantoro, Dipo H., Wisnalmawati, & Istanto, Yuni. Pengaruh Kualitas Produk Terhadap Loyalitas Pelanggan Dimediasi Oleh Kepuasan Pelanggan Pada Coffee Shop Fihi Pekanbaru. The 4th Conference on Innovation and Application of Science and Technology (CIASTECH 2021), ISSN Cetak : 2622-1276 ISSN Online : 2622-1284, 2020. 2021), ISSN Cetak : 2622-1276 ISSN Online : 2622-1284, 2020. Firmansyah, M Anang. Pemasaran (Dasar Dan Konsep). Surabaya: CV. Penerbit Qiara Media, 2019. 2019. Firmansyah, M Anang. Pemasaran Produk dan Merek (Planning&Strategy). Surabaya: CV. Penerbit Qiara Media, 2019. Hardani et al., Metode Penelitian Kualitatif & Kuantitatif, ed. Husna Abadi. Yogyakarta: CV. Pustaka Ilmu, 2020. Hidayati, Aprihatiningrum Wa Ode Sifatu,, Maddinsyah, Denok Sunarsi, dan Jasmani, Loyalitas dan Kepuasan Konsumen: Tinjauan Teoritik, Surabaya: Cipta Media Nusantara (CMN), 2021. Ibrahim, Metode Penelitian Kualitatif. Bandung:Alfa Beta, 2015. Ibrahim, Tatang dan Rusdiana, A. Manajemen Mutu Terpadu. Bandung: Yrama Widya, 2021. Iswanto, Juni. Puji Winarko, and Zainatun Munawarah, “Analisis Hukum Islam Terhadap Strategi Pemasaran Usaha Konveksi CV. Nuratina Collection Dalam Menghadapi Persaingan Bisnis Di Era Industri 4.0,” Jurnal Dinamika Ekonomi Syariah 08, no. 01 (2021): 48–62, http://ejurnal.iaipd-nganjuk.ac.id/index.php/es/index. Kasmir. Kewirausahaan. Jakarta : PT. Raja Grafindo Persada, 2017. Kementrian Agama Republik Indonesia. Al-Qur’an dan terjemahan. Jakarta: Kemenag Mahmudati, Rina dan Indrawati, Ragil Tri. Strategi Peningkatan Pemasaran Produk Kopi Dusun Bogoran Kecamatan Sapuran. Jurnal PPKM, Vol. 6, No. 3, 172 – 175, ISSN (print): 2354- 869X 2021. Moleong, Lexy J. Metodologi Penelitian Kualitatif, Cet. 3 Eds. Bandung: PT Remaja Rosdakarya, 2018. Kotler, Phillip dan Keller, Kevin Lane. Marketing Management. England: Pearson Education Iswanto– Strategi Optimaisasi Kualitas 73 Limited, 2016. Limited, 2016. P d k S il P di M d P li i k li if Y k UNY P 2017 Primadiana Yunita, Struktur Tata Kelola Global Value Chains Produk Kopi Dalam Perdagangan Kopi Global: Studi Komparatif Kopi Indonesia Dan Kopi Vietnam, Jurnal Indonesia Sosial Sains 2, no. 5 (2021), Rosnaini Daga, Citra, Kualitas Produk, dan Kepuasan Pelanggan. Gowa: Global-RCI, Simarmata, H. M.P., Revida, Erika dan Kato, Iskandar Manajemen Perilaku Konsumen dan Loyalitas. Medan: Yayasan Kita Menulis, 2021. Sonatasia, D., Onsardi, dan Arini, Ety. Strategi Meningkatkan Loyalitas Konsumen Makanan Khas Kota Curup Kabupaten Rejang Lebong, JMIB: Jurnal Manajemen Modal Insani dan Bisnis Vol 1, No 1 (2020) Vol 1, No 1 (2020) Sugiyono. Metode Penelitian Pendidikan (Pendekatan Kuantitatif, Kualitatif,dan R&D). Bandung: Alfabeta, 2015. Suryanti, Lilik.Manajemen Pemasaran, Strategi dalam meningkatkan Loyalitas Pelanggan.Sleman : Deepublish,2015. Syaickhu, A., & Guntur Hendratri, B. Analisis Peningkatan Kualitas Produksi Abon Dalam Rangka Meningkatkan Pendapatan Pada Ud. Tunas Mulya. Jurnal Dinamika Ekonomi Syariah, 9 (No.1), 2022. https://doi.org/10.53429/jdes.v9iNo.1.331 Wanardi & Triyono. Manajemen Pemasaran. Sleman: CV. Budi Utama, 2019. Yusuf, A Muri.Metode Penelitian Kuantitatif, Kualitatif & Penelitian Gabungan, ed. Suwito, Kencana, Cetakan Ke (Jakarta: Kencana, 2017. Zaki, Khozin. Manajemen Syariah : Viral Marketing Dalam Perspektif Pemasaran Syariah Studi Kasus Pada Perusahaan Start Up Sosial. Purwokerto: Penerbit Amerta Media, 2020.
https://openalex.org/W4390509184	https://implementationscience.biomedcentral.com/counter/pdf/10.1186/s13012-023-01330-y	English	null	Protocol for a hybrid type 3 effectiveness-implementation trial of a pragmatic individual-level implementation strategy for supporting school-based prevention programming	Implementation science	2,024	cc-by	12,000	© The Author(s) 2023. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecom‑ mons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Lyon et al. Implementation Science (2024) 19:2 https://doi.org/10.1186/s13012-023-01330-y Lyon et al. Implementation Science (2024) 19:2 https://doi.org/10.1186/s13012-023-01330-y Implementation Science Open Access Protocol for a hybrid type 3 effectiveness‑implementation trial of a pragmatic individual‑level implementation strategy for supporting school‑based prevention programming Aaron R. Lyon1* , Clayton R. Cook2, Madeline Larson2, Maria L. Hugh3, Alex Dopp4, Corinne Hamlin5, Peter Reinke5, Mahasweta Bose5, Amy Law1, Roger Goosey1, Annie Goerdt5, Nicole Morrell5, Alisha Wackerle‑Hollman5 and Michael D. Pullmann1 Aaron R. Lyon1* , Clayton R. Cook2, Madeline Larson2, Maria L. Hugh3, Alex Dopp4, Corinne Hamlin5, Peter Reinke5, Mahasweta Bose5, Amy Law1, Roger Goosey1, Annie Goerdt5, Nicole Morrell5, Alisha Wackerle‑Hollman5 and Michael D. Pullmann1 Abstract Background For approximately one in five children who have social, emotional, and behavioral (SEB) challenges, accessible evidence-based prevention practices (EBPPs) are critical. In the USA, schools are the primary setting for children’s SEB service delivery. Still, EBPPs are rarely adopted and implemented by front-line educators (e.g., teach‑ ers) with sufficient fidelity to see effects. Given that individual behavior change is ultimately required for success‑ ful implementation, focusing on individual-level processes holds promise as a parsimonious approach to enhance impact. Beliefs and Attitudes for Successful Implementation in Schools for Teachers (BASIS-T) is a pragmatic, multifac‑ eted pre-implementation strategy targeting volitional and motivational mechanisms of educators’ behavior change to enhance implementation and student SEB outcomes. This study protocol describes a hybrid type 3 effectiveness- implementation trial designed to evaluate the main effects, mediators, and moderators of the BASIS-T implementa‑ tion strategy as applied to Positive Greetings at the Door, a universal school-based EBPP previously demonstrated to reduce student disruptive behavior and increase academic engagement. Methods This project uses a blocked randomized cohort design with an active comparison control (ACC) condition. We will recruit and include approximately 276 teachers from 46 schools randomly assigned to BASIS-T or ACC condi‑ tions. Aim 1 will evaluate the main effects of BASIS-T on proximal implementation mechanisms (attitudes, subjective norms, self-efficacy, intentions to implement, and maintenance self-efficacy), implementation outcomes (adoption, reach, fidelity, and sustainment), and child outcomes (SEB, attendance, discipline, achievement). Aim 2 will examine how, for whom, under what conditions, and how efficiently BASIS-T works, specifically by testing whether the effects of BASIS-T on child outcomes are (a) mediated via its putative mechanisms of behavior change, (b) moderated by teacher factors or school contextual factors, and (c) cost-effective. Correspondence: Aaron R. Lyon lyona@uw.edu Full list of author information is available at the end of the article Correspondence: Aaron R. Lyon lyona@uw.edu Full list of author information is available at the end of the article *Correspondence: Aaron R. Lyon lyona@uw.edu Full list of author information is available at the end of the article Addressing children’s social, emotional and behavioral health At least one in five children experience social, emo- tional, and behavioral/mental health (SEB) challenges [1], making accessible SEB prevention and intervention programming a high priority. When children experience SEB challenges, they are at an increased risk of academic and social difficulties in school and long-term experi- ence with the judicial system, substance use problems, and unemployment [1]. In contrast, children who receive preventive SEB support experience social and academic gains into adulthood [2–6]). Given the widespread need to address children’s SEB, exacerbated by the COVID-19 pandemic [7, 8], accessible prevention programming is a high priority. Most implementation frameworks include critical indi- vidual implementer factors [30, 31]. Indeed, front-line professionals—such as teachers—are ultimately respon- sible for the adoption and delivery of EBPPs and present their own set of implementation determinants such as attitudes, beliefs, and intentions to implement an EBPP [30, 32, 33]. Some school-based research has docu- mented that individual-level determinants can be more predictive of EBPP implementation than organizational factors, such as climate or culture [32, 34], even within supportive implementation contexts [35, 36]. © The Author(s) 2023. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecom‑ mons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Page 2 of 14 (2024) 19:2 Lyon et al. Implementation Science Lyon et al. Implementation Science (2024) 19:2 Discussion This study will provide a rigorous test of BASIS-T—a pragmatic, theory-driven, and generalizable imple‑ mentation strategy designed to target theoretically-derived motivational mechanisms—to increase the yield of stand‑ ard EBPP training and support strategies. Trial registration ClinicalTrials.gov ID: NCT05989568. Registered on May 30, 2023. Trial registration ClinicalTrials.gov ID: NCT05989568. Registered on May 30, 2023. Keywords Individual determinants, Implementation strategy, Theory of planned behavior, Health action p Keywords Individual determinants, Implementation strategy, Theory of planned behavior, Health action process approach, Education sector, Mental and behavioral health Need for improved implementation supports p p pp As with other service sectors, EBPPs for school settings are adopted inconsistently and frequently delivered with poor fidelity [21–23]. In the education sector, this imple- mentation gap has been resistant to change despite inter- vention at federal and state policy levels [24]. Even studies of quality implementation strategies, such as coaching and consultation, demonstrate that many EBPPs fail to be adopted by school-based implementers [25–27]. As a result, the potential public health impact of SEB-focused EBPPs is greatly diminished [28]. • Findings will fill a gap in the literature surrounding the utility of pragmatic, individual-level strategies for pre- ventive interventions as well as the variables (mecha- nisms) through which the strategies operate, and under which conditions the strategies work. Contributions to the literature school-delivered SEB practices have increasingly been prioritized in policy and legislation [12–14]. In particular, there are a variety of universal evidence-based preven- tion practices (EBPPs) that exist to address children’s SEB challenges [15, 16]. Among these are high-quality, effec- tive, universal EBPPs delivered at the classroom or school level to support student SEB health [17–20]. • Evidence-based prevention programs are often not adopted or delivered with sufficient fidelity to address children’s social, emotional, and behavioral health needs, even when conducive organizational supports are in place. • This study tests a parsimonious and pragmatic indi- vidually focused implementation strategy delivered to frontline professionals (teachers) to enhance their implementation of an evidence-based prevention pro- gram. We will evaluate the effects on proximal out- comes/mechanisms (e.g., attitudes) as well as imple- mentation (e.g., adoption) and children’s outcomes (e.g., social, emotional, and behavioral status). • This study tests a parsimonious and pragmatic indi- vidually focused implementation strategy delivered to frontline professionals (teachers) to enhance their implementation of an evidence-based prevention pro- gram. We will evaluate the effects on proximal out- comes/mechanisms (e.g., attitudes) as well as imple- mentation (e.g., adoption) and children’s outcomes (e.g., social, emotional, and behavioral status). Implementation determinants in schools Like other health service sectors, schools are multi-level implementation contexts with myriad priorities, deci- sion-makers, implementers, and recipients of the inter- vention. Across sectors, organizational influences on implementation have been the subject of considerable research, but creating organizational change is time-con- suming and expensive, often lasting years [29]. Further- more, even with appropriate implementation support to address organizational-level barriers or enact organiza- tional facilitators, educators’ EBPP implementation can still be stilted. School‑based prevention Schools are the most common setting in which children and adolescents in the USA receive both preventive and indicated care for SEB concerns [9–11]. As a result, Page 3 of 14 Lyon et al. Implementation Science (2024) 19:2 Lyon et al. Implementation Science (2024) 19:2 Lyon et al. Implementation Science Individual‑level implementation strategiesi Beliefs and attitudes for successful implementation in schools for teachers (BASIS‑T) Beliefs and attitudes for successful implementation in schools for teachers (BASIS‑T) Despite significant research on cornerstone implemen- tation strategies such as training and consultation [37, 38], additional approaches are often needed to change behavior. Individual theories of behavior change can be leveraged to further facilitate EBPP use. The current study applies the theory of planned behavior (TPB) [39] and the health action process approach (HAPA) [40]. The TPB states that an individual’s subjective norms (their perceptions of the social importance of performing the behavior), attitudes (appraisals of the behavior), and task self-efficacy (perceived behavioral control and confidence in their ability to implement the behavior), cumulatively predict intentions to perform a behavior [41]. Intention is a strong predictor of behavior change [42–44]. The HAPA augments the TPB by (1) positing that the inten- tion to engage in a behavior is influenced by one’s out- come expectancies and perceived risks (beliefs about the possible consequences of a behavior and risks of not engaging in a behavior, factors that we cluster with “atti- tudes” in the TPB) and (2) emphasizing individual voli- tion (initial action planning and planning for coping with barriers) that increase an implementers’ maintenance self-efficacy (the belief that one is capable of overcom- ing barriers while implementing the behavior) and facili- tate the link between intentions and behavior. The most common implementation strategies, such as workshops, coaching, and consultation, primarily target knowledge and skills while often neglecting to explicitly attend to norms, attitudes, intentions, and volition. BASIS-T is designed to address individual-level mecha- nisms of behavior change (e.g., self-efficacy) often miss- ing from standard EBPP training that relate to motivation prior to receiving training and volition after training. It is an EBPP-agnostic implementation strategy designed to be delivered within the preparation/adoption phase, immediately prior to active implementation [45]. BASIS- T targets behavioral intentions via improvement in atti- tudes, subjective norms, and self-efficacy. Our theory of change (Fig. 1) shows the core BASIS-T components, mechanisms of change (volitional and motivational), implementation outcomes, and resulting child SEB out- comes within the current study. The BASIS-T strategy is grounded in the TPB and in the HAPA strategies of action planning (specifying the “when,” “where”, and “how” of implementing the EBPP) and problem-solving planning (generating solutions to specific barriers that one anticipates encountering when adopting a new prac- tice) to overcome barriers to implementation. Preliminary BASIS‑T studiesh The BASIS strategy was developed via an iterative user- centered design approach [46]. Initial pilots of BASIS-T and a version designed for school-based mental health clinicians (BASIS-C) have demonstrated promise in enhancing participants’ attitudes, subjective norms, self-efficacy, and adoption [30, 33, 47]. The BASIS-T pilot study on which the current project is based was BASIS Volitional Components BASIS Motivational Components Covariates Implementation Outcomes Attitudes Self-Efficacy Subjective norms Intentions to Implement Reach Strategic education (e.g., benefits of implementing EBPPs) Social influence (e.g., social proofing; inducing cognitive dissonance) Adoption Teacher (i.e., SES, race) School-level: Leadership, Climate Motivational Interviewing (e.g., elicit self-motivational statements) Fidelity Maintenance Self-Efficacy Student Behavioral Outcomes Sustainment Action Planning Problem-Solving Planning Fig. 1 BASIS-T theory of change: components, hypothesized mechanisms of change and target outcomes BASIS Motivational Components Strategic education (e.g., benefits of implementing EBPPs) Attitudes Fidelity Social influence (e.g., social proofing; inducing cognitive dissonance) Intentions to Implement Subjective norms Student Behavioral Outcomes Reach Sustainment Motivational Interviewing (e.g., elicit self-motivational statements) Self-Efficacy Maintenance Self-Efficacy Fig. 1 BASIS-T theory of change: components, hypothesized mechanisms of change and target outcomes Page 4 of 14 Lyon et al. Implementation Science (2024) 19:2 Lyon et al. Implementation Science (2024) 19:2 Lyon et al. Implementation Science an attention-control randomized trial conducted with 82 elementary school teachers implementing a univer- sal, classroom-based EBPP [48–50]. That study found a statistically significant positive impact on implementa- tion task self-efficacy and outcome expectancy imme- diately after training and significantly less decline in task self-efficacy than attention control throughout the academic year [33]. Positive attitudes towards evidence- based practices decreased for both groups over time but with a marginal time trend in favor of less decrease for the BASIS-T condition (p = 0.08). Except for ownership/ role, all other mechanisms (outcome expectancies, sub- jective norms, self-efficacy, and intentions to implement) deteriorated for both groups after the post-BASIS time- point (i.e., during active implementation), yet all effect sizes were in favor of less deterioration in the BASIS-T condition. Significantly higher proportions of teachers in BASIS-T immediately adopted the EBPP (74% of BASIS- T condition, 40% of attention control). With marginal significance, fidelity to the EBPP remained steady for the BASIS-T group over time but deteriorated for attention control (p = 0.052), and the BASIS-T condition engaged in the EBPP more frequently (p = 0.097). Aim 1: Experimentally evaluate the effects of BASIS-T versus active comparison control (ACC) Aim 1 will evaluate the main effects of BASIS-T on proximal mechanisms (attitudes, subjective norms, self-efficacy, intentions to implement, and maintenance self-efficacy), implementation outcomes (adoption, reach, fidelity, and sustainment), and student outcomes (classroom aggregated grades, test scores, attendance, and teacher ratings of classroom on-task behavior, dis- ruptive behavior, and prosocial behavior). Research question (RQ) 1a. Is BASIS-T more effec- tive than the ACC condition at producing changes in proximal mechanisms of behavior change? RQ 1b. Is BASIS-T more effective than the ACC condition in promoting implementation outcomes? RQ 1c. Is BASIS-T more effective than the ACC condition in promoting meaningful changes in stu- dent SEB and academic outcomes? Aim 2: Evaluate how, for whom, under what con- ditions, and how efficiently BASIS-T works to improve outcomes Aim 2 will evaluate the effects of BASIS-T on stu- dent outcomes via the mechanisms of implementation behavior change and if those effects are moderated by teacher factors and school contextual factors. We will also explore how mechanisms are linked to implemen- tation outcomes for “hypothesis-defying residuals” (i.e., teachers whose attitudes, subjective norms, and self-efficacy surrounding EBPP implementation are inconsistent with their documented implementation behaviors). Preliminary BASIS‑T studiesh A prelimi- nary analysis estimated the cost of BASIS-T at $256 per teacher based on time and material costs. Notwithstand- ing the promise of these findings, this was an under- powered pilot trial designed to lay a foundation for the current study. behavior [53]. However, consistent with more general research on universal EBPPs, results also suggested that some of the teachers delivering PGD struggled with ini- tial adoption, with two of the five teachers in the PGD condition (40%) requiring extra consultative support due to initial low levels of implementation. Aim 1: Experimentally evaluate the effects of BASIS-T versus active comparison control (ACC) Aim 1: Experimentally evaluate the effects of BASIS-T versus active comparison control (ACC) RQ 2a. Are the effects of BASIS-T mediated via mechanisms of behavior change? RQ 2b. Are the effects of BASIS-T on implementa- tion and student outcomes moderated by teacher- level factors (e.g., demographics, stress, baseline intentions to implement) and school-level factors (e.g., geographic location, school demographics, supportive leadership, implementation climate)? RQ 2c. What explains “residual” teachers whose implementation behaviors are not accounted for by the mediation model? RQ 2d. What are the costs and cost-effectiveness of BASIS-T? RQ 2d. What are the costs and cost-effectiveness of BASIS-T? Objectives and aims Th ll This project will conduct a hybrid type 3 effectiveness- implementation randomized trial to evaluate the effects of BASIS-T on implementation mechanisms and out- comes when applied to positive greetings at the door (PGD), a low-burden, universal EBPP that has been found to reduce disruptive behavior and increase aca- demic engagement [51–53], both important indicators of positive SEB functioning [54–57]. We will also examine for whom, under what conditions, and how efficiently BASIS-T works to improve outcomes. RQ 2a. Are the effects of BASIS-T mediated via mechanisms of behavior change?f PGD is a preventive classroom management strategy based on three major themes: (a) classroom climate, (b) pre-correction, and (c) positive reinforcement [53]. PGD has been found to be well aligned with school settings and effective at addressing SEB needs. Multi- ple studies have found increases in on-task behavior in middle school students, and reductions in latency- to-task engagement in high school students [51, 52]. These findings were replicated in a longitudinal rand- omized controlled efficacy trial conducted with 203 students across 10 classrooms, finding improvements in academic engagement and decreases in disruptive RQ 2d. What are the costs and cost-effectiveness of BASIS-T? Lyon et al. Implementation Science (2024) 19:2 Page 5 of 14 Page 5 of 14 Lyon et al. Implementation Science (2024) 19:2 Lyon et al. Implementation Science Methodh This hybrid type 3 effectiveness-implementation trial employs a blocked randomized cohort design with an active comparison control (ACC) condition to pro- vide a rigorous initial test of the efficacy of BASIS-T in authentic elementary school settings (see Additional file 1 for SPIRIT checklist). Schools will be randomized to BASIS-T or ACC conditions (see Fig. 2) and data will be gathered at the teacher/classroom level. There will be two cohorts of participants—one for each of two academic years—and multiple time points of data col- lection over 18 months across implementation and sus- tainment phases. Institutional review board approval has been obtained (Additional file 2), which includes plans for de-identification and secure data storage as well as tracking and reporting of adverse events or pro- tocol modifications if needed. Teachers from schools in the USA will be recruited to participate. Inclusion criteria include being a teacher at an elementary (typically K–5th grade or K–8th) school and not having been trained or supervised in delivering PGD in the past 5 years. We will recruit approximately 276 teachers from 46 elementary schools in the USA. The final balance of teachers and schools is dependent on recruitment, with the goal of meeting minimum sta- tistical power. Schools will be approached to participate via multiple routes, including leveraging existing rela- tionships and networks, educational listservs, and post- ing on social media. Interested school representatives may respond using the interest survey linked in flyers or by contacting the research team via email. The general- izer (generalizer.org), a free, web-based tool for select- ing schools for randomized controlled designs that are Fig. 2 CONSORT diagram Lyon et al. Implementation Science (2024) 19:2 Lyon et al. Implementation Science Page 6 of 14 the classroom, (c) providing specific pre-corrective state- ments privately to individual students who have difficulty self-regulating their behavior, and (d) providing spe- cific praise and encouragement to students to reinforce desired behaviors [53]. statistically representative of a chosen inference popu- lation, will also be used to help assure the representa- tiveness of our sample [58]. Strata will be created on urbanicity, school race/ethnicity, school percent female, school percent free and reduced lunch, school size, school percentage English language learners, and the number of schools in the district. Table 1 BASIS-T strategy components Table 1 BASIS-T strategy components Motivational components (TBP mechanisms) Strategic education (attitudes) Connecting EBPP to student success Maintaining an internal locus of control Social influence (subjective norms) Providing normative information Testimonials from experts Testimonials from similar others Evoking public commitments Motivational interviewing (self-efficacy) Elicit-provide-elicit structure Professional values clarification activity Pros and cons activity to illicit change talk Anticipating implementation barriers Values-directed goal setting “Ruler questions” (e.g., how confident are you) Volitional components (HAPA mechanism) Action planning and problem-solving planning (Maintenance self-efficacy) Action planning to initiate implementation Problem-solving planning to overcome barriers Implementation conditions BASIS‑T strategy Once schools have been selected, teacher recruit- ment will proceed for both cohorts of schools with the assistance of site administrators. Principals will provide us with the email addresses of teachers in their schools. Research staff or school leadership will contact eligi- ble participants by email or phone to describe the pur- pose of the study, research procedures, and incentives. Informed consent will be collected online prior to train- ing (Additional file 3) and teachers will be free to decline participation. Monetary incentives will be provided to participating teachers and schools. BASIS‑T motivational components The BASIS-T imple- mentation strategy integrates three core motivational components (Table 1). First, the BASIS-T facilitator pro- vides strategic education about implementing EBPP and overcoming barriers via maintaining an internal locus of control to improve attitudes. The second component is social influence techniques to alter perceptions of sub- jective norms, which consists of two broad approaches: (1) social proofing messages using data or testimonials to describe the behavior or attitudes of others and (2) techniques to induce cognitive dissonance. Social proofs have been used to reduce problem behaviors, includ- ing alcohol/drug use and disordered eating behaviors [62–64]. Techniques to induce cognitive dissonance operate on the premise that individuals strive for consist- ency between attitudes and actions [65]. Desired behav- iors can be increased by evoking commitments that are active, public (vs. private), and voluntary (vs. coerced) [66]. Third, motivational interviewing (MI) is used to Motivational components (TBP mechanisms) Strategic education (attitudes) Connecting EBPP to student success Randomizationh This study will employ a randomized cluster-blocked cohort design with random assignment at the school level to eliminate the possibility of condition contami- nation among teachers. Schools will be blocked within the district by a number of teachers participating in the study, school enrollment, % of non-White students, % of students who qualify for free/reduced lunch, and mean teacher baseline BASIS-T mechanisms of change (e.g., self-efficacy). These variables were chosen because they are associated with EBPP use and student academic out- comes [59]. We will create matched pairs using the near- est neighbor approach [60] and randomly assign schools within pairs to condition. Randomization will occur to the greatest extent possible, although there may be some situations where trainers in only one condition are avail- able when school staff are available to be trained; these situations will be considered essentially random. School personnel and participants will be masked to condition. Table 1 BASIS-T strategy components Intervention All participants will receive standard tele-delivered train- ing on PGD from educational consultants with expertise in its implementation and in school-based EBPP train- ing more generally. PGD is a proactive classroom man- agement strategy [61] that takes a prevention-based approach to responding to behavioral needs in the classroom [53]. Research shows that PGD can increase student-level outcomes such as on-task behavior and decrease disruptive behavior [51–53]. PGD was designed to facilitate smoother transitions in the classroom by (a) connecting with each student by greeting them by name, (b) using pre-corrective statements with the entire class to communicate expected behaviors for transitions into Volitional components (HAPA mechanism) Lyon et al. Implementation Science (2024) 19:2 Lyon et al. Implementation Science Page 7 of 14 of potential solutions to implementation barriers from which they can select when encountering challenges and set value-congruent goals related to the upcoming EBPP training. enhance self-efficacy. MI is a collaborative, person-cen- tered approach to elicit and strengthen motivation to change [67]. MI has been adapted as a brief intervention with strong evidence, feasibility, and acceptability among school-based mental health clinicians [68]; shown to improve self-efficacy and implementation among teach- ers and primary care providers [68, 69]; and used in group contexts to promote change [47]. The BASIS-T facilitator uses group MI techniques by adopting an empathic, non- directive, and person-centered style to elicit self-moti- vational statements, encourage discussion of potential changes (“change talk”) and enhance self-efficacy. The BASIS-T post-training session (75 min) provides protected time and a structured experience to develop action plans and problem-solving plans. Teachers will be provided with an action planning template to detail what PGD components they will use, how, with whom, where/when, and the environmental cues and resources needed to serve as prompts to deliver PGD with fidelity. The problem-solving plan involves teachers anticipating situational barriers and generating solutions to overcome those barriers to develop personalized “if–then” plans for dealing with specific barriers. Teachers share their plans with colleagues to receive input and feedback and to pub- licly set values-based goals for implementation. BASIS‑T volitional components To address the inten- tion-behavior gap, BASIS-T includes volitional planning interventions to increase the likelihood that teachers will maintain their self-efficacy by enacting specific imple- mentation behaviors associated with an EBPP. These strategies have been shown to help people enact health behaviors they are already motivated to perform [70–75]. Intervention Moreover, these planning interventions have demon- strated success in improving teacher adoption of inter- ventions for student behavior [76]. These interventions support the translation of intentions into actions through detailed planning of how to perform desired behaviors in specific contexts. Solutions are generated to situational and internal (e.g., cognitive) barriers to facilitate follow- through with the action plan. In combination, action planning and problem-solving planning increase the like- lihood of behavior change [77]. BASIS‑T fidelity The BASIS-T pre- and post-training intervention sessions will be recorded and independently coded by two trained research assistants with disagree- ments resolved through consensus dialogue [78, 79]. The research assistants will use the BASIS-T fidelity tool [47] to assess the fidelity of its delivery. Active comparison control Teachers assigned to the ACC will receive pre- and post- training experiences designed to mirror those received in the BASIS-T condition. These training experiences will be virtual (again via Zoom or similar) and approximately the same length as BASIS-T but will not contain any of the BASIS-T content or mechanisms of change. The ACC pre-training experience will define, describe, and advocate for EBPP implementation and fidelity of EBPP use in schools. Content will be delivered in modes that mirror that of BASIS-T with video content, workbooks, and didactic training. The ACC thus controls for dose, information provided, and delivery mode effects. Some trainers will provide ACC and BASIS-T to reduce the potential for trainer effects. BASIS‑T structure BASIS-T is delivered in a group- based format shortly before and after receiving EBPP (e.g., PGD) training. BASIS-T will be delivered via tele- facilitation (via Zoom or another similar video con- ferencing platform), pre-recorded video content, and electronic sharing of documents to promote scalability. BASIS-T facilitators are experienced school-based pro- fessionals. A pre-training session (75 min) targets atti- tudes, social norms, and perceived behavioral control. The pre-training opens with the facilitator engaging teachers in an activity to clarify their professional val- ues (an MI component). The facilitator uses open-ended questions to elicit change talk and reflects, summarizes, and draws themes across participant responses. The pre-training session is designed to help participants (a) explore their professional values and goals and make connections between those and EBPP training opportu- nities, (b) link EBPP delivery to improved outcomes for students, and (c) recognize common cognitive shortcuts that leave individuals vulnerable to adopting non-EBPPs. Teachers collaboratively develop an individualized menu Data analytic plan B i d i Basic data screening and descriptives will be conducted for all quantitative variables. We will explore and sta- tistically adjust for baseline equivalence between condi- tions on all individual outcomes and all school, teacher, and student variables following established guide- lines [83]. For all longitudinal modeling, the statisti- cal adjustment will use baseline intercepts as random terms; for dissimilar outcome domains, non-equivalent baseline variables will be included as covariates. Data missing at random will be modeled using full informa- tion maximum likelihood estimation or multiple impu- tation as appropriate. Quantitative surveys low implementation outcomes, but high TPB constructs) will be invited to a qualitative interview at the end of the active implementation phase to explore additional imple- mentation determinants. These teachers will be identified at the end of their first year of participation based on the results of statistical modeling, balanced between adopters and non-adopters, and BASIS-T and ACC conditions (15 to 19 interviewees total). Semi-structured phone inter- views (approximately 60 min) will be conducted at a con- venient time for identified teachers and audio-recorded for transcription and coding purposes. low implementation outcomes, but high TPB constructs) will be invited to a qualitative interview at the end of the active implementation phase to explore additional imple- mentation determinants. These teachers will be identified at the end of their first year of participation based on the results of statistical modeling, balanced between adopters and non-adopters, and BASIS-T and ACC conditions (15 to 19 interviewees total). Semi-structured phone inter- views (approximately 60 min) will be conducted at a con- venient time for identified teachers and audio-recorded for transcription and coding purposes. y Participants will complete secure web-based surveys at 13 timepoints spanning three phases: preparation (times 1–4), active implementation (times 5–11), and sustain- ment (times 12–13 follow-up during the next academic year) [45]. Time 1 will occur during informed consent, time 2 in the days immediately before BASIS-T/ACC pre-training, time 3 immediately after pre-training, time 4 immediately after post-training, time 5 two weeks after training, time 6 four weeks after training, times 7–11 monthly until the end of the academic year, time 12 at the beginning of the subsequent academic year, and time 13 during the spring of that academic year. Teach- ers will self-report demographics, perceptions about BASIS-T (when relevant) and PGD training, BASIS-T mechanisms, implementation outcomes, organizational moderators (implementation climate, leadership), and time and resources used for PGD implementation. From times 7–11, to reduce the respondent burden we will use a random item planned missingness design for measures of attitudes (1 item selected per subscale), action self- efficacy (2 items selected), and subjective norms (1 item selected per each scale). Teacher and school data collection Teacher data collection will span both active implemen- tation and sustainment phases (18 months per cohort). Data will include teacher quantitative web-based surveys and qualitative interviews, each of which will be incentiv- ized. To promote data integrity, key items will be forced to choose to prevent unplanned missingness and out-of- range responses. All data will be de-identified and stored securely. Detailed information about all study measures (including citations) can be found in Additional file 4. Page 8 of 14 Lyon et al. Implementation Science (2024) 19:2 Lyon et al. Implementation Science (2024) 19:2 Lyon et al. Implementation Science PGD fidelity T h l Teachers will complete self-reported PGD fidelity assess- ments, aligned with recommendations to gather reli- able and valid data, monthly in the time 4–13 surveys. Additionally, observations to assess the PGD fidelity of the implementing teachers will be conducted by trained school-based personnel using a standard PGD fidelity tool. Cost assessments Cost data collection will occur with all participating teachers to capture major costs of PGD delivery with and without BASIS-T, using activity-based costing to focus on key expenses (e.g., teacher and staff time, materials) [80, 81]. We will measure costs from the payor (i.e., school system) perspective, since the primary costs and associ- ated decision-making would be within the implement- ing school system in real-world implementation. These data will be collected along with other study measures depending on how often an item needs to be measured (e.g., one-time training prep in time 4 vs. PGD delivery collected monthly alongside the fidelity measures). Fol- lowing expert guidance to use mixed methods in imple- mentation cost studies [82], we will include open-ended items in each survey asking teachers to identify unex- pected resources they have needed for BASIS-T or PGD. ACC costs will be excluded from the cost estimate for the comparison condition; PGD implementation-as-usual is the ideal counterfactual for the cost-effectiveness of BASIS-T because it represents “business as usual” for PGD and, unlike BASIS-T, ACC would never be deliv- ered outside of a research project. To develop complete cost estimates for BASIS-T versus PGD implementation- as-usual, we will combine the teacher-reported data with information from other sources, such as BASIS-T train- ing records and meetings with school partners. Data collection regarding students will focus on class- room behavior, behavioral discipline, attendance, and academic performance and will be collected in aggre- gate at the classroom level, without individual identifiers. Teachers will complete the first secure online survey on student classroom behavior at time 4, as earlier data col- lection would prohibit teachers from having an adequate sample of student behavior. At the end of each school year, academic and behavioral data (attendance, discipli- nary actions, grades, and standardized test scores) will be requested for all students who were in classrooms of par- ticipating teachers; these data will be obtained in aggre- gate at the classroom level and no individual identifiers will be provided to the research team. Qualitative interviews Impact on PGD fidelity will be analyzed using mixed effects mod- els with sessions within teachers, testing for the main effects of condition on adherence and par- ticipant responsiveness ratings derived from both PGD observational and self-report fidelity data. We will test the effects of BASIS-T on PGD sustain- ment and delayed implementation using the mul- tilevel longitudinal analytic approach described for RQ1a. mixed effects model will examine whether the pro- portion of teachers in the BASIS-T condition who adopt PGD (i.e., initiate PGD) is higher than the proportion in the ACC condition. Second, Kaplan– Meier time-to-event analyses will be used to com- pare conditions on the number of days between training and PGD initiation. Reach will be ana- lyzed using mixed effects models comparing BASIS- T vs. ACC on the proportion of students in each classroom (out of those eligible based on whether their classroom teacher participated in PGD train- ing) who received PGD practices. Impact on PGD fidelity will be analyzed using mixed effects mod- els with sessions within teachers, testing for the main effects of condition on adherence and par- ticipant responsiveness ratings derived from both PGD observational and self-report fidelity data. We will test the effects of BASIS-T on PGD sustain- ment and delayed implementation using the mul- tilevel longitudinal analytic approach described for RQ1a. RQ1c: BASIS-T impact on student SEB and aca- demic outcomes will be tested via mixed effects mul- tilevel models as described in RQ1a, with academic data aggregated to the classroom level. Dependent variables will include post-intervention scores on teacher ratings of student behavior. g f RQ2a: Whether mechanisms of behavior change mediate the impact of BASIS-T on implementa- tion outcomes will be analyzed using path analysis [89] extending traditional mediation modeling to a multi-level framework for nested data [90, 91]. We will use thegeneralizer.org to compare our sample to national demographics to examine the generalizabil- ity of our findings [58]. If not generalizable, we will use inverse probability weighting to increase the represent- ativeness of estimates [87]. RQ2b: To test whether the effect of BASIS-T on implementation and student outcomes is moder- ated by teacher factors (e.g., demographics, baseline intentions to implement) and school-level factors, we will add moderators and interaction terms to the analytic approach described in RQs 1a-c. Qualitative interviews Teachers whose implementation behavior is insufficiently accounted for by our mediation model (e.g., teachers with favorable implementation outcomes, but who demon- strate low levels of TPB constructs and/or teachers with Page 9 of 14 Lyon et al. Implementation Science (2024) 19:2 Lyon et al. Implementation Science (2024) 19:2 Unless otherwise indicated, quantitative analyses will employ mixed-effects models using time points nested within classrooms within the school. Standard model-building procedures will be used [84, 85]. Class- room and time trends will be allowed to randomly vary. Piecewise time models will estimate slopes from T1 to T4 (training), T5 through T6 (first month of post-train- ing implementation), T7–T11 (first academic year), and T12–T13 (sustainment during the subsequent year). Variables for condition and condition x time will be added, and iterative models with possible control variables will be tested. Covariates not contributing at p < 0.10 based on likelihood ratio tests will be removed. Level 2 and 3 predictors will be fit and excluded for the same reasons. We will obtain estimates of whether there were statistically significant differences among the groups on the rate of change over time (i.e., slope) and whether there are statistically significant group dif- ferences in the average score on each outcome variable at the final timepoint of each piecewise segment. Mod- els will be generalized, with appropriate link functions (e.g., log-link, Poisson) applied based on distributional form (e.g., dichotomous, zero-inflation). Estimation will be fit using full maximum likelihood. Models will be assessed for possible violations of assumptions. Goodness-of-fit will be evaluated using likelihood ratios, deviance statistics, and fit criteria. The infer- ence will be evaluated relative to p < 0.05. For RQs with multiple DVs, we will adjust for the false discovery rate using Benjamini and Hochberg’s procedure [86]. mixed effects model will examine whether the pro- portion of teachers in the BASIS-T condition who adopt PGD (i.e., initiate PGD) is higher than the proportion in the ACC condition. Second, Kaplan– Meier time-to-event analyses will be used to com- pare conditions on the number of days between training and PGD initiation. Reach will be ana- lyzed using mixed effects models comparing BASIS- T vs. ACC on the proportion of students in each classroom (out of those eligible based on whether their classroom teacher participated in PGD train- ing) who received PGD practices. Qualitative interviews RQ 1a: A series of 3-level piecewise mixed effects models will be used to test our primary hypoth- eses: (1) the BASIS-T condition will show steeper gains than the ACC condition from T1 to T2 (e.g., more favorable attitudes towards EBPPs, increased social norms; enhanced task and maintenance self-efficacy; stronger intentions to implement) and (2) both groups will decline after training (T3–T5) as has been found in past TPB research [88], and in the BASIS-T pilot trial [33], but the BASIS-T condition will have longer sustained between- condition effects after training from T3 to T13. Dependent variables will include subscale scores on attitudes, social norms, self-efficacy, intentions, and maintenance self-efficacy. y RQ2c: To explore what explains “residual” teachers whose implementation behaviors are not accounted for by the mediation model, we will analyze quali- tative interviews with teachers who have a differ- ence between predicted (based on BASIS-T putative mechanisms) and actual implementation behav- ior of ≥ 1 SD. Data will be coded using an inte- grated directed and conventional content analysis [92] approach as certain codes will be conceptual- ized during the interview guide development and driven by the exploration, preparation, implemen- tation, sustainment [93] framework (i.e., deductive approach) which will allow for the examination of influences on implementation across multiple levels and phases. Other codes will be developed through a fi RQ 1b: BASIS-T impact on behavioral implemen- tation outcomes will be tested in two ways. First, a Page 10 of 14 Page 10 of 14 Lyon et al. Implementation Science (2024) 19:2 Lyon et al. Implementation Science Lyon et al. Implementation Science (2024) 19:2 teachers per school (after 15% attrition), school ICCs of 0.10 (consistent with our pilot data), and 13 timepoints. With the same assumptions as above, plus an assumption that 50% of the schools will be in a given moderator sub- group, our MDES for detecting any single school-level moderator variable is 0.48 and teacher-level moderator variable is 0.45. For mediator analyses, PowerUp! iden- tifies power using the Sobel test, joint test, and Monte Carlo simulations. Across all tests, we will have a power of greater than 0.80 to detect reasonable and likely effects, based on our pilot trial and standard interpreta- tions of effect sizes. Qualitative interviews We will detect an MDES equivalent to Cohen’s d = 0.60 for the treatment-mediator pathways (e.g., BASIS-T to implementation intention) which is lower than our actual obtained effect size from the pilot trial (which ranged from d = 0.61 to 1.16 for our primary mediators), an MDES of Pearson’s r = 0.3 for the media- tor-outcome path (e.g., implementation intention to stu- dent behavior, a small effect), and an MDES of d = 0.10 for the direct path from treatment to student outcome. close reading of an initial subset of transcripts (i.e., inductive approach). These themes will provide a way of identifying and understanding the most sali- ent factors that impact implementation and extend beyond the existing BASIS-T mechanisms and the- ory of change. After a stable set of codes is developed, a consensus process will be used in which all review- ers independently code and compare their coding to arrive at consensus judgments through open dia- logue [78, 79, 94]. RQ2d: We will process cost data by assigning mon- etary values to each cost. We will use CostOut [95], a web-based program for conducting cost-effectiveness analysis in education, to identify nationally repre- sentative unit prices for ingredients. For qualitative data, we will rapidly analyze responses on an ongoing basis [96, 97] and incorporate newly identified costs into future surveys for quantitative measurement. Once cost data collection is complete, we will calculate the costs of BASIS-T versus PGD implementation-as- usual based on the unit price and amount of each cost category (e.g., hours spent, items purchased). We will use CostOut to standardize dollar values, including an index year for inflation; cost-of-living adjustments; and dis- counting costs from different years to account for prefer- encing of delayed over immediate costs [80, 95]. We will generate descriptive statistics describing typical costs (i.e., means, standard deviations) for each condition and incre- mental costs of BASIS-T over implementation as usual. Discussion Innovationh This hybrid type 3 trial will contribute to the literature on pragmatic implementation strategies, as well as nascent but expanding literature on implementation mechanisms [101–103]. Aside from BASIS-T’s counterpart strategy, BASIS-C [104], no studies of strategies have been explic- itly designed to impact TPB and HAPA mechanisms while testing those mechanisms via mediation models. Recent systematic reviews [103] indicate that much more work is necessary surrounding implementation strategy mechanisms to allow for the development of streamlined, pragmatic approaches to improving implementation out- comes that can be generalized across EBPPs. The current study also contributes significantly to theory-building in implementation by exploring factors beyond TPB and HAPA that help to explain hypothesis-defying residuals’ relationships between mechanisms and behavior. Once the cost analysis is complete, we will use CostOut to calculate the cost-effectiveness [81] of BASIS-T ver- sus PGD implementation as usual. This will involve cal- culating a series of incremental cost-effectiveness ratios for each implementation and student outcome measure, representing the incremental costs of BASIS-T divided by its incremental benefit (i.e., effect size). Themes about BASIS-T mechanisms and outcomes from the qualita- tive teacher interviews will allow for mixed-method cost-effectiveness evaluation [82] in which participants’ views help determine whether the results were worth the cost. For both cost and cost-effectiveness analyses, we will conduct sensitivity analyses that vary key sources of uncertainty in the models to examine the robustness of our estimates [98, 99]. In addition, this project will examine the costs and cost- effectiveness of the BASIS-T strategy. Cost-effectiveness is an important driver of adoption decisions, especially at system and policy levels [105]. Examination of cost- effectiveness is particularly critical for implementation strategies designed with pragmatism in mind. Efficient delivery and impact are key components of pragmatism, as are clear links between prioritized implementation determinants (e.g., self-efficacy in BASIS-T) and strategy components [106]. Availability of data and materials Availability of data and materials Please contact the lead author for more information. Availability of data and materials Please contact the lead author for more information. Limitationsh the study recruitment and data collection. CRC, ML, ARL, MLH, and AL have worked to refine and update the BASIS-T implementation strategy in prepara‑ tion for deployment. All authors contributed to the development, drafting, or review of the manuscript. All authors approved the final manuscript. the study recruitment and data collection. CRC, ML, ARL, MLH, and AL have worked to refine and update the BASIS-T implementation strategy in prepara‑ tion for deployment. All authors contributed to the development, drafting, or review of the manuscript. All authors approved the final manuscript. This study uses a block randomization approach in which all participating teachers in each school are randomized to either BASIS-T or the ACC condition. Although our team considered randomizing at the indi- vidual teacher level to align with the individual focus of the BASIS-T strategy, we opted against this because it presented a significant risk of contamination because of the extent to which some of BASIS-T’s putative mecha- nisms are likely to be socially influenced (especially subjective social norms). Power Our planned sample, including attrition, will provide suf- ficient power to test linear effects on teacher- and class- room school-level variables with a minimum detectable effect size (MDES) of small to medium effects, d = 0.37. We used PowerUp! [100] assuming clustering, final sam- ples of 46 schools evenly randomized to condition, 6 Page 11 of 14 Page 11 of 14 Lyon et al. Implementation Science (2024) 19:2 Lyon et al. Implementation Science (2024) 19:2 Lyon et al. Implementation Science Declarations The current study will provide evidence of the efficacy and cost-effectiveness of applying BASIS-T in an educa- tional setting alongside EBP training to improve imple- mentation outcomes. Trial results will be disseminated via publications, presentations via traditional (e.g., press releases) and social media, and through networks of prac- titioners. Positive findings from this trial would support the generalizability of BASIS-T to additional universal, school-based EBPPs for social, emotional, and behavio- ral health. More generally, if effective, it will add to the growing evidence for pragmatic implementation strate- gies and the mechanisms through which they operate. Consent for publication Not applicable. Consent for publication Not applicable. Received: 27 October 2023 Accepted: 12 December 2023 Received: 27 October 2023 Accepted: 12 December 2023 HAPA Health action process approach MI Motivational interviewing MDES Minimal detectable effect size PGD Positive greetings at the door SEB Social, emotional, and behavioral health TAU Treatment-as-usual TPB Theory of planned behavior Supplementary Information The online version contains supplementary material available at https://doi. org/10.1186/s13012-023-01330-y. Additional file 1. SPIRIT 2013 Checklist. Additional file 2. IRB Determination. Additional file 3. Consent Form for Teachers. Additional file 4. Detailed Study Measures [107–115]. Acknowledgements None. Authors’ contributions ARL, CRC, ML, AD, and MDP developed the overarching scientific aims and design of the project. CH, PR, MB, and RG assisted in the development and operationalization of the study methods and worked with the investigators to obtain institutional review board approval. AG, NM, and AWH are supporting Competing interests The authors declare that they have no competing interests. References 1. Chang X, Jiang X, Mkandarwire T, Shen M. Associations between adverse childhood experiences and health outcomes in adults aged 18–59 years. PLos One. 2019;14(2):e0211850. 2. Colizzi M, Lasalvia A, Ruggeri M. Prevention and early intervention in youth mental health: is it time for a multidisciplinary and trans- diagnostic model for care? Int J Ment Heal Syst. 2020;14(1):23. 3. Durlak JA, Wells AM. Primary prevention mental health programs for children and adolescents: a meta-analytic review. Am J Community Psychol. 1997;25(2):115–52. Supplementary Information The online version contains supplementary material available at https://doi. org/10.1186/s13012-023-01330-y. Additional file 1. SPIRIT 2013 Checklist. Additional file 2. IRB Determination. Additional file 3. Consent Form for Teachers. Additional file 4. Detailed Study Measures [107–115]. 4. Mendelson T, Eaton WW. Recent advances in the preven‑ tion of mental disorders. Soc Psychiatry Psychiatr Epidemiol. 2018;53(4):325–39. 5. Tennant R, Goens C, Barlow J, Day C, Stewart-Brown S. A systematic review of reviews of interventions to promote mental health and prevent mental health problems in children and young people. J Public Ment Health. 2007;6(1):25–32. 6. Webster-Stratton C, Rinaldi J, Jamila MR. Long-term outcomes of incredible years parenting program: predictors of adolescent adjust‑ ment. Child Adolesc Ment Health. 2011;16(1):38–46. 7. Meherali S, Punjani N, Louie-Poon S, Abdul Rahim K, Das JK, Salam RA, et al. Mental health of children and adolescents amidst COVID-19 and past pandemics: a rapid systematic review. Int J Environ Res Public Health. 2021;18(7):3432. Ethics approval and consent to participate This project was approved by the University of Washington Institutional Review Board (IRB). Author details 1 University of Washington, 6200 NE 74Th Street, Suite 100, Seattle, WA 98115, USA. 2 Character Strong, 4227 S Meridian, Puyallup, WA 98373, USA. 3 University of Kansas, 1122 W Campus Rd, Lawrence, KS 66045, USA. 4 RAND Corporation, 1776 Main Street, Santa Monica, CA 91604, USA. 5 University of Minnesota, 1954 Buford Avenue, Suite 425, St. Paul, MN 55108, USA. Abbreviations ACC Active control condition BASIS-T Beliefs and attitudes for successful implementation in schools EBPP Evidence-based prevention programs HAPA Health action process approach MI Motivational interviewing MDES Minimal detectable effect size PGD Positive greetings at the door SEB Social, emotional, and behavioral health TAU Treatment-as-usual TPB Theory of planned behavior Supplementary Information The online version contains supplementary material available at https://doi. org/10.1186/s13012-023-01330-y. Additional file 1. SPIRIT 2013 Checklist. Additional file 2. IRB Determination. Additional file 3. Consent Form for Teachers. Additional file 4. Detailed Study Measures [107–115]. Acknowledgements None. Authors’ contributions ARL, CRC, ML, AD, and MDP developed the overarching scientific aims and design of the project. CH, PR, MB, and RG assisted in the development and Abbreviations ACC Active control condition BASIS-T Beliefs and attitudes for successful implementation in schools EBPP Evidence-based prevention programs HAPA Health action process approach MI Motivational interviewing MDES Minimal detectable effect size PGD Positive greetings at the door SEB Social, emotional, and behavioral health TAU Treatment-as-usual TPB Theory of planned behavior Funding g This publication was supported by grant R305A220481, awarded by the Institute of Education Sciences (IES;https://ies.ed.gov/). The content is solely the responsibility of the authors and does not necessarily represent the official views of the IES. IES had no role in study design; data collection, management, analysis, or interpretation; writing of the manuscript; or the decision to submit the manuscript for publication. Availability of data and materials Authors’ contributions Merle JL, Cook CR, Pullmann MD, Larson MF, Hamlin CM, Hugh ML, et al. Longitudinal effects of a motivationally focused strategy to increase the yield of training and consultation on teachers’ adoption and fidelity of a universal program. Sch Ment Heal. 2023;15(1):105–22. y 13. An act to reauthorize the Elementary and Secondary Education Act of 1965 to ensure that every child achieves. U.S. Government Publishing Office; 2015. Available from: https://www.govinfo.gov/app/details/ PLAW-114publ95. Cited 2023 Apr 11. 34. Locke J, Lawson GM, Beidas RS, Aarons GA, Xie M, Lyon AR, et al. Individual and organizational factors that affect implementation of evidence-based practices for children with autism in public schools: a cross-sectional observational study. Implement Sci. 2019;14(1):29. 14. Burns MK, Jimerson SR, Van Der Heyden AM, Deno SL. Toward a unified response-to-intervention model: multi-tiered systems of support. In: Handbook of response to intervention: The science and practice of multi-tiered systems of support. 2nd ed. New York, NY, US: Springer Science + Business Media; 2016. p. 719–32. 35. Kincaid D, Childs K, Blase KA, Wallace F. Identifying barriers and facilita‑ tors in implementing schoolwide positive behavior support. J Posit Behav Interv. 2007;9:174–84. 36. Sanford Derousie RM, Bierman KL. Examining the sustainability of an evidence-based preschool curriculum: the REDI program. Early Child Res Q. 2012;27(1):55–6. 15. Greenberg MT, Abenavoli R. Universal interventions: fully exploring their impacts and potential to produce population-level impacts. J Res Educ Effect. 2017;10(1):40–67. f 16. Fazel M, Hoagwood K, Stephan S, Ford T. Mental health interventions in schools 1: Mental health interventions in schools in high-income countries. Lancet Psychiatry. 2014;1(5):377–87. 37. Beidas RS, Kendall PC. Training therapists in evidence-based practice: a critical review of studies from a systems-contextual perspective. Clin Psychol (New York). 2010;17(1):1–30. 17. Horner RH, Sugai G, Smolkowski K, Eber L, Nakasato J, Todd AW, et al. A randomized, wait-list controlled effectiveness trial assessing school- wide positive behavior support in elementary schools. J Posit Behav Interv. 2009;11:133–44. 38. McHugh RK, Barlow DH. The dissemination and implementation of evidence-based psychological treatments. A review of current efforts. Am Psychol. 2010;65(2):73–84. 39. Ajzen I. From intentions to actions: a theory of planned behavior. In: Kuhl J, Beckmann J, editors. Action Control: From Cognition to Behavior. Berlin, Heidelberg: Springer; 1985. p. 11–39. (SSSP Springer Series in Social Psychology). Available from: https://doi.org/10.1007/978-3-642- 69746-3_2. Cited 2023 Apr 11.fi 18. Durlak JA, Weissberg RP, Dymnicki AB, Taylor RD, Schellinger KB. Authors’ contributions Authors’ contributions ARL, CRC, ML, AD, and MDP developed the overarching scientific aims and design of the project. CH, PR, MB, and RG assisted in the development and operationalization of the study methods and worked with the investigators to obtain institutional review board approval. AG, NM, and AWH are supporting 8. Leeb RT, Bitsko RH, Radhakrishnan L, Martinez P, Njai R, Holland KM. Mental health-related emergency department visits among children aged <18 years during the COVID-19 pandemic - United Page 12 of 14 Lyon et al. Implementation Science (2024) 19:2 Lyon et al. Implementation Science (2024) 19:2 into community settings. J Autism Dev Disord. 2013;43(12):2970. https://doi.org/10.1007/s10803-013-1826-7. into community settings. J Autism Dev Disord. 2013;43(12):2970. https://doi.org/10.1007/s10803-013-1826-7. States, January 1-October 17, 2020. MMWR Morb Mortal Wkly Rep. 2020;69(45):1675–80. 28. Durlak JA, DuPre EP. Implementation matters: a review of research on the influence of implementation on program outcomes and the factors affect‑ ing implementation. Am J Community Psychol. 2008;41(3–4):327–50. 9. Burns BJ, Costello EJ, Angold A, Tweed D, Stangl D, Farmer EM, et al. Children’s mental health service use across service sectors. Health Aff (Millwood). 1995;14(3):147–59. 10. Duong MT, Bruns EJ, Lee K, Cox S, Coifman J, Mayworm A, et al. Rates of mental health service utilization by children and adolescents in schools and other common service settings: a systematic review and meta- analysis. Adm Policy Ment Health. 2021;48(3):420–39. 29. Glisson C. The organizational context of children’s mental health ser‑ vices. Clin Child Fam Psychol Rev. 2002;5(4):233–53. 30. Cook CR, Lyon AR, Kubergovic D, Wright DB, Zhang Y. A supportive beliefs intervention to facilitate the implementation of evidence-based practices within a multi-tiered system of supports. School Mental Health: A Multidiscipl Res Pract J. 2015;7:49–60. 11. Merikangas KR, He JP, Burstein M, Swendsen J, Avenevoli S, Case B, et al. Service utilization for lifetime mental disorders in U.S. adolescents: results of the National Comorbidity Survey-Adolescent Supplement (NCS-A). J Am Acad Child Adolesc Psychiatry. 2011;50(1):32–45. 31. Nilsen P. Making sense of implementation theories, models and frame‑ works. Implement Sci. 2015;10(1):53. 12. Final Report of the Federal Commission on School Safety \| National Center on Safe Supportive Learning Environments (NCSSLE). Federal Commission on School Safety; 2018. Available from: https://safesuppor tivelearning.ed.gov/resources/final-report-federal-commission-school- safety. Cited 2023 Apr 11. 32. Locke J, Kang-Yi C, Frederick L, Mandell DS. Individual and organiza‑ tional characteristics predicting intervention use for children with autism in schools. Autism. 2020;24(5):1152–63. 33. Authors’ contributions The impact of enhancing students’ social and emotional learning: a meta-analysis of school-based universal interventions. Child Dev. 2011;82(1):405–32. 40. Schwarzer R. Self-efficacy in the adoption and maintenance of health behaviors: theoretical approaches and a new model. In: Self-efficacy: Thought control of action. Washington, DC, US: Hemisphere Publishing Corp; 1992. p. 217–43. 19. García-Moya I. The importance of student-teacher relationships for wellbeing in schools. the importance of connectedness in student- teacher relationships: insights from the teacher connectedness project. 2020:1–25. 20. Borgmeier C, Loman SL, Hara M. Teacher self-assessment of evidence- based classroom practices: preliminary findings across primary, inter‑ mediate and secondary level teachers. Teach Dev. 2016;20(1):40–56. 41. Ajzen I. The theory of planned behavior. Organ Behav Hum Decis Process. 1991;50(2):179–211. 42. Ajzen I, Manstead ASR. Changing health-related behaviours: an approach based on the theory of planned behaviour. In: The scope of social psychology: Theory and applications. New York, NY, US: Psychol‑ ogy Press; 2007. p. 43–63. 21. Evans SW, Weist MD. Implementing empirically supported treat‑ ments in the schools: what are we asking? Clin Child Fam Psychol Rev. 2004;7(4):263–7. 22. Lyon AR, Bruns EJ. From evidence to impact: Joining our best school mental health practices with our best implementation strategies. School Ment Health. 2019;11(1):106–14. 43. Eccles MP, Hrisos S, Francis J, Kaner EF, Dickinson HO, Beyer F, et al. Do self- reported intentions predict clinicians’ behaviour: a systematic review. Implement Sci. 2006;1:28. 23. Wilson DB, Gottfredson DC, Najaka SS. School-based preven‑ tion of problem behaviors: a meta-analysis. J Quant Criminol. 2001;17(3):247–72. 44. Steinmetz H, Knappstein M, Ajzen I, Schmidt P, Kabst R. How effective are behavior change interventions based on the theory of planned behavior? A three-level meta-analysis. Zeitschrift für Psychologie. 2016;224:216–33. 24. Vernez G, Karam RT, Mariano LT, DeMartini C. Evaluating comprehensive school reform models at scale: focus on implementation. RAND Corpo‑ ration; 2006. Available from: https://www.rand.org/pubs/monographs/ MG546.html. Cited 2023 Apr 11. 45. Moullin JC, Dickson KS, Stadnick NA, Rabin B, Aarons GA. Systematic review of the exploration, preparation, implementation, sustainment (EPIS) framework. Implement Sci. 2019;14(1):1. 25. Locke J, Olsen A, Wideman R, Downey MM, Kretzmann M, Kasari C, et al. A tangled web: the challenges of implementing an evidence-based social engagement intervention for children with autism in urban public school settings. Behav Ther. 2015;46(1):54–67. 46. Duong MT, Cook CR, Lee K, Davis CJ, Vázquez-Colón CA, Lyon AR. Authors’ contributions User testing to drive the iterative development of a strategy to improve implementation of evidence-based practices in school mental health. Evid-Based Pract Child Adolesc Mental Health. 2020;5(4):414–25. 26. Stahmer AC, Reed S, Lee E, Reisinger EM, Connell JE, Mandell DS. Train‑ ing teachers to use evidence-based practices for autism: examining procedural implementation fidelity. Psychol Sch. 2015;52(2):181–95. 47. Larson M, Cook CR, Brewer SK, Pullmann MD, Hamlin C, Merle JL, et al. Examining the effects of a brief, group-based motivational implemen‑ tation strategy on mechanisms of teacher behavior change. Prev Sci. 2021;22(6):722–36. 27. Suhrheinrich J, Stahmer AC, Reed S, Schreibman L, Reisinger E, Mandell D. Implementation challenges in translating pivotal response training Page 13 of 14 Lyon et al. Implementation Science (2024) 19:2 Lyon et al. Implementation Science 48. Flower A, McKenna JW, Bunuan RL, Muething CS, Vega R Jr. Effects of the good behavior game on challenging behaviors in school settings. Rev Educ Res. 2014;84:546–71. motivation to adopt and implement school-based interventions. Adv School Ment Health Promot. 2013;6:158–73. 70. Sanetti LMH, Kratochwill TR, Long ACJ. Applying adult behavior change theory to support mediator-based intervention implementation. Sch Psychol Q. 2013;28(1):47–62. 49. Joslyn PR, Donaldson JM, Austin JL, Vollmer TR. The good behavior game: a brief review. J Appl Behav Anal. 2019;52(3):811–5. 71. Adriaanse MA, Vinkers CDW, De Ridder DTD, Hox JJ, De Wit JBF. Do implementation intentions help to eat a healthy diet? A system‑ atic review and meta-analysis of the empirical evidence. Appetite. 2011;56(1):183–93. 50. Barrish HH, Saunders M, Wolf MM. Good behavior game: effects of indi‑ vidual contingencies for group consequences on disruptive behavior in a classroom. J Appl Behav Anal. 1969;2(2):119–24. 51. Allday RA, Pakurar K. Effects of teacher greetings on student on-task behavior. J Appl Behav Anal. 2007;40(2):317–20. 72. Bélanger-Gravel A, Godin G, Amireault S. A meta-analytic review of the effect of implementation intentions on physical activity. Health Psychol Rev. 2013;7(1):23–54. 52. Allday RA, Hinkson-Lee K, Hudson T, Neilsen-Gatti S, Kleinke A, Russel C. Training general educators to increase behavior-specific praise: effects on students with EBD. Behavioral Disorders. 2012; Available from: https://ir.stthomas.edu/celc_ed_speced_pub/13. 73. Sheeran P, Orbell S. Implementation intentions and repeated behaviour: augmenting the predictive validity of the theory of planned behaviour. Eur J Soc Psychol. 1999;29(2–3):349–69. 53. Cook CR, Davis C, Brown EC, Locke J, Ehrhart MG, Aarons GA, et al. Confirmatory factor analysis of the evidence-based practice attitudes scale with school-based behavioral health consultants. Authors’ contributions Implement Sci. 2018;13(1):116. 74. Orbell S, Sheeran P. Motivational and volitional processes in action initiation: a field study of the role of implementation intentions. J Appl Soc Psychol. 2000;30:780–97. y 75. Orbell S, Hodgkins S, Sheeran P. Implementation intentions and the theory of planned behavior. Pers Soc Psychol Bull. 1997;23(9):945–54. 54. Ialongo N, Poduska J, Werthamer L, Kellam S. The distal impact of two first-grade preventive interventions on conduct problems and disorder in early adolescence. J Emot Behav Disord. 2001;9:146–60. 76. Sanetti LMH, Collier-Meek MA, Long ACJ, Kim J, Kratochwill TR. Using implementation planning to increase teachers’ adherence and quality to behavior support plans. Psychol Sch. 2014;51:879–95. 55. Antisocial behavior in children and adolescents: a developmental analysis and model for intervention. Washington, DC, US: American Psychological Association; 2002. xi, 337 p. (Reid JB, Patterson GR, Snyder J, editors. Antisocial behavior in children and adolescents: A develop‑ mental analysis and model for intervention). 77. Hagger MS, Luszczynska A. Implementation intention and action plan‑ ning interventions in health contexts: state of the research and propos‑ als for the way forward. Appl Psychol Health Well Being. 2014;6(1):1–47. 78. DeSantis L, Ugarriza DN. The concept of theme as used in qualitative nursing research. West J Nurs Res. 2000;22(3):351–72. 56. Datu JAD, King RB. Subjective well-being is reciprocally associated with academic engagement: a two-wave longitudinal study. J Sch Psychol. 2018;69:100–10. 79. Hill CE, Thompson BJ, Williams EN. A guide to conducting consensual qualitative research. Couns Psychol. 1997;25(4):517–72. 57. Smith ND, Suldo S, Hearon B, Ferron J. An application of the dual-factor model of mental health in elementary school children: examining academic engagement and social outcomes. J Positive School Psychol. 2020;4(1):49–68. 80. Hollands FM, Pratt-Williams J, Shand R. Cost analysis standards & guidelines 1.1. Cost Analysis in Practice (CAP) Project. Revised. Grantee Submission. 2021. Available from: https://eric.ed.gov/?id=ED624010. Cited 2023 Apr 18. 58. Tipton E, Miller K. The generalizer: a webtool for improving the general‑ izability of results from experiments. Available from: https://www.thege neralizer.org/. Cited 2023 Apr 11. 81. Levin HM, McEwan PJ, Belfield C, Bowden AB, Shand R. Economic evalu‑ ation in education: cost-effectiveness and benefit-cost analysis. SAGE Publications; 2017. 212 p. 59. Lyon AR, Whitaker K, Locke J, Cook CR, King KM, Duong M, et al. The impact of inter-organizational alignment (IOA) on implementation outcomes: evaluating unique and shared organizational influences in education sector mental health. Implement Sci. 2018;13(1):24. 82. Authors’ contributions Dopp AR, Mundey P, Beasley LO, Silovsky JF, Eisenberg D. Mixed-method approaches to strengthen economic evaluations in implementation research. Implement Sci. 2019;14(1):2. 60. Dong N, Lipsey M. The statistical power of the cluster randomized block design with matched pairs--a simulation study. Society for Research on Educational Effectiveness. Society for Research on Educational Effec‑ tiveness; 2010. Available from: https://eric.ed.gov/?id=ED512728. Cited 2023 Apr 18. 83. What Works Clearinghouse. What works clearinghouse standards handbook, Version 4.1. U.S. Department of Education, Institute of Edu‑ cation Sciences, National Center for Education Evaluation and Regional Assistanc; 2020. Available from: https://ies.ed.gov/ncee/wwc/handb ooks. Cited 2023 Apr 11. 61. Rathvon N. Effective school interventions: evidence-based strategies for improving student outcomes. 2nd ed. New York, NY, US: Guilford Press; 2008. p. xvii–460 (Effective school interventions: Evidence-based strate‑ gies for improving student outcomes, 2nd ed). 84. Raudenbush SW, Bryk AS. Hierarchical linear models. Second. SAGE Publications; 2023. (Advanced Quantitative Techniques in the Social Sci‑ ences; vol. 1). Available from: https://us.sagepub.com/en-us/nam/hiera rchical-linear-models/book9230. Cited 2023 Apr 11. 62. Chassin L, Presson CC, Sherman SJ, Corty E, Olshavsky RW. Predicting the onset of cigarette smoking in adolescents: a longitudinal study. J Appl Soc Psychol. 1984;14(3):224–43. 85. Singer JD, Willett JB. Survival analysis. In: Handbook of psychology: research methods in psychology, Vol 2. Hoboken, NJ, US: John Wiley & Sons, Inc.; 2003. p. 555–80. y 63. Perkins HW. The contextual effect of secular norms on religiosity as moderator of student alcohol and other drug use. Res Soc Scient Study Religion. 1994;6:187–208. 86. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J Roy Stat Soc: Ser B (Meth‑ odol). 1995;57(1):289–300. 64. Perkins HW, Meilman PW, Leichliter JS, Cashin JR, Presley CA. Mispercep‑ tions of the norms for the frequency of alcohol and other drug use on college campuses. J Am Coll Health. 1999;47(6):253–8. 87. Seaman SR, White IR. Review of inverse probability weighting for deal‑ ing with missing data. Stat Methods Med Res. 2013;22(3):278–95. 88. Helfrich CD, Kohn MJ, Stapleton A, Allen CL, Hammerback KE, Chan KCG, et al. Readiness to change over time: change commitment and change efficacy in a workplace health-promotion trial. Front Public Health. 2018;6:110. 65. Petrova PK, Cialdini RB, Sills SJ. Consistency-based compliance across cultures. J Exp Soc Psychol. 2007;43:104–11. 66. Pratkanis AR. An invitation to social influence research. In: The science of social influence: advances and future progress. Authors’ contributions New York, NY, US: Psychology Press; 2007. p. 1–15 (Frontiers of social psychology). 89. Loehlin JC, Beaujean AA. Latent variable models: an introduction to fac‑ tor, path, and structural equation analysis, fifth edition. Taylor & Francis; 2016. 391 p. 67. Miller WR, Rollnick S. Motivational interviewing: helping people change. Guilford Press; 2012. 497 p. 90. Kenny DA, Korchmaros JD, Bolger N. Lower level mediation in multilevel models. Psychol Methods. 2003;8(2):115–28. 68. Reinke WM, Herman KC, Sprick R. Motivational interviewing for effective classroom management: the classroom check-up. Guilford Press: Practi‑ cal Intervention in the Schools Series. Guilford Publications; 2011. 91. Bauer DJ, Preacher KJ, Gil KM. Conceptualizing and testing random indi‑ rect effects and moderated mediation in multilevel models: new proce‑ dures and recommendations. Psychol Methods. 2006;11(2):142–63. 69. Frey AJ, Lee J, Small JW, Seeley JR, Walker HM, Feil EG. The motivational interviewing navigation guide: a process for enhancing teachers’ Page 14 of 14 Lyon et al. Implementation Science (2024) 19:2 Lyon et al. Implementation Science (2024) 19:2 92. Hsieh HF, Shannon SE. Three approaches to qualitative content analysis. Qual Health Res. 2005;15(9):1277–88. sector: confirmatory factor analysis of measures of implementation leadership, climate, and citizenship. Implement Sci. 2018;13:1–4. sector: confirmatory factor analysis of measures of implementation leadership, climate, and citizenship. Implement Sci. 2018;13:1–4. 93. Aarons GA, Hurlburt M, Horwitz SM. Advancing a conceptual model of evidence-based practice implementation in public service sectors. Adm Policy Ment Health. 2011;38(1):4–23. 113. Lyon AR, Corbin CM, Brown EC, Ehrhart MG, Locke J, Davis C, Picozzi E, Aarons GA, Cook CR. Leading the charge in the education sector: development and validation of the School Implementation Leadership Scale (SILS). Implement Sci. 2022;17(1):48. y 94. Hill CE, Knox S, Thompson BJ, Williams EN, Hess SA, Ladany N. Consen‑ sual qualitative research: an update. J Couns Psychol. 2005;52:196–205. 114. Thayer AJ, Cook CR, Davis C, Brown EC, Locke J, Ehrhart MG, Aarons GA, Picozzi E, Lyon AR. Construct validity of the school-implementation cli‑ mate scale. Implement Res Pract. 2022;3:26334895221116064. https:// doi.org/10.1177/26334895221116065. 95. Hollands FM, Hanisch-Cerda B, Levin HM, Belfield CR, Menon A, Shand R, Pan Y, Bakir I, Cheng H. CostOut-the CBCSE cost tool kit. Center for Benefit-Cost Studies of Education, Teachers College: Columbia Univer‑ sity. www. cbcsecosttoolkit. org; 2015. 115. Kilgus SP, Chafouleas SM, Riley-Tillman TC, Welsh ME. Direct behavior rating scales as screeners: a preliminary investigation of diagnostic accuracy in elementary school. Sch Psychol Q. 2012;27(1):41. 96. Authors’ contributions Hamilton AB, Finley EP. Qualitative methods in implementation research: an introduction. Psychiatry Res. 2019;280:112516. 97. Taylor B, Henshall C, Kenyon S, Litchfield I, Greenfield S. Can rapid approaches to qualitative analysis deliver timely, valid findings to clini‑ cal leaders? A mixed methods study comparing rapid and thematic analysis. BMJ Open. 2018;8(10):e019993. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in pub‑ lished maps and institutional affiliations. Springer Nature remains neutral with regard to jurisdictional claims in pub‑ lished maps and institutional affiliations. 98. Briggs AH, Gray AM. Handling uncertainty in economic evaluations of healthcare interventions. BMJ. 1999;319(7210):635–8. 99. Levin HM, Belfield CR. Chapter 15: cost-effectiveness and educational efficiency. In: Handbook of Contemporary Education Economics. 2nd ed. Edward Elgar Publishing; 2017. p. 338–56. Available from: https:// www-elgaronline-com.ezp1.lib.umn.edu/display/edcoll/9781785369 063/9781785369063.00020.xml. Cited 2023 Apr 18. p 100. Dong N, Maynard R. PowerUp!: a tool for calculating minimum detectable effect sizes and minimum required sample sizes for experimental and quasi-experimental design studies. J Res Educ Effect. 2013;6(1):24–67. 101. Lewis CC, Klasnja P, Powell BJ, Lyon AR, Tuzzio L, Jones S, et al. From classification to causality: advancing understanding of mechanisms of change in implementation science. Front Public Health. 2018;6:136. 102. Lewis CC, Klasnja P, Lyon AR, Powell BJ, Lengnick-Hall R, Buchanan G, et al. The mechanics of implementation strategies and measures: advancing the study of implementation mechanisms. Implement Sci Commun. 2022;3(1):114. 103. Lewis CC, Boyd MR, Walsh-Bailey C, Lyon AR, Beidas R, Mittman B, et al. A systematic review of empirical studies examining mechanisms of implementation in health. Implement Sci. 2020;15(1):21. 104. Lyon AR, Pullmann MD, Dorsey S, Levin C, Gaias LM, Brewer SK, et al. Protocol for a hybrid type 2 cluster randomized trial of trauma-focused cognitive behavioral therapy and a pragmatic individual-level imple‑ mentation strategy. Implement Sci. 2021;16(1):3. 105. Walker SC, Lyon AR, Aos S, Trupin EW. The consistencies and vagaries of the Washington State inventory of evidence-based practice: the defini‑ tion of “evidence-based” in a policy context. Admin Pol Mental Health Mental Health Serv Res. 2017;44:42–54. 106. Powell BJ, Fernandez ME, Williams NJ, Aarons GA, Beidas RS, Lewis CC, et al. Enhancing the impact of implementation strategies in healthcare: a research agenda. Front Public Health. 2019;7:3. 107. Merle JL, Cook CR, Locke JJ, Ehrhart MG, Brown EC, Davis CJ, Lyon AR. Teacher attitudes toward evidence-based practices: exploratory and confirmatory analyses of the school-adapted evidence-based practice attitude scale. Implement Res Pract. 2023;4:26334895221151024. https://doi.org/10.1177/26334895221151026. p g 108. Francis J, Eccles MP, Johnston M, et al. Constructing questionnaires based on the theory of planned behaviour: a manual for health services researchers. Published online 2004. http://openaccess.city.ac.uk/1735/. • fast, convenient online submission • thorough peer review by experienced researchers in your ﬁeld • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year • At BMC, research is always in progress. Learn more biomedcentral.com/submissions Ready to submit your research Ready to submit your research ? Choose BMC and benefit from: ? Choose BMC and benefit from: Publisher’s Note • fast, convenient online submission • thorough peer review by experienced researchers in your ﬁeld • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year • At BMC, research is always in progress. Learn more biomedcentral.com/submissions Ready to submit your research Ready to submit your research ? Choose BMC and benefit from: ? Choose BMC and benefit from: • fast, convenient online submission • thorough peer review by experienced researchers in your ﬁeld • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year • At BMC, research is always in progress. Learn more biomedcentral.com/submissions Ready to submit your research Ready to submit your research ? Choose BMC and benefit from: ? Choose BMC and benefit from: 109. Ramsay CR, Thomas RE, Croal BL, Grimshaw JM, Eccles MP. Using the theory of planned behaviour as a process evaluation tool in ran‑ domised trials of knowledge translation strategies: a case study from UK primary care. Implement Sci. 2010;5(1):71. https://doi.org/10.1186/ 1748-5908-5-71. 110. Schmitz GS, Schwarzer R. Selbstwirksamkeitserwartung von Lehrern: Längsschnittbefunde mit einem neuen Instrument. Zeitschrift für Päda‑ gogische Psychologie/German J Educ Psychol. 2000. https://doi.org/10. 1024/1010-0652.14.1.12. 111. Kortteisto T, Kaila M, Komulainen J, Mäntyranta T, Rissanen P. Healthcare professionals’ intentions to use clinical guidelines: a survey using the theory of planned behaviour. Implement Sci. 2010;5:1. 112. Lyon AR, Cook CR, Brown EC, Locke J, Davis C, Ehrhart M, Aarons GA. Assessing organizational implementation context in the education 112. Lyon AR, Cook CR, Brown EC, Locke J, Davis C, Ehrhart M, Aarons GA. Assessing organizational implementation context in the education
https://openalex.org/W3083515883	https://scholarlypublications.universiteitleiden.nl/access/item%3A3146951/view	English	null	Lifestyle Enclaves in the Instagram City?	Social media + society	2,020	cc-by	8,788	SI: Studying Instagram Beyond Selfies https://doi.org/10.1177/205630512094069 Social Media + Society July-September 2020: 1–10 © The Author(s) 2020 Article reuse guidelines: sagepub.com/journals-permissions DOI: 10.1177/2056305120940698 journals.sagepub.com/home/sms Instagram, social media, segregation, urban space, integration Instagram, social media, segregation, urban space, integration Although social media corporations and their representatives continue to argue that their platforms incubate a global com- munity, commentators and scholars nowadays stress that social media are at risk of undermining cohesion and democ- racy. The behemoth Facebook is a case in point: CEO Mark Zuckerberg published a lengthy manifesto about his compa- ny’s role in fostering global community at the start of 2017, just a little over a year before he was called to testify before Congress about the company’s role in the American political crisis. Danah boyd’s (2017) verdict was merciless. She called Zuckerberg “naive as hell” for believing in the dream that “he could build the tools that would connect people at unprecedented scale” (n.p.). A large and growing literature documents how social media’s penchant for reinforcing assortative ties results in polarization, balkanization, echo chambers, and filter bubbles (Del Vicario et al., 2016; Pariser, 2011). Such concerns over social media-induced fragmenta- tion dovetail with anxieties about geographic segregation. While titles like American Apartheid (Massey & Denton, 1998) signal long-standing concerns about racial segrega- tion, more recently commentators have expressed worries about lifestyle segregation. In the United States, progressive coastal states are pitted against the conservative Midwest and South. Looking at a lower level of granularity, stark divi- sions between progressive inner-cities and conservative suburbs light up. The city of Amsterdam, our case study area, also exhibits plain contrasts between overwhelmingly left- leaning inner-city areas and more right-leaning outer bor- oughs. These differences in political preferences are tightly coupled with differences in lifestyles and identities, resulting in “lifestyle enclaves” (Bellah et al., 1985; DellaPosta et al., 2015). It seems plausible, perhaps even inescapable, that processes of self-segregation online and offline work together to generate increasingly fragmented social landscapes. This article addresses these concerns by studying social relations of Instagram users in Amsterdam, examining how they form groups, segregate, and claim different places within the city. While we find some evidence of “lifestyle enclaves” among Amsterdam’s Instagram users, we also highlight connections between groups and processes of inte- gration. In short, we try to answer the question to what degree and in what ways processes of fragmentation and integration shape the relations of Amsterdam-based Instagram users. In 1Leiden University, The Netherlands 2University of Amsterdam, The Netherlands 1Leiden University, The Netherlands 2University of Amsterdam, The Netherlands Corresponding Author: John D. Lifestyle Enclaves in the Instagram City? John D. Boy1 and Justus Uitermark2 Abstract Commentators and scholars view both social media and cities as sites of fragmentation. Since both urban dwellers and social media users tend to form assortative social ties, so the reasoning goes, identity-based divisions are fortified and polarization is exacerbated in digital and urban spaces. Drawing on a dataset of 34.4 million interactions among Amsterdam Instagram users over half a year, this article seeks to gauge the level of fragmentation that occurs at the interface of digital and urban spaces. We find some evidence for fragmentation: users form clusters based on shared tastes and leisure activities, and these clusters are embedded in four distinct lifestyle zones at the interface of social media and the city. However, we also find connections that span divisions. Similarly, places that are tagged by Instagram users generally include a heterogeneity of clusters. While there is evidence that Instagram users sort into groups, there is no evidence that these groups are isolated from one another. In fact, our findings suggest that Instagram enables ties across different groups and mitigates against particularity and idiosyncrasy. These findings have important implications for how we should understand and study social media in the context of everyday life. Scholars should not only look for evidence of division through standard network analytic techniques like community detection, but also allow for countervailing tendencies. Creative Commons CC BY: This article is distributed under the terms of the Creative Commons Attribution 4.0 License (https://creativecommons.org/licenses/by/4.0/) which permits any use, reproduction and distribution of the work without further permission provided the original work is attributed as specified on the SAGE and Open Access pages (https://us.sagepub.com/en-us/nam/open-access-at-sage). 698 SMSXXX10.1177/2056305120940698Social Media <span class="symbol" cstyle="Mathematical">+</span> SocietyBoy and Uit 698 SMSXXX10.1177/2056305120940698Social Media <span class="symbol" cstyle="Mathematical">+</span> SocietyBoy and Uit Instagram, social media, segregation, urban space, integration Boy, Institute of Anthropology and Sociology, Leiden University, Wassenaarseweg 52, P.O. Box 9555, Leiden 2333 AK, The Netherlands. Email: j.d.boy@fsw.leidenuniv.nl Creative Commons CC BY: This article is distributed under the terms of the Creative Commons Attribution 4.0 License (https://creativecommons.org/licenses/by/4.0/) which permits any use, reproduction and distribution of the work without further permission provided the original work is attributed as specified on the SAGE and Open Access pages (https://us.sagepub.com/en-us/nam/open-access-at-sage). Social Media + Society 2 the following, we start by presenting our perspective on frag- mentation and integration. We specifically focus on the inter- face of Instagram and the city, examining whether Instagram users indeed self-segregate online and in the city. Using con- cepts and methods that are widely used in contemporary debates on polarization and social media, we then empiri- cally show that Instagram users in Amsterdam do, in fact, sort into groups with specific appearances and lifestyles. We further demonstrate that these groups construct zones at the online–offline interface, that is, symbolic and material domains that serve as stages for the enactment of identity and the performance of status. While our analyses, therefore, confirm that tendencies toward fragmentation are present, we go on to complicate this conclusion by dissecting its con- ceptual and methodological premises. To put it bluntly, we find fragmentation, but only if we neglect any countervailing processes. The second part of the empirical analysis, there- fore, takes a different angle in analyzing our dataset and demonstrates that there are indeed formidable processes of integration at play. The Instagram city, we argue, may be a much more integrated, and much more boring, place than the tidings of fragmentation and conflict would suggest. village community suppresses differences, the city rein- forces and amplifies them (Fischer, 1982; Wirth, 1938). This process makes for extraordinarily vibrant environments of highly diverse subcultures, but also results in fragmentation and its associated evils of anomie, collective paralysis, and failures of empathy. It is no coincidence that contemporary anxiety over social media echoes historical concerns over cities. The move to cities and the development of modern communication tech- nologies are essentially two sides of the same coin: both developments emancipate people from the communities they were born into and allow them to associate with people of their own choosing. Understanding Fragmentation and Integration in the Instagram City As social media allow us to associate with like-minded peo- ple, so communis opinio holds, we are inadvertently yet ineluctably drawn into echo chambers or filter bubbles. To make things worse, algorithms reinforce our propensity to associate with those like us by suggesting we befriend our friends’ friends or read more from the blogs we just visited. Social media, then, feed on our differences and reinforce them, resulting in fragmentation (Pariser, 2011; Sunstein, 2001). While we are increasingly connected to others who are just like us, the distance to others grows. American com- munication scholar danah boyd (2017, n.p.) sums up this pessimistic diagnosis: g g g However, the evidence of a connection between social media use and fragmentation is moderate and mixed, even in the United States context, on which most of the work in this field is focused (Boxell et al., 2017; Garrett, 2009). We therefore need to develop a perspective that allows not just the possibility of fragmentation but also of its opposite, that is, integration. In addition, concerns over fragmentation are typically voiced in relation to Facebook and especially Twitter (Tufekci, 2014), much less in relation to other plat- forms, including Instagram. Since Instagram is much more visual than either Facebook or Twitter, we may expect that it is less likely to induce acrimonious debate. The budding research literature on Instagram emphasizes that the platform is not a staging ground for symbolic resistance (Manovich, 2016, p. 23) but gives ample space to corporate-sponsored influencers to shape tastes and desires (Abidin, 2016; Boy et al., 2018) while compelling users to enact idealized selves and hide stress and strains (Duffy & Hund, 2015). Whereas many studies on Twitter highlight polarization, this literature on Instagram conjures up the image of users connecting in an environment where beauty, wealth, and success are cele- brated and estheticized (cf. Boy & Uitermark, 2017; Marwick, 2015). Ironically, in a world in which we have countless tools to connect, we are also watching fragmentation, polarization, and de-diversification happen en masse. The American public is self-segregating, and this is tearing at the social fabric of the country. Ironically, in a world in which we have countless tools to connect, we are also watching fragmentation, polarization, and de-diversification happen en masse. The American public is self-segregating, and this is tearing at the social fabric of the country. Instagram, social media, segregation, urban space, integration It is plausible that, when the city and social media become intertwined, differences are multiplied, reinforcing mechanisms of fragmentation (Bastos et al., 2018; De Waal, 2014; Graham, 2005; Kitchin & Dodge, 2011; Wang et al., 2018). In this article, we pursue this line of thought by examining how subcultures emerge at the interface of cities and social media (Boy & Uitermark, 2016, 2017). We show how different social groups claim their space and mark their territory. Places figure into this story as stages for expressing individual status and group belonging. Social media users generally do not picture quotidian activi- ties like visiting the supermarket, but rather share experi- ences of places for aspirational consumption (see also Boy & Uitermark, 2016, 2017; cf. Currid-Halkett, 2017). Social media, in this line of thinking, are hyper-segregated: by selectively displaying where social media users are, they reflect and reinforce segregation on the ground. Methods and Data Our analysis in this article is based on a corpus of 709,348 geotagged Instagram posts gathered over half a year, between 1 December 2015 and 31 May 2016. On Instagram, users can opt into geotagging (attaching a location to posts) on a post- by-post basis. Since our main interest is in how city dwellers use locative social media in their everyday lives, our corpus only includes posts by users with at least two geotagged posts at least 4 weeks apart to eliminate likely tourists. The total number of users in our corpus is 78,207, equivalent to about one-tenth of Amsterdam’s population. On 1 June 2016, Instagram severely restricted the data that could be accessed through its application programming interface (API), which is why we focus on the period up to 31 May 2016. We should thus take seriously the possibility that the social media promote integration and breed conformity instead of amplifying difference. There are prima facie rea- sons to consider the scenario. A number of commentators have observed how, in the age of social media, radical and deviant subcultures have withered away. For instance, Jessa Lingel (2017) has written about several groups that used to make a home on the open web during the 1990s, only to be gradually displaced by the ascendant social media platforms when the 2000s came along. One of her studies focuses on extreme body modification, a subculture that for many years had a virtual meeting place at Body Modification Ezine (BME). The founders of BME positioned the platform as an “online haven for outsiders” (Lingel, 2017) where members could share their experiences with face tattoos, scarification, subincisions, stretched ears, piercings, flesh pulls, split tongues, and the like (p. 37). The platform flourished in the late 1990s but faltered as Facebook rose to prominence. Although the story of BME’s decline is complicated, it is clear social media feature prominently. The promise of a wider audience pulled members away from BME and onto social media like Facebook where body modifiers’ sense of alterity and community dwindled. With our corpus, we can investigate how social media is implicated in the creation of subcultures and social divisions. Our main data are the 34.4 million “likes” and comments among the users in our dataset. Understanding Fragmentation and Integration in the Instagram City These concerns about fragmentation sound all too famil- iar to students of the city. The scholars of the Chicago School of sociology argued in their classic works that the disintegrating forces of modernity—as observed by first- generation sociologists like Ferdinand Tönnies and Émile Durkheim—reached their apex in cities. Robert Park (1915) wrote of cities that “every social group tends to create its own milieu . . . The processes of segregation establish moral distances which make the city a mosaic of little worlds which touch but do not interpenetrate” (p. 608). While the When we speak of “integration,” we therefore do not nec- essarily mean a benign process where people with different preferences and interests come together in harmony. We Boy and Uitermark 3 view social media as stages for the expression of status that are characterized by mutual monitoring, the collapse of pri- vate–public distinction, and, most importantly, stratified sys- tems of rank (Boy & Uitermark, 2019). What is essential to social media is not that people share images of posts, but that these contributions are appraised by proximate and distant others. “Integration” means that people come to depend on and surveil one another. Although this process may be har- monious, it can also involve competitive individualism and breed anxiety as social media users are compelled to antici- pate appraisals by others. ethnographic case studies thus suggest that social media do not solely amplify difference but could also facilitate integra- tion and promote conformity. We therefore examine both fragmentation and integration at the interface of Instagram and the city. Before we do so, we discuss our methods and data. view social media as stages for the expression of status that are characterized by mutual monitoring, the collapse of pri- vate–public distinction, and, most importantly, stratified sys- tems of rank (Boy & Uitermark, 2019). What is essential to social media is not that people share images of posts, but that these contributions are appraised by proximate and distant others. “Integration” means that people come to depend on and surveil one another. Although this process may be har- monious, it can also involve competitive individualism and breed anxiety as social media users are compelled to antici- pate appraisals by others. Methods and Data While we acknowledge that the meaning of likes and comments varies across contexts and situations, we pragmatically consider a reciprocated tie (I comment on, or like, your post and you comment on, or like, my post) as a proxy of affinity between users. Out of the interactions in our data, 130,665 are reciprocated, and we use these mutual ties to identify groups and the relations between them. Since we use different methods to study fragmentation and integration, respectively, we provide further details on these methods in the following, empirical sections. Writing on Rotterdam’s gay scene, Ferrie Weeda (2018) also relates the ascendancy of social media, and specifically the dating app Grindr, to the disappearance of a subcultural milieu. After the dating app Grindr allowed its users to seek hook-ups and partners online, the number of gay meeting places dropped precipitously. One after another, gay bars and clubs have closed down. While LGBTQ folks may benefit from the efficiency of the dating app to arrange a tête-à-tête, more radical and collective expressions of difference lose their place within the city. Grindr allows its users to search for specific “tribes” (otter, bear, geek, twink, trans, etc.), but its effect is less to reinforce difference than to have LGBTQ people retreat into privacy or blend into the public. Fragmentation There is plenty of evidence of fragmentation if we look for it. The network of reciprocated likes or comments has a modu- larity score of circa 0.6, a relatively high score (Newman, 2006), suggesting that interactions tend to occur within rela- tively more densely knit subgroups of users and providing prima facie evidence of the fragmentation often associated with city life and social media. We can identify subgroups thanks to a procedure called community detection. By apply- ing the Louvain method of modularity optimization to the undirected network of reciprocated ties (Blondel et al., 2008), we find a total of 31 clusters with at least 200 users. We subsequently engage in qualitative coding. We identify the 10 most central users in each cluster, look for common- alities among them, and characterize the clusters accord- ingly. For instance, when we spot pictures of people flexing their muscles in gyms and references to “personal body plans” in bios, we label the cluster as “fitness enthusiasts.” These examples suggest social media may spur the inte- gration of deviant groups into the mainstream and contribute to the dissolution of subcultural milieus. Difference does not so much disappear but comes to be expressed through ever more subtle and individual strategies of distinction—a pro- cess Elias (1994) captures with the phrase “diminishing con- trasts, increasing varieties” (pp. 382–386). The ambiguous evidence of social media-induced polarization as well as 4 Social Media + Society Table 1. The nine largest clusters of Instagram users. # Label Users Posts Median number of followers 1 City Consumers 4,642 86,925 684 2 Hedonist Lifestyles 4,412 57,879 785 3 City Imageers 3,977 81,482 562 4 Rich Kids 3,196 20,225 552 5 Gay Performers 2,843 31,904 550 6 Refined Lifestyles 2,097 33,682 857 7 Clubbing 2,073 22,234 707 8 Beliebers 1,952 11,740 498 9 Fitness 1,773 23,301 520 Table 1. The nine largest clusters of Instagram users. When users within the same clusters use places as staging grounds for their status displays, we assume a link between those places; the more links between places, the more likely they are part of the same zone. We construct a proximity matrix of places based on how frequently they are tagged by people in the same cluster. We then apply the same method of community detection as discussed above, first turning the proximity matrix into a co-occurrence network. Fragmentation This yields four clusters of places, three of which are of roughly equal size (between 500 and 600 places), and one that is smaller. These zones are not contiguous areas, as in Parks’ “natural areas” or Burgess’ concentric zones model (Park & Burgess, 1925), but sets of places in the city that figure into status displays on Instagram. See Table 2 for an overview of the four zones, which also presents some additional data on the places we sourced from Yelp, the popular social reviews site.2 Before considering in greater detail how these zones and the places that make them up are bound up with status displays in the city, we first describe their features. Through such qualitative coding, we are able to character- ize these clusters in terms of shared interests and lifestyles. In addition to the nine largest clusters shown in Table 1, we identify several smaller ones. They include, in order of decreasing size, international students (1,120 users), Amsterdammers of Turkish descent (821), coffee aficiona- dos (769), Russian-speaking expats (519), evangelicals (395), CrossFit adherents (350), and electronic dance music enthusiasts (304). Table 1 shows that users in the cluster organized around “refined lifestyles” have somewhat more followers than users in the other clusters. However, the great- est inequality is found within clusters rather than among them: all the clusters have very uneven distributions of both likes and comments. We thus find that Instagram users self- organize into clusters of different sizes but all with a median number of followers between 403 (the CrossFit adherents) and 857 (the cognoscenti of refined lifestyles). While there are differences among the clusters in terms of follower count or activity, what stands out are qualitative differences in terms of interests and lifestyles. The nightlife zone consists, at its core, of places associ- ated with the city’s clubbing scenes. According to Yelp, the review site that we mined for additional data on places tagged by Instagram users, a typical closing time for places in the nightlife zone is 6:00 a.m. Concert venues like Paradiso and Melkweg and clubs like Jimmy Woo, Bitterzoet, and Club Air are at the center. Generally, we find most of the city’s dance clubs and a high number of bars and cafes within this zone. Fragmentation Footwear and sportswear stores carrying local brands also rank highly, suggesting that some sartorial and con- sumer choices predominate in the city’s clubbing scenes and serve as a source of subcultural capital (Thornton, 1996). Images tend to show performers and groups swept up in the action. More than in any other zone, the images taken here show moments of collective enjoyment. If the nightlife zone is about dancing, the lifestyle zone appears to be about eating. The yuppie’s favorite meal, brunch, is an important occasion to frequent places in this zone, in which hotels, cafes, and restaurants that serve brunch staples like poached eggs and pancakes predominate. At the center of this zone, we find a number of upmarket hotels— the Conservatorium, The Hoxton, and W Hotel—where patrons like to picture beautifully plated French toast and bespoke cocktails. Boutique coffee places, such as Coffee & Coconuts in De Pijp, as well as a slew of restaurants serving various cuisines also form part of this zone. Several locations on the city’s luxury shopping street, the P.C. Hooftstraat, are among the lifestyle locations, as are other places associated with fashion and design, such as the Dutch headquarters of Hearst, publishers of Elle, Harper’s Bazaar, Esquire, and Cosmopolitan; a showroom of Dutch design; and a fashion retailer specializing in “the good things in life.” Looking again at Yelp reviews, places in this zone have comparatively high ratings and high prices.3 Images foreground moments of consumption, often conspicuous, or at least indicative of sophistication (Currid-Halkett, 2017; Veblen, 1899/1934). Since we are especially interested in the interface of online and urban spaces, we subsequently examine which places these different groups tag. A place tag is a predefined location name that can be attached to a post. This is a form of metadata that enables geographic exploration on Instagram, but also enriches individual posts with additional information.1 In line with the perspective we outlined, we do not inter- pret place tags as “trace data” that can be used to track users’ trajectories through the city, but as features of status displays. Users typically do not post about their daily shopping at the supermarket or their ride to work, but selectively and strate- gically use Instagram as a platform to live out their identities and showcase their social contacts, sense of style, achieve- ments, or new purchases (cf. Fragmentation This aspirational dimension of Instagram use comes out not only in the places that are tagged but also in what is por- trayed in these places and how. We could start with the most notorious genre of social media post, the selfie. Generally, commentary on the selfie is out of proportion to its actual prominence on most social media, and in our data, too, self- ies account for only a small proportion of the total volume (cf. Boy & Uitermark, 2017; Manovich et al., 2014). In the fitness zone, however, posts bearing a #selfie hashtag (or ironic variations like #shamelessgymselfie) can be found much more frequently, which speaks to the centrality of the physical self to status displays staged in this zone.4 In the culture zone, architectural details and outside views predom- inate because here recognition hinges on one’s identification with the branded image of Amsterdam. In the lifestyle zone, bands, performers, and party people literally take center stage—here, esteem is rewarded on evidence of hedonistic pursuits. Unlike in the selfie-saturated fitness zone, portraits are more likely to show groups than individuals. Finally, in the lifestyle zone, still-lifes of desirable items—especially food, fashion, and furniture—speak to the ways in which conspicuous consumption continues to be an avenue toward prestige. In short, by combining community detection with geographic analysis, we can show how subcultural groups use digital technologies to mark their territory within the city, demonstrating that fragmentation occurs not only through residential segregation but also through more complex Finally, the smallest of the four is the fitness zone, which unsurprisingly revolves around sports clubs and gyms. Amsterdam’s CrossFit gyms and yoga studios can be found in this zone along with Yoghurt Barn franchises. This zone comprises not only places appealing to the health-conscious but also establishments that cater to other practices involving the body, such as tattoo parlors and a cryotherapy center (where customers can subject their bodies to temperatures of 110°C below zero for 3 min). More than in other zones, the ideal of expressive individualism shines through in the dis- plays from this zone (Turner, 2011). The uneven presence of clusters in different territories supports the assumption we made at the outset of this inves- tigation: that places are used strategically for displays that play to different social scenes, garnering rewards in the form of esteem or recognition. Fragmentation Boy & Uitermark, 2017; Hochman & Manovich, 2013; Zasina, 2018). Different kinds of places offer resources (props and audiences) for different kinds of displays that garner esteem and prestige in different social scenes. Places are not just physical settings, but also social situations that encourage contextually appropriate expressions of conformity and distinction. 5 Boy and Uitermark Table 2. Zones at the interface of the city and Instagram. Nightlife Lifestyle Culture Fitness Tagged places 604 588 533 91 Average latest closing time 6:00 a.m. 6:00 p.m. 5:00 p.m. 6:30 p.m. Most common hashtags #music #party #paradiso #food #love #coffee #rijksmuseum #netherlands #iamsterdam #fitness #workout #gym Posts 71,312 49,431 42,151 10,976 Users 19,912 13,085 15,648 3,945 Focal areas Leidseplein, Rembrandtplein Herengracht, De Pijp Museumplein, Amsterdam Arena Zuid, Noord Standout places Paradiso, Jimmy Woo The Hoxton, Conservatorium Hotel Rijksmuseum, Eye Film Museum Changing Life Hub, Vondelgym Typical images Bands, performers, dancing Food, clothes, group shots, selfies Art, architecture, outside views Groups, action shots, selfies, outfits Table 2. Zones at the interface of the city and Instagram. The cultural zone revolves around the city’s museums, with the iconic Rijksmuseum at the helm. Alongside it are other well-known landmarks and cultural institutions—the zoo, botanical garden, opera house as well as the public library’s central branch—which are frequented by the city’s cultural connoisseurs and pictured for distant audiences who appreciate images of Amsterdam cityscapes and sights. Users tagging these locations frequently invoke the city brand #iamsterdam, which in its sculpture form is an inescapable sight on Instagram, serving as a metonymy for the city as a whole (which lacks instantly recognizable landmarks on par with the Eiffel Tower or Big Ben). They are tagged by expats and people with a more international audience, suggesting they signal well to these international audiences as markers of being in Amsterdam. clear, albeit less pronounced, tendencies. Nightlife locations are tagged by users in the Hedonist Lifestyles cluster and the Clubbing cluster, while locations in the culture zone are tagged by the cluster of City Imageers as well as internation- ally oriented clusters of expats and foreign-exchange stu- dents. Lifestyle locations are tagged by various clusters of apparel and fashion enthusiasts and the CrossFit cluster (which branches out from the fitness zone into other territo- ries). Unsurprisingly, we also find the cognoscenti of refined lifestyles represented here. Integration There is a large literature that directs our attention to divi- sions online and in the city. Part of the impetus for this litera- ture is the concern that society will fall apart into different groups that at best live past one another, and at worst will clash. The smartphone, as a territory device (De Waal, 2014), is believed to further buttress these divisions by algorithmi- cally reinforcing and fortifying urban dwellers’ propensity to find kindred spirits. We have a range of tools and measures at our disposal to capture such tendencies toward fragmenta- tion. For instance, like many other researchers, we use com- munity detection to identify different groups. However, we need to appreciate that a community detection algorithm will identify communities even in random networks. Identifying groups and their places within the city, as we have done above, risks overemphasizing the differences and fragmenta- tion. We need to look more closely, and we may also need to look differently, if we are to understand the relational struc- tures that social media users spawn. One might counter that the maps in Figure 1 are not at the right scale. Perhaps different clusters all tend to post from the city center but from different places within it. The mixing of different groups in the city center would, in this scenario, reveal profound processes of segregation operating at a lower scale where people might keep out others by constructing of a parochial realm (Lofland, 1998). Although members of dif- ferent clusters traverse the same spaces in the city’s center, they might ultimately self-segregate into different places—a pattern referred to in the literature as “social tectonics” (Robson & Butler, 2001; see also Jackson & Butler, 2015). However, when analyzing at the level of places, we do not find strong support for this scenario. If we look at the 100 places that are tagged the most (in at least 309 posts), we find not a single place where posts originate from one cluster only. While some places are more parochial than others, as a general rule, members of different clusters rub shoulders in bars, squares, restaurants, parks, clubs, or boutiques. There are good reasons to not only look at fragmentation but also integration. While we preferentially connect with like-minded people on social media, strangers are never far away and our audiences are always multiple, at least poten- tially. Fragmentation The fitness zone, in particular, stands out as the preferred domain of several clusters that revolve around fitness and tattooing, suggesting that esteem in these groups is bound up with particular places and their affordances—in this case, getting and maintaining an attrac- tive, fit, and healthy body. The three larger zones which revolve around nightlife, culture, and lifestyle also show Social Media + Society 6 e 1. Heat maps of Instagram activity for different groups centered on the same coordinates. Hotter colors indicate more pla ce tag and the location from which a message is posted do not necessarily coincide—a user may append a place tag for “Club Vividio” while from their homes. Figure 1. Heat maps of Instagram activity for different groups centered on the same coordinates. Hotter colors indicate more place tags. The place tag and the location from which a message is posted do not necessarily coincide—a user may append a place tag for “Club Vividio” while posting from their homes. 7 Boy and Uitermark spatial sorting on the interface of social media and the city (Graham, 2005; Wang et al., 2018). gravity in the center of the city. There are a couple of clusters that also show a lot of activity in Amsterdam South East because of the concert venues in that neighborhood but this is hardly distinctive. If there is one cluster that stands out, it is a cluster with Amsterdammers of Turkish descent (16) that shows a lot of activity in the Western part of the city. However, this cluster, too, gravitates to the center of the city. While both classic writings on the city and contemporary writing on social media would lead us to expect stark differ- ences, we do not at all find that groups sort into internally homogeneous “natural areas.” Integration As we explained above, social media insert users into systems of standardized rank and into each others’ purview (Boy & Uitermark, 2019). Social media users know this all too well and tend to adjust their posts accordingly. Through their exposure to multiple audiences, social media users have to cope with or internalize different kinds of expectations and pressures. Acknowledging that social media bring differ- ent people together, we can write a different story of the same network, using different measures and data points. While it is now clear that members of different clusters traverse the same spaces and rub shoulders in places, perhaps segregation operates in still more subtle and insidious ways. The literature suggests that urbanites who move around in the same neighborhood and even frequent the same places still may have little to no contact. They may “live together apart” as they use digital devices to carve out their parochial domain (De Waal, 2014). If this would be so, there should be little online interaction between members of different clus- ters. This is, again, contrary to what we actually find. Although the network of interactions has a relatively high modularity score of 0.6 (as reported above), remarkably a whopping two out of three interactions are between, not within, clusters. The clusters observed through communities may be distinct, but they are also perforated and intercon- nected. As we can see in Figure 2, even though nodes within each community cluster together, nodes of different colors are also interspersed, indicating that there are numerous ties between communities. Similarly, while we can identify zones that serve as the domain of specific groups, we should also tell a different story. The place network—where differ- ent places are connected when they are tagged by people in the same cluster—has a modularity of only 0.15, signaling that there are many connections between places. For instance, while we followed convention by character- izing clusters according to their most central nodes, we could also look at randomly selected rank and file users. When we know which cluster they belong to, it is usually not so diffi- cult to see why that would be so. For example, users in the cluster of “gay performers” might present themselves as gay or performers. Integration However, gay people and performers are also to be found in other clusters, which means it is very difficult to guess which clusters randomly designed users belong to, suggesting that the communities we find through community detection are not status groups in Weber’s sense or tribes in Mafessoli’s sense (Maffesoli, 1988/1996; Weber, 1921/2010). What is true for community detection also holds for our strategy of identifying zones: it is a method designed to high- light difference by filtering out similarity. When we look at the spatial footprint of the different clusters, as we do in Figure 1, we get a different sense of how Instagram users are positioned within the city. What is remarkable is that the heat maps are so much alike: all clusters have their center of For instance, while we followed convention by character- izing clusters according to their most central nodes, we could also look at randomly selected rank and file users. When we know which cluster they belong to, it is usually not so diffi- cult to see why that would be so. For example, users in the cluster of “gay performers” might present themselves as gay or performers. However, gay people and performers are also to be found in other clusters, which means it is very difficult to guess which clusters randomly designed users belong to, suggesting that the communities we find through community detection are not status groups in Weber’s sense or tribes in Mafessoli’s sense (Maffesoli, 1988/1996; Weber, 1921/2010). What is true for community detection also holds for our strategy of identifying zones: it is a method designed to high- light difference by filtering out similarity. When we look at the spatial footprint of the different clusters, as we do in Figure 1, we get a different sense of how Instagram users are positioned within the city. What is remarkable is that the heat maps are so much alike: all clusters have their center of Communities are neither bubbles nor bounded fields, and interactions frequently span across clusters. Amsterdam Social Media + Society 8 we can discern distinct groups around specific foci like hob- bies, professions, or lifestyles. We further identify distinct zones: sets of interconnected places that serve as the domain for particular kinds of groups. And yet, that is not the whole story. Our findings do not conform to the dystopian image of deep and algorithmically fortified divisions. Integration Nodes are colored according to the clusters they belong to (see Table 1 for a description) and scaled according to their eigenvector centrality. Figure 2. A graph representation of Instagram users in our dataset. Nodes are colored according to the clusters they belong to (see Table 1 for a description) and scaled according to their eigenvector centrality. Instagrammers organize into clusters according to their life- styles and backgrounds but neither on Instagram nor in the city are they far removed from others. People may have a primary reference group that is most consequential for how they understand and comport themselves, but this primary reference group is not apart from rest of the social world. While they associate with people with similar interests and lifestyles, they generally do not form enclaves. The entan- gled networks documented in this article are the structural backdrop of the cross-pressures that users experience as they consider posting to the platform. The aggregate result of these cross-pressures is that Instagram breeds conformity: the platform is used for a range of purposes by different groups but it nevertheless has aesthetic and social norms that all users have to reckon with (Manovich, 2016). Instagrammers organize into clusters according to their life- styles and backgrounds but neither on Instagram nor in the city are they far removed from others. People may have a primary reference group that is most consequential for how they understand and comport themselves, but this primary reference group is not apart from rest of the social world. While they associate with people with similar interests and lifestyles, they generally do not form enclaves. The entan- gled networks documented in this article are the structural backdrop of the cross-pressures that users experience as they consider posting to the platform. The aggregate result of these cross-pressures is that Instagram breeds conformity: the platform is used for a range of purposes by different groups but it nevertheless has aesthetic and social norms that all users have to reckon with (Manovich, 2016). While we do not take issue with this interpretation, we do want to consider another. There certainly are important varia- tions between platforms, but it is nevertheless possible to dis- cern trends. While the internet initially functioned and felt like an alternate reality, it is now increasingly woven into everyday life. Integration Social media accounts make the internet more personal- ized, intimate, and visual, while also making interdependen- cies more extensive, differentiated, and dense. The relational patterns we identify here emerge within this structural context: as we construct our personae and connections through social media, we are compelled to take into account the views of proximate and distant others. The processes and mechanisms we identify on Instagram may be less salient on other Integration Even when users socialize in a community of CrossFit fanatics, they are never far from users with other interests, such as Beliebers or coffee aficionados. Users coalesce into groups, so much is true, but the boundaries of such groups are fuzzy. This casts social media in a different light; perhaps they are best seen as vehicles of integration rather fragmentation. Figure 2. A graph representation of Instagram users in our dataset. Nodes are colored according to the clusters they belong to (see Table 1 for a description) and scaled according to their eigenvector centrality. g g We come to our conclusions based on computational anal- ysis of a slice of data produced by a specific population in a particular place using a platform designed to facilitate visual communication. In grounding our study in Amsterdam, we have chosen a location that at least historically has resisted tendencies toward disintegration and growing inequality (Uitermark, 2009). Today Amsterdam is known as a liberal city, and perhaps our results would have looked differently if we had focused on a more divided city. Considering these specificities, we concede that our case is likely not represen- tative of other populations or platforms. But the same is true for other research based on data sourced from, say, Twitter or Facebook, that informs tidings of fragmentation. One way to account for the differences between our observations of Instagram and others’ observations of Twitter and Facebook is to trace them back to the affordances of different platforms (cf. Van Dijck, 2013; Wellman et al., 2003). On this reading, the patterns of interconnection and pressures toward confor- mity we observe are peculiar to Instagram and the specific functions the application offers its users. Although we read- ily agree that platforms have different affordances, we never- theless feel this kind of argument is limited by its privileging of the technological underpinnings of social relations. Theorizing of affordances originated from the need to move beyond technological determinism and explicitly acknowl- edge that the same technological set-up allows for different kinds of social relations to emerge. And yet, technology remains the starting and end point of analysis—whatever happens, happens because technology affords it, leading researchers to scrutinize design decisions in minute detail. Figure 2. A graph representation of Instagram users in our dataset. Notes 1. Place tags are often rather generic. For instance, users could tag their picture with “Amsterdam” or “Amsterdam West.” In the analysis that follows, we restrict our analysis to clearly defined places on the assumption that they convey a status signal when they are tagged. We eliminate places with a lot of variation in the rooftop coordinates associated with them. This includes place tags for large parks, long streets, or entire neighborhoods. We manually verified the remaining places, keeping 51.2% of tagged locations. There are 1,750 places in Amsterdam that people across clusters tag. Using this list, we look at the co-occurrence of places within these clusters. 2. 2. We were able to gather Yelp data for just over half of the places in our database. Yes, there are radical or outlandish views even on Instagram,5 but there are also powerful pressures toward conformity that render countercultures precarious. Users pursue distinction, but in a conformist way—they know what the norms are and they abide by them. In this article, we bring into view the wider set of relations through which norms are maintained: the fine-grained and cross-cutting linkages within and between communities. The sorts of com- munities discussed at the beginning of this article require a degree of closure to shield its participants from the dominant gaze and have low chances of survival within this constella- tion of the fine-grained and cross-cutting linkages. Instagram users in Amsterdam form an integrated, albeit differentiated, social world. Social media are the interface through which we negotiate what is acceptable, exceptional, or beautiful. The aggregate outcome of these processes of negotiation is not an online space partitioned into a wide range of commu- nities that each have their own ideas or norms, but an expand- ing web of relations that bring people into dependence and implores them to take others’ views into account. 3. 3. The average Yelp rating of 4.15 is noticeably higher than the overall average (4.0), indicating that places in this zone are viewed favorably not just by the Instagrammers who tag them, but by Yelp reviewers as well. Second, Yelp indicates how costly establishments are through the use of repeated dollar signs ($, $$, $$$, and $$$$). Again, the lifestyle establishments score highest, with an average of 2.4 dollar signs, compared to an overall average of just 2.28. Notes In both cases, the differences are slight but significant and further support the impression that the lifestyle zone comprises high-status establishments. We performed a t-test of statistical significance (p < .01). We performed a t-test of statistical significance (p < .01) 4. In the fitness territory, 3.4% of post captions contain “selfie,” as opposed to 0.9% overall. 4. In the fitness territory, 3.4% of post captions contain “selfie,” as opposed to 0.9% overall. 5. At the time of writing, several high-profile members of far- right groups continued to be present on Instagram after having been banned from other platforms; see Sommer (2018). 5. At the time of writing, several high-profile members of far- right groups continued to be present on Instagram after having been banned from other platforms; see Sommer (2018). Acknowledgements We are grateful to Petter Törnberg and Willem R. Boterman for taking the time to read this article and provide us with helpful feed- back and suggestions. We also thank Special Issue editor Alessandro Caliandro for his gracious and incisive comments, and the anony- mous reviewers for numerous suggestions for improving the article. Bellah, R. N., Madsen, R., Sullivan, W. M., Swidler, A., & Tipton, S. M. (1985). Habits of the heart: Individualism and commit- ment in American life. University of California Press. Blondel, V. D., Guillaume, J. L., Lambiotte, R., & Lefebvre, E. (2008). Fast unfolding of communities in large networks. Journal of Statistical Mechanics: Theory and Experiment, 2008(10), 10008. https://doi.org/10.1088/1742-5468/2008/10/P10008 References Abidin, C. (2016). “Aren’t these just young, rich women doing vain things online?” Influencer selfies as subversive frivolity. Social Media + Society, 2(2), 1–17. Bastos, M., Mercea, D., & Baronchelli, A. (2018). The geographic embedding of online echo chambers: Evidence from the Brexit campaign. PLOS ONE, 13(11), Article e0206841. Declaration of Conflicting Interests Boxell, L., Gentzkow, M., & Shapiro, J. M. (2017). Greater internet use is not associated with faster growth in political polarization among us demographic groups. Proceedings of the National Academy of Sciences of the United States of America, 114(40), 10612–10617. The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. Funding Boy, J. D., & Uitermark, J. (2016). How to study the city on Instagram. PLOS ONE, 11(6), Article e0158161. https://doi. org/10.1371/journal.pone.0158161 The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: We gratefully acknowledge funding from the ODYCCEUS Project as financed through the European Union’s Horizon 2020 program (grant agreement no. 732942). Boy, J. D., & Uitermark, J. (2017). Reassembling the city through Instagram. Transactions of the Institute of British Geographers, 42(2), 612–624. Boy, J. D., & Uitermark, J. (2019). Theorizing social media with Elias: Status displays, mutual monitoring, and the genesis of new sensibilities. https://doi.org/10.31235/osf.io/phm5x Conclusion Both social media and the city are widely seen as spaces of fragmentation. In these spaces, commentators expect and fear, people will flock to each other, forming enclaves or bubbles, losing touch with the wider society. Departing from this perspective, this article traces the formation of groups at the interface of the city and Instagram. We indeed find that Boy and Uitermark 9 platforms, but we surmise they are present there, too. So what is the use of our case? We would suggest that Instagram pro- vides an alternative starting point for theorizing. Where researchers of political communication on Twitter or Facebook use their specific cases to theorize about fragmentation, we can use our study of Instagram to highlight mechanisms of integration. Where researchers of political communication view social media posts as expressions of opinion, our case pushes us to consider them as status displays. Our theoretical perspective applied to a specific set of data enables us to iden- tify processes and dimensions that may not have caught the attention of researchers working from a different theoretical perspective and studying different platforms. ORCID iD John D. Boy https://orcid.org/0000-0003-2118-4702 10 Social Media + Society Park, R. E., & Burgess, E. W. (1925). The city. University of Chicago Press. Boy, J. D., Uitermark, J., & Wiersma, L. (2018). Trending #hijab- fashion: Using big data to study religion and the online–urban interface. Nordic Journal of Religion and Society, 31(1), 22–40. Robson, G., & Butler, T. (2001). Coming to terms with London: Middle class communities in a global city. International Journal of Urban and Regional Research, 25(1), 70–86. boyd, d. (2017). Why America is self-segregating. https://points.data- society.net/why-america-is-self-segregating-d881a39273ab Currid-Halkett, E. (2017). The sum of small things: A theory of the aspirational class. Princeton University Press. Sommer, W. (2018). Instagram is the alt-right’s new favorite haven. The Daily Beast. https://www.thedailybeast.com/instagram-is- the-alt-rights-new-favorite-haven DellaPosta, D., Shi, Y., & Macy, M. (2015). Why do liberals drink lattes? American Journal of Sociology, 120(5), 1473–1511. Sunstein, C. R. (2001). Republic.com. Princeton University Press. Del Vicario, M., Vivaldo, G., Bessi, A., Zollo, F., Scala, A., Caldarelli, G., & Quattrociocchi, W. (2016). Echo chambers: Emotional contagion and group polarization on Facebook. Scientific Reports, 6, 37825. https://doi.org/10.1038/srep37825 Thornton, S. (1996). Club cultures: Music, media, and subcultural capital. Wesleyan University Press. Tufekci, Z. (2014). Big questions for social media big data: Representativeness, validity and other methodological pitfalls. In Proceedings of the eighth international Association for the Advancement of Artificial Intelligence conference on weblogs and social media (ICWSM ’14). https://arxiv.org/ftp/arxiv/ papers/1403/1403.7400.pdf De Waal, M. (2014). The city as interface: How digital media are changing the city. Nai010. Duffy, B. E., & Hund, E. (2015). “Having it all” on social media: Entrepreneurial femininity and self-branding among fashion bloggers. Social Media + Society, 1(2). Turner, B. S. (2011). Religion and modern society: Citizenship, secularisation and the state. Cambridge University Press. Elias, N. (1994). The civilizing process. Blackwell. Uitermark, J. (2009). An in memoriam for the just city of Amsterdam. City, 13(2–3), 347–361. Fischer, C. (1982). To dwell among friends: Personal networks in town and country. University of Chicago Press. Van Dijck, J. (2013). The culture of connectivity: A critical history of social media. Oxford University Press. Garrett, R. K. (2009). Echo chambers online? Politically motivated selective exposure among internet news users. Journal of Computer-Mediated Communication, 14(2), 265–285. Veblen, T. (1934). The theory of the leisure class: An economic study of institutions. The Modern Library. (Original work pub- lished 1899) Graham, S. (2005). Software-sorted geographies. Progress in Human Geography, 29(5), 562–580. ORCID iD Wang, Q., Phillips, N. E., Small, M. L., & Sampson, R. J. (2018). Urban mobility and neighborhood isolation in America’s 50 largest cities. Proceedings of the National Academy of Sciences of the United States of America, 115(30), 7735–7740. Hochman, N., & Manovich, L. (2013). Zooming into an Instagram city: Reading the local through social media. First Monday, 18(7). https://doi.org/10.5210/fm.v18i7.4711 Jackson, E., & Butler, T. (2015). Revisiting “social tectonics”: The middle classes and social mix in gentrifying neighbourhoods. Urban Studies, 52(13), 2349–2365. Weber, M. (2010). The distribution of power within the commu- nity: Classes, stände, parties. Journal of Classical Sociology, 10(2), 137–152. (Original work published 1921) Kitchin, R., & Dodge, M. (2011). Code/space: Software and every- day life. MIT Press. Weeda, F. (2018). Hoe ik Rotterdam heb zien veranderen via datin- gapps als Grindr. Vers Beton. https://versbeton.nl/2018/09/hoe- ik-rotterdam-heb-zien-veranderen-via-datingapps-als-grindr/ Lingel, J. (2017). Digital countercultures and the struggle for com- munity. MIT Press. Lofland, L. (1998). The public realm: Exploring the city’s quintes- sential social territory. Aldine De Gruyter. Wellman, B., Quan-Haase, A., Boase, J., Chen, W., Hampton, K., Díaz, I., & Miyata, K. (2003). The social affordances of the internet for networked individualism. Journal of Computer-Mediated Communication, 8(3). https://doi. org/10.1111/j.1083-6101.2003.tb00216.x Maffesoli, M. (1996). The time of tribes: The decline of individual- ism in mass society (D. Smith, Trans.). SAGE. (Original work published 1988) org/10.1111/j.1083-6101.2003.tb00216.x Manovich, L. (2016). Designing and living Instagram photogra- phy. http://manovich.net/index.php/projects/instagram-and- contemporary-image Wirth, L. (1938). Urbanism as a way of life. American Journal of Sociology, 44(1), 1–24. Manovich, L., Stefaner, M., Yazdani, M., Baur, D., Goddemeyer, D., Tifentale, A., & Chow, J. (2014). Selfiecity. http://selfiec- ity.net/ Zasina, J. (2018). The Instagram image of the city: Insights from Lodz, Poland. Bulletin of Geography: Socio-Economic Series, 42(42), 213–225. https://doi.org/10.2478/bog-2018-0040 Marwick, A. (2015). Instafame: Luxury selfies in the atten- tion economy. Public Culture, 27(1), 137–160. https://doi. org/10.1215/08992363-2798379 Author biographies Massey, D. S., & Denton, N. A. (1998). American apartheid: Segregation and the making of the underclass. Harvard University Press. John D. Boy (PhD, City University of New York) is an assistant professor of sociology in the Institute of Cultural Anthropology and Development Sociology at Leiden University. His research inter- ests include urban studies, digital sociology, the sociology of reli- gion, and social and cultural theory. Newman, M. E. J. (2006). Modularity and community structure in networks. Proceedings of the National Academy of Sciences, 103(23), 8577–8582. Justus Uitermark (PhD, University of Amsterdam) is a professor of Urban Geography at the Faculty of Social and Behavioural Sciences of the University of Amsterdam (UvA). His research interests include digital platforms, rapid urbanization, and urban policy. Pariser, E. (2011). The filter bubble: What the internet is hiding from you. Penguin Books. Park, R. E. (1915). The city: Suggestions for the investigation of human behavior in the city environment. American Journal of Sociology, 20(5), 577–612.
https://openalex.org/W3115549735	https://www.researchsquare.com/article/rs-28355/v1.pdf?c=1631855919000	English	null	Automatic strategy to identify the risk of malnutrition in primary care by means of Controlling Nutritional (CONUT) Score: a large population study.	Research Square (Research Square)	2,020	cc-by	4,702	Automatic strategy to identify the risk of malnutrition in primary care by means of Controlling Nutritional (CONUT) Score: a large population study. María Salinas (  salinas_mar@gva.es ) Emilio Flores Hospital Universitari Sant Joan d'Alacant Emilio Flores Hospital Universitari Sant Joan d'Alacant Automatic strategy to identify the risk of malnutrition in primary care by means of Controlling Nutritional (CONUT) Score: a large population study. María Salinas (  salinas_mar@gva.es ) Hospital Universitari Sant Joan d'Alacant https://orcid.org/0000-0003-3420-5509 Maite López-Garrigós Hospital Universitari Sant Joan d'Alacant Emilio Flores Hospital Universitari Sant Joan d'Alacant Carmen Puche Hospital Universitari Sant Joan d'Alacant Pablo Leiva-Salinas Universidad Miguel Hernandez de Elche Facultad de Medicina Ana Santo-Quiles Hospital Universitari Sant Joan d'Alacant Carlos Leiva-Salinas Department Radiology, University of Missouri Research article Keywords: Risk of Malnutrition, Laboratory Organization and Management, CONUT, Albumin, Strategy Posted Date: May 27th, 2020 DOI: https://doi.org/10.21203/rs.3.rs-28355/v1 License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License Results of the 74743 requests from primary care, CONUT was calculated in 9353 patients. 2256 (24.1%) showed risk of malnutrition according CONUT; 630 (6.8%) through albumin. Albumin alone identified 17.5% of mild cases based on CONUT, but every severe case. The risk of malnutrition increased with patient age. In one year, through an expense of 5430 € in measuring the serum non requested test, we could have screened for risk of malnutrition the entire primary care population that underwent laboratory testing. Research article Keywords: Risk of Malnutrition, Laboratory Organization and Management, CONUT, Albumin, Strategy Posted Date: May 27th, 2020 DOI: https://doi.org/10.21203/rs.3.rs-28355/v1 License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License Page 1/11 Page 1/11 Page 1/11 Background Risk of malnutrition can be evaluated through Controlling Nutritional (CONUT) Score based on total peripheral lymphocytes, serum albumin, and total cholesterol. Our aim was to automatically calculate CONUT in primary care when involved markers were requested and compare its performance with albumin to identify patients at risk of malnutrition. CONUT was evaluated according to patient age, and calculated the potential expenses to calculate CONUT, when serum markers were no requested, and measured. Methods We calculated CONUT when the three laboratory markers were requested, compared CONUT and albumin, considering risk of malnutrition CONUT ≥ 2, and albumin < 35 g/L, and counted the number of albumin and/or cholesterol tests that would have been measured to calculate CONUT in patients with total lymphocytes and serum availability, and calculated hypothetical cost. Conclusion CONUT score calculated at no cost when requested involved laboratory markers, detected more patients at risk of malnutrition than albumin, increased with patient age, and could be calculated in all primary care patients by measuring unrequested tests, at a very affordable cost. Setting and subjects The clinical laboratory is located in a 370-bed suburban University Community Hospital that serves the population of the Health Department (HD) (234551 inhabitants). It receives samples from inpatients and outpatients and primary care patients that are phlebotomized in nine different primary care centres. Samples are transported by couriers to the laboratory reception desk. Requests are made by general practitioners (GPs) via computerized order entry, and results are also automatically sent to the electronic medical record. In this study, we included all community inhabitants over 18 years old within the HD covered by the clinical laboratory. The study was approved by the Hospital Research Committee. Study Design A cross sectional study was designed from January 1st 2018 to December 31st 2018. In consensus with GPs, the CONUT score was automatically calculated by our Laboratory Information System (LIS) for every primary care request when the three involved laboratory markers were requested, in accordance with the tool described in Table 1. Background Malnutrition is an unfavorable prognostic factor associated with increased morbidity and mortality, known to increase hospital stays, readmissions, and resources consumption (1). Albumin, the most extensively studied protein to identify patients at risk of malnutrition, is a useful indicator of general nutrition status for older adults in non-acutely ill states (2). Nutritional status evaluated through NUTritional COntrol (CONUT) based on total peripheral lymphocytes, serum albumin Page 2/11 Page 2/11 and total cholesterol has been shown as a valuable tool for the early detection and monitoring of clinical under nutrition. Since its introduction in 2002 (3), its main application has been to evaluate and quantify the prevalence of malnutrition at hospital admission (4). Since then, the CONUT score has been extensively studied as an efficient tool in different types of patients and contexts and the results in different clinical scenarios have been of most relevance (5–18). All the above research was conducted in hospitalized patients, and no meaningful information regarding the use of CONUT in primary care is available. However, it is crucial to look for the feasibility of implementing routine nutritional screening in general practices (19). However, it is crucial to look for the feasibility of implementing routine nutritional screening in general practices (19). Our aim was to automatically calculate the CONUT score in a large number of primary care patients when all three involved laboratory markers were requested, and compare its performance with albumin to identify risk of malnutrition. We planned to evaluate CONUT according to patient age, and evaluate the costs of a potential strategy involving CONUT as a screening tool in all primary care patients that had a laboratory request. We hypothesized that CONUT score could be a valid first line tool to screen for patients at risk of malnutrition in primary care to later be confirmed / discarded with proven nutrition evaluation tools. Data collection For all primary care requests, we collected patient’s demographic data, CONUT and albumin value from the LIS (iGestlab®), and compared CONUT and albumin values to evaluate for concordance, considering risk of malnutrition CONUT ≥ 2, and albumin < 35 g/L. We classified results by patient’s age, and compared them for the different age brackets. We classified results by patient’s age, and compared them for the different ag For the potential strategy involving CONUT as a screening tool in all laboratory requests from primary care, we counted the number of non-demanded albumin and/or total cholesterol, when measured total number of lymphocytes and serum availability. We calculated the hypothetical economic cost based on both tests prices (0.09 € for total cholesterol, and 0.08 € for albumin). Statistical analysis Demographic summary statistics are reported as median (IQR) for continuous and frequency (%) for categorical variables. Together with the requests’ basic information, differences were assessed using Chi- square, U-Mann Whitney test and Kruskal-Wallis test where applicable. A two-sided p ≤ 0.05 rule was utilized as the criterion for rejecting the null hypothesis of no difference. Statistical anlyses were done using Statistical Package for the Social Sciences (SPSS), Version 22 (IBM Corp., Chicago, Illinois, USA). Laboratory methods The blood samples were collected using BD vacutainer tubes (Beckton Dickinson, NJ, USA). Page 3/11 Albumin was measured through an immunoturbidimetric assay and total cholesterol through an enzymatic colorimetric method (Cobas 8000, Roche, Mannheim, Germany). Total lymphocyte were counted with clinical hematology analyzers (Sysmex, Kobe, Japan). D t ll ti Total lymphocyte were counted with clinical hematology analyzers (Sysmex, Kobe, Japan). Data collection Discussion This is the first study that investigates the relationship between CONUT, albumin and the risk of malnutrition in primary care. The CONUT score, based on three widely used laboratory markers, detected more patients at risk of malnutrition, than albumin alone. The intervention to automatically report the CONUT score when all three necessary laboratory tests were requested was successful, and at a zero cost. The number of patients at risk of malnutrition through the CONUT score increased with patient age. By measuring non-requested albumin and/or total cholesterol tests needed for CONUT calculation when serum was available and lymphocytes counted, it would be possible to assess the risk of malnutrition of all primary care patients undergoing laboratory testing, at a very affordable and reasonable cost, for later to be confirmed/discarded with proven nutritional tools. When comparing serum albumin levels < 35 g/L as the unique marker for risk malnutrition with CONUT score, it would result in up to 80% of the subjects being misdiagnosed (20), as albumin alone identified every severe case, but missed a significant number of patients at risk of malnutrition (mild and moderate cases) through CONUT score. However, CONUT score identified the risk of malnutrition in 24.1% of the studied patients, figures that may seem too high for a primary care population. It might be explained for two facts. First, the patients with the three involved markers, and consequently evaluated for risk of malnutrition through both methods, had a median (IQR) age of 68. These numbers are not representative of a primary care population but of a primary care elderly population. Second, there could be patients with clinical suspicion of risk of malnutrition, since the GP ordered serum albumin. Consequently, in our study, CONUT score could have been calculated in a primary care elderly population at risk of malnutrition. Moreover, and to try to explain the differences in the number of patients identified at risk of malnutrition through CONUT and serum albumin, it has been reported that serum albumin cutoff of 35 g/L under- diagnoses malnutrition as defined by validated nutrition screening tools such as Mini Nutritional Assessment (MNA), Nutritional Risk Screening 2002 (NRS-2002), Mini Nutritional Assessment-Short Form (MNA-SF), and Geriatric Nutritional Risk Index (GNRI)) (21). Moreover, the utility of serum albumin and the traditional cutoff (35 g/l) in physically impaired elderly older people has also been reported that is questionable (22). Results During the study period, the clinical laboratory received 74743 requests from primary care. CONUT score was calculated in 9353 patients, as the 3 involved markers were requested, with a median (IQR) age of 68 (54–82) and 35.6% of whom were men. Of those, 2256 (24.1%) showed risk of malnutrition according to the CONUT score (CONUT score ≥ 2), and 630 (6.7%) based on albumin (albumin < 35 g/L). Table 2 shows their demographic information and also albumin values according to CONUT score. Higher CONUT scores were significantly associated with male sex (P < .05), older age (P < .05) and lower albumin value (P < .05). As shown in Fig. 1, through both markers, the risk of malnutrition, increased as did patient age. As shown in Fig. 1, through both markers, the risk of malnutrition, increased as did patient age. 100% was the ability of albumin alone to identify cases of severe malnutrition, but only 17.5% of mild based on CONUT score (Fig. 2). 100% was the ability of albumin alone to identify cases of severe malnutrition, but only 17.5% of mild based on CONUT score (Fig. 2). From the 74743 total patients, in 65390 the score could have been calculated, as they had total peripheral lymphocytes, and also serum availability to measure the missing non-requested test, either cholesterol or albumin. The missing test was total cholesterol in 3351 and albumin in 64107. Through an expense of Page 4/11 Page 4/11 Page 4/11 5430 €, the laboratory could have screened the entire HD primary care population that underwent a laboratory test. 5430 €, the laboratory could have screened the entire HD primary care population that underwent a laboratory test. Discussion Based on the above, it might be reasonable to replace albumin for CONUT score. It could be a better laboratory indicator to identify patients at risk of malnutrition, especially key in overcrowded primary care. Malnutrition will be confirmed/discarded as soon as possible with proven nutritional assessment tools, to establish the appropriate measures for its correction and subsequent monitoring. Our study shows that the CONUT score increased with aging, as we detected more subjects at risk of malnutrition in the elderly population. Malnutrition remains a significant and highly prevalent public Page 5/11 Page 5/11 health problem of developed countries, especially in old population. Earlier identification and appropriate nutrition support may help to reverse or halt the malnutrition trajectory and the negative outcomes associated with poor nutritional status (23). Healthcare professionals should identify the malnutrition risk, and take action, as early as possible (24). Detecting malnutrition in the general population through traditional tools could be too expensive. Our study results of the potential intervention involving CONUT as a screening tool in all primary care patients who had a laboratory request, by measuring serum albumin and/or total cholesterol when not requested, taking advantage of the serum sample availability and the affordable reagent cost, seems cost-effective. Especially if considering the examined population of more than 70000 primary care patients. The results of our study provide us the opportunity to develop a large-scale opportunistic detection of the risk of malnutrition in the general population, to be subsequently confirmed/discarded through traditional malnutrition evaluation tools. The study had some limitations. First, patients with CONUT score showing risk of malnutrition were not studied with the nutrition traditional tools, however it has been confirmed in different settings the validity of the CONUT system for early detection and monitoring of clinical undernutrition (25). Second, the calculated economic investments may not apply to other countries or settings, since our laboratory belongs to the Public Health Network, where reagent prices tend to be low. CONUT score, calculated at no cost when requested involved laboratory markers, detected more patients at risk of malnutrition than albumin, increased with patient age, and could be calculated in all primary care patients by measuring unrequested tests, at a very affordable cost. The early detection of risk of malnutrition at initial stages through the CONUT score, that could be confirmed/discarded through traditional nutritional tools, could greatly improve primary care patient’s outcome. Conclusions CONUT score calculated at no cost when requested involved laboratory markers, detected more patients at risk of malnutrition than albumin, increased with patient age, and could be calculated in all primary care patients by measuring unrequested tests, at a very affordable cost. Abbreviations CONUT Nutritional Control GNRI Geriatric Nutritional Risk GPs General practitioners HD Page 6/11 Page 6/11 Health Department LIS Laboratory Information System MNA Mini Nutritional Assessment MNA-SF Mini Nutritional Assessment-Short Form SPSS Statistical Package for the Social Sciences Declarations - Ethics approval and consent to participate: Written consent was obtained, the study was approved by the Research Committee of the Hospital Universitario San Juan de Alicante. - Consent for publication: Not applicable - Availability of data and materials: The datasets used and/or analysed during the current study are available from the corresponding author on request. - Competing interests: The authors declare that they have no competing interests - Funding: This work did not receive any funding. - Authors' contributions: All authors researched literature and conceived the study. MLG, EF, CP, ASQ and PLS were involved in protocol development and data analysis. MS wrote the first draft of the manuscript. Authors' contributions: All authors researched literature and conceived the s - Authors' contributions: All authors researched literature and conceived the study. MLG, EF, CP, ASQ and - Authors' contributions: All authors researched literature and conceived the study. MLG, EF, CP, ASQ and PLS were involved in protocol development and data analysis. MS wrote the first draft of the manuscript. CLS reviewed and edited the manuscript. All authors read and approved the final manuscript. - Authors' contributions: All authors researched literature and conceived the study. MLG, EF, CP, ASQ and PLS were involved in protocol development and data analysis. MS wrote the first draft of the manuscript. CLS reviewed and edited the manuscript. All authors read and approved the final manuscript. LS were involved in protocol development and data analysis. MS wrote the PLS were involved in protocol development and data analysis. MS wrote the first draft of the manuscript. CLS reviewed and edited the manuscript All authors read and approved the final manuscript CLS reviewed and edited the manuscript. All authors read and approved the final manuscript. LS reviewed and edited the manuscript. All authors read and approved the f - Acknowledgements: The authors would like to express their deep gratitude to all the clinical laboratory staff. - Acknowledgements: The authors would like to express their deep gratitude to all the clinical laboratory staff. References 1. Lobo Támer G, Ruiz López MD, Pérez de la Cruz AJ. [Hospital malnutrition: relation between the hospital length of stay and the rate of early readmissions]. Med Clin (Barc) [Internet]. 2009 Mar 21 [cited 2019 Feb 8];132(10):377–84. Available from: https://linkinghub.elsevier.com/retrieve/pii/S0025775308000547. 2. Cabrerizo S, Cuadras D, Gomez-Busto F, Artaza-Artabe I, Marín-Ciancas F, Malafarina V. Serum albumin and health in older people: Review and meta analysis. Maturitas [Internet]. 2015 May [cited 2019 Jan 24];81(1):17–27. Available from: https://linkinghub.elsevier.com/retrieve/pii/S0378512215000572. Page 7/11 Page 7/11 3. de Ulíbarri Pérez JI, González-Madroño Giménez A, González Pérez P, Fernández G, Rodríguez Salvanés F, Mancha Alvarez-Estrada A, et al. [New procedure for the early detection and control of hospital malnutrition]. Nutr Hosp [Internet]. [cited 2019 Jan 24];17(4):179–88. Available from: http://www.nutricionhospitalaria.com/pdf/3323.pdf. 4. Ignacio de Ulíbarri J, González-Madroño A, de Villar NGP, González P, González B, Mancha A, et al. CONUT: a tool for controlling nutritional status. First validation in a hospital population. Nutr Hosp [Internet]. [cited 2019 Jan 28];20(1):38–45. Available from: http://scielo.isciii.es/pdf/nh/v20n1/original1.pdf. 4. Ignacio de Ulíbarri J, González-Madroño A, de Villar NGP, González P, González B, Mancha A, et al. CONUT: a tool for controlling nutritional status. First validation in a hospital population. Nutr Hosp [Internet]. [cited 2019 Jan 28];20(1):38–45. Available from: http://scielo.isciii.es/pdf/nh/v20n1/original1.pdf. 5. Nishikawa H, Yoh K, Enomoto H, Ishii N, Iwata Y, Takata R, et al. The Relationship between Controlling Nutritional (CONUT) Score and Clinical Markers among Adults with Hepatitis C Virus Related Liver Cirrhosis. Nutrients [Internet]. 2018 Aug 29 [cited 2019 Jan 28];10(9):1185. Available from: http://www.ncbi.nlm.nih.gov/pubmed/30158477. 5. Nishikawa H, Yoh K, Enomoto H, Ishii N, Iwata Y, Takata R, et al. The Relationship between Controlling Nutritional (CONUT) Score and Clinical Markers among Adults with Hepatitis C Virus Related Liver Cirrhosis. Nutrients [Internet]. 2018 Aug 29 [cited 2019 Jan 28];10(9):1185. Available from: http://www.ncbi.nlm.nih.gov/pubmed/30158477. 6. Kang HW, Seo SP, Kim WT, Yun SJ, Lee S-C, Kim W-J, et al. Prognostic Impact of Nutritional Status Assessed by the Controlling Nutritional Status (CONUT) Score in Patients with Surgically Treated Renal Cell Carcinoma. Nutr Cancer [Internet]. 2018 Aug 18 [cited 2019 Jan 28];70(6):886–94. Available from: http://www.ncbi.nlm.nih.gov/pubmed/30235010. 6. Kang HW, Seo SP, Kim WT, Yun SJ, Lee S-C, Kim W-J, et al. Prognostic Impact of Nutritional Status Assessed by the Controlling Nutritional Status (CONUT) Score in Patients with Surgically Treated Renal Cell Carcinoma. Nutr Cancer [Internet]. 2018 Aug 18 [cited 2019 Jan 28];70(6):886–94. Available from: http://www.ncbi.nlm.nih.gov/pubmed/30235010. 7. Iseki Y, Shibutani M, Maeda K, Nagahara H, Ohtani H, Sugano K, et al. Impact of the Preoperative Controlling Nutritional Status (CONUT) Score on the Survival after Curative Surgery for Colorectal Cancer. Chu P-Y, editor. PLoS One [Internet]. 2015 Jul 6 [cited 2019 Jan 28];10(7):e0132488. Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4492767/. 8. http://doi.wiley.com/10.1111/j.1872-034X.2012.01055.x Taniguchi E, Kawaguchi T, Otsuka M, Uchida Y, Nagamatsu A, Itou M, et al. Nutritional assessments for ordinary medical care in patients with chronic liver disease. Hepatol Res [Internet]. Page 7/11 2017 Dec [cited 2019 Jan 28];36(6):1681–5. Available from: https://linkinghub.elsevier.com/retrieve/pii/S0261561416313139. 13. Kono T, Sakamoto K, Shinden S, Ogawa K. Pre-therapeutic nutritional assessment for predicting severe adverse events in patients with head and neck cancer treated by radiotherapy. Clin Nutr [Internet]. 2017 Dec [cited 2019 Jan 28];36(6):1681–5. Available from: https://linkinghub.elsevier.com/retrieve/pii/S0261561416313139. 14. Liang R-F, Li J-H, Li M, Yang Y, Liu Y-H. The prognostic role of controlling nutritional status scores in patients with solid tumors. Clin Chim Acta [Internet]. 2017 Nov [cited 2019 Jan 28];474:155–8. Available from: https://linkinghub.elsevier.com/retrieve/pii/S0009898117303777. 14. Liang R-F, Li J-H, Li M, Yang Y, Liu Y-H. The prognostic role of controlling nutritional status scores in patients with solid tumors. Clin Chim Acta [Internet]. 2017 Nov [cited 2019 Jan 28];474:155–8. Available from: https://linkinghub.elsevier.com/retrieve/pii/S0009898117303777. 15. La Rovere MT, Maestri R, Olmetti F, Paganini V, Riccardi G, Riccardi R, et al. Additional predictive value of nutritional status in the prognostic assessment of heart failure patients. Nutr Metab Cardiovasc Dis [Internet]. 2017 Mar [cited 2019 Jan 28];27(3):274–80. Available from: https://linkinghub.elsevier.com/retrieve/pii/S0939475316301648. 15. La Rovere MT, Maestri R, Olmetti F, Paganini V, Riccardi G, Riccardi R, et al. Additional predictive value of nutritional status in the prognostic assessment of heart failure patients. Nutr Metab Cardiovasc Dis [Internet]. 2017 Mar [cited 2019 Jan 28];27(3):274–80. Available from: https://linkinghub.elsevier.com/retrieve/pii/S0939475316301648. 16. Iwakami N, Nagai T, Furukawa TA, Sugano Y, Honda S, Okada A, et al. Prognostic value of malnutrition assessed by Controlling Nutritional Status score for long-term mortality in patients with acute heart failure. Int J Cardiol [Internet]. 2017 Mar 1 [cited 2019 Jan 28];230:529–36. Available from: https://linkinghub.elsevier.com/retrieve/pii/S0167527316345521. 16. Iwakami N, Nagai T, Furukawa TA, Sugano Y, Honda S, Okada A, et al. Prognostic value of malnutrition assessed by Controlling Nutritional Status score for long-term mortality in patients with acute heart failure. Int J Cardiol [Internet]. 2017 Mar 1 [cited 2019 Jan 28];230:529–36. Available from: https://linkinghub.elsevier.com/retrieve/pii/S0167527316345521. 17. 10.1007/s00384-016-2668-5 Tokunaga R, Sakamoto Y, Nakagawa S, Ohuchi M, Izumi D, Kosumi K, et al. CONUT: a novel independent predictive score for colorectal cancer patients undergoing potentially curative resection. Int J Colorectal Dis [Internet]. 2017 Jan 10 [cited 2019 Jan 28];32(1):99–106. Available from: http://link.springer.com/10.1007/s00384-016-2668-5. 17. 10.1007/s00384-016-2668-5 Tokunaga R, Sakamoto Y, Nakagawa S, Ohuchi M, Izumi D, Kosumi K, et al. CONUT: a novel independent predictive score for colorectal cancer patients undergoing potentially curative resection. Int J Colorectal Dis [Internet]. 2017 Jan 10 [cited 2019 Jan 28];32(1):99–106. Available from: http://link.springer.com/10.1007/s00384-016-2668-5. 18. Page 7/11 2013 Feb [cited 2019 Jan 28];43(2):192–9. Available from: http://doi.wiley.com/10.1111/j.1872-034X.2012.01055.x. 8. http://doi.wiley.com/10.1111/j.1872-034X.2012.01055.x Taniguchi E, Kawaguchi T, Otsuka M, Uchida Y, Nagamatsu A, Itou M, et al. Nutritional assessments for ordinary medical care in patients with chronic liver disease. Hepatol Res [Internet]. 2013 Feb [cited 2019 Jan 28];43(2):192–9. Available from: http://doi.wiley.com/10.1111/j.1872-034X.2012.01055.x. 9. Narumi T, Arimoto T, Funayama A, Kadowaki S, Otaki Y, Nishiyama S, et al. Prognostic importance of objective nutritional indexes in patients with chronic heart failure. J Cardiol [Internet]. 2013 Nov [cited 2019 Jan 28];62(5):307–13. Available from: https://linkinghub.elsevier.com/retrieve/pii/S0914508713001688. 9. Narumi T, Arimoto T, Funayama A, Kadowaki S, Otaki Y, Nishiyama S, et al. Prognostic importance of objective nutritional indexes in patients with chronic heart failure. J Cardiol [Internet]. 2013 Nov [cited 2019 Jan 28];62(5):307–13. Available from: https://linkinghub.elsevier.com/retrieve/pii/S0914508713001688. 10. Solórzano-Pineda OM, Rivera-López FA, Rubio-Martínez B. [Malnutrition incidence in surgical diabetic and non diabetic patients in general surgery department]. Nutr Hosp [Internet]. [cited 2019 Jan 28];27(5):1469–71. Available from: http://scielo.isciii.es/scielo.php?script=sci_arttext&pid=S0212- 16112012000500015. 10. Solórzano-Pineda OM, Rivera-López FA, Rubio-Martínez B. [Malnutrition incidence in surgical diabetic and non diabetic patients in general surgery department]. Nutr Hosp [Internet]. [cited 2019 Jan 28];27(5):1469–71. Available from: http://scielo.isciii.es/scielo.php?script=sci_arttext&pid=S0212- 16112012000500015. 11. Wang X-B, Chen J, Xiang B-D, Wu F-X, Li L-Q. High CONUT score predicts poor survival and postoperative HBV reactivation in HBV-related hepatocellular carcinoma patients with low HBV-DNA levels. Eur J Surg Oncol [Internet]. 2018 Nov 10 [cited 2019 Jan 28]; Available from: https://www.ejso.com/article/S0748-7983(18)31998-X/fulltext. 11. Wang X-B, Chen J, Xiang B-D, Wu F-X, Li L-Q. High CONUT score predicts poor survival and postoperative HBV reactivation in HBV-related hepatocellular carcinoma patients with low HBV-DNA levels. Eur J Surg Oncol [Internet]. 2018 Nov 10 [cited 2019 Jan 28]; Available from: https://www.ejso.com/article/S0748-7983(18)31998-X/fulltext. 12. Kato Y, Yamada S, Suenaga M, Takami H, Niwa Y, Hayashi M, et al. Impact of the Controlling Nutritional Status Score on the Prognosis After Curative Resection of Pancreatic Ductal 12. Kato Y, Yamada S, Suenaga M, Takami H, Niwa Y, Hayashi M, et al. Impact of the Controlling Nutritional Status Score on the Prognosis After Curative Resection of Pancreatic Ductal Page 8/11 Page 8/11 Adenocarcinoma. Pancreas [Internet]. 2018 Aug [cited 2019 Jan 28];47(7):823–9. Available from: http://insights.ovid.com/crossref?an=00006676-201808000-00005. Adenocarcinoma. Pancreas [Internet]. 2018 Aug [cited 2019 Jan 28];47(7):823–9. Available from: http://insights.ovid.com/crossref?an=00006676-201808000-00005. 13. Kono T, Sakamoto K, Shinden S, Ogawa K. Pre-therapeutic nutritional assessment for predicting severe adverse events in patients with head and neck cancer treated by radiotherapy. Clin Nutr [Internet]. Page 7/11 Screening for Malnutrition in Older People. Clin Geriatr Med. 2015 Aug;31(3):429–37. 24. http://doi.wiley.com/10.1111/jan.13297 Bauer S, Halfens RJG, Lohrmann C. Changes in nutritional status in nursing home residents and associated factors in nutritional status decline: a secondary data analysis. J Adv Nurs [Internet]. 2017 Oct [cited 2019 Jun 21];73(10):2420–9. Available from: http://doi.wiley.com/10.1111/jan.13297. 24. http://doi.wiley.com/10.1111/jan.13297 Bauer S, Halfens RJG, Lohrmann C. Changes in nutritional status in nursing home residents and associated factors in nutritional status decline: a secondary data analysis. J Adv Nurs [Internet]. 2017 Oct [cited 2019 Jun 21];73(10):2420–9. Available from: http://doi.wiley.com/10.1111/jan.13297. 25. González-Madroño A, Mancha A, Rodríguez FJ, Culebras J, de Ulibarri JI. Confirming the validity of the CONUT system for early detection and monitoring of clinical undernutrition: comparison with two logistic regression models developed using SGA as the gold standard. Nutr Hosp. 2012 Mar- Apr;27(2):564 – 71 Available from: http://scielo.isciii.es/pdf/nh/v27n2/33_original_23.pdf. Tables Due to technical limitiations, Tables 1-2 are provided in the Supplementary Files section. Page 7/11 10.1007/s00268-016-3549-3 Yoshida N, Baba Y, Shigaki H, Harada K, Iwatsuki M, Kurashige J, et al. Preoperative Nutritional Assessment by Controlling Nutritional Status (CONUT) is Useful to estimate Postoperative Morbidity After Esophagectomy for Esophageal Cancer. World J Surg [Internet]. 2016 Aug 24 [cited 2019 Jan 28];40(8):1910–7. Available from: http://link.springer.com/10.1007/s00268-016-3549-3. 18. 10.1007/s00268-016-3549-3 Yoshida N, Baba Y, Shigaki H, Harada K, Iwatsuki M, Kurashige J, et al. Preoperative Nutritional Assessment by Controlling Nutritional Status (CONUT) is Useful to estimate Postoperative Morbidity After Esophagectomy for Esophageal Cancer. World J Surg [Internet]. 2016 Aug 24 [cited 2019 Jan 28];40(8):1910–7. Available from: http://link.springer.com/10.1007/s00268-016-3549-3. 19. Hamirudin AH, Charlton K, Walton K, Bonney A, Potter J, Milosavljevic M, Hodgkins A, Albert G, Ghosh A, Dalley A. Feasibility of implementing routinenutritional screening for older adults in Australian general practices: amixed-methods study. BMC Fam Pract. 2014 Nov 25;15:186. doi: 10.1186/s12875-014-0186-5. PubMed PMID: 25421546; PubMed Central PMCID:PMC4247715. 19. Hamirudin AH, Charlton K, Walton K, Bonney A, Potter J, Milosavljevic M, Hodgkins A, Albert G, Ghosh A, Dalley A. Feasibility of implementing routinenutritional screening for older adults in Australian general practices: amixed-methods study. BMC Fam Pract. 2014 Nov 25;15:186. doi: 10.1186/s12875-014-0186-5. PubMed PMID: 25421546; PubMed Central PMCID:PMC4247715. 20. Bharadwaj S, Ginoya S, Tandon P, Gohel TD, Guirguis J, Vallabh H, et al. Malnutrition: Laboratory markers vs nutritional assessment. Gastroenterology Report. 2016. 20. Bharadwaj S, Ginoya S, Tandon P, Gohel TD, Guirguis J, Vallabh H, et al. Malnutrition: Laboratory markers vs nutritional assessment. Gastroenterology Report. 2016. 21. Zhang Z, Pereira SL, Luo M, Matheson EM. Evaluation of Blood Biomarkers Associated with Risk of Malnutrition in Older Adults: A Systematic Review and Meta-Analysis. Nutrients [Internet]. 2017 Aug 3 [cited 2019 Jan 28];9(8):829. Available from: http://www.mdpi.com/2072-6643/9/8/829. 21. Zhang Z, Pereira SL, Luo M, Matheson EM. Evaluation of Blood Biomarkers Associated with Risk of Malnutrition in Older Adults: A Systematic Review and Meta-Analysis. Nutrients [Internet]. 2017 Aug 3 [cited 2019 Jan 28];9(8):829. Available from: http://www.mdpi.com/2072-6643/9/8/829. 22. Kuzuya M, Izawa S, Enoki H, Okada K, Iguchi A. Is serum albumin a good marker for malnutrition in the physically impaired elderly? Clin Nutr. 2007. 22. Kuzuya M, Izawa S, Enoki H, Okada K, Iguchi A. Is serum albumin a good marker for malnutrition in the physically impaired elderly? Clin Nutr. 2007. Page 9/11 Page 9/11 23. Guyonnet S, Rolland Y. Screening for Malnutrition in Older People. Clin Geriatr Med. 2015 Aug;31(3):429–37. 23. Guyonnet S, Rolland Y. Figure 2 Concordance in the detection of risk of malnutrition through CONUT and albumin Concordance in the detection of risk of malnutrition through CONUT and albumin Figures Page 10/11 Page 10/11 Page 10/11 Requests found to have CONUT and albumin values showing risk of malnutrition Requests found to have CONUT and albumin values showing risk of malnutrition Requests found to have CONUT and albumin values showing risk of malnutrition Figure 2 Concordance in the detection of risk of malnutrition through CONUT and albumin Supplementary Files Requests found to have CONUT and albumin values showing risk of malnutrition Figure 2 Page 11/11 gure 2 oncordance in the detection of risk of malnutrition through CONUT and albumin Supplementary Files his is a list of supplementary files associated with this preprint. Click to download. ETHICSAPPROVAL.pdf Table2CONUT.pdf Table1CONUT.pdf Supplementary Files This is a list of supplementary files associated with this preprint. Click to download. ETHICSAPPROVAL.pdf Table2CONUT.pdf Table1CONUT.pdf ETHICSAPPROVAL.pdf Table2CONUT.pdf Table1CONUT.pdf Page 11/11
https://openalex.org/W3124505768	https://scindeks-clanci.ceon.rs/data/pdf/0354-8724/2020/0354-87242004285A.pdf	English	null	The use of geocoding for home healthcare application and management an epidemic situation: Case of COVID-19 virus outbreak	Geographica Pannonica	2,020	cc-by	5,516	Abstract The lack of an addressing system is one of the problems of urban management in Algeria, which makes it hard to find the addresses concerned, especially in case of crisis where the decision-makers need ac- curate data in real-time. Like many countries, Algeria follows up the world health organization guide- lines that declared the COVID-19 virus as pandemic and recommended the full quarantine and reduces the social contact as much as possible; however, these procedures weren’t enough to control the in- creasing number of confirmed cases, which exceeded the hospital’s capacities. To face up the outbreak of this pandemic, the Algerian health professionals decided to treat most coronavirus cases at home. This study aims to use a geocoding tool developed in C# programming language and ArcGIS Software Development Kit (SDK) to help in the epidemiological control operation in Ain Touta city and simplifies the interventions using a spatial approach. These problems are addressed by a tool to collect, analyze, store, and process archiving of the geographic data using a geodatabase server. Keywords: Geocoding; COVID-19; quarantine; address; home healthcare Introduction The Algerian cities are mosaics with random and un- controlled urban evolution for many reasons (eco- nomics, politics, security, historic…etc.), so the ter- ritory control and data flow such as the addresses is difficult to handle due lack of accurate information, the local government of many cities finds problems to implement coherent management, weaknesses of the territory control, and the lack of coordination (Akak- ba & Filali 2017). rus in Ain Touta and recording more than 200 con- firmed cases, the local hospital reaches its maximum capacity (50 beds). This epidemiological situation in Ain Touta forced the decision-makers to choose the home healthcare for stable Coronavirus cases where the control of their health condition is done by send- ing the medical staff to houses for monitoring the health condition of the patients. Our paper suggests using a graphical user interface (GUI) tool programmed on the C# programming lan- guage and ArcGIS SDK to use the geocoding technol- ogy, and take the advantages of the addresses to sim- plify the home healthcare, at the same time limiting the spread of the epidemic by applying the social iso- lation principle. ISSN 0354-8724 (hard copy) \| ISSN 1820-7138 (online) ISSN 0354-8724 (hard copy) \| ISSN 1820-7138 (online) A University of Batna 2, Laboratory of Natural Hazards and Spatial Planning (LRNAT), Batna, Algeria, a.akakba@univ-batna2.dz; b.lahmar@univ-batna2.dz; ORCID ID: 0000-0002-8479-8982 * Corresponding author: Ahmed Akakba; e-mail: a.akakba@univ-batna2.dz, Tel:(+231)772135306 The use of Geocoding for Home Healthcare Application and Management an Epidemic Situation. Case of COVID-19 Virus Outbreak Ahmed AkakbaA, Belkacem LahmarA Received: August 20, 2020 \| Revised: October 21, 2020 \| Accepted: November 27, 2020 doi: 10.5937/gp24-28062 Ahmed AkakbaA, Belkacem LahmarA Received: August 20, 2020 \| Revised: October 21, 2020 \| Accepted: November 27, 2020 doi: 10.5937/gp24-28062 Literature review compared each other in many studies such (Zandber- gen, 2014) who compares the previous address data models and find that out of 163,886 original samples of addresses, 36% weren’t used in the final analysis be- cause they could not be reliably geocoded using either street or parcel geocoding. Also, according to (Zand- bergen, 2007) that used geocoding to detect quality on environmental exposure assessment of children liv- ing near high traffic roads (measure pollution on in- dividual address, the result is very accurate) conclude that ”Results of the case-study presented here strong- ly suggest that typical street geocoding is insufficient for fine-scale analysis and more accurate alternatives need to be considered”. The process to convert addresses into geographic loca- tions is known as geocoding, which is a basic opera- tion in geographic information systems (GIS) (Long- ley et al. 2005), it’s widely using in space management (Davis & Fonseca, 2007; Edwards et al., 2014; Zand- bergen, 2014), healthcare and epidemiologic (Krieger et al., 2002; Mazumdar et al., 2008; Oliver et al., 2005), even in civil security (Ratcliffe 2004).f f Today, the Internet is affecting GIS in three major areas: access to GIS data is now simpler than before, spatial information dissemination, and GIS mode- ling/processing. It provides GIS users easy access to GIS data from different data sources and improve data “infrastructure” for providing a basis for valid GIS studies (Yang et al., 2004), because of GIS servers available on the internet such as GeoServer, MapServ- er, Mapnik, MapGuide, QGIS server, etc. There are many potential problems with street ge- ocoding, which have been well described by (Ratcliffe, 2001; Ward et al., 2005; Whitsel et al., 2004; 2006), sev- eral studies (Bonner et al., 2003; Cayo & Talbot, 2003; Dearwent et al., 2001; Karimi et al., 2004) have deter- mined to quantitative estimates of the positional accu- racy of street geocoding by comparing the street geo- coded locations with the ‘true’ location based on taking field measurements using a Global Positioning Sys- tems (GPS) unit, The results in urban areas are gener- ally more accurate than in rural areas. This suggests that the positional error of street geocoding can be sub- stantial and needs to be characterized in a meaningful manner relevant to the use of the geocoded locations. Abstract By March 2020, Algeria was detected the first con- firmed COVID-19 case, after few days the government announces the province of Blida under full quaran- tine for 24/24 hours and 7/7 days and till now (by 24th July 2020) over 80% of the country under the health quarantine. After the outbreak of the COVID-19 vi- Geographica Pannonica • Volume 24, Issue 4, 271–284 (December 2020) 285 The use of Geocoding for Home Healthcare Application and Management an Epidemic Situation. Case of COVID-19 Virus Outbreak Literature review Online geocoding become a good example of this development and increase its functionality to the Inter- net users (Roongpiboonsopit & Karimi 2010). all these servers are open source servers i.e. freely available. Arc- GIS also provides server but it is not available for free, but with many extra on features. The most used servic- es are Google maps and OpenStreetMap but there is a difference in their quality (Präger et al., 2019). f Geocoding is available to implement in three ad- dresses models: as a points, parcels, and as a street networks models; but the geocoding quality have been Geographica Pannonica • Volume 24, Issue 4, 285–293 (December 2020) Address data modelling There are many challenges to building a good refer- ence data (Arctur & Zeiler, 2004; Yang et al., 2004), and technically the addresses can be associated with many kinds of feature classes in a reference database; for example, road centerlines, parcel boundaries, ad- dress points, building structures (Zandbergen, 2007). We can organize geocoding model into four different models as follow: the geographic model, street net- work model, parcel data model, and address point model street. It also supports cases where streets have mul- tiple address ranges and names. Some additional at- tribute characteristics include the use of full block address ranges for major roads. While this results in a better spatial location for known valid address- es, this can also be problematic. When approximated addresses are geocoded against the centerline the re- cords fail to match since the value does not fit into the existing range. The street address in Ain Touta is col- lected from OpenStreetMap server, local government department and from field. The geographic unit model can consist of postal codes (such as ZIP codes in the United States), coun- ties, cities, census enumeration areas, or any other ge- ographic boundary considered meaningful. In the ge- ocoding process, the location assigned to a particular address is the polygon (or the polygon centroid) repre- senting the geographic unit. Location within the unit is not specified, but analyses can be carried out us- ing data associated with the geographic unit. In the case of our study area the geographic unit address is refereeing to the administrative boundary and to the neighborhoods (represented by a polygon). i Parcel boundaries is another address model which is traditionally the most spatially accurate data with address information available. Geocoding against parcels allows for matching against individual plots of land (or the centroids of those polygons) rather than interpolating against a street centerline. This is par- ticularly useful in areas where parcels are not regu- larly addressed (such as on roads with mixed parity) or those parcels that may be quite a distance from the centerline. Study area The city of Ain Touta is located at the intersection be- tween several important roads, particularly Nation- al Road number 03 (RN03), connecting Constantine and Biskra, National Road number 28 (RN28) who linking M’sila and Batna city, as well as the Constan- tine-Biskra railway line. It’s the third most populated city in the province with more than 100000 inhabit- ants in the city. Figure 1. Geographic location of the study area and the user interface of the tool Ain Touta is at 450 km in the South-East of the Al- giers capital, between 6° 7’to 6°13’ east longitude and 35°34’ to 35°31 north latitude, it’s located between two mountain ranges with an altitude ranging between 900 and 950 meters. Ain Touta has one hospital with a capacity of only 50 beds to serve about 100,000 people, as a result, the local health sector of the city suffers from high pres- sure, which is negatively affecting the quality of the health service, in addition to the spread of the Coro- navirus, that threatens the global health system, not to mention the local ones. Figure 1. Geographic location of the study area and the user interface of the tool Geographica Pannonica • Volume 24, Issue 4, 285–293 (December 2020) 286 Ahmed Akakba, Belkacem Lahmar Figure 2. Spread of covid-19 between 1 June – 1 July 2020 in Ain Touta The current outbreak caused pressure on local gov- ernments, as a result the health infrastructures be- come unable to take care all coronavirus cases, since the first case recorded in the city on June 1, 2020, rais- ing the number on June 14 to 11 cases, and by the end of the month of June, hundreds of cases recorded. And hundreds of unconfirmed cases. i By 20 June the local hospital reaches its maximum capacity and the local health department proposed the “Home healthcare” for Coronavirus cases, and the medical staff check the patients periodically at home. f To handle the home healthcare management, we propose to create a geographic database to store at- tributes of every confirmed case and managed by an application that using the geocoding technic. Figure 2. Spread of covid-19 between 1 June – 1 July 2020 in Ain Touta Geographica Pannonica • Volume 24, Issue 4, 285–293 (December 2020) Address data modelling • What is the preferred model to standardize ad- dresses in Algeria? • What is the preferred model to standardize ad- dresses in Algeria? • What is the preferred model to standardize ad- dresses in Algeria? • the street model that store the house number in both side if any, To answer these questions, the approach used con- sists of analyzing the main models of address geoc- oding and the possibility of composing a standard model suitable for our cities. This data contains in- direct textual references about location, such as place names (toponyms) or postal addresses. To map such data, each item needs to be geocoded, i.e. assigned with coordinates through the matching of an indi- rect spatial reference with entities identified in a ge- ographical data source (e.g. a map georeferenced in a well-known coordinate reference system) (Goldberg et al., 2007).h • partially, the parcel model as this address model is new and not full adopted officially; Geocoding in Algeria Geocoding is the ability to transform the indirect spa- tial reference of a piece of information into a direct spatial reference, many geocoding web services have been developed to fulfill this need (Google Geocoding API, Mapzen1), public agencies or from the open- source community (OSM Nominatim3, Gisgraphy4). Geocoding in Alegria is classed on the 3rd level on the quality of the address which is the lightest-shaded countries, this level provides a fair geocoding experi- ence. Address searches sometimes result in matches to point address and street address levels, but more of- ten match to the street name and admin levels (“Geoc- ode coverage—ArcGIS REST API: World Geocoding Service \| ArcGIS for Developers”, 2020). This paper aims to develop a geocoding tool to: This paper aims to develop a geocoding tool to: h • Use ArcGIS SDK. Net framework to handle geoco- ding information. • Create a graphic user interface (GUI) to add, edit, update, and find the address. i • A qualification of geocoding results in textual and spatial terms. To understand the address in Algeria and improve its quality, it is essential to answer the following ques- tions: • Synchronize database between online server and offline user. fl • Solve no-structuring address data and simplify manage the urban space. Address data modelling The parcel boundaries in our study area are not accurate due historical and planning reasons (The parcels are mostly out of control of the local gov- ernment and was located far from the city at first time so they were not integrated into the official plan of the city and didn’t benefit from an official addressing sys- tem till now, of course there is a new geocoding sys- tem proposed by the Algerian government to solve this situation started earlier this year 2020). The street network model is most widely employed address data model and based on street network data. In this approach, a street network is represented as street line segments that hold street names and the range of house numbers and block numbers on each side of the street. Address geocoding is accomplished by first matching the street name, then the segment that contains the house numbers, and finally placing a point along the segment based on a linear interpola- tion within the range of house numbers.h To overcome the limitations of parcels for geoco- ding, address points have emerged as the main address data model. The address point data model can also be constructed from several existing data layers such as parcel data. The point address model is the most accu- rate model because it can be constructed from several existing address data and merged into one geocoding The street network address model facilitates stor- ing different names and address ranges for differ- ent sides of the street and enables validation of cas- es where there is no address range for one side of the 287 The use of Geocoding for Home Healthcare Application and Management an Epidemic Situation. Case of COVID-19 Virus Outbreak database, this is called the composite address. This is the preferred model to solve no-structuring address in our study, which is constructed from: • What are the obstacles preventing the use of Geoc- oding in Algeria? • What are the obstacles preventing the use of Geoc- oding in Algeria? • What are the obstacles preventing the use of Geoc- oding in Algeria? • What type of model should it adopt in Algeria? • What type of model should it adopt in Algeria? • the geographic address model that represent the administrative and neighborhoods boundaries, • What is the preferred model to standardize ad- dresses in Algeria? Locator addressh The locator address is generated previously using GIS software; it’s a composite geocoding that merges: The locator address is generated previously using GIS software; it’s a composite geocoding that merges: • Open source extension. t • District locator address: this locator contained the district number of neighborhoods. • Compatible with OGC standard. • Broadly implemented (GDAL, QGIS, R, Python, Esri, …) • Neighborhood name locator; this is simple locator created for demonstrate purpose, it contains one neighborhood, but we can include all neighbor- hoods in Ain Touta city. • More lightweight than a real geodatabase, but as fast as a geodatabase • Everything is contained in a single file and easier file management with GeoPackage than with Esri Shapefile • Bloc number: a locator represents the bloc number of every bloc. • The city locator is to represent the city, but as men- tioned before in literature review, we can use any i • Faster work-flow than with Shapefiles.h • There are nearly no limitations • There are nearly no limitations Resultants and discussions The proposed tool has geographical part and descriptive part as we explain before in the approach. The proposed tool has geographical part and descriptive part as we explain before in the approach. – The geocoding (.loc) file: a file store offline geoc- oding data in case of unavailable of an online ge- ocoding service. The source code and functions using is explained as follow:i – Esri Basemaps and online resources, Approaches Model and approach ESRI base map Data base service Geo package and geocode ﬁles Graphic Use Inteface Tool ArcGIS Sdk.net funkctionalities Display/analyse online Archiving - Add new address - Address information - Find address - Update addresses Requested data and service (input) Provider functionalities (outputs) • Outputs: are a referenced geographically dataset that select the location they are most often ren- dered into simple two-dimensional points associ- ated with inputs. • The processing algorithm: consists in finding the best-matching element from the reference dataset. • The processing algorithm: consists in finding the best-matching element from the reference dataset. Figure 4. Model and approach Open source file It’s possible to use traditional database file extension (.shp, geodatabase) in the SDK, but we prefer to con- vert all files to an (. gpkg) file. Because Geopackage allow the following features: Collecting Data h online geocoding server to get the geographic ad- dress (state, zip, city…etc.). The data provided in this this paper are collected from different sources and it’s a personnel field survey that started from 01st June 2020 until 1st of July 2020. The exact Coronavirus confirmed number in the city and the most important the location (the addresses) of pa- tient are officially unavailable because of the patient privacy policy. The feature to represent the address geographical- ly is the point, according to previous cited studies, the point is the most accurate data and give us very ac- curate result compared to street and parcel and geo- graphic address. It should be mentioned that the personal data was hidden to protect the privacy of the infected patients, but in a way that does not negatively affect the scien- tific value of this scientific paper. Approaches This approach is to simplify the management of the addresses geographically as follow: • Create Geo-server database: provided the request- ed data such as base maps. • Use ArcGIS SDK and C# programming language to create a desktop user interface (UI) and assure the connection between the UI and server database. • Use ArcGIS SDK and C# programming language to create a desktop user interface (UI) and assure the connection between the UI and server database. • The desktop user interface has 4 mains parts: • The desktop user interface has 4 mains parts: a) Geographical map that display the map and lo- cations of addresses. b) Part to display related information current loca- tion, current address selected. Figure 3. Different parts of the User Interfaces (UI) Tool c) A part that can add, modify addresses d) Assure synchronize data between offline-online geodatabase. • GeoServer: it hosts all addresses information and locations for every single address, it’s stocked on- line to make it possible to work online or archiving, or save a copy of it offline file (.gpkg). The following figure illustrates the approach taken. The data is directly associated with geographical coordinates with inputs, output, and algorithms as below: • Inputs: are the addresses text information such as:12 Mustakbal street, stand, Ain Touta, Algeria) or incomplete information such (al-moujahidine, Figure 3. Different parts of the User Interfaces (UI) Tool 288 Geographica Pannonica • Volume 24, Issue 4, 285–293 (December 2020) 288 Geographica Pannonica • Volume 24, Issue 4, 285–293 (December 2020) 288 Ahmed Akakba, Belkacem Lahmar Figure 4. Model and approach ESRI base map Data base service Geo package and geocode ﬁles Graphic Use Inteface Tool ArcGIS Sdk.net funkctionalities Display/analyse online Archiving - Add new address - Address information - Find address - Update addresses Requested data and service (input) Provider functionalities (outputs) Ahmed Akakba, Belkacem Lahmar Figure 4. Model and approach ESRI base map Data base service Geo package and geocode ﬁles Graphic Use Inteface Tool ArcGIS Sdk.net funkctionalities Display/analyse online Archiving - Add new address - Address information - Find address - Update addresses Requested data and service (input) Provider functionalities (outputs) the full name) or (rue louchen taher.) or simply re- fer to a landmark (theatre de Batna, radio ourass, rahba…etc). Figure 4. 2. Add new address and save it as (.gpkg) file. 2. Add new address and save it as (.gpkg) file. 1. Initiate all necessary files yi In this initiate part (See ‘Initiate database and nec- essary files‘ section in annex), it uploads all the necessary files and databases, framework…etc. in- cluding:i i In this part we create an attribute table program- matically (See ‘Add new address and save it as (.gpkg) file’ section in annex), this table contains five attributes to store necessaire address parts (ad- mirative address, a neighborhoods name, door number, apartment number), the table also stores the full name as a string of character and every new input is assigned automatically to an ID. i In this part we create an attribute table program- matically (See ‘Add new address and save it as (.gpkg) file’ section in annex), this table contains five attributes to store necessaire address parts (ad- mirative address, a neighborhoods name, door number, apartment number), the table also stores the full name as a string of character and every new input is assigned automatically to an ID. – Address Geo-package file with (.gpkg) exten- sion it’s a free open source file that hold all at- tribute table. Geographica Pannonica • Volume 24, Issue 4, 285–293 (December 2020) 289 The use of Geocoding for Home Healthcare Application and Management an Epidemic Situation. Case of COVID-1 The use of Geocoding for Home Healthcare Application and Management an Epidemic Situation Case of COVID-19 Virus The use of Geocoding for Home Healthcare Application The use of Geocoding for Home Healthcare Application and Management an Epidemic Situation. Case of COVID-19 Virus Outbreak Figure 5. Add new address from field Figure 6. Find the patient by address and display information as needed Figure 5. Add new address from field Figure 6. Find the patient by address and display information as needed Figure 6. Find the patient by address and display information as needed Figure 5. Add new address from field Figure 5. Add new address from field The next figure (Figure 5) shows the result of exe- cuting the previous command. pressure on the hospital, and help in epidemiologi- cal development analysis. 4. Archiving and real-time informationh The main purpose of this tool is to standardize ad- dresses and in the same time help the medical stuff to work in the field using a mobile device to collect data, especially in cases like the COVID-19 pan- demic, where the ability of collected data from the field and make it available to display, to make anal- yses…etc., is very essential for the decision-mak- er. This tool is hosted by a local cartographic server that created for this purpose. Also, collecting data operation is synchronized with decision-maker who controls all operations from the control center. The search process code (See ‘Show information about selected item’ section in the annex) is dy- namic so we can search by name or by address and the result will show: the corresponding full name, the complete address (the door number, par- cel number, street name and neighborhood ) , ge- ographic location, and any other information as needed (health status, date of recovery, last med- ical examination…etc.), this part of the code will facilitate patient access for medical intervention, increase healthcare quality as a result of reducing The main purpose of this tool is to standardize ad- dresses and in the same time help the medical stuff to work in the field using a mobile device to collect data, especially in cases like the COVID-19 pan- demic, where the ability of collected data from the field and make it available to display, to make anal- yses…etc., is very essential for the decision-mak- er. This tool is hosted by a local cartographic server that created for this purpose. Also, collecting data operation is synchronized with decision-maker who controls all operations from the control center. The search process code (See ‘Show information about selected item’ section in the annex) is dy- namic so we can search by name or by address and the result will show: the corresponding full name, the complete address (the door number, par- cel number, street name and neighborhood ) , ge- ographic location, and any other information as needed (health status, date of recovery, last med- ical examination…etc.), this part of the code will facilitate patient access for medical intervention, increase healthcare quality as a result of reducing Geographica Pannonica • Volume 24, Issue 4, 285–293 (December 2020) References Journal, 1–15. https://doi.org/10.1007/s10708-020- 10189-1 Journal, 1–15. https://doi.org/10.1007/s10708-020- 10189-1 Arctur, D., & Zeiler, M. (2004). Designing Geodatabas- es: case studies in GIS data modeling. ESRI, Inc. Longley, P. A., Goodchild, M. F., Maguire, D. J., & Rhind, D. W. (2005). Geographic information sys- tems and science. John Wiley & Sons. Bonner, M. R., Han, D., Nie, J., Rogerson, P., Vena, J. E., & Freudenheim, J. L. (2003). Positional accura- cy of geocoded addresses in epidemiologic research. Epidemiology, 14(4), 408–412. Mazumdar, S., Rushton, G., Smith, B. J., Zimmerman, D. L., & Donham, K. J. (2008). Geocoding accuracy and the recovery of relationships between environ- mental exposures and health. International Journal of Health Geographics, 7(1), 13. Cayo, M. R., & Talbot, T. O. (2003). Positional error in automated geocoding of residential addresses. In- ternational journal of health geographics, 2(1), 10. Davis, C. A., & Fonseca, F. T. (2007). Assessing the certainty of locations produced by an address geoc- oding system. Geoinformatica, 11(1), 103–129. Oliver, M. N., Matthews, K. A., Siadaty, M., Hauck, F. R., & Pickle, L. W. (2005). Geographic bias related to geocoding in epidemiologic studies. Internation- al Journal of Health Geographics, 4(1), 29. Dearwent, S. M., Jacobs, R. R., & Halbert, J. B. (2001). Locational uncertainty in georeferencing public health datasets. Journal of Exposure Science & En- vironmental Epidemiology, 11(4), 329–334. Präger, M., Kurz, C., Böhm, J., Laxy, M., & Maier, W. (2019). Using data from online geocoding servic- es for the assessment of environmental obesogen- ic factors: a feasibility study. International journal of health geographics, 18(1), 13.f Edwards, S. E., Strauss, B., & Miranda, M. L. (2014). Geocoding Large Population-level Administrative Datasets at Highly Resolved Spatial Scales. Trans- actions in GIS, 18(4), 586–603. Ratcliffe, J. H. (2001). On the accuracy of TIGER-type geocoded address data in relation to cadastral and census areal units. International Journal of Geo- graphical Information Science, 15(5), 473–485. Karimi, H. A., Durcik, M., & Rasdorf, W. (2004). Eval- uation of uncertainties associated with geocoding techniques. Computer-Aided Civil and Infrastruc- ture Engineering, 19(3), 170–185. g f Ratcliffe, J. H. (2004). Geocoding crime and a first es- timate of a minimum acceptable hit rate. Interna- tional Journal of Geographical Information Science, 18(1), 61–72. Krieger, N., Chen, J. T., Waterman, P. D., Soobader, M.- J., Subramanian, S., & Carson, R. (2002). Conclusion The geocoding tool is used for crisis management and detect proximal service for healthcare which is very im- portant in the Algerian case (Lahmar et al., 2020). In this paper we try to expand the use of geocoding as an impor- tant technic in health geography, to simplify the follow- ing data through one common platform that simplifies the exchange of information and search for it in the real- time. As the main result of this paper is to use the geoco- ding to handle the quarantine crisis using home health- care. Geocoding was also used to monitor the pandemic situation in Ain Touta city and help in assembling and management of the local health resources. Downloading the application using smartphones and tablets which help the medical staff to work directly on the ground and make it easy to do statistics like cen- sus and update data periodically, so that every interven- tion in the field is recorded and stored the process on the server at the level of the crisis management center. Among most important results are:i Among most important results are:i Among most important results are:i Among most important results are:i • Contribute significantly to the success of the quar- antine and social isolation. • Using the home healthcare improved the health system flexibility and extended its capacity from capacity 40 patients in the hospital to over 200 pa- tients using home healthcare. To achieve the goals of the research, we based on a digital approach by accomplished a geographic data- base that was localized in a cartographic server and exploiting by developing an interface (application) us- ing the C# programming language to access the data- base in the cartographic server. • Using objective inquiry to determine the needs of the patients. • Determine the type of disease and medication for those with chronic diseases to benefit from period- ic home healthcare. Geographica Pannonica • Volume 24, Issue 4, 285–293 (December 2020) 290 Ahmed Akakba, Belkacem Lahmar Ahmed Akakba, Belkacem Lahmar • Using the spatial statistics to know the number and distribution of patients in each neighborhood, and to determine the geographical area of each health facility. Conclusion This paper is treated 4 main issues at the same time: Crisis management, database, and data flow management in the territory, and simplify the ter- ritory problems; in our case, we handle the unstruc- tured addresses, using a server to store, display, and analyses data in order to generalize this approach throughout any territory, it is necessary to fill and preserve information and ensure that data input in real-time. y • Using network analysis and geocoding to deter- mine addresses and travel paths between the health facility and patient’s homes. • Increase principle of the proximity in services so every facility assures the need of the nearest popu- lation, as a result the territory will be more hierarchy. References Geocoding and monitoring of US socioeconomic inequalities in mortality and cancer incidence: does the choice of area-based measure and geographic level matter? the Public Health Disparities Geocoding Project. American journal of epidemiology, 156(5), 471–482. Roongpiboonsopit, D., & Karimi, H. A. (2010). Com- parative evaluation and analysis of online geoco- ding services. International Journal of Geographi- cal Information Science, 24(7), 1081–1100. Ward, M. H., Nuckols, J. R., Giglierano, J., Bonner, M. R., Wolter, C., Airola, M., et al. (2005). Positional accuracy of two methods of geocoding. Epidemiol- ogy, 16(4), 542–547. La Direction Technique Chargée des Statistiques Ré- gionales, l’Agriculture et de la C. (2008). Collec- tions Statistiques n° 163/2011 Série S : Statistiques Sociales Armature urbaine - RGPH 2008 -. Office National des Statistiques. http://www.ons.dz/IMG/ pdf/armature_urbaine_2008.pdf Whitsel, E. A., Quibrera, P. M., Smith, R. L., Catellier, D. J., Liao, D., Henley, A. C., & Heiss, G. (2006). Ac- curacy of commercial geocoding: assessment and implications. Epidemiologic Perspectives & Innova- tions, 3(1), 8. Lahmar, B., Dridi, H., & Akakba, A. (2020). Territo- rial health approach outputs of geo-governance of health facilities: case study of Batna, Algeria. Geo- 29 Geographica Pannonica • Volume 24, Issue 4, 285–293 (December 2020) 291 The use of Geocoding for Home Healthcare Application and Management an Epidemic Situation. Case of COVID-19 Virus Outbreak Zandbergen, P. A. (2007). Influence of geocoding quality on environmental exposure assessment of children living near high traffic roads. BMC pub- lic health, 7(1), 37. Whitsel, E. A., Rose, K. M., Wood, J. L., Henley, A. C., Liao, D., & Heiss, G. (2004). Accuracy and repeata- bility of commercial geocoding. American journal of epidemiology, 160(10), 1023–1029. Zandbergen, P. A. (2014). Ensuring confidentiality of geocoded health data: Assessing geographic mask- ing strategies for individual-level data. Advances in medicine, 2014. Zandbergen, P. A. (2014). Ensuring confidentiality of geocoded health data: Assessing geographic mask- ing strategies for individual-level data. Advances in medicine, 2014. Yang, D.-H., Bilaver, L. M., Hayes, O., & Goerge, R. (2004). Improving geocoding practices: evalua- tion of geocoding tools. Journal of medical systems, 28(4), 361–370. ANNEX The use of Geocoding for Home Healthcare Application and Management an Epidemic Situation. Case of COVID-19 Virus Outbreak Add new address and save it as (.gpkg) file Add new address and save it as (.gpkg) file 2 G g hi P i V l 2 I 28 293 (D b 2020) G hi P i V l I 8 (D b ) 292 Geographica Pannonica • Volume 24, Issue 4, 285–293 (December 2020) Ahmed Akakba, Belkacem Lahmar Ahmed Akakba, Belkacem Lahmar Show information about selected item Show information about selected item 29 Geographica Pannonica • Volume 24, Issue 4, 285–293 (December 2020) 293 Geographica Pannonica • Volume 24, Issue 4, 285–293 (December 2020)
https://openalex.org/W4283797460	https://munin.uit.no/bitstream/10037/28542/2/article.pdf	English	null	How to quantify bilingual experience? Findings from a Delphi consensus survey	Bilingualism	2,022	cc-by	12,415	How to quantify bilingual experience? Findings from a Delphi consensus survey Bilingualism: Language and Cognition 1Initiatives such as the BLC mini-series (Luk & Esposito, 2020) which aim to gather systematic collections of tools used to document bilingual experiences are a step in the right direction, but they do not aim to enhance the comparability of measures used across bilingualism research and across sectors. Abstract While most investigations of bilingualism document participants’ language background, there is an absence of consensus on how to quantify bilingualism. The high number of different language background questionnaires used by researchers and practitioners jeopardises data comparability and cross-pollination between research and practice. Using the Delphi consen- sus survey method, we asked 132 panellists (researchers, speech and language therapists, tea- chers) from 29 countries to rate 124 statements on a 5-point agreement scale. Consensus was pre-defined as 75% agreement threshold. After two survey rounds, 79% of statements reached consensus. The need for common measures to quantify bilingualism was acknowledged by 96% of respondents. Agreement was reached to document: language exposure and use, lan- guage difficulties, proficiency (when it cannot be assessed directly), education and literacy, input quality, language mixing practices, and attitudes (towards languages and language mix- ing). We discuss the implications of these findings for the creation of a new tool to quantify bilingual experience. cambridge.org/bil Cécile De Cat1 , Draško Kašćelan2, Philippe Prévost3, Ludovica Serratrice4 , Laurie Tuller3, Sharon Unsworth5 and The Q-BEx Consortium* Research Article *The Q-BEx Consortium: Nouhad Abou Melhem, Ashley Adams, Eva Aguilar-Mediavilla, Nadia Ahufinger, Shanley Allen, Llorenc Andreu, Effrosyni Froso Argyri, Sharon Armon-Lotem, Jacqueline Arsenault, Heather Baer, Colin Baxter, Lisa Bedore, Elma Blom, Mirjam Blumenthal, Ute Bohnacker, Claudine Bowyer-Crane, Krista Byers-Heinlein, Allegra Cattani, Shravasti Chakravarty, Vicky Chondroggiani, Dayna Camilleri Clarke, Joanne Cook, Daniel Cubilla-Bonnetier, Ineta Dabasinskiene, Kankan Das, Rachael Davis, Angela de Bruin, Annick De Houwer, Martine Diab, Maaike Diender, Helen Drummond, Heli Elovaara, Victoria Farrell, Karina Fascinetto-Zago, Giulia Filippi, Caroline Floccia, Max Freeman, Silke Fricke, Margaret Friend, Carolyn Gelenter, Samina Ghafoor, Kleanthes K. Grohmann, Matt Hall, Ewa Haman, Cornelia Hamann, Riikka Härkönen, Jeni Harrison, Carly Hartshorn, Maha Hassan, Erika Hoff, Nayr Ibrahim, Sladjana Indjin, Katy Isaac, Susan Joffe, Holly Joseph, Maria Kambanaros, Kathryn Kashyap, Olga Kepinska, Andrea Kewin, Ekaterini Klepousniotou, Tanja Kupisch, Francesca La Morgia, Gabrielle Lai, Nienke Lam, Jane Le Roux, Mia Le Roux, Annina Manninen, Virginia Marchman, Theodoros Marinis, Prisca Martin, Lourdes Martinez Nieto, Chantal Mayer-Crittenden, Natalia Meir, Liz Metcalfe, Eléonore Morin, Lina Mukhopadhyay, Victoria A. Murphy, Vishnu KK Nair, Marie Newton, Arja Nieminen, Silvia Nieva, Audrey Noel, Ewelina Urszula O’Donnell, Therese O’Sullivan, Ciara O’Toole, Weronika Ozpolat, Felicity Parry, Michelle Pascoe, Rupam Patel, Vrishali Patil Ingle, Sean Pert, Anne-Gaelle Piller, Eva Poort, Christine Potter, Eveliina Rantanen, Saiqa Riasat, Lidia Rodriguez, Wiebke Scharff Rethfeldt, Bethany Faye Schwartz, Irina Sekerina, Miquel Serra Raventós, Saleh Shaalan, Jinder Singh, Beth Skelton, Vicky Slaughter, Anne Margaret Smith, Sini Smolander, Neal Snape, Marina Sokolova, Antonella Sorace, Lisa Spinney-Hutton, Frances Jane Stokes, Luke Swift, Vasim Salim Tamboli, Elena Theodorou, Elin Thordardottir, Ianthi Tsimpli, Rachael Tuckley, Olga Urek, Liza van den Bosch, Klarien van der Linde, Laetitia Vanbruwaene, Josje Verhagen, Virve Vihman, Adriana Weisleder, Gillian Wigglesworth 1University of Leeds and UiT Arctic University of Norway; 2University of Leeds; 3Université de Tours; 4University of Reading and UiT Arctic University of Norway and 5Radboud University Keywords: quantifying bilingualism in children; Delphi consensus survey; researchers; speech and language therapists; teachers 1. Introduction Most investigations of bilingualism have moved away from classifying participants as bilingual without some documentation of language history and experience. There are several groups of professionals, such as researchers, teachers, and speech and language therapists (SLTs), who often have to document bilingual experience. Both within and across these groups, document- ing bilingualism might be required for a range of different purposes: assessing children’s devel- opment in each language, assessing the risk of a language disorder, assessing their learning potential, assessing their cognitive development, assessing their well-being, etc. This documen- tation is usually based on language background questionnaires. When studying children, the information is mostly obtained from caregivers, sometimes teachers, or even children themselves. Cite this article: De Cat C, Kašćelan D, Prévost P, Serratrice L, Tuller L, Unsworth S, The Q-BEx Consortium (2023). How to quantify bilingual experience? Findings from a Delphi consensus survey. Bilingualism: Language and Cognition 26, 112–124. https://doi.org/10.1017/S1366728922000359 In a recent review of 48 questionnaires used to quantify bilingualism in children, Kašćelan, Prévost, Serratrice, Tuller, Unsworth and De Cat (2021) observed substantial variation in the documentation of key dimensions of bilingualism, such as language exposure and use, activ- ities in each language, and language skills. For instance, across these questionnaires, exposure and use were documented with particular interlocutors, in specific contexts, during various activities, or as a combination of these. The level of detail varied greatly depending on the interlocutors, contexts, and activities specified in the questionnaire, and depending on whether informants could add categories (e.g., interlocutors) relevant to their circumstances. Language exposure and language use were usually treated as separate constructs but they were not always documented separately, depending on the wording of the relevant questions (which asked, for example, about the frequency of “speaking or hearing” each language). Frequency was usually documented on a Likert scale, but questionnaires used scales of different lengths and with dif- ferent labels (e.g., quantifying adverbs, percentage-based points, a combination of frequency adverbs and percentages). Some questionnaires estimated frequencies through open-ended questions. Several questionnaires documented the time that the child spent with each inter- locutor, so that the frequency of language exposure from that interlocutor could be adjusted for the proportion of the time spent with the child. Bilingualism: Language and Cognition 26, 112–124. https://doi.org/10.1017/S1366728922000359 Received: 29 January 2021 Revised: 5 May 2022 Accepted: 5 May 2022 First published online: 4 July 2022 Address for correspondence: Address for correspondence: Cécile De Cat, University of Leeds, Michael Sadler Building, Woodhouse, Leeds LS2 9JT, United Kingdom; c.decat@leeds.ac.uk Cécile De Cat, University of Leeds, Michael Sadler Building, Woodhouse, Leeds LS2 9JT, United Kingdom; c.decat@leeds.ac.uk © The Author(s), 2022. Published by Cambridge University Press. This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (http://creativecommons.org/ licenses/by/4.0/), which permits unrestricted re-use, distribution and reproduction, provided the original article is properly cited. p p p This considerable variability between questionnaires constitutes one of the major hurdles in making the resulting datasets truly comparable.1 In line with Marian and Hayakawa (2020), we argue in Kašćelan et al. (2021) that bilingualism research is in need of a consensus on what aspects of bilingual experience to document and how. To achieve such a consensus, an 2. Methodology 2. Methodology Our aim was to identify the broad consensus on what needs to be documented rather than ask panellists to prioritise which aspects to document (as this would have depended on their particular area of specialism). To maximise the comparability of measures used across studies and practice, we aim to develop a customisable questionnaire allowing professionals to select what is relevant to their purpose from a large set of consensus-informed question- naire components. The paper is organised as follows. First, we briefly explain the Delphi method and review relevant literature. Second, we outline the methodology adopted in our study, and present the results obtained as well as the list of statements reaching consensus among the three stakeholder groups. Finally, we discuss the find- ings and outline future steps towards the creation of a common set of measures to quantify bilingual experience in children. Ethical approval for the study was granted by the University of Leeds. Our procedure was as follows. To generate the first set of statements for the online survey, we organised a workshop with a group of experts. Subsequently, we conducted a two-round online survey in which a larger group of panellists rated the state- ments on a 5-point scale, with a possibility to leave comments. It is a common concern in the Delphi approach that repeated itera- tions can lead to increased panellist attrition (Walker & Selfe, 1996). Therefore, the end of round 2 of the online survey was set as a stopping criterion, no matter the number of statements that reached consensus. This gave panellists the opportunity to change their mind once, and only for those statements in the proximal zone. https://doi.org/10.1017/S1366728922000359 Published online by Cambridge University Press https://doi.org/10.1017/S1366728922000359 Published online by Cambridge University Press 113 Bilingualism: Language and Cognition Pipon-Young (2009), and Thangaratinam and Redman (2005), as well as the review by Diamond, Grant, Friedman, Pencharz, Ling Moore and Wales (2014). We modelled parts of our approach on Bishop, Snowling, Thompson, Greenhalgh and the CATALISE consortium (2016), as well as on Langlands, Jorm, Kelly and Kitchener (2008), and Spain and Happé (2019). Pipon-Young (2009), and Thangaratinam and Redman (2005), as well as the review by Diamond, Grant, Friedman, Pencharz, Ling Moore and Wales (2014). We modelled parts of our approach on Bishop, Snowling, Thompson, Greenhalgh and the CATALISE consortium (2016), as well as on Langlands, Jorm, Kelly and Kitchener (2008), and Spain and Happé (2019). inclusive and collaborative approach involving both researchers and practitioners (teachers and SLTs) is required. Such an approach can maximise the potential for cross-pollination between research and practice, and enable a construction of com- mon tools to gather the relevant information about children’s bilingualism. g We conducted an international Delphi consensus survey, with the aim of informing the creation of a modular tool for quantify- ing bilingual experience and achieve consensus between different groups (researchers and practitioners). In doing so, we contribute to the literature by “taking the pulse” of current research and practice, gathering experts’ opinions on how bilingualism should be quantified. The survey was completed by 132 researchers, tea- chers and SLTs from 29 different countries who had worked with bilingual children of various ages and from different bi/multilin- gual contexts. 1.1 The Delphi method The Delphi approach is commonly used across disciplines to explore the diversity of opinions among a group of stakeholders or lead them towards a consensus (Iqbal & Pipon-Young, 2009). The stakeholders (also known as panellists) are a represen- tative group of experts on the relevant topic of interest. The tech- nique involves an iterative process, in which the panellists express their opinions on a series of statements through two or more sur- vey rounds. The initial set of statements is sometimes itself gener- ated with an open-ended survey (Thangaratinam & Redman, 2005). p In the next two sections, we present the panellist characteristics (section 2.1) and the survey design and procedure (section 2.2). How consensus was defined is explained in the results (section 3.1). https://doi.org/10.1017/S1366728922000359 Published online by Cambridge University Press 2.1 Panellists/Stakeholders ) The initial set of statements is presented to the panellists via an online survey, in which they are asked to rate their level of agree- ment with each statement and (optionally) provide comments to justify their ratings. The following round includes a reduced set of statements, depending on the ratings and comments of the previ- ous round. From one round to the next, the statements which have not yet reached agreement or which are in a predefined proximal zone (i.e., grey area) are retained. If required, they can be reformulated. New statements can also be added if necessary. In the second (and subsequent) round(s), each statement is presented to panellists along with the distribution of responses from the preceding round as well as the panellist’s own ratings. This allows them to reflect on their responses in light of group views, and give them the opportunity to maintain or change their rating as they see fit. This process can be repeated until the consensus is reached, until the predetermined number of rounds is completed, or until it is clear that greater agreement is not possible. The initial workshop (which will be described in the next section) brought together researchers (n = 22) and practitioners (n = 14) who have worked with bilingual children and extensively used (and often designed) questionnaires to document bilingual experi- ence. We tried to be as inclusive and representative as possible. The workshop participants included experts in typical as well as atypical language development, and had different lengths of experience in bilingualism research. Their expertise spanned dif- ferent types of bilingual populations, including speakers of heri- tage languages with various levels of societal prestige, speakers of two majority languages with equal or unequal societal status, and bidialectal speakers. The researchers came from 11 countries: the UK (8), Germany (3), Canada (2), the Netherlands (2), the US (2), France (1), India (1), Israel (1), Poland (1), South Africa (1), and Sweden (1). The practitioners were mostly local (12 from the UK, 1 from Lebanon, and 1 from France), and they ranged from early career to more experienced. The initial set of statements is presented to the panellists via an online survey, in which they are asked to rate their level of agree- ment with each statement and (optionally) provide comments to justify their ratings. 2.2 Survey design and procedure A Delphi consensus survey requires three types of contributors: panellists, moderators and an independent administrator. The panellists included the workshop participants who had informed the initial generation of statements, as well as those recruited through expression of interest forms. The Q-BEx team members (i.e., the main authors of this paper) acted as panellists but also as moderators: they organised and participated in the initial work- shop, designed and administered the online survey, and analysed the data. To guarantee anonymity of participation, an independ- ent administrator handled the correspondence with panellists at round 2 (as this included individualised reports). Anonymisation of the data ensured the moderators could not attribute any rating or comment to a particular panellist. Table 1 summarises the distribution of round 1 panellists across self-assigned categories (allowing identification with mul- tiple categories). They came from 30 countries: the UK (52), the US (17), the Netherlands (14), Canada (9), India (7), Spain (7), Finland (6), France (6), Germany (6), South Africa (6), Norway (4), the United Arab Emirates (4), Cyprus (3), Israel (3), Ireland (2), Italy (2), Lithuania (2), Reunion Island (2)2, Australia (1), Egypt (1), Estonia (1), Greece (1), Japan (1), Lebanon (1), Malta (1), Mexico (1), Panama (1), Poland (1), Singapore (1), and Sweden (1). We outline the survey design and procedure below. The full protocol we followed is schematised in Figure 1. An important aim of the design was bias limitation. This was implemented by inviting a group of experts to inform the generation of the initial statements, and by adopting pre-defined procedures for moder- ation and analysis (as will be explained in the analytic strategy section below). The choice of what topics to include in the Delphi survey was informed by an in-depth review of existing questionnaires (Kašćelan et al., 2021) and by current research and practice (via the workshop). In round 2, we invited the 164 panellists who had completed round 1. Of these, 132 completed round 2 (response rate: 80% of the round 2 panellist set, or 67% of the round 1 panellist set). This compares favourably with other Delphi consensus sur- veys: in a review of 100 Delphi studies, Diamond et al. (2014) reported that only five studies had ≥100 panellists in the final round, while the rest of the studies either had fewer participants (n = 90 studies) or the number was not specified (n = 5 studies). 2.1 Panellists/Stakeholders Categories Sub-categories Distribution by sub-category, Number (%) Distribution by category, Number (%) Researchers 68 (41%) 68 (41%) Practitioners Speech and language therapist/ pathologist/ logopedist 38 (23%) 58 (35%) Teacher 20 (12%) Speech and language therapist/ pathologist/ logopedist, Teacher 0 Researchers/ Practitioners Researcher, Speech and language therapist/ pathologist/ logopedist 24 (15%) 38 (23%) Researcher, Teacher 12 (7%) Researcher, Speech and language therapist/ pathologist/ logopedist, Teacher 2 (1%) TOTAL 164 (6), Spain (6), France (5), India (5), Cyprus (3), Israel (3), South Africa (3), Ireland (2), Norway (2), Reunion Island (2), United Arab Emirates (2), Australia (1), Egypt (1), Estonia (1), Italy (1), Japan (1), Lebanon (1), Lithuania (1), Malta (1), Mexico (1), Panama (1), Poland (1), Singapore (1), and Sweden (1). the Literacy Association of South Africa, the National Association for Language Development in the Curriculum in the UK). The expression of interest forms contained questions about the prospective panellists’ demographic background and about their experience working with bilingual children. Using these forms, registrations for the survey were submitted by 82 researchers, 61 SLTs, and 27 teachers. We applied the following exclusion criteria: having less than a year of work experience, having never worked with bilinguals, not being able to commit to both rounds of the Delphi survey. This resulted in 13 exclusions (1 researcher, 10 SLTs and 2 teachers), after which 157 stakeholders were retained. In addition, 22 researchers and 14 practitioners from the work- shop, as well as 3 researchers who were invited to the workshop but could not attend, were invited to participate in the first round of the Delphi survey. Of the 196 invited panellists, 164 completed round 1 of the survey (response rate: 83%). 2.2 Survey design and procedure Table 2 summarises the distribution of round 2 panellists across self-assigned categories (allowing identification with multiple cat- egories). Note that in this round, 10 panellists did not select iden- tical stakeholder labels to describe themselves as in round 1. Round 2 panellists came from 29 countries: the UK (43), the US (15), the Netherlands (9), Canada (7), Finland (6), Germany 2Although Reunion Island is an overseas department and region of the French repub- lic, we counted it separately due to the geographical distance and potentially diverse experiences of the stakeholders in comparison to continental France. 2.1 Panellists/Stakeholders The following round includes a reduced set of statements, depending on the ratings and comments of the previ- ous round. From one round to the next, the statements which have not yet reached agreement or which are in a predefined proximal zone (i.e., grey area) are retained. If required, they can be reformulated. New statements can also be added if necessary. In the second (and subsequent) round(s), each statement is presented to panellists along with the distribution of responses from the preceding round as well as the panellist’s own ratings. This allows them to reflect on their responses in light of group views, and give them the opportunity to maintain or change their rating as they see fit. This process can be repeated until the consensus is reached, until the predetermined number of rounds is completed, or until it is clear that greater agreement is not possible. y Following the workshop, we used online expression of interest forms (separate ones for researchers, teachers, and speech and language therapists) to expand and diversify the sample of panel- lists for the online Delphi survey. The expression of interest forms were advertised on our project website and on social media (Facebook, Twitter). We also emailed 247 individual researchers, practitioners, and relevant organisations (e.g., Comité Permanent De Liaison Des Orthophonistes-Logopèdes de l’UE, Audiology and Speech-Language Pathology Associations around the world, Anonymity is key. It guarantees parity among panellists by giv- ing an equal voice to all, promotes freedom of expression, and limits the risk of bias. The online format allows greater inclusivity, as panellists can be recruited from different geographic areas and complete the survey at their own convenience (within a set time limit). The design of the present study was informed by guidelines from Hasson, Keeney and McKenna (2000), Iqbal and Cécile De Cat et al. 114 Table 1. Round 1 panellist breakdown per stakeholder group. Table 1. Round 1 panellist breakdown per stakeholder group. https://doi.org/10.1017/S1366728922000359 Published online by Cambridge University Press 3Erika Hoff was supported by the Eunice Kennedy Shriver National Institute of Child Health and Human Development Grant HD068421. 4We generally avoided negative statements, to prevent the complication of double negatives, which are hard to interpret (“I strongly disagree that this is not the case.”). The only exceptions were statements s.5, s.13, s.35, s.45 and s.47. 5A complete list of round 1 and round 2 statements, panellists’ response distribution and comments are available via the Open Science Framework: https://osf.io/2pd65/ Delphi survey round 1 b. Using experience data to inform the assessment of risk of atypical development (Sharon Armon-Lotem) Delphi survey round 1 The first round of the online survey was administered in April and May 2020. Panellists were given five weeks to complete it (including a one-week extension). In addition to a personalised link to the survey, each panellist was emailed a Briefing Document (see supplement 1) explaining the aim of the study, and a Glossary of the technical terms appearing in the survey (see supplement 2). The panellists were asked to score the state- ments on a 5-point scale (1 = strongly disagree, 2 = disagree, 3 = I don’t know, 4 = agree, 5 = strongly agree). Two statements were different: they asked panellists to indicate the optimal amount of time needed to complete a short and a long version of a bilingual experience questionnaire. Options for a short ver- sion included 5 minutes, 10 minutes, and 15 minutes, while options for the long version included 20 minutes, 30 minutes, 40 minutes, 50 minutes, and 60 minutes. There was a space for optional open-ended comments following each statement. c. Measuring family socioeconomic status in studies of bilingual development (Erika Hoff)3 d. Language mixing (Elma Blom) e. Input quality in relation to input quantity (Johanne Paradis) Emerging themes were identified and discussed, leading to the generation of statements (individually and in small groups). Participants were then invited to contribute what they considered uncontroversial as well as controversial statements. With the practitioners, the discussions were informed by a crit- ical review of current practice and reflections on practitioners’ needs. This also led to the generation of statements. At the end of the workshop, a combined list of 197 statements had been compiled.4 Following the workshop, the moderators excluded duplicates and unclear statements. Similar statements were merged, and some were reformulated for clarity, based on our notes from the workshop. This resulted in 53 statements, some of which con- sisted of several parts (see example (2)). To assess agreement with each part more precisely, every part of an overarching state- ment was assessed as a separate statement. Therefore, in what fol- lows, we will refer to each part (i.e., sub-component of the original 53 overarching statements) as a separate statement. In this way, altogether, 112 statements5 were included in round 1 of the Delphi survey. https://doi.org/10.1017/S1366728922000359 Published online by Cambridge University Press Workshop The three-day workshop was organised in Leeds in January 2020. The first two days of the workshop were attended by researchers, and the third one by practitioners. One practitioner and five researchers attended all three days. To inform the generation of statements for the online survey, the workshop was organised around thematic presentations by leading experts, a review of the state-of-the-art in bilingual experience questionnaires, issues raised by the participants, and guidance on the principles of ques- tionnaire creation and validation by an expert in psychometrics (Kate Harvey). Throughout the workshop, participants were https://doi.org/10.1017/S1366728922000359 Published online by Cambridge University Press Bilingualism: Language and Cognition Bilingualism: Language and Cognition 115 Table 2. Round 2 panellist breakdown per stakeholder group. Table 2. Round 2 panellist breakdown per stakeholder group. Categories Sub-categories Distribution by sub-category, Number (%) Distribution by category, Number (%) Researchers Researcher 57 (43%) 57 (43%) Practitioners Speech and language therapist/ pathologist/ logopedist 26 (20%) 40 (30%) Teacher 13 (10%) Speech and language therapist/ pathologist/ logopedist, Teacher 1 (1%) Researchers/ Practitioners Researcher, Speech and language therapist/ pathologist/ logopedist 24 (18%) 35 (27%) Researcher, Teacher 9 (7%) Researcher, Speech and language therapist/ pathologist/ logopedist, Teacher 2 (2%) TOTAL 132 invited to contribute their views verbally, in writing (using post- its and interactive virtual whiteboards), and through live polling (using Mentimeter). The thematic presentations focused on the themes in (1), and led to group discussions. The survey was piloted by a group including the six moderators (i.e., the main authors of this paper), as well as six additional researchers and one SLT (the latter seven did not participate in the online study). The aim of the pilot was to check for any errors, assess the clarity of the statements, and optimise their order of presentation. (1) Thematic presentations a. Capturing linguistic diversity (Ianthi Tsimpli) b d f h 3.1 Analytic strategy and between-rounds moderation process In their guideline paper on the Delphi methodology, Hasson et al. (2000) reported that values between 51% and 80% of agreement have been used as cut-off points for consensus in the literature. In order to determine the consensus cut-off point for our survey, we followed a review by Diamond et al. (2014). This review was based on 100 Delphi studies randomly selected from various dis- ciplines that had been published between 2000 and 2009. Among those using a percentage or a proportion to define consensus, the median consensus threshold was 75%. Consequently, we applied the same criterion in our survey. Figure 2 presents the distribution of the statements by agree- ment rating in round 1. It shows that the pre-defined thresholds for consensus (75% agreement) or proximal zone (60% to 74% agreement) do not correspond to discrete breaks in the agreement distribution. The distribution is skewed (towards high agreement levels) but continuous. (Dis)agreement was defined as follows. Whenever a panellist selected ‘4 = agree’ or ‘5 = strongly agree’, we marked this as agree- ment with the statement. The round 1 statements reaching an agreement rate ≥75% were considered as having reached consen- sus and were excluded from round 2 (n = 74 statements), that is, they were immediately included in the final set of agreement- reaching statements (which we analyse in section 3.6). In addition, as an adaptation from Langlands et al. (2008) and Spain and Happé (2019), the round 1 items with an agreement score between 60% and 74%7 were re-rated in round 2. The purpose of this approach was to reconsider what we refer to as the ‘proximal zone statements’ (i.e., those statements which were not that far from the designated consensus threshold) rather than dismiss them straight away. In round 2, the panellists were invited to reconsider their own scores for these statements, in light of the distribution of round 1 average scores and relevant comments: they could either confirm or modify their score based on this information. Twenty statements met the above ‘proximal zone’ criterion for inclusion in round 2. Finally, we conducted a thematic analysis of the comments gathered in round 1. This qualitative analysis technique aims to identify the themes that emerge in a dataset inductively. 8One of these eight statements (statement 39.2) was excluded from round 2 as its related and preceding statement (i.e., statement 39.1) reached a very low overall agree- ment (32%). Consequently, it did not make sense to include statement 39.2 in round 2 for reconsideration. 6To illustrate our procedure, we adapted a flowchart Figure 1 from Bishop et al.’s (2016) CATALISE Delphi consensus study. Delphi survey round 2 This round was administered in June and July 2020 and the panel- lists were given six weeks to complete it (including a one-week extension). Apart from a link to the survey, each panellist was emailed a personalised report of round 1, containing the distribu- tion of responses for each statement, as well as their own scores for each statement (see supplement 3). In addition, the panellists were emailed a list of round 1 comments (see supplement 4). Finally, the panellists were sent Clarifications and Instructions for completing round 2 of the survey (see supplement 5). In round 2, the panellists were asked to re-rate 27 of the round 1 statements, and to rate 10 reformulated statements (each one presented immediately following their original formulation), and four new statements. As explained below, we only reformu- lated statements at round 2 if there was reason to suspect that the lack of consensus at round 1 was due to lack of clarity (deter- mined by participant comments). https://doi.org/10.1017/S1366728922000359 Published online by Cambridge University Press 116 Cécile De Cat et al. Figure 1. Procedure for the Delphi consensus survey, outlining the role of panellists and moderators6 116 Cécile De Cat et al. ey, sus survey, ors6 Figure 1. Procedure for the Delphi consensus survey, outlining the role of panellists and moderators6 3. Results perspectives, we decided to give stronger weight to their round 1 ratings in the following way: statements reaching proximal zone agreement levels (60%-74%) in the dual interest group were auto- matically included in round 2, even if they were below the 60% threshold in the panel as a whole. Eight statements were thus iden- tified for inclusion to the proximal zone pool (in addition to the 20 mentioned above).8 Statements reaching an overall agreement rate lower than 60% in round 1 were excluded from round 2 (n = 8). 7The upper limit was 79% in Langlands et al. (2008) and in Spain and Happé (2019). https://doi.org/10.1017/S1366728922000359 Published online by Cambridge University Press 3.1 Analytic strategy and between-rounds moderation process In this approach, a theme is to be understood as a concept central to a mind map of related topics (i.e., a theme is underpinned by a core organising concept, related to other concepts – Braun & Clarke, 2019; Clarke & Braun, 2017). A theme can therefore cap- ture a diversity of inter-connected meanings. The importance of a theme is determined by this web of relationships rather than just frequency of occurrence in the data. The flexibility of this tech- nique allowed us not only to focus on the semantic content of the data, but also to consider the latent levels (i.e., going beyond the semantic content). Consequently, the need for new statements and reformulations could be identified even if they were not expli- citly requested. For a further discussion of this approach, see Braun, Clarke and Hayfield (2019). In round 1, there were 2,486 comments. One of the moderators selected a random subset of 11 statements (1.1, 1.2, 10.1, 10.2, 11, 12, 13, 14, 15, 16.1, 16.2) which in total contained 430 comments A subset of panellists identified themselves as both a researcher and a practitioner (n = 38). We refer to these panellists as the ‘dual interest’ group. Since these panellists had insights from both https://doi.org/10.1017/S1366728922000359 Published online by Cambridge University Press Bilingualism: Language and Cognition 117 Figure 2. Distribution of round 1 statements by agreement rating (defined as the pro- portion of participants choosing “agree” or “strongly agree”). The horizontal lines indicate thresholds (for consensus: upper line, and for proximal zone: lower line). After both rounds, consensus was reached for 79% of state- ments (98/124). As seen in Figure 3, the distribution of statements varied across agreement bands (with agreement defined as the proportion of panellists expressing agreement or strong agree- ment for a particular statement). The three highest agreement bands included approximately 85% of statements. By contrast, the 60-70% agreement band included only 6% of statements, and 8% of statements received agreement below 60%. None received less than 20% agreement. g Round 2 included 27 statements in their original formulation, 10 reformulations, and 4 new statements. Most of the original statements (93%, i.e., 25/27) were rated higher in round 2 than in round 1. However, this was not always sufficient for consensus to be reached (requiring a rating as “agree” or “strongly agree” by at least 75% of round 2 panellists). versions of the questionnaire. These 2 statements are excluded from the analyses below. For the distribution of responses on these 2 statements, see supplement 7. 3.3 Attrition analysis Out of the 164 panellists in round 1, 32 did not respond to round 2. To assess the risk of bias at round 2 as a result of panellist attri- tion, we reanalysed the data from round 1 by including only responses from panellists who had taken part in both rounds. The results from this subset of panellists differed from the results from the whole round 1 panel in three respects. 3.1 Analytic strategy and between-rounds moderation process • 36.2 [The literacy practices of parents] need to be documented independently of parental education and socioeconomic status. In total, round 2 included 27 statementsto be re-rated from round 1(i.e.,proximalzonestatementsbasedontheoverallorthedualinter- est group ratings), 10 reformulations of the round 1 statements, and four new statements. Round 2 statements reaching an agreement rate ≥75% were considered to have reached a consensus. 3.1 Analytic strategy and between-rounds moderation process Only 52% of the original state- ments from round 1 (i.e., 14/27) reached consensus in round 2. From the reformulated statements, 7/10 yielded a higher agree- ment rating in round 2 than the round 1 originals, while 6/10 reached consensus. Two of these six reformulated statements (statements 31.1 and 49.1) reached consensus both in their round 1 formulation (proximal zone statements) and in their round 2 reformulation. All the new round 2 statements (n = 4) reached consensus, bringing the overall round 2 consensus to 59% (24/41). Figure 2. Distribution of round 1 statements by agreement rating (defined as the pro- portion of participants choosing “agree” or “strongly agree”). The horizontal lines indicate thresholds (for consensus: upper line, and for proximal zone: lower line). (approx. 17% of the data). Based on the reading of these com- ments, 19 themes were identified and defined inductively. Two other moderators checked the thematic classification of 126 of these comments for consistency, and the list of themes was refined accordingly. The final list of themes and their definitions (pro- vided in supplement 6) were used by all six moderators to classify the remainder of the comments. Each comment could be assigned up to three themes. The comments were then filtered by theme and analysed to identify opinions not yet represented by the sur- vey statements. This resulted in 10 reformulations of the round 1 statements, as well as 4 new statements. Note that all 10 reformu- lations were rewordings of round 1 proximal zone statements. Consequently, in round 2, the panellists had to re-rate the original round 1 proximal zone statement as well as its reformulation. Figure 4 shows the difference in agreement ratings of proximal zone statements between the rounds. The statements are pre- sented in the ascending order of the difference in ratings. From the 14 proximal zone statements that reached consensus following round 2, the following three had a jump higher than 15% between the rounds: • 42. There should be a question on attitudes to language mixing (b) within the local community (including school). • 42. There should be a question on attitudes to language mixing (b) within the local community (including school). • 5. The questionnaire should not aim to measure the child’s lan- guage proficiency. This should be done by other means. 9Of these, 124 statements were rated on a 5-point agreement scale (strongly disagree, disagree, I don’t know, agree, strongly agree). The remaining 2 statements were rated on a time-length scale, as they inquired about preferred time lengths of the short and the long https://doi.org/10.1017/S1366728922000359 Published online by Cambridge University Press 3.2 Post-round 2 descriptive analyses Overall, there were 55 overarching statements (53 in Round 1 and an additional two in Round 2). Many of these consisted of two or more parts (or sub-components), as in (2). First, out of the 74 statements that reached consensus at round 1 in our original analysis, four of them (22, 25, 27e, and 51.1.a) remained marginally below the consensus threshold in the subset analysis (reaching 71.21%, 73.48%, 74.24%, and 72.72% agree- ment respectively) when the panellists who responded in round 1 only were excluded. While there is no way of knowing whether re-rating these 4 statements in round 2 would pass the 75% con- sensus threshold, these high levels of agreement (71%-74%) sug- gest that it is likely that they would. Indeed, as shown in Figure 4, (2) The language(s) used at school should be documented as: a. language(s) used by teachers; b. language(s) used by the child; c. language(s) used by playmates. For the purposes of our analysis, each part (i.e., sub- component) was counted separately, leading to a total of 126 statements (across the two rounds).9 9Of these, 124 statements were rated on a 5-point agreement scale (strongly disagree, disagree, I don’t know, agree, strongly agree). The remaining 2 statements were rated on a time-length scale, as they inquired about preferred time lengths of the short and the long https://doi.org/10.1017/S1366728922000359 Published online by Cambridge University Press 118 Cécile De Cat et al. Figure 3. Percentage of statements per agreement band following the two rounds of the Delphi survey Figure 4. Difference in agreement ratings of proximal zone statements between the rounds (ordered ascendingly by difference size). The dotted horizontal line indicates the consensus threshold. Figure 3. Percentage of statements per agreement band following the two rounds of the Delphi survey Figure 4. Difference in agreement ratings of proximal zone statements between the rounds (ordered ascendingly by difference size). The dotted horizontal line indicates the consensus threshold. Figure 4. Difference in agreement ratings of proximal zone statements between the rounds (ordered ascendingly by difference size). The dotted horizontal line indicates the consensus threshold. Figure 4. Difference in agreement ratings of proximal zone statements between the rounds (ordered ascendingly by difference size). The dotted horizontal line i di h h h ld atings of proximal zone statements between the rounds (ordered ascendingly by difference size). 3.5 Polarisation analysis To ascertain whether there was any polarisation of opinion in the statements that reached consensus, we inspected the distribution of responses for those statements. The data can be seen in Figure S3 (supplement 8), in which statements are ordered according to the level of disagreement (with the would-be contro- versial statements appearing in the top-most part of the plot). None of the consensus-reaching statements was rated as “strongly disagree” by more than 6% of panellists. We conclude that there 3.2 Post-round 2 descriptive analyses The dotted horizontal line most statements that were re-rated in round 2, increased in the agreement rate by 5% or more. reached consensus at round 1 even if only researchers were included. For 92% of statements, there was less than a 5% differ- ence in agreement level between researchers and the panel as a whole. The average difference was 3.4%, and the maximum 10.6%. The direction of the difference was almost evenly balanced (with 53% of statements achieving a higher level of agreement among researchers compared to the panel as a whole, and 47% thus showing a lower level of agreement among the researchers). Differences greater than 5% were always towards greater agree- ment among researchers than the group as a whole. Second, out of the 28 proximal zone statements identified above, 2 statements (49.1 and 7) already reached consensus at round 1 in the subset analysis (reaching 77.27% and 75% agree- ment respectively). This is of no consequence, as the re-rating of these statements at round 2 allowed them to reach 86.36% and 82.57% agreement respectively. Finally, statement 31.2 reached the proximal zone in the original analysis, but not in the subset analysis. Again, this is of no consequence as the re-rating of this statement at round 2 did not reach consensus, even in its reformulation (62.87% and 68.18% agreement respect- ively). Therefore, we conclude that attrition did not increase the risk of bias at round 2. https://doi.org/10.1017/S1366728922000359 Published online by Cambridge University Press 3.4 Subgroup analysis To assess the extent to which the views of researchers aligned with those of the panel as a whole, we analysed the level of consensus among researchers at round 1 (including those from the dual interest group). This revealed that the exact same set of statements https://doi.org/10.1017/S1366728922000359 Published online by Cambridge University Press https://doi.org/10.1017/S1366728922000359 Published online by Cambridge University Press 119 Bilingualism: Language and Cognition Figure 5. Consensus by stage of the Delphi survey Figure 5. Consensus by stage of the Delphi survey 2. For ease of reference, the statement numbers below are as per those in the online survey. A short commentary is provided at the end of some sections as a reflection on the consensus. The complete datasets from both rounds of the online survey can be found via the Open Science Framework: https://osf.io/ 2pd65/ was no substantial polarisation with respect to the consensus- reaching statements. 3.6 What is the consensus? In this section, we present the 98 statements that reached consen- sus following both rounds of the Delphi survey (i.e., with an agreement rate ≥75%). We list them in sections grouped around specific topics, along with the level of agreement reached (expressed as the proportion of panellists who agreed or strongly agreed). We also indicate if the statement was the original round 1 statement, a reformulation, or a new statement added in round Mandate for a new tool 1.1 There needs to be a set of common measures of children’s bilingual language experience, to allow comparability across studies and to facilitate communication across sectors (research, education, therapy). [96%, original] https://doi.org/10.1017/S1366728922000359 Published online by Cambridge University Press https://doi.org/10.1017/S1366728922000359 Published online by Cambridge University Press 120 Cécile De Cat et al. 9.2 Frequencies of use should be documented separately for each type of interlocutor. [84%, original] 10.1 The languages used outside school should be documented as: (a) languages used with parents [99%, original] (b) languages used with siblings [99%, original] (c) languages used with other carers in the home [96%, original] (d) languages used with friends of the family [80%, original] (e) languages used amongst the child’s friendship groups [93%, original] (f) languages used within the neighbourhood [78%, original] 10.2 Frequencies of use should be documented separately for each type of interlocutor. [84%, original] 11. The amount of overheard speech (between parents) needs to be estimated. [76%, original] 12. The child’s digital language exposure and use needs to be mea- sured (e.g., Internet, social media, gaming). [90%, original] 14. Changes in the child’s language exposure over time should be documented. [95%, original] 16.1 The child’s first exposure to each language should be docu- mented, in terms of: 9.2 Frequencies of use should be documented separately for each type of interlocutor. [84%, original] 10.1 The languages used outside school should be documented as: (a) languages used with parents [99%, original] (b) languages used with siblings [99%, original] (c) languages used with other carers in the home [96%, original] (d) languages used with friends of the family [80%, original] (e) languages used amongst the child’s friendship groups [93%, original] (f) languages used within the neighbourhood [78%, original] 10.2 Frequencies of use should be documented separately for each type of interlocutor. [84%, original] 11. The amount of overheard speech (between parents) needs to be estimated. [76%, original] 12. The child’s digital language exposure and use needs to be mea- sured (e.g., Internet, social media, gaming). [90%, original] 14. Changes in the child’s language exposure over time should be documented. [95%, original] 16.1 The child’s first exposure to each language should be docu- mented, in terms of: (a) age [97%, original] 1.2 These measures should be applicable to children who speak more than two languages. [90%, original] 1.2 These measures should be applicable to children who speak more than two languages. [90%, original] 2. Language difficulties 4.1 The questionnaire should contain a section to identify chil- dren who might be at risk for a Developmental Language Disorder. [77%, original] 4.2 This should include: (a) early language milestones [84%, original] (b) hearing difficulties [80%, original] (d) issues related to trauma, attachment, prematurity [79%, original] (e) family history of learning difficulties or speech/language delays [92%, new] (a) age [97%, original] (b) context (e.g., childminder, pre-school, etc.) [95%, original] (a) age [97%, original] (b) context (e.g., childminder, pre-school, etc.) [95%, original] 16.2 A list of potential contexts should be provided. [89%, original] 17.1 Time spent in school should be quantified. [89%, original] 17.2 This should be done in: 7. The questionnaire should ask about difficulties the child may have (had) with language, in order to identify what might require further assessment by specialists. [83%, original] 7. The questionnaire should ask about difficulties the child may have (had) with language, in order to identify what might require further assessment by specialists. [83%, original] (a) hours per week [78%, original] (a) hours per week [78%, original] 55. If the questionnaire includes a section documenting language and/or developmental difficulties, this should under no circum- stances be used as a diagnostic tool. [87%, new] In addition to these statements, it is important to highlight statement 13, which did not reach consensus: the precise measur- ing of what happens during holidays is unnecessary. It is enough to document whether or not the child travels to the “home” coun- try or has ties with “manifestations of the home country” (e.g., regular contacts online in the home language with dispersed fam- ily members). As the statement had a low agreement rate (47%), the documentation of exposure- and use-related practices during holidays might need to be considered. https://doi.org/10.1017/S1366728922000359 Published online by Cambridge University Press Mandate for a new tool The questionnaire should be accompanied by a tool yielding automatic calculation of objective scores (in each language) of: (a) current language exposure [95%, original] (b) current language use [96%, original] (c) cumulative language exposure [91%, original] (d) cumulative language use [91%, original] 3. The tool should provide clear guidance about how to interpret the data it produces (where possible). [93%, original] 53. Questions should be as concrete as possible (e.g., asking about daily routines rather than asking the respondent to estimate frequencies in percentages). [87%, original] 3. The tool should provide clear guidance about how to interpret the data it produces (where possible). [93%, original] 53. Questions should be as concrete as possible (e.g., asking about daily routines rather than asking the respondent to estimate frequencies in percentages). [87%, original] 10.2 Frequencies of use should be documented separately for each type of interlocutor. [84%, original] 11. The amount of overheard speech (between parents) needs to be estimated. [76%, original] 12. The child’s digital language exposure and use needs to be mea- sured (e.g., Internet, social media, gaming). [90%, original] 14. Changes in the child’s language exposure over time should be documented. [95%, original] 16.1 The child’s first exposure to each language should be docu- mented, in terms of: 10.2 Frequencies of use should be documented separately for each type of interlocutor. [84%, original] 11. The amount of overheard speech (between parents) needs to be estimated. [76%, original] 12. The child’s digital language exposure and use needs to be mea- sured (e.g., Internet, social media, gaming). [90%, original] 14. Changes in the child’s language exposure over time should be documented. [95%, original] 16.1 The child’s first exposure to each language should be docu- mented, in terms of: Child’s proficiency p y 5. The questionnaire should not aim to measure the child’s language proficiency. This should be done by other means. [75%, original] 6. The child’s proficiency should be documented for the language (s) that cannot be tested directly. [81%, original] 5. The questionnaire should not aim to measure the child’s language proficiency. This should be done by other means. [75%, original] 6. The child’s proficiency should be documented for the language (s) that cannot be tested directly. [81%, original] The consensus was that questionnaires should document a child’s proficiency only when the languages in question cannot be tested directly. Further support for this interpretation comes from the lack of consensus on the reformulation of statement 5 (The questionnaire should not aim to document the child’s lan- guage proficiency [in any of the languages]. This should be done by other means) and of statement 6 (The child’s proficiency should be documented for all their languages). 18. Any prolonged period in which the child did not attend for- mal education should be documented. [91%, original] g 19. The questionnaire should ask if the child attended school in another country. [92%, original] y g 21. The child’s frequency of reading in each language should be measured. [86%, original] Exposure and use 8. Exposure and use should be measured (for each language): (a) over an average week [85%, original] (c) over holiday and school periods separately [80%, original] (d) over home and school separately [92%, original] Exposure and use 8. Exposure and use should be measured (for each language): g 22. Child’s frequency of writing in each language should be mea- sured. [75%, original] g 23. The questionnaire should document any home-language classes that children are attending: (a) over an average week [85%, original] (c) over holiday and school periods separately [80%, original] (d) over home and school separately [92%, original] (a) in school [94%, original] (b) outside school [96%, original] (a) in school [94%, original] (b) outside school [96%, original] (d) over home and school separately 9.1 The language(s) used at school should be documented as: 36.1 Any literacy activities that the parents engage in with the child should be documented. [86%, reformulation] (a) language(s) used by teachers [96%, original] (b) language(s) used by the child [95%, original] (c) language(s) used by playmates [91%, original] 36.2 This needs to be done independently of parental education and socioeconomic status. Child’s proficiency The following demographic information should be collected: (a) child’s date of birth [99%, original] (b) the date of filling in the questionnaire [98%, original] (c) child’s sex [92%, original] (d) child’s birth order [85%, original] (e) child’s arrival into the country (if relevant) [98%, original] (f) period of the child’s life spent in other countries (if relevant) [98%, original] (g) name of each country in which the child has lived (if rele- vant) [90%, original] (a) each parent [90%, original] (b) any siblings [88%, original] (c) other members of the households [77%, original] (d) playmates [77%, original] (e) teachers [75%, original] 28. The language proficiency of the child’s interlocutors should be estimated in relation to typical and representative members of the population/region in which the child lives. [76%, reformulation] 29.1 The types of activity carried out in each language should be documented (e.g., storytelling, video games, play, etc.). [81%, original] 29.2 A predefined list of activities should be provided to ensure comparability. [79%, original] 37. Both mother’s and father’s education need to be documented. [86%, original] 38. If a parent was educated in more than one language, education i h l h ld b d t d t l g 38. If a parent was educated in more than one language, education in each language should be documented separately. The statements above demonstrate a general requirement across the sectors for estimating child’s input quality. While statement 26 outlines the aspects which according to the panellists are indicative of input quality, some of the suggestions did not reach consensus. These were: interlocutor accent in each language (34%), language mixing (52%), and digital exposure (73%). 45. The questionnaire shouldnotlabellanguagesassocietal, heritage, minority or majority languages. The language names (provided by the respondents) should be used throughout the questionnaire to identify these languages where relevant. Labels can be applied post- hoc by the researcher or practitioner as required. [94%, original] 46. A range of labels for the child’s carers should be allowed by the questionnaire, to better document different family constella- tions (e.g., other than mother + father). [86%, original] Language mixing 31.1 Language mixing should be estimated (in terms of exposure and use). [77%, original] 31.1 Language mixing should be estimated (in terms of exposure and use). [77%, original] 31.1 Language mixing (heard or produced by the child) should be estimated as part of the child’s language exposure and language use. [78%, reformulation] 32. Language mixing that the child is exposed to should be docu- mented separately from the language mixing that the child pro- duces. [77%, reformulation] 33. The questionnaire should ask if the parents use one language in conversation and the child responds in the other. 31.1 Language mixing (heard or produced by the child) should be estimated as part of the child’s language exposure and language use. [78%, reformulation] In addition to the above, only 24% of the panellists agreed with statement 47 (The questionnaire should not document the immi- gration history of the child). Such an outcome could be expected as immigration history often complements or is a part of the data on the child’s language exposure and use. 33. The questionnaire should ask if the parents use one language in conversation and the child responds in the other. [92%, original] [92%, original] Questionnaire versions Child’s proficiency [79%, original] (b) language(s) used by the child (c) language(s) used by playmates [91%, original (c) language(s) used by playmates Bilingualism: Language and Cognition 121 Input quality 24. Input quality should be measured as far as this is feasible. [82%, original] 25. There needs to be agreement on a global/composite measure of input quality. [76%, original] 26. The following aspects are indicative of input quality: (a) parental education [78%, original] (b) interlocutor proficiency in each language [91%, original] (e) pre-literacy and literacy activities [82%, original] (g) playing with peers [82%, original] (h) number of interlocutors who interact with the child in each language [86%, new] third of stakeholders agreed with the related statement on mixing in highly multilingual contexts (i.e., in densely multilingual societies, language mixing need not be measured (because it is so prevalent)). 34. There should be a question probing whether the child is unwilling to speak one of his/her languages. [84%, original] 40. There should be a question on attitudes towards each of the child’s languages (a) parental education [78%, original] (b) interlocutor proficiency in each language [91%, original] (e) pre-literacy and literacy activities [82%, original] (g) playing with peers [82%, original] (h) number of interlocutors who interact with the child in each language [86%, new] (a) within the family (at home) [90%, original] (b) within the local community (including school) [86%, original] (c) within the broader society [79%, original] (a) within the family (at home) [90%, original] (b) within the local community (including school) [86%, original] (c) within the broader society [79%, original] 27. The language proficiency of the child’s interlocutors should be documented (based on the respondent’s estimate). This should include estimates for: 41.1 There should be a question about the status of each of the child’s languages 41.1 There should be a question about the status of each of the child’s languages (a) within the local community (including school) [79%, original] (b) in the “home” country (if applicable) [77%, original] 42. There should be a question on attitudes to language mixing (a) within the family (at home) [84%, original] (b) within the local community (including school) [78%, original] 43.1 Parents should be asked if they feel pressurised to speak the societal/majority language. [87%, original] Background information 41.2 The child’s languages should be identified precisely (e.g., var- iety, dialect). [87%, original] 44. https://doi.org/10.1017/S1366728922000359 Published online by Cambridge University Press 4. Discussion This Delphi consensus survey aimed to ascertain the level of agreement among researchers and practitioners regarding how bilingual experience should be documented. Our aim was to iden- tify the broadest consensus possible, in order to inform the design of a new, customisable questionnaire allowing researchers and practitioners to select the components relevant to their purpose. p p p p The round 1 statements were formulated based on the out- comes of an exploratory workshop with 22 researchers and 14 practitioners. In two rounds of the online survey, a diverse set of 132 panellists from 29 countries rated a total of 126 state- ments10 (112 original statements, four new statements, and 10 statement reformulations). Furthermore, 27 original round 1 statements were re-rated in round 2. Several measures were adopted to limit the risk of bias: we used a diversification strategy for the recruitment of panellists; the data was anonymised upon collection; we used pre-defined criteria for consensus and for selecting statements to be re-rated; and we carried out post-hoc bias analyses. While it is impossible to avoid bias entirely, we believe these measures give credence to our approach. We are confident we have captured a range of opinions reflecting a variety of centres of interests among an international community of researchers and practitioners. Another aspect eliciting diverse reactions was language profi- ciency. Round 2 reformulations revealed that this was due to a caveat: if language proficiency could not be assessed by other means, the agreement was that it should be documented via a questionnaire. Similarly, as long as the questionnaire is not seen as a diagnostic tool, the consensus was that the child’s difficulties with language should be documented. The need for a flexible or modular questionnaire elicited strong consensus. The constant tension between the level of detail aspired to and the constraints inherent to data collection will often result in having to forgo the documentation of some aspects. A modular questionnaire will allow for this. The exclusion of some aspects will also partly depend on foci of interest and the purpose of documentation (e.g., screening for language difficulties vs. informing a study on a particular aspect of bilingual experi- ence). Different versions of the questionnaire were also considered necessary to adapt to different types of respondents (e.g., care- giver, child, teacher). 10Two of these were not rated on an agreement scale and were excluded from the con- sensus calculations below. Questionnaire versions It is worth noting that in contrast to the statements on lan- guage mixing above, where consensus was reached, only about a 48. The questionnaire should be available in an online version, in a paper version, and as an interview protocol. [95%, original] https://doi.org/10.1017/S1366728922000359 Published online by Cambridge University Press https://doi.org/10.1017/S1366728922000359 Published online by Cambridge University Press 122 Cécile De Cat et al. 49.1 The questionnaire should be available in different lengths. [86%, original] 51.1 The questionnaire should be available in different versions for different respondents. This should include: (a) a version to be used with 5-7-year-old children [75%, original] (b) a version to be used with 8-12-year-old children [82%, original] (c) a version to be used with parents/carers (also usable by adolescents) [90%, original] (d) a version to be used with teachers [82%, original] fairly representative of those of the wider, more diverse panel par- ticipating in the online survey, including a solid proportion of uncontroversial views but also some more controversial ones. Questionnaire modularity 49.1 The questionnaire should contain thematic sections (e.g., on language exposure/use, on proficiency, on attitudes, etc.). Each sec- tion should be optional, and it should be up to the researchers/ practitioners to select which section to use. [87%, reformulation] 50. Some sections of the questionnaire should be optional. The researcher/practitioner should be able to exclude some sections. [86%, original] The variability observed with respect to other aspects is likely a reflection of the fact that they have hitherto been researched less systematically. The survey results indicate that researchers across the board agree these aspects are likely to be important but require more supporting evidence and/or more scrutiny. For example, input quality and language mixing are starting to attract more attention in bilingualism research (see, e.g., Unsworth, 2016 and Byers-Heinlein, 2013, respectively), and this is reflected in a set of quality-related and language mixing statements which reached consensus. Nonetheless, several statements relating to the documentation of language mixing did not pass the consensus threshold. For instance, there was no agreement about the docu- mentation of language mixing for each interlocutor in interactions with the child (statement 30); nor was consensus reached on doc- umenting language mixing through examples of different types of mixing or their frequency (statement 31.2 and its reformulation). This might reflect scepticism as to whether questionnaires can reliably document these aspects of language mixing. Further research will be necessary to elucidate these points. g 54. Some parts of the questionnaire should not be administered if asking these specific questions is not adequate (for safety, polit- ical, personal, or any other ethical or relevance reasons). ) [88%, new] [88%, new] 4.1 The consensus (a) a version to be used with 5-7-year-old children [75%, original] (b) a version to be used with 8-12-year-old children [82%, original] (c) a version to be used with parents/carers (also usable by adolescents) [90%, original] (d) a version to be used with teachers [82%, original] (a) a version to be used with 5-7-year-old children [75%, original] (b) a version to be used with 8-12-year-old children [82%, original] (c) a version to be used with parents/carers (also usable by adolescents) [90%, original] (d) a version to be used with teachers [82%, original] There was an almost unanimous call for a set of common mea- sures of bilingual experience, enabling greater comparability across studies, and facilitating exchanges and cross-pollination across sectors. Also, almost unanimous was the desire for an auto- mated calculator of language exposure and language use, and the need to allow for measuring multilingual (not only bilingual) experience. 51.2 The child-focused versions will need to be complemented by a brief parental questionnaire. [80%, original] 51.2 The child-focused versions will need to be complemented by a brief parental questionnaire. [80%, original] Consensus was also reached regarding the need to document a number of aspects of bilingual experience. This includes, for each language: language exposure and use, language mixing, language difficulties experienced, proficiency (if it cannot be assessed by other means), education and literacy, indices of input quality, lan- guage mixing practices, and attitudes (towards each language and towards language mixing). Consensus levels were the highest in relation to language exposure and use, and the need to document them in detail (i.e., across interlocutors, in different contexts, and over time). p q g 52. The questionnaire in which the child is the respondent should be administered as an interview with the child. [76%, original] https://doi.org/10.1017/S1366728922000359 Published online by Cambridge University Press 4. Discussion Overall, the level of consensus was high: 79% of statements (i.e., 98/124) reached consensus across rounds, and only 8% of state- ments (10/124) received less than a 60% agreement rate. This sug- gests that the views of the experts from the initial workshop were The apparent contradiction between the call for a flexible and modular tool and the acknowledgement of the necessity to https://doi.org/10.1017/S1366728922000359 Published online by Cambridge University Press 123 Bilingualism: Language and Cognition document a wide range of aspects of bilingual experience brings us to the question of what is ‘core’ vs. what can be considered optional in the documentation. While this question was not asked directly in the survey, we believe variations in level of con- sensus can be interpreted as useful indicators. Language exposure and use, as well as some indicators of language difficulties and some indicators of input quality, thus seem to emerge as core aspects of the quantification of bilingual experience. Ultimately, though, the identification of an essential ‘core’ is an empirical question, which will require comparing data from large and diverse groups of bilinguals and multilinguals, using identical measures. And this empirical investigation will need to ascertain not only which aspects of bilingual experience are part of the ‘core’, but also the minimum level of detail required to measure (or document) them reliably. Indeed, the cognitive load involved in completing bilingual experience questionnaires can be quite complex, as respondents are required not only to recall language practices over long periods of time and many different contexts, but also to estimate their frequency. In addition, they might not have been direct witnesses of these practices (e.g., in the case of a parent estimating what happens at school), or they might not have been fully aware of them (e.g., in the case of language mix- ing). This is likely to result in unavoidably high levels of error in the measurements. It is therefore not necessarily the case that a more detailed questionnaire will elicit more precise information. Here too, empirical investigation will be necessary to identify the optimal level of detail to be targeted by bilingual experience questionnaires. their value: (1) provide reproducible criteria for panellist selection, (2) state the number of rounds performed, (3) provide clear cri- teria for dropping the items, and (4) clarify whether there is a stopping criteria other than the number of rounds specified. 11The design phase took place while the manuscript for this paper was under review. The new online tool is now available for free at https://q-bex.org. 4. Discussion We reported on each of them in this manuscript. The ultimate limitation of the Delphi method is that, in spite of all attempts at bias limitation, it remains possible that the cor- rect answer or opinion was not identified (Hasson et al., 2000; Iqbal & Pipon-Young, 2009). Opinion is also likely to evolve as bilingualism research progresses. However, we believe the findings of this survey could enable a step-change in bilingualism research through the adoption of a common method to document bilin- gual experience, thereby enabling greater comparability across studies (and across populations), and increased synergy between practitioners and researchers. The validity of the tool developed on the basis of the Delphi consensus will need to be assessed based on new empirical evidence. It may turn out that some of the aspects of bilingual experience which were deemed necessary to document (as per the consensus reached by the Delphi survey) can in fact not be reliably documented via questionnaires. If that is the case, it will indicate that opinions need to change. This paper represents the first step in this long process, that is, the documentation of current opinions. p There are also limitations inherent to the quantification of (aspects of) the bilingual experience, which was at the heart of this Delphi survey. Such an approach might not be sufficient to capture key aspects of bilingualism associated with variation across socio-cultural contexts, such as non-industrialised coun- tries without a robust education system. Ethnographic approaches (not considered in this survey) might offer insightful alternatives in some contexts. It is possible that standardised questionnaires are inadequate to document bilingualism in some populations. The extent to which bi/multilingualism in these populations can be conceptualised along the same lines as in other populations is an important question beyond the scope of this paper and will require further research. 4.2 Limitations While the Delphi method provides a flexible approach to accom- modate the needs of various fields, a clear limitation is the lack of agreed standards on what should count as consensus and how it should be interpreted (Iqbal & Pipon-Young, 2009). To limit the risk of bias, we adopted a pre-defined criterion informed by Diamond et al.’s (2014) review, which identified 75% agreement among panellists as the median consensus threshold among the publications in which consensus was defined as a percentage or proportion. Furthermore, we adopted a pre-defined criterion to identify round 1 statements in the proximal agreement zone (i.e., the proximal zone statements): informed by the approach of Langlands et al. (2008) and Spain and Happé (2019), any state- ment reaching agreement between 60% and 74% of round 1 panellists was automatically selected for re-rating at round 2. We also pre-defined a dual interest group (identifying as both researchers and practitioners), whose proximal zone state- ments (i.e., statements with the agreement rate 60%-74%) were automatically selected for round 2, in addition to those of the panel as a whole. These procedures enabled us to mitigate the risk of a potentially too conservative consensus threshold. https://doi.org/10.1017/S1366728922000359 Published online by Cambridge University Press 5. Concluding remarks The need for standardisation in how bilingualism is characterised and categorised has become uncontroversial. This Delphi consen- sus survey has highlighted the readiness of bilingualism research- ers and practitioners from both clinical and educational settings to adopt common methods for the documentation of bilingual experience, to enhance the generalisability of research findings and facilitate exchanges between research and practice. Byers-Heinlein K (2013) Parental language mixing: Its measurement and the relation of mixed input to young bilingual children’s vocabulary size. Bilingualism: Language and Cognition 16, 32–48. https://doi.org/10.1017/ S1366728912000120 Clarke V and Braun V (2017) Thematic analysis. The Journal of Positive Psychology 12(3), 297–298, DOI: 10.1080/17439760.2016.1262613 DeVellis RF (2017) Scale Development Theory and Applications (4th ed.). SAGE Publications, Inc. Several new profiling measures have been proposed recently, as global indices of bilingualism: for example, the LSBQ composite score (Anderson, Mak, Keyvani Chahi & Bialystok, 2018; Anderson, Hawrylewicz & Bialystok, 2020), language entropy (Gullifer & Titone, 2020), or a possible bilingualism quotient (Marian & Hayakawa, 2020). Independent of their conceptual valid- ity, the reliability of these measures will be determined by how their componentsaredocumentedandquantified.Thisinturnwillrequire comparabilityofmeasures,whichwillonlybepossibleifsimilardocu- mentation tools are used across research teams. The results of this Delphi consensus survey constitute a first step in that direction. Diamond IR, Grant RC, Friedman BM, Pencharz PB, Ling SC, Moore AM and Wales PW (2014) Defining consensus: A systematic review recom- mends methodologic criteria for reporting of Delphi studies. Journal of Clinical Epidemiology 67, 401–409. Gullifer JW and Titone D (2020) Characterizing the social diversity of bilin- gualism using language entropy. Bilingualism: Language and Cognition 23 (2), 283–294. https://doi.org/10.1017/S1366728919000026 Hasson F, Keeney S and McKenna H (2000) Research guidelines for the Delphi survey technique. Journal of Advanced Nursing 32(4), 1008–1015. Iqbal S and Pipon-Young L (2009) The Delphi method. Methods 22(7), 598–601. Kašćelan D, Prévost P, Serratrice L, Tuller L, Unsworth S and De Cat C (2021) A review of questionnaires quantifying bilingual experience in chil- dren: Do they document the same constructs? Bilingualism: Language and Cognition, 1–13. https://doi.org/10.1017/S1366728921000390 Acknowledgments. Warm thanks to the Q-BEx Consortium for contribut- ing to the Delphi survey (twice!), and to the workshop participants for their insights and enthusiasm. We are grateful to Bissera Ivanova for helping us communicate anonymously with the panellists. We thank the reviewers for inviting us to go deeper into the discussion and to clarify the remit of our enterprise. 4.3. Next steps 4.3. Next steps The main authors of this paper are currently designing an online questionnaire11 and background calculator meeting the require- ments identified via the Delphi consensus survey and informed by a review of the state of the art in bilingualism questionnaires (Kašćelan et al., 2021), and by methodological insights from the literature on psychometrics (e.g., DeVellis, 2017). This question- naire will allow some level of customisability, so that professionals can choose the components relevant to their research objectives, clinical practice, or educational aims. To limit the burden on respondents, we will also carry out a cost-benefit analysis aiming to identify the optimal balance between informativity and error margin. This is particularly important as lengthy questionnaires are likely to be too burdensome for some marginalised communi- ties, which would result in less representative population samples. As is the case with the creation of any standardised tool, the domain of applicability of this new questionnaire will need to be assessed empirically. In the long term, a multi-team approach will be necessary for full validation and optimisation. In spite of our diversification strategy, the panel remained pre- dominantly Western-centric (both in the workshop and in the online survey). The use of English and the use of an online plat- form to conduct the study aimed to increase inclusivity, but at the same time it discriminated against non-English speaking stake- holders and individuals without access to the internet. Future work should therefore seek to improve representation, as many significant voices have likely not been included. The creation of a new tool raises several methodological chal- lenges. First, the constructs that make up bilingual experience g y We also note that there are no validated quality indicators of Delphi studies. However, Diamond et al. (2014) proposed a set of four elements to include in Delphi publications to increase Cécile De Cat et al. 124 bilingualism in a diverse population. Behavior Research Methods 50, 250–263. https://doi.org/10.3758/s13428-017-0867-9 need to be operationalised precisely and concretely to avoid reifi- cation. Second, there is a risk of normalisation inherent to the cre- ation of a standard questionnaire. This risk must be limited by embedding into the questionnaire design the intention to capture the diversity of bilingual experiences. This will however need to be balanced with the need to keep the questionnaire sufficiently short and clear for respondents. bilingualism in a diverse population. Behavior Research Methods 50, 250–263. 5. Concluding remarks This project is funded by the Economic and Social Research Council (grant reference: ES/S010998/1), which is gratefully acknowledged. Langlands RL, Jorm AF, Kelly CM and Kitchener BA (2008) Journal of Affective Disorders 105, 157–165. Luk G and Esposito A (2020) BLC mini series: Tools to document bilingual experiences. Bilingualism: Language and Cognition 23(5), 927–928. Supplementary Material. For supplementary material accompanying this paper, visit https://doi.org/10.1017/S1366728922000359 Marian V and Hayakawa S (2020) Measuring bilingualism: the quest for a “bilingualism quotient”. Applied Psycholinguistics, 1–22. doi:10.1017/ S0142716420000533 Data availability statement. The data that supports the findings of this study are openly available via the Open Science Framework at https://osf.io/ 2pd65/. Spain D and Happé F (2019) How to Optimise Cognitive Behaviour Therapy (CBT) for People with Autism Spectrum Disorders (ASD): A Delphi Study. Journal of Rational-Emotive & Cognitive-Behavior Therapy. https://doi.org/ 10.1007/s10942-019-00335-1 Competing interests declaration. The authors declare none. Thangaratinam S and Redman C (2005) The Delphi technique. The Obstetrician & Gynaecologist 7(2), 120–125. Unsworth S (2016) Quantity and Quality of Language Input in Bilingual Language Development. In Nicoladis E and Montanari S (Eds.), Bilingualism Across the Lifespan: Factors Moderating Language Proficiency (pp. 103–121). De Gruyter Mouton. 4.3. Next steps https://doi.org/10.3758/s13428-017-0867-9 Bishop DVM, Snowling MJ, Thompson PA, Greenhalgh T CATALISE consortium (2016) CATALISE: A Multinational and Multidisciplinary Delphi Consensus Study. Identifying Language Impairments in Children. PLoS ONE 11(7), 1–26. Braun V and Clarke V (2019) Reflecting on reflexive thematic analysis. Qualitative Research in Sport, Exercise and Health 11(4), 589–597, DOI: 10.1080/2159676X.2019.1628806 Braun V, Clarke V and Hayfield N (2019) ‘A starting point for your journey, not a map’: Nikki Hayfield in conversation with Virginia Braun and Victoria Clarke about thematic analysis. Qualitative Research in Psychology, DOI: 10.1080/14780887.2019.1670765 References Anderson JAE, Hawrylewicz K and Bialystok E (2020) Who is bilingual? Snapshots across the lifespan. Bilingualism: Language and Cognition 23, 929–937. https://doi.org/10.1017/S1366728918000950 Anderson JAE, Mak L, Keyvani Chahi A and Bialystok E (2018) The language and social background questionnaire: Assessing degree of Anderson JAE, Hawrylewicz K and Bialystok E (2020) Who is bilingual? Snapshots across the lifespan. Bilingualism: Language and Cognition 23, 929–937. https://doi.org/10.1017/S1366728918000950 (pp. 103–121). De Gruyter Mouton. Walker AM and Selfe J (1996) The Delphi technique: a useful tool for the allied health researcher. British Journal of Therapy and Rehabilitation 3, 677–680. Anderson JAE, Mak L, Keyvani Chahi A and Bialystok E (2018) The language and social background questionnaire: Assessing degree of https://doi.org/10.1017/S1366728922000359 Published online by Cambridge University Press https://doi.org/10.1017/S1366728922000359 Published online by Cambridge University Press
https://openalex.org/W1988390113	https://www.scielo.br/j/rbpm/a/TGDNfz8fT6JJmNHn9b7CFsQ/?lang=pt&format=pdf	Portuguese	null	Caracterização morfológica e agronômica de acessos de jambu (Spilanthes oleracea L.) nas condições do Norte de Minas Gerais	Revista brasileira de plantas medicinais	2,012	cc-by	3,151	Caracterização morfológica e agronômica de acessos de jambu (Spilanthes oleracea L.) nas condições do Norte de Minas Gerais MARTINS, C.P.S.; MELO, M.T.P.; HONÓRIO, I.C.G.; D’ÁVILA, V.A.; CARVALHO JÚNIOR, W.G.O. Universidade Federal de Minas Gerais - Departamento de Fitotecnia, Avenida Universitária, 1000, Bairro Universitário, CEP: 39404-006, Montes Claros-Brasil martinscps@yahoo.com.br MARTINS, C.P.S.; MELO, M.T.P.; HONÓRIO, I.C.G.; D’ÁVILA, V.A.; CARVALHO JÚNIOR, W.G.O. Universidade Federal de Minas Gerais - Departamento de Fitotecnia, Avenida Universitária, 1000, Bairro Universitário, CEP: 39404-006, Montes Claros-Brasil martinscps@yahoo.com.br RESUMO: O objetivo do presente trabalho foi caracterizar os aspectos morfológicos e agronômicos de três acessos de jambu (Spilanthes oleracea L.) nas condições do Norte de Minas Gerais. O estudo foi desenvolvido em casa de vegetação, no período de abril a junho de 2009, no Instituto de Ciências Agrárias da Universidade Federal de Minas Gerais (ICA/UFMG), localizado na cidade de Montes Claros, Minas Gerais, Brasil. O delineamento experimental foi o inteiramente casualizado, com três tratamentos e oito repetições. Os tratamentos foram os acessos de jambu (1- Montes Claros/MG; 2- Pará, Norte do Brasil; 3- Cristália/MG). Dois meses após o transplantio, foram avaliados o comprimento da inflorescência (cm), o comprimento e a largura das folhas do terceiro nó de cada planta (cm) e a matéria fresca e seca da parte aérea e das inflorescências (g). Foram também caracterizadas a coloração dos caules e os tipos de inflorescências. A cor dos caules foi distinguida visualmente, variando de roxo intenso, acessos de Cristália e Montes Claros, a verde claro, acesso do estado do Pará. Os subtipos de inflorescência foram caracterizadas como simples ou geminadas e alongadas ou globóides. Observou-se que o acesso de Jambu, proveniente do estado do Pará, apresentou os melhores resultados para a maioria das características avaliadas. Esse acesso apresentou inflorescências dos tipos simples alongada,e, ocasionalmente, geminadas, tanto globóides como alongadas, enquanto os acessos de Cristália e Montes Claros apresentaram inflorescências simples globóides. Pode-se concluir que os três acessos são equivalentes comercialmente. Entretanto, para produção de sementes, recomenda-se o cultivo do acesso de jambu proveniente do estado do Pará. Palavras-chave: Spilanthes oleracea L., caracteres morfo-agronômicos, planta medicinal, hortaliça não-convencional ABSTRACT: Morphological and agronomic characterization of Jambu (Spilanthes oleracea L.) accessions under the conditions of North Minas Gerais State, Brazil. The aim of this study was to characterize morphological and agronomic aspects of three Jambu (Spilanthes oleracea L.) accessions under the conditions of North Minas Gerais State, Brazil. NOTA PRÉVIA NOTA PRÉVIA 410 Recebido para publicação em 09/03/2010 Aceito para publicação em 12/01/2012 INTRODUÇÃO Federal de Minas Gerais (ICA/UFMG), localizado na cidade de Montes Claros, norte de Minas Gerais (16º41’00’’S e 43º50’00’’W), no período de abril a junho de 2009. Essa região caracteriza-se fisionomicamente pelo predomínio do Cerrado sensu stricto. Na classificação de Koeppen (1948), o clima da região é o Aw, com estação seca bem definida e chuvas concentradas nos meses de novembro a janeiro. As sementes foram coletadas de três acessos de jambu, em diferentes regiões do Brasil, a saber: Acesso 1 - proveniente da cidade de Montes Claros, Norte de Minas Gerais, Acesso 2 - proveniente do estado do Pará, região Norte do Brasil e Acesso 3 - proveniente da cidade de Cristália, Norte de Minas Gerais. O jambu (Spilanthes oleracea L.) é uma Asteraceae nativa da região amazônica, hortaliça herbácea perene, semi-ereta e de ramos decumbentes. As inflorescências são pequenas e amareladas, dispostas em capítulos. O consumo da espécie no estado do Pará é bastante difundido, compondo diversos pratos, como pato no tucupi e tacacá, sendo também muito utilizada em saladas. É também conhecido por agrião do Pará, agrião do Brasil, agrião do Norte, jabuaçu, erva maluca, jaburama, botão de ouro, entre outros (Coutinho et al., 2006). Além de muito utilizado na culinária, o jambu possui importância medicinal, por possuir princípios ativos como óleo essencial, saponinas, espilantinas, afinina, filoesterina, colina, triterpenóides e, principalmente, o espilantol, que teve a estrutura desvendada por Yasuda et al. (1980). Além dessas propriedades, também é utilizado como matéria-prima em cosméticos anti-rugas. As inflorescências quando mastigadas provocam sensação de dormência nos lábios e na língua. Estudo realizado por Saraf & Dixit (2002) evidenciou que o espilantol apresenta grande atividade inseticida contra os mosquitos Anopheles culicifacies (Giles, 1910), Culex quiquefasciatus (Say, 1823) e Aedes aegypti (Linnaeus, 1762). Essa hortaliça apresenta- se como importante fonte de renda para pequenos produtores dos municípios do Pará, uma vez que é planta de múltiplo uso (condimentar, medicinal e ornamental), reunindo elementos essenciais para formação de um sistema sustentável (Gusmão et al., 2003). As sementes foram colocadas para germinar em bandejas de polietileno expandido, sendo duas sementes por célula, preenchida com substrato comercial Plantmax® sob condições de casa de vegetação com cobertura plástica. Caracterização morfológica e agronômica de acessos de jambu (Spilanthes oleracea L.) nas condições do Norte de Minas Gerais The study was carried out in a greenhouse from April to June 2009 in the Institute for Agrarian Sciences of Federal University of Minas Gerais (ICA/UFMG), located in the city of Montes Claros, Minas Gerais, Brazil. The experimental design was completely randomized, with three treatments and eight replicates. Treatments were Jambu accessions (1- from Montes Claros City, Minas Gerais; 2- from Pará State, North of Brazil; 3- from Cristália City, Minas Gerais). Two months after transplanting, the inflorescence length (cm), the length and width of leaves of the third node of each plant (cm) and the fresh and dry matter of shoot and inflorescences (g) were assessed. The color of stems and the type of inflorescences were also characterized. Stem color was visually distinguished, varying from intense purple, Cristália and Montes Claros accessions, to light green, Pará State accession. Inflorescence subtypes were characterized as simple or geminated and elongated or globoid. Jambu accession from Pará State presented the best results for most assessed characteristics. This accession presented elongated simple and, occasionally, both globoid and elongated geminated inflorescences, while Cristália and Montes Claros accessions presented globoid simple inflorescences. It can be concluded that all three accessions are commercially equivalent. However, for seed production, cultivation of Jambu accession from Pará State is recommended. Key words: Spilanthes oleracea L., morpho-agronomic traits, medicinal plant, non-conventional vegetable Rev. Bras. Pl. Med., Botucatu, v.14, n.2, p.410-413, 2012. 411 INTRODUÇÃO Após 45 dias da semeadura, foram transplantadas duas mudas por vaso, com capacidade de três litros, contendo terra de subsolo, esterco bovino curtido e areia, na proporção de 3:2:1, respectivamente, sendo esses dispostos aleatoriamente em bancadas de madeira em casa de vegetação. Diariamente, realizavam-se tratos culturais como capina manual de plantas espontâneas e irrigação. Dois meses após o transplantio, foram analisadas as variáveis: comprimento das inf lorescências (cm), comprimento e largura das folhas do terceiro nó de cada planta (cm), medido com paquímetro digital, matéria fresca e seca da parte aérea e das inflorescências (g) pesadas em balança analítica com precisão de três casas decimais. A caracterização morfológica e agronômica de populações é uma das etapas iniciais dos programas de melhoramento, que visam ao desenvolvimento de populações com potencial agronômico, como produção, hábito de crescimento, altura da planta, resistência a pragas, entre outras características (Rosa et al., 2006). Vários trabalhos vêm sendo desenvolvidos com o propósito de identificação de genótipos superiores ou seleções regionais que atendam aos interesses de mercado, como os de Meletti et al. (2005) e de Jesus et al. (2004). Visualmente, também foram distinguidos a cor dos caules das plantas, que variavam de roxo intenso a verde claro, e o subtipo das inflorescências (tipo capítulo), classificados neste trabalho de acordo com a presença/ ausência de geminação, podendo ser simples (único capítulo em um eixo caulinar) ou geminado (dois capítulos em um único eixo caulinar), e também de acordo com o formato da inflorescência, podendo ser globóide ou alongado. Na determinação da matéria seca, as amostras foram mantidas em estufa de circulação forçada de ar a 65ºC, até massa constante. Dada a importância econômica e medicinal do jambu e à intenção de se iniciar programa de melhoramento da espécie, o presente trabalho visou caracterizar morfológica e agronomicamente três diferentes acessos de Spilanthes oleracea L., sendo dois nas condições do norte de Minas Gerais, e um proveniente do estado do Pará. O delineamento experimental utilizado foi o inteiramente casualizado, com três tratamentos (Acessos) e oito repetições. Os resultados foram submetidos à análise de variância e, em seguida, aplicou-se o teste de Scott & Knott, a 5% de probabilidade de erro, para a comparação das médias. Para a análise estatística utilizou-se o RESULTADO E DISCUSSÃO De acordo com os resultados obtidos, as variáveis matéria fresca e seca da parte aérea não apresentaram diferença estatística entre os tratamentos. Isso pode ser explicado pelo fato dos acessos 1 (Montes Claros/MG) e 3 (Cristália/ MG) apresentarem tamanhos de folhas aproximados e, visualmente, quantidades semelhantes dessas. Entretanto, apesar desses acessos apresentarem folhas de menor comprimento e largura (Tabela 1) que o acesso do Pará (2), essa diferença pode ter sido compensada pela menor quantidade de folhas desse. De acordo com Coutinho et al. (2006), a parte de maior uso dessa espécie, são as folhas para o uso culinário. Assim, apesar das diferenças encontradas, comercialmente os três acessos não se diferem. Para melhor diferenciação, além das folhas do terceiro nó, algumas características foram constatadas visualmente. Primeiramente, referente à cor das hastes. Enquanto a haste do jambu, proveniente do Pará (Acesso 2), possui coloração verde claro, os de Cristália (Acesso 3) e de Montes Claros (Acesso 1) possuem coloração roxo intenso (Figura 1). Pimentel (1985) caracterizou como verde clara a cor das hastes dessa planta, por outro lado, segundo Gusmão et al. (2003), em estudos realizados em Belém do Pará, relataram que em pequenas quantidades das plantas da região, aproximadamente 10%, a coloração roxo intenso nas hastes também é encontrada, porém diferenciado pelo nome popular de jambu roxinho. Entretanto, apesar desses acessos apresentarem folhas de menor comprimento e largura (Tabela 1) que o acesso do Pará (2), essa diferença pode ter sido compensada pela menor quantidade de folhas desse. De acordo com Coutinho et al. (2006), a parte de maior uso dessa espécie, são as folhas para o uso culinário. Assim, apesar das diferenças encontradas, comercialmente os três acessos não se diferem. O acesso 2 apresentou as melhores médias em todas as variáveis avaliadas que se diferiram estatisticamente. Por outro lado, os acessos 1 e 3 não deferiram entre si em todos os aspectos avaliados. Isso pode ser justificado pelo fato desses dois últimos serem procedentes da mesma região, enquanto o acesso 2 é proveniente de uma região com características edafoclimáticas distintas. O formato da inflorescência também é fator de diferenciação, por isso justifica-se os maiores resultados do Acesso 2 em relação ao comprimento dessa parte da planta. MATERIAL E MÉTODO O presente ensaio foi conduzido no Instituto de Ciências Agrárias da Universidade Rev. Bras. Pl. Med., Botucatu, v.14, n.2, p.410-413, 2012. 412 programa SAEG - Sistema para Análises Estatísticas e Genéticas (Ribeiro Júnior, 2001). novamente, o acesso 2 apresentou maiores médias em comparação aos outros dois acessos. Como o peso total das folhas não apresentou diferença, esse aspecto tem funcionalidade apenas na diferenciação dos acessos. RESULTADO E DISCUSSÃO Enquanto os acessos de Cristália e Montes Claros apresentaram inflorescências simples globóides, as inflorescências do jambu do Pará são, em sua maioria, simples alongadas, apresentando ocasionalmente, inflorescências geminadas alongadas e geminadas globóides (Figura 1). De acordo com Ferreira et al. (2007), uma das formas mais eficientes de propagação do jambu é a seminífera. Visto que o acesso do Pará apresentou maior comprimento das inflorescências (Figura 1), consequentemente maior número de sementes, pode-se dizer que esse acesso possui também maior probabilidade de sucesso reprodutivo, quando comparado com os outros acessos. O acesso proveniente do Pará difere fenotipicamente dos acessos de Montes Claros e Cristália, que apresentam características morfológicas e agronômicas semelhantes. Analisando os dados, recomenda-se para produção de sementes, o acesso proveniente do Pará. No entanto, são necessários estudos mais aprofundados para se recomendar a melhor espécie para a extração do princípio ativo com fim nutricional e medicinal. Evidenciou-se também neste trabalho, a diferença nos comprimentos e larguras das folhas do terceiro nó das plantas avaliadas. E, TABELA 1. Comprimento das inflorescências (COMP. INFL.), matéria fresca e seca das inflorescências (MF. INFL e MS. INFL), comprimento e largura das folhas do terceiro nó das plantas (COMP. F e LARG. F) e matéria fresca e seca da parte aérea (MF. PA e MS. PA) dos três acessos de jambu avaliados. Montes Claros/MG. TABELA 1. Comprimento das inflorescências (COMP. INFL.), matéria fresca e seca das inflorescências (MF. INFL e MS. INFL), comprimento e largura das folhas do terceiro nó das plantas (COMP. F e LARG. F) e matéria fresca e seca da parte aérea (MF. PA e MS. PA) dos três acessos de jambu avaliados. Montes Claros/MG. Acessos COMP.INFL. MF.INFL. MS.INFL. COMP.F. LARG. F. MF.PA MS.PA (mm) (g) (g) (mm) (mm) (g) (g) M. Claros 11,1846 b 6,2152 b 1,2277 b 64,7300 b 48,7800 b 40,1250 a 6,9481 a Pará 18,7052 a 16,5802 a 2,9643 a 78,5038 a 61,7125 a 48,8750 a 6,6357 a Cristália 10,4801 b 7,5329 b 1,5323 b 53,6550 b 42,1937 b 53,2857 a 9,0359 a CV(%) 26,284 64,345 59,619 24,145 26,769 41,951 41,831 As médias seguidas da mesma letra nas colunas não se diferem estatisticamente pelo Teste de Scott & Knott, a 5% de probabilidade. Rev. Bras. Pl. Med., Botucatu, v.14, n.2, p.410-413, 2012. Rev. Bras. Pl. Med., Botucatu, v.14, n.2, p.410-413, 2012. 413 FIGURA 1. RESULTADO E DISCUSSÃO Caracterização e variações entre as inflorescências e coloração do caule de acessos de jambu (Spilanthes oleracea L.). A- Inflorescências simples globóides do acesso de Montes Claros/MG (1). B- Inflorescências simples alongadas do acesso do Estado do Pará (2). C- Inflorescências simples globóides do acesso de Cristália/ MG (3). D - Detalhe do caule do acesso 1. E - Detalhe do caule do acesso 2. F- Detalhe do caule do acesso 3. G- Detalhe da inflorescência geminada alongada do acesso 2. H - Detalhe da inflorescência geminada globóide do acesso 2. Barra = 10 mm. FIGURA 1. Caracterização e variações entre as inflorescências e coloração do caule de acessos de jambu (Spilanthes oleracea L.). A- Inflorescências simples globóides do acesso de Montes Claros/MG (1). B- Inflorescências simples alongadas do acesso do Estado do Pará (2). C- Inflorescências simples globóides do acesso de Cristália/ MG (3). D - Detalhe do caule do acesso 1. E - Detalhe do caule do acesso 2. F- Detalhe do caule do acesso 3. G- Detalhe da inflorescência geminada alongada do acesso 2. H - Detalhe da inflorescência geminada globóide do acesso 2. Barra = 10 mm. REFERÊNCIA MELETTI, L.M.M. et al. Caracterização fenotípica de três seleções de maracujazeiro-roxo. Revista Brasileira de Fruticultura, v.27, n.2, p.268-72, 2005. Disponível em: <http://www.scielo.br/pdf/rbf/v27n2/a20v27n2.pdf>. Acesso em: 10 Jun. 2009. MELETTI, L.M.M. et al. Caracterização fenotípica de três seleções de maracujazeiro-roxo. Revista Brasileira de Fruticultura, v.27, n.2, p.268-72, 2005. Disponível em: <http://www.scielo.br/pdf/rbf/v27n2/a20v27n2.pdf>. Acesso em: 10 Jun. 2009. COUTINHO,L.N. et al. Galhas e deformações em Jambu (Spilanthes oleraceae L.) causadas por Tecaphora spilanthes (Ustilaginales). Summa Phytopathology, v.32, n.3, p.283-5, 2006. Botucatu. Disponível em: <http:// www.scielo.br/scielo.php?script=sci_arttext&pid= S0100-54052006000300013>. Acesso em: 18 Abr 2009. FERREIRA, A.M.S.D. et al. Avaliação de diferentes substratos na germinação do jambu (Spilanthes oleraceae L.). Revista Brasileira de Horticultura, v.25, n.1, p.20-4, 2007. Disponível em: <http://www.abhorticultura. com.br/biblioteca/arquivos/Download/Biblioteca/46_ 0593.pdf>. Acesso em: 1 Jul. 2009. GUSMÃO, S.A.L. et al. Caracterização do cultivo de jambu nas áreas produtoras que abastecem a grande Belém. CONGRESSO BRASILEIRO DE OLERICULTURA, 2003. Disponível em: <http://www.abhorticultura.com.br/ biblioteca/arquivos/Biblioteca/olfg4074C.pdf>. Acesso em: 10 Abr. 2009. JESUS, N. et al. Caracterização de quatro grupos de jabuticabeira, nas condições de Jaboticabal-SP. Revista Brasileira de Fruticultura, v.26, n.3, p.482-5, 2004. Disponível em: <http://www.scielo.br/pdf/rbf/26n3/ 23150.pdf >. Acesso em: 10 Jun. 2009. COUTINHO,L.N. et al. Galhas e deformações em Jambu (Spilanthes oleraceae L.) causadas por Tecaphora spilanthes (Ustilaginales). Summa Phytopathology, v.32, n.3, p.283-5, 2006. Botucatu. Disponível em: <http:// www.scielo.br/scielo.php?script=sci_arttext&pid= S0100-54052006000300013>. Acesso em: 18 Abr 2009. FERREIRA, A.M.S.D. et al. Avaliação de diferentes substratos na germinação do jambu (Spilanthes oleraceae L.). Revista Brasileira de Horticultura, v.25, n.1, p.20-4, 2007. Disponível em: <http://www.abhorticultura. com.br/biblioteca/arquivos/Download/Biblioteca/46_ 0593.pdf>. Acesso em: 1 Jul. 2009. Ã PIMENTEL, A.A.M.P. Olericultura no trópico úmido: hortaliças na Amazônia. São Paulo: Agro Ceres, 1985. 322p. ROSA,M.S. et al. Caracterização agromorfológica interpopulacional em Oryza glumaepatula. Bragantia, v.65, n.1, p.1-10, 2006. Disponível em: <http://www.scielo.br/ s c i e lo . p h p ? s c r i p t = s c i _ ar t t e xt & p id = S 0 0 0 6 - 87052006000100002>. Acesso em: 10 Jun. 2009. SARAF,D.K. et al. Spilanthes acmella Murr.: Study on its extract spilanthol as larvicidal compound. Asian Journal of Experience Sciences, v.16, n.1/2, p.9-19, 2002. Disponível em: <http://ajes.in/PDFs/02/2%20VK%20Dixit.pdf>. Acesso em: 05 Abr. 2009. ROSA,M.S. et al. Caracterização agromorfológica interpopulacional em Oryza glumaepatula. Bragantia, v.65, n.1, p.1-10, 2006. Disponível em: <http://www.scielo.br/ s c i e lo . p h p ? REFERÊNCIA s c r i p t = s c i _ ar t t e xt & p id = S 0 0 0 6 - 87052006000100002>. Acesso em: 10 Jun. 2009. ROSA,M.S. et al. Caracterização agromorfológica interpopulacional em Oryza glumaepatula. Bragantia, v.65, n.1, p.1-10, 2006. Disponível em: <http://www.scielo.br/ s c i e lo . p h p ? s c r i p t = s c i _ ar t t e xt & p id = S 0 0 0 6 - 87052006000100002>. Acesso em: 10 Jun. 2009. SARAF,D.K. et al. Spilanthes acmella Murr.: Study on its extract spilanthol as larvicidal compound. Asian Journal of Experience Sciences, v.16, n.1/2, p.9-19, 2002. Disponível em: <http://ajes.in/PDFs/02/2%20VK%20Dixit.pdf>. Acesso em: 05 Abr. 2009. YASUDA, I. et al. The geometric structure of spilanthol. Chemical & Pharmaceutical Bulletin, v.28, n.7, p.2251-3, 1980. Disponível em: <http://nels.nii.ac.jp/els/ 110003662647.pdf?id= ART000420 1392&type=pdf&lang= en&host=cinii&order_no=&ppv_type=0&lang_sw= &no= 1246836584&cp=>. Acesso em: 10 Abr. 2009. FERREIRA, A.M.S.D. et al. Avaliação de diferentes substratos na germinação do jambu (Spilanthes oleraceae L.). Revista Brasileira de Horticultura, v.25, n.1, p.20-4, 2007. Disponível em: <http://www.abhorticultura. com.br/biblioteca/arquivos/Download/Biblioteca/46_ 0593.pdf>. Acesso em: 1 Jul. 2009. SARAF,D.K. et al. Spilanthes acmella Murr.: Study on its extract spilanthol as larvicidal compound. Asian Journal of Experience Sciences, v.16, n.1/2, p.9-19, 2002. Disponível em: <http://ajes.in/PDFs/02/2%20VK%20Dixit.pdf>. Acesso em: 05 Abr. 2009. GUSMÃO, S.A.L. et al. Caracterização do cultivo de jambu nas áreas produtoras que abastecem a grande Belém. CONGRESSO BRASILEIRO DE OLERICULTURA, 2003. Disponível em: <http://www.abhorticultura.com.br/ biblioteca/arquivos/Biblioteca/olfg4074C.pdf>. Acesso em: 10 Abr. 2009. YASUDA, I. et al. The geometric structure of spilanthol. Chemical & Pharmaceutical Bulletin, v.28, n.7, p.2251-3, 1980. Disponível em: <http://nels.nii.ac.jp/els/ 110003662647.pdf?id= ART000420 1392&type=pdf&lang= en&host=cinii&order_no=&ppv_type=0&lang_sw= &no= 1246836584&cp=>. Acesso em: 10 Abr. 2009. JESUS, N. et al. Caracterização de quatro grupos de jabuticabeira, nas condições de Jaboticabal-SP. Revista Brasileira de Fruticultura, v.26, n.3, p.482-5, 2004. Disponível em: <http://www.scielo.br/pdf/rbf/26n3/ 23150.pdf >. Acesso em: 10 Jun. 2009. JESUS, N. et al. Caracterização de quatro grupos de jabuticabeira, nas condições de Jaboticabal-SP. Revista Brasileira de Fruticultura, v.26, n.3, p.482-5, 2004. Disponível em: <http://www.scielo.br/pdf/rbf/26n3/ 23150.pdf >. Acesso em: 10 Jun. 2009. Rev. Bras. Pl. Med., Botucatu, v.14, n.2, p.410-413, 2012.
https://openalex.org/W2536805108	https://europepmc.org/articles/pmc5078490?pdf=render	English	null	Genetic Ablation of CXCR2 Protects against Cigarette Smoke-Induced Lung Inflammation and Injury	Frontiers in pharmacology	2,016	cc-by	6,201	ORIGINAL RESEARCH published: 25 October 2016 doi: 10.3389/fphar.2016.00391 Edited by: Narasaiah Kolliputi, University of South Florida, USA Narasaiah Kolliputi, University of South Florida, USA Reviewed by: M. Firoze Khan, University of Texas Medical Branch, USA Anup Srivastava, Yale University, USA Yohannes T. Ghebre, Baylor College of Medicine, USA Correspondence: Irfan Rahman irfan_rahman@urmc.rochester.edu †Present Address: Wei Lei, Department of Respiratory Medicine, The First Afﬁliated Hospital of Soochow University, Suzhou, China Reviewed by: M. Firoze Khan, University of Texas Medical Branch, USA Anup Srivastava, Yale University, USA Yohannes T. Ghebre, Baylor College of Medicine, USA Reviewed by: M. Firoze Khan, University of Texas Medical Branch, USA Anup Srivastava, Yale University, USA Yohannes T. Ghebre, Baylor College of Medicine, USA Correspondence: Irfan Rahman irfan_rahman@urmc.rochester.edu Specialty section: This article was submitted to Respiratory Pharmacology, a section of the journal Frontiers in Pharmacology Received: 18 August 2016 Accepted: 06 October 2016 Published: 25 October 2016 Keywords: cigarette smoke, inﬂammation, DNA damage, NF-κB, emphysema Genetic Ablation of CXCR2 Protects against Cigarette Smoke-Induced Lung Inﬂammation and Injury Chad A. Lerner, Wei Lei †, Isaac K. Sundar and Irfan Rahman* Department of Environmental Medicine, University of Rochester Medical Center, Rochester, NY, USA Antagonism of CXCR2 receptors, predominately located on neutrophils and critical for their immunomodulatory activity, is an attractive pharmacological therapeutic approach aimed at reducing the potentially damaging effects of heightened neutrophil inﬂux into the lung. The role CXCR2 in lung inﬂammation in response to cigarette smoke (CS) inhalation using the mutant mouse approach is not known. We hypothesized that genetic ablation of CXCR2 would protect mice against CS-induced inﬂammation and DNA damage response. We used CXCR2−/−deﬁcient/mutant (knock-out, KO) mice, and assessed the changes in critical lung inﬂammatory NF-κB-driven chemokines released from the parenchyma of CS-exposed mice. The extent of tissue damage was assessed by the number of DNA damaging γH2AX positive cells. CXCR2 KO mice exhibited protection from heightened levels of neutrophils measured in BALF taken from mice exposed to CS. IL-8 (KC mouse) levels in the BALF from CS-exposed CXCR2 KO were elevated compared to WT. IL-6 levels in BALF were refractory to increase by CS in CXCR2 KO mice. There were no signiﬁcant changes to MIP-2, MCP-1, or IL-1β. Total levels of NF-κB were maintained at lower levels in CS-exposed CXCR2 KO mice compared to WT mice exposed to CS. Finally, CXCR2 KO mice were protected from lung cells positive for DNA damage response and senescence marker γH2AX. CXCR2 KO mice are protected from heightened inﬂammatory response mediated by increased neutrophil response as a result of acute 3 day CS exposure. This is also associated with changes in pro-inﬂammatory chemokines and reduced incursion of γH2AX indicating CXCR2 deﬁcient mice are protected from lung injury. Thus, CXCR2 may be a pharmacological target in setting of inﬂammation and DNA damage in the pathogenesis of COPD. Ethics Statement and Scientiﬁc Rigor/Reproducibility g Mice were injected intraperitoneally with 100 mg/kg body weight of pentobarbiturate (Abbott laboratories, Abbott Park, IL) and then sacriﬁced by exsanguination. The lungs were lavaged 3 times with 0.6 ml of saline via a cannula inserted into the trachea. The aliquots were combined, centrifuged, and the BAL inﬂammatory cell pellet was resuspended in saline. The total cell number was determined with a hemocytometer, and cytospin slides (Thermo Shandon, Pittsburgh, PA) were prepared using 50,000 cells per slide. Diﬀerential cell counts (∼500 cells/slide) were performed on cytospin-prepared slides stained with Diﬀ- Quik (Dade Behring, Newark, DE). All experiments for animal studies were performed in accordance with the standards established by the United States Animal Welfare Act, as set forth by the National Institutes of Health guidelines. The research protocol for mouse studies was approved by the University Committee on Animal Research Committee of the University of Rochester. We used a rigorous/robust and unbiased approach throughout the experimental plans (e.g., in vivo mouse model) and during data analysis so as to ensure that our data are reproducible along with full and detailed reporting of both methods and raw/analyzed data. All the key biological and/or chemical resources that are used in this study were validated and authenticated (methods and resources) and are of scientiﬁc standard from commercial sources. Our results adhere to NIH standards of reproducibility and scientiﬁc rigor. Citation: Cigarette smoke (CS) is predominately the driving factor in the etiology of chronic obstructive pulmonary disease (COPD). COPD is characterized by destruction of alveolar wall, inﬂammatory response, and premature lung aging or cellular senescence (Nyunoya et al., 2006; Moriyama et al., 2010; Yao et al., 2012; Ahmad et al., 2015). Pro-inﬂammatory mediators such as IL-8 (mouse KC) and MIP-2, act as CXCR2 ligands, which are critical for recruitment of peripheral neutrophils. Elevated neutrophil exudate found in bronchoalveolar lavage ﬂuid (BALF) or neutrophils observed Lerner CA, Lei W, Sundar IK and Rahman I (2016) Genetic Ablation of CXCR2 Protects against Cigarette Smoke-Induced Lung Inﬂammation and Injury. Front. Pharmacol. 7:391. doi: 10.3389/fphar.2016.00391 October 2016 \| Volume 7 \| Article 391 Frontiers in Pharmacology \| www.frontiersin.org CXCR2 in Smoke-Induced Lung Inﬂammation Lerner et al. amongst the parenchyma in histological sections are a feature of COPD and may contribute to tissue destruction due to lack of inﬂammatory resolution. The CXCR2 chemokine receptor is diﬀerentially expressed on the surface of certain myeloid and lymphoid cell types and plays a major role in peripheral neutrophil mediated inﬂammation. Chemotactic cytokines that bind to CXC family receptors mediate recruitment of the immune cells expressing them toward the injured or infected tissue (Belperio et al., 2002; Wareing et al., 2007; Nagarkar et al., 2009). The CXCR2−/−deﬁcient/mutant (knock-out, KO) mouse is defective in neutrophil function and exhibits reduced neutrophil lung inﬁltration following infection, injury, and exposure to ozone (Johnston et al., 2005; Reutershan et al., 2006). Thus, targeting of CXCR2 has been sought after as a potential therapy to quell inﬂammation in response to lung injury and infection in order to reduce the potential for excessive tissue damage mediated by neutrophil inﬂammation (Lomas- Neira et al., 2004; Nomellini et al., 2008; Zarbock et al., 2008; Russo et al., 2009; Leaker et al., 2013). However, the role of CXCR2 in mediating neutrophil inﬂammatory response to CS is not well understood. We hypothesize exposure of lungs to CS lays the ground work for progression toward COPD through activation of CXCR2 and excessive recruitment of neutrophils. To elucidate the role of CXCR2 in neutrophil recruitment and lung damage in response to CS, we utilized CXCR2−/−mice in our experiments. Cytokine Analysis in Bronchoalveolar Lavage The level of proinﬂammatory mediators, such as the chemokine keratinocyte chemoattractant (KC), macrophage inﬂammatory protein 2 (MIP-2), interleukin 6 (IL-6), monocyte chemotatic protein (MCP)-1 and, interleukin 1 beta (IL-1β) in BAL ﬂuid were measured by enzyme-linked immunosorbent assay (ELISA) using respective dual-antibody kits (R&D Systems, Minneapolis, MN) according to the manufacturer’s instructions. The results were expressed as pg/ml. Citation: This model allowed us to further assess the CS mediated response of critical lung inﬂammatory cytokines, the γH2AX DNA damage signal in the CXCR2−/−background, and examine how NF-κB as a master regulator of lung inﬂammatory cytokines, is aﬀected by CS in the absence of CXCR2 expression in vivo. 12:12 light-dark (LD) cycle with lights on at 6 a.m. and fed with a regular diet and water ad libitum unless otherwise indicated. For CS exposure, mice were kept in a standard 12:12 (LD) cycle with lights on from 6 a.m. to 6 p.m. throughout the experiment. All of the procedures described in this study were approved by the University Committee on Animal Research at the University of Rochester, Rochester, NY. CS exposure Eight to ten weeks old mice were exposed to acute (3 days) CS using Baumgartner-Jaeger CSM2082i cigarette smoking machine (CH Technologies, Westwood, NJ) in the Inhalation Core Facility at the University of Rochester. For acute CS exposure, mice were placed in individual compartments of a wire cage, which was placed inside a closed plastic box connected to the smoke source. The smoke was generated from 3R4F research cigarettes containing 10.9 mg of total particulate matter (TPM), 9.4 mg of tar, and 0.726 mg of nicotine, and carbon monoxide 11.9 mg per cigarette (University of Kentucky, Lexington, KY). Mice received two 1-h exposures per day, 1 h apart, according to the Federal Trade Commission protocol (1 puﬀ/min of 2-s duration and 35 mL volume) for 3 days (acute exposure). Mainstream CS was diluted along with ﬁltered air and directed into the exposure chamber. Monitoring of CS exposure (TPM per cubic meter of air) was done in real time using a MicroDust Pro-aerosol monitor (Casella CEL, Bedford, UK) and veriﬁed daily by gravimetric sampling immediately after the exposure was completed. By adjusting the number of cigarettes used to produce smoke and the ﬂow rate of the dilution air, the concentration of smoke was set at a nominal value (∼300 mg/m3 TPM). Western Blot Analysis for NF-κB Levels in the Lungs e u gs Proteins (25 µg) from lung tissue homogenates, were separated on a 7.5% sodium dodecyl sulfate (SDS)-polyacrylamide gel, transferred onto nitrocellulose membranes (Amersham, Arlington Heights, IL), and blocked using 5% bovine serum albumin (BSA) for 1 h at room temperature. The membranes were then probed with NF-κB (sc-109, Santa Cruz, CA), NF-κB- Phospho-serine536 (sc-101752, Santa Cruz, CA), and GAPDH (sc-365062, Santa Cruz, CA) primary antibody (1:1000 dilution in 5% BSA in phosphate-buﬀered saline [PBS] containing 0.1% Tween 20) at 4◦C for overnight. After three 10-min washing steps, the membrane was probed with suitable secondary anti- rabbit, or anti-mouse, or anti-goat antibody (1:10,000 dilution in 5% BSA) linked to horseradish peroxidase for 1 h, and detected using the enhanced chemiluminescence method (Perkin Elmer, Waltham, MA) and images were taken with Bio-Rad ChemiDoc MP, Imaging system. Equal loading of the gel was determined by quantitation of protein as well as by reprobing the same membranes for GAPDH. Animals Male C57BL/6J (C57) and CXCR2 knockout/deﬁcient (referred to as CXCR2−/−or KO) mice were purchased from the Jackson Laboratory (Bar Harbor, ME). These mice were housed under a October 2016 \| Volume 7 \| Article 391 Frontiers in Pharmacology \| www.frontiersin.org 2 CXCR2 in Smoke-Induced Lung Inﬂammation Lerner et al. Neutrophil Inﬂux in CXCR2−/−Mouse Exposed to Acute Cigarette Smoke Neutrophil Inﬂux in CXCR2−/−Mouse Exposed to Acute Cigarette Smoke CXCR2−/−mice exposed with CS for 3 days are refractory to neutrophil inﬂux into the BALF. In contrast, wild type (WT) mice exposed to acute CS exhibit robust neutrophil transmigration into BAL ﬂuid in the lungs compared to WT air group. These results show that CS is a potent activator of neutrophil recruitment to the lung and the process is highly dependent on the expression of CXCR2 (Figure 1). Level of Pro-Inﬂammatory Cytokines in CXCR2−/−Mice Exposed to Acute Cigarette Smoke To assess if the absence of CXCR2 inﬂuences levels of pro- inﬂammatory cytokines that may play a role in neutrophil recruitment in response to acute CS exposure, the Cxcr2 ligands KC and MIP2 were measured in BALF. CXCR2−/−mice exposed to CS exhibit increased levels of KC compared to CXCR2−/−air group and WT air group. Levels of MIP2 are not signiﬁcantly aﬀected by CS in the WT and CXCR2−/−mice. IL-6 was signiﬁcantly increased in WT mice exposed to CS. However, in CXCR2−/−mice, IL-6 is resistant to increase by FIGURE 1 \| Elevated neutrophils in BALF via cigarette smoke are blocked in CXCR2−/−mice. Mice were exposed to acute CS [300 mg/m3; total particulate matter (TPM)] for 2 h per day for 3 days. Mice were sacriﬁced 24 h following last day of CS exposure. Data are shown as the mean ± SEM (WT-Air; n = 4, WT-CS; n = 5, CXCR2; KO-Air n = 3, CXCR2 KO-CS; n = 3). P < 0.01 signiﬁcant for WT-CS compared to WT-Air; ##P < 0.01 signiﬁcant for CXCR2 KO-CS compared to WT-CS. Immunohistochemistry for γH2AX Levels in the Lungs Immunostaining was performed on formalin-ﬁxed, paraﬃn- embedded lung tissue. Paraﬃn sections (4 µm thick) were deparaﬃnized and then rehydrated through series of xylene and graded ethanol. Antigen retrieval was performed by heating in citrate buﬀer (10 mM Citric acid, 0.05% Tween 20, pH 6.0). Primary antibody was incubated overnight at 4◦C with rabbit anti-γH2AX antibody (05-636, EMD Millipore, Darmstadt, Germany) Appropriate ﬂuorescently labeled secondary antibodies (FITC-conjugated anti-mouse 2◦ antibodies) were used to detect the immune complexes before tissues sections were counterstained with 4′,d-diamidino-2- phenylindole (dapi). FIGURE 1 \| Elevated neutrophils in BALF via cigarette smoke are blocked in CXCR2−/−mice. Mice were exposed to acute CS [300 mg/m3; total particulate matter (TPM)] for 2 h per day for 3 days. Mice were sacriﬁced 24 h following last day of CS exposure. Data are shown as the mean ± SEM (WT-Air; n = 4, WT-CS; n = 5, CXCR2; KO-Air n = 3, CXCR2 KO-CS; n = 3). P < 0.01 signiﬁcant for WT-CS compared to WT-Air; ##P < 0.01 signiﬁcant for CXCR2 KO-CS compared to WT-CS. Protein Extraction from Lung Tissues and Quantiﬁcation of the number of ﬂuorescent punctate nuclei present in γH2AX immunohistochemistry, and densitometry of Western blot analysis. These results are shown as the mean ± SEM. ∗P < 0.05, ∗∗P <0.01, ∗∗∗P <0.001 which were considered as statistically signiﬁcant. One lobe of the lung tissue (∼50 mg) was homogenized (Pro 200 homogenizer, at maximum speed, 5th gear for 40 s) in 0.5 mL of ice-cold RIPA buﬀer containing complete protease inhibitor cocktail (Sigma). The tissue homogenate was then incubated on ice for 45 min to allow total cell lysis. The homogenate was then centrifuged at 13,000 × g for 5 min at 4◦C to separate the protein fraction from the cell/tissue debris. The supernatant containing protein was aliquoted and stored at −80◦C for Western blotting. This fraction was taken for protein analysis by bicinchoninic acid (BCA) colorimetric assay (Thermo Scientiﬁc, Rockford, IL) using BSA as a standard. DNA Damage Signaling in CXCR2−/− Mouse Lung Lung tissue is prone to DNA damage by exposure to CS. In addition, since neutrophil inﬂux is suggested to contribute lung DNA damage through inducing genotoxic stress (van Berlo et al., 2010), we sought to determine if CXCR2 status, in its ability to inﬂuence CS mediated neutrophil inﬂux into BALF (Figure 1) might aﬀect DNA damage response, which also intersects inﬂammatory signaling pathways (McCool and Miyamoto, 2012). In lung tissue sections from mice exposed to acute CS, WT mice exhibit an increased number of γH2AX positive cells by immunoﬂuorescence compared to air group. The CXCR2−/−mouse lung was highly devoid of γH2AX positive cells following CS exposure and exhibited only a marginal increase in γH2AX compared to CXCR2−/−air group (Figures 4A,B). These data indicate the γH2AX DNA damage signal is associated with increased BALF neutrophils in response to CS exposure, which are both dependent on expression of CXCR2. Statistical Analysis Statistical analysis of signiﬁcance was calculated using one-way analysis of variance (ANOVA) followed by Tukey’s post-hoc test for multigroup comparisons using the StatView software or GraphPad Prism. Image J software (Version 1.47, National Institutes of Health, Bethesda, MD) was used for quantiﬁcation Frontiers in Pharmacology \| www.frontiersin.org October 2016 \| Volume 7 \| Article 391 3 CXCR2 in Smoke-Induced Lung Inﬂammation Lerner et al. WT or CXCR2−/−mice (data not shown). These results suggest that by ablating CXCR2, NF-κB expression in response to CS exposure is attenuated. WT or CXCR2−/−mice (data not shown). These results suggest that by ablating CXCR2, NF-κB expression in response to CS exposure is attenuated. CS. Levels of MCP-1 and IL-1β in CS-exposed CXCR2−/−was not signiﬁcantly diﬀerent compared to CS or air exposed WT mice and this was conﬁrmed for IL-1β in plasma (Figure 2). We conclude key cytokines IL-6 and KC both involved in mediating inﬂammation and neutrophil activity in response to acute CS exposure are altered due to the absence of CXCR2. Lack of neutrophil inﬂux in response to CS results in enhanced KC and blunted IL-6. NF-κB Expression in CXCR2−/−Mouse Lung Homogenates Exposed to Acute Cigarette Smoke To determine if NF-κB is aﬀected in CXCR2−/−mouse lung as it is a master regulator of IL-6, we prepared whole lung homogenates after acute air and CS exposure and measured both the relative levels of total NF-κB and phospho-Ser536 on NF-κB to assess its activation. Following CS exposure in WT mice, total NF-κB levels are increased compared to air group. Conversely, CS is not able to induce lung NF-κB expression to similar levels in CXCR2−/−mice compared to WT CS-exposed mice (Figures 3A,B). Levels of Ser536 phospho-NF-κB relative to total NF-κB remained unchanged between air and CS-exposed FIGURE 2 \| The levels of proinﬂammatory mediators in BALF of WT and CXCR2−/−after acute CS exposure. Mice were exposed to acute CS [300 mg/m3; total particulate matter (TPM)] for 2 h per day for 3 days. The levels of proinﬂammatory mediators in BALF; MCP-1, IL-6, IL-1β, KC, and MIP-2 after 3 days CS exposure and IL-1β in plasma were measured by ELISA. Data are shown as the mean ± SEM (WT-Air; n = 3–4, WT-CS; n = 5, CXCR2 KO-Air; n = 3, CXCR2 KO-CS; n = 3). P < 0.05, P < 0.01 signiﬁcant for WT-CS compared to WT-Air; #P < 0.05 signiﬁcant for CXCR2 KO-CS compared to WT-CS. FIGURE 2 \| The levels of proinﬂammatory mediators in BALF of WT and CXCR2−/−after acute CS exposure. Mice were exposed to acute CS [300 mg/m3; total particulate matter (TPM)] for 2 h per day for 3 days. The levels of proinﬂammatory mediators in BALF; MCP-1, IL-6, IL-1β, KC, and MIP-2 after 3 days CS exposure and IL-1β in plasma were measured by ELISA. Data are shown as the mean ± SEM (WT-Air; n = 3–4, WT-CS; n = 5, CXCR2 KO-Air; n = 3, CXCR2 KO-CS; n = 3). P < 0.05, *P < 0.01 signiﬁcant for WT-CS compared to WT-Air; #P < 0.05 signiﬁcant for CXCR2 KO-CS compared to WT-CS. October 2016 \| Volume 7 \| Article 391 Frontiers in Pharmacology \| www.frontiersin.org 4 CXCR2 in Smoke-Induced Lung Inﬂammation Lerner et al. FIGURE 3 \| NF-κB levels in WT and CXCR2−/−mice after acute CS exposure. Mice were exposed to acute CS [300 mg/m3; total particulate matter (TPM)] for 2 h per day for 3 days. Total proteins isolated from lung homogenates were resolved on SDS-PAGE gel for immunoblotting. (A) Immunoblot of total NF-κB. GAPDH was used as a housekeeping control. NF-κB Expression in CXCR2−/−Mouse Lung Homogenates Exposed to Acute Cigarette Smoke P < 0.05 signiﬁcant for WT-CS compared to WT-Air; #P < 0.05 signiﬁcant for CXCR2 KO-CS compared to WT-CS. FIGURE 3 \| NF-κB levels in WT and CXCR2−/−mice aft IL-β is a potent inﬂammatory mediator. Stevenson et al. observed a time-dependent increase in IL-1β in BALF in rat model at 24 h post-CS exposure (Stevenson et al., 2005). In our mice, 3 days acute CS exposure did not elicit an appreciable increase in IL-1β in the lavage ﬂuid in WT mice and there were no remarkable changes in the CXCR2−/−mice exposed to air or CS. We cannot conclude for certain IL-β response is unaﬀected by CS or deﬁciency in CXCR2, rather the acute exposure we employed in this study, may lead to acclimation or impairment for certain inﬂammatory targets. NF-κB Expression in CXCR2−/−Mouse Lung Homogenates Exposed to Acute Cigarette Smoke Immunoblots are representative of 2 independent experiments. (B) Densitometry for quantitation of relative differences in band intensity for total NF-κB normalized to GAPDH in (A). Measurements are shown as arbitrary units (A.U.). Data are shown as the mean ± SEM. P < 0.05 signiﬁcant for WT-CS compared to WT-Air; #P < 0.05 signiﬁcant for CXCR2 KO-CS compared to WT-CS. complex dynamics during heightened tissue inﬂammation. CXCR2 internalization in addition to canonical G-protein coupled receptor desensitization mechanisms and autocrine feedback responses may further limit excessive immunological activity on cells expressing CXC family receptors, a dynamic that requires further investigation. Although IL-6 is not a CXCR2 ligand, the CXCR2−/−mice when exposed to either air or CS exhibited reduced IL-6 in the BALF which may extend anti-inﬂammatory eﬀects beyond limited neutrophil recruitment. An interesting dynamic between IL-6 expression and CXCR2 activity is beginning to emerge in studies involving modulating CXC type receptor activity. The CXCR2 antagonist SCH-N is predominantly a CXCR2 antagonist with partial CXCR1 antagonism in primates and almost exclusively a CXCR2 antagonist in rodent (Chapman et al., 2007). Though SCH-N CXCR2 antagonism did not appreciably aﬀect CS-exposed mouse BALF IL-6 levels (Thatcher et al., 2005), mice treated with a nasal instillation of barn dust in conjunction with the dual CXCR1/CXCR2 antagonist CXCL8 (3-74) K11R/G31P exhibited similar results to CXCR2−/−mice BALF IL-6 levels (Schneberger et al., 2015). Thus, it is possible dual CXCR1/2 antagonism is aﬀecting IL-6 regulation and requires more extensive characterization of how other CXC receptors are aﬀected in the CXCR2−/−mice models. The interesting eﬀect of CXCR2 antagonism by antileukinate on PMN extravasation to the lung under septic challenge leads to reduced IL-6 plasma levels (Lomas-Neira et al., 2004). Thus, further assessment will require optimizing therapeutic approaches for CXC receptor family targeting diﬀerent inﬂammatory insults in context with cytokine tissue locale. FIGURE 3 \| NF-κB levels in WT and CXCR2−/−mice after acute CS exposure. Mice were exposed to acute CS [300 mg/m3; total particulate matter (TPM)] for 2 h per day for 3 days. Total proteins isolated from lung homogenates were resolved on SDS-PAGE gel for immunoblotting. (A) Immunoblot of total NF-κB. GAPDH was used as a housekeeping control. Immunoblots are representative of 2 independent experiments. (B) Densitometry for quantitation of relative differences in band intensity for total NF-κB normalized to GAPDH in (A). Measurements are shown as arbitrary units (A.U.). Data are shown as the mean ± SEM. Frontiers in Pharmacology \| www.frontiersin.org DISCUSSION The increased neutrophilic environment induced by CS may contribute to COPD etiology which is typically associated with COPD. COPD is frequently accompanied by bacterial and viral infections that further exacerbate symptoms and accelerate pathogenesis. We utilized CXCR2−/−mice to assess neutrophil recruitment, pro-inﬂammatory cytokines, NF-κB activity, and the DNA damage signal γH2AX in mouse lungs exposed to acute CS. MIP-2 and KC are primarily CXCR2 ligands. In our CXCR2−/−mice exposed to acute CS, KC levels were signiﬁcantly altered. The up-regulation of KC in CXCR2−/− mice exposed to CS is consistent with a previous report which showed inhibition of CS-induced lung inﬂammation by a CXCR2 antagonist (Thatcher et al., 2005), though no gene deletion approach was used without implicating DNA damaging response. CXCR2 is desensitized by very high levels of KC, which may allude to its regulation under inﬂammatory states where KC is potentially internalized (Wiekowski et al., 2001; Rose et al., 2004). We indeed observed KC increase in CS-exposed WT mice. However, in the CXCR2−/−background, we ﬁnd KC expression increases further in response to CS which may allude to the possibility that KC levels are regulated by receptor-ligand The major role NF-κB plays in modulating inﬂammation in response to injury and infection in multiple tissues was of interest to us in the CXCR2−/−mice exposed to CS, particularly since we observed diﬀerential regulation in KC and IL-6. In NF-κB deﬁcient mice, CXCR2 mediated neutrophil inﬂux into the lung, is reported to be enhanced indicating a regulatory interaction (though potentially indirect) may exist between CXCR2 and NF- κB (von Vietinghoﬀet al., 2010). Prostate cancer cells cultured under hypoxia, also appear to require NF-κB to upregulate CXCR2 RNA (Maxwell et al., 2007). The CXCR2−/−in our study appears to have a reduced capacity to express total levels of NF- κB, however, the activity of NF-κB indicated by phosphorylation of Serine 536 shows it retains competency in its ability to modulate signaling eﬀects despite reduced total expression levels. The γH2AX DNA damage signal is frequently associated with senescent positive cell. CXCR2 signaling essentially promotes senescence which is hypothesized to be a factor in the Frontiers in Pharmacology \| www.frontiersin.org October 2016 \| Volume 7 \| Article 391 5 CXCR2 in Smoke-Induced Lung Inﬂammation Lerner et al. FIGURE 4 \| Assessment of γH2AX positive cells in lung tissues from WT and CXCR2−/−mice after acute CS exposure. DISCUSSION Mice were exposed to acute CS [300 mg/m3; total particulate matter (TPM)] for 2 h per day for 3 days. Parafﬁn embedded lung tissue sections from 24 h following last day of CS exposure were used for immunohistochemistry. (A) Representative immunoﬂuorescent images at 20x showing γH2AX positive nuclei (green) overlaid with DAPI stained nuclei (Blue). White arrows point to γH2AX positive regions. (B) Quantitation of the number of γH2AX positive nuclei per image for each condition. Data are shown as the mean ± SEM (WT-Air; n = 14, WT-CS; n = 15, CXCR2 KO-Air; n = 10, CXCR2 KO-CS; n = 5). P < 0.01 signiﬁcant for WT-CS compared to WT-Air; P < 0.001 signiﬁcant for CXCR2 KO-CS compared to CXCR2 KO-Air; ###P < 0.001 signiﬁcant for CXCR2 KO-CS compared to WT-CS. FIGURE 4 \| Assessment of γH2AX positive cells in lung tissues from WT and CXCR2−/−mice after acute CS exposure. Mice were exposed to acute CS [300 mg/m3; total particulate matter (TPM)] for 2 h per day for 3 days. Parafﬁn embedded lung tissue sections from 24 h following last day of CS exposure were used for immunohistochemistry (A) Representative immunoﬂuorescent images at 20x showing γH2AX positive nuclei (green) overlaid with DAPI stained nuclei (Blue) White FIGURE 4 \| Assessment of γH2AX positive cells in lung tissues from WT and CXCR2−/−mice after acute CS exposure. Mice were exposed to acute CS [300 mg/m3; total particulate matter (TPM)] for 2 h per day for 3 days. Parafﬁn embedded lung tissue sections from 24 h following last day of CS exposure were used for immunohistochemistry. (A) Representative immunoﬂuorescent images at 20x showing γH2AX positive nuclei (green) overlaid with DAPI stained nuclei (Blue). White arrows point to γH2AX positive regions. (B) Quantitation of the number of γH2AX positive nuclei per image for each condition. Data are shown as the mean ± SEM (WT-Air; n = 14, WT-CS; n = 15, CXCR2 KO-Air; n = 10, CXCR2 KO-CS; n = 5). P < 0.01 signiﬁcant for WT-CS compared to WT-Air; ***P < 0.001 signiﬁcant for CXCR2 KO-CS compared to CXCR2 KO-Air; ###P < 0.001 signiﬁcant for CXCR2 KO-CS compared to WT-CS. pathogenesis of COPD and lung cancer (Acosta et al., 2008). The γH2AX signal is one facet of DNA damage response (DDR), and is frequently used as a corollary to other senescence assays. REFERENCES short-term cigarette smoke exposure. Acta Pharmacol. Sin. 29, 1432–1439. doi: 10.1111/j.1745-7254.2008.00899.x Acosta, J. C., O’Loghlen, A., Banito, A., Guijarro, M. V., Augert, A., Raguz, S., et al. (2008). Chemokine signaling via the CXCR2 receptor reinforces senescence. Cell 133, 1006–1018. doi: 10.1016/j.cell.2008.03.038 Nie, L., Xiang, R., Zhou, W., Lu, B., Cheng, D., and Gao, J. (2008b). Attenuation of acute lung inﬂammation induced by cigarette smoke in CXCR3 knockout mice. Respir. Res. 9:82. doi: 10.1186/1465-9921-9-82 p Nomellini, V., Faunce, D. E., Gomez, C. R., and Kovacs, E. J. (2008). An age- associated increase in pulmonary inﬂammation after burn injury is abrogated Ahmad, T., Sundar, I. K., Lerner, C. A., Gerloﬀ, J., Tormos, A. M., Yao, H., et al. (2015). Impaired mitophagy leads to cigarette smoke stress-induced cellular senescence: implications for chronic obstructive pulmonary disease. FASEB J. 29, 2912–2929. doi: 10.1096/fj.14-268276 Nomellini, V., Faunce, D. E., Gomez, C. R., and Kovacs, E. J. (2008). An age- associated increase in pulmonary inﬂammation after burn injury is abrogated by CXCR2 inhibition. J. Leukoc. Biol. 83, 1493–1501. doi: 10.1189/jlb.1007672 associated increase in pulmonary inﬂammation after burn injury is abrogated by CXCR2 inhibition. J. Leukoc. Biol. 83, 1493–1501. doi: 10.1189/jlb.1007672 by CXCR2 inhibition. J. Leukoc. Biol. 83, 1493–1501. doi: 10.1189/jlb.1007672 Nyunoya, T., Monick, M. M., Klingelhutz, A., Yarovinsky, T. O., Cagley, J. R., and Hunninghake, G. W. (2006). Cigarette smoke induces cellular senescence. Am. J. Respir. Cell Mol. Biol. 35, 681–688. doi: 10.1165/rcmb.2006-0169OC Belperio, J. A., Keane, M. P., Burdick, M. D., Londhe, V., Xue, Y. Y., Li, K., et al. (2002). Critical role for CXCR2 and CXCR2 ligands during the pathogenesis of ventilator-induced lung injury. J. Clin. Invest. 110, 1703–1716. doi: 10.1172/JCI0215849 Respir. Cell Mol. Biol. 35, 681–688. doi: 10.1165/rcmb.2006-0169O J. Respir. Cell Mol. Biol. 35, 681–688. doi: 10.1165/rcmb.2006-0169OC Reutershan, J., Morris, M. A., Burcin, T. L., Smith, D. F., Chang, D., Saprito, M. S., et al. (2006). Critical role of endothelial CXCR2 in LPS-induced neutrophil migration into the lung. J. Clin. Invest. 116, 695–702. doi: 10.1172/JCI27009 Reutershan, J., Morris, M. A., Burcin, T. L., Smith, D. F., Chang, D., Saprito, M. Reutershan, J., Morris, M. A., Burcin, T. L., Smith, D. F., Chang, D., Saprito, M. S., et al. (2006). Critical role of endothelial CXCR2 in LPS-induced neutrophil migration into the lung. J. Clin. Invest. 116, 695–702. doi: 10.1172/JCI27009 Chapman, R. W., Minnicozzi, M., Celly, C. S., Phillips, J. E., Kung, T. T., Hipkin, R. REFERENCES W., et al. (2007). A novel, orally active CXCR1/2 receptor antagonist, Sch527123, inhibits neutrophil recruitment, mucus production, and goblet cell hyperplasia in animal models of pulmonary inﬂammation. J. Pharmacol. Exp. Ther. 322, 486–493. doi: 10.1124/jpet.106.119040 Rose, J. J., Foley, J. F., Murphy, P. M., and Venkatesan, S. (2004). On the mechanism and signiﬁcance of ligand-induced internalization of human neutrophil chemokine receptors CXCR1 and CXCR2. J. Biol. Chem. 279, 24372–24386. doi: 10.1074/jbc.M401364200 Russo, R. C., Guabiraba, R., Garcia, C. C., Barcelos, L. S., Roﬀê, E., Souza, A. L., et al. (2009). Role of the chemokine receptor CXCR2 in bleomycin-induced pulmonary inﬂammation and ﬁbrosis. Am. J. Respir. Cell Mol. Biol. 40, 410–421. doi: 10.1165/rcmb.2007-0364OC Eash, K. J., Greenbaum, A. M., Gopalan, P. K., and Link, D. C. (2010). CXCR2 and CXCR4 antagonistically regulate neutrophil traﬃcking from murine bone marrow. J. Clin. Invest. 120, 2423–2431. doi: 10.1172/JCI41649 Guo, H., Liu, Z., Xu, B., Hu, H., Wei, Z., Liu, Q., et al. (2013). Chemokine receptor CXCR2 is transactivated by p53 and induces p38-mediated cellular senescence in response to DNA damage. Aging Cell. 12, 1110–1121. doi: 10.1111/acel.12138 Schneberger, D., Gordon, J. R., DeVasure, J. M., Boten, J. A., Heires, A. J., Romberger, D. J., et al. (2015). CXCR1/CXCR2 antagonist CXCL8(3- 74)K11R/G31P blocks lung inﬂammation in swine barn dust-instilled mice. Pulm. Pharmacol. Ther. 31, 55–62. doi: 10.1016/j.pupt.2015.02.002 Johnston, R. A., Mizgerd, J. P., and Shore, S. A. (2005). CXCR2 is essential for maximal neutrophil recruitment and methacholine responsiveness after ozone exposure. Am. J. Physiol. Lung Cell. Mol. Physiol. 288, L61–L67. doi: 10.1152/ajplung.00101.2004 Stevenson, C. S., Coote, K., Webster, R., Johnston, H., Atherton, H. C., Nicholls, A., et al. (2005). Characterization of cigarette smoke-induced inﬂammatory and mucus hypersecretory changes in rat lung and the role of CXCR2 ligands in mediating this eﬀect. Am. J. Physiol. Lung Cell. Mol. Physiol. 288, L514–L522. doi: 10.1152/ajplung.00317.2004 Leaker, B. R., Barnes, P. J., and O’Connor, B. (2013). Inhibition of LPS-induced airway neutrophilic inﬂammation in healthy volunteers with an oral CXCR2 antagonist. Respir. Res. 14:137. doi: 10.1186/1465-9921-14-137 Thatcher, T. H., McHugh, N. A., Egan, R. W., Chapman, R. W., Hey, J. A., Turner, C. K., et al. (2005). Role of CXCR2 in cigarette smoke-induced lung inﬂammation. Am. J. Physiol. Lung Cell. Mol. Physiol. 289, L322–L328. doi: 10.1152/ajplung.00039.2005 Lomas-Neira, J. L., Chung, C. S., Grutkoski, P. S., Miller, E. J., and Ayala, A. (2004). DISCUSSION The hypothesis that COPD etiology integrates accumulation of senescent cells through aging, smoking, or a combination of both explains that the senescent associated secretory phenotype (SASP) contributes to onset or progression of COPD. CXCR2 has been shown to be critical for senescence as many of the SASP factors are CXCR2 ligands (Acosta et al., 2008). More recently CXCR2 is shown to be involved in DDR mediated senescence (Guo et al., 2013). Our results along with others indicate DDR is engaged within the WT lung parenchyma by CS exposure as indicated by increased γH2AX, increased p53 activity, and changes to CXCR2 dynamics and their ligands (Tiwari et al., 2016). The CXCR2-DDR-senescence connection is intriguing and our results show CXCR2 deﬁciency may protect from DDR in response to acute CS exposure. However, it is not clear how targeting this axis inﬂuences the pathogenesis of COPD (along with overcoming steroid resistance) as it is diﬃcult to target CXCR2 while retaining critical host defense and tissue repair capacity. Other CXC family members such as CXCR3 have also been shown to attenuate CS mediated lung inﬂammation in a CXCR3−/−mouse model by reducing CD8+ T cell toxicity which further suggests targeting of multiple CXC family members may be an approach to optimize in treatment of inﬂammatory lung diseases (Nie et al., 2008a,b). In the CXCR2−/−background, neutrophils are further impaired from extravasation from bone marrow into circulation which underscores the severity of the genetic deﬁcit (Eash et al., 2010). In contrast, the pharmacological targeting and new precise genetic methods such as CRISPR/Cas9 models of CXCR2 allows direct oversight into dose, location of administration, protein activity, and timing which will also help to determine how chemokine October 2016 \| Volume 7 \| Article 391 Frontiers in Pharmacology \| www.frontiersin.org 6 CXCR2 in Smoke-Induced Lung Inﬂammation Lerner et al. the data; CL, WL, IS, and IR wrote and revised/edited the manuscript. proﬁles are aﬀected by CXCR2 antagonism and which are most therapeutic while retaining host immune defenses. Nevertheless, our data show that CXCR2 may be a pharmacological target in setting of inﬂammation and DNA damage in the pathogenesis of COPD. ACKNOWLEDGMENTS We thank Ms. Janice Gerloﬀfor her technical assistance. This study was supported by the NIH 1R01HL085613 (to IR), American Lung Association RG-305393 (to IS), pulmonary training grant T32 HL066988, and NIEHS Environmental Health Science Center grant P30-ES01247. AUTHOR CONTRIBUTIONS CL, WL, IS, and IR conceived and designed the experiments; WL, IS, and CL performed the experiments; IS and WL analyzed Wiekowski, M. T., Chen, S. C., Zalamea, P., Wilburn, B. P., Kinsley, D. J., Sharif, W. W., et al. (2001). Disruption of neutrophil migration in a conditional transgenic model: evidence for CXCR2 desensitization in vivo. J. Immunol. 167, 7102–7110. doi: 10.4049/jimmunol.167.12.7102 Zarbock, A., Allegretti, M., and Ley, K. (2008). Therapeutic inhibition of CXCR2 by Reparixin attenuates acute lung injury in mice. Br. J. Pharmacol. 155, 357–364. doi: 10.1038/bjp.2008.270 Yao, H., Chung, S., Hwang, J. W., Rajendrasozhan, S., Sundar, I. K., Dean, D. A., et al. (2012). SIRT1 protects against emphysema via FOXO3-mediated reduction of premature senescence in mice. J. Clin. Invest. 122, 2032–2045. doi: 10.1172/JCI60132 Frontiers in Pharmacology \| www.frontiersin.org REFERENCES CXCR2 inhibition suppresses hemorrhage-induced priming for acute lung injury in mice. J. Leukoc. Biol. 76, 58–64. doi: 10.1189/jlb. 1103541 Maxwell, P. J., Gallagher, R., Seaton, A., Wilson, C., Scullin, P., Pettigrew, J., et al. (2007). HIF-1 and NF-κB-mediated upregulation of CXCR1 and CXCR2 expression promotes cell survival in hypoxic prostate cancer cells. Oncogene 26, 7333–7345. doi: 10.1038/sj.onc.1210536 Tiwari, N., Marudamuthu, A. S., Tsukasaki, Y., Ikebe, M., Fu, J., and Shetty, S. (2016). p53- and PAI-1-mediated induction of C-X-C chemokines and CXCR2: importance in pulmonary inﬂammation due to cigarette smoke exposure. Am. J. Physiol. Lung Cell. Mol. Physiol. 310, L496–L506. doi: 10.1152/ajplung.00290.2015 McCool, K. W., and Miyamoto, S. (2012). DNA damage-dependent NF-κB activation: NEMO turns nuclear signaling inside out. Immunol. Rev. 246, 311–326. doi: 10.1111/j.1600-065X.2012.01101.x van Berlo, D., Wessels, A., Boots, A. W., Wilhelmi, V., Scherbart, A. M., Gerloﬀ, K., et al. (2010). Neutrophil-derived ROS contribute to oxidative DNA damage induction by quartz particles. Free Radic. Biol. Med. 49, 1685–1693. doi: 10.1016/j.freeradbiomed.2010.08.031 Moriyama, C., Betsuyaku, T., Ito, Y., Hamamura, I., Hata, J., Takahashi, H., et al. (2010). Aging enhances susceptibility to cigarette smoke-induced inﬂammation through bronchiolar chemokines. Am. J. Respir. Cell Mol. Biol. 42, 304–311. doi: 10.1165/rcmb.2009-0025OC von Vietinghoﬀ, S., Asagiri, M., Azar, D., Hoﬀmann, A., and Ley, K. (2010). Defective regulation of CXCR2 facilitates neutrophil release from bone marrow causing spontaneous inﬂammation in severely NF-κB-deﬁcient mice. J. Immunol. 185, 670–678. doi: 10.4049/jimmunol.1000339 Nagarkar, D. R., Wang, Q., Shim, J., Zhao, Y., Tsai, W. C., Lukacs, N. W., et al. (2009). CXCR2 is required for neutrophilic airway inﬂammation and hyperresponsiveness in a mouse model of human rhinovirus infection. J. Immunol. 183, 6698–6707. doi: 10.4049/jimmunol.0900298 Wareing, M. D., Shea, A. L., Inglis, C. A., Dias, P. B., and Sarawar, S. R. (2007). CXCR2 is required for neutrophil recruitment to the lung during inﬂuenza virus infection, but is not essential for viral clearance. Viral Immunol. 20, 369–378. doi: 10.1089/vim.2006.0101 Nie, L., Xiang, R. L., Liu, Y., Zhou, W. X., Jiang, L., Lu, B., et al. (2008a). Acute pulmonary inﬂammation is inhibited in CXCR3 knockout mice after October 2016 \| Volume 7 \| Article 391 Frontiers in Pharmacology \| www.frontiersin.org 7 CXCR2 in Smoke-Induced Lung Inﬂammation CXCR2 in Smoke-Induced Lung Inﬂammation Lerner et al. Conﬂict of Interest Statement: The authors declare that the research was conducted in the absence of any commercial or ﬁnancial relationships that could be construed as a potential conﬂict of interest. Copyright © 2016 Lerner, Lei, Sundar and Rahman. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Zarbock, A., Allegretti, M., and Ley, K. (2008). Therapeutic inhibition of CXCR2 by Reparixin attenuates acute lung injury in mice. Br. J. Pharmacol. 155, 357–364. doi: 10.1038/bjp.2008.270 October 2016 \| Volume 7 \| Article 391 Frontiers in Pharmacology \| www.frontiersin.org 8
https://openalex.org/W2801681446	https://europepmc.org/articles/pmc6005818?pdf=render	English	null	Adipogenic Differentiation of Muscle Derived Cells is Repressed by Inhibition of GSK-3 Activity	Frontiers in veterinary science	2,018	cc-by	4,566	Edited by: Fausto Cremonesi, Università degli Studi di Milano, Italy Edited by: Fausto Cremonesi, Università degli Studi di Milano, Italy Reviewed by: So-ichiro Fukada, Osaka University, Japan Ruchi Sharma, Stemnovate ltd, United Kingdom Correspondence: Paul Loughna Paul.loughna@nottingham.ac.uk Reviewed by: So-ichiro Fukada, Osaka University, Japan Ruchi Sharma, Stemnovate ltd, United Kingdom Conclusions: Studies in vivo have suggested that the Wnt pathway is a major regulator of whole body adiposity. In this study we have shown that the ability of cells derived from porcine skeletal muscle to differentiate along an adipogenic lineage, in vitro, is severely impaired by mimicking the action of this pathway. This was done by inactivation of GSK- 3β by the use of Lithium Chloride. Specialty section: This article was submitted to Veterinary Regenerative Medicine, a section of the journal Frontiers in Veterinary Science Adipogenic Differentiation of M Derived Cells is Repressed by Inhibition of GSK-3 Activity Zoe Redshaw 1,2 and Paul Thomas Loughna 1 1 School of Veterinary Medicine and Science, The University of Nottingham, Loughborough, United Kingdom, 2 Faculty of Health and Life Sciences, De Montfort University, Leicester, United Kingdom Zoe Redshaw 1,2 and Paul Thomas Loughna 1* 1 School of Veterinary Medicine and Science, The University of Nottingham, Loughborough, United Kingdom, 2 Faculty of Health and Life Sciences, De Montfort University, Leicester, United Kingdom Background: Intramuscular fat is important in large animal livestock species in regard to meat quality and in humans is of clinical significance in particular in relation to insulin resistance. The canonical Wnt signalling pathway has been implicated at a whole body level in regulating relative levels of adiposity versus lean body mass. Previously we have shown that pig muscle cells can undergo adipogenic differentiation to a degree that is dependent upon the specific muscle source. In this work we examine the role of the canonical Wnt pathway which acts through inactivation of glycogen synthase kinase-3 (GSK-3) in the regulation of adipogenic differentiation in muscle cells derived from the pig semimembranosus muscle. Results: The application of lithium chloride to muscle derived cells significantly increased the phosphorylation of GSK-3β and thus inhibited its activity thus mimicking Wnt signaling. This was associated with a significant decrease in the expression of the adipogenic transcription factor PPARγ and an almost complete inhibition of adipogenesis in the cells. The data also suggest that GSK-3α plays, at most, a small role in this process. Keywords: Wnt, skeletal muscle, GSK-3β, PPARγ, adipogenesis Introduction Skeletal muscle is the largest organ of the body comprising >45% of overall mass and is the major contributor to whole-body protein metabolism. This tissue also demonstrates remarkable plasticity and may be subject to rapid changes in mass. The size of multinucleate muscle fibers is dictated by a number of factors the predominant being nutrition and mechanical loading. Muscle fibers also have a remarkable ability to regenerate following trauma or disease. Both muscle hypertrophic growth and regenerative capacity are, at least in part, dependent upon resident stem cell populations. Received: 21 March 2017 Accepted: 04 May 2018 Published: 12 June 2018 Original Research published: 12 June 2018 doi: 10.3389/fvets.2018.00110 Original Research Semi-Quantitative RT-PCR Total RNA was isolated using the Qiagen mini prep kit (Qiagen, UK) and reverse transcription performed using superscript™ first strand synthesis system (Invitrogen) according to the manufacturers guidelines (input RNA: 500 ng). Samples were treated with DNA free dnase kit (Applied Biosystems, UK) and sample purification was ascertained via nanodrop, with all samples having a 260/280 OD ratio greater than 2.0. All primers were designed against known pig mRNA sequences (Ensembl Genome Browser) using free software Primer3 (http://frodo.wi.mit.edu/). Primer sequences: Wnt10b (111 bp) forward: AATGCGAATCCACAACAACA, reverse: CTCCAGCACGTCTTGAACTG. Amplification conditions as described in (8). Citation: The muscles of Wnt 10b -/- mice show impaired regenerative capacity following injury with replacement of muscle by adipose tissue (12). Satellite cells isolated from obese Zucker rats expressed an increased propensity to adipogenesis (as measured by oil red O) when compared to those from lean rats and this correlated with reduced Wnt 10b expression (13). These studies although suggestive do not however directly demonstrate a role for canonical Wnt signalling in the trans- differentiation of muscle stem cells along an adipogenic lineage. In the present study we have employed a large animal (porcine) model to study the effects of activation of the Wnt canonical signalling pathway upon adipogenesis in muscle derived cells (incubated at 37°C, 5% CO2 and air). At this stage, GM was substituted for either myogenic (Myo, DMEM, 2% horse serum, 2 mM L-glutamine, 100 IU/ml penicillin, 100 µg/ml streptomycin, 3 µg/ml amphotericin B) or adipogenic differentiation media (Ad, DMEM, 10% horse serum (Invitrogen), 1 µM dexamethasone (DEX, Sigma), 50 µM IBMX (Sigma), 10 µU insulin (Eli Lilly & Co Ltd, UK) and L-glutamine, penicillin, streptomycin and amphotericin B as for Myo). For RT-PCR, cells were cultured up to +24 h differentiation prior to RNA isolation, or for LiCl studies up to day 6 of differentiation, with Ad media added as described for 3 days, followed by 3 days in Ad media minus DEX and IBMX). For drug treated cells, 20 mM LiCl (Sigma) was added to Ad media on days 0 and 3 only. Following differentiation at day 6, cells were initially fixed in 70% ethanol. Image Analysis All images used for the quantification of Oil Red O staining, were analysed using Image Pro version 6.3. Phase contrast images were prepared as tiff documents for 100× magnification, taken for all experiments carried out at n = 6, with 15 fields of view (FOV) per replicate. For individual staining analyses, a macro was created to include the range of colour intensities denoting a positive result and applied to all grouped images. Error bars represent SD. Comparison of Serum Type on Adipogenic Differentiationf Adipogenic differentiation was induced [as previously described (7, 8)] using Ad media containing either 10% foetal bovine serum (FBS, Invitrogen) or 10% horse serum. All other conditions remained identical for both treatments. Oil Red O Staining g Prior to Oil Red O staining, cells were additionally fixed in 10% buffered formalin for 20 min and lipid was visualised via Oil Red O (0.5%, Sigma) staining as described previously (7, 8). Citation: Redshaw Z and Loughna PT (2018) Adipogenic Differentiation of Muscle Derived Cells is Repressed by Inhibition of GSK-3 Activity. Front. Vet. Sci. 5:110. doi: 10.3389/fvets.2018.00110 Redshaw Z and Loughna PT (2018) Adipogenic Differentiation of Muscle Derived Cells is Repressed by Inhibition of GSK-3 Activity. Front. Vet. Sci. 5:110. doi: 10.3389/fvets.2018.00110 Intermuscular adipose tissue (IMAT) has been described to include fat that is located beneath the muscle fascia and within the muscle itself (1). This intramuscular fat, deposited between skeletal muscle fibres and bundles of muscle fibers is highly evident in large animals and humans although insignificant June 2018 \| Volume 5 \| Article 110 Frontiers in Veterinary Science \| www.frontiersin.org 1 Adipogenic Differentiation of Muscle Derived Cells Redshaw and Loughna in healthy mice and rats. In livestock the degree of intramuscular adipose tissue (marbling) varies between species and breeds within a species and it is a major contributor to meat quality. In humans there is increased accretion of adipose tissue within skeletal muscle with age, inactivity and disease conditions such as diabetes (1, 2) . The increase in intramuscular adipose tissue in humans is visible on MRI images and has a clinical impact with regard to insulin resistance and cardio-metabolic syndrome. There are number of different cell types present in skeletal muscle that could contribute to fat cell formation including satellite cells, mesenchymal progenitors, fibro/ adipogenic progenitors (FAPs) and pericytes (1, 3, 4). In vitro studies have shown that satellite cells isolated from skeletal muscle or resident beneath the basil lamina of isolated muscle fibers can undergo adipogenic differentiation when subjected to high glucose levels or thiazolidinediones which are potent activators of the adipogenic marker PPARγ (5, 6). We have recently shown that cellsderived from different porcine skeletal muscles differ distinctly in their adipogenic potential. Cells derived from the diaphragm had a dramatically reduced ability to undergo adipogenic differentiation when compared to those derived from the hindlimb semi-membranosus muscle (7, 8). The Wnt family of secreted protein growth factors play a significant role in the differentiation and growth of skeletal muscle cells (9). Activation of the canonical Wnt pathway through the administration of lithium chloride has been shown to induce myotube hypertrophy (10). In other cell types Wnt signalling has been shown to have a significant role in differentiation with reduced signalling causing a shift in cell fate of pre-osteoblasts from osteoblasts to adipocytes (11). Materials and Methods Porcine muscle derived stem cells were isolated from the diaphragm (DIA) and hind limb semi-membranosus (SM) muscles, as previously described (7, 8). Freshly isolated cells were cultured on type I collagen (0.01%, Sigma, UK) coated plastic, in Memα growth media [GM, 20% FBS (Invitrogen, UK), 2 mM L-glutamine (Sigma), 100 IU/ml penicillin, 100 µg/ml streptomycin, 3 µg/ ml amphotericin B (Invitrogen)] and incubated at 37°C in a humidified atmosphere in 5% CO2, 90% Nitrogen and 20% O2 until approximately 60% confluent before cryogenic preservation in FBS and 10% DMSO (invitrogen). SDS-PAGE/western Blotting Myogenic and Adipogenic Differentiation Cells were seeded at a density of 2.6 × 104 cells/cm2 on collagen coated plastic and initially cultured in GM until ~80% confluent Results and Discussion Muscle stem cells including the predominant satellite cell population have been shown to have the capability to trans- differentiate along an adipogenic lineage including on isolated muscle fibers. In contrast a study in mice by Joe et al (4). strongly suggested that FAPs are the main source of adipocytes in adult tissue and another study, also in mice, showed a major role for mesenchymal progenitors in fat cell formation (3) A recent lineage study on myofibers isolated from mouse masseter and hind-limd muscles suggested that non-myogenic fates are a result of myofiber- associated progenitors rather than mesenchymal stem cells (15). Further study will be needed to define the cell types involved in the pig cells.We have recently shown that cells derived from different muscles have a distinctly different capacities to undergo adipogenic differentiation; with those isolated from the hind-limb semi- membranosus (SM) muscle being able to undergo significantly greater adipogenic differentiation when compared to those isolated from the diaphragm (DIA) (7, 8). This is in keeping with previous observations that different skeletal muscles respond differently to the same stimulus (16, 17). FIGURE 1 \| Wnt Expression and The influence of serum type on adipogenesis. (A) Expression of Wnt 10b by RT-PCR in muscle stem cells derived from diaphragm (DIA) and semi-membranosus (SM) muscles: undifferentiated and +24 h myogenic and adipogenic differentiation. (B) Lipid accumulation was measured by Oil Red O staining at Day 6 of differentiation, for cells grown in media containing either foetal bovine serum (FBS) or horse serum (HS). (n = 6, 15 FOV/rep). Error bars represent SD. GURE 1 \| Wnt Expression and The influence of serum type on FIGURE 1 \| Wnt Expression and The influence of serum type on adipogenesis. (A) Expression of Wnt 10b by RT-PCR in muscle stem cells derived from diaphragm (DIA) and semi-membranosus (SM) muscles: undifferentiated and +24 h myogenic and adipogenic differentiation. (B) Lipid accumulation was measured by Oil Red O staining at Day 6 of differentiation, for cells grown in media containing either foetal bovine serum (FBS) or horse serum (HS). (n = 6, 15 FOV/rep). Error bars represent SD. the expression of the Wnt 10b transcript in cells derived from the DIA and SM. Statistical Analysis Where applicable, student’s t-tests were performed on data sets (unpaired, two-tailed). Error bars represent SD. Results and Discussion Cells derived from the SM (which have a greater adipogenic potential) have a lower level of the 10b transcript just prior exposure to DM (i.e., at 80% confluence) compared to those from the DIA (Figure 1A). In cells exposed to myogenic DM for 24 h this higher expression in the DIA derived cells was maintained. In cells exposed to adipogenic DM, however, the transcript was not detectable in cells derived from either muscle (Figure 1A). These data suggest that the adipogenic potential of a muscle cell population(s) might be inversely correlated to levels of endogenous activation of the canonical Wnt pathway. If this is the case then activation of the canonical Wnt The Wnt (wingless type mouse mammary tumour virus integration-site family) family of growth factors play a crucial role during development in cell specification. The Wnt genes encode a large family of proteins that can signal through both canonical and non-canonical pathways (18). Of particular interest is Wnt 10b acting through the canonical pathway. This protein binds to the Frizzled/LRP receptor complex which leads to the inactivation of GSK3-β and the accumulation of cytosolic β-catenin which then translocates to the nucleus to activate transcription of Wnt responsive genes which include the myogenic regulatory factors MyoD and myogenin (19). In Wnt 10b−/− mice regeneration of muscle after induced muscle injury is associated with accumulation of large amounts of lipid and myoblasts from these animals express high levels of adipogenic markers such as PPAR-γ (12). Furthermore Wnt 10b mRNA levels are reduced in myoblasts from 24 month old compared to 8-month-old mice with cells from the older mice also exhibiting a greater adipogenic potential (12). It is not known, however, whether this factor plays any role in the markedly different adipogenic potential of stem cells derived from different muscles that we have described previously (7, 8). We therefore examined FIGURE 2 \| The effect of LiCl on adipogenesis. Lipid accumulation was measured by Oil Red O staining at Day 6 of differentiation, following treatment with LiCl (n = 6, 15 FOV/rep). Inset image representative of Oil Red O staining for control cells. Error bars represent SD. FIGURE 2 \| The effect of LiCl on adipogenesis. Lipid accumulation was measured by Oil Red O staining at Day 6 of differentiation, following treatment with LiCl (n = 6, 15 FOV/rep). Inset image representative of Oil Red O staining for control cells. Myogenic and Adipogenic Differentiation Cell lysates were prepared and electrophoresed as described in (7, 14). Protein was generated at four time points: undifferentiated (70% confluence), Day 1, Day 3 and Day 6 following induction of Cells were seeded at a density of 2.6 × 104 cells/cm2 on collagen coated plastic and initially cultured in GM until ~80% confluent June 2018 \| Volume 5 \| Article 110 Frontiers in Veterinary Science \| www.frontiersin.org 2 Adipogenic Differentiation of Muscle Derived Cells Redshaw and Loughna differentiation, in triplicate for each treatment. Primary antibodies were rabbit monoclonal antibodies to PPARγ (1:300, Abcam), GSK3β (1:1,000, Cell Signalling, UK), pGSK3β (1:1,000, Cell Signalling), pGSK3α (1:1,000, Cell Signalling) and mouse monoclonal antibody α-Tubulin (1:1,000, Abcam). Secondary antibodies were anti-rabbit HRP and anti-mouse HRP (1:2,000, Cell Signalling). Quantification of protein expression was analysed via densitometry using ImageQuant TL (version 2005, Amersham Biosciences). FIGURE 1 \| Wnt Expression and The influence of serum type on adipogenesis. (A) Expression of Wnt 10b by RT-PCR in muscle stem cells derived from diaphragm (DIA) and semi-membranosus (SM) muscles: undifferentiated and +24 h myogenic and adipogenic differentiation. (B) Lipid accumulation was measured by Oil Red O staining at Day 6 of differentiation, for cells grown in media containing either foetal bovine serum (FBS) or horse serum (HS). (n = 6, 15 FOV/rep). Error bars represent SD. Frontiers in Veterinary Science \| www.frontiersin.org Results and Discussion Error bars represent SD. Frontiers in Veterinary Science \| www.frontiersin.org Frontiers in Veterinary Science \| www.frontiersin.org June 2018 \| Volume 5 \| Article 110 3 Adipogenic Differentiation of Muscle Derived Cells Redshaw and Loughna FIGURE 3 \| Expression of total GSK3-β (A), pGSK3-β (B), pGSK2-α (C) and α-Tubulin (D) proteins following LiCl treatment, at various time points of differentiation (n = 3). Inset images representative of western blot raw data. Error bars represent SD. FIGURE 3 \| Expression of total GSK3-β (A), pGSK3-β (B), pGSK2-α (C) and α-Tubulin (D) proteins following LiCl treatment, at various time points of differentiation (n = 3). Inset images representative of western blot raw data. Error bars represent SD. pathway should suppress the adipogenic transdifferentiation of these cells. differentiation, we found that choice of serum did significantly influence the degree of lipid accumulation, with more observed in HS treated cultures (Figure 1B, p =< 0.0001 n = 6, 15 FOV/rep). All subsequent work was performed in 10% HS. Subsequent studies were carried out on cells derived from the SM only as we have shown previously that cells derived from the diaphragm have a very limited potential for adipogenic differentiation. Cells were isolated from the SM muscle and the predominant satellite cell population identitied verified by the expression of myogenic regulatory factors Pax7 and Pax3 (8). The majority of cells expressed both markers, which we typically find to be greater than a 98% pure population [(7, 8) and data not shown]. Having previously shown that cells from SM muscle are capable of adipogenic differentiation (7, 8), we wished to further enhance this response. We therefore carried out cell culture studies at 20% oxygen concentration rather than at lower (~5%) more physiological levels which although enhancing myogenic retards adipogenic differentiation (7). We also compared the effect of serum type, within the adipogenic DM, on the ability of cells to generate lipid. Many adipogenic-inducing protocols use media containing 10% foetal bovine serum (FBS) as standard, compared with myogenic differentiation where horse serum (HS) is preferentially used to induce terminal differentiation. Following adipogenic To investigate the signalling pathways involved in the adipogenic cell fate of SM cells, we inhibited glycogen synthase kinase-3 (GSK- 3) activity, effectively stimulating Wnt activation by addition of lithium chloride (LiCl). GSK-3 exists as two isoforms α and β both of which are expressed in skeletal muscle. Frontiers in Veterinary Science \| www.frontiersin.org Funding This research was funded by the University of Nottingham. 4. Joe AW, Yi L, Natarajan A, Le Grand F, So L, Wang J, et al. Muscle injury activates resident fibro/adipogenic progenitors that facilitate myogenesis. Nat Cell Biol (2010) 12(2):153–63. doi: 10.1038/ncb2015 Conclusions A number of studies have demonstrated that in both muscle cell lines (such as the murine C2C12 cells) and in primary stem cells derived from skeletal muscle that there is a limited potential for adipogenic trans-differentiation in vitro. We have previously shown that cells derived from different porcine skeletal muscles differ in their adipogenic potential and that furthermore this potential may be affected by the oxygen environment in which the cells are cultured. In the present study we have further optimised the conditions to promote adipogenic differentiation in porcine skeletal muscle derived stem cells. This adipogenic differentiation was almost completely prevented by the administration of lithium chloride (LiCl) which is a widely characterised inhibitor of GSK-3. This compound is in effect a mimic of Wnt pathway activation and the present study supports in vivo evidence that the Wnt pathway is inhibitory to adipose tissue formation. FIGURE 4 \| Effect of LiCl treatment on PPARγ expression. Expression of PPARγ protein following various time points of adipogenic differentiation (n = 3). Inset image representative of western blot raw data. Error bars represent SD. p =< 0.0001 n = 6, 15 FOV/rep). Subsequently we examined the expression of both GSK3β and the inactivated phosphorylated form (pGSK3β), at the protein level, over several time points of differentiation: undifferentiated, Day 1, Day 3 and Day 6 in DM. Results for the level of total GSK3β showed no significant difference between treatment and control, at any time point (Figure 3A). In contrast, however, following initiation of differentiation, pGSK3β expression markedly increased in LiCl treated cells for all time points (Figure 3B, Day 1 p =< 0.001, Day 3 p =< 0.01, Day 6 p =< 0.05, n = 3), supporting the ORO results. Expression of the phosphorylated alpha form of GSK3 (pGSK3α) increased over time for both treatments but was only significantly higher in LiCl treated cells at Day 3 (Figure 3C, p =< 0.05, n = 3). Loading control, α-Tubulin expression was similar for samples at most time points (Figure 3D). Of note, we observe variable expression between treatments of many loading control proteins during early time points of myogenic and adipogenic differentiation including GAPDH, β-tubulin and β-actin). Following LiCl treatment, the key adipogenic regulatory transcription factor Peroxisome Results and Discussion These two isoforms show extensive homology and although they have similar functions they are not functionally redundant (20). It has been demonstrated that they have distinct roles in cardiogenesis with GSK-3α required for myocyte survival and GSK-3β modulating left right symmetry (21). To date GSK-3α is the much less studied of the two isoforms although some studies have shown it to be the more potent isoform. We have therefore also examined this isoform in the present study. Inhibition of GSK-3β by Lithium chloride has been shown to induce myotube hypertrophy in cells cultured in myogenic DM (10). In the present study in which cells were cultured in adipogenic DM lipid formation was virtually ablated in all treated cells, which was quantified by Oil Red O (ORO) staining (Figure 2B, June 2018 \| Volume 5 \| Article 110 4 Adipogenic Differentiation of Muscle Derived Cells Redshaw and Loughna FIGURE 4 \| Effect of LiCl treatment on PPARγ expression. Expression of PPARγ protein following various time points of adipogenic differentiation (n = 3). Inset image representative of western blot raw data. Error bars represent SD. Proliferator Activator Protein gamma (PPARγ) was significantly down regulated from Day 3 of differentiation onwards, compared to the control (Figure 4, Day 3 p = 0.05, Day 6 p =< 0.001, n = 3). The fact that PPARγ was significantly down-regulated from day 3 and that ORO almost completely ablated by day 6 suggests that it is GSK3β and not GSK3α that plays the major role in this process. The down regulation of canonical wnt signalling with age and in certain disease processes suggests that this pathway may play a major role in the accumulation of adipose and its replacement of skeletal muscle tissue, in these conditions, in both large animals and humans. Author Contributions Both Authors contributed to the design of the study, the interpretation of the data and the writing of the paper. References 1. Vettor R, Milan G, Franzin C, Sanna M, de Coppi P, Rizzuto R, et al. The origin of intermuscular adipose tissue and its pathophysiological implications. Am J Physiol Endocrinol Metab (2009) 297(5):E987–98. doi: 10.1152/ajpendo.00229.2009 5. Aguiari P, Leo S, Zavan B, Vindigni V, Rimessi A, Bianchi K, et al. High glucose induces adipogenic differentiation of muscle-derived stem cells. Proc Natl Acad Sci U S A (2008) 105(4):1226–31. doi: 10.1073/pnas.0711402105 6. de Coppi P, Milan G, Scarda A, Boldrin L, Centobene C, Piccoli M, et al. Rosiglitazone modifies the adipogenic potential of human muscle satellite cells. Diabetologia (2006) 49(8):1962–73. doi: 10.1007/s00125-006- 0304-6 2. Hilton TN, Tuttle LJ, Bohnert KL, Mueller MJ, Sinacore DR. Excessive adipose tissue infiltration in skeletal muscle in individuals with obesity, diabetes mellitus, and peripheral neuropathy: association with performance and function. Phys Ther (2008) 88(11):1336–44. doi: 10.2522/ ptj.20080079 7. Redshaw Z, Loughna PT. Oxygen concentration modulates the differentiation of muscle stem cells toward myogenic and adipogenic fates. Differentiation (2012) 84(2):193–202. doi: 10.1016/j.diff.2012.06.001f 3. Uezumi A, Fukada S, Yamamoto N, Takeda S, Tsuchida K. Mesenchymal progenitors distinct from satellite cells contribute to ectopic fat cell formation in skeletal muscle. Nat Cell Biol (2010) 12(2):143–52. doi: 10.1038/ncb2014 3. Uezumi A, Fukada S, Yamamoto N, Takeda S, Tsuchida K. Mesenchymal progenitors distinct from satellite cells contribute to ectopic fat cell formation in skeletal muscle. Nat Cell Biol (2010) 12(2):143–52. doi: 10.1038/ncb2014 f 8. Redshaw Z, Mcorist S, Loughna P. Muscle origin of porcine satellite cells affects in vitro differentiation potential. Cell Biochem Funct (2010) 28(5):403–11. doi: 10.1002/cbf.1670 June 2018 \| Volume 5 \| Article 110 Frontiers in Veterinary Science \| www.frontiersin.org 5 Adipogenic Differentiation of Muscle Derived Cells Redshaw and Loughna 9. Brack AS, Conboy IM, Conboy MJ, Shen J, Rando TA. A temporal switch from notch to Wnt signaling in muscle stem cells is necessary for normal adult myogenesis. Cell Stem Cell (2008) 2(1):50–9. doi: 10.1016/j.stem.2007.10.006 17. Loughna PT, Brownson C. Two myogenic regulatory factor transcripts exhibit muscle-specific responses to disuse and passive stretch in adult rats. FEBS Lett (1996) 390(3):304–6. doi: 10.1016/0014-5793(96)00681-3 10. Vyas DR, Spangenburg EE, Abraha TW, Childs TE, Booth FW. GSK-3beta negatively regulates skeletal myotube hypertrophy. Am J Physiol Cell Physiol (2002) 283(2):C545–51. doi: 10.1152/ajpcell.00049.2002 18. Li F, Chong ZZ, Maiese K. Winding through the WNT pathway during cellular development and demise. Histol Histopathol (2006) 21(1):103–24. doi: 10.14670/HH-21.103 11. References Song L, Liu M, Ono N, Bringhurst FR, Kronenberg HM, Guo J. Loss of wnt/ β-catenin signaling causes cell fate shift of preosteoblasts from osteoblasts to adipocytes. J Bone Miner Res (2012) 27(11):2344–58. doi: 10.1002/jbmr.1694 19. Rochat A, Fernandez A, Vandromme M, Molès JP, Bouschet T, Carnac G, et al. Insulin and wnt1 pathways cooperate to induce reserve cell activation in differentiation and myotube hypertrophy. Mol Biol Cell (2004) 15(10):4544–55. doi: 10.1091/mbc.e03-11-0816 12. Vertino AM, Taylor-Jones JM, Longo KA, Bearden ED, Lane TF, Mcgehee RE, et al. Wnt10b deficiency promotes coexpression of myogenic and adipogenic programs in myoblasts. Mol Biol Cell (2005) 16(4):2039–48. doi: 10.1091/mbc. e04-08-0720 20. Freland L, Beaulieu JM. Inhibition of GSK3 by lithium, from single molecules to signaling networks. Front Mol Neurosci (2012) 5:14. doi: 10.3389/ fnmol.2012.00014 13. Scarda A, Franzin C, Milan G, Sanna M, dal Prà C, Pagano C, et al. Increased adipogenic conversion of muscle satellite cells in obese Zucker rats. Int J Obes (2010) 34(8):1319–27. doi: 10.1038/ijo.2010.47 21. Lee HC, Tsai JN, Liao PY, Tsai WY, Lin KY, Chuang CC, et al. Glycogen synthase kinase 3 alpha and 3 beta have distinct functions during cardiogenesis of zebrafish embryo. BMC Dev Biol (2007) 7:93. doi: 10.1186/1471-213X-7- 93 21. Lee HC, Tsai JN, Liao PY, Tsai WY, Lin KY, Chuang CC, et al. Glycogen synthase kinase 3 alpha and 3 beta have distinct functions during cardiogenesis i of zebrafish embryo. BMC Dev Biol (2007) 7:93. doi: 10.1186/1471-213X-7- 93 14. Atherton PJ, Szewczyk NJ, Selby A, Rankin D, Hillier K, Smith K, et al. Cyclic stretch reduces myofibrillar protein synthesis despite increases in FAK and anabolic signalling in L6 cells. J Physiol (2009) 587(Pt 14):3719–27. doi: 10.1113/jphysiol.2009.169854 Conflict of Interest Statement: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. 15. Segev E, Shefer G, Adar R, Chapal-Ilani N, Itzkovitz S, Horovitz I, et al. Muscle-bound primordial stem cells give rise to myofiber-associated myogenic and non-myogenic progenitors. PLoS One (2011) 6(10):e25605. doi: 10.1371/ journal.pone.0025605 Copyright © 2018 Redshaw and Loughna. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). Frontiers in Veterinary Science \| www.frontiersin.org References The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. 16. Brownson C, Loughna P. Alterations in the mRNA levels of two metabolic enzymes in rat skeletal muscle during stretch-induced hypertrophy and disuse atrophy. Pflugers Arch (1996) 431(6):990–2. doi: 10.1007/ s004240050097 June 2018 \| Volume 5 \| Article 110 Frontiers in Veterinary Science \| www.frontiersin.org 6
https://openalex.org/W2987436765	https://bora.uib.no/bora-xmlui/bitstream/1956/22592/3/2019_Kimmich_Participatory%2bmodeling%2bupdates%2bexpectations.pdf	English	null	Participatory Modeling Updates Expectations forIndividuals and Groups, Catalyzing BehaviorChange and Collective Action inWater‐Energy‐Food NexusGovernance	Earth's future	2,019	cc-by	16,648	©2019. The Authors. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. Received 12 JUL 2019 Accepted 25 OCT 2019 Plain Language Summary Our research contributes to understanding actionable knowledge for sustainability using a before‐after intervention with ﬁshing and farming community representatives in a situation of conﬂicting water, energy, food, and livelihoods priorities in rural Cambodia. We explain why reducing uncertainty and building consensus on action through participatory research could potentially catalyze new behavior that promotes sustainability and test how this happens in our intervention. The result is a new and much needed evaluation framework and method for behavioral change outcomes in sustainability interventions. Author Contributions: Conceptualization: C. Kimmich, L. Gallagher, B. Kopainsky, M. Dubois, C. Bréthaut Data curation: C. Kimmich, B. Kopainsky, M. Dubois, C. Sovann Formal analysis: C. Kimmich, L. Gallagher Funding acquisition: L. Gallagher, C. Bréthaut Investigation: B. Kopainsky, M. Dubois, C. Sovann, C. Buth (continued) Author Contributions: Conceptualization: C. Kimmich, L. Gallagher, B. Kopainsky, M. Dubois, C. Bréthaut Data curation: C. Kimmich, B. Kopainsky, M. Dubois, C. Sovann Formal analysis: C. Kimmich, L. Gallagher Funding acquisition: L. Gallagher, C. Bréthaut Investigation: B. Kopainsky, M. Dubois, C. Sovann, C. Buth (continued) Citation: Citation: Kimmich, C., Gallagher, L., Kopainsky, B., Dubois, M., Sovann, C., Buth, C., & Bréthaut, C. (2019). Participatory modeling updates expectations for individuals and groups, catalyzing behavior change and collective action in water‐energy‐food nexus governance. Earth's Future, 7. https:// doi.org/10.1029/2019EF001311 Correspondence to: L. Gallagher, louise.gallagher@unige.ch Correspondence to: L. Gallagher, louise.gallagher@unige.ch Correspondence to: L. Gallagher, louise.gallagher@unige.ch Key Points: • Participatory modeling could be fundamental to understanding complex social‐ecological systems with high uncertainty and conﬂicting interests C. Kimmich1 , L. Gallagher2 , B. Kopainsky3 , M. Dubois4, C. Sovann5, C. Buth6, and C Bréthaut2 • Engagement in participatory modeling can change expectations and thereby increase individual and collective agency in facing sustainability challenges C. Kimmich1 , L. Gallagher2 , B. Kopainsky3 , M. Dubois4, C. Sovann5, C. Buth6, an C. Bréthaut2 1Faculty of Social StudiesDepartment of Environmental Studies, Masaryk University, Brno, Czechia, 2Institute for Environmental Sciences, University of Geneva, Geneva, Switzerland, 3System Dynamics Group, Department of Geography, University of Bergen, Bergen, Norway, 4WorldFish Myanmar – Fish CGIAR Research Program, 5Department of Environmental Science, Royal University of Phnom Penh, Phnom Penh, Cambodia, 6Mekong Flooded Forest Landscape Team, World Wide Fund for Nature, Kratie Province, Cambodia • Our method provides a framework to evaluate and compare participatory modeling approaches for their behavior change potential • Our method provides a framework to evaluate and compare participatory modeling approaches for their behavior change potential Supporting Information: Supporting Information: Supporting Information: • Supporting Information S1 • Data Set S1 • Supporting Information S1 • Data Set S1 • Supporting Information S1 Abstract Participatory modeling is a potentially high‐impact approach for catalyzing fundamental sustainability transformations. We test if participation in a group system dynamics modeling exercise increases participants' agency through a novel method to evaluate potential behavioral change using expectation measures. A water‐energy‐food nexus—a functionally interdependent but underconceptualized system with low consensus and high scientiﬁc uncertainty—was mapped, and its evolution simulated by 46 participants in three interventions in a region undergoing hydropower infrastructure development in Northeastern Cambodia. Participants' system‐related expectations were measured before and after the interventions. Our results suggest that participants became signiﬁcantly more optimistic about their individual agency to increase agricultural and ﬁshing income and, interestingly, less likely to participate in local government development planning procedures. Findings also reveal how some uncertainties for multiple variables were reduced within and across the groups. Such converging expectations suggest that participatory modeling could contribute to making collective solutions and institutionalized agreements more likely. This research contributes to innovation in sustainability because it unpacks some underlying mechanics of how participatory processes can lead to new adaptive capacities, shared perspectives, and collective actions. • Data Set S1 • Data Set S1 Author Contributions: Conceptualization: C. Kimmich, L. Gallagher, B. Kopainsky, M. Dubois, C. Bréthaut Data curation: C. Kimmich, B. Kopainsky, M. Dubois, C. Sovann Formal analysis: C. Kimmich, L. Gallagher Funding acquisition: L. Gallagher, C. Bréthaut Investigation: B. Kopainsky, M. Dubois, C. Sovann, C. Buth (continued) 1. Introduction First, that knowledge produced through participatory processes is more likely to identify transformational leverage points (Star & Griesemer, 1989) and feasible responses for tackling complex sus- tainability challenges because the knowledge is coproduced from multiple perspectives (Cash et al., 2003). Second, such participatory research can directly and indirectly inﬂuence behavior change required for rea- lizing sustainability pathways (Gerritsen et al., 2013). However, epistemological, ontological, and methodo- logical issues for integrating knowledge and engaging with values, power, and politics remain so signiﬁcant (Jordan et al., 2018; Miller & Wyborn, 2018; Wesselink et al., 2013) that such assumptions need to be con- structively challenged if we are to innovate a scalable, deliberate, and thoughtful practice for participatory research in sustainability. Resources: L. Gallagher, M. Dubois, C. Buth Supervision: L. Gallagher, B. Validation: C. Kimmich, L. Gallagher, B. Kopainsky, M. Dubois, C. Buth Visualization: C. Kimmich, B. Kopainsky Participatory modeling is suited to address less structured problems where consensus is low, uncertainty is high, and collaboration, co‐decision, and joint action are needed (Basco‐Carrera et al., 2017; Hurlbert & Gupta, 2015). Compliance‐based thinking dominates in rationalizing uses of participatory modeling (Basco‐Carrera et al., 2017), suggesting the primary reason for stakeholder participation in model building is to increase acceptability, ownership, and legitimacy of modeling results. In sustainability science, partici- patory modeling is celebrated for social learning about systems and root causes of problems, providing multi- ple perspectives in decision support for policy design, increasing adaptive capacity, and enabling coordinated action (Hedelin et al., 2017; Voinov et al., 2016). Participation in model development has received far less attention as a ﬁeld of practice and research, compared to focus groups, citizen panels, or participatory rural appraisals, however (Scott et al., 2016b). We argue this overlooks the potential contribution to behavior change of the process of participatory modeling itself. Writing ‐ original draft: C. Kimmich, L. Gallagher Participatory modeling has been called a cultural process (Crane, 2010) that can reveal key uncertainties in play and resolve some of these through facilitating information sharing, group problem and action analysis, deliberation, and negotiation (Rouwette et al., 2011; Rouwette & Vennix, 2006). Comparative analyses of participatory modeling techniques attest to effects on social capital (Davies et al., 2015) and on participant mental models and attitudes (Scott et al., 2016b), which are theorized as key determinants of behavior (Denzau & North, 1994; Rodrik, 2014). 1. Introduction Investigation: B. Kopainsky, M. Dubois, C. Sovann, C. Buth (continued) Sustainability transformations necessarily happen under huge uncertainty about future pathways and actions to take (Abson et al., 2017). Empowerment is a critical component of adaptive capacity, agency, and behavior change needed to navigate such uncertainty. Collaborating with stakeholders in research across and beyond disciplines (Pahl‐Wostl et al., 2013; Pohl et al., 2010) produces new knowledge that can address some uncertainties (Addor et al., 2015; Yung et al., 2019) and empower communities and individuals to develop new pathways (Gerritsen et al., 2013; Miller & Wyborn, 2018). Progress has been made on better understanding knowledge coproduction (Dubois et al., 2016) and its outcomes (DeLorme et al., 2016; Miller & Wyborn, 2018). However, we know much less about speciﬁc mechanisms that may induce participants in these processes to change behavior (Carr et al., 2012; Smajgl & Ward, 2015) and why this new behavior might catalyze sustainability outcomes. 1 KIMMICH ET AL. KIMMICH ET AL. Earth's Future 10.1029/2019EF001311 Methodology: C. Kimmich, L. Gallagher, B. Kopainsky, M. Dubois, C. Sovann, C. Bréthaut Project administration: L. Gallagher, C. Bréthaut Resources: L. Gallagher, M. Dubois, C. Buth Supervision: L. Gallagher, B. Kopainsky, M. Dubois, C. Sovann, C. Bréthaut Validation: C. Kimmich, L. Gallagher, B. Kopainsky, M. Dubois, C. Buth Visualization: C. Kimmich, B. Kopainsky Writing ‐ original draft: C. Kimmich, L. Gallagher Writing – review & editing: C. Kimmich, L. Gallagher, B. Kopainsky, M. Dubois, C. Sovann, C. Buth, C. Bréthaut Though participatory research is framed differently across sustainability domains, two core assumptions seem to dominate. First, that knowledge produced through participatory processes is more likely to identify transformational leverage points (Star & Griesemer, 1989) and feasible responses for tackling complex sus- tainability challenges because the knowledge is coproduced from multiple perspectives (Cash et al., 2003). Second, such participatory research can directly and indirectly inﬂuence behavior change required for rea- lizing sustainability pathways (Gerritsen et al., 2013). However, epistemological, ontological, and methodo- logical issues for integrating knowledge and engaging with values, power, and politics remain so signiﬁcant (Jordan et al., 2018; Miller & Wyborn, 2018; Wesselink et al., 2013) that such assumptions need to be con- structively challenged if we are to innovate a scalable, deliberate, and thoughtful practice for participatory research in sustainability. Though participatory research is framed differently across sustainability domains, two core assumptions seem to dominate. 1. Introduction However, there have been no explorations of their impact on parti- cipants' expectations even though expectation changes are strong predictors of behavior change (Delavande et al., 2011b; Jensen, 2010; Kimmich & Fischbacher, 2016; McKenzie et al., 2013) and are likely inﬂuenced by changes in both individual and shared mental models (Glynn et al., 2017). Expectations have been shown to predict behavior in a number of different areas and ﬁeld settings, including expected income and education (Manski, 2004), income and migration (McKenzie et al., 2013), coffee prices and labor allocation (Hill, 2010), wood prices and industrial organization (Kimmich & Fischbacher, 2016), or monsoon onset and planting decisions (Giné et al., 2009). See Delavande et al. (2011b) for a review. This research contributes to understanding the actionable dimension of “actionable knowledge” produc- tion by testing behavioral change efﬁcacy in participatory systems modeling using a before‐after inter- vention design in ﬁeld conditions. We tested behavior change in a participatory systems thinking and modeling procedure called causal loop diagram (CLD) mapping and simulation (Kopainsky et al., 2017) conducted with ﬁshing and farming community stakeholders in Kratie Province, the site of bur- geoning water‐energy‐food (WEF) nexus tensions in Northeastern Cambodia in the Mekong River Basin. Theoretically, we provide explanations for the mechanisms of behavioral change resulting from systems thinking‐ and modeling‐based interventions. Empirically, our ﬁeld results show support for a recent pro- position in the behavioral literature that intervening to change expectations can drive changes in beha- vior. Methodologically, we contribute a novel and needed evaluation framework for behavioral outcome evaluation in sustainability interventions. Practically, we derive lessons for participatory sustainability research and practice. These four contributions have implications for design and evaluation of pro- grams, projects, networks, and partnerships in sustainability. 2.2. Case Selection: Why Focus on Participatory Processes in WEF Nexus Governance? 2. Case Selection: Why Focus on Participatory Processes in WEF Nexus Governance? We chose to test our hypotheses in the nexus context using participatory CLD mapping and simulation. hree reasons led us to choose this method and context: First, with increasing pressures on natural resources, and externalities and trade‐offs between sectors becoming more visible (Al‐Saidi & Elagib, 2017), the WEF nexus has emerged as an important phenomenon and theme in sustainability (Gallagher et al., 2016; Liu et al., 2018). While a vast number of WEF nexus fra- meworks have emerged within less than a decade, agreement on key ingredients of this socioecological sys- tem has not been achieved (Albrecht et al., 2018; Allan et al., 2015; Allouche et al., 2015; Villamayor‐Tomas et al., 2015). In this respect, low consensus on methods (Kaddoura & El Khatib, 2017), conﬂicting interests (Fox & Sneddon, 2019; Lebel & Lebel, 2018), and scientiﬁc uncertainty make the WEF nexus a paradigmatic case for sustainability science (Boyd et al., 2015; Cash et al., 2003; Hurlbert & Gupta, 2015). Second, model‐based decision support is an area ripe for innovation in sustainability research (Basco‐ Carrera et al., 2017; Gerritsen et al., 2013; van Voorn et al., 2016). New insights on how this form of knowl- edge production and use could translate to behavior change have potential to scale given its common usage. The WEF nexus makes a good case precisely because its computational modeling frameworks and meth- odologies (Albrecht et al., 2018) are currently debated (Kaddoura & El Khatib, 2017) and evolving to opera- tionalize the WEF nexus usefully in governance and decision making (Shannak et al., 2018; Weitz et al., 2017). Finally, WEF nexus governance research is increasingly concerned with questions of participation. WEF nexus governance requires complex adaptation among a large set of stakeholders with all the attendant uncertainties, politics, and power dynamics (Allouche et al., 2015; Foran, 2015; Gallagher et al., 2019). We now understand that tracing causes of problems and potential responses in WEF nexus governance requires transdisciplinary systems thinking (Hagemann & Kirschke, 2017; Howarth & Monasterolo, 2017). Public participation has been extensively addressed by research on challenges of inclusion of different stakeholder groups in decision making and the varying degrees of legitimacy and success implied by such processes (see, among others, Fung, 2006; Irvin & Stansbury, 2004; Reed, 2008; Heijden & Heuvelhof, 2012). 2.1. Hypotheses We deﬁne efﬁcacy of participatory system modeling interventions as actual and prospective behavioral change that was intended and induced by the intervention. This includes all types of individual behaviors, either at the household and farm scale, or as part of organizations. KIMMICH ET AL. KIMMICH ET AL. 2 Earth's Future Earth's Future 10.1029/2019EF001311 We formulate two hypotheses to evaluate whether these processes are likely to elicit new behavioral change in participants. We hypothesize that participant expectations concerning (1) the likelihood of certain future events and (2) their own individual agency to respond to these events will change because of the intervention. We assume that optimism or pessimism regarding future outcomes will be affected by these expectations and, in turn, inﬂuence future choices in behavior. This activation hypothesis relates to the theory on self‐efﬁcacy in bringing about speciﬁc expected outcomes (Gittelsohn et al., 2012) and empowerment outcomes from par- ticipatory modeling (Basco‐Carrera et al., 2017). Self‐efﬁcacy is a key component of empowerment (Diener & Biswas‐Diener, 2005). The hypothesis suggests that participants become more optimistic about their indivi- dual opportunities under their control, rather than external outcomes that they cannot inﬂuence. We also postulate that participatory modeling reduces uncertainty and enables convergence in expectations across members of the group. Shared expectations are an important prerequisite for collective action (Runge, 1986) and therefore fundamental to institutional innovation and making choices about shared resources. Once participants converge on the likelihood of an event or outcome, support for an institutional solution becomes common knowledge, which facilitates collective choice. We assume that joint exploration of system properties in a participatory modeling protocol leads to shared mental models (Scott et al., 2016b) and hypothesize that convergence dominates divergence in expectations‐forming across participants. We focus on uncertainty at the group level rather than changes in individuals' uncertainty, which would require expectation elicitation of multiple intervals, exceeding survey time constraints when multiple expectation variables are measured. 2.3. Place‐Based Case Selection, Sampling Procedure, and Participant Identiﬁcation 2.3. Place‐Based Case Selection, Sampling Procedure, and Participant Identiﬁcation The Mekong River Basin is the site of large‐scale and largely uncoordinated hydropower development in a situation of rapid climate and socioeconomic changes, where the Mekong River Commission is the primary intergovernmental body for cooperation between the lower basin riparian countries (Viet Nam, Thailand, Lao PDR, and Cambodia) and upstream partners (China and Myanmar). Hydropower is viewed as a poverty reduction measure in this region, and yet these infrastructures are also expected to impact signiﬁcantly on ecosystem integrity, ﬁsheries and agricultural systems, and, subsequently, nature‐based livelihoods and food production (Mekong River Commission, 2017; Molle et al., 2012). Mekong ﬁshers and farmers are and will continue to operate in a complex situation, making decisions under a large degree of future uncertainty. Navigating this WEF nexus has been an ongoing challenge for the trans- boundary governance of this globally signiﬁcant river basin. The CGIAR WLE (CGIAR's Research Program on Water, Land and Ecosystems) Challenge program on water and food (2002–2013) and subsequent research investments on various dimensions of the Mekong WEF nexus (Gallagher et al., 2019; Foran, 2015; Grafton et al., 2016; Lebel & Lebel, 2018; Molle, 2009; Middleton et al., 2015; Orr et al., 2012; Pittock et al., 2016; Smajgl & Ward, 2013; Smajgl et al., 2015; Villamayor‐Tomas et al., 2016) mean this is one of the best researched WEF nexus cases globally. Our Mekong WEF nexus case study is a WEF situation in Kratie Province, Cambodia (Liu et al., 2018). Local communities in this region will be impacted by two major hydropower projects in Cambodia: Stung Treng dam in Stung Treng Province, a proposed mainstem (central river channel) gravity dam being built for energy export to Thailand; and Sambor dam in Kratie Province, the largest proposed dam in the entire river basin (Wild et al., 2019; Wu et al., 2010). Kratie and Stung Treng Provinces are also experiencing other forms of development that impact the agricultural and ﬁshing activities that currently support local populations' subsistence and livelihoods in different ways. Rubber plantation expansion, river bed sand mining, and road network infrastructure are all in progress, alongside a burgeoning ecosystem‐based tourism sector driven by the presence of rare and endangered Irrawaddy freshwater dolphins (Kratie Provincial Department of Planning, 2014). 2.2. Case Selection: Why Focus on Participatory Processes in WEF Nexus Governance? A particular subset of WEF nexus governance literature has shown that participatory modeling holds some potential for anticipatory governance (Boyd et al., 2015; Guston, 2014) in river basin WEF nexus situations (Sendzimir et al., 2007; Smajgl et al., 2015; Smajgl & Ward, 2013). It can be useful to address the complexity of stakes in WEF nexus governance (Halbe et al., 2015; Howarth & Monasterolo, 2017; Smajgl & Ward, 2013). It has been shown to increase understanding of empirical reality and intricacy by integrating different perspectives and narratives related to WEF nexus intersectoral challenges (Gallagher et al., 2019). It KIMMICH ET AL. 3 KIMMICH ET AL. Earth's Future Earth's Future 10.1029/2019EF001311 facilitates knowledge sharing and integration (Hagemann & Kirschke, 2017) and has been used to open new formal or informal deliberations between WEF nexus actors, as well as reinforcing existing science‐policy interfaces (Smajgl et al., 2015). Yet the link between participation and anticipatory governance outcomes is underdeveloped (Smajgl & Ward, 2015). Exploring the effect of participatory modeling may have on actor behavior in WEF nexus governance systems is a critical next step in this strand of research. facilitates knowledge sharing and integration (Hagemann & Kirschke, 2017) and has been used to open new formal or informal deliberations between WEF nexus actors, as well as reinforcing existing science‐policy interfaces (Smajgl et al., 2015). Yet the link between participation and anticipatory governance outcomes is underdeveloped (Smajgl & Ward, 2015). Exploring the effect of participatory modeling may have on actor behavior in WEF nexus governance systems is a critical next step in this strand of research. 2.3. Place‐Based Case Selection, Sampling Procedure, and Participant Identiﬁcation With natural resource exploitation closely linked to fraught domestic politics (Milne & Mahanty, 2015; Un & Sokbunthoeun, 2009) and poor local participation in large dam planning (Siciliano et al., 2015; Sithirith, 2016), local government development planning processes are the major formal planning mechanisms open to people in places like Kratie Province. Local‐level priority and concerns are identiﬁed at commune level (the third level of local government administration in Cambodia, below district and provincial administra- tions) by Commune Councils in a 3‐year rolling Commune Development Plan (CDP), which is then reviewed annually to produce a Commune Investment Plan (CIP) for the year. The CDP/CIP dominates local formal development planning (Plummer & Tritt, 2012) as an institution dating back to French colonial times and the ﬁrst level of government administration to be reestablished following the civil war (Öjendal & Lilja, 2009). In principle, it is a good process that relies on and encourages citizen participation to build legiti- macy. There are many concerns about how it works in practice, however (Milne & Mahanty, 2015; Plummer & Tritt, 2012; Vuković & Babović, 2018). Underfunded, and conducted in the context of speciﬁc cultural and sociopolitical dynamics of Cambodia's rural areas, the annual CIP rarely challenges the CDP or other plans generated higher up in the political hier- archy (Öjendal & Sedara, 2006; Sedara, 2012), and infrastructure (often roads and local irrigation projects) is identiﬁed repeatedly as the main priority without any evidence as to its effectiveness in poverty reduction. A process of decentralization and deconcentration of government functions (hereafter “D&D reforms”) is shifting government toward more transversal service delivery goals with new local planning procedures to encompass district and commune plans (Sedara, 2012; Royal Decree No. NS/RKM/1208/1429 2008; Royal Decree no. NS/RKM/1014/1174 2014). However, the D&D reforms face their own implementation challenges (Cambodia Development Resource Institute, 2004; National Committee for Sub‐National Democratic Development, 2010, 2017), and with no current alternatives, local people engage in CIP meetings. 4 KIMMICH ET AL. KIMMICH ET AL. Earth's Future 10.1029/2019EF001311 We chose to select participants from villages in communes within Kratie Province that are in proximate dis- tance to the Mekong River channel to assure that a majority of our participants were involved in the local manifestation of complex intersectoral issues and involved at the scale of which response planning will likely take place. 2.3. Place‐Based Case Selection, Sampling Procedure, and Participant Identiﬁcation Drawing on the Cambodian Commune Council Database (National Committee for Sub‐National Democratic Development, 2015), we instrumentally selected (Gerring, 2007) two villages along the Mekong in Cambodia where some of the WEF nexus dynamics are paramount (Allouche et al., 2015; Lebel & Lebel, 2018; Pittock et al., 2016) based on criteria that made it more likely we would capture WEF nexus dimen- sions in the models. Dependence on natural resources and proximity to water bodies had to be high. We wanted some recent technology adoption to be present, to capture potential emerging adaptive dynamics as a region that has had limited electricity connections but seeks to increase energy access. Furthermore, alternative livelihoods, including resource‐dependent and resource‐independent incomes, had to be present. Therefore, we chose one village with solar photovoltaic system (PV) use (Damrae, G1), while the other has ecotourism (Koh Phdao, G2). Representative sampling within each community is hard to achieve. Although community‐level data are available to identify numbers of people of age cohorts, employment, and so on, there is no name list that could be used for randomized sampling. In addition, we needed a subsample of farmers and ﬁshers. Given that real‐world group sampling procedures are frequently nonrandom by design and can yield valid infer- ences if carefully analyzed (Harrison & List, 2004; Smith, 1983), we decided to target heterogeneity on the basis of age, gender, and education, including village chiefs' subjective judgments about participants' knowl- edge about farm and ﬁsher livelihoods. Among the villagers, we selected participants from the ﬁshing and farming communities, with some of them ofﬁcially employed and involved in political processes at the com- mune level. We aimed to sample a maximum heterogeneity along the dimensions of gender, age, education, and wealth. The objective was to integrate a variety of perspectives into the procedure with the underlying assumption being that demographic heterogeneity is related to variation in expectations and therefore strengthens the hypothesized effects of the intervention. We did not test for the effects of diversity but rather controlled for some of the potential effects of demographics through related survey items (see the supporting information section S3 for an analysis of group and gender effects). A purposefully heterogeneous composition was achieved by stratiﬁed sampling to guide selection to include members of farming and ﬁshing households of both genders and different ages. 2.3. Place‐Based Case Selection, Sampling Procedure, and Participant Identiﬁcation The village chief's local knowledge, together with local ﬁeld expert interviews, and review of demographic information and existing infrastructure in the Cambodia Commune Database (National Committee for Sub‐National Democratic Development, 2015) and Commune Investment Plan 2014 (Kratie Provincial Department of Planning, 2014) supported the selection. The data sources are formal government data and are not always freely avail- able or up to date. We accessed them through personal contacts in the Ministry of Interior and the Provincial Administration in Kratie. 2.4. The Participatory Modeling Intervention Our intervention protocol describes and simulates the current livelihood and environmental condition situa- tion in the region through explicit scenario modeling, using participants' knowledge and facilitating dialo- gue about future risks and potential actions to mitigate or adapt to these. The type of participatory modeling matters (Davies et al., 2015). We employed a CLD procedure that involves stakeholders selecting essential variables, developing causal effects between them, and then simulating development of these relationships into the future through scenario analysis (Kopainsky et al., 2017). Though simulation is rarer, CLD mapping has been applied in many cases (Hovmand et al., 2012). A com- bination of socio‐ecological system indicator selection and participatory simulation has been recently applied to coastal management in the Dutch Wadden Sea (Vugteveen et al., 2015). A matrix approach to facilitating causal link development has been applied to coastal management in Egypt (Sanò et al., 2014), and individual model aggregation has been addressed in a participatory CLD mapping and simulation pro- cess on the Volta River Basin (Kotir et al., 2017). The optimal group size has been reported at around 15 participants in the literature (Phillips & Phillips, 1993). Accordingly, two CLD mapping and simulation exercises were conducted with two groups of 15 par- ticipants from the two different villages sampled for the intervention. The third intervention was a mixed KIMMICH ET AL. KIMMICH ET AL. 5 Earth's Future 10.1029/2019EF001311 group of 16 participants consisting of inhabitants from both villages, without including any participant from the other two groups. With this mixed group, we tested for the effects of integration between the two village perspectives (PV use and ecotourism) on the CLD intervention (see supporting information for details of group comparison). group of 16 participants consisting of inhabitants from both villages, without including any participant from the other two groups. With this mixed group, we tested for the effects of integration between the two village perspectives (PV use and ecotourism) on the CLD intervention (see supporting information for details of group comparison). The CLD intervention design was grounded in the participatory system dynamics literature but further adapted to groups at the community level with low formal educational background (Kopainsky et al., 2017). We avoided computer simulation and mapped with physical materials (Hovmand et al., 2012) that included a hands‐on simulation exercise (Kopainsky et al., 2017). Participants engaged in identifying key variables, mapping relationships between them. 3.1. Intervention Design for Hypothesis Testing and Data Collection We developed a novel evaluation procedure based on an empirically supported assumption that peoples' expectations about the future drive their behavior (Delavande et al., 2011b; Jensen, 2010; Kimmich & Fischbacher, 2016; McKenzie et al., 2013), hypothesizing that the learning enabled by participatory model- ing (Hedelin et al., 2017) reduces uncertainty by updating participant expectations and increasing their agency to respond to future change. Our design involves a within‐subjects treatment with before and after surveys. We measured expectations before and after the intervention, building on behavioral economic the- ory (Manski, 2004) and recent empirical evidence (Delavande et al., 2011b; Jensen, 2010; Kimmich & Fischbacher, 2016; McKenzie et al., 2013). The intervention design and survey were pretested in February and conducted in March 2018. All participants received the same treatment, which took approximately 6 hr in total. We tested general treatment efﬁcacy with the difference between a pretreatment and posttreatment mea- surement of expectations as behavioral proxies, provided by the surveys. Differential efﬁcacy was measured as relative efﬁcacy between the three groups, comparing the two villages and the mixed group. Note that we did not measure differential efﬁcacy between the CLD intervention and any other form of intervention, which would require respective control groups. We elicited subjective expectations to measure prospective behavioral change. The method of eliciting expec- tations with subjective probabilities is preferable over Likert scales or other elicitation methods (Manski, 2004). In the developing economy context, even with illiterate subjects, subjective probability measures have been shown to be understood, internally consistent, and frequently highly accurate when compared to actual realizations and past experiences (Delavande et al., 2011a). Visual aids help improve accuracy of returns, especially in mixed cultural and linguistic contexts (Delavande et al., 2011b). We used predeﬁned binary classes and used coffee beans as a measure to support elicitation. An elicitation pretest suggested that the tactile dimension of using coffee beans leads to a more reﬂected weighing than percentage number elicitations. A test for potential biases is still missing, however. We did not incentivize questions with payments to expectations answers, which is in line with common practice and supported by empirical ﬁndings (Delavande et al., 2011b). We tested whether participants understood the concept. The ﬁrst three questions and related expectations items tested the two extremes and equal probabilities. Table 2 provides descriptive statistics for the expecta- tions items. 2.4. The Participatory Modeling Intervention The current situation and related problems were then simu- lated using glasses of water to depict stocks and ﬂows to represent time‐dependent change in stock levels to develop and explore future scenarios collectively. Afterward, solutions were identiﬁed and grouped accord- ing to actions by individual and commune‐level actors and then simulated for future outcomes. Using scripts that serve as replicable protocols (Hovmand et al., 2012), the intervention was led by one group facilitator and one modeler/technical advisor and guided by several research assistants acting as process coa- ches (Hovmand, 2014). The CLD procedure lasted between 3 and 4 hr. Figure 1 provides an example of a CLD produced by Group 3 (G3). The protocol, the underlying scripts, and the Group 1 (G1) and Group 2 (G2) outcomes are included in the supporting information. 3.1. Intervention Design for Hypothesis Testing and Data Collection Figures 1 and 2 provide box plots of the distributions before and after the treatment and the changes, respectively. The results were supportive. The median corresponded to the expected value in all three questions (see Table 2). The mean indicates that there was nevertheless a share of farmers who did KIMMICH ET AL. KIMMICH ET AL. 6 Earth's Future 10.1029/2019EF001311 Figure 1. A causal loop diagram produced by Group 3 in our intervention in approximately 45 min of discussion and deliberation (see Script S4 in the supporting information). It depicts key variables and the cause‐and‐effect relationships linking them, as identiﬁed by the group participants with the support of facilitators. Figure 1. A causal loop diagram produced by Group 3 in our intervention in approximately 45 min of discussion and deliberation (see Script S4 in the supporting information). It depicts key variables and the cause‐and‐effect relationships linking them, as identiﬁed by the group participants with the support of facilitators. not expect the sun to rise with certainty and did expect the river to run dry. There was also a considerable share of participants (42.86%) who expected the next child in their household to be a boy more or less likely than did 50%. Overall, however, there has been a learning effect, as the mean moved toward the median and the standard deviation (SD) decreased for each of the three questions. We also tested whether participants understood nested probabilities using an adapted set of questions con- cerning the occurrence of a drought in Cambodia (like the one in 2015/2016) within 5 and 10 years, respec- tively. Two farmers had inconsistent expectations, stating that a drought within 5 years was more likely than was a drought within 10 years. This deviation of 4.35% is slightly higher than what has been found in another consistency test (Delavande et al., 2011b). We elicited the following variables to control for demographic and household characteristics: A set of ques- tions concerning gender, age, and education controls for basic demographics. We also included a set of ques- tions concerning household size and position of the surveyed person in the household. Sources of household income were differentiated to capture income from farming, ﬁshing, tourism, and public and private employment. We also include items on food self‐provision and expenditure for food to account for house- holds' semi‐subsistence. 3.1. Intervention Design for Hypothesis Testing and Data Collection Households' vulnerability to food insecurity and food quality was covered to control for potential effects on expectation measures. For example, Indian farmers' monsoon predictions have been shown to be inﬂuenced by their dependence on rain‐fed irrigation (Giné et al., 2009). Finally, we also tested whether the CLD intervention could inﬂuence more qualitative expectations dimen- sions unlikely to appear in the model. This allowed us to separate the effects of the CLD intervention from communication effects or information sharing that happens naturally while building the CLD. 4.1. Individual Empowerment Results Recalling that our activation hypothesis proposes that affecting self‐ efﬁcacy in participatory modeling brings about speciﬁc empowerment outcomes (Basco‐Carrera et al., 2017; Diener & Biswas‐Diener, 2005; Gittelsohn et al., 2012), we test if participants become more optimistic about the opportunities directly under their control. “Irrigation pump”: likelihood that someone else or the respondent would purchase a water pump in the next 5 years; “Dam/dike”: small dikes or dams will be constructed by respondent's commune within the next 5 years and anticipated implications for agricultural production; “CIP”: likelihood that the respondent will participate in the CIP process within the next 5 years, that CIP decisions will have a positive effect on household income of the respondent, and that the respondent could inﬂuence a CIP decision. We asked the participants: “How likely do you think it will be that your income from agricultural production (and ﬁshing, respectively) within the next ﬁve years will increase?” Using the nonparametric Wilcoxon signed‐rank test, we found statistical support for the general income expectation variables over which participants have control. Participants were signiﬁcantly more optimistic concerning their farm‐ and ﬁshing‐ ntion; that is, we can reject the null hypothesis for agricultural income (P ﬁshing (P > 0.002). We asked the participants: “How likely do you think it will be that your income from agricultural production (and ﬁshing, respectively) within the next ﬁve years will increase?” Using the nonparametric Wilcoxon signed‐rank test, we found statistical support for the general income expectation variables over which participants have control. Participants were signiﬁcantly more optimistic concerning their farm‐ and ﬁshing‐ related incomes after the intervention; that is, we can reject the null hypothesis for agricultural income (P > 0.008) as well as income from ﬁshing (P > 0.002). y l production; “CIP”: likelihood that CIP process within the next 5 years, effect on household income of the ould inﬂuence a CIP decision. signed rank test, we found statistical support for the general income expectation variables over which participants have control. Participants were signiﬁcantly more optimistic concerning their farm‐ and ﬁshing‐ related incomes after the intervention; that is, we can reject the null hypothesis for agricultural income (P > 0.008) as well as income from ﬁshing (P > 0.002). We also tested whether such increased optimism could derive from speciﬁc agricultural techniques, rather than from a more general empowerment effect. 4. Results and Discussion Figure 2. Box plots of the changes in expectations for our key variables. “Drought”: likelihood of a drought like in 2015/2016 reoccurring; “Take over by child”: likelihood that the family farm or ﬁshing activity will be taken over by someone from the family; “Agr. Income”: likelihood that household income earned through agricultural activities will increase; “Fish income”: likelihood that household income earned through ﬁshing activities will increase; “Food expenditure”: likelihood that the share of food bought rather than produced by the household will increase in the next 5 years; “Corn”: likelihood the household would earn more from planting corn than from rice in the next season, and then under drought conditions; “Organic fert.”: likelihood that yields would be higher with organic fertilizers than with conventional (chemical) fertilizers, and expected effects on food qual- ity; “Fishpond”: likelihood that a new ﬁshpond will be constructed in the commune within the next 5 years, by the respondent or someone else, and the anticipated implications for food quality and household income; “Irrigation pump”: likelihood that someone else or the respondent would Instrumental sample selection provided an equal distribution along gen- der and heads of households and a large variation in terms of age and edu- cation (Table S1). Standard consistence tests used in expectation elicitation procedures, testing extremes, equal, and nested probabilities assured that participants understood the concept (see consistence items in Table 1 and section 2). 3.2. Data Analysis Strategy Note that due to the small sample size, the power of the tests is relatively low. The statistical test results provide only preliminary support for the hypotheses. Wilcoxon signed‐rank test (Harris & Hardin, 2013), which supported pre- vious ﬁndings. Note that due to the small sample size, the power of the tests is relatively low. The statistical test results provide only preliminary support for the hypotheses. One assumption of the Wilcoxon test is independent sampling within and between groups. This assumption is violated with clustered data of the type we produced in our three intervention groups. Newson's method to calculate conﬁdence intervals and Somers' D can account for clustered data, however (Newson, 2002). We therefore compared our results with the results from Newson's method, comparing them with both clustered and nonclustered results. Table S3 in the supporting information reports Somers' D without clustering, which can be compared with the Wilcoxon signed‐rank test results in the main text. 3.2. Data Analysis Strategy Employing nonparametric tests instead of t tests is recommended with small samples and when t‐test distri- butional assumptions are likely to be violated. A Shapiro‐Wilk test for normality (see Table S2 in the support- ing material) showed that there is a statistically signiﬁcant deviation from normality for the tests not only for extremes (sunrise and river) but also for expectations related to ﬁsh income, organic fertilizer effects on income and food quality, others' and own decision to invest into a pump, and participation in the annual CIP. We therefore chose nonparametric tests, using the Wilcoxon signed‐rank test for paired samples, which tests for a zero difference of the median before and after the treatment. We compare the results with the exact KIMMICH ET AL. 7 KIMMICH ET AL. Earth's Future 10.1029/2019EF001311 related incomes after the interve Figure 2. Box plots of the changes in expectations for our key variables. “Drought”: likelihood of a drought like in 2015/2016 reoccurring; “Take over by child”: likelihood that the family farm or ﬁshing activity will be taken over by someone from the family; “Agr. Income”: likelihood that household income earned through agricultural activities will increase; “Fish income”: likelihood that household income earned through ﬁshing activities will increase; “Food expenditure”: likelihood that the share of food bought rather than produced by the household will increase in the next 5 years; “Corn”: likelihood the household would earn more from planting corn than from rice in the next season, and then under drought conditions; “Organic fert.”: likelihood that yields would be higher with organic fertilizers than with conventional (chemical) fertilizers, and expected effects on food qual- ity; “Fishpond”: likelihood that a new ﬁshpond will be constructed in the commune within the next 5 years, by the respondent or someone else, and the anticipated implications for food quality and household income; “Irrigation pump”: likelihood that someone else or the respondent would purchase a water pump in the next 5 years; “Dam/dike”: small dikes or dams will be constructed by respondent's commune within the next 5 years and anticipated implications for agricultural production; “CIP”: likelihood that the respondent will participate in the CIP process within the next 5 years, that CIP decisions will have a positive effect on household income of the respondent, and that the respondent could inﬂuence a CIP decision. Wilcoxon signed‐rank test (Harris & Hardin, 2013), which supported pre- vious ﬁndings. 4.1. Individual Empowerment Results Neither ﬁshpond income (P > 0.92) nor effects of the CIP on income (P > 0.75) contributed to this optimism. With the help of Somers' D, we can also interpret the effect size. Concerning agricultural income, for example, participants are between 11% and 42% more likely to expect their income from agriculture to increase than to decrease after the CLD intervention (see Table 2). The Wilcoxon signed‐rank test ﬁndings are supported by Newson's method (see section 2), accounting for clustered data across the three intervention groups (see Table 2). While agricultural and income expectations changed signiﬁcantly, the source of this optimism is neither directly related to ﬁshpond nor related to CIP income expectations, where no signiﬁcant changes could be observed. Increasing optimism about children taking over farming or ﬁshing activities is in line with more optimistic income expectations. We also found statistically signiﬁcant changes in other expectation variables modeled by the participants. We found a decrease in expected likelihood that somebody in the community will invest in a pump and that the commune will invest in a ﬁshpond. The ﬁndings concerning raised expectations for earnings from corn planting during droughts are not robust when compared with other model speciﬁcations (see nonclustered results in Table S3). KIMMICH ET AL. KIMMICH ET AL. 8 Earth's Future 10.1029/2019EF001311 Table 1 Summary Statistics of the Expectation Variables Before and After the Intervention for Three Consistence Tests and 21 Key Variables Variable N Before After Mean SD Med Min Max Mean SD Med Min Max consistence: sunrise 42 7.17 4.15 10 0 10 8.19 3.12 10 0 10 consistence: river 42 2.10 3.47 0 0 10 0.98 2.40 0 0 10 consistence: boy 42 4.45 1.98 5 0 10 4.48 1.71 5 0 9 drought (5y) 46 4.98 2.13 5 0 10 5.20 1.68 5 1 10 drought (10y) 46 6.02 2.71 6 0 10 6.24 1.99 6 3 10 takeover by child 46 5.09 2.78 5 0 10 5.76 2.46 5 1 10 agr. income (5y) 46 5.15 2.25 5 0 10 6.13 1.68 6 3 10 ﬁsh income (5y) 46 2.70 2.18 3 0 8 4.02 2.15 4 0 8 food expenditure (5y) 46 5.04 2.39 5 0 10 4.91 1.56 5 1 7 corn: earnings 46 3.70 2.23 4 0 8 4.33 2.12 4 0 9 corn: earn. 4.1. Individual Empowerment Results with drought 46 3.80 2.62 3.5 0 10 3.89 2.49 4 0 8 organic fert.: yield 46 7.35 2.51 8 0 10 7.00 2.42 8 0 10 organic fert.: food quality 46 8.76 1.90 10 0 10 8.30 1.58 8 4 10 ﬁshpond: commune 46 5.50 2.18 5 0 10 5.02 2.18 5 0 10 ﬁshpond: own 46 5.17 3.10 5 0 10 5.17 2.60 5 0 10 ﬁshpond: food quality 46 6.98 1.89 7 2 10 6.57 1.60 7 2 9 ﬁshpond: income 46 6.61 1.89 7 2 10 6.57 2.12 7 1 10 irrigation pump: others 46 6.78 2.70 7 0 10 6.15 2.61 7 0 10 irrigation pump: own 46 6.74 3.30 7 0 10 7.17 2.96 8 0 10 dam/dike commune 46 3.96 2.81 5 0 10 4.24 2.58 5 0 10 dam/dike yield increase 46 6.30 2.05 7 1 10 6.43 2.21 7 0 10 CIP: participation 46 7.46 2.33 8 0 10 7.13 2.17 7.5 2 10 CIP: income effects 46 6.20 2.48 6.5 1 10 6.30 2.00 7 2 10 CIP: own inﬂuence 46 5.33 2.39 5 0 10 5.48 2.27 5 1 10 We also tested for variables that were unlikely to be modeled explicitly in a CLD framework to identify the potential effects of communication during the modeling process. We found a signiﬁcant decrease in expected food quality from the application of organic fertilizers and suppressed expectations about food quality from ﬁshpond investments (although less statistically robust, see Table S3). Given that food quality was not mod- eled by any of the groups (see section 2), this result shows that the CLD mapping and simulation was not the only instrument shaping expectations during the intervention. Participants went beyond the boundary object of the CLD intervention to form new expectations independently for variables not discussed by the group. In brief, we found support for the activation hypothesis. The optimism concerning agricultural and ﬁshing income seems to be general and cannot be traced to any speciﬁc agricultural method, such as pump or ﬁsh- pond investments, or earnings from corn. Therefore, we cannot trace this optimism to any speciﬁc produc- tion. Importantly, this increase in optimism is not in line with other ﬁndings on agricultural income expectations in Cambodia (Grafton et al., 2016; Scheidel et al., 2014). 4.1. Individual Empowerment Results Rather, this optimism appears to be a more general phenomenon resulting from having participated in the CLD procedure. 4.2. Shared Expectations and Collective Action Results The table shows the results of testing the hypothesis that participants become more optimistic about variables on which they have direct inﬂuence, as com- pared to those variables that are subject to external forces (activation hypothesis). Note. The table shows the results of testing the hypothesis that participants become more optimistic about variables on pared to those variables that are subject to external forces (activation hypothesis). intervention to analyze expectation convergence. Overall, we found support for the shared expectations hypothesis. Table 2 indicates that the SD was reduced in almost all cases, except for expected income increase from ﬁshponds. We used Levene's test to identify whether the variances changed after the intervention, including the median modiﬁcations of the Brown‐Forsythe test. The null hypothesis was moderately rejected for food expenditure (PW0 > 0.056; PW50 > 0.055; PW10 > 0.066), but not for any of the other items. intervention to analyze expectation convergence. Overall, we found support for the shared expectations hypothesis. Table 2 indicates that the SD was reduced in almost all cases, except for expected income increase from ﬁshponds. We used Levene's test to identify whether the variances changed after the intervention, including the median modiﬁcations of the Brown‐Forsythe test. The null hypothesis was moderately rejected for food expenditure (PW0 > 0.056; PW50 > 0.055; PW10 > 0.066), but not for any of the other items. Another proxy variable supports the argument that collective action at the commune level was affected by the CLD mapping and simulation intervention. Participants judged it less likely that they would participate in the CIP again in their posttest survey (see Table 3). It may be that an empowerment effect related to self‐ organization possibilities creates pessimism about their individual and collective agency in the CIP. This argument could be supported by recent ﬁndings concerning the CIP in Cambodia (Plummer & Tritt, 2012), but the political‐economic dimension of this needs further theorization in WEF nexus research (Foran, 2015). This is particularly relevant for local perceptions of provincial administration capacities for natural resources management, which has been recently devolved to provinces under D&D reforms (Vuković & Babović, 2018). There is limited application of risk‐based management in Cambodia with low availability and sharing of local risk information (Mochizuki et al., 2015), suggesting that it was our intervention that reduced uncer- tainty about sources of potential risk for our groups, though not uniformly, across variables. 4.2. Shared Expectations and Collective Action Results Shared expectations are an important prerequisite for collective action (Runge, 1986) and therefore funda- mental for institutional innovation. Once participants converge on the likelihood of an event or outcome, support for an institutional solution becomes common knowledge that facilitates collective choice. We assume that joint exploration of system properties in a participatory modeling protocol leads to shared men- tal models (Scott et al., 2016a) and hypothesize that convergence dominates divergence in expectations‐ forming across participants. We focus on uncertainty at the group level rather than changes in individuals' uncertainty, which would require expectation elicitation of multiple intervals, exceeding survey time con- straints when multiple expectation variables are measured. The results show that the CLD intervention has contributed to some convergence in expectations and a related reduction in uncertainty in our groups. We compared the SD of each variable before and after the KIMMICH ET AL. KIMMICH ET AL. 9 9 Earth's Future 10.1029/2019EF001311 Table 2 Within‐Cluster Somers' D With Standard Errors Adjusted for the Three Groups CLD_treatment Coefﬁcient Jknife SEs z P > z 95% conﬁdence interval drought (5y) 0.0623229 0.0935694 0.67 0.505 −0.1210698 0.2457156 drought (10y) −0.0028329 0.0790037 −0.04 0.971 −0.1576772 0.1520115 takeover by child 0.1331445 0.0738423 1.80 0.071 −0.0115837 0.2778727 agr. income (5y) 0.2677054 0.0790918 3.38 0.001 0.1126882 0.4227225 ﬁsh income (5y) 0.3541076 0.1022549 3.46 0.001 0.1536918 0.5545235 food expenditure (5y) −0.0524079 0.1248944 −0.42 0.675 −0.2971965 0.1923806 corn: earnings 0.1657224 0.1362876 1.22 0.224 −0.1013964 0.4328411 corn: earn. with drought 0.0410765 0.0233324 1.76 0.078 −0.0046541 0.0868071 organic fert.: yield −0.0963173 0.1450005 −0.66 0.507 −0.3805131 0.1878785 organic fert.: food quality −0.2294618 0.0467748 −4.91 0.000 −0.3211387 −0.1377848 ﬁshpond: commune −0.0991501 0.0470523 −2.11 0.035 −0.191371 −0.0069293 ﬁshpond: own −0.0212465 0.0159911 −1.33 0.184 −0.0525885 0.0100955 ﬁshpond: food quality −0.1458924 0.0834862 −1.75 0.081 −0.3095223 0.0177376 ﬁshpond: income 0.0254958 0.0967985 0.26 0.792 −0.1642259 0.2152174 irrigation pump: others −0.1203966 0.0479123 −2.51 0.012 −0.2143031 −0.0264901 irrigation pump: own 0.0538244 0.0553522 0.97 0.331 −0.0546641 0.1623128 dam/dike: commune 0.0694051 0.0799335 0.87 0.385 −0.0872616 0.2260718 dam/dike: yield increase 0.0609065 0.0625659 0.97 0.330 −0.0617204 0.1835335 CIP: participation −0.1033994 0.0360144 −2.87 0.004 −0.1739864 −0.0328125 CIP: income effects 0.0354108 0.0355212 1.00 0.319 −0.0342095 0.1050311 CIP: own inﬂuence −0.0042493 0.0297723 −0.14 0.887 −0.0626019 0.0541033 Note. The table shows the results of testing the hypothesis that participants become more optimistic about variables on which they have direct inﬂuence, as com- pared to those variables that are subject to external forces (activation hypothesis). Note. 4.3. Discussion—The Implications of Changes and Convergence in Expectations for Nexus Governance in Mekong We set out to test the idea that reducing uncertainty and building consensus on action through participatory modeling could potentially catalyze new behaviors with this research. We hypothesized that participant expectations for (1) the likelihood of certain future events and (2) their own individual agency to respond to these events will change because of the CLD intervention. We also postulate that participatory modeling reduces uncertainty and enables convergence in expectations across members of the group. Our expectations‐testing results suggest that these ideas may be true in some interesting ways. The research team achieved a new understanding of WEF nexus risk anticipation in local stakeholders. We entered into KIMMICH ET AL. KIMMICH ET AL. 10 Earth's Future 10.1029/2019EF001311 Table 3 Mean, Median, and SD Differences Across Groups With Regard to Changes in Expectations Variable Mean Median SD_d G1 G2 G3 G1 G2 G3 G1 G2 G3 consistence: sunrise −0.27 2.92 −0.10 0.00 0.00 0.00 −0.60 −2.02 0.01 consistence: river −0.36 −2.17 0.27 0.00 0.00 0.00 −0.68 −1.60 −1.03 consistence: boy 0.00 0.00 0.09 0.00 0.00 0.00 −0.07 −0.20 −0.50 drought (5y) 0.20 −0.25 0.73 0.00 0.00 1.00 −0.39 −0.29 −0.43 drought (10y) 0.00 0.12 0.53 0.00 0.50 1.00 −0.31 −0.78 −1.16 takeover by child 0.40 0.19 1.47 0.00 0.00 1.00 −0.69 −0.04 −0.07 agr. income (5y) 1.07 0.25 1.67 1.00 0.00 1.00 −0.72 −0.71 −0.41 ﬁsh income (5y) 0.73 1.94 1.27 0.00 2.00 2.00 0.21 −0.07 −0.04 food expenditure (5y) −0.67 −0.44 0.73 0.00 −0.50 0.00 −0.65 −1.00 −0.59 corn: earnings 1.60 0.38 −0.07 1.00 0.00 0.00 0.17 −0.53 0.02 corn: earn. 4.3. Discussion—The Implications of Changes and Convergence in Expectations for Nexus Governance in Mekong Moreover, participants became signiﬁcantly more optimistic concerning their individual agency to cope (i.e., increase agricultural and ﬁshing income) in the face of such risks and, conversely, less enthusiastic about the likelihood that they will participate in their local upcoming CIP processes. The interpretation we offer is that the CIP results reﬂect the ongoing issue of the state's inability to provide communities' access to key resources for their development that the D&D reforms have been established to remedy. Local citizens participate in the CDP/CIP processes because there is no alternative. But if, as our results suggest, the sense of individual and collective agency is heightened, people will perhaps decide not to participate in the CIP. The convergence in expectations for multiple variables across the groups suggests that while collective solutions and institutionalized agreements are more likely to emerge, they are less likely to be realized through the CIP given expectations about reduced participation. Other community planning activities, both formal (e.g., Department of Agricultural Extension (DAE), 2012) and informal, may hold more promise for local WEF nexus governance. this work understanding that dams were at the heart of the WEF nexus in this region. Yet hydroelectric dam construction was not the primary risk identiﬁed for our participants. Local villagers consider themselves to be facing more immediate challenges to food security, like illegal ﬁshing and locally degraded ﬁsh stocks, while dam developments are far‐off concerns. Moreover, participants became signiﬁcantly more optimistic concerning their individual agency to cope (i.e., increase agricultural and ﬁshing income) in the face of such risks and, conversely, less enthusiastic about the likelihood that they will participate in their local upcoming CIP processes. The interpretation we offer is that the CIP results reﬂect the ongoing issue of the state's inability to provide communities' access to key resources for their development that the D&D reforms have been established to remedy. Local citizens participate in the CDP/CIP processes because there is no alternative. But if, as our results suggest, the sense of individual and collective agency is heightened, people will perhaps decide not to participate in the CIP. The convergence in expectations for multiple variables across the groups suggests that while collective solutions and institutionalized agreements are more likely to emerge, they are less likely to be realized through the CIP given expectations about reduced participation. 4.3. Discussion—The Implications of Changes and Convergence in Expectations for Nexus Governance in Mekong with drought −0.07 0.06 0.27 0.00 0.00 0.00 −0.34 0.29 −0.38 organic fert.: yield 0.20 −1.06 −0.13 1.00 0.00 −1.00 −0.20 −0.60 0.55 organic fert.: food quality −0.40 −0.19 −0.80 0.00 0.00 −1.00 0.16 −1.24 0.22 ﬁshpond: commune −0.60 −0.31 −0.53 0.00 0.00 0.00 0.16 −0.57 0.63 ﬁshpond: own 0.20 −0.12 −0.07 0.00 0.00 0.00 −0.49 −0.49 −0.56 ﬁshpond: food quality −0.67 −0.75 0.20 −1.00 −1.00 0.00 −0.62 −0.34 −0.05 ﬁshpond: income 0.13 −0.44 0.20 0.00 −0.50 0.00 0.37 −0.02 0.29 irrigation pump: others −0.53 −0.25 −1.13 0.00 0.00 0.00 0.71 −0.37 −0.36 irrigation pump: own −0.20 0.00 1.53 0.00 0.00 0.00 −1.48 0.10 0.31 dam/dike commune −0.80 0.69 0.93 0.00 0.00 1.00 0.63 −0.17 −0.90 dam/dike yield increase 0.60 0.19 −0.40 0.00 0.00 0.00 0.72 0.11 −0.58 CIP: participation −0.33 −0.06 −0.60 0.00 0.00 0.00 0.24 −0.44 −0.15 CIP: income effects −0.13 0.25 0.20 0.00 0.00 0.00 −0.33 −0.50 −0.63 CIP: own inﬂuence 0.27 0.12 0.07 0.00 0.00 0.00 −0.31 0.29 −0.31 Note. Mean and median values of the change are reported, while SD_d reports the difference in standard deviations before and after the intervention. N = 15 for G1 and G2 and N = 16 for G3. The table summarizes the results of differences in expectation changes across the groups (shared expectations hypothesis), with changes in standard deviations provided in the last three columns. Table 3 Mean, Median, and SD Differences Across Groups With Regard to Changes in Expectations Note. Mean and median values of the change are reported, while SD_d reports the difference in standard deviations before and after the intervention. N = 15 for G1 and G2 and N = 16 for G3. The table summarizes the results of differences in expectation changes across the groups (shared expectations hypothesis), with changes in standard deviations provided in the last three columns. this work understanding that dams were at the heart of the WEF nexus in this region. Yet hydroelectric dam construction was not the primary risk identiﬁed for our participants. Local villagers consider themselves to be facing more immediate challenges to food security, like illegal ﬁshing and locally degraded ﬁsh stocks, while dam developments are far‐off concerns. 4.3. Discussion—The Implications of Changes and Convergence in Expectations for Nexus Governance in Mekong Other community planning activities, both formal (e.g., Department of Agricultural Extension (DAE), 2012) and informal, may hold more promise for local WEF nexus governance. Our evidence regarding the activation and convergence hypotheses is in line with previous studies from a wide range of application domains that evaluate the effectiveness of participatory system dynamics (e.g., Rouwette, 2003; Rouwette et al., 2011; Scott et al., 2016b). Those studies focused on more traditional settings for participatory modeling such as corporate or public policy applications in board room environments. The observed alignment of mental models seems to hold also in the context of groups at the community level with low formal educational background (convergence hypothesis). Similar to those of Rouwette (2003) and Rouwette et al. (2011), our participants became signiﬁcantly more optimistic about individual opportu- nities under their control. While existing studies found no change in perceived behavioral control, our 11 KIMMICH ET AL. KIMMICH ET AL. Earth's Future 10.1029/2019EF001311 results point at an overall increase in the likelihood of individual as well as collective action. Future research will have to address the question whether these differences in outcomes are an artifact of somewhat different evaluation criteria used (perceived behavioral control versus expectations regarding future individual and collective action) or whether they are a function of either our participatory modeling protocol or the speciﬁc WEF nexus context in which we were operating. results point at an overall increase in the likelihood of individual as well as collective action. Future research will have to address the question whether these differences in outcomes are an artifact of somewhat different evaluation criteria used (perceived behavioral control versus expectations regarding future individual and collective action) or whether they are a function of either our participatory modeling protocol or the speciﬁc WEF nexus context in which we were operating. Some caveats are important to interpreting our ﬁndings. In our experience, the CLD process generated dis- cussion about other things than what were modeled, and so there is some leakage into issues and topics that were not controlled for in the process. For example, some expectation effects were observed for food quality from organic fertilizer use and related to aquaculture, which were not actually modeled in the CLD proce- dures. Secondly, differences across our groups were considerable. Table 3 provides differences across groups concerning median and mean changes in expectations. 5. Conclusions Our starting point in this paper was that two assumptions drive much participatory research in sustainabil- ity: that knowledge produced through participatory processes is more likely to identify transformational leverage points (Cash et al., 2003; Star & Griesemer, 1989) and that such knowledge and processes can directly and indirectly inﬂuence behavior change for sustainability (Gerritsen et al., 2013). Given the various issues for knowledge integration while engaging with values, power, and politics (Jordan et al., 2018; Miller & Wyborn, 2018; Wesselink et al., 2013), we argued that such assumptions need testing. Our analysis contributes to explaining one possible mechanism of behavioral change resulting from systems thinking‐ and modeling‐based interventions. Speciﬁcally, sustainability research suggests that participatory modeling can trigger self‐efﬁcacy outcomes and reduce perceived uncertainties in addressing collective action challenges. Our empirical results provide preliminary support for this claim and for a recent proposi- tion in behavioral sciences that intervening to change expectations can drive changes in behavior. Methodologically, we tested a novel evaluation framework for behavioral outcome evaluation in sustainabil- ity interventions. We measured positive and negative changes in expectations of participants in our proce- dure in line with the hypothesis that the participatory CLD mapping and simulation could affect their perceived self‐efﬁcacy (activation hypothesis). Our participants also demonstrated convergence in some key expectations about likelihood of future events and capacities to respond to these (shared expectations hypothesis). This is signiﬁcant because reducing uncertainties at the group level removes some barriers impeding the complex adaptation among a large set of stakeholders with all the attendant uncertainties, pol- itics, and power dynamics that WEF nexus governance demands. Our intervention effectively coproduced a new awareness of future individual and local shared risks that may follow global environmental change, regional and national energy strategies, and technical decisions on hydropower dam siting and design and ideas for potential responses to manage these in our case study participants, despite driver factors for the coming changes lying largely outside of local control. These results suggest three implications for participa- tory, model‐based sustainability research and practice in general, and WEF nexus governance in speciﬁc. The ﬁrst signiﬁcant point for designing participatory processes that are intended to generate governance out- comes is how updating participants' expectations does not rely on outputs produced and published after interventions. It appears to be at least as much about the process as the end product. 4.3. Discussion—The Implications of Changes and Convergence in Expectations for Nexus Governance in Mekong Although SD measures decreased across groups in most cases, there were differences across groups, and SD of some groups actually increased for several vari- ables. This suggests that the development of shared expectations is by no means an unconditional result of the CLD intervention but is more likely contingent upon participants' contributions that are in turn inﬂu- enced by internal group dynamics, or major external events, rather than intended intervention effects within the scope of the intervention protocol itself. Our cases were all treated on the same day in the same location with the same team, so we controlled for some inﬂuences; but participants did come from different commu- nities, and group dynamics changed in each group with personalities. Finally, we can only say something about the uncertainties that we actually measured and cannot infer results for reducing uncertainties con- cerning variable coverage and model structure. 5. Conclusions The conceptual system interpretation and dynamic simulation in our CLD procedure changes some conditions (risk awareness and income expectations) that can affect future behavior of those participating during the 6‐hr workshop process. This suggests a second insight on the importance of paying close attention to how expectations are being KIMMICH ET AL. 12 KIMMICH ET AL. Earth's Future 10.1029/2019EF001311 updated during the group activity in case these are wrong. For example, the individual increases in optimism for income in our groups are not in line with other ﬁndings on agricultural income expectations in Cambodia (Grafton et al., 2016; Scheidel et al., 2014). Not only did our intervention change expectations for the future, but it changed them in a way that could be risky for participants if they are wrong. The impli- cation of this is that mechanisms to evaluate and navigate expectations updating, convergence, and diver- gence are needed during participatory modeling procedures, not afterward. The method we used could reduce the risk of unknowingly changing expectations against contradicting evidence while also supporting sustainability researchers to link their intervention and future potential behavior change impacts. updated during the group activity in case these are wrong. For example, the individual increases in optimism for income in our groups are not in line with other ﬁndings on agricultural income expectations in Cambodia (Grafton et al., 2016; Scheidel et al., 2014). Not only did our intervention change expectations for the future, but it changed them in a way that could be risky for participants if they are wrong. The impli- cation of this is that mechanisms to evaluate and navigate expectations updating, convergence, and diver- gence are needed during participatory modeling procedures, not afterward. The method we used could reduce the risk of unknowingly changing expectations against contradicting evidence while also supporting sustainability researchers to link their intervention and future potential behavior change impacts. A third and ﬁnal key insight for WEF nexus governance is that convergence and divergence in expectations point to clear areas of certainty and uncertainty in groups, which is critical to catalyzing opportunities and conditions for collective action. Group interpretation is an epistemic nudge, entailing deliberative processes of empowerment and so potentially strengthening agency and enabling collective action (John, 2018). This information can identify individuals primed for interventions, for example, technical training or further negotiations, on issues where uncertainty is a barrier. 5. Conclusions It also means that while heterogeneity in group mod- eling is advised to stimulate discourse, any participatory modeling exercise should carefully weigh advan- tages and disadvantages where collective action is the goal. In too homogenous groups, convergence may be achieved without surfacing and challenging assumptions. Conversely, groups too heterogeneous by design can produce unwarranted divergence in expectations, increasing uncertainty and reducing agreement. One important caveat in our results is permanence of change and convergence in expectations. Structural power issues will undermine empowerment gains made in our process without further interventions. In this regard, future research could explore if different forms of research designs, participatory approaches, and group compositions have different effects on expectations, paying attention to power dynamics. Note that without a systematic experimental comparison of group compositions and intervention methods, the exter- nal validity of our ﬁndings remains limited and open to further analyses. Testing the effectiveness of differ- ent group selection procedures, including a random selection baseline, and their relation to participatory approaches, could increase external validity. For example, we chose an optimal demographic composition and group size, but future studies could experimentally test inﬂuence of group heterogeneity, leadership, group sizes, and many other variables on behavioral effects of CLD interventions, or any other workshop technique or epistemic approach. Acknowledgments The authors declare no competing interests that could directly undermine or be perceived to undermine the objectivity, integrity, and value of this primary research article. Author contributions: C. K. co‐lead on conception and design of this research, data analysis, interpretation of data and drafting the manuscript and supporting information; L. G. co‐lead on conception and design of this research, interpretation of data, drafting the manuscript and supporting information and is the corresponding author; B. K., M. D., C. S., and C. B. contributed to research design, data acquisition, interpretation of data, and drafting the manuscript; and C.B. contributed to conceptualization of the research and to substantial review of the manuscript. The authors gratefully acknowledge the Nomis Foundation and the Mava Foundation as funding partners in this research and in‐kind support from CGIAR Research Program on Fish Agri‐Food Systems (FISH) led by WorldFish. We thank the participants from Kratie Province and the students from the Royal University of Phnom Penh who conducted the survey work during the ﬁeld interventions. We also thank Andrea Betancourt for her project management support and comments on an earlier version of the manuscript. 5. Conclusions The data supporting the results and conclusions in the paper can be found in Tables 1, in Figure 2, and the supporting information Table S4. Key references supporting conclusions include Delavande et al. (2011b) and Runge (1986). We conclude that participatory CLD mapping and simulation is likely to be one effective implementation method for WEF nexus governance processes. We also conclude that expectations measures can be used, with caution, to evaluate the likelihood of future behavioral change following participatory processes. Our contribution lies in developing and testing one mechanism for measuring changes in expectations and show- ing that it can be done in challenging ﬁeld conditions. Abson, D. J., Fischer, J., Leventon, J., Newig, J., Schomerus, T., Vilsmaier, U., et al. (2017). Leverage points for sustainability transforma- tion. Ambio, 46(1), 30–39. https://doi.org/10.1007/s13280‐016‐0800‐y Allan, T., Keulertz, M., & Woertz, E. (2015). The water‐food‐energy nexus: An introduction to nexus concepts and some conceptual and operational problems. International Journal of Water Resources Development, 31(3), 301–311. https://doi.org/10.1080/ , , , , , , Ç , , g , , , ( ) p p interdisciplinary and iterative dialogue in a changing climate. Earth's Future, 3(8), 289–297. https://doi.org/10.1002/2015EF000303 Albrecht, T. R., Crootof, A., & Scott, C. A. (2018). The water‐energy‐food nexus: A systematic review of methods for nexus assessment. Environmental Research Letters, 13(4), 043002. References M., Eckley, N., Guston, D. H., et al. (2003). Knowledge systems for sustainable devel- opment. Proceedings of the National Academy of Sciences, 100(14), 8086–8091. https://doi.org/10.1073/pnas.1231332100 Crane, T. A. (2010). Of models and meanings: Cultural resilience in social–ecological systems. Ecology and Society, 15(4). Davies, K. K., Fisher, K. T., Dickson, M. E., Thrush, S. F., & Le Heron, R. (2015). Improving ecosystem service frameworks to address wicked problems. Ecology and Society, 20(2). Delavande, A., Giné, X., & McKenzie, D. (2011a). Eliciting probabilistic expectations with visual aids in developing countries: How s sitive are answers to variations in elicitation design? Journal of Applied Econometrics, 26(3), 479–497. https://doi.org/10.1002/jae.1 Delavande, A., Giné, X., & McKenzie, D. (2011b). Measuring subjective expectations in developing countries: A critical review and n evidence Journal of Development Economics 94(2) 151163 https://doi org/10 1016/j jdeveco 2010 01 008 Delavande, A., Giné, X., & McKenzie, D. (2011b). Measuring subjective expectations in developing countries: A critical review and new A., Giné, X., & McKenzie, D. (2011b). Measuring subjective expectat Delavande, A., Giné, X., & McKenzie, D. (2011b). Measuring subjective expectations in developing countries: A evidence. Journal of Development Economics, 94(2), 151–163. https://doi.org/10.1016/j.jdeveco.2010.01.008 evidence. Journal of Development Economics, 94(2), 151–163. https://doi.org/10.1016/j.jdeveco.2010.01.008 DeLorme, D. E., Kidwell, D., Hagen, S. C., & Stephens, S. H. (2016). Developing and managing transdisciplinary and transformative research on the coastal dynamics of sea level rise: Experiences and lessons learned. Earth's Future, 4(5), 194–209. https://doi.org/ 10.1002/2015EF000346 Lorme, D. E., Kidwell, D., Hagen, S. C., & Stephens, S. H. (2016). Developing and managing transdisciplinary and transformative research on the coastal dynamics of sea level rise: Experiences and lessons learned. Earth's Future, 4(5), 194–209. https://doi.org/ DeLorme, D. E., Kidwell, D., Hagen, S. C., & Stephens, S. H. (2016). Developing and managing transdisciplinary and transformative research on the coastal dynamics of sea level rise: Experiences and lessons learned. Earth's Future, 4(5), 194–209. https://doi.org/ 10.1002/2015EF000346 Denzau, A. T., & North, D. C. (1994). Shared mental models: Ideologies and institutions. Kyklos, 47(1), 3–31. https://doi.org/10.1111/j.1467 6435.1994.tb02246.x Department of Agricultural Extension (DAE), Ministry of Agriculture, forestry and ﬁsheries (MAff), Cambodia and CGIAr Challenge Department of Agricultural Extension (DAE), Ministry of Agriculture, forestry and ﬁsheries (MAff), Cambodia and CGIAr Challenge program on Water and food (CpWf). 2012. Commune agroecosystem analysis in Cambodia: A guidance manual. References Phnom Penh, Cambodia: Ministry of Agriculture, forestry and ﬁsheries, Department of Agricultural Extension (DAE); Colombo, Sri Lanka: CGIAr Challenge program on Water and food (CpWf). 118p. partment of Agricultural Extension (DAE), Ministry of Agriculture, forestry and ﬁsheries (MAff), Cambodia and CGIAr Challenge program on Water and food (CpWf). 2012. Commune agroecosystem analysis in Cambodia: A guidance manual. Phnom Penh, Cambodia: Ministry of Agriculture forestry and ﬁsheries Department of Agricultural Extension (DAE); Colombo Sri Lanka: CGIAr Department of Agricultural Extension (DAE), Ministry of Agriculture, forestry and ﬁsheries (MAff), Cambodia and CGIAr Challenge program on Water and food (CpWf). 2012. Commune agroecosystem analysis in Cambodia: A guidance manual. Phnom Penh, Cambodia: Ministry of Agriculture, forestry and ﬁsheries, Department of Agricultural Extension (DAE); Colombo, Sri Lanka: CGI program on Water and food (CpWf). 2012. Commune agroecosystem analysis in Cambodia: A guidance manual. Phnom Penh, Cambodia: Ministry of Agriculture, forestry and ﬁsheries, Department of Agricultural Extension (DAE); Colombo, Sri Lanka: CGIAr Challenge program on Water and food (CpWf). 118p. Cambodia: Ministry of Agriculture, forestry and ﬁsheries, Department of Agricultural Extension (DAE); Colombo, Sri Lanka: CGIAr Challenge program on Water and food (CpWf). 118p. g p g ( p ) p Diener, E., & Biswas‐Diener, R. (2005). Psychological empowerment and subjective well‐being. In D. Narayan (Ed.), Measuring empow- erment: Cross‐disciplinary perspectives, (pp. 125–140). Washington DC: World Bank. p p p pp g Dubois, M., Hadjmichael, M., & Raakjaer, J. (2016). The rise of the scientiﬁc ﬁsherman: Mobilising knowledge and negotiating user rig in the Devon inshore brown crab ﬁshery, UK. Marine Policy, 65, 48–55. Foran, T. (2015). Node and regime: Interdisciplinary analysis of water‐energy‐food nexus in the Mekong region. Water Alternatives, 8(1), 655–674. Fox, C., & Sneddon, C. (2019). Political borders, epistemological boundaries, and contested knowledges: Constructing dams and narrat in the Mekong river basin. Water, 11(3), 413. https://www.mdpi.com/2073‐4441/11/3/413 Fung, A. (2006). Varieties of participation in complex governance. Public Administration Review, 66(1), 66–75. https://doi.org/10.1111 j.1540‐6210.2006.00667.x Gallagher, L., Dalton, J., Bréthaut, C., Allan, T., Bellﬁeld, H., Crilly, D., et al. (2016). The critical role of risk in setting directions for water, food and energy policy and research. Current Opinion in Environmental Sustainability, 23, 12–16. https://doi.org/10.1016/j. cosust.2016.10.002 Gerring, J. (2007). Case study research: Principles and practices. New York: Cambridge University Press. Retrieved from. http://www.loc. gov/catdir/enhancements/fy0703/2006051819‐d.html Gerritsen, A. L., Stuiver, M., & Termeer, C. J. A. M. (2013). References Abson, D. J., Fischer, J., Leventon, J., Newig, J., Schomerus, T., Vilsmaier, U., et al. (2017). Leverage points for sustainability transforma- tion. Ambio, 46(1), 30–39. https://doi.org/10.1007/s13280‐016‐0800‐y Addor, N., Ewen, T., Johnson, L., Çöltekin, A., Derungs, C., & Muccione, V. (2015). From products to processes: Academic events to fo interdisciplinary and iterative dialogue in a changing climate. Earth's Future, 3(8), 289–297. https://doi.org/10.1002/2015EF000303 Albrecht, T. R., Crootof, A., & Scott, C. A. (2018). The water‐energy‐food nexus: A systematic review of methods for nexus assessment. Environmental Research Letters, 13(4), 043002. Allan, T., Keulertz, M., & Woertz, E. (2015). The water‐food‐energy nexus: An introduction to nexus concepts and some conceptual a operational problems International Journal of Water Resources Development 31(3) 301311 https://doi org/10 1080/ Allan, T., Keulertz, M., & Woertz, E. (2015). The water‐food‐energy nexus: An introduction to nexus concepts and some conceptual and operational problems. International Journal of Water Resources Development, 31(3), 301–311. https://doi.org/10.1080/ 07900627.2015.1029118 operational problems. International Journal of Water Resources Development, 31(3), 301–311. https://doi.org/10.1080/ 07900627.2015.1029118 Allouche, J., Middleton, C., & Gyawali, D. (2015). Technical veil, hidden politics: Interrogating the power linkages behind the nexus. Water Alternatives, 8(1). Allouche, J., Middleton, C., & Gyawali, D. (2015). Technical veil, hidden politics: Interrogating the power linkages behind the nexus. Water Alternatives, 8(1). Al‐Saidi, M., & Elagib, N. A. (2017). Towards understanding the integrative approach of the water, energy and food nexus. Science of Total Environment, 574, 1131–1139. https://doi.org/10.1016/j.scitotenv.2016.09.046 Basco‐Carrera, L., Warren, A., van Beek, E., Jonoski, A., & Giardino, A. (2017). Collaborative modeling or participatory modeling? A fra- mework for water resources management. Environmental Modeling & Software, 91(Supplement C), 95–110. https://doi.org/10.1016/j. envsoft.2017.01.014 Boyd, E., Nykvist, B., Borgström, S., & Stacewicz, I. A. (2015). Anticipatory governance for social‐ecological resilience. Ambio, 44(1), 149–161. https://doi.org/10.1007/s13280‐014‐0604‐x Bréthaut, C., Gallagher, L., Dalton, J., & Allouche, J. (2019). Power dynamics and integration in the water‐energy‐food nexus: Learning lessons for transdisciplinary research in Cambodia. Environmental Science & Policy, 94, 153–162. https://doi.org/10.1016/j. envsci.2019.01.010 KIMMICH ET AL. 13 KIMMICH ET AL. Earth's Future Earth's Future 10.1029/2019EF001311 Cambodia Development Resource Institute. (2004). The challenges of decentralisation design in Cambodia. October 2004. Phnom Penh. Cambodia Development Resource Institute. (2004). The challenges of decentralisation design in Cambodia. October 2004. Phnom Penh. rr, G., Blöschl, G., & Loucks, D. (2012). Evaluating participation in water resource management: A review. Water Resources Research, 48, W11401. https://doi.org/10.1029/2011WR011662 g Cash, D. W., Clark, W. C., Alcock, F., Dickson, N. References Knowledge governance: An exploration of principles, impact, and barrier Science and Public Policy, 40(5), 604–615. https://doi.org/10.1093/scipol/sct012 Giné, X., Townsend, R., & Vickery, J. (2009). Forecasting when it matters: Evidence from semi‐arid India. Manuscript, FRB of NY. Gittelsohn, J., Rowan, M., & Gadhoke, P. (2012). Interventions in small food stores to change the food environment, improve diet, and reduce risk of chronic disease. Preventing Chronic Disease, 9, E59. Glynn, P. D., Voinov, A. A., Shapiro, C. D., & White, P. A. (2017). From data to decisions: Processing information, biases, and beliefs improved management of natural resources and environments. Earth's Future, 5(4), 356–378. Grafton, R. Q., McLindin, M., Hussey, K., Wyrwoll, P., Wichelns, D., Ringler, C., et al. (2016). Responding to global challenges in foo energy, environment and water: Risks and options assessment for decision‐making. Asia & the Paciﬁc Policy Studies, 3(2), 275–299 energy, environment and water: Risks and options assessment for decision‐making. Asia & the Paciﬁc Policy Studies, 3(2 Guston, D. H. (2014). Understanding “anticipatory governance”. Social Studies of Science, 44(2), 218–242. gy p g ﬁ Guston, D. H. (2014). Understanding “anticipatory governance”. Social Studies of Science, 44(2), 218–242. Hagemann, N., & Kirschke, S. (2017). Key issues of interdisciplinary NEXUS governance analyses: Lessons learned from research on integrated water resources management. Resources, 6(1). https://doi.org/10.3390/resources6010009 Halbe, J., Pahl‐Wostl, C., Lange, M. A., & Velonis, C. (2015). Governance of transitions towards sustainable development—The wate energy–food nexus in Cyprus. Water International, 40(5–6), 877–894. https://doi.org/10.1080/02508060.2015.1070328 Harris, T., & Hardin, J. W. (2013). Exact Wilcoxon signed‐rank and Wilcoxon Mann‐Whitney ranksum tests. Stata Journal, 13(2), 337–343. Harrison, G. W., & List, J. A. (2004). Field experiments. Journal of Economic Literature, 42(4), 1009–1055. https://doi.org/10.1257/ 0022051043004577 Harris, T., & Hardin, J. W. (2013). Exact Wilcoxon signed rank and Wilcoxon Mann Whitney ranksum tests. Stata Journal, 13(2), 337 343. Harrison, G. W., & List, J. A. (2004). Field experiments. Journal of Economic Literature, 42(4), 1009–1055. https://doi.org/10.1257/ 0022051043004577 Hedelin, B., Evers, M., Alkan‐Olsson, J., & Jonsson, A. (2017). Participatory modeling for sustainable development: Key issues derived from ﬁve cases of natural resource and disaster risk management. Environmental Science & Policy, 76(Supplement C), 185–196. https://doi. org/10.1016/j.envsci.2017.07.001 Heijden, J., & Heuvelhof, E. (2012). The mechanics of virtue: Lessons on public participation from implementing the Water Fram Directive in the Netherlands. Environmental Policy and Governance, 22(3), 177–188. https://doi.org/10.1002/eet.1583 Heijden, J., & Heuvelhof, E. (2012). References Retrieved from: http://www.mrcmekong.org/news‐and‐events/events/the‐ power projects. Mekong River Commission, Vientiane. Retrieved from: http://www.mrcmekong.org/news‐and‐events/events/the‐ regional‐stakeholder‐forum‐on‐the‐council‐study‐updated‐pdg‐and‐sustainable‐hydropower‐development‐strategy/ National Committee for Sub National Democratic Development. (2010). National program for sub national democratic development (NP SNDD) 2010–2019. Retrieved from http://ncdd.gov.kh/wp‐content/uploads/2017/10/NP‐SNDD‐Final‐28‐05‐2010‐_En_.pdf National Committee for Sub‐National Democratic Development. (2015). Commune council database. Retrieved from http://db.ncdd.gov kh/ccd/home/index.castle p ( ) p g p SNDD) 2010–2019. Retrieved from http://ncdd.gov.kh/wp‐content/uploads/2017/10/NP‐SNDD‐Final‐28‐05‐2010‐_En_.pdf SNDD) 2010–2019. Retrieved from http://ncdd.gov.kh/wp‐content/uploads/2017/10/NP‐SNDD‐Final‐28‐05‐2010‐_En_.pdf N ti l C itt f S b N ti l D ti D l t (2015) C il d t b R t i d f htt //db dd SNDD) 2010–2019. Retrieved from http://ncdd.gov.kh/wp‐content/uploads/2017/10/NP‐SNDD‐Final‐28‐05‐2010‐_En_.pdf National Committee for Sub‐National Democratic Development. (2015). Commune council database. Retrieved from http://db.ncdd.gov SNDD) 2010–2019. Retrieved from http://ncdd.gov.kh/wp‐content/uploads/2017/10/NP‐SNDD‐Final‐28‐05‐2010‐_En_.pdf National Committee for SubNational Democratic Development (2015) Commune council database Retrieved from http://db ncdd gov National Committee for Sub‐National Democratic Development. (2015). Commune council database. Retrieved from http://db.ncdd. kh/ccd/home/index.castle ational Committee for Sub‐National Democratic Development. (2017). 2017 NCDD annual workplan and budget. Retrieved from http:// ncdd.gov.kh/wp‐content/uploads/2017/10/1‐2017‐AWPB‐2017.pdf Newson, R. (2002). Parameters behind “nonparametric” statistics: Kendall's tau, Somers' D and median differences. Stata Journal, 2( 45–64. Öjendal, J., & Lilja, M. (Eds) (2009). Beyond democracy in Cambodia: Political reconstruction in a post‐conﬂict society. Democracy in A Series 12. Copenhagen: NIAS. Öjendal, J., & Sedara, K. (2006). Korob, Kaud, Klach: In search of agency in rural Cambodia. Journal of Southeast Asian Studies, 37(3), 507–526. https://doi.org/10.1017/S0022463406000762 Orr, S., Pittock, J., Chapagain, A., & Dumaresq, D. (2012). Dams on the Mekong River: Lost ﬁsh protein and the implications for land and water resources. Global Environmental Change, 22(4), 925–932. https://doi.org/10.1016/j.gloenvcha.2012.06.002 Pahl‐Wostl, C., Giupponi, C., Richards, K., Binder, C., de Sherbinin, A., Sprinz, D., et al. (2013). Transition towards a new global change science: Requirements for methodologies, methods, data and knowledge. Environmental Science & Policy, 28, 36–47. https://doi.org/ 10.1016/j.envsci.2012.11.009 j Phillips, L. D., & Phillips, M. C. (1993). Faciliated work groups: Theory and practice. Journal of the Operational Research Society, 44(6 533–549. Pittock, J., Dumaresq, D., & Bassi, A. M. (2016). Modeling the hydropower–food nexus in large river basins: A Mekong case study. Water, 8(10), 425. Plummer, J., & Tritt, G. (2012). Voice, choice and decision: A study of local governance processes in Cambodia. Washington, DC: World Ba Plummer, J., & Tritt, G. (2012). Voice, choice and decision: A study of local governance processes in Cambodia. Washington, DC: World Bank. Pohl, C., Rist, S., Zimmermann, A., Fry, P., Gurung, G. References Public–private partnerships (PPP), build‐operate‐transf (BOT) and large hydropower dams in the Mekong region In N Matthews & K Geheb (Eds ) Hydropower development in the Mek , , , , , ( ) y p p p ( ), p (BOT) and large hydropower dams in the Mekong region. In N. Matthews, & K. Geheb (Eds.), Hydropower development in the Mekong region: Political, socio‐economic and environmental perspectives, (pp. 224–237). London: Earthscan. (BOT) and large hydropower dams in the Mekong region. In N. Matthews, & K. Geheb (Eds.), Hydropower development in the Me region: Political, socio‐economic and environmental perspectives, (pp. 224–237). London: Earthscan. ller, C. A., & Wyborn, C. (2018). Co‐production in global sustainability: Histories and theories. Environmental Science & Policy. https:// doi.org/10.1016/j.envsci.2018.01.016 lne, S., & Mahanty, S. (Eds) (2015). Conservation and development in Cambodia: Exploring frontiers of change in nature, state and society. London and New York: Routledge. g Mochizuki, J., Vitoontus, S., Wickramarachchi, B., Hochrainer‐Stigler, S., Williges, K., Mechler, R., & Sovann, R. (2015). Operationalizing iterative risk management under limited information: Fiscal and economic risks due to natural disasters in Cambodia. International Mochizuki, J., Vitoontus, S., Wickramarachchi, B., Hochrainer‐Stigler, S., Williges, K., Mechler, R., & Sovann, R. (2015). Operationalizing iterative risk management under limited information: Fiscal and economic risks due to natural disasters in Cambodia. International Journal of Disaster Risk Science, 6(4), 321–334. https://doi.org/10.1007/s13753‐015‐0069‐y Mochizuki, J., Vitoontus, S., Wickramarachchi, B., Hochrainer‐Stigler, S., Williges, K., Mechler iterative risk management under limited information: Fiscal and economic risks due to natu Journal of Disaster Risk Science, 6(4), 321–334. https://doi.org/10.1007/s13753‐015‐0069‐y iterative risk management under limited information: Fiscal and economic risks due to natur Journal of Disaster Risk Science, 6(4), 321–334. https://doi.org/10.1007/s13753‐015‐0069‐y Journal of Disaster Risk Science, 6(4), 321–334. https://doi.org/10.1007/s13753‐015‐0069‐y Molle, F. (2009). River‐Basin Planning and Management: The Social Life of a Concept. Geoforum, 40(3), 484–94. https://doi.org/10.101 geoforum.2009.03.004 Molle, F., Foran, T., & Käkönen, M. (Eds) (2012). Contested waterscapes in the Mekong region: Hydropower, livelihoods and governanc London: Earthscan. MRC (2017). The study on sustainable development and management of the Mekong river, including the impacts of mainstream hydr power projects. Mekong River Commission, Vientiane. Retrieved from: http://www.mrcmekong.org/news‐and‐events/events/the‐ regional‐stakeholder‐forum‐on‐the‐council‐study‐updated‐pdg‐and‐sustainable‐hydropower‐development‐strategy/ MRC (2017). The study on sustainable development and management of the Mekong river, including the impacts of mainstream h power projects. Mekong River Commission, Vientiane. Retrieved from: http://www.mrcmekong.org/news‐and‐events/events/th regional‐stakeholder‐forum‐on‐the‐council‐study‐updated‐pdg‐and‐sustainable‐hydropower‐development‐strategy/ ( ) y p g g , g p y power projects. Mekong River Commission, Vientiane. References Environmental Science & Policy, 90, 164– https://doi.org/10.1016/j.envsci.2017.08.015 Lebel, L., & Lebel, B. (2018). Nexus narratives and resource insecurities in the Mekong region. Environmental Science & Policy, 90, 164– https://doi.org/10.1016/j.envsci.2017.08.015 Liu, J., Hull, V., Godfray, H. C. J., Tilman, D., Gleick, P., Hoff, H., et al. (2018). Nexus approaches to global sustainable development. Nature Sustainability, 1(9), 466–476. https://doi.org/10.1038/s41893‐018‐0135‐8 Liu, J., Hull, V., Godfray, H. C. J., Tilman, D., Gleick, P., Hoff, H., et al. (2018). Nexus approaches to global sustainable development. Nature Sustainability, 1(9), 466–476. https://doi.org/10.1038/s41893‐018‐0135‐8 y ( ) p g Manski, C. F. (2004). Measuring expectations. Econometrica, 72(5), 1329–1376. https://doi.org/10.1111/j.1468‐0262.2004.00537.x McKenzie, D., Gibson, J., & Stillman, S. (2013). A land of milk and honey with streets paved with gold: Do emigrants have over‐optim expectations about incomes abroad? Migration and Development 102 116127 https://doi org/10 1016/j jdeveco 2012 01 001 McKenzie, D., Gibson, J., & Stillman, S. (2013). A land of milk and honey with streets paved with gold: Do emigrants have over‐optimistic expectations about incomes abroad? Migration and Development, 102, 116–127. https://doi.org/10.1016/j.jdeveco.2012.01.001 ekong River Commission (2017). The study on sustainable development and management of the Mekong River, including the impacts of mainstream hydropower projects Vientiane: Mekong River Commission Retrieved from: http://www mrcmekong org/news and Mekong River Commission (2017). The study on sustainable development and management of the Mekong River, including the impacts of mainstream hydropower projects. Vientiane: Mekong River Commission. Retrieved from:. http://www.mrcmekong.org/news‐and‐ events/events/the‐regional‐stakeholder‐forum‐on‐the‐council‐study‐updated‐pdg‐and‐sustainable‐hydropower‐development‐strategy/ Mekong River Commission (2017). The study on sustainable development and management of the Mekong River, including the impacts of mainstream hydropower projects. Vientiane: Mekong River Commission. Retrieved from:. http://www.mrcmekong.org/news‐and‐ e ents/e ents/the regional stakeholder for m on the co ncil st d pdated pdg and s stainable h dropo er de elopment strateg / g y p g f g g p f mainstream hydropower projects. Vientiane: Mekong River Commission. Retrieved from:. http://www.mrcmekong.org/news‐and‐ Middleton, C., Matthews, N., & Mirumachi, N. (2015). Whose risky business? Public–private partnerships (PPP), build‐operate‐transfer (BOT) and large hydropower dams in the Mekong region. In N. Matthews, & K. Geheb (Eds.), Hydropower development in the Mekong region: Political, socio‐economic and environmental perspectives, (pp. 224–237). London: Earthscan. Middleton, C., Matthews, N., & Mirumachi, N. (2015). Whose risky business? References Earth's Future, 6(8), 1046–1057. https://doi.org/10.1029/2018EF000841 f g g p p y g g Jordan, R., Gray, S., Zellner, M., Glynn, P. D., Voinov, A., Hedelin, B., et al., & National Socio‐Environmental Synthesis Center Participatory Modeling Pursuit Working Group (2018). Twelve questions for the participatory modeling community. Earth's Future, 6(8), 1046–1057. https://doi.org/10.1029/2018EF000841 Kaddoura, S., & El Khatib, S. (2017). Review of water‐energy‐food nexus tools to improve the nexus modeling approach for integrated policy making. Environmental Science & Policy, 77, 114–121. https://doi.org/10.1016/j.envsci.2017.07.007 Kaddoura, S., & El Khatib, S. (2017). Review of water‐energy‐food nexus tools to improve the nexus modeling approach for integrated policy making. Environmental Science & Policy, 77, 114–121. https://doi.org/10.1016/j.envsci.2017.07.007 g g j Kimmich, C., & Fischbacher, U. (2016). Behavioral determinants of supply chain integration and coexistence. Journal of Forest Econom 25, 55–77. https://doi.org/10.1016/j.jfe.2016.08.001 p g j j Kopainsky, B., Hager, G., Herrera, H., & Nyanga, P. H. (2017). Transforming food systems at local levels: Using participatory system dynamics in an interactive manner to reﬁne small‐scale farmers' mental models. Ecological Modeling, 362, 101–110. https://doi.org/ 10.1016/j.ecolmodel.2017.08.010 p g j j Kopainsky, B., Hager, G., Herrera, H., & Nyanga, P. H. (2017). Transforming food systems at local levels: Using participatory system dynamics in an interactive manner to reﬁne small‐scale farmers' mental models. Ecological Modeling, 362, 101–110. https://doi.org/ 10.1016/j.ecolmodel.2017.08.010 Kopainsky, B., Hager, G., Herrera, H., & Nyanga, P. H. (2017). Transforming food systems at local levels: Using participatory system dynamics in an interactive manner to reﬁne small‐scale farmers' mental models. Ecological Modeling, 362, 101–110. https://doi.org/ 10.1016/j.ecolmodel.2017.08.010 j Kotir, J. H., Brown, G., Marshall, N., & Johnstone, R. (2017). Systemic feedback modeling for sustainable water resources management and agricultural development: An application of participatory modeling approach in the Volta River Basin. Environmental Modeling & Software, 88(Supplement C), 106–118. https://doi.org/10.1016/j.envsoft.2016.11.015 j Kotir, J. H., Brown, G., Marshall, N., & Johnstone, R. (2017). Systemic feedback modeling for sustainable water resources management and agricultural development: An application of participatory modeling approach in the Volta River Basin. Environmental Modeling & Software, 88(Supplement C), 106–118. https://doi.org/10.1016/j.envsoft.2016.11.015 f pp p g j Kratie Provincial Department of Planning. (2014). Commune investment plan 2014. al Department of Planning. (2014). Commune investment plan 2014 ratie Provincial Department of Planning. (2014). Commune investm Kratie Provincial Department of Planning. (2014). Commune investment plan 2014. Lebel, L., & Lebel, B. (2018). Nexus narratives and resource insecurities in the Mekong region. References The mechanics of virtue: Lessons on public participation from implementing the Water Framework Directive in the Netherlands. Environmental Policy and Governance, 22(3), 177–188. https://doi.org/10.1002/eet.1583 Hill, R. V. (2010). Liberalisation and producer price risk: Examining subjective expectations in the Ugandan coffee market. Journal African Economies, 19(4), 433–458. f Hovmand, P. S. (2014). Community based system dynamics. New York, Heidelberg, Dordrecht, London: Springer. Hovmand, P. S., Andersen, D. F., Rouwette, E., Richardson, G. P., Rux, K., & Calhoun, A. (2012). Group model‐bui ll b ti l i t l S t R h d B h i l S i 29(2) 179 193 htt //d i /10 1002/ Hovmand, P. S., Andersen, D. F., Rouwette, E., Richardson, G. P., Rux, K., & Calhoun, A. (2012). Group model‐bu collaborative planning tool. Systems Research and Behavioral Science, 29(2), 179–193. https://doi.org/10.1002/sr Hovmand, P. S., Andersen, D. F., Rouwette, E., Richardson, G. P., Rux, K., & Calhoun, A. (2012). Group model‐building ‘scrip , , , , , , , , , , , ( ) p g p collaborative planning tool. Systems Research and Behavioral Science, 29(2), 179–193. https://doi.org/10.1002/sres.2105 Howarth, C., & Monasterolo, I. (2017). Opportunities for knowledge co‐production across the energy‐food‐water nexus: Making interdis ciplinary approaches work for better climate decision making. Environmental Science & Policy, 75, 103–110. https://doi.org/10.1016/j envsci.2017.05.019 Hurlbert, M., & Gupta, J. (2015). The split ladder of participation: A diagnostic, strategic, and evaluation tool to assess when participation is necessary. Environmental Science & Policy, 50, 100–113. Irvin, R. A., & Stansbury, J. (2004). Citizen participation in decision making: Is it worth the effort? Public Administration Review, 64(1), 55–65. https://doi.org/10.1111/j.1540‐6210.2004.00346.x Irvin, R. A., & Stansbury, J. (2004). Citizen participation in decision making: Is it worth the effort? Public Administration Review, 64(1), 55–65. https://doi.org/10.1111/j.1540‐6210.2004.00346.x Jensen, R. (2010). The (perceived) returns to education and the demand for schooling. The Quarterly Journal of Economics, 125(2), 515–548. p g j Jensen, R. (2010). The (perceived) returns to education and the demand for schooling. The Quarterly Journal of Economics, 125(2), 515– KIMMICH ET AL. 14 Earth's Future 10.1029/2019EF001311 John, P. (2018). How far to nudge?: Assessing behavioral public policy. Cheltenham: Edward Elgar Publishing. , ( ) f g g p p y g g Jordan, R., Gray, S., Zellner, M., Glynn, P. D., Voinov, A., Hedelin, B., et al., & National Socio‐Environmental Synthesis Center Participatory Modeling Pursuit Working Group (2018). Twelve questions for the participatory modeling community. References Moving from theory to practice in the water–energy–food nexus: an evaluation of existing models and frameworks. Water‐Energy Nexus, 1(1), 17–25. https://doi.org/10.1016/j.wen.2018.04.001 Siciliano, G., Urban, F., Kim, S., & Lonn, P. D. (2015). Hydropower, social priorities and the rural–urban development divide: The cas large dams in Cambodia. Energy Policy, 86, 273–285. Sithirith, M. (2016). Dams and state security: Damming the 3S rivers as a threat to Cambodian state security. Asia Paciﬁc Viewpoint, 57 60–75. Smajgl, A., & Ward, J. (2013). The water‐food‐energy nexus in the Mekong region: Assessing development strategies considering cross‐sect and transboundary impacts. New York: Springer. y p p g Smajgl, A., & Ward, J. R. (2015). Evaluating participatory research: Framework, methods and implementation results. Journal of Smajgl, A., & Ward, J. R. (2015). Evaluating participatory research: Framework, methods and Environmental Management, 157, 311–319. https://doi.org/10.1016/j.jenvman.2015.04.014 Smajgl, A., & Ward, J. R. (2015). Evaluating participatory research: Framework, methods and i Environmental Management, 157, 311–319. https://doi.org/10.1016/j.jenvman.2015.04.014 Smajgl, A., Ward, J. R., Foran, T., Dore, J., Ward, J., & Larson, S. (2015). Visions, beliefs, and transformation: Exploring cross‐sector and transboundary dynamics in the wider Mekong region. Ecology and Society, 20(2), 15. Retrieved from. https://www.ecologyandsociety. org/vol20/iss2/art15/ Smith, T. M. F. (1983). On the validity of inferences from non‐random sample. Journal of the Royal Statistical Society. Series A (Gener 146(4), 394. https://doi.org/10.2307/2981454 Star, S., & Griesemer, J. (1989). Ecology, ‘translations’ and boundary objects: Amateurs and professionals in Berkeley's Museum of Star, S., & Griesemer, J. (1989). Ecology, ‘translations’ and boundary objects: Amateurs and professionals in Berkeley's Museum of Vertebrate Zoology, 1907‐39. Social Studies of Science, 19(3), 387–420. https://doi.org/10.1177/030631289019003001 Un, K., & Sokbunthoeun, S. (2009). Politics of natural resource use in Cambodia. Asian Affairs: An American Perspective, 36, 123–13 Star, S., & Griesemer, J. (1989). Ecology, ‘translations’ and boundary objects: Amateurs and professionals in Berkeley's Muse Vertebrate Zoology, 1907‐39. Social Studies of Science, 19(3), 387–420. https://doi.org/10.1177/030631289019003001 U K & S kb h S ( ) P li i f l i C b di A i Aff i A A i P i gy y j p y Vertebrate Zoology, 1907‐39. Social Studies of Science, 19(3), 387–420. https://doi.org/10.1177/030631289019003001 van Voorn, G. A. K., Verburg, R. W., Kunseler, E.‐M., Vader, J., & Janssen, P. H. M. (2016). A checklist for model credibility, salience, legitimacy to improve information transfer in environmental policy assessments. Environmental Modeling & Software, van Voorn, G. A. K., Verburg, R. References N., Ramos‐Martin, J., Giampietro, M., & Mayumi, K. (2014). Land poverty a Insights from Kampot province. Environment, Development and Sustainability, 16(4), 823–840. Scheidel, A., Farrell, K. N., Ramos‐Martin, J., Giampietro, M., & Mayumi, K. (2014). Land poverty and emergin Insights from Kampot province. Environment, Development and Sustainability, 16(4), 823–840. Scheidel, A., Farrell, K. N., Ramos‐Martin, J., Giampietro, M., & Mayumi, K. (2014). Land pover Insights from Kampot province. Environment, Development and Sustainability, 16(4), 823–840 Scott, R. J., Cavana, R. Y., & Cameron, D. (2016a). Mechanisms for understanding mental model c Scott, R. J., Cavana, R. Y., & Cameron, D. (2016a). Mechanisms for understanding mental model change in group model bui R h d B h i l S i ( ) h //d i / / Scott, R. J., Cavana, R. Y., & Cameron, D. (2016a). Mechanisms for understanding mental model change in group model building. Systems Research and Behavioral Science, 33(1), 100–118. https://doi.org/10.1002/sres.2303 Scott, R. J., Cavana, R. Y., & Cameron, D. (2016a). Mechanisms for understanding mental model change in group model building. Systems Research and Behavioral Science, 33(1), 100–118. https://doi.org/10.1002/sres.2303 Research and Behavioral Science, 33(1), 100–118. https://doi.org/10.1002/sres.2303 h and Behavioral Science, 33(1), 100–118. https://doi.org/10.1002/sre , ( ), p // g/ / Scott, R. J., Cavana, R. Y., & Cameron, D. (2016b). Recent evidence on the effectivene Operational Research 249(3) 908 918 https://doi org/10 1016/j ejor 2015 06 078 g Scott, R. J., Cavana, R. Y., & Cameron, D. (2016b). Recent evidence on the effectiveness of group model building. European Journal o Operational Research, 249(3), 908–918. https://doi.org/10.1016/j.ejor.2015.06.078 , , , , , ( ) Operational Research, 249(3), 908–918. https://doi.org/10.1016/j.ejor.2015.06.078 Sedara, K. (2012). Democracy in Action: Decentralization in post‐conﬂict Cambodia. PhD dissertation in Peace and Development Research, School of Global Studies, University of Gothenburg. Sedara, K. (2012). Democracy in Action: Decentralization in post‐conﬂict Cambodia. PhD dissertation in Peace and Development Research School of Global St dies Uni ersit of Gothenb rg Sedara, K. (2012). Democracy in Action: Decentralization in post Sedara, K. (2012). Democracy in Action: Decentralization in post‐ Research, School of Global Studies, University of Gothenburg. Research, School of Global Studies, University of Gothenburg. Sendzimir, J., Magnuszewski, P., Balogh, P., & Vári, A. (2007). Anticipatory modeling of biocomplexity in the Tisza River Basin: First steps to establish a participatory adaptive framework. Environmental Modeling & Software, 22(5), 599–609. https://doi.org/10.1016/j. envsoft.2005.12.032 Shannak, S., Mabrey, D., & Vittorio, M. (2018). References S., Schneider, F., et al. (2010). Researchers' roles in knowledge co‐production: Experience from sustainability research in Kenya, Switzerland, Bolivia and Nepal. Science and Public Policy, 37(4), 267–281. https://doi. ( ) y f g p g Pohl, C., Rist, S., Zimmermann, A., Fry, P., Gurung, G. S., Schneider, F., et al. (2010). Researchers' roles in knowledge co‐production Experience from sustainability research in Kenya, Switzerland, Bolivia and Nepal. Science and Public Policy, 37(4), 267–281. https:// org/10.3152/030234210X496628 g Reed, M. S. (2008). Stakeholder participation for environmental management: A literature review. Biological Conservation, 141(10), 2417–2431. https://doi.org/10.1016/j.biocon.2008.07.014 Rodrik, D. (2014). When ideas trump interests: Preferences, worldviews, and policy innovations. Journal of Economic Perspectives, 28 189–208. Rouwette, E. A. J. A. (2003). Group model building as mutual persuasion. Nijmegen: Wolf Legal Publishers. KIMMICH ET AL. 15 Earth's Future Earth's Future 10.1029/2019EF001311 Rouwette, E. A. J. A., Korzilius, H., Vennix, J. A. M., & Jacobs, E. (2011). Modeling as persuasion: The impact of group model building on attitudes and behavior. System Dynamics Review, 27(1), 1–21. https://doi.org/10.1002/sdr.441 Rouwette, E. A. J. A., Korzilius, H., Vennix, J. A. M., & Jacobs, E. (2011). Modeling as persuasion: The impact of group model building attitudes and behavior. System Dynamics Review, 27(1), 1–21. https://doi.org/10.1002/sdr.441 Rouwette, E. A. J. A., & Vennix, J. A. M. (2006). System dynamics and organizational interventions. Systems Research and Behavioral Science, 23(4), 451–466. Royal Decree No. NS/RKM/1208/1429 (2008). on Establishment of National Committee for Sub‐National Democratic Development da on December 31, 2008. Royal Decree no. NS/RKM/1014/1174 (2014). on Modiﬁcation and Supplement of New Members to National Committee for Sub‐Nat Democratic Development dated on October 6, 2014. Runge, C. F. (1986). Common property and collective action in economic development. World Development, 14(5), 623–635. Sanò, M., Richards, R., & Medina, R. (2014). A participatory approach for system conceptualization and analysis applied to coa agement in Egypt. Environmental Modeling & Software, 54, 142–152. https://doi.org/10.1016/j.envsoft.2013.12.009 g gyp g f , , p g j Scheidel, A., Farrell, K. N., Ramos‐Martin, J., Giampietro, M., & Mayumi, K. (2014). Land poverty and emerging ruralities in Cambodia: I i ht f K t i E i D l d S i bili 16(4) 823 840 Scheidel, A., Farrell, K. N., Ramos‐Martin, J., Giampietro, M., & Mayumi, K. (2014). Land poverty and emerging ruralities in Cambodia: Insights from Kampot province. Environment, Development and Sustainability, 16(4), 823–840. Scheidel, A., Farrell, K. References W., Kunseler, E.‐M., Vader, J., & Janssen, P. H. M. (2016). A checklist for model credibility, salience, and legitimacy to improve information transfer in environmental policy assessments. Environmental Modeling & Software, 83(Supplement C), 224–236. https://doi.org/10.1016/j.envsoft.2016.06.003 van Voorn, G. A. K., Verburg, R. W., Kunseler, E. M., Vader, J., & Janssen, P. H. M. (2016). A checklist for model credibility, salience, and legitimacy to improve information transfer in environmental policy assessments. Environmental Modeling & Software, 83(Supplement C) 224 236 https://doi org/10 1016/j envsoft 2016 06 003 pp p g j Villamayor‐Tomas, S., Avagyan, M., Firlus, M., Helbing, G., & Kabakova, M. (2016). Hydropower vs. ﬁsheries conservation: A test of institutional design principles for common‐pool resource management in the lower Mekong basin social‐ecological system. Ecology and Society, 21(1), 3. Retrieved from:. https://www.ecologyandsociety.org/vol21/iss1/art3/ y p gy y g Villamayor‐Tomas, S., Grundmann, P., Epstein, G., Evans, T., & Kimmich, C. (2015). The water‐energy‐food security nexus through the lenses of the value chain and the Institutional Analysis and Development frameworks. Water Alternatives, 8(1), 735–755. Voinov, A., Kolagani, N., McCall, M. K., Glynn, P. D., Kragt, M. E., Ostermann, F. O., et al. (2016). Modeling with stakeholders—Next generation. Environmental Modeling & Software, 77(Supplement C), 196–220. https://doi.org/10.1016/j.envsoft.2015.11.016 Vugteveen, P., Rouwette, E., Stouten, H., van Katwijk, M. M., & Hanssen, L. (2015). Developing social‐ecological system indicators using group model building. Ocean and Coastal Management, 109(Supplement C), 29–39. https://doi.org/10.1016/j.ocecoaman.2015.02.011 Vuković, D., & Babović, M. (2018). The trap of neo‐patrimonialism: Social accountability and good governance in Cambodia. Asian Studies Review, 1–17. Weitz, N., Strambo, C., Kemp‐Benedict, E., & Nilsson, M. (2017). Closing the governance gaps in the water‐energy‐food nexus: Insights from integrative governance. Global Environmental Change, 45, 165–173. https://doi.org/10.1016/j.gloenvcha.2017.06.006 Weitz, N., Strambo, C., Kemp‐Benedict, E., & Nilsson, M. (2017). Closing the governance gaps in the water‐energy‐food nexus: Insights from integrative governance. Global Environmental Change, 45, 165–173. https://doi.org/10.1016/j.gloenvcha.2017.06.006 Wesselink, A., Buchanan, K. S., Georgiadou, Y., & Turnhout, E. (2013). Technical knowledge, discursive spaces and politics at the science– policy interface SI: Environmental and Developmental Discourses: Technical Knowledge Discursive Spaces and Politics 30 19 https:// Wesselink, A., Buchanan, K. S., Georgiadou, Y., & Turnhout, E. (2013). Technical knowledge, discursive spaces and politics at the science– policy interface. SI: Environmental and Developmental Discourses: Technical Knowledge, Discursive Spaces and Politics, 30, 1–9. https:// doi.org/10.1016/j.envsci.2012.12.008 Wild, T. B., Reed, P. M., Loucks, D. P., Mallen‐Cooper, M., & Jensen, E. D. (2019). Wild, T. B., Reed, P. M., Loucks, D. P., Mallen‐Cooper, M., & Jensen, E. D. (2019). Balancing hydropower development and ecological impacts in the Mekong: Tradeoffs for Sambor mega dam. Journal of Water Resources Planning and Management, 145(2), 05018019. Wu, S., Ishidaira, H., & Sun, W. (2010). Potential impact of Sambor dam project on interaction between Mekong River and Tonle Sap Lake. Annual Journal of Hydraulic Engineering, 54, 109–114. Retrieved from:. http://library.jsce.or.jp/jsce/open/00028/2010/54‐0019.pdf Yung, L., Louder, E., Gallagher, L., Jones, K., & Wyborn, C. (2019). How methods for navigating uncertainty connect science and policy at the water‐energy‐food nexus. Frontiers in Environmental Science: Freshwater Science, 7(37). https://doi.org/10.3389/fenvs.2019.00037 References Balancing hydropower development and ecological impacts in the Mekong: Tradeoffs for Sambor mega dam. Journal of Water Resources Planning and Management, 145(2), 05018019. Wu, S., Ishidaira, H., & Sun, W. (2010). Potential impact of Sambor dam project on interaction between Mekong River and Tonle Sap Lake. Annual Journal of Hydraulic Engineering, 54, 109–114. Retrieved from:. http://library.jsce.or.jp/jsce/open/00028/2010/54‐0019.pdf Yung, L., Louder, E., Gallagher, L., Jones, K., & Wyborn, C. (2019). How methods for navigating uncertainty connect science and policy at Wu, S., Ishidaira, H., & Sun, W. (2010). Potential impact of Sambor dam project on interaction between Mekong River and Tonle Sap Lake. Annual Journal of Hydraulic Engineering, 54, 109–114. Retrieved from:. http://library.jsce.or.jp/jsce/open/00028/2010/54‐0019.pdf Yung, L., Louder, E., Gallagher, L., Jones, K., & Wyborn, C. (2019). How methods for navigating uncertainty connect science and policy at the water‐energy‐food nexus. Frontiers in Environmental Science: Freshwater Science, 7(37). https://doi.org/10.3389/fenvs.2019.00037 KIMMICH ET AL. KIMMICH ET AL. 16
https://openalex.org/W3104645803	https://ieeexplore.ieee.org/ielx7/8784343/8889549/09220164.pdf	English	null	Wildfire Risk Mitigation: A Paradigm Shift in Power Systems Planning and Operation	IEEE open access journal of power and energy	2,020	cc-by	8,545	JOHN W. MUHS1 (Member, IEEE), MASOOD PARVANIA 1 (Senior Member, IEEE), AND MOHAMMAD SHAHIDEHPOUR 2 (Fellow, IEEE) (Invited Paper) p 1Department of Electrical and Computer Engineering, The University of Utah, Salt Lake City, UT 84112 USA 2Department of Electrical and Computer Engineering, Illinois Institute of Technology, Chicago, IL 60616 USA CORRESPONDING AUTHOR: M. PARVANIA (masood.parvania@utah.edu) ABSTRACT Managing the risk of wildﬁres has been arguably the biggest recent challenge of electric util- ities with infrastructure located in the wildland-urban interface. Utilities are deploying solutions for wildﬁre risk mitigation, such as public safety power shutoffs, which are counter-intuitive from a reliability-centric operation paradigm. This article presents an overview of the challenges, implications, and potential strategies for wildﬁre risk mitigation in power systems, and introduces the vision for a wildﬁre-resilient power system. The wildﬁre risk management strategies presented in this article range from fault prevention methods such as structural hardening, vegetation management and implementing advanced protection systems, to arc-suppression and ignition prevention methods. This article also identiﬁes relevant research opportunities associated with implementing wildﬁre mitigation techniques on power systems. INDEX TERMS Wildﬁre risk mitigation, power grid resilience, proactive operation, wildﬁre-resilient power grid. Power system resilience has been broadly deﬁned as ‘‘the ability of the power grid to prepare for and adapt to changing conditions and withstand and recover rapidly from disruptions’’ [4]. Resilience differs from reliability in that it focuses on high impact low probability events rather than mitigating disruptions to electricity service dur- ing normal operating conditions. Notable resilience studies have encompassed impacts to the power grid due to severe weather [5], [6], cyber/physical attacks [7], [8], and cascading blackouts [9], [10]. This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/ ensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/ Wildfire Risk Mitigation: A Paradigm Shift in Power Systems Planning and Operation JOHN W. MUHS1 (Member, IEEE), MASOOD PARVANIA 1 (Senior Member, IEEE), AND MOHAMMAD SHAHIDEHPOUR 2 (Fellow, IEEE) (Invited Paper) 1Department of Electrical and Computer Engineering, The University of Utah, Salt Lake City, UT 84112 USA 2Department of Electrical and Computer Engineering, Illinois Institute of Technology, Chicago, IL 60616 USA CORRESPONDING AUTHOR: M. PARVANIA (masood.parvania@utah.edu) Received 19 January 2020; revised 20 July 2020 and 2 October 2020; accepted 3 October 2020. Date of publication 12 October 2020; date of current version 9 November 2020. Digital Object Identiﬁer 10.1109/OAJPE.2020.3030023 Received 19 January 2020; revised 20 July 2020 and 2 October 2020; accepted 3 October 2020. Date of publication 12 October 2020; date of current version 9 November 2020. Digital Object Identiﬁer 10.1109/OAJPE.2020.3030023 I. INTRODUCTION such as lightning or arson [30]. Power system infrastructure is often located among vulnerable communities who live in ﬁre prone areas in the wildland-urban interface. As communities in the wildland-urban interface rapidly expand and grow, it is expected that power system wildﬁre ignitions will cause higher property damage and economic losses [31], [32]. load curtailed, or number of power system components made unavailable over the course of a catastrophic event. Resilience metrics quantify different aspects of power grid performance during extreme events, including the rate at which the system reaches its most severely degraded state, the extent of impact caused by a catastrophic event, the duration of time at which the power grid remains at a degraded state, and the rate at which electricity service is restored, among other metrics. g p p y g The objective of this article is to provide a comprehensive review of planning and operation techniques that contribute to create a wildﬁre-resilient power system. In this article, a wildﬁre-resilient power system is visioned as one that not only possesses the qualities of a traditionally deﬁned resilient grid, but also that mitigates catastrophic wildﬁre ignitions through a combination of planning, operation, and response strategies. This article focuses on mitigating strategies in power distribution systems, as a majority of faults, failures, and ﬁres occur at the distribution level. This article identiﬁes the myriad ways that wildﬁres and their subsequent impact can be mitigated, which are categorized as: 1) methods that prevent faults, 2) methods that prevent sustained ignitions of surrounding combustible fuel beds, and 3) methods that mitigate the impact of wildﬁres if an ignition caused by the power system occurs. This article also identiﬁes a variety of planning and operation research opportunities in the area of wildﬁre mitigation that can be aligned with current utility grid modernization goals. Policy and research literature has identiﬁed wildﬁres as a major natural disaster threat to the power system, and has also identiﬁed important challenges in mitigating the impacts caused by wildﬁres approaching power systems. The authors in [15] model the temperature increase of overhead conduc- tors as a result of an approaching wildﬁre by use of heat trans- fer equations. The impacts of approaching wildﬁres on power systems operation have been formulated as an optimization problem for transmission systems and distribution systems in [16] and [17], respectively. I. INTRODUCTION T HOUSANDS of miles of power lines have been developed in arid wildland-urban interface to provide electricity service to rural and remote communities. Power infrastructure in these locations is often aged and at risk of igniting catastrophic wildﬁres —endangering utility infras- tructure, the natural environment, and the lives and property of nearby communities. For decades, power systems have been designed and operated so as to limit the frequency and duration of disruptions to power service as assessed by metrics such as the System Average Interruption Dura- tion Index (SAIDI) and System Average Interruption Fre- quency Index (SAIFI). However, recent catastrophic events have challenged this reliability-centric paradigm and have exposed the need for new metrics and operating principles in the power industry. The concept of resilience has been developed and explored in power grids with the objective of quantifying and ultimately enhancing the ability of the grid to prevent and respond to low probability catastrophic events that may have a high or long-lasting impact on power grids performance [1]–[3]. T A plethora of research literature explores methods to quan- tify power system resilience through performance evaluation approaches. The authors in [11] propose a method to quan- tify the resilience of an infrastructure system by measuring its ability to: 1) resist or prevent hazards, 2) absorb initial damage, and 3) recover to normal operation. This process has been expended in [12]–[14], and a series of resilience metrics have been developed based on measuring the power system’s performance over time. The impacts of an extreme event to the power grid can be quantiﬁed by considering, for example, the number of customers without power, amount of 366 VOLUME 7, 2020 VOLUME 7, 2020 Muhs et al.: Wildfire Risk Mitigation: A Paradigm Shift in Power Systems Planning and Operation TABLE 1. Notable wildfires believed to have been ignited by power systems. TABLE 1. Notable wildfires believed to have been ignited by power systems. load curtailed, or number of power system components made unavailable over the course of a catastrophic event. Resilience metrics quantify different aspects of power grid performance during extreme events, including the rate at which the system reaches its most severely degraded state, the extent of impact caused by a catastrophic event, the duration of time at which the power grid remains at a degraded state, and the rate at which electricity service is restored, among other metrics. I. INTRODUCTION Additionally, the authors in [18] demonstrate that, through proactive operation and selective power shutoffs through microgrid operations, wildﬁre risk could be effectively mitigated as a wildﬁre approaches a power distribution system. Many of the studies in the literature consider wildﬁre risk to power systems to be a similar impact as other natural disasters —an externally caused, impending, and predictable phenomena from which power grid assets must be protected. However, recent wildﬁre events in Australia and California highlight the need for resilience-centric research literature to not only consider planning and operation strategies based on approaching wildﬁres, but also investigate mitigating catas- trophic wildﬁre ignitions caused by power systems. In the last two decades, a number of notable wildﬁres have been ignited by power systems as shown in Table 1. The remainder of this article is organized as follows: Section II provides a technical overview of wildﬁre ignition caused by power systems, and introduces a three pronged approach to wildﬁre mitigation covered in Sections III, IV, and V, which are described in detail subsequently. Section VI concludes the paper and provides an in-depth exploration for future research on wildﬁre-resilient power systems. II. WILDFIRE IGNITION BY DISTRIBUTION SYSTEMS While power line ignition events occur relatively infre- quently compared to other ignition sources (approximately 1% of wildﬁre ignitions in California are started by power systems), ﬁres ignited by power lines tend to burn signiﬁ- cantly more land area as compared to other ignition sources Although faults, failures, and ﬁres occur in nearly every subsystem within the power system, a vast majority of power system-related wildﬁres are caused by the distribution system. Distribution lines around the globe are scattered VOLUME 7, 2020 367 TABLE 2. Classification of wildfire mitigation techniques. TABLE 2. Classification of wildfire mitigation techniques. TABLE 2. Classification of wildfire mitigation techniques. throughout arid wildland terrain that is susceptible to ﬁre ignition [31]. In most known cases, wildﬁre ignition from power systems is initiated by an unintended fault or other form of catastrophic failure [33]. When power lines make contact with an external object, electricity ﬂowing through that line may ﬁnd an alternative path to ground causing a fault in the system. When this path to ground is realized, a high amount of current travels from the energized line to the ground through this path, inducing a high fault current in the line. Faulting causes an energy release along the path to ground, forming a high energy arc that can reach temperatures of up to 2240◦C [34]. System protection devices such as circuit breakers and fuses are deliberately coordinated to detect fault current, and de-energize the faulted power line as quickly as possible. Known wildﬁre ignition cases demonstrate a variety of ways whereby arcs can cause sustained ignition of a wildﬁre. Arcs can ignite surrounding fuels through direct contact, ignite a foreign object such as vegetation that makes con- tact with the energized line, and/or melt the metal conduc- tor (typically made of aluminium and steel) causing molten metal particles or ﬁrebrands to be ejected from the power line onto the ground. These phenomena, as they relate to power system caused ﬁre ignitions, have been studied in literature [34]–[36]. categorize the beneﬁt of each wildﬁre mitigation technique and comprise the three-pronged approach to wildﬁre mitiga- tion presented in this article, which are described in detail in Sections III, IV, and V, respectively. III. FAULT PREVENTION Wildﬁre mitigation begins with a precedent of high qual- ity and well maintained power system infrastructure. Over time, load bearing elements on distribution systems can fatigue due to prolonged wind loading causing wire-downs, loose equipment, or conductor failure [38]. Thorough inspec- tion, prognostics, and incremental replacements/upgrades of distribution system components, occurrence rates of catastrophic faults can be drastically reduced. Major fault prevention strategies include structural hardening, asset man- agement and inspection, and utilizing advanced protection systems and vegetation management, which are discussed next. 1) Fault Prevention methods such as structural hardening, asset management and inspection, advanced protection systems and vegetation management are utilized to reduce the likelihood of faults or failures in distribution systems. 2) Ignition Prevention methods such as sensitive protec- tion schemes, recloser disabling, and resonant ground- ing are implemented to reduce the likelihood of wildﬁre ignition in the case that a fault or failure does occur. II. WILDFIRE IGNITION BY DISTRIBUTION SYSTEMS Thus, the overall likelihood that a wildﬁre will ignite on a section of distribution feeder has to do both with: 1) the likelihood that the line will experience a fault, and 2) that the resulting fault will ignite surrounding fuel beds, each of which is an unlikely event that depends on a variety of both controllable and uncontrollable factors [36], [37]. As such, the wildﬁre mitigation techniques explored in this article are categorized as follows, based on their role in mitigating wildﬁre occurrences in power systems: B. ASSET INSPECTIONS AND TRACKING The improbability of these events highlights the need for advanced protection systems for detecting faults in distribu- tion systems. Advanced waveform analytics, such as ones presented in [45], [46], use advanced modeling techniques to detect abnormal behavior in distribution systems. More recent prognostics systems are able to not only detect but discriminate between fault causes, and instruct the system operator on corrective actions [47]. By use of advanced waveform analytics, asset owners and operators can identify small problems in distribution systems and repair them before they result in a catastrophic failure. Similar technologies can be used to recognize high impedance faults resulting from vegetation interference or a downed conductor. Many distribution systems in the United States were built over 30 years ago and continue to fatigue today. It is notable that the expected life of distribution infrastructure is approx- imately 40-50 years [42]. Thus, many of these assets have seen adverse conditions during their lifetime, which inﬂuence the rate of deterioration, and consequently the likelihood of a catastrophic fault. When distribution systems equipment fail, low cost solutions such as line splices are implemented to ﬁx the problem. By accumulating detailed knowledge of power distribution assets, utilities can optimize maintenance, and reduce catastrophic fault probability. Namely, mechani- cal elements such as clamps and line splices are commonly carefully tracked, and plans are developed in advance to replace worn, stressed, or fatiguing infrastructure. Although structural requirements for distribution poles and conductors are rated by the National Electrical Safety Code [43], and are designed to withstand storm level wind speeds, a vast majority of failures occur at wind speeds much lower than this critical wind speed [40]. This relationship implies that, while failure probability is inextricably related to wind speed, distribution system faults can occur independent of high wind speeds, and that deterioration of structural elements may contribute to failure rates of power lines even at low wind speeds. High impedance faults are similarly difﬁcult to detect. High impedance faults occur when a conductor comes into contact with an external object that does not conduct elec- tricity well (i.e., the path to ground has a high impedance) and may only partially or intermittently disrupt power ﬂow through the line. High impedance faults are troublesome because resulting fault current may not be sufﬁcient to trip the line via time overcurrent protection methods. A. STRUCTURAL HARDENING Structural hardening includes methods of design and engi- neering implemented to prevent faults in distribution sys- tems [2]. By assessing the structural needs of power lines in densely vegetated areas, utilities select upgraded mate- rials and pole designs that effectively harden their power lines and reduce fault probability. A study by electric power research institute found that replacing wooden poles with steel or ﬁberglass may not be the most cost effective method to strengthen power line structures [39]. Rather, utilities can seek wooden poles with higher top circumference for greater mechanical strength [39]. Distribution line failure is often considered as a ‘‘weakest link’’ failure, so upgrading several consecutive poles (especially adjacent to those that carry automated switches and reclosers) is a commonplace 3) Fire Response and Impact Mitigation efforts such as maintaining situational awareness, and establishing communication with infrastructure partners, ﬁre crews, and customers are employed in the unlikely case that a wildﬁre ignition occurs. Wildﬁre mitigation techniques can perform one or more of the aforementioned roles. For example, burying a distribution line underground not only largely eliminates interference with vegetation (reducing fault probability), but in the case that a fault does occur, it is typically not at risk of igniting surrounding fuel beds (reducing ignition probability). A clas- siﬁcation of most notable wildﬁre mitigation techniques is shown in Table 2. These three categories were selected to 368 VOLUME 7, 2020 VOLUME 7, 2020 Muhs et al.: Wildfire Risk Mitigation: A Paradigm Shift in Power Systems Planning and Operation hardening strategy [39], [40]. Additionally, spacer cables can help to reduce tree impact related outages [39]. represent a major ongoing cost to utility companies. In many cases, distribution feeders are inspected subannually, with detailed inspections occurring every 3 to 5 years. How- ever, in California, as regulated by the California Public Utilities Commission (CPUC), feeder patrols should occur annually in Extreme or Very High Threat Fire Districts (HTFDs) [44]. Interference with trees cannot be altogether avoided on overhead distribution lines. Thus, some utilities choose to cover conductors with insulation to protect the conductor from vegetation interference. Covering conductors, however, can be costly as conductor insulation can drastically increase the weight of the conductor, requiring smaller tower spans. Aerial Bundled Conductors (ABCs) have been implemented in Australia since the Ash Wednesday bushﬁres of 1983. However, these conductors are not commonly used in the United States [41]. C. ADVANCED WAVEFORM ANALYTICS AND PROTECTION SYSTEMS While visual inspections of distribution feeders may allow line patrols to identify visibly obvious issues (e.g., broken elements, vegetation interference), many issues may not be visible to the naked eye, or may not be occurring at the time of inspection. For example, arcs may be so small that they are not visible to patrol crews, temporary arcs may only last for a few milliseconds, and/or arcing may occur very infrequently —making these events nearly impossible to detect via visual inspection. While small events may not result in catastrophic failure or wildﬁre ignition, over time, they degrade the equipment and increase the probability that a catastrophic failure may occur in the future. As a more permanent (albeit costly) solution, distribu- tion lines can be strategically buried underground to elim- inate interference with vegetation above ground. Since the Black Saturday Fires in 2009, the state of Victoria, Australia has allocated over A$200 million toward line underground- ing. However, burying distribution lines underground can be costly at a large scale, costing hundreds of thousands of dollars per mile of line, and increasing maintenance costs and fault response times for the lifetime of the line [2]. B. ASSET INSPECTIONS AND TRACKING Therefore, specialized high impedance fault detection methods are com- monly introduced in distribution circuits with dense vege- tation, high impedance soil types, and areas with high ﬁre danger. A review of high impedance detection methods can be found in [48], [49]. To abate deterioration of distribution components, asset owners have traditionally scheduled routine inspection sched- ules for visual inspections. By patrolling distribution circuits, inspectors search for visible deterioration along the line. Intrusive inspections are also carried out whereby wooden distribution poles are inspected for decay, termite damage, or structural fatigue, and soil samples around the pole may be taken. Although line inspections can be carried out concurrently with other business, feeder patrols still A. ARC-FAULT SUPPRESSION Arc duration has a high correlation with probability of sus- tained ignition from power lines [36]. The probability of sustained ignition of surrounding fuel beds increases dras- tically over the course of milliseconds if an arc persists. Thus, any method to reduce the duration of (or eliminate) arcing in lines represents a key factor in wildﬁre-resilient grid operation. 3) RESONANT GROUNDING Although many of the preceding techniques can be taken to reduce the number of faults on overhead power dis- tribution systems, faults are not completely preventable. Therefore, it is necessary to investigate methods that reduce ignition likelihood in the case that a fault does occur in the system. Two methods are investigated in this section: 1) Arc-Fault Suppression, and 2) Proactive Operation Strategies. System neutral grounding practice differs around the world, and can have a signiﬁcant impact on fault current levels. On isolated neutral systems, resonant grounding systems (such as Peterson Coils) shift voltages on the faulted phase during a fault such that arcs are extinguished within millisec- onds [55]. However, there are two issues with this method. First, these systems can cause transient over-voltage, which can cause dangerous fault states that can damage equipment along the distribution circuit. Secondly, taps on Petersen Coils must be constantly adjusted to match distribution cir- cuit capacitance, requiring changes to Petersen Coil taps any time distribution topology changes. The cost and regulatory barriers for implementing resonant grounding in the United States make it uncommon in distribution systems. However, resonant grounding systems have been widely implemented in Victoria Australia’s power system as a response to the 2009 Victoria bushﬁres [25]. B. PROACTIVE OPERATION One straightforward method of reducing arc duration is to trip faults more quickly with existing time overcurrent protection. Tripping on distribution systems occurs by circuit breakers and fuses, that can be located throughout a distribution cir- cuit. Circuit breakers and fuses are coordinated to detect and trip downstream faults while leaving the upstream grid energized. Where applicable and reasonable, more sensitive circuit breaker settings or fuse ratings can be set to trip faults in less time —resulting in less arc contact with vegetation and/or less molten material ejected into combustible fuel beds. Additionally, some fuse types can emit sparks when blown, so careful consideration should be taken to replace these fuses with non-sparking alternatives in high threat ﬁre districts [54]. While utility companies work to ensure safe operation of their power lines, their efforts would likely not eliminate the chances of wildﬁre ignition for the foreseeable future. Thus, there is a need for operational solutions that can be carried out on systems as they exist today. Proactive operation strategies, such as proactive de-energization and network reconﬁgura- tion, prevent ignition on power lines by de-energizing lines before a catastrophic fault, failure, or ﬁre occurs. 2) DISABLED RECLOSERS model tree locations, heights, and approximate proximity to distribution conductors to enhance situational awareness and conduct more efﬁcient condition based trimming. As a result, utilities are able to more efﬁciently dispatch vegetation management crews along their distribution corridors to target high-risk areas. 2) DISABLED RECLOSERS According to many industry statistics, 60-70% of faults on distribution systems are self-clearing. If and when the exter- nal path to ground is cleared, lines can be re-energized with automated reclosers. However, in cases under which faults do not self-clear, attempting to re-energize distribution lines can cause a second arc event that is equally or more likely to result in sustained ignition compared to the initial fault. In many cases, re-energization attempts result in higher fault current than the initial fault event if the fault is not cleared. Additionally, recloser restrikes that occur within 5 seconds after the initial fault event are more likely to ignite surround- ing combustible materials. Time delays in reclosing attempts of 30 seconds or more can reduce ignition probability in cases in which the fault has not cleared. However, it remains unclear if reclosers on the market today have the ability to reclose on a time delay [36]. In California, reclosing is com- monly completely discontinued in HTFDs during ﬁre season. However, this action requires manual inspection of power lines before re-energization can occur, negatively impacting reliability metrics. Although common practice in the US allows utility com- panies to trim up to 12 feet (or more) of clearance from distri- bution lines [51], [52], case studies show that trees from up to 100 feet away from energized conductors can cause damage via airborne debris [53]. Grasses, nearby trees, underbrush, and other ignitable material surrounding power lines may not only interfere with power lines via vegetation growth or as airborne debris, but also could ignite as a result of ejected material from an arc. Therefore, distribution corridors can also be maintained and controlled for long-term vegetation growth to not only prevent faults, but also reduce the likeli- hood of ignition if a fault does occur by clearing dead and dry debris in proximity to power lines. Herbicides are commonly applied underneath the poles to limit vegetation growth [52]. D. VEGETATION MANAGEMENT Vegetation management represents one of the costliest and most challenging tasks to wildﬁre mitigation [50]. Histori- cally, vegetation management has been carried out on ﬁxed schedules to areas that may or may not need trimming. How- ever, advances in aerial imaging (e.g., drones, LIDAR) enable utilities to recognize areas that need trimming, efﬁciently VOLUME 7, 2020 369 1) PROACTIVE DE-ENERGIZATION Proactive de-energization (also known as Public Safety Power Shutoff) has been suggested as a wildﬁre mitigation strategy in many of the California Public Utility Wildﬁre 370 VOLUME 7, 2020 VOLUME 7, 2020 VOLUME 7, 2020 Muhs et al.: Wildfire Risk Mitigation: A Paradigm Shift in Power Systems Planning and Operation Mitigation Plans, and has been utilized in multiple high-risk ﬁre weather scenarios in Australia [54], [56]. Since permitted by CPUC, only a handful of de-energization events have been reported [57]. Proactive de-energization is permitted when high-risk ﬁre weather conditions present themselves in areas where power infrastructure is located, and keeping lines energized presents a signiﬁcant safety threat to surrounding communities. When considering de-energization as a ﬁre mit- igation technique, utilities consider several factors, including: the current state of power infrastructure, the state of surround- ing vegetation, seasonal weather and precipitation levels, and the impact that de-energization will have on the surrounding communities. At time of writing, the decision to de-energize has not been automated based on quantitative criteria alone. De-energization decisions largely relies on decisions by expe- rienced professionals who clearly understand the risks and trade-offs associated with de-energization [54]. restoration (FLISR) systems [58]. Rural distribution grids are typically radial in nature, but isolating the fault allows downstream sections of the grid to become energized if so designed. Power can be partially restored to meet critical loads by either grid network reconﬁguration or islanded oper- ation [59]–[62]. During network reconﬁguration, an alternate or backup network topology is utilized to provide an alternate path to grid sourced power for the islanded section of the grid. If distributed energy resources (DER) such as backup dispatchable generators, renewable energy sources, or energy storage systems exist in isolated sections of the grid, they can be utilized to provide the capability to temporarily or intermittently restore power to downstream customers. A study focused in Victoria, Australia found that solar PV and lithium-ion battery energy storage used in conjunction with power line de-energization could reduce comparable levels of ﬁre risk for approximately 10% of the cost of burying rural distribution lines underground [56]. In October 2018, a California utility company made the decision to enact its ﬁrst proactive de-energization based on forecasted high-risk ﬁre weather [57]. 1) ADVANCED WEATHER MODELING Weather plays a vital role in wildﬁre situational aware- ness. Historical weather trends can be characterized via cumulative distribution functions allowing meteorologists to recognize atypical weather patterns in utility service terri- tory. Likewise, detailed high resolution weather forecasts can be performed up to 48 hours in advance, although these forecasts become uncertain beyond 6 to 12 hours in the future. Weather forecasts are one of the most important inputs to the proactive de-energization deci- sion making process, and add value to the system oper- ators in a variety of other ways. In addition to installing remote automated weather stations, advanced weather modeling techniques are also being widely used by California FIGURE 1. Proactive de-energization event in October 2018. V. FIRE RESPONSE AND IMPACT MITIGATION Utilities with infrastructure located in high threat ﬁre districts are expected to allocate more resources and efforts into ﬁre response and impact mitigation. While electric utilities are not necessarily directly involved in ﬁre suppression, strate- gies such as maintaining situational awareness, and devel- oping protocols for mutual assistance and communication with customers have proven to be vital in cases in which catastrophic wildﬁre ignition occurs. This is also empha- sized by the regulatory agencies; for instance, the CPUC recently highlighted that utility activities in California should ‘‘be focused on preventing catastrophic wildﬁres, not simply ignitions’’ [63]. FIGURE 1. Proactive de-energization event in October 2018. A. WILDIRE RISK MODELING Maintaining situational awareness is a major priority for utility companies as highlighted in recent Wildﬁre Mitigation Plans [54]. Wildﬁre situational awareness is deﬁned as a dis- tribution system operator’s ability to monitor and understand the wildﬁre environment in their service territory [52]. 1) PROACTIVE DE-ENERGIZATION After the event, the utility company highlighted the extensive wind related damage to the de-energized system, including: 18 damaged spans of conductor, 5 damaged cross-arms, 3 damaged insu- lators, 2 damaged fuses, 1 damaged transformer, and 1 dam- aged pole. While the decision to de-energize caused loss of power to approximately 60,000 customers, this event stands as a prime example wherein weather forecasts and proactive de-energization were utilized to mitigate wildﬁre risk. Cus- tomer outage data were collected from a post-event report and are plotted in Fig. 1. Customers were forewarned up to 48 hours in advance, and feeders were repaired and restored in sections for two days following the de-energization event. A. CHARACTERIZING A WILDFIRE-RESILIENT POWER SYSTEM The deﬁnition of resilience should be clariﬁed as it relates to catastrophic wildﬁres. Existing resilience literature treats wildﬁres similarly to other natural disasters: an external and predictable event. However, recent cases show that decisions made by power system planners and operators affect power system ﬁre risk. The concept of wildﬁre-resilience power sys- tem should be expanded to encompass the unique relationship between power systems and catastrophic wildﬁres. Namely, that the grid should be designed and operated in such a way that power system assets are not only protected from external wildﬁre threats, but also that catastrophic wildﬁres are not ignited by power system infrastructure. If high-risk ﬁre weather is forecasted to occur, observa- tion crews may be dispatched to remote feeders to provide additional situational awareness and report potential hazards. Utilities strive to warn customers of weather related impacts to electrical service in advance. Utilities reach out to cus- tomers via phone, email, and social media (among other methods), and take special care to make contact with medical baseline customers. Utility companies coordinate with other key stakeholders such as municipalities, state government, and operators of natural gas, transportation, and water net- works if an impact to electrical service is expected [73]. Planning within the utility company also occurs during this time to ensure stafﬁng needs are met during the anticipated service interruption. B. UTILITY EMERGENCY PLANNING Utilities play an increasingly critical role in emergency man- agement during wildﬁre incidents. Emergency management is generically comprised of pre-event planning, real-time response, and long-term infrastructure restoration. 2) WILDFIRE DETECTION AND SPREAD MODELING 2) WILDFIRE DETECTION AND SPREAD MODELING High-Deﬁnition cameras installed in utility service terri- tory allow for early spotting and triangulation of wildﬁres. By installing high-deﬁnition cameras at strategic vantage points, utility companies are granted visual observation capa- bility to remote distribution feeders. To automate the pro- cess of wildﬁre detection, machine learning techniques are being developed to recognize ﬁres in camera footage by identifying visible features of ﬁres such as bright spots and grey smoke plumes [65]–[68]. If a wildﬁre is detected or reported, understanding likely spread scenarios is vital to protecting power infrastructure and conducting emer- gency response. Fire spread is predominately determined by a combination of land use type, slope, wind speed, and wind direction. A broad spectrum of ﬁre spread modeling tools exist in forestry research. The most complex models require the use of high performance computing to model a dynamic ﬁre boundary [69]–[71]. Other theoretical ﬁre spread models such as the minimum travel time method have been presented as computationally efﬁcient and sufﬁcient for research uses [72]. Each wildﬁre event occurs for different reasons, and impacts communities differently. Ultimately, each wildﬁre event teaches the power system community a lesson that serves to inform utility practice for future catastrophic wild- ﬁre events. VI. CONCLUSION AND WILDFIRE MITIGATION RESEARCH DIRECTION This article provided a comprehensive overview of wildﬁre risk mitigation techniques for power systems. The risk miti- gation techniques are categorized according to their function in mitigating wildﬁres as: 1) Fault Prevention, 2) Arc-Ignition Prevention, and 3) Fire Response and Impact Mitigation. For each solution, relevant literature and use cases are presented. The concept of mitigating catastrophic wildﬁres presents a promising area of future research, and offers many problems that can spark innovation in research and practice commu- nities. Three major directions for future wildﬁre mitigation research are summarized next. 2) NETWORK RECONFIGURATION AND DISTRIBUTED ENERGY RESOURCES Following disruptions in power service caused by proactive de-energizations, the distribution grid can be automatically sectionalized using fault location, isolation, and service 371 VOLUME 7, 2020 utilities, and are quickly becoming more computationally efﬁcient and accurate [64]. service personnel may be evacuated from the area until the ﬁre is contained. Once utility personnel obtain permission to enter a ﬁre zone, a thorough damage assessment of the power system is conducted, and repairs are coordinated so as to restore power to as many customers as possible within the shortest time – depending on the extent of damage and complexity of the repair work. 2) WILDFIRE DETECTION AND SPREAD MODELING B. QUANTIFYING WILDFIRE RISK AND EFFECTIVENESS OF MITIGATION TECHNIQUES Modeling and quantifying the wildﬁre risk is crucial to utility companies’ ability to understand wildﬁre risk and strategically harden the power grid. An example of wild- ﬁre risk modeling techniques is an approach that was used to strategically harden Victoria’s power grid as a result of the Victoria Royal Commission following the 2009 Black Saturday Bushﬁres [74]. In this analysis, wildﬁre ignition likelihood was calculated as a function of historical fault data, and consequence was determined using a Monte Carlo-based spread model that took into account historical wind direc- tions. Several ﬁres were simulated along every node in the system, and damage was quantiﬁed by the number of If a ﬁre incident occurs, system operators and deci- sion makers convene with other key stakeholders in local operation centers to coordinate ﬁre response. Fire suppres- sion ground and aerial crews are dispatched once ignition occurs and the wildﬁre location is conﬁrmed. Power lines, although unaffected by the wildﬁre incident, may need to be de-energized to allow ﬁre crews to operate in their vicin- ity safely. Fire boundaries and anticipated weather changes are closely monitored to ensure response crew safety. Some utilities provide backup generation to hospitals and com- munity centers during prolonged service interruptions [54]. Depending on the severity of the ﬁre, communities and utility VOLUME 7, 2020 372 VOLUME 7, 2020 Muhs et al.: Wildfire Risk Mitigation: A Paradigm Shift in Power Systems Planning and Operation problems based on wildﬁre mitigation effectiveness, timeline for technology implementation, and budgetary constraints. The authors in [58] utilized mixed integer linear pro- gramming to determine where to implement sectionalizing switches in distribution feeders with a constrained budget. Similar approaches can be taken to consider the various other technologies available to mitigate wildﬁre ignition probabil- ity such as waveform prognostics sensors, high impedance fault detection technologies, and the underlying infrastructure required to enable these technologies on rural distribution systems. A prerequisite to these analyses, however, is a clear understanding of the relative beneﬁts of each technology to mitigate wildﬁre. buildings destroyed. By combining ignition likelihood with consequence of spread, a risk value was generated for nodes across Victoria’s power grid, allowing the system planners to identify areas in the regional power grid at which infrastruc- ture upgrades would result in the most wildﬁre risk reduction. B. QUANTIFYING WILDFIRE RISK AND EFFECTIVENESS OF MITIGATION TECHNIQUES pg Further, quantifying the effectiveness of wildﬁre risk mit- igation techniques represents one of the foremost challenges in wildﬁre mitigation planning. Currently, there is no uniﬁed method of quantifying the complex nature of wildﬁre risk in power systems. However, by categorizing wildﬁre mitigation methods by their function (e.g., preventing faults, reduce arc- ing probability, etc.), the beneﬁts of these various techniques can ultimately be quantiﬁed. As faults, failures and ﬁres on power distribution systems are all rare events, probabilistic methods will likely need to be utilized. Once the beneﬁt of various wildﬁre mitigation techniques is successfully quanti- ﬁed, these measures can be used in cost-beneﬁt analysis and selected for utilities based on their speciﬁc needs. Understanding and modeling of high-risk ﬁre weather presents several opportunities to make decisions based on trade-offs both on and off power systems. Optimal power ﬂow objectives such as minimizing load curtailment in public safety power shutoff scenarios through use of distributed energy resources (DER) has already been studied, though not in the context of wildﬁre risk mitigation (see e.g., [60], [62]). As DER become more cost-effective, a variety of opera- tional optimization techniques could be used to guide oper- ational practices, which would lead to a lower likelihood of wildﬁre ignition caused by power systems. In known ﬁre weather scenarios, there has often been a 24-48 hour lead time during which precarious ﬁre weather was fore- casted. During this time, DER, mobile generators, and other resilience-enhancing technologies can be coordinated and dispatched as a comprehensive utility response strategy. Other operational challenges such as condition-based line inspection and maintenance may also inform utility practices in the area of mitigating catastrophic wildﬁres. REFERENCES [1] J.-P. Watson et al., ‘‘Conceptual framework for developing resilience met- rics for the electricity oil and gas sectors in the United States,’’ Sandia Nat. Laboratories, Albuquerque, NM, USA, Tech. Rep., 2014. Laboratories, Albuquerque, NM, USA, Tech. Rep., 2014. [2] ‘‘Enhancing distribution resiliency–Opportunities for applying innovative technologies,’’ Electr. Power Res. Inst., Washington, DC, USA, Tech. Rep., 2013. FIGURE 2. Trade-off between wildfire mitigation solutions. [3] Y. Wang, C. Chen, J. Wang, and R. Baldick, ‘‘Research on resilience of power systems under natural disasters—A review,’’ IEEE Trans. Power Syst., vol. 31, no. 2, pp. 1604–1613, Mar. 2016. The trade-off between cost, timeline for implementation, and wildﬁre mitigation effectiveness that are central to the wildﬁre mitigation asset planning problem is shown in Fig. 2. For example, proactive de-energization is effective at prevent- ing arc-ignitions, and can be carried out on power grids as it exists today (i.e., short implementation timeline), but conse- quently carries a high cost to customers. However, hardening measures such as burying lines underground also effectively prevent arc-ignitions, but burying lines comes at a high cost and may take years to implement across utility service terri- tories. In addition to these three trade-offs, efforts to reduce wildﬁre risk should consider the changing nature of the grid, and be aligned with other long-term grid modernization goals. [4] Executive Ofﬁce of the President, ‘‘Economic beneﬁts of increasing electric grid resilience to weather outages,’’ Executive Ofﬁce of the President, Tech. Rep., 2013. [Online]. Available: https://www.energy.gov/ sites/prod/ﬁles/2013/08/f2/Grid%20Resiliency%20Report_FINAL.pdf p y p p [5] J. W. Muhs and M. Parvania, ‘‘Stochastic spatio-temporal hurricane impact analysis for power grid resilience studies,’’ in Proc. IEEE Power Energy Soc. Innov. Smart Grid Technol. Conf. (ISGT), Feb. 2019, pp. 1–5. [6] M. Panteli, C. Pickering, S. Wilkinson, R. Dawson, and P. Mancarella, ‘‘Power system resilience to extreme weather: Fragility modeling, proba- bilistic impact assessment, and adaptation measures,’’ IEEE Trans. Power Syst., vol. 32, no. 5, pp. 3747–3757, Sep. 2017. [7] T. Koppel, Lights Out: A Cyberattack, a Nation Unprepared, Surviving Aftermath. New York, NY, USA: Broadway Books, 2015. [8] M. M. S. Khan, A. Palomino, J. Brugman, J. Giraldo, S. K. Kasera, and M. Parvania, ‘‘The cyberphysical power system resilience testbed: Archi- tecture and applications,’’ Computer, vol. 53, no. 5, pp. 44–54, May 2020. pp p pp y [9] P. Hines, J. Apt, and S. Talukdar, ‘‘Trends in the history of large blackouts in the United States,’’ in Proc. C. PLANNING AND OPERATION OF WILDFIRE-RESILIENT POWER SYSTEMS From a power system planning standpoint, there are major trade-offs between the various wildﬁre mitigation solutions that have been highlighted in this article. Wide ranging factors such as utility history and location, customer needs, previous design choices and standards all affect the decision on how to mitigate wildﬁres. Understanding the potential costs of taking no mitigation action may aid decision makers in allocating the necessary funding for wildﬁre mitigation improvements. FIGURE 2. Trade-off between wildfire mitigation solutions. REFERENCES IEEE Power Energy Soc. Gen. Meeting- Convers. Del. Electr. Energy 21st Century, Jul. 2008, pp. 1–8. A variety of opportunities exist for deploying advanced optimization techniques to solve power system planning VOLUME 7, 2020 VOLUME 7, 2020 VOLUME 7, 2020 373 [10] L. Dueñas-Osorio and S. M. Vemuru, ‘‘Cascading failures in com- plex infrastructure systems,’’ Struct. Saf., vol. 31, no. 2, pp. 157–167, Mar. 2009. [35] A. C. Fernandez-Pello et al., ‘‘Spot ﬁre ignition of natural fuel beds by hot metal particles, embers, and sparks,’’ Combustion Sci. Technol., vol. 187, nos. 1–2, pp. 269–295, Jan. 2015. pp [36] D. Coldham, A. Czerwinski, and T. Marxsen, ‘‘Probability of bushﬁre ignition from electric arc faults,’’ HRL Technol. Pty Ltd., Melbourne, VIC, Australia, Tech. Rep. HRL/2010/195, 2011. [11] M. Ouyang, L. Dueñas-Osorio, and X. Min, ‘‘A three-stage resilience analysis framework for urban infrastructure systems,’’ Structural Saf., vols. 36–37, pp. 23–31, May 2012. [37] Q. Chen, ‘‘The probability, identiﬁcation, and prevention of rare events in power systems,’’ Ph.D. dissertation, Dept. Elect. Comput. Eng., Iowa State Univ., Ames, IA, USA, 2004. [12] E. Vugrin, A. Castillo, and C. Silva-Monroy, ‘‘Resilience metrics for the electric power system: A performance-based approach,’’ Sandia Nat. Lab., Albuquerque, NM, USA, Tech. Rep. SAND2017-1493, 2017. [38] J. W. Mitchell, ‘‘Power lines and catastrophic wildland ﬁre in Southern California,’’ in Proc. 11th Int. Conf. Fire Mater., 2009, pp. 225–238. [13] M. Panteli, P. Mancarella, D. N. Trakas, E. Kyriakides, and N. D. Hatziargyriou, ‘‘Metrics and quantiﬁcation of operational and infrastructure resilience in power systems,’’ IEEE Trans. Power Syst., vol. 32, no. 6, pp. 4732–4742, Nov. 2017. , f , , pp [39] P. Zahodiakin, ‘‘Making distribution grids stronger, more resilient,’’ EPRI J., vol. 4, pp. 4–8, Aug. 2016. [39] P. Zahodiakin, ‘‘Making distri J., vol. 4, pp. 4–8, Aug. 2016. [40] J. W. Mitchell, ‘‘Power line failures and catastrophic wildﬁres under extreme weather conditions,’’ Eng. Failure Anal., vol. 35, pp. 726–735, Dec. 2013. pp [14] M. M. Hosseini and M. Parvania, ‘‘Quantifying impacts of automation on resilience of distribution systems,’’ IET Smart Grid, vol. 3, no. 2, pp. 144–152, Apr. 2020. [41] C. Williamson, M. Garner, K. Barber, and R. Elford, ‘‘Development and introduction of aerial bundled cables in Australia,’’ in Proc. Int. Conf. Overhead Line Design Construct., Theory Pract., Nov. 1988, pp. 113–117. [15] E. I. Koufakis, P. T. Tsarabaris, J. S. Katsanis, C. G. Karagiannopoulos, and P. D. REFERENCES Bastrop Fires Began With Electric Power Line Sparks. [Online]. Available: https://www.houstonchronicle. com/news/houston-texas/article/Bastrop-ﬁres-began-with-electric-power- line-2180666.php [49] B. Aucoin and B. Russell, ‘‘Distribution high impedance fault detection utilizing high frequency current components,’’ IEEE Trans. Power App. Syst., vol. PAS-101, no. 6, pp. 1596–1606, Jun. 1982. [24] Las Conchas Fire Fact Sheet. Accessed: Aug. 2019. [Online]. Available: https://swﬁreconsortium.org/wp-content/uploads/2012/11/Las-Conchas- Factsheet_bsw.pdf [50] J. A. Wischkaemper, C. L. Benner, and B. D. Russell, ‘‘Electrical character- ization of vegetation contacts with distribution conductors-investigation of progressive fault behavior,’’ in Proc. IEEE/PES Transmiss. Distrib. Conf. Expo., Apr. 2008, pp. 1–8. [25] B. Teague, R. McLeod, and S. Pascoe, Victorian Bushﬁres Royal Commis- sion Final Report. Melbourne, VIC, Australia: Parliament Victoria, 2009. p p pp [51] General Order 95, Appendix e, Clearance of Poles, Towers and Struc- tures From Railroad Tracks. Accessed: Aug. 2019. [Online]. Available: https://www.cpuc.ca.gov/gos/GO95/go_95_appendix_e.html [26] W. Ekard, H. Tuck, and R. Lane, San Diego County Firestorms After Action Report. San Diego, CA, USA: Ofﬁce Emergency Services, 2007. [27] CAL Fire Incident Report, Rice Fire. Accessed: Aug. 2019. [Online]. Avail- able: https://www.ﬁre.ca.gov/incidents/2007/10/22/rice-ﬁre/#incident- contacts [52] San Diego Gas & Electric Company’s (u 902 e)Wildﬁre Mitigation Plan. Website. Accessed: Oct. 25, 2018. [Online]. Available: https://www. sdge.com/sites/default/ﬁles/regulatory/R.18-10-007%20SDG% [28] M. Groves. (May 2013). Edison Agrees to Pay $37 Million in 2007 Wildﬁre. [Online]. Available: https://www.latimes.com/local/la-xpm-2013-may-21- la-me-0521-malibu-ﬁre-settlement-20130521-story.html [53] S. D. Guikema, R. A. Davidson, and H. Liu, ‘‘Statistical models of the effects of tree trimming on power system outages,’’ IEEE Trans. Power Del., vol. 21, no. 3, pp. 1549–1557, Jul. 2006. pp [54] CPUC Wildﬁre Mitigation Plans From SB 901. Accessed: Aug. 2019. [Online]. Available: http://www.cpuc.ca.gov/sb901/ [29] Ash Wednesday Bushﬁre, Australian Institute for Disaster Resilience. Accessed: Aug. 2019. [Online]. Available: https:// knowledge.aidr.org.au/resources/bushﬁre-ash-wednesday-1983/ [55] (Aug. 2014). REFCL (Rapid Earth Fault Current Limiter) Trial–Ignition Tests in ‘Wire on Ground’ Faults Under Worst Case Fire Weather Conditions. ch 8 Appendices. Website. [Online]. Available: https://www. researchgate.net/publication/270893674_REFCL_Rapid_Earth_Fault_ Current_Limiter_Trial_-_ignition_tests_in_’wire_on_ground’_faults_ under_wo rst_case_ﬁre_weather_conditions_Ch_8_Appendices, doi: 10.13140/2.1.2930.9761. [30] C. Miller et al., ‘‘Electrically caused wildﬁres in Victoria, Australia are over-represented when ﬁre danger is elevated,’’ Landscape Urban Plan- ning, vol. 167, pp. 267–274, Nov. 2017. [31] V. C. Radeloff et al., ‘‘Rapid growth of the US wildland-urban inter- face raises wildﬁre risk,’’ Proc. Nat. Acad. Sci. USA, vol. 115, no. 13, pp. 3314–3319, Mar. 2018. [56] M. Williamson, ‘‘Application of distributed solar photovoltaics and energy storage to mitigate bushﬁre risk in Victoria, Australia,’’ Ph.D. dissertation, School Mech., Elect. Manuf. REFERENCES Bourkas, ‘‘A wildﬁre model for the estimation of the temperature rise of an overhead line conductor,’’ IEEE Trans. Power Del., vol. 25, no. 2, pp. 1077–1082, Apr. 2010. [42] X. Zhang, E. Gockenbach, V. Wasserberg, and H. Borsi, ‘‘Estimation of the lifetime of the electrical components in distribution networks,’’ IEEE Trans. Power Del., vol. 22, no. 1, pp. 515–522, Jan. 2007. [16] M. Choobineh, B. Ansari, and S. Mohagheghi, ‘‘Vulnerability assessment of the power grid against progressing wildﬁres,’’ Fire Saf. J., vol. 73, pp. 20–28, Apr. 2015. [43] A. Crosby, ‘‘Special research topic report on current practice in utility distribution poles and light poles,’’ Tech. Rep., 2011. [Online]. Available: https://docplayer.net/14511014-Special-research-topic-report-on-current- practice-in-utility-distribution-poles-and-light-poles-adam-crosby-date-5- 05-11.html [17] S. Mohagheghi and S. Rebennack, ‘‘Optimal resilient power grid operation during the course of a progressing wildﬁre,’’ Int. J. Electr. Power Energy Syst., vol. 73, pp. 843–852, Dec. 2015. [18] D. N. Trakas and N. D. Hatziargyriou, ‘‘Optimal distribution system oper- ation for enhancing resilience against wildﬁres,’’ IEEE Trans. Power Syst., vol. 33, no. 2, pp. 2260–2271, Mar. 2018. [44] General Order 165: Inspection Requirements for Electric Distribution and Transmission Facilities. Accessed: Aug. 2019. [Online]. Available: https://www.cpuc.ca.gov/gos/GO165/GO_165(III).html [19] Camp Fire Incident Report, CAL Fire. Accessed: Aug. 2019. [Online]. Available: https://ﬁre.ca.gov/incident/?incident=75dafe80-f18a-4a4a- 9a37-4b564c5f6014 p p g g [45] B. D. Russell and C. L. Benner, ‘‘Intelligent systems for improved reliabil- ity and failure diagnosis in distribution systems,’’ IEEE Trans. Smart Grid, vol. 1, no. 1, pp. 48–56, Jun. 2010. [20] R. Ramos, ‘‘Sce woolsey letter to CPUC,’’ Southern California Edison, Rosemead, CA, USA, Tech. Rep., 2018. [46] J. A. Wischkaemper, C. L. Benner, B. D. Russell, and K. Manivannan, ‘‘Application of waveform analytics for improved situational awareness of electric distribution feeders,’’ IEEE Trans. Smart Grid, vol. 6, no. 4, pp. 2041–2049, Jul. 2015. [21] C. Saqui, ‘‘Thomas ﬁre investigation report,’’ California Dept. Forestry Fire Protection San Luis Obispo Unit, San Luis Obispo, CA, USA, Tech. Rep., 2019. pp [47] C. L. Benner and B. D. Russell, ‘‘Practical high-impedance fault detec- tion on distribution feeders,’’ IEEE Trans. Ind. Appl., vol. 33, no. 3, pp. 635–640, May/Jun. 1997. [22] S. Ahillen. (Dec. 2016). How the Historic Gatlinburg Fires Happened; a Humble Start. [Online]. Available: https://www.govtech. com/em/disaster/How-the-historic-Gatlinburg-ﬁres-happened-a-humble- start.html pp y [48] C. G. Wester, ‘‘High impedance fault detection on distribution systems,’’ in Proc. Rural Electr. Power Conf. Presented 42nd Annu. Conf., Apr. 1998, pp. 1–5 [23] R. B. Brazziel. (Sep. 2011). [10] L. Dueñas-Osorio and S. M. Vemuru, ‘‘Cascading failures in com- plex infrastructure systems,’’ Struct. Saf., vol. 31, no. 2, pp. 157–167, Mar. 2009. REFERENCES Eng., Loughborough Univ., Loughborough, U.K., 2015. pp [32] D. Thomas, D. Butry, S. Gilbert, D. Webb, and J. Fung, ‘‘The costs and losses of wildﬁres,’’ Nat. Inst. Standards Technol., Tech. Rep. 1215, 2017. [33] B. D. Russell, C. L. Benner, and J. A. Wischkaemper, ‘‘Distribution feeder caused wildﬁres: Mechanisms and prevention,’’ in Proc. 65th Annu. Conf. Protective Relay Eng., Apr. 2012, pp. 43–51. [57] PG&E Wildﬁre Mitigation Plans From SB 901. Accessed: Aug. 2019. [Online]. Available: https://www.cpuc.ca.gov/uploadedFiles/CPUC_ Public_Website/Content/Utili ties_and_Industries/Energy_-_Electricity_ and_Natural_Gas/PGE%20PSPS%20Report%20Letter%2020181031.pdf y g p pp [34] C. D. Zak, ‘‘The effect of particle properties on hot particle spot ﬁre ignition,’’ Ph.D. dissertation, Dept. Mech. Eng., Univ. California, Berkeley, Berkeley, CA, USA, 2015. 374 VOLUME 7, 2020 Muhs et al.: Wildfire Risk Mitigation: A Paradigm Shift in Power Systems Planning and Operation [58] A. Abiri-Jahromi, M. Fotuhi-Firuzabad, M. Parvania, and M. Mosleh, ‘‘Optimized sectionalizing switch placement strategy in distribution sys- tems,’’ IEEE Trans. Power Del., vol. 27, no. 1, pp. 362–370, Jan. 2012. [66] O. Günay, K. Tasdemir, B. U. Töreyin, A. E. Çetin, ‘‘Video based wildﬁre detection at night,’’ Fire Saf. J., vol. 44, no. 6, pp. 860–868, Aug. 2009. [67] R. Allison, J. Johnston, G. Craig, and S. Jennings, ‘‘Airborne optical and thermal remote sensing for wildﬁre detection and monitoring,’’ Sensors, vol. 16, no. 8, p. 1310, Aug. 2016. [59] M. M. Hosseini, A. Umunnakwe, and M. Parvania, ‘‘Automated switching operation for resilience enhancement of distribution systems,’’ in Proc. IEEE Power Energy Soc. Gen. Meeting (PESGM), Aug. 2019, pp. 1–5. [68] M. M. Hosseini, A. Umunnakwe, M. Parvania, and T. Tasdizen, ‘‘Intelli- gent damage classiﬁcation and estimation in power distribution poles using unmanned aerial vehicles and convolutional neural networks,’’ IEEE Trans. Smart Grid, vol. 11, no. 4, pp. 3325–3333, Jul. 2020. [60] H. T. Nguyen, J. W. Muhs, and M. Parvania, ‘‘Assessing impacts of energy storage on resilience of distribution systems against hurricanes,’’ J. Modern Power Syst. Clean Energy, vol. 7, no. 4, pp. 731–740, Jul. 2019. Smart Grid, vol. 11, no. 4, pp. 3325–3333, Jul. 2020. y gy pp [61] M. M. Hosseini and M. Parvania, ‘‘Computationally efﬁcient formula- tions for fault isolation and service restoration in distribution systems,’’ in Proc. IEEE Power Energy Soc. Gen. Meeting (PESGM), Aug. 2020, pp. 1–5. [69] C. Tymstra et al., ‘‘Development and structure of prometheus: The Cana- dian Wildland ﬁre growth simulation model,’’ Can. REFERENCES Forest Service, North- ern Forestry Centre, Edmonton, AB, Canada, Inf. Rep. NOR-X-417, 2010. [70] A. K. Kochanski, M. A. Jenkins, J. Mandel, J. D. Beezley, and S. K. Krueger, ‘‘Real time simulation of 2007 Santa Ana ﬁres,’’ Forest Ecology Manage., vol. 294, pp. 136–149, Apr. 2013. pp [62] H. Nguyen, J. Muhs, and M. Parvania, ‘‘Preparatory operation of automated distribution systems for resilience enhancement of critical loads,’’ IEEE Trans. Power Del., early access, Oct. 14, 2020, doi: 10.1109/TPWRD.2020.3030927. pp p [71] J. Mandel, J. D. Beezley, and A. K. Kochanski, ‘‘Coupled atmosphere- wildland ﬁre modeling with WRF-ﬁre,’’ 2011, arXiv:1102.1343. [Online]. Available: http://arxiv.org/abs/1102.1343 [63] Decision 19-05-036: Guidance Decision on 2019 Wildﬁre Mitigation Plans Submitted Pursuant to Senate Bill 901 (Rulemaking 18-10-007). Accessed: Mar. 6, 2019. [Online]. Available: http://docs.cpuc.ca. gov/PublishedDocs/Published/G000/M296/K577/296577466.PDF [72] M. A. Finney, ‘‘Fire growth using minimum travel time methods,’’ Can. J. Forest Res., vol. 32, no. 8, pp. 1420–1424, Aug. 2002. [73] M. Ouyang and Z. Wang, ‘‘Resilience assessment of interdependent infras- tructure systems: With a focus on joint restoration modeling and analysis,’’ Rel. Eng. Syst. Saf., vol. 141, pp. 74–82, Sep. 2015. [64] B. K. Blaylock, J. D. Horel, and C. Galli, ‘‘High-resolution rapid refresh model data analytics derived on the open science grid to assist wildland ﬁre weather assessment,’’ J. Atmos. Ocean. Technol., vol. 35, no. 11, pp. 2213–2227, Nov. 2018. [74] D. Huston, G. Towns, A. Stephenson, S. Woodman, and A. Hunt, ‘‘Assess- ing the role of electricity networks in bushﬁre ignitions: Estimation of current and prospective performance,’’ in Proc. Int. Congr. Modelling Simulation, vol. 21, 2015, pp. 1–4. pp [65] B. U. Toreyin and A. E. Cetin, ‘‘Wildﬁre detection using LMS based active learning,’’ in Proc. IEEE Int. Conf. Acoust., Speech Signal Process., Apr. 2009, pp. 1461–1464. 375 VOLUME 7, 2020 VOLUME 7, 2020 375
https://openalex.org/W4228996682	https://www.granthaalayahpublication.org/journals/granthaalayah/article/download/4567/4674	English	null	THE EFFECT OF INFORMATION TECHNOLOGY CAPABILITY, AND STRATEGIC PLANNING ON KNOWLEDGE MANAGEMENT MEDITED ORGANIZATIONAL PERFORMANCE IN THE ARMY CENTRAL IMPLEMENTING AGENCY	International journal of research - granthaalayah	2,022	cc-by	7,008	Original Article ISSN (Online): 2350-0530 ISSN (Print): 2394-3629 Original Article ISSN (Online): 2350-0530 ISSN (Print): 2394-3629 Original Article ISSN (Online): 2350-0530 ISSN (Print): 2394-3629 International Journal of Research - GRANTHAALAYAH April 2022 10(4), 99-111 1, 2, 3 Faculty of Economics, University Trisakti Jakarta, Indonesia 1, 2, 3 Faculty of Economics, University Trisakti Jakarta, Indonesia Received 15 March 2022 Accepted 15 April 2022 Published 04 May 2022 Corresponding Author Eko Daryanto, daryanyoeko@gmail.com DOI 10.29121/granthaalayah.v10.i4.2022 .4567 Received 15 March 2022 Accepted 15 April 2022 Published 04 May 2022 Corresponding Author Eko Daryanto, daryanyoeko@gmail.com DOI 10.29121/granthaalayah.v10.i4.2022 .4567 How to cite this article (APA): Daryanto, E., Jasfar, F., and Kusnadi. (2022). The Effect of Information Technology Capability, and Strategic Planning on Knowledge Management Medited Organizational Performance in The Army Central Implementing Agency. International Journal of Research - GRANTHAALAYAH, 10(4), 99-111. doi: 10.29121/granthaalayah.v10.i4.2022.4503 Keywords: Information Technology Capability, Strategic Planning, Knowledge Management, Organizational Performance ABSTRACT The purpose of this study is to provide input to the leadership of the Indonesian Army National Armed Forces Headquarters regarding the importance of knowledge management on technological developments because this knowledge is one way to create and improve work professionalism in the military, especially the Indonesian Armed Forces. The research design and method used in this research is hypothesis testing. This research is a quantitative non-experimental, using a questionnaire given to 93 respondents. This research was conducted at the Central Implementing Agency for the Indonesian National Armed Forces. The results of this study indicate that: The effect of Information Technology Capability, Strategic Planning on Organizational Performance is positive and significant. The influence of Knowledge Management on Organizational Performance is positive and significant. The influence of Knowledge Management mediating Information Technology Capability, Strategic Planning on Organizational Performance is positive and significant. It means that the higher/positive mediation of Knowledge Management on Information Technology Capability, Strategic Planning, the higher/positive Organizational Performance. Theoretical implications: Knowledge Management as a mediating of Information Technology Capability, Strategic Planning is able to improve Organizational Performance, by increasing Knowledge Management through increasing its dimensions it will be able to improve the Organizational Performance of the Central Implementing Body of the Indonesian Armed Forces. Received 15 March 2022 Accepted 15 April 2022 Published 04 May 2022 DOI 10.29121/granthaalayah.v10.i4.2022 .4567 Funding: This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors. Copyright: © 2022 The Author(s). This work is licensed under a Creative Commons Attribution 4.0 International License Copyright: © 2022 The Author(s). This work is licensed under a Creative Commons Attribution 4.0 International License. With the license CC-BY, authors retain the copyright, allowing anyone to download, reuse, re-print, modify, distribute, and/or copy their contribution. The work must be properly attributed to its author. THE EFFECT OF INFORMATION TECHNOLOGY CAPABILITY, AND STRATEGIC PLANNING ON KNOWLEDGE MANAGEMENT MEDITED ORGANIZATIONAL PERFORMANCE IN THE ARMY CENTRAL IMPLEMENTING AGENCY Eko Daryanto 1 , Farida Jasfar 2, Kusnadi 3 1. INTRODUCTION The Indonesian Armed Forces of the Army has gradually implemented technology through rejuvenating the equipment and main tools of its weapons system on a computerized basis. The working mechanism in units of the Indonesian Army's National Armed Forces has used information systems through application development to simplify and speed up tasks and can provide more accurate data. Based on Decree number 66/2019 concerning the Organizational Structure of the Indonesian Army National Army. Then the Central Implementing Agency is a work 99 The Effect of Information Technology Capability, and Strategic Planning on Knowledge Management Medited Organizational Performance in The Army Central Implementing Agency The Effect of Information Technology Capability, and Strategic Planning on Knowledge Management Medited Organizational Performance in The Army Central Implementing Agency unit at the Indonesian National Armed Forces Headquarters level and the Army Headquarters which serves as staff and implementers of activities at the central level within the scope of the Indonesian National Armed Forces Headquarters or the Army Headquarters. The Organization of the Central Implementing Body of the Indonesian Army National Armed Forces is an organization formed by overseeing other organizations including: The Army Information Service was ratified based on the Army Chief of Staff Regulation No. 6 of 2015. The main task is to carry out the Information function in order to support the main tasks of the Army. The Center for Passwords and Cyber is established based on Army Chief of Staff Regulation No. 26 of 2019. The main task is to carry out personnel development as well as password and cyber functions in order to support the tasks of the Indonesian Army and the Indonesian National Army. The Army Information and Data Processing Service is established based on the Army Chief of Staff Regulation No. 24 of 2017. The task of the Army Information and Data Processing Service is to organize information system development and data processing which includes the construction, development, maintenance and preparation of the Indonesian Armed Forces Information system. The Army electronically covering the fields of intelligence, operations, personnel, logistics/material, territorial, finance, force development planning, programs and budgets, methods and maintenance in order to support the tasks of the Indonesian Army. International Journal of Research - GRANTHAALAYAH 1. INTRODUCTION The Indonesian Army National Army carries out its activities in a competitive environment, and if the ranks of the Indonesian Army National Army want to survive and exist, they must be prepared to face many challenges whose impacts will affect performance both positively and negatively. Organizations that can face challenges by using resources as much and efficiently as possible will be able to maintain and even improve their business performance. Organizational performance is a vital indicator of organizational success or failure. The challenge faced by the ranks of the Indonesian Armed Forces at this time is the unstoppable development of industry 4.0 so that an organizational reaction is needed to prepare all organizational resources which include knowledge and capabilities of Human Resources, capabilities in the field of information technology, capabilities in the field of Human Resource management and leadership. An empirical study conducted by Jain (2016) proves that an increase in the use of technology significantly increases the need for human resource development such as formal and informal training and education (training and development). Purdue further, (2016) Purdue (2016) explains that in an environment where knowledge and technology continue to grow and develop, training and development are key factors in organizational success. Smart organizations invest in employee training and development to improve job performance, increase productivity and boost morale all of which are aimed at improving the overall performance of the organization. This is in line with Roth's research, (2016) Roth (2016) which suggests that technology development significantly improves personnel performance in supporting organizational performance. Military organizations are divided into systems, natural systems, rational systems and open systems. Natural system means that military organizations struggle to defend themselves and protect interests in an environment of scarce resources and conflict. The rational system means that military organizations aim to increase efficiency in how to secure the country. A military organization like this will determine a strategy based on the International Journal of Research - GRANTHAALAYAH 100 Eko Daryanto, Farida Jasfar, and Kusnadi geography, technological development, and behavior of its adversary. Meanwhile, the open system means that the military organization is a manifestation of strong institutional rules and myths that bind its members. 1. INTRODUCTION The research gap found is that there is still limited comprehensive research in Indonesia related to the influence of Information Technology Capability in the military, Strategic Planning and Knowledge Management on the Organizational Performance of the Indonesian Army's military, as well as the nature of information on the military whose characteristics are limited and confidential. the first perception related to military information technology which is scientifically reviewed and published to the public. This can be seen from previous research: "Process of strategic planning and cognitive diversity as determinants of cohesiveness and performance" Nowak (2020). This study found a negative effect of cognitive differences among employees on unit cohesiveness and performance and a positive moderating effect of the strategic planning process on these relationships. Pham and Hoang (2019); Nowak (2020). The novelty in this research is, Information Technology Capability and Strategic Planning, namely strengthening Information Technology Capability which is supported by the right plan and the placement of Knowledge Management to improve the Organizational Performance of the Indonesian Army's military. International Journal of Research - GRANTHAALAYAH organizational needs (Wade (2013); Barua and Mitragotri (2014); Acemoglu (2015); Mathis et al. (2017); Parrish (2018); Van Waeyenberge (2019)). The implementation of knowledge management is related to technology, organizational adjustment, evaluation of knowledge management activities and / or knowledge resources, regulating / managing activities and / or knowledge resources, employee motivation, and external factors can improve organizational performance (Holsapple et al. (2014)) . Synergy in knowledge management practices is needed so that it can have a significant impact on organizational performance. (Turulja and Bajgoric (2019)) shows that information technology capabilities will improve Human Resource management capabilities and improve information technology knowledge management capabilities. As a result, information technology knowledge management capabilities and Human Resources management capabilities will improve the organization's business performance. In addition, there is an interaction effect of knowledge management capability and Human Resource management capability on business performance. This study is in line with the results of research (Ahmed et al. (2018)) showing that the orientation of knowledge management capabilities strategies that are different and complementary to one another is an important relationship in the relationship between knowledge management strategies and organizational performance, but this study also shows there is no direct influence of knowledge management strategy orientation on organizational performance. Strategic Planning is the approach of the Indonesian Army's National Armed Forces in carrying out the main organizational tasks carried out by the Indonesian Army's National Armed Forces Headquarters through the Indonesian Army's Central Implementing Body Organization. Coordinate with the organizational division of the Army Information Service, the Center for Passwords and Cyber, the Army Information and Data Processing Service. Organizational performance is influenced by several factors, including Strategic Planning (Bolland and Bolland (2020)). Strategic Planning is a systemic process that is agreed upon by the organization and builds engagement among stakeholders, especially regarding the essential priorities for its mission and responsiveness to the operational environment (Barron and Chou (2017)). The Department of Human Resource Management considers military experience a plus, in part, due to the mission-oriented work ethic of ex-military soldiers and quality management training. The military is related to the knowledge management program used in developing training quite comprehensively. The private sector can learn more from the military, including more effective ways of applying information technology to remote learning and teaching. (Frey and Osborne (2017); Garcia-Lausin et al. (2019)). International Journal of Research - GRANTHAALAYAH 2. LITERATURE REFERENCES This study tries to develop theories about Information Technology Capabilities which is carried out to see how Information Technology Capabilities help human resources so that they will improve organizational performance. Alshaher (2020); Zheng et al. (2020). Strategic Planning, which is carried out for organizational strategic planning, is found in several studies from (Barron Chou (2017); Esfahani et al. (2018); Dlamini et al. (2019); Al-Dhaafri and Alosani (2020). Knowledge Management also depends on the human aspect where some previous and recent studies that have become references are according to several experts including, (Inkinen (2016); Allam and Mohamed (2017); Yee et al. (2019)). Organizational Performance according to Antony & Bhattacharyya (2018); Oyewobi et al. (2019) The potential success of an organization depends to a large extent on performance, which is related to its ability to effectively implement strategies to achieve organizational goals. (Baird et al. (2019); Chen et al. (2020)). Information Technology capabilities are the main drivers that support organizational management in doing business in the digital era (Turulja and Bajgoric (2019)), while knowledge and Human Resources (HR) are increasingly considered as the main drivers of global competitive advantage in a dynamic and complex business environment (Klimkiewicz and Oltra (2017); Stathakis et al. (2017). Human resource management and knowledge management are considered as important segments of organizational management to achieve competitiveness and superior business performance. Knowledge has become the most strategically significant resource of an organization due to its heterogeneity and immobility (Gupta et al. (2017)). According to the Resource Based View theory, information technology resources can improve business performance, but only if accompanied by other resources such as (1) an effective organizational structure, (2) a productive culture, and (3) skills that support improving information technology assets. for International Journal of Research - GRANTHAALAYAH 101 The Effect of Information Technology Capability, and Strategic Planning on Knowledge Management Medited Organizational Performance in The Army Central Implementing Agency The Effect of Information Technology Capability, and Strategic Planning on Knowledge Management Medited Organizational Performance in The Army Central Implementing Agency organizational needs (Wade (2013); Barua and Mitragotri (2014); Acemoglu (2015); Mathis et al. (2017); Parrish (2018); Van Waeyenberge (2019)). organizational needs (Wade (2013); Barua and Mitragotri (2014); Acemoglu (2015); Mathis et al. (2017); Parrish (2018); Van Waeyenberge (2019)). organizational needs (Wade (2013); Barua and Mitragotri (2014); Acemoglu (2015); Mathis et al. (2017); Parrish (2018); Van Waeyenberge (2019)). Knowledge Management has several views related to knowledge, namely: (i) The first group views human capital as the owner of knowledge, (ii) The second group views knowledge without involving Human Resources, (iii) The third group discusses Human Resources Management (human capital) and Knowledge Management that a) knowledge depends on people and b) the practice of Human Resources Management is fundamental to Knowledge Management in organizations (Klimkiewicz and Oltra (2017)). Military Knowledge faces many of the same challenges as the private sector in dealing with the realities of the Information age. The main focus of the military, centered on the development of new strategies that enable it to maintain a dominant military position in spite of funding constraints (Ruel et al. (2018)). International Journal of Research - GRANTHAALAYAH 102 Eko Daryanto, Farida Jasfar, and Kusnadi Organizational Performance of the Indonesian National Armed Forces The Army wants to give birth to the benefits of its intellectual capital, so the role of human capital should be seen as a strategic resource, because only humans can create knowledge. Organizations should pay attention to the dimensions of knowledge and the differences in value creation activities related to human capital as a superior resource for the organization. Organizations should be able to form an organizational physical environment that can encourage creativity and an organizational culture that has shared values and openness. With such a focus, organizations must be able to develop more significant programs, which can encourage innovation (Anwar et al. (2016)); (Pang, and Lu (2018)); (Pham and Hoang (2019). International Journal of Research - GRANTHAALAYAH 3. METHODOLOGY International Journal of Research - GRANTHAALAYAH 103 The Effect of Information Technology Capability, and Strategic Planning on Knowledge Management Medited Organizational Performance in The Army Central Implementing Agency The Effect of Information Technology Capability, and Strategic Planning on Knowledge Management Medited Organizational Performance in The Army Central Implementing Agency The Effect of Information Technology Capability, and Strategic Planning on Knowledge Management Medited Organizational Performance in The Army Central Implementing Agency Questionnaires were distributed to respondents who are military leadership officers of the Indonesian Army at the level of strategic or operational policy makers. The number of samples in this study that was successfully obtained would be respondents, using primary data and distributing questionnaires directly. The sample used is as many as 120 people. The sampling technique is purposive sampling. From this amount, this study will take a sample of respondents from the category of leaders who can make decisions. The results of these respondents are planned to be processed using SMART PLS/SEM. Primary data collection in this study was carried out through a survey process. The survey was conducted using a questionnaire instrument. Questionnaires are the most commonly used data collection instrument in business research (Cooper and Schindler (2014)). The questionnaire made did not ask for the name of the respondent and the company where he worked to ensure that the respondent gave an objective answer. The cover letter also states that the data collected will be kept confidential and for academic purposes only, so to ensure its representation the probability sampling technique chosen is stratified random sampling. Another method that can be referred to to get the sample size is according to the Krejcie and Morgan Table. 3. METHODOLOGY The qualitative method used in this study is to raise the facts that exist in the organization, namely to see the Capability of Information Technology (Zheng et al. (2020)) and Strategic Planning (Al-Dhaafri and Alosani (2020)) as independent and Organizational Performance (Chen et al. (2020)) as the dependent variable, while Knowledge Management (Pour et al. (2019)) as the mediating variable. In this study, using a descriptive and verification approach (Hair et al. (2015); Voler Hoffman (2015)). Descriptive approach is a method in examining the status of human groups, objects, conditions and systems of thought. The purpose of descriptive research is to make a systematic, factual and accurate description, picture or painting of the facts, characteristics and relationships of the phenomena being investigated. The verification approach is a research method that aims to determine the causal relationship between variables through a hypothesis test through a statistical calculation so that evidence is obtained that shows the hypothesis is rejected or accepted (Hair et al. (2015); Voler Hoffman (2015)). The research method used in this study is a qualitative method. Research using qualitative methods, namely researchers using questionnaires in conducting data collection techniques. The qualitative method is a research method based on the philosophy of positivism, used to examine the condition of natural objects, (as opposed to experimentation) where the researcher is the key instrument, data collection techniques are carried out by triangulation (combined), data analysis is inductive/qualitative, and the results of qualitative research emphasize meaning rather than generalizations (Rahi (2017)). This study consists of Independent Variables, namely Information Technology Capability (Zheng et al. (2020)) and Strategic Planning (Al-Dhaafri and Alosani (2020)). The dependent variable is Organizational Performance (Chen et al. (2020)) and Knowledge Management variables. (Pour et al. (2019)) as a mediating variable. The following is a conceptual definition and variable indicator. The types of data used in this study are classified into two types, namely: 1) Primary data, namely data obtained directly through the distribution of questionnaires to respondents. 2) Secondary data, namely data obtained indirectly by obtaining references in the form of literature books, scientific articles, research results, and other sources. In this study, two data collection techniques were used, namely questionnaires conducted by giving a set of questions or written statements to respondents, and observations made directly to research objects, especially those concerning the variables studied. 4. RESULT AND DISCUSSION Descriptive statistics is a process that can be carried out in research data in tabulated form so that it can be easily understood and interpreted. Descriptive statistical data analysis aims to analyze data by summarizing and describing numerical data regarding gender, age, status, position. Based on the results of descriptive analysis of respondents' data based on gender who participated in the study. That there are 85 male respondents (91.4%) and 8 female respondents (8.6%). These results indicate that more men are decision makers as officers and echelon officers in the Army Headquarters environment compared to women. Respondent data by age. Table 1 Respondent Data by Age Age Frekwensi Persentase 21- 30 2 2,2 % 31- 40 8 8,6 % 41- 50 73 78,5 % 51- 60 10 10,8 % Total 93 100% Source: Processed data (2022) Table 1 Respondent Data by Age Age Frekwensi Persentase 21- 30 2 2,2 % 31- 40 8 8,6 % 41- 50 73 78,5 % 51- 60 10 10,8 % Total 93 100% Source: Processed data (2022) Table 1 shows that there are 2 people aged 21-30 with a percentage of 2.2% indicating that at this age the number of personnel who are appointed as decision- making officials is still in the process of leading a good career. Furthermore, for the age of 31-40 (8.6%) indicates the age that is starting to enter the process of increasing position towards a better and more productive level, namely 8 people. At this age the career path begins to show the organization that personnel have begun to increase. For ages 41-50 (78.5%) shows the maturity of decision makers at the Army Headquarters, Indonesian National Army Officers, Army staff and military leadership of the Indonesian Army, through the Central Implementing Body Organization of the Indonesian Army National Army. Coordinate with the International Journal of Research - GRANTHAALAYAH 104 Eko Daryanto, Farida Jasfar, and Kusnadi organizational division of the Army Information Service, the Center for Passwords and Cyber. Army Information and Data Processing Service, for personnel who are experienced in their fields as many as 73 leaders, followed by the age of 51-60 (10.8%) with 10 people, the age of starting to retire at work because of a long career in the army. Military Organizations, especially the Army Headquarters. Respondents in this study also have different levels of education. 4. RESULT AND DISCUSSION Respondent data based on education Table 2 Respondent Data Based on Education Education Frekwensi Persentase Senior High School 3 3,2% Diploma/D3 4 4,3% Bachelor/S1 36 38,7% Magister/S2 50 53,8% Doktor/S3 0 0 Total 93 100% Source: Processed data (2022) Table 2 Respondent Data Based on Education Table 2 shows the average level of education of respondents in Senior High School as much as 3 people (3.2%) indicating that decision makers who graduated from Senior High School or equivalent have started a little because they have started to understand education in organizational activities, and decision makers who They are still educated at Senior High School or equivalent because they have been loyal to work in military organizations for a long time, while continuing to pursue higher education is forgotten. There are 4 Diploma/D3 graduates (4.3%), this also shows that at the level of decision makers who pass a Diploma or equivalent, only a few people show that decision makers have started well to plan their future education. For undergraduate/S1 graduation as many as 36 people (38.7%). At the undergraduate level, many people begin to notice that education is important for making decisions and changing one's mindset. For graduation at the Masters/S2 level, 50 people (53.8%) indicate the quality of employees at the decision-making level at the Army Headquarters, both Army staff and military leadership of the Indonesian Army Army through the Central Executive Agency Organization of the Indonesian Army Army. Coordinate with the organizational division of the Army Information Service, the Center for Crypto and Cybersecurity, the Army Information and Data Processing Service, the ability to analyze and provide good service to other parties is also getting better, especially in the use of expertise based on education, especially military organizations and this level of education. demonstrate the ability of the strategic division, especially the decision makers to be able to analyze the concept and development of military organizations well. The loading factor value for each indicator from the data processing results must be more than 0.5 and the average extracted variance (AVE) value must be more than 0.5 to assess convergent validity. The results of the reliability test to see the value of Cronbach's alpha and composite reliability. The requirements for the value of Cronbach's alpha and composite reliability are 0.7. So that all variables in this study are said to be reliable. International Journal of Research - GRANTHAALAYAH International Journal of Research - GRANTHAALAYAH 5. CONCLUSION The results of this study provide a conclusion that in general, Information Technology Capability and Strategic Planning are variables that affect Organizational Performance and also Knowledge Management as a mediating variable. Likewise Knowledge Management has a good influence on Organizational Performance, as well as the Knowledge Management variable as a mediation on the Information Technology Capability variable and Strategic Planning has a strong influence on the Organizational Performance of Army Headquarters, this will have a positive impact on the performance of the Agency Implementing the Indonesian National Army Center for the Army. Coordinate with the organizational division of the Army Information Service, the Center for Crypto and Crypto, the Army Information and Data Processing Service. There needs to be a maximum change if the Indonesian National Army, especially the Army is getting better, especially in improving the quality of Human Resources as well as improving organizational infrastructure, so that in the face of change it will be more prepared and keep up with the times and increasingly advanced technology. Cooperation that must be carried out from every part of the organization of the Indonesian National Armed Forces, especially the Army, can be well established. The conclusions of the variables in this study are as follows: Information Technology Capability has a positive effect on Organizational Performance, illustrating that the organization of resources in the Central Executive Agency of the Indonesian Armed Forces. greatly affects organizational performance, affecting all officers and echelon officials in making decisions so that they are able to provide maximum policy in maintaining the level and ability that will play a role in technological capacity (Baird et al. (2019); Chen et al. (2020)) Strategic planning has a positive effect on Organizational Performance, this positive effect gives the conclusion that Strategic planning is able to show good planning for organizational performance, is the goal of the Central Implementing Agency for the Indonesian Army National Army in producing quality resources, as well as good capabilities so that the direction of quality must meet the dimensions that serve as a reference for the quality of the performance of the Indonesian Army's Central Implementing Body for the Indonesian Army, this is still the main task if the Indonesian Army's Central Implementing Body wants to continue to be a quality military organization. (Wright et al. 4. RESULT AND DISCUSSION The results of reliability testing of the variables of this study are outlined in Table 3 International Journal of Research - GRANTHAALAYAH 105 105 The Effect of Information Technology Capability, and Strategic Planning on Knowledge Management Medited Organizational Performance in The Army Central Implementing Agency Table 3 Reliability Test Results Variabel Dimensi Cronbachs Alpha Composite Reliability Information Technology Capability Knowledge Operation infrastructure 0.811 0.856 Strategic Planning Top management Facilities and Infrastructure Fund Strategic Decision 0.855 0.906 Knowledge Management. Altruistic Calling Emotional Healing Persuasive Mapping 0.88 0.909 Organizational Performance Work Productivity Responsive Responsibilty 0.815 0.871 Source: Processed data (2022) The results of testing the hypothesis of the influence of each variable are as follows: H1. Information Technology capability has a T-Statistic value of 8.776 > 1.96 and a P-Value of 0.000 < 0.050. This figure shows that Information Technology Capabilities have a positive significant effect on Organizational Performance H1. Information Technology capability has a T-Statistic value of 8.776 > 1.96 and a P-Value of 0.000 < 0.050. This figure shows that Information Technology Capabilities have a positive significant effect on Organizational Performance H2. Strategic Planning has a T-Statistic value of 5.409 > 1.96 and a P-Value of 0.000 < 0.050. This figure shows that Strategic Planning has a significant positive effect on Organizational Performance. H3. Information Technology Capability has a T-Statistic value of 16,261 > 1.96 and a P-Value of 0.000 < 0.050. This figure shows that Information Technology Capabilities have a significant positive effect on Knowledge Management. H4. Strategic Planning has a T-Statistic value of 4.043 > 1.96 and a P-Value of 0.000 < 0.050. This figure shows that Strategic Planning has a significant positive effect on Knowledge Management. H5. Knowledge Management has a T-Statistic value of 5.028 > 1.96 and a P- Value of 0.000 < 0.050. This figure shows that Knowledge Management has a positive significant effect on Organizational Performance H6. Information Technology Capability, has a T-Statistic value of 5.045 > 1.96 and a P-Value of 0.000 < 0.050. This figure shows that Information Technology Capability has a positive effect on Organizational Performance mediated by Knowledge Management H7. Strategic Planning has a T-Statistic value of 5.881 > 1.96 and a P-Value of 0.000 < 0.050. This figure shows that Strategic Planning has a positive effect on Organizational Performance mediated by Knowledge Management. International Journal of Research - GRANTHAALAYAH 106 Eko Daryanto, Farida Jasfar, and Kusnadi International Journal of Research - GRANTHAALAYAH 5. CONCLUSION (2018); Al-Dhaafri and Alosani (2020)) Information Technology Capabilities, significantly positive effect on Knowledge Management, the meaning of this positive influence is that Information Technology Capabilities received can give a positive reaction to Knowledge Management. Improving Information Technology Capability should be done by increasing the dimensions of the organization of resources in the Central Implementing Agency of the Indonesian Armed Forces. Coordinate with the organizational division of the Army Information Service, Crypto and Crypto Center, Army Information and Data Processing Service. (Baird et al. (2019); Chen et al. (2020)) Strategic planning that has a positive effect on Knowledge Management illustrates that the purpose of the human resource planning strategy is to meet the standards expected for the progress of the Central Implementing Agency for the Indonesian Army National Army so that good planning is carried out, especially in addressing all synergy policies from officials related to other resources. Antonym and Bhattacharyya (2018); Dlamini et al. (2019); Al-Dhaafri and Alosani (2020)) International Journal of Research - GRANTHAALAYAH 107 107 The Effect of Information Technology Capability, and Strategic Planning on Knowledge Management Medited Organizational Performance in The Army Central Implementing Agency The Effect of Information Technology Capability, and Strategic Planning on Knowledge Management Medited Organizational Performance in The Army Central Implementing Agency Effect of Information Technology Capability, and Strategic Planning on Knowledge Management Medited Organizational Performance in The Army Central Implementing Agency Knowledge Management has a positive effect on Organizational Performance, a significant positive effect on Organizational Performance, this illustrates that management knowledge greatly influences the performance of the Central Executing Agency for the Indonesian Armed Forces. Maximum management knowledge is the extent to which the benefits of an outcome are felt in accordance with what is expected. So that Knowledge Management is a strong variable that affects organizational performance, it can be seen from how all parties involved work together to increase the maximum value for the needs and desires of the Central Executive Board of the Indonesian Armed Forces (Allam and Mohamed (2017); Yee et al. (2019)). Information Technology Capability, on Organizational Performance mediated by Knowledge Management. This shows that the existence of Knowledge Management that mediates the Information Technology Capabilities provided by the Central Implementing Agency of the Indonesian Army National Army can increase the influence on the performance of officers and echelon officials at the Central Implementing Agency of the Indonesian Army National Army. International Journal of Research - GRANTHAALAYAH International Journal of Research - GRANTHAALAYAH 5. CONCLUSION Good organizational performance will have a maximum impact on the progress of the Indonesian Army's Central Executing Agency for the Army in each section, in maintaining Information Technology Capability, organizational readiness can create the performance of the Indonesian Army's Central Executing Agency for the Indonesian Army. (Baird et al. (2019); Chen et al. (2020)). Strategic planning, on Organizational Performance mediated by Knowledge Management. This shows that this illustrates that strategic planning has an effect on organizational performance, so with Knowledge Management as a mediation, it will further increase the positive influence of strategic planning on the performance of the Central Executive Board of the Indonesian Army National Army. Good resources will bring good quality also for an organization that maximally develops all aspects for the advancement of the Central Implementing Agency of the Indonesian Army National Army. (Al-Dhaafri and Alosani (2020)) Based on the results of the research conducted, the theoretical implications related to the development of Knowledge Management theory and Organizational Performance of Information Technology Capability and Strategic planning for the Central Implementing Body of the Indonesian Armed Forces are as follows: Information Technology Capability and Strategic Planning The results of this study strengthen the positive and significant influence of Information Technology and Strategic Planning Capabilities on the Organizational Performance of the Central Executive Agency of the Indonesian Army, in line with previous research. (Baird et al. (2019); Chen et al. (2020)). Knowledge Management as a mediating on Information Technology Capability and Strategic Planning is able to improve the performance of the Central Implementing Agency of the Indonesian Armed Forces. (Allam and Mohamed (2017); Yee et al. (2019)) The results of the study conclude that increasing Knowledge Management through increasing its dimensions will be able to improve the performance of the Central Implementing Agency for the Indonesian Army, it can increase the trust and loyalty of related parties. Improvement and development of the quality of human resources must first, pay attention to policies and strategic plans. The development will provide a good indication of the performance of the Indonesian Army's Central Executing Agency for the Indonesian National Armed Forces. Yee et al. (2019) International Journal of Research - GRANTHAALAYAH 108 Eko Daryanto, Farida Jasfar, and Kusnadi REFERENCES https://doi.org/10.1108/ijpl-10-2019-0062 International Journal of Research - GRANTHAALAYAH 109 The Effect of Information Technology Capability, and Strategic Planning on Knowledge Management Medited Organizational Performance in The Army Central Implementing Agency Esfahani, P., Mosadeghrad, A. M., & Akbarisari, A. (2018). The success of strategic planning in health care organizations of Iran. International Journal of Health Care Quality Assurance, 31(6), 563–574. https://doi.org/10.1108/IJHCQA- 08-2017-0145 Frey, C. B., & Osborne, M. A. (2017). The future of employment: How susceptible are jobs to computerisation? Technological Forecasting and Social Change. https://doi.org/10.1016/j.techfore.2016.08.019 Garcia-Lausin, L., Perez-Botella, M., Duran, X., Rodríguez-Pradera, S., Gutierrez- Martí, M. J., & Escuriet, R. (2019). Relation between Epidural Analgesia and severe perineal laceration in childbearing women in Catalonia. Midwifery. https://doi.org/10.1016/j.midw.2018.12.007 Gupta, K., Li, Q., Fan, J. J., Fong, E. L. S., Song, Z., Mo, S., Tang, H., Ng, I. C., Ng, C. W., Pawijit, P., Zhuo, S., Dong, C. Y., Low, B. C., Wee, A., Dan, Y. Y., Kanchanawong, P., So, P., Viasnoff, V., & Yu, H. (2017). Actomyosin contractility drives bile regurgitation as an early response during obstructive cholestasis. Journal of Hepatology. https://doi.org/10.1016/j.jhep.2017.01.026 Hair, J. F., Celsi, M., Money, A. H., Samouel, P., & Page, M. J. (2015). Essentials of business research methods (3rd ed.). Routledge. https://www.amazon.com/Essentials-Business-Research- Methods/dp/0765646137 Holsapple, C., Lee-Post, A., & Pakath, R. (2014). A unified foundation for business analytics. Decision Support Systems. https://doi.org/10.1016/j.dss.2014.05.013 Inkinen, H. (2016). Review of empirical research on knowledge management practices and firm performance. Journal of Knowledge Management, 20(2), 230–257. https://doi.org/10.1108/JKM-09-2015-0336 Jain, R. (2016). Jain Mathematics. Anusandhaan - Vigyaan Shodh Patrika. https://doi.org/10.22445/avsp.v4i1.4380 Kala Kamdjoug, J. R., Nguegang Tewamba, H. J., & Fosso Wamba, S. (2019). IT capabilities, firm performance and the mediating role of ISRM: A case study from a developing country. Business Process Management Journal, 25(3), 476–494. https://doi.org/10.1108/BPMJ-11-2017-0297 Klimkiewicz, K., & Oltra, V. (2017). Does CSR Enhance Employer Attractiveness? The Role of Millennial Job Seekers’ Attitudes. Corporate Social Responsibility and Environmental Management. https://doi.org/10.1002/csr.1419 Mathis, M. W., Mathis, A., & Uchida, N. (2017). Somatosensory Cortex Plays an Essential Role in Forelimb Motor Adaptation in Mice. Neuron. https://doi.org/10.1016/j.neuron.2017.02.049 Nowak, R. (2020). “Process of strategic planning and cognitive diversity as determinants of cohesiveness and performance ” Business Process Esfahani, P., Mosadeghrad, A. M., & Akbarisari, A. (2018). The success of strategic planning in health care organizations of Iran. International Journal of Health Care Quality Assurance, 31(6), 563–574. https://doi.org/10.1108/IJHCQA- 08-2017-0145 Frey, C. B., & Osborne, M. A. (2017). International Journal of Research - GRANTHAALAYAH REFERENCES Acemoglu, D. (2015). Why nations fail? In Pakistan Development Review. Econ papers Vol. 54, issue 4, 301-312 https://econpapers.repec.org/article/pidjournl/v_3a54_3ay_3a2015_3ai_3 a4_3ap_3a301-312.htm Ahmed, A., Ramachandram, S., & Rahman Khan, K. U. (2018). Privacy Preserving Dynamically Indexed Multi-Phrase Search over Encrypted Data. 2018 International Conference on Advances in Computing, Communications and Informatics, ICACCI 2018. https://doi.org/10.1109/ICACCI.2018.8554618 Al-Dhaafri, H., & Alosani, M. (2020). Closing the strategic planning and implementation gap through excellence in the public sector: empirical investigation using SEM. Measuring Business Excellence, 24(4), 553–573. https://doi.org/10.1108/MBE-12-2019-0128 Allam, A., & Mohamed, E. (2017). SMART KM model: The integrated knowledge management framework for organisational excellence. World Journal of Science, Technology and Sustainable Development, 14(2/3), 172–193. https://doi.org/doi:10.1108/WJSTSD-01-2017-0001 Alshaher, A. (2020). IT capabilities as a fundamental of electronic government system success in developing countries from users perspectives. Transforming Government: People, Process and Policy. https://doi.org/10.1108/TG-05-2020-0080 Antony, J. P., & Bhattacharyya, S. (2018). Measuring organizational performance and organizational excellence of SMEs – Part 1: A conceptual framework. Measuring Business Excellence, 14(2), 3–11. https://doi.org/10.1108/13683041011047812 Anwar, M., Law, R., & Schier, J. (2016). Kratom (Mitragyna speciosa) exposures reported to poison centers — United States, 2010-2015. In Morbidity and Mortality Weekly Report. https://doi.org/10.15585/mmwr.mm6529a4 y y p p // g/ / Baird, K., Su, S., & Munir, R. (2019). Levers of control, management innovation and organisational performance. Pacific Accounting Review, 31(3), 358–375. https://doi.org/10.1108/PAR-03-2018-0027 p // g/ / Barron, K., & Chou, S. Y. (2017). Toward a spirituality mode of firm sustainability strategic planning processes. Society and Business Review, 12(1), 46–62. https://doi.org/10.1108/sbr-01-2016-0008 Barua, S., & Mitragotri, S. (2014). Challenges associated with penetration of nanoparticles across cell and tissue barriers: A review of current status and future prospects. In Nano Today. https://doi.org/10.1016/j.nantod.2014.04.008 Bolland, E. J., & Bolland, E. J. (2020). Evolution of Strategy: Origin, Planning, Strategic Planning and Strategic Management. Strategizing, 25–48. https://doi.org/10.1108/978-1-78973-697-720201003 Chen, M. Y. C., Lam, L. W., & Zhu, J. N. Y. (2020). Should companies invest in human resource development practices? The role of intellectual capital and organizational performance improvements. Personnel Review. https://doi.org/10.1108/PR-04-2019-0179 p // g/ / Cooper, D. R., & Schindler, P. S. (2014). Business Research Methods - Donald R. Cooper. In McGraw-Hill. Dlamini, N., Mazenda, A., Masiya, T., & Nhede, N. T. (2019). Challenges to strategic planning in public institutions. International Journal of Public Leadership, 16(1), 109–124. REFERENCES The future of employment: How susceptible are jobs to computerisation? Technological Forecasting and Social Change. https://doi.org/10.1016/j.techfore.2016.08.019 Garcia-Lausin, L., Perez-Botella, M., Duran, X., Rodríguez-Pradera, S., Gutierrez- Martí, M. J., & Escuriet, R. (2019). Relation between Epidural Analgesia and severe perineal laceration in childbearing women in Catalonia. Midwifery. https://doi.org/10.1016/j.midw.2018.12.007 Gupta, K., Li, Q., Fan, J. J., Fong, E. L. S., Song, Z., Mo, S., Tang, H., Ng, I. C., Ng, C. W., Pawijit, P., Zhuo, S., Dong, C. Y., Low, B. C., Wee, A., Dan, Y. Y., Kanchanawong, P., So, P., Viasnoff, V., & Yu, H. (2017). Actomyosin contractility drives bile regurgitation as an early response during obstructive cholestasis. Journal of Hepatology. https://doi.org/10.1016/j.jhep.2017.01.026 Hair, J. F., Celsi, M., Money, A. H., Samouel, P., & Page, M. J. (2015). Essentials of business research methods (3rd ed.). Routledge. https://www.amazon.com/Essentials-Business-Research- Methods/dp/0765646137 Holsapple, C., Lee-Post, A., & Pakath, R. (2014). A unified foundation for business analytics. Decision Support Systems. https://doi.org/10.1016/j.dss.2014.05.013 Inkinen, H. (2016). Review of empirical research on knowledge management practices and firm performance. Journal of Knowledge Management, 20(2), 230–257. https://doi.org/10.1108/JKM-09-2015-0336 Jain, R. (2016). Jain Mathematics. Anusandhaan - Vigyaan Shodh Patrika. https://doi.org/10.22445/avsp.v4i1.4380 Kala Kamdjoug, J. R., Nguegang Tewamba, H. J., & Fosso Wamba, S. (2019). IT capabilities, firm performance and the mediating role of ISRM: A case study from a developing country. Business Process Management Journal, 25(3), 476–494. https://doi.org/10.1108/BPMJ-11-2017-0297 g Klimkiewicz, K., & Oltra, V. (2017). Does CSR Enhance Employer Attractiveness? The Role of Millennial Job Seekers’ Attitudes. Corporate Social Responsibility and Environmental Management. https://doi.org/10.1002/csr.1419 Mathis, M. W., Mathis, A., & Uchida, N. (2017). Somatosensory Cortex Plays an Essential Role in Forelimb Motor Adaptation in Mice. Neuron. https://doi.org/10.1016/j.neuron.2017.02.049 Nowak, R. (2020). “Process of strategic planning and cognitive diversity as determinants of cohesiveness and performance.” Business Process Management Journal, 27(1), 55–74. https://doi.org/10.1108/BPMJ-09- 2019-0401 Oyewobi, L. O., Oke, A. E., Adeneye, T. D., & Jimoh, R. A. (2019). Influence of organizational commitment on work–life balance and organizational performance of female construction professionals. Engineering, Construction and Architectural Management, 26(10), 2243–2263. https://doi.org/10.1108/ECAM-07-2018-0277 Pang, K., & Lu, C. S. (2018). Organizational motivation, employee job satisfaction and organizational performance: An empirical study of container shipping companies in Taiwan. Maritime Business Review, 3(1), 36–52. https://doi.org/10.1108/MABR-03-2018-0007 International Journal of Research - GRANTHAALAYAH 110 Eko Daryanto, Farida Jasfar, and Kusnadi Parrish, T. (2018). Postmodernism. In Herman Melville in Context. https://doi.org/10.1017/9781316755204.032 Parrish, T. (2018). Postmodernism. In Herman Melville in Context https://doi.org/10.1017/9781316755204.032 Pham, L. International Journal of Research - GRANTHAALAYAH REFERENCES T., & Hoang, H. V. (2019). The relationship between organizational learning capability and business performance. Journal of Economics and Development, 21(2), 259–269. https://doi.org/10.1108/jed-10-2019-0041 Pour, M. J., Matin, H. Z., Yazdani, H. R., & Zadeh, Z. K. (2019). A comprehensive investigation of the critical factors influencing knowledge management strategic alignment. Knowledge Management and E-Learning. https://doi.org/10.34105/j.kmel.2019.11.011 Prof. Dr. Voler Hoffman. (2015). Strategic Management. 0–55. Purdue. (2016). The Lewis Definition of Acids and Bases. Bodner Research Web Rahi, S (2017). Research design and methods: A systematic review of research paradigms, sampling issues and instruments development. International Journal of Economics & Management Sciences, 6(2), 1–5. https://www.researchgate.net/publication/316701205_Research_Design_ and_Methods_A_Systematic_Review_of_Research_Paradigms_Sampling_Issu es_and_Instruments_Development Roth, B. L. (2016). DREADDs for Neuroscientists. In Neuron https://doi.org/10.1016/j.neuron.2016.01.040 Ruel, E., Wagner, W. E., & Gillespie, B. J. (2018). Introduction to Survey Research. In The Practice of Survey Research: Theory and Applications. https://doi.org/10.4135/9781483391700.n1 Stathakis, E., Brachos, K., Abatzianis, C., Bandekas, D. V., & Mouroutsos, S. G. (2017). A smart approach to measuring the performance-efficiency, effectiveness and productivity - of the general university hospital of alexandroupolis- G.U.H.A. Using special indicators. Journal of Engineering Science and Technology Review. https://doi.org/10.25103/jestr.102.17 Tarifa-Fernández, J., de-Burgos-Jimenez, J., & Cespedes-Lorente, J. (2019). Absorptive capacity as a confounder of the process of supply chain integration. Business Process Management Journal, 25(7), 1587–1611. https://doi.org/10.1108/BPMJ-12-2017-0340 Turulja, L., & Bajgoric, N. (2019). Innovation, firms’ performance and environmental turbulence: is there a moderator or mediator? European Journal of Innovation Management. https://doi.org/10.1108/EJIM-03-2018-0064 Van Waeyenberge, B. (2019). Bridging the gap. In Nature Materials. https://doi.org/10.1038/s41563-019-0411-7 Wade, W. G. (2013). The oral microbiome in health and disease. In Pharmacological Research. https://doi.org/10.1016/j.phrs.2012.11.006 Wright, P. M., Nyberg, A. J., & Ployhart, R. E. (2018). A research revolution in SHRM: New challenges and new research directions. Research in Personnel and Human Resources Management, 36, 141–161. https://doi.org/10.1108/S0742-730120180000036004 Yee, Y. M., Tan, C. L., & Thurasamy, R. (2019). Back to basics: building a knowledge management system. Strategic Direction, 35(2), 1–3. https://doi.org/10.1108/SD-07-2018-0163 Zheng, J., Liu, H., & Zhou, J. (2020). High-performance work systems and open innovation: moderating role of IT capability. Industrial Management and Data Systems, 120(8), 1441–1457. https://doi.org/10.1108/IMDS-09- 2019-0475 International Journal of Research - GRANTHAALAYAH 111 111
https://openalex.org/W2064514339	https://e-journals.ku.lt/journal/TBB/article/1920/file/pdf	Lithuanian	null	TARPREGIONINIO EUROPOS SĄJUNGOS IR LOTYNŲ AMERIKOS BENDRADARBIAVIMO DINAMIKA	Tiltai	2,021	cc-by	5,651	TARPREGIONINIO EUROPOS SĄJUNGOS IR LOTYNŲ AMERIKOS BENDRADARBIAVIMO DINAMIKA ISSN 1392-3137. TILTAI, 2014, 2 Mindaugas Norkevičius Vytauto Didžiojo universitetas DOI: http://dx.doi.org/10.15181/tbb.v67i2.845 DOI: http://dx.doi.org/10.15181/tbb.v67i2.845 Anotacija Straipsnyje analizuojamas tarpregioninis Lotynų Amerikos ir Europos Sąjungos (ES) bendradarbia- vimas, siekiant atskleisti regioninės integracijos procesus tarptautiniu mastu. Aptariamos pagrindinės ES ir Lotynų Amerikos tarpregioninio bendradarbiavimo priežastys, apimančios vyraujančią sociali- nę nelygybę ir skurdą, demokratijos deficitą, augantį Lotynų Amerikos ekonominį potencialą ir atsi- veriančias naujas rinkas. ES skatinama ekonominė, socialinė ir institucinė integracija prisideda prie regionalizacijos procesų, kurie vyksta Lotynų Amerikoje. Tarpregioninis bendradarbiavimas papras- tai vyksta per tarptautines organizacijas – Rio grupę, MERCOSUR, UNASUR, CAN. Straipsnyje na- grinėjami pagrindiniai tarpregioninio bendradarbiavimo tikslai, regioninio bendradarbiavimo sritys ir santykiai tarp ES bei Lotynų Amerikos regioniniame, subregioniniame ir dvišaliame lygmenyje, kurie tarpusavyje persipynę veiklos sričių apibrėžtumu. PAGRINDINIAI ŽODŽIAI: regionalizmas, tarpregioninis bendradarbiavimas, subregionalizmas, Lotynų Amerikos regionas, Europos Sąjungos vystymo politika, tarptautinės organizacijos. Abstract This article analyses inter-regional colloboration between Latin America and European Union (EU) in order to find out the processes of regional integration to international extent. Basic reasons of inter-regional collaboration between EU and Latin America which are related with social inequality and destitution, shortfall of democracy and growth of Latin America’s economical potential as well as appearance of various new markets are excluded in this text. EU encourages economical, social and institutional integration which is associated with regionalization processes in the area of Latin America. Inter-regional collaboration happens through international organizations such as Rio group, MERCOSUR, UNASUR, CAN. This article deals with general aims of inter-regional collaboration, various branches of it and relations between EU and Latin America in regional, sub-regional and bilateral level which are intermingled among themselves. KEYWORDS: regionalism, interregional cooperation, Latin America, European union’s develo- pment policy, international organizations. TARPREGIONINIO EUROPOS SĄJUNGOS IR LOTYNŲ AMERIKOS... TARPREGIONINIO EUROPOS SĄJUNGOS IR LOTYNŲ AMERIKOS... Siekiant atlikti ES ir Lotynų Amerikos bendradarbiavimo analizę straipsnyje remiamasi K. Arts, A. Dickson „EU development cooperation: from model to sym- bol“ (2004), A. Gordillo „The future of Latin America: can the EU help?“ (2003), J. A. Sanahuja „América Latina: las visiones y políticas de Europa“ (2006) darbais, kuriuose aptariami teoriniai ir tiriamieji ES bei Lotynų Amerikos tarpregioninio bendradarbiavimo pavyzdžiai. ES institucijų ir Lotynų Amerikos tarptautinių or- ganizacijų interneto svetainių duomenys padėjo įvertinti bendradarbiavimo mastą, ES institucijų vaidmenį vykdant plėtros politiką Lotynų Amerikos regione. Pirmoje straipsnio dalyje teoriškai apibrėžiamos regioninės politikos ir tar- pregioninių santykių tendencijos bei funkcijos. Antroje dalyje identifi kuojama ES vykdoma politika Lotynų Amerikos regione, aptariamos tarpregioninio bendradar- biavimo kryptys ir strategijos, įvertinamas ES institucijų vaidmuo. Trečioje dalyje analizuojamas ES ir Lotynų Amerikos bendradarbiavimas, analizė atliekama apta- riant ES santykius su tam tikromis regioninėmis, subregioninėmis organizacijomis ir pavienėmis Lotynų Amerikos valstybėmis. Įvadas Tarpegioniniai santykiai tarp ES ir Lotynų Amerikos užmegzti dar šeštajame XX amžiaus dešimtmetyje, tai gali būti laikoma tarpregioninio bendradarbiavi- mo pradžia. Lotynų Amerikos ir Europos žemynus saisto įvairūs politiniai, isto- riniai, kultūriniai ir religiniai ryšiai (Keen, Haynes, 2009, p. 7). Tarpregioninis bendradarbiavimas suintensyvėja XX amžiaus devintajame dešimtmetyje. Iki tol 35 Mindaugas Norkevičius ES domėjimasis Lotynų Amerikos regionu labiau apsiribojo dvišaliais santykiais su pavienėmis šio regiono valstybėmis. XXI amžiuje ES ir Lotynų Amerikos santykiai užsienio politikos srityje labai glaudūs. Pagrindiniai ES vykdomi užsienio politikos tikslai Lotynų Amerikoje yra regioninės integracijos, prekybinių ir ekonominių santykių skatinimas, socia- linės gerovės idėjų skaida regione, politinio bendradarbiavimo ryšių stiprinimas. Portugalijai ir Ispanijai tapus ES narėmis, santykiai su Lotynų Amerika dar labiau suintensyvėjo. Šias ES valstybes ir Lotynų Amerikos regioną glaudžiai sieja isto- riniai, kultūriniai, religiniai ir politiniai veiksniai. ES subregioniniame lygmenyje bendradarbiauja su MERCOSUR valstybėmis (isp. Mercado Común del Sur), Cen- trinės Amerikos organizacijos šalimis (isp. Unión de Naciones Suramericanas – UNASUR), Andų bendrija (isp. Comunidad Andina – CAN). Regioniniame ly- gmenyje santykiai plėtojami su Rio grupės valstybėmis. Aktyviausias tarpregioni- nis Lotynų Amerikos ir ES bendradarbiavimas XXI amžiuje pastebimas su pavie- nėmis valstybėmis – Meksika, Čile ir Brazilija. Tarpregioniniame lygmenyje ES yra antroji Lotynų Amerikos regiono prekybos ryšių partnerė, ją lenkia tik JAV. ES yra didžiausia investuotoja Lotynų Amerikos regione. ES intensyviai skatina re- gioninės integracijos procesus, remdama ekonominio bendradarbiavimo ir plėtros santykius prekybos srityje, demokratizacijos procesus Lotynų Amerikoje. Tyrimo problema susijusi su augančiu Lotynų Amerikos regiono vaidmeniu pa- saulio politinėje ir ekonominėje erdvėje, tai verčia ES aktyviau puoselėti tarpregio- ninius santykius. Lotynų Amerika laikoma svarbiu ekonominiu ir politiniu partneriu. Tyrimo tikslas – išanalizuoti ES ir Lotynų Amerikos tarpregioninio bendra- darbiavimo dinamiką. Tyrimo tikslas – išanalizuoti ES ir Lotynų Amerikos tarpregioninio bendra- darbiavimo dinamiką. Tyrimo tikslui pasiekti išsikelti uždaviniai: • nustatyti pagrindines ES ir Lotynų Amerikos tarpregioninio bendradarbia- vimo priežastis; • išanalizuoti ES institucijų vaidmenį kuriant užsienio politiką Lotynų Ame- rikoje; • išnagrinėti ES vykdomą bendradarbiavimą su regioninėmis grupėmis ir pavienėmis Lotynų Amerikos valstybėmis; • įvertinti Europos Sąjungos Lotynų Amerikos regione vykdomą plėtros po- litiką. Straipsnyje taikoma keletas mokslinių tyrimo metodų: taikant kokybinę tu- rinio analizę analizuoti ES ir Lotynų Amerikos tarpregioninį bendradarbiavimą apibrėžiantys dokumentai; taikant aprašomąją interpretacinę problemos analizę siekiama pateikti informaciją apie ES ir Lotynų Amerikos tarpregioninį bendra- darbiavimą teoriniame lygmenyje. 36 1. Regioninės politikos ir tarpregioninių santykių tendencijos Vykdoma regioninė politika ir tarpregioniniai santykiai prisideda prie regiona- lizacijos procesų, todėl svarbu tinkamai apibrėžti jų sampratas ir tendencijas. Es- minius regioninės politikos bruožus galima analizuoti aptariant regioninę politiką tarptautiniame lygmenyje. Tarptautinių santykių kontekste regioninės politikos ir regiono sampratos yra panašios. Šiuolaikiniame pasaulyje regioninė politika pri- sideda prie regioninės integracijos procesų nacionalinių valstybių, tarpvalstybinių santykių skatinimo ir tarpregioninių santykių plėtojimo vaidmens (žr. 1 pav.). 37 1 pav. Regioninės politikos vaidmuo Pastaba: sudaryta autoriaus. 37 37 Mindaugas Norkevičius Regioninės politikos teoretikas B. Goodal tai apibūdina kaip nacionalinių vy- riausybių politiką, kuri veikia pagrindinių ekonominių išteklių paskirstymą ben- droje valstybių erdvėje ar už jos ribų. Regioninės politikos objektas yra regionas (Goodal, 1987, p. 235). Tokios politikos samprata neatsiejama nuo ekonominių veiksnių. Tai apibūdina M. Skinerio pateiktas apibrėžimas, kuriame autorius at- skleidžia nacionalinių vyriausybių siekį likviduoti tarp regionų esančius ekono- minius ir socialinius skirtumus, pritraukiant investicijas į menkiau išsivysčiusias teritorijas (Skiner, 1999, p. 197). Regioninė politika skatina spręsti netolygaus ekonominio išsivystymo problemas valstybės regionuose ar subnacionaliniuose teritoriniuose vienetuose. Tarpregioninių santykių tendencijos išryškėjo XX a. 9-ojo dešimtmečio pabai- goje. Teoretikai skirtingai apibrėžia tarpregioninius santykius. Remiantis B. Het- tne nuomone, procesas įvardijamas kaip būdas pertvarkyti esamą pasaulį (Hwee, Vidal, 2008, p. 45). J. Gilson tarpregioniniams santykiams apibrėžti vartoja kons- truktyvizmo teorinę paradigmą, procesus suvokiant kaip vykstančias socialines sąveikas, paremtas regioniniais veiksniais (Gilson, 2005, p. 309). Pasak R. Ro- loff, tarpregioniniai santykiai vysto politinius, ekonominius ir socialinius procesus tarptautinėje arenoje (Hanggi, Roloff, Ruland, 2006, p. 18). Vis dėlto lieka neaiš- kios tarpregioninių santykių klasifikacijos ir pagrindinės funkcijos. Galima daryti prielaidą, kad tarpregioniniai santykiai suvokiami įvairiapusiškai, apima pasaulio tvarkos pakeitimo ir integracinių procesų sklaidą regionuose (plačiau apie tarpre- gioninių santykių funkcijas žr. 2 pav.). Regionalizmas sudaro tarpregioninių santykių analizės pagrindą. Procesui bū- dingas daugiafunkcis kompleksiškumas, apimantis politiką, ekonomiką, saugumą ir kultūrą. Siekiant sistemingai išanalizuoti tarpregioninius santykius, reikia tin- kamai apsibrėžti vykstančiam procesui įtakos turinčius subjektus. Tarpregioninius santykius galima nagrinėti skirtingais lygmenimis – regioniniu, subregioniniu ir dvišalių santykių kontekste. 38 TARPREGIONINIO EUROPOS SĄJUNGOS IR LOTYNŲ AMERIKOS... 2 pav. Tarpregioninių santykių funkcijos Pastaba: sudaryta straipsnio autoriaus, remiantis M. Doidge (2007, p. 18). 2 pav. Tarpregioninių santykių funkcijos Pastaba: sudaryta straipsnio autoriaus, remiantis M. Doidge (2007, p. 18). 2. Europos Sąjungos politika Lotynų Amerikoje 1999 m. prasidėjo aukščiausio lygio susitikimai, organizuojami ES ir Lotynų Amerikos regiono. Pagrindinės problemos, lėmusios tarpregioninį bendradarbia- vimą, – demokratijos plėtra, skurdo ir socialinės atskirties mažinimas, regioninės politikos puoselėjimas, aplinkos apsaugos problemos (Arts, Dickson, 2004, p. 74). Lotynų Amerika ir ES sudarė nemažai susitarimų, kuriuose dominuoja ekonomi- nio bendradarbiavimo perspektyvos, politinio dialogo skatinimas, tarpregioninių prekybos santykių plėtra. 39 39 Mindaugas Norkevičius Mindaugas Norkevičius ES intensyviausiai bendradarbiauja su dviem Lotynų Amerikos subregio- nais – Pietų ir Centrine Amerika. Dauguma Lotynų Amerikos regiono valsty- bių – buvusios kolonijinės santvarkos šalys, pradėjusios kurti subregionines or- ganizacijas. Tarpregioninį bendradarbiavimą galima skirti į tarpregioninius santy- kius ir politinį bendradarbiavimą subregioniniu lygiu. Pagrindiniai ES ir Lotynų Amerikos tarpregioninių santykių tikslai, kuriuos siekiama įgyvendinti visuose santykių lygmenyse, yra politinių ryšių stiprinimas, prekybos santykių plėtojimas, socialinės gerovės ir demokratijos rėmimas Lotynų Amerikos šalyse bei regioni- nės integracijos sklaida. Sudaryti ES ir Centrinės Amerikos bendradarbiavimo ir politinio dialogo plėto- jimo susitarimai leidžia 2003 m. laikyti tarpregioninio bendradarbiavo atsinaujini- mo laikotarpiu. Tais pačiais metais pasirašytas tarpregioninio bendradarbiavimo su- sitarimas tarp ES ir Andų bendrijos valstybių (sutartis kol kas neįsigaliojusi). Svar- biausios tarpregioninio bendradarbiavimo veiklos sritys apibrėžiamos ES ir Lotynų Amerikos atstovų susitikimuose: ES valstybių ir Lotynų Amerikos bei Karibų jūros (LAKR) šalių vyriausybių vadovų aukščiausio lygio susitikime, Rio grupės ir ES valstybių ministrų susitikimuose, Europos Sąjungos Trejeto ir Centrinės Amerikos šalių ministrų konferencijoje (San Chosė dialogas) (Gordillo, 2003, p. 141). Vieni svarbiausių – ES valstybių ir Lotynų Amerikos bei Karibų jūros (LAKR) šalių vyriausybių vadovų aukščiausio lygio susitikimai, vykstantys kas dveji metai (ES ryšių su Lotynų Amerika strategija, 2010). Susitikimų tikslas – stiprinti tar- pregioninius politinius, ekonominius ir kultūrinius ryšius. Kiti svarbūs susitikimai vyksta tarp Rio grupės (apima visas Lotynų Amerikos ir Karibų jūros regiono ša- lis) ir ES valstybių ministrų. Ši grupė siekia plėtoti ES ir Lotynų Amerikos regiono šalių bendradarbiavimą, užtikrinti demokratijos skaidą, socialinės politikos, kuri ypač svarbi siekiant spręsti skurdo ir socialinės atskirties problemas, įgyvendinimą (Sanahuja, 2006, p. 2). Puoselėti tolygius tarpregioninius santykius, kurie apima prekybos ryšius ir politinį dialogą. Aktyviai bendradarbiaujama su šalimis, kurios Lotynų Amerikos regione turė- jo kolonijų, tai Ispanija, Portugalija, Prancūzija. Naujosios ES šalys, ypač Lenkija ir Čekija, palaiko artimus prekybos ryšius su Lotynų Amerikos regionu, šalių vers- lininkai regione valdo stambias pramonės srities įmones. Tarpregioninį bendradar- biavimą skatina ES finansinė parama Lotynų Amerikai: infrastruktūrai, energeti- kai, kovai su skurdu ir socialine atskirtimi bei aplinkosaugos problemoms spręsti. 2.1. Tarpregioninio bendradarbiavimo kryptys ir strategijos Galima skirti pagrindines tarpregioninio bendradarbiavimo sritis, kurios sufor- muluotos bendradarbiavimo strategijoje. 2007–2013 metais regionai bendradar- 40 TARPREGIONINIO EUROPOS SĄJUNGOS IR LOTYNŲ AMERIKOS... biavo, remdamiesi regionine ES ir Lotynų Amerikos bendradarbiavimo strategija, kurios finansinės paramos sritis – skurdo mažinimas ir pastangos spręsti regione kylančias problemas, vykdant įvairias programas. Naujoje strategijoje pagrindi- nės bendradarbiavimo veiklos sritys išlieka nepakitusios. Siekiant geriau įvertinti tarpregioninio bendradarbiavimo prioritetus, reikia skirti vykdomas ir vystomas programas, kurios skirtos įgyvendinti ekonominį ir politinį bendradarbiavimą bei užtikrinti socialinę gerovę. Sukurta keletas viešojo administravimo ir socialinių reikmių programų, tokių kaip URB-AL, kuri skirta plėtoti decentralizuotą bendradarbiavimą tarp vietos val- džios institucijų. EUrosociAL programa skirta plėtoti socialinės sanglaudos politi- ką ir mažinti skurdą Lotynų Amerikos regione. EUrocLIMA programa skirta įgy- vendinti aplinkos apsaugos idėjas ir spręsti su tuo susijusias problemas. Siekiant skatinti tarpregioninį ekonominį bendradarbiavimą sukurta AL-Invest programa, kuri orientuota į tarpregioninį įmonių bendradarbiavimą. ATLAS programa skir- ta prekybos rūmų bendradarbiavimui, @LIS programos funkcija – skatinti infor- macinių technologijų taikymą. ALURE programa skirta plėtoti bendradarbiavimą energetikos sektoriuje tarpregioniniame lygmenyje (European Commission Deve- lopment and cooperation – EuropeAid, 2011). ES ir Lotynų Amerikos regionai švietimo žmogiškųjų išteklių srityje bendra- darbiauja vykdydami ALFA universitetų bendradarbiavimo programą. ALBAN programa skirta sudaryti aukštosios pakopos studijų stipendijų skyrimo ir moks- linių tyrimų atlikimo galimybes. Erasmus Mundus programa orientuota į platesnį studentų mainų mastą regionų valstybėse. ES vystomosios bendradarbiavimo pa- galbos teikimo Lotynų Amerikai pagrindiniai tikslai – įveikti socialinę nelygybę ir skurdą, stiprinti teisinę ir demokratinę valstybę, plėtoti ekonominius santykius (Seidelmann, 2005, p. 35). Tokio pobūdžio politika veikia tarpregioninių ir dviša- lių susitarimų dėka ir apima kultūros, politikos bei prekybos sritis. Lotynų Amerika dalyvauja vykdant ir kitas ES politikos programas, kurios su- sijusios su mokslinių tyrimų, technologijų ir inovacijų, migracijos ir prieglobsčio politika, partneryste su NVO sektoriumi, žmogaus teisių įgyvendinimu, pilietinės visuomenės skatinimu. ES ir Lotynų Amerikos regioninis bendradarbiavimas re- miasi įvairaus pobūdžio programomis, kurios skatina abipusį bendradarbiavimą. Didėjantis bendradarbiavimo su Lotynų Amerika ekonomikoje ir politikoje porei- kis skatina investicinių bendradarbiavimo projektų plėtrą. 2. ES institucijų veikla Lotynų Amerikos regione: ES Parlamento vaidmu ES institucijos vykdo aktyvų politinį dialogą su Lotynų Amerikos regionų par- tneriais. Regioninių santykių plėtra labiausiai pastebima po 2004 metais vykusių 41 Mindaugas Norkevičius Europos Parlamento (EP) rinkimų. EP, atsižvelgdamas į susitarimus su Lotynų Amerikos regionu, subūrė delegacijas tarpregioniniams ryšiams su MERCOSUR valstybėmis, Centrinės Amerikos šalių organizacija, Andų bendrija, Meksika ir Čile stiprinti. EP vaidmuo svarbus palaikant glaudžius ryšius su Lotynų Amerikos Parla- mentu (Parlatino), MERCOSUR jungtiniu parlamentiniu komitetu (CPM), Cen- trinės Amerikos Parlamentu (Parlacen), Andų bendrijos Parlamentu (Parlandino), Meksikos ir Čilės valstybių kongresais (The European Parliament, 2013). Lotynų Amerikos parlamentinės institucijos kartu su EP organizuoja konferencijas, kurios padeda užmegzti tarpregioninio bendradarbiavimo dialogą. Reikšmingu EP žings- niu galima laikyti 2006 metų rezoliucijoje iškeltą tikslą – sukurti bendrą ES ir Lo- tynų Amerikos bei Karibų jūros regiono šalių strategiją (Sanahuja, 2006, p. 10). EP pabrėžia žmogaus teisių užtikrinimą, demokratijos principų taikymą, teisinės vals- tybės kūrimą Lotynų Amerikos regione, nusako pagrindinius bendradarbiavimo tikslus, išskiriant politikos, ekonomikos (prekybos) ir socialinės apsaugos sritis. Kalbant apie bendradarbiavimą politikos srityje galima paminėti ES ir Lotynų Amerikos transatlantinės asamblėjos (EUROLAT) sukūrimą. Jos tikslas – politinio (parlamentinio) dialogo puoselėjimas plėtojant tarpregioninę politiką, bendradar- biavimą saugumo ir gynybos srityse, siekis įsteigti regioninių konfliktų preven- cijos centrą. Todėl siekiama pasirašyti ES ir Lotynų Amerikos taikos ir saugumo chartiją, kurioje būtų apibrėžti pagrindiniai aspektai, svarbūs puoselėjant tarpre- gioninius santykius saugumo srityje (Polski Instytut Spraw Międzynarodowych, 2010). Kitas EP siekis – sukurti ES ir Lotynų Amerikos laisvosios prekybos zoną, vykdyti susitarimus su partneriais – MERCOSUR valstybių, Andų bendrijos bei Centrinės Amerikos šalių. Regioninio solidarumo fondo įkūrimas, kurio pagrin- dinis tikslas būtų skurdo ir socialinės atskirties mažinimas, rodo EP rūpinimąsi problemomis, kurios susijusios su socialine gerove Lotynų Amerikos regione. ES vystomojo bendradarbiavimo pagalbos teikimas paremtas kova su socialine nely- gybe ir skurdu, teisinės bei demokratinės valstybės stiprinimu, ekonominių santy- kių plėtra (Seidelmann, 2005, p. 35). 2.3. EUROLAT tikslai ir funkcijos ES ir Lotynų Amerikos šalių parlamentinė asamblėja EUROLAT įkurta 2006 metais. Asamblėja įsteigta parlamentinės institucijos pagrindu ir apima strategi- nę ES ir Lotynų Amerikos regionų partnerystę (The European Parliament, 2013). ES ir Lotynų Amerikos šalių parlamentinės asamblėjos tikslas – pateikti ir pri- imti rekomendacijas, rezoliucijas institucijoms bei organizacijoms, atsakingoms 42 TARPREGIONINIO EUROPOS SĄJUNGOS IR LOTYNŲ AMERIKOS... už tarpregioninę ES ir Lotynų Amerikos plėtrą. Jungtinę daugiašalę parlamenti- nę asamblėją sudaro 150 narių: 75 atstovauja EP, 75 – Lotynų Amerikos šalims. Asamblėjoje dalyvauja Lotynų Amerikos Parlamento, Andų bendrijos Parlamento, Centrinės Amerikos organizacijos Parlamento ir MERCOSUR Parlamento atsto- vai (Euro-Latin American Parliamentary Assembly rezoliution, 2011). Atstovau- jama ir Meksikos bei Čilės kongresams, nes sudaryti ES ir Meksiko, ES ir Čilės jungtiniai parlamentiniai komitetai. EUROLAT nariai skiriami vadovaujantis kie- kvieno parlamento vidaus taisyklėmis, siekiant atskleisti politinių frakcijų įvairovę ir apimti šalis, kurioms atstovauja kiekviename šių parlamentų ar kongresų. Parlamentinės asamblėjos struktūrą sudaro du pirmininkai: vienas atstovauja Lotynų Amerikos regionui, kitas – ES. Asamblėjos institucinę struktūrą sudaro: vykdomasis biuras, nuolatiniai komitetai, darbo grupės, sekretoriatas. Vykdomą- jį komitetą sudaro pirmininkai iš kiekvieno regiono, turintys po 7 pavaduotojus. Komitetas koordinuoja parlamentinės asamblėjos darbą, palaiko ryšius su ES ir LAKR aukščiausio lygio vadovais, Rio grupe, vykstant San Chosė konferenci- joms. Pagrindines funkcijas atskleidžia asamblėją sudarantys komitetai, kurie at- sakingi už: 1) politikos, saugumo ir žmogaus teisių sritis; 2) prekybą ir ekonominį bendradarbiavimą; 3) švietimo, kultūros, socialinę ir aplinkos apsaugą (Stravridis, Ajenjo, 2010). Komitetuose priimami rekomendacijų ir rezoliucijų projektai, kurie svarstomi parlamentinėje asamblėjoje. EUROLAT išleistos rezoliucijos apima įvairius aspektus: 2009 metais išleistos rezoliucijos aptarė finansų krizės, prekybos ir klimato kaitos priežastis. 2008 me- tais didžiausias dėmesys skirtas energetikos politikos analizei, skurdo ir socialinės atskirties problemoms. Pagrindinis ES ir Lotynų Amerikos šalių parlamentinės asamblėjos dėmesys 2011 metų gegužės mėnesį išleistoje rezoliucijoje krypsta į tarpregioninių santykių aptarimą saugumo ir gynybos srityje. Asamblėja skatina bendradarbiauti gynybos politikos sektoriuje (skatina taikos palaikymą, tarptauti- nio saugumo priemonių plėtrą, bendradarbiavimą sprendžiant karinius ir gynybos klausimus, veiksmų koordinavimą ginklų eskorto srityje) (Euro-Latin American Parliamentary Assembly rezoliution, 2011). Vienas pagrindinių strateginės partne- rystės tikslų rezoliucijoje – siekis uždrausti branduolinio ginklo platinimą. 3. ES santykiai su regionų grupėmis ir pavienėmis Lotynų Amerikos valstybėmis ES ir Lotynų Amerikos santykiai puoselėjami trijuose skirtinguose lygmenyse (plačiau žr. 3 pav.). Regioninis lygmuo apima bendradarbiavimą su Rio grupės šalimis. ES subregioniniame lygmenyje bendradarbiauja su MERCOSUR valsty- bėmis, Centrinės Amerikos organizacija, Andų bendrija. Dvišalių santykių lygmuo 43 Mindaugas Norkevičius labiausiai regimas santykiuose su Meksika, Brazilija ir Čile. ES ir Lotynų Ame- rikos susitarimai apima ekonominį, technologinį ir mokslinį bendradarbiavimą, demokratijos deficito mažinimą, žmogaus teisių sklaidą. Tokios prioritetinės sritys bendradarbiaujant su subregioninėmis grupėmis Lotynų Amerikoje išryškėjo pa- sirašius deklaraciją dėl politinio dialogo su Andų bendrija 1983 m., po vienerių metų – ir su Centrinės Amerikos šalių organizacija (Borzel, Risse, 2009, p. 10–11). Ekonominis bendradarbiavimas ir politinio dialogo puoselėjimas tarpregio- niniuose santykiuose ne visada išreiškė aiškų siekį skatinti regioninę integraciją. Tarpregioninio Lotynų Amerikos ir ES bendradarbiavimo susitarimuose, kuriuos ES pradėjo pasirašinėti nuo 1990 m., numatyta tiesioginė parama kurti regionines institucijas. Taip ES prisideda prie institucinės integracijos plėtros ir sklaidos Lo- tynų Amerikoje. 3 pav. Tarpregioniniai Europos Sąjungos ir Lotynų Amerikos bendradarbiavimo lygmenys Pastaba: sudaryta autoriaus. 3 pav. Tarpregioniniai Europos Sąjungos ir Lotynų Amerikos bendradarbiavimo lygmenys Pastaba: sudaryta autoriaus. 3.1. ES ir Rio grupės santykiai Pagrindinės tarpregioninio bendradarbiavimo tendencijos regioniniame ly- gmenyje pastebimos su Rio grupe. Ši grupė apima beveik visą Lotynų Amerikos regioną. 1990 m. priimta deklaracija rodo ES ir Rio grupės santykių pradžią (Sei- delmann, 2005, p. 35–37). Regioninės grupės yra patariamojo pobūdžio organi- zacija, kurios tikslas – konsultacijų ekonominiais, politiniais ir socialiniais klau- simais rengimas. Kitas organizacijos bruožas – integracijos procesų skatinimas, paremtas bendradarbiavimo sklaida. Rio grupės veikla tarpregioninių santykių 44 TARPREGIONINIO EUROPOS SĄJUNGOS IR LOTYNŲ AMERIKOS... kontekste paprastai apima švelnios galios (angl. soft power) politikos sritis: aplin- kosauga, demokratijos idėjų sklaida, žmogaus teisių užtikrinimas, kultūros ir so- cialinė politika. Bendradarbiavimo atgimimu galima laikyti 2008 m. Limoje vykusį aukščiau- sio lygio ES ir Rio grupės valstybių atstovų susitikimą, kur apibrėžtos pagrindinės tarpregioninio bendradarbiavimo sritys, pažymėtas siekis užtikrinti daugiašalį ES ir Lotynų Amerikos regiono bendradarbiavimą (The European Council, 2013). Pa- grindinės sritys apima socialinės politikos plėtrą regione, siekiant sumažinti socia- linės nelygybės ir atskirties problemas. Kita sritis – ekonominis regiono vystyma- sis sprendžiant aplinkos apsaugos problemas. Galiausiai tarpregioninis bendradar- biavimas apima siekį stiprinti regioninės integracijos procesus Lotynų Amerikoje. 3.2. ES ir MERCOSUR valstybių bendradarbiavimas ES ir MERCOSUR valstybių santykiai įteisinti 1995 m. bendruoju sutarimu, kuris lėmė tolesnį tarpregioninį bendradarbiavimą politikos ir prekybos srityse. ES, pasirašiusi paramos sutartis su MERCOSUR valstybėmis, žengė žingsnį, lė- musį tolesnius tarpregioninio ekonominio bendradarbiavimo ir politinio dialogo skatinimo etapus. 2002 m. Madrido aukščiausiojo lygio susitikime ES ir MERCOSUR valstybių atstovai atnaujino derybas ekonomikos srityje. Šalys sutarė dėl darbotvarkės ir de- rybų procedūrų, numatė bendradarbiabimo planą būsimo susitarimo pagrindu. Ne- palanki tarptautinė situacija – 2001–2002 m. krizė Argentinoje neigiamai paveikė derybas. 2003 m. lapkričio 12 d. ES ir MERCOSUR ministrų susitikime nuspręsta 2004 m. spalio mėnesį Lisabonoje užbaigti derybas dėl asociacijos ir laisvosios prekybos susitarimo, tačiau to padaryti nepavyko. Susitikimo nesėkmė parodė, kad pagrindinė sutarimo kliūtis – žemės ūkis. Nors Gvadalacharos, Vienos ir Limos aukščiausio lygio susitikimai nuteikė opti- mistiškai, derybos vis dar nesibaigia (Developments in the EU-Mercosur Associa- tion Agreement, 2004). Derybų atnaujinimo sąlygos ir darbotvarkė priklauso nuo Pasaulio prekybos organizacijos (PPO) derybų ir prekybos liberalizavimo pažan- gos (The European Parliament, 2013). ES ir MERCOCUR santykiai daugiausia remiasi ekonominio bendradarbiavi- mo puoselėjimu. Tokias tendencijas galima formuluoti siejant tai su tuo, kad ES yra pagrindinė investuotoja ekonomikos ir finansų srityse, o MERCOSUR – vie- na didžiausių ekonominių sąjungų pasaulyje. MERCOSUR galima laikyti Lotynų Amerikos ekonominio potencialo pavyzdžiu. 45 45 Mindaugas Norkevičius 3.3. Bendradarbiavimas su Centrinės Amerikos valstybių organizacija ES paramą Centrinės Amerikos regiono valstybėms skiria nuo 1984 m. Tais pačiais metais ES ir Centrinės Amerikos atstovai susitiko San Chosė (Kosta Rika) ieškoti būdų, kaip susilpninti regioną ištikusią krizę (Mold, 2007, p. 155). ES kartu su Centrinės Amerikos valstybėmis 1993 m. pasirašė bendradarbiavimo sutartį, kuri sutvirtino tarpregioninius ryšius. 2002 m. Madride vykusioje ministrų kon- ferencijoje nuspręsta parengti naują tarpregioninio bendradarbiavimo programą, kuri pakeistų vykdytą 1993 m. Naujasis bendradarbiavimo etapas – 2003 m. spalio mėnuo, kai ES ir Centri- nės Amerikos valstybės sutarė dėl bendradarbiavimo ir politinio dialogo puoselė- jant tarpregioninius santykius. Susitarimas pasirašytas 2003 m. gruodžio mėnesį Romoje vykusio aukščiausio lygio vadovų susitikimo metu (EU-Central America Agreement, 2003). Galima skirti pagrindines gaires, kuriomis remiasi ES, teik- dama paramą Centrinės Amerikos valstybėms: ekonominis bendradarbiavimas ir socialinis vystymasis, politinio stabilumo (demokratijos) užtikrinimas, žmogaus teisių skaida regione, kova su terorizmu ir imigracijos kontrolės didinimas. 2010 m. atnaujintas ir parengtas susitarimas tarp ES ir Centrinės Amerikos valstybių. 2009 m. to padaryti nepavyko dėl Hondūre įvykusio perversmo ir prasi- dėjusios konstitucinės krizės. Prekybos sektoriaus liberalizavimas ES ir Centrinės Amerikos atžvilgiu lemia didėjančią subregiono įtaką analizuojant tarpregioninius santykius Lotynų Amerikoje. Taip ES prisideda prie ekonominės integracijos plė- tros Lotynų Amerikos regione. Tokia vykdoma tarpregioninė politika atskleidžia prekybos ir ekonomikos li- beralizavimą tarp ES ir Centrinės Amerikos. Šios subregioninės grupės vaidmuo tarptautinėje arenoje didėja, pasaulyje matomas vis didesnis bendradarbiavimo su Centrinės Amerikos valstybėmis poreikis. 3.4. ES ir Andų bendrijos santykiai ES ir Andų bendrijos santykiai plėtojami nuo 1969 m., kai Lotynų Ameriko- je susibūrė ši subregioninė grupė. 1983 m. ES su Andų bendrijos šalimis paren- gė tarpregioninio bendradarbiavimo susitarimą, kurį atnaujino 1993 m. (Europos Tarybos informacija, 2013). Pagrindiniai susitarimo aspektai: bendradarbiavimas prekybos srityje, politinio dialogo puoselėjimas. Andų bendrija siekė parengti įvai- rių bendradarbiavimo sričių susitarimą, kuris apimtų ekonomines ir politines sritis. 2003 m. pasirašytas susitarimas, atnaujinęs regioninį bendradarbiavimą tarp Andų bendrijos valstybių ir ES. Naujasis susitarimas turėjo praplėsti laisvosios prekybos regioniniame lygmenyje nuostatas, kurios būtų pagrįstos regioninės integracijos 46 TARPREGIONINIO EUROPOS SĄJUNGOS IR LOTYNŲ AMERIKOS... procese pasiekta pažanga ir Pasaulio prekybos organizacijos Dohos derybų roundo rezultatais (Smith, 2006, p. 220). Susitarime taip pat numatyta kova su terorizmu ir nelegalia imigracija. Esminiu bendradarbiavimo laikotarpiu galima laikyti 2006 m., kai Europos Komisija pateikė siūlymų Europos Tarybai pradėti ES ir Andų bendrijos derybas dėl bendradarbiavimo sutarčių pasirašymo. Tačiau Andų bendrija šiuo klausimu nebuvo vieninga. Vienoje pusėje buvo Bolivija ir Ekvadoras, kurios nurodė, kad derėtų atsisakyti kai kurių susitarimo punktų, konkrečiai – prekybos liberaliza- cijos kriterijų, kurie apima investicijų, viešųjų pirkimų, paslaugų ir intelektinės nuosavybės sritis. Kitoje pusėje buvo Peru ir Kolumbija – ekonominiu požiūriu pažangesnės valstybės Lotynų Amerikoje, siekusios glaudesnio ekonominio ben- dradarbiavimo tarpregioniniame lygmenyje. Daugiašalio politinio dialogo ir susitarimų tarp ES ir Andų bendrijos parengti nepavyko, tačiau 2009 m. pradėta skatinti prekybos su Peru, Kolumbija ir Ekva- doru politika. Deja, stringantys susitarimai su Andų bendrija ir skirtingi Lotynų Amerikos valstybių interesai neleidžia aktyviai puoselėti ekonominio bendradar- biavimo bendrame subregioniniame lygmenyje. 3.5. ES bendradarbiavimas su Čile, Meksika ir Brazilija ES puoselėja santykius ir su Lotynų Amerikos valstybėmis Meksika, Čile ir Brazilija. Bendradarbiavimo santykiai su Meksika užmegzti 1997 m. Tais metais sudarytas ES ir Meksikos susitarimas, apimantis regioninio bendradarbiavimo, ekonominės partnerystės, politinio dialogo ir koordinavimo aspektus. Dėl regio- ninės politikos plėtros: nuo 2003 m. Meksikos pramonės produkcija gali laisvai patekti į ES rinkas. Meksikos rinka ES eksportui atsivėrė nuo 2007 m. ES eks- portas į Meksiką išaugo 30 %, o Meksikos eksportas į ES padvigubėjo (Sanahuja, 2006, p. 31). Susitarimai lėmė laisvosios prekybos zonos sukūrimą, demokratijos idėjų sklaidą regione, žmogaus teisių įteisinimą ir tarpregioninį politinį dialogą. Plėtojama ne tik regioninė prekyba, 2004 m. pasirašyta bendradarbiavimo sutartis mokslo ir technologijų srityse. Šalys rengia projektus, siekdamos vykdyti bendrus susitarimus švietimo ir mokslo srityse. Aktyvus ES bendradarbiavimas su Čile prasidėjo 2002 m., nors jau 1996 m. parengtas susitarimas tarp šalių, o 1999 m. pradėtos derybos dėl asociacijos su- sitarimo. ES vykdomos politikos dėl Čilės principai apima prekybos skatinimą, regioninį bendradarbiavimą ir politinį dialogą. Susitarimas lėmė prekybos libe- ralizavimą ir ekonomikos atsivėrimą. Pagrindine bendradarbiavimo ašimi galima laikyti tai, kad nuo 2013 m. 97 % Čilėje pagaminamų produktų gali laisvai patekti į ES rinkas. Šis susitarimas, EP atstovų nuomone, laikomas XXI amžiaus susitarimu 47 Mi d N k iči Mindaugas Norkevičius (The European Parliament, 2013). Tai naujas žingsnis, nes ES pirmą kartą pasirašė tokį susitarimą su šalimi, kuri nepretenduoja tapti ES nare. Brazilija pirmoji iš visų Pietų Amerikos valstybių 1960 m. užmezgė diploma- tinius regioninio bendradarbiavimo santykius su EEB. Spartus ekonominio sekto- riaus augimas Brazilijoje paskatino ES 2007 m. pasirašyti strateginės partnerystės sutartį (The European Council, 2013). Bendradarbiavimo tikslai: žmogaus teisių sklaida, ekonominio bendradarbiavimo perspektyvos, mokslo, technologijų ir ino- vacijų plėtra, socialinės politikos klausimai ir regioninės integracijos aplinkybės. Č ES, pasirašydama bendradarbiavimo sutartis su Meksika, Čile ir Brazilija, galėjo džiaugtis puikiais rezultatais politikoje, ekonomikoje ir prekyboje. Lotynų Amerikos regione ES teikia nuolatinę pagalbą vykstant įvairiems konfliktams, sti- chinėms nelaimėms, tai vykdo Europos humanitarinės pagalbos centras. ES, kaip tarpregioninio bendradarbiavimo partnerė, prisideda prie regionalizacijos procesų Lotynų Amerikos regione, skatinant ekonominę, socialinę ir institucinę integraciją. Išvados Pagrindiniai ES ir Lotynų Amerikos regioninio bendradarbiavimo tikslai yra vyraujanti socialinė nelygybė ir skurdas Lotynų Amerikos regione, demokratijos deficitas bei augantis Lotynų Amerikos ekonominis potencialas ir atsiveriančios naujos rinkos. Išanalizavus ES vykdomą užsienio politiką Lotynų Amerikos re- gione galima daryti prielaidą, kad ES bendradarbiavimas su Lotynų Amerika pa- prastai vyksta per subregionines organizacijas – MERCOSUR, CAN, UNASUR. Europos Sąjungos Lotynų Amerikoje vykdoma užsienio politika daugiausia plėto- jama kultūros, politinio dialogo puoselėjimo ir prekybos srityse. ES institucijų vaidmuo kuriant užsienio politiką Lotynų Amerikos regione pastaraisiais metais išaugo. EP, atsižvelgdamas į susitarimus su Lotynų Amerikos regionu, subūrė delegacijas regioniniams ryšiams su MERCOSUR valstybėmis, Centrinės Amerikos organizacija, Andų bendrija, Meksika ir Čile stiprinti. EP pa- brėžia būtinybę užtikrinti žmogaus teises, diegti demokratijos, teisinės valstybės pagrindus Lotynų Amerikos regione. Svarbiausios sritys, kuriose veikia EP, orien- tuodamasis į Lotynų Amerikos regioną: politinio dialogo skatinimas, ekonominių (prekybos) ryšių puoselėjimas, socialinės apsaugos (skurdo ir socialinės atskirties mažinimas) užtikrinimas. ES ir Lotynų Amerikos šalių parlamentinės asamblėjos EUROLAT tikslas – pateikti ir priimti rekomendacijas, rezoliucijas dėl institucijų bei organizacijų, kurios atsakingos už tarpregioninę ES ir Lotynų Amerikos plėtrą. Tarpregioninis bendradarbiavimas vyksta trijuose skirtinguose lygmenyse. Regioninis bendradarbiavimas su Rio grupe pasižymi patariamojo pobūdžio po- litika. Pagrindinės bendradarbiavimo sritys: socialinė gerovė, aplinkos apsaugos 48 TARPREGIONINIO EUROPOS SĄJUNGOS IR LOTYNŲ AMERIKOS... klausimai ir demokratinių idėjų sklaida. Subregioninio bendradarbiavimo priori- tetinės sritys: ekonominės, politinės ir socialinės veiklos skatinimas, integracijos procesai. Pasirašytos ekonominio bendradarbiavimo sutartys su Andų bendrija ir Centrinės Amerikos organizacija leidžia aktyviau išnaudoti Lotynų Amerikos eko- nominį potencialą. Tačiau sutarties tarp ES ir MERCOSUR, kuri turi didžiausią ekonominį potencialą ne tik Lotynų Amerikos regione bet ir pasaulyje, nebuvimas rodo nevisiškai išnaudotas abipusių santykių bendradarbiavimo galimybes. Aktyviausiai dvišaliai santykiai plėtojami su Meksika, Brazilija ir Čile. Dvi- šaliai santykiai su šiomis valstybėmis netolygūs. Santykiai su Brazilija mažiau skatinami ir plėtojami nei su Meksika ir Čile, kurios traktuojamos kaip strategiškai didesnį potencialą turinčios partnerės. Augantis Lotynų Amerikos potencialas ska- tina ES aktyviai bendradarbiauti su šiuo besivystančiu regionu, plėtojant politinį dialogą ir ekonominį bendradarbiavimą. Tarpregioninių santykių tarp ES ir Lotynų Amerikos regiono puoselėjimas skatina gilesnius regioninės integracijos procesus, kurie apima ekonomines, politines, institucines ir socialines sritis Lotynų Ameri- kos regione. Gauta 2014 04 05 Pasirašyta spaudai 2014 07 01 Literatūra A History of Latin America: Independence to the Present. Boston: Cengage Lear- ning. Mold, A. (2007). EU development policy in a changing world: challenges for the 21st century. Amsterdam Uni- versity Press. lski Instytut Spraw Międzynarodowych. (2010). Europe and Latin America: looking at each other? Varšav Sanahuja, A. (2006). América Latina: las visiones y políticas de Europa. Foreign Affairs en español, vol. 6, nr. 1: 2–31. Seidelmann, R. (2005). Relations between the European Union and Latin America; Biregionalism in a Changing Global System, p. 279–318. y p Smith, M. (2006). The European Union’s roles in international politics: concepts and analysis. New York. Stavridis, S., Ajenjo, N. (2010). EU-Latin American parliamentary relations: some preliminary comments on EUROLAT. Jean Monnet/Robert Schuman Paper Series, vol. 10, nr. 3, April: 29–34. e EU’s relations with Central America. Prieiga internete: http://eeas.europa.eu/ca/index_en.htm; The European Union and Latin America: Global Players in Partnership. Prieiga internete: http://www.eeas.eu- ropa.eu. киннер, М., Редферн, Д., Фармер, Д. (1999). География А–Я: словарь-справочник. Москва: Файр-Прес Literatūra Arts, K., Dickson, A. (2004). EU development cooperation: from model to symbol. UK: Manchester University Press. Arts, K., Dickson, A. (2004). EU development cooperation: from model to symbol. UK: Manchester University Press. Borzel, A., Risse, T. (2009). Regionalism: the EU as a Model of Regional Integration. KFG Working Paper, nr. 7: 10–11. Borzel, A., Risse, T. (2009). Regionalism: the EU as a Model of Regional Integration. KFG Working Paper, nr. 7: 10–11. Communication from the commission to the European Parliament and the Council. (2009). The European Union and Latin America: Global Players in Partnership, Brussels. Prieiga internete: http://www.eeas.europa.eu/la/ docs/com09 495 en.pdf. Communication from the commission to the European Parliament and the Council. (2009). The European Union and Latin America: Global Players in Partnership, Brussels. Prieiga internete: http://www.eeas.europa.eu/la/ docs/com09_495_en.pdf. Communication from the Commissionto the Council and the European Parliament. A stronger partnership betwe- en the European Union and Latin America. Prieiga internete: europa.eu.int/comm/external_relations/index. htm. Developments in the EU-Mercosur Association Agreement. Prieiga internete: http://cap2020.ieep.eu/2011 developments-in-the-eu-mercosur-association-agreement. developments-in-the-eu-mercosur-association-agreement. Doidge, M. (2007). From development regionalism to development interregionalism. NCRE Working Paper, 7 (1) 18 idge, M. (2007). From development regionalism to development interregionalism. NCRE Working Pape nr. 7 (1): 18. ES ryšių su Lotynų Amerika strategija: Europos Parlamento rezoliucija dėl ES santykių su Lotynų Amerika stra- tegijos. (2010). Prieiga internete: http://eurlex.europa.eu/LexUriServ/LexUriServ.do?uri=CELEX:21996A0 319(02):EN:NOT. ( ) EU-Andean Community. Prieiga internete: http://www.consilium.europa.eu/press/press-releases/latest-press-rele- ases/newsroomrelated.aspx?bid=102&grp=5680&lang=1&cmsid=360. EU-Central America Agreement. Prieiga internete: http://ec.europa.eu/world/agreements/prepareCreateTreaties- Workspace/treatiesGeneralData.do?step=0&redirect=true&treatyId=2121. Euro-Latin American Parliamentary Assembly rezoliution. Relations between the European Union and Latin America and the Caribbean in the field of security and defence. (2011). Urugvajus. Prieiga internete: http:// www.europarl.europa.eu/intcoop/eurolat. 49 Mindaugas Norkevičius European Commission Development and cooperation – EuropeAid. Prieiga internete: http://ec.europa.eu/euro- peaid. European Union external extion. Prieiga internete: http://eeas.europa.eu/mercosur/index_en.htm. The European Parliament. Prieiga internete: ec.europa.eu/external_relations/index_en.htm. The European Council. Prieiga internete: http://www.consilium.europa.eu/policies/foreign-policy/third-coun- tries-and-regions/eu-latin-america-relations.aspx?lang=lt. Gilson, J. (2005). New Interregionalism? The EU and East Asia. European integration, nr. 27 (3): 309. Goodal, B. (1987). Dictionary of human geography. The Penguin. oodal, B. (1987). Dictionary of human geography. The Penguin. Gordillo, A. (2003). The future of Latin America: can the EU help? London. Hanggi, H., Roloff, R., Ruland, J. (2006). Interregionalism and International Relations. Routledge, Taylor & Francis e-Library. Hwee, Y. L., Vidal, L. (2008). Regionalism and Interregionalism in the ASEM context: current dynamics and theo- retical approaches. CIDOB foundation. Prieiga internete: www.cidob.org/es/content/download/9506/96818/ file/doc asia 23.pdf. Interregional Framework Cooperation Agreement. Prieiga internete: http://circa.europa.eu/irc/opoce/fact_sheets/ info/data/relations/relations/article_7247_lt.ht. Keen, B., Haynes, K. (2009). Dynamics of Inter-regional Cooperation of European Union and Latin America Mindaugas Norkevičius Summary Regional processes are not new sphere in international relationships. Regio- nalism is understandable as integration of countries into specific regions. Ambi- valence and complexity of regional processes are demonstrated by the variety of theories which includes the conception of old and new regionalism. Regions are defined by economical or social aspects and individual structures of regionalism. 50 TARPREGIONINIO EUROPOS SĄJUNGOS IR LOTYNŲ AMERIKOS... International regional processes can be decribed as dynamic processes which try to make regions as separate geopolitical items. Latin America is related with linguistic, historical and religious attitudes. Different regional associations promote regional processes. There are two associa- tions in regional: the group of Rio and CELAC. In a sub-regional level: MERCO- SUR, CAN, UNASUR, ALBA and CARICOM. In order to reveal the processes of regional integration in an international extent, the assessment of interregional collaboration between Latin America and European Union was done. The basic reasons of regional collaboration between Latin America and European Union are social inequality, poverty, lack of democracy, growing economy in Latin America and new markets which are emerging there. Such international organizations as MERCOSUR, UNASUR, CAN pursue the interregional collaboration in that area. Promotion of economical, social and institutional integrations joins regional pro- cesses in Latin America. It was established that this region demonstrates position in global stage of po- litics. Connections and opportunities of economical colloboration and functions of political processes join the forming tendencies and processes of separate regions. Regional integration in the area of Latin America is directed towards economi- cal and social collaboration, development of economics, forming of identities and effective work of institutions. The basic reasons of inter-regional collaboration between EU and Latin Ame- rica are: social inequality and destitution in Latin America, shortfall of democracy, growing economical potential of Latin America and various new markets. Analysis of EU foreign policy in Latin America allows to make an assumption that EU col- laborates with Latin America mostly through sub-regional organizations such as MERCOSUR, CAN, UNASUR. EU foreign policy in Latin America is mostly associated with the development of cultural and political dialogue between two countries and trade bussiness. Po- wer of EU institutions in forming foreign polity in Latin America has grown in recent years. European Parliament (EP) created separate delegations in order to strengthen regional relations with countries of MERCOSUR, Central America‘s organization, community of Andes, Mexico and Chile. Summary EP emphasizes human rights, democracy, creation of legal nation‘s basement in the region of Latin Ame- rica. EP and the region of Latin America focuses on encouragement of political dialogue, development of economical (market) contacts, social security (reduction of deprivation and social disjuncture). EUROLAT is parliamentary assembly of EU and Latin America. The aim of this organization is to represent and accept recomendations, resolutions of institutions and organizations which are responsi- ble for inter-regional develoment between EU and Latin America. 51 Mindaugas Norkevičius Mindaugas Norkevičius Inter-regional collaboration happens in three different levels. Regional colla- boration with Rio group is known for deliberative politics. Social welfare, environ- ment protection and democracy are the general spheres of it. The most important areas in sub-regional collaboration are: encouragement of economical, political and social activities, development of integration processes. Economical collabora- tion treaties with community of Andes and organization of Central America give an opportunity to use economical potential of Latin Ametica more actively. Howe- ver, absence of association treaty between EU and MERCOSUR, which has the biggest economical potential not only in Latin America but in whole world as well, demontrates that not all collaboration possibilities were fully used. The most active development of bilateral relations happens with seperate countries such as Mexico, Brasil and Chile. The position of these countries in inter- regional relation context shows oneness on one side and unsuccessful negotiation in sub-regional level on the other side. Bilateral relations among EU and these countries are not equal. Due to the fact that Mexico and Chile has bigger potenti- al in successful collaboration, relations with these two countries are encouraged more than with Brasil. Growing potential of Latin America encourages EU to pur- sue active collaboration with this region. Development of inter-regional relations between EU and Latin Americaʼs region encourages deeper processes of regional integration which include economical, political, institutional and social areas in the region of Latin America. 52
https://openalex.org/W3004159222	https://europepmc.org/articles/pmc6987175?pdf=render	English	null	Risk of active tuberculosis development in contacts exposed to infectious tuberculosis in congregate settings in Korea	Scientific reports	2,020	cc-by	7,796	www.nature.com/scientificreports www.nature.com/scientificreports Risk of active tuberculosis development in contacts exposed to infectious tuberculosis in congregate settings in Korea Shin Young Park, Sunmi Han, Young-Man Kim, Jieun Kim, Sodam Lee, Jiyeon Yang, Un-Na Kim* & Mi-sun Park* Shin Young Park, Sunmi Han, Young-Man Kim, Jieun Kim, Sodam Lee, Jiyeon Yang, Un-Na Kim* & Mi-sun Park* Contact investigation is an important and effective active case-finding strategy, but there is a lack of research on congregate settings in countries with an intermediate incidence. This study determined the incidence of and risk factors for tuberculosis (TB) development after exposure in congregate settings. This retrospective cohort study included 116,742 contacts identified during the investigation of 2,609 TB cases diagnosed from January to December 2015. We searched the Korean National Tuberculosis Surveillance System TB registry to identify contacts that developed active TB during follow-up until May 2018. During the mean observation period of 2.9 years, 499 of 116,742 contacts (0.4%) developed new active TB. From these contacts, 404 (81.0%) developed TB within 2 years after exposure. The 2-year Kaplan-Meier cumulative risk for TB was the highest in contacts aged ≥65 years [1%; 95% confidence interval (CI), 0.8–1.3]. Contacts with LTBI who completed chemoprophylaxis exhibited a lower risk of active TB development than those without chemoprophylaxis (adjusted hazard ratio, 0.16; 95% CI, 0.08–0.29). Aggressive contact investigation is effective for the early detection and prevention of TB in congregate settings. The risk of progression to active TB among contacts with LTBI can be minimised by the completion of chemoprophylaxis. Tuberculosis (TB) imposes a high global disease burden, with more than 10 million new patients and 1.6 million annual deaths worldwide1. The global burden of latent TB infection (LTBI) was 23.0%, amounting to approxi- mately 1.7 billion people. WHO South-East Asia, Western-Pacific, and Africa regions had the highest prevalence and accounted for approximately 80% of those with LTBI1,2. pp y Although the incidence of TB in the Republic of Korea (ROK) has decreased from 89 per 100,000 in 2013 to 70 per 100,000 in 2017, it remains greater than that in any other country within the Organization for Economic Cooperation and Development nations1,3. p p Generally, 5–10% individuals with LTBI develop active TB in their lifetime, with 50% developing active dis- ease within 2 years after infection4–6. Thus, the management of recent contacts of patients with infectious TB is important for the overall management of TB. Division of TB Investigation, Korea Centers for Disease Control and Prevention, Osong, Republic of Korea. *email: unnakim@korea.kr; pmsun4686@naver.com Scientific Reports \| (2020) 10:1306 \| https://doi.org/10.1038/s41598-020-57697-1 Risk of active tuberculosis development in contacts exposed to infectious tuberculosis in congregate settings in Korea The incidence of TB in contacts was 146 per 100,000 person-years; the incidence per 100,000 person-years was 414 within 3 months. The 2-year risk of TB in contacts was 0.2% [95% confidence interval (CI), 0.1–0.2] in those aged 0–18 years, 0.3% (95% CI, 0.2–0.4) in those aged 19–35 years, 0.5% (95% CI, 0.4–0.6) in those aged 36–64 years, and 1.0% (95% CI, 0.8–1.3) in those aged ≥65 years. The risk was the highest in individuals aged ≥65 years (Fig. 2A). Months After Notification Index Case Contacts with TB, No. (Cumulative %a) No. of Contacts Person-Years of Observation, No. Cumulative Incidence/100,000 (95% CI) Incidence Rate/100,000b (95% CI) Total 499 116,742 340,961 427 (391.2–467.0) 146 (133.5–159.2) 0–3 119 (23.8) 116,623 28,769 102 (84.9–122.5) 414 (339.3–488.0) 4–6 85 (40.9) 116,538 28,744 73 (58.6–90.6) 296 (232.9–358.6) 7–9 50 (50.9) 116,488 28,726 43 (32.1–57.0) 174 (125.8–222.3) 10–12 48 (60.5) 116,440 30,312 41 (30.7–55.1) 158 (113.6–203.1) 13–24 102 (81.0) 116,338 116,390 88 (71.8–106.9) 88 (70.6–104.6) 25–42 95 (100) 9,958 108,020 954 (776.6–1,170.0) 88 (70.2–105.6) Table 1. Rates of active tuberculosis and time to diagnosis for contacts of index cases in congregate settings. TB, tuberculosis; CI, confidence interval. aCalculated as the proportion of all contacts with tuberculosis (denominator, 499). bIncidence rate in cases per 100,000 person-years observation. Table 1. Rates of active tuberculosis and time to diagnosis for contacts of index cases in congregate settings. TB, tuberculosis; CI, confidence interval. aCalculated as the proportion of all contacts with tuberculosis d i 499) bI id i 100 000 b i Table 1. Rates of active tuberculosis and time to diagnosis for contacts of index cases in congregate settings. TB, tuberculosis; CI, confidence interval. aCalculated as the proportion of all contacts with tuberculosis (denominator, 499). bIncidence rate in cases per 100,000 person-years observation. Risk of active tuberculosis development in contacts exposed to infectious tuberculosis in congregate settings in Korea The World Health Organization (WHO) recommends that tests and treatments for LTBI should be prioritised for contacts of TB patients7.h The Korea Centers for Disease Control and Prevention (KCDC) established a TB epidemic investigation team in 2013. Between 2013 and 2017, there were 12,447 investigations with about 700,000 contacts8. However, follow-up research addressing TB incidence among contacts residing in areas with intermediate-level TB inci- dence is lacking, and most studies have addressed household contacts, with a lack of discussion regarding the TB incidence and relevant risk factors among contacts in congregate settings9–11.h g g g g The present study aimed to analyse the TB incidence in individuals who had contact with TB cases in con gate settings and evaluate the risk factors that influence the development of TB among these contacts. www.nature.com/scientificreports/ Figure 1. Flow chart for contacts of index cases of tuberculosis (TB) in congregate settings (January 2013 to May 2018). TB, tuberculosis; PT, preventive therapy; LTBI, latent tuberculosis infection. Figure 1. Flow chart for contacts of index cases of tuberculosis (TB) in congregate settings (January 2013 to May 2018). TB, tuberculosis; PT, preventive therapy; LTBI, latent tuberculosis infection. Months After Notification Index Case Contacts with TB, No. (Cumulative %a) No. of Contacts Person-Years of Observation, No. Cumulative Incidence/100,000 (95% CI) Incidence Rate/100,000b (95% CI) Total 499 116,742 340,961 427 (391.2–467.0) 146 (133.5–159.2) 0–3 119 (23.8) 116,623 28,769 102 (84.9–122.5) 414 (339.3–488.0) 4–6 85 (40.9) 116,538 28,744 73 (58.6–90.6) 296 (232.9–358.6) 7–9 50 (50.9) 116,488 28,726 43 (32.1–57.0) 174 (125.8–222.3) 10–12 48 (60.5) 116,440 30,312 41 (30.7–55.1) 158 (113.6–203.1) 13–24 102 (81.0) 116,338 116,390 88 (71.8–106.9) 88 (70.6–104.6) 25–42 95 (100) 9,958 108,020 954 (776.6–1,170.0) 88 (70.2–105.6) Table 1. Rates of active tuberculosis and time to diagnosis for contacts of index cases in congregate settings. TB, tuberculosis; CI, confidence interval. aCalculated as the proportion of all contacts with tuberculosis (denominator, 499). bIncidence rate in cases per 100,000 person-years observation. 2 Scientific Reports \| (2020) 10:1306 \| https://doi.org/10.1038/s41598-020-57697-1 Results Cumulative TB Risk among Contacts. In total, 116,742 contacts of 2,609 TB cases reported between January and December 2015 were included in this study (Fig. 1). On average, the contacts were followed-up for 2.9 years until 31 May, 2018, and 499 (0.4%) developed active TB (Table 1). From these, 119 (23.8%) and 404 (81.0%) developed TB within 3 months and 2 years, respectively. Results l Of the 116,623 evaluated contacts, 380 (0.3%) developed incident TB (Table 4). The risks of incident TB were higher when index cases were male [adjusted hazard ratio (aHR), 1.48; 95% CI, 1.15–1.90] and when cavitary lesions were present on chest radiographs (aHR, 1.97; 95% CI, 1.59–2.45]. The risks were also higher in male contacts (aHR, 1.54; 95% CI, 1.22–1.95) and contacts aged ≥65 years (aHR, 1.96; 95% CI, 1.14–3.35) than in those aged 0–18 years. Risks were higher in healthcare facilities (aHR, 3.34; 95% CI, 2.18–5.13) and social welfare facilities (aHR, 3.14; 95% CI, 1.90–5.18) than in schools. Furthermore, the risks of incident TB were higher in close contacts than in casual contacts (aHR, 1.76; 95% CI, 1.41–2.20), while they were lower in contacts with negative LTBI screening results (aHR, 0.12; 95% CI, 0.08–0.17), those who were not tested for LTBI (aHR, 0.40; 95% CI, 0.30–0.53), and those who initiated chemoprophylaxis (aHR, 0.50; 95% CI, 0.33–0.74) than in those with positive LTBI screening results who did not receive chemoprophylaxis. Incident TB risk among contacts with LTBI. Of 11,596 contacts who tested positive for LTBI, 45 were diagnosed with co-prevalent TB and 118 developed incident TB (Table 5). The risks of incident TB were higher when index cases were male (aHR, 1.91; 95% CI, 1.19–3.07) and when cavitary lesions were present on chest radiographs (aHR, 2.58; 95% CI, 1.76–3.77), while they were lower in contacts aged 36–64 years (aHR, 0.29; 95% CI, 0.15–0.57) than in those aged 0–18 years. The risks of incident TB were also lower in those who completed treatment for LTBI than in those who did not receive treatment (aHR, 0.16; 95% CI, 0.08–0.29).h The incidence of TB in individuals treated for LTBI, those who initiated but did not complete the treatment, and those who did not receive treatment was 119 (95% CI, 58.9–179.6), 612 (95% CI, 356.4–867.9), and 460 (95% CI, 360.1–560.6) per 100,000 person-years, respectively. In contacts aged ≤35 years, the TB incidence in indi- viduals treated for LTBI was 98 (95% CI, 30.0–165.4) per 100,000 person-years, 766 (95% CI, 364.9–1,168.0) per 100,000 person-years in those who did not complete the treatment, and 842 (95% CI, 570.6–1,113.0) per 100,000 person-years in those who were not treated. Results l Cumulative TB Risk among Contacts. In total, 116,742 contacts of 2,609 TB cases reported between January and December 2015 were included in this study (Fig. 1). On average, the contacts were followed-up for 2.9 years until 31 May, 2018, and 499 (0.4%) developed active TB (Table 1). From these, 119 (23.8%) and 404 (81.0%) developed TB within 3 months and 2 years, respectively. The incidence of TB in contacts was 146 per 100,000 person-years; the incidence per 100,000 person-years was 414 within 3 months. The 2-year risk of TB in contacts was 0.2% [95% confidence interval (CI), 0.1–0.2] in those aged 0–18 years, 0.3% (95% CI, 0.2–0.4) in those aged 19–35 years, 0.5% (95% CI, 0.4–0.6) in those aged 36–64 years, and 1.0% (95% CI, 0.8–1.3) in those aged ≥65 years. The risk was the highest in individuals aged ≥65 years (Fig. 2A). Scientific Reports \| (2020) 10:1306 \| https://doi.org/10.1038/s41598-020-57697-1 www.nature.com/scientificreports/ Figure 2. Cumulative risk of tuberculosis among contacts after notification of index cases, by age, in congregate settings. (A) 116,742 contacts (B) 11,596 contacts with latent tuberculosis infection (LTBI). Figure 2. Cumulative risk of tuberculosis among contacts after notification of index cases, by age, in congregate settings. (A) 116,742 contacts (B) 11,596 contacts with latent tuberculosis infection (LTBI). Cumulative TB risk among contacts with LTBI. Of the 11,596 contacts with positive test results for LTBI, 163 (1.4%) developed active TB (Table 2), with 45 (27.6%) and 138 (84.7%) diagnosed within 3 month and 2 years, respectively. The incidence of TB in these contacts was 483 per 100,000 person-years; the incidence per 100,000 person-years was 1,577 within 3 months. The 2-year risk of TB was 2.3% (95% CI, 1.2–4.3) in contacts aged ≥65 years, 1.9% (95% CI, 1.4–2.6) in those aged 0–18 years, 1.6% (95% CI, 1.1–2.2) in those aged 19–35 years, and 0.7% (95% CI, 0.5–0.9) in those aged 36–64 years. Thus, contacts aged 36–64 years exhibited the lowest risk (Fig. 2B). Co-prevalent TB risk among contacts. Of the 116,742 evaluated contacts, 119 (0.1%) exhibited co-prevalent TB (Table 3). The risks of co-prevalent TB were higher when index cases were male [adjusted odds ratio (aOR), 1.56; 95% CI, 1.05–2.33] and sputum smears were positive (aOR, 2.29; 95% CI, 1.45–3.59). The risks were also higher when contact occurred in healthcare facilities than when contact occurred in schools (aOR, 2.56; 95% CI, 1.62–4.06). Incident TB risk among contacts. Results l Among contacts aged ≥36 years, the TB incidence per 100,000 person-years was 159 (95% CI, 41.3–277.6), 453 (95% CI, 139.1–766.5), and 333 (95% CI, 234.9–431.9) in those who were treated, those who did not complete the treatment, and those who were not treated, respectively, with no statistically significant differences (Table 6). Discussion Of the 163 contacts with LTBI who eventually developed active TB, 84.7% were diagnosed within 2 years. A Canadian study that followed up contacts in congregate settings and household contacts for an average of 6 years reported that 86% contacts who developed active TB were diagnosed within 2 years, similar to our findings16. Although differing in terms of the follow-up duration and participants, recent studies, including the present one13,14,16–18, have reported that the 2-year incidence of TB in contacts was 63%–94.9%, which is higher than that reported in earlier studies5,19. According to the ROK guidelines on follow-up of contacts, follow-up chest radiography was conducted at 3 months and 9 months after exposure in contacts who were negative for LTBI. When contacts who were positive for LTBI were not treated, they underwent follow-up chest radiography every 3 months for 2 years20. Considering that 60.5% and 81.0% contacts developed TB within 1 and 2 years, respectively, it would be advisable to develop stratified strategies for prolonging the follow-up duration for each contact category beyond that required by the current ROK guidelines.ht q y g The Korean War has often been considered the cause of the high TB incidence in the ROK, despite the devel- oped economic status in this country21. It is estimated that many Koreans were infected with TB in the poor post-war environments in the 1950s and 60 s. The LTBI rate in the ROK was 64.2% in 1960, 59.3% in 1975, 44.4% in 1990, and 33.2% in 201622. The number of new TB cases aged ≥65 years has increased annually, accounting for 45.5% of all TB patients in 20183. In the present study, the LTBI rate was 5.9% in individuals aged 0–18 years, 11.2% in those aged 19–35, 37.4% in those aged 36–64, and 44.5% in those aged ≥65 years (see supplementary information). The risk of incident TB was higher in those aged ≥65 years than in those aged 0–18 years, possi- bly because the cases were previously infected with TB and developed active TB as they aged and their immune functions decreased. In contrast, among individuals with LBTI, the risks of incident TB were lower in those aged 36–64 years than in those aged 0–18 years, opposite to the trend observed for the LTBI rate. Discussion The overall incidence in contacts was almost seven times higher than that in the general population of the ROK in 2015 (63 per 100,000 persons)3, and the incidence in contacts who tested positive for LTBI was 22 times higher than that in the general population.i g g In contrast to our findings, studies conducted in the US and the Netherlands, which have a low TB incidence and burden, reported a higher TB incidence of 1.2%–3.5% in contacts12–14. Studies conducted in Taiwan9 and Hong Kong10, which have an intermediate TB burden, similar to that in the ROK, reported incidences of 0.7% and 1.7%, respectively. The relatively lower incidence found in the present study could be attributed to the fact that these previous studies9,10,12–14 included only close contacts of TB cases or household contacts.i p y Moreover, the definition of contacts in the ROK is more comprehensive than that used in other countries. Whereas the mean number of contacts per index case was 45 in this study, it was six in studies from the US12,14, 15 in a study from the Netherlands13, and three in a study from Hong Kong10. According to the 2015 ROK national guidelines, the entire school population was included as contacts when two or more TB cases were confirmed within 6 months or when three or more TB cases were confirmed among students in the same year. Although such aggressive contact investigations may be less efficient than investigations of targeted groups of contacts, the prevention of new active TB by finding and treating LTBI cases could be an important TB management strategy in countries with a relatively high TB incidence, such as the ROK15. Indeed, the TB incidence has been decreasing each year by 5% since 2011, with the number of TB cases among individuals aged <20 years exhibiting a marked decrease. The number of new TB cases among individuals aged <20 years was 1,501 in 2013, accounting for 4.2% of all cases. However, this number had decreased to 508 in 2018, accounting for 1.9% of all cases. Thus, the percentage of younger patients, who are often exposed to congregate settings, has been decreasing in particular3, possibly because of intervention involving contacts in congregate settings. p y g g g g Of the 499 contacts who developed active TB in this study, 81.0% were diagnosed within 2 years. Discussion In this study, we analysed the TB incidence in individuals who had contact with active TB cases in congregate settings in the ROK and sought to confirm the risk factors that influenced the development of TB in these cases. When 116,742 contacts of TB cases reported in 2015 were followed up for an average of 2.9 years, until 31 May, 2018, 499 (0.4%) had newly developed active TB. Of 499 TB cases, 421 (84.4%) were pulmonary tuberculosis and 78 (15.6%) were extra-pulmonary tuberculosis (see supplementary information). The incidence of TB in contacts was 427 per 100,000 person-years, with an incidence of 1,406 per 100,000 person-years in those who Scientific Reports \| (2020) 10:1306 \| https://doi.org/10.1038/s41598-020-57697-1 www.nature.com/scientificreports/ Months After Notification Index Case Contacts with TB, No. (Cumulative %a) No. of Contacts Person-Years of Observation, No. Cumulative Incidence/100,000 (95% CI) Incidence Rate/100,000b (95% CI) Total 163 11,596 33,731 1,406 (1,203.0–1,641.0) 483 (409.1–557.4) 0–3 45 (27.6) 11,551 2,854 390 (287.6–525.8) 1,577 (1,116.0–2,037.0) 4–6 35 (49.1) 11,516 2,843 304 (215.0–427.4) 1,231 (852.3–1,639.0) 7–9 19 (60.7) 11,497 2,837 165 (102.4–263.3) 670 (368.6–970.8) 10–12 14 (69.3) 11,483 2,990 122 (69.4–210.0) 468 (223.0–713.5) 13–24 25 (84.7) 11,458 11,470 218 (144.3–327.1) 218 (132.5–303.4) 25–42 25 (100) 3,063 10,737 816 (540.3–1,221.0) 233 (141.6–324.1) Table 2. Rates of active tuberculosis and time to diagnosis for contacts with LTBI in congregate settings. LTBI, latent tuberculosis infection; TB, tuberculosis; CI, confidence interval. aCalculated as the proportion of all contacts with tuberculosis (denominator, 163). bIncidence rate in cases per 100,000 person-years observation. Table 2. Rates of active tuberculosis and time to diagnosis for contacts with LTBI in congregate settings. LTBI, atent tuberculosis infection; TB, tuberculosis; CI, confidence interval. aCalculated as the proportion of all i h b l i (d i 163) bI id i 100 000 b i Table 2. Rates of active tuberculosis and time to diagnosis for contacts with LTBI in congregate settings. LTBI, latent tuberculosis infection; TB, tuberculosis; CI, confidence interval. aCalculated as the proportion of all contacts with tuberculosis (denominator, 163). bIncidence rate in cases per 100,000 person-years observation. Table 2. Rates of active tuberculosis and time to diagnosis for contacts with LTBI in congregate settings. LTBI, latent tuberculosis infection; TB, tuberculosis; CI, confidence interval. aCalculated as the proportion of all contacts with tuberculosis (denominator, 163). bIncidence rate in cases per 100,000 person-years observation. tested positive for LTBI. Scientific Reports \| (2020) 10:1306 \| https://doi.org/10.1038/s41598-020-57697-1 Discussion Here, individuals aged 0–18 years were likely to have been infected with Mycobacterium tuberculosis relatively more recently and were more likely to develop active TB within 2 years.h y p y The risks of co-prevalent and incident TB were higher in healthcare facilities and social welfare facilities than in schools. This may be because most contacts belonging to medical and social welfare facilities have acute or chronic conditions relative to those in other settings. In the present study, the mean age of contacts in these set- tings was 52.1 ± 18.3 years (see supplementary information), an age group with a relatively high prevalence of diabetes and chronic renal failure23–25, which are known TB risk factors. However, because we could not confirm the presence of underlying diseases in contacts, follow-up research is required. Scientific Reports \| (2020) 10:1306 \| https://doi.org/10.1038/s41598-020-57697-1 www.nature.com/scientificreports/ Contacts Without Contacts With Unadjusted Adjusted Co-prevalent TB, No. (%) Co-prevalent TBa, No. Discussion (%) Odds Ratio (95% CI) p– value Odds Ratio (95% CI) p–value Total 116,623 119 Index factors Sex Male 70,744 (99.9) 83 (0.1) 1.49 (1.01–2.21) 0.044 1.56 (1.05–2.33) 0.026 Female 45,879 (99.9) 36 (0.1) Reference Reference Age, years 0–18 41,007 (99.9) 26 (0.1) Reference 19–35 36,147 (99.9) 36 (0.1) 1.57 (0.94–2.60) 0.079 36–64 23,815 (99.8) 37 (0.2) 2.45 (1.48–4.04) 0.000 ≥65 15,654 (99.9) 20 (0.1) 2.01 (1.12–3.61) 0.019 Sputum smear status Smear-positive 62,265 (99.9) 89 (0.1) 2.75 (1.80–4.21) 0.000 2.29 (1.45–3.59) 0.000 Smear-negative 53,999 (99.9) 28 (0.1) Reference Reference Unknown 359 (99.4) 2 (0.6) 10.74 (2.55–45.27) 0.001 6.80 (1.59–29.12) 0.010 Cavities on chest radiograph Cavities 29,264 (99.9) 43 (0.1) 1.67 (1.14–2.44) 0.008 No Cavities 80,966 (99.9) 71 (0.1) Reference Unknown 6,393 (99.9) 5 (0.1) 0.89 (0.36–2.21) 0.805 Contact factors Sex Male 62,402 (99.9) 70 (0.1) 1.24 (0.86–1.78) 0.246 Female 54,221 (99.9) 49 (0.1) Reference Age, years 0–18 45,103 (99.9) 30 (0.1) Reference 19–35 34,087 (99.9) 30 (0.1) 1.32 (0.79–2.19) 0.278 36–64 29,183 (99.9) 40 (0.1) 2.06 (1.28–3.30) 0.003 ≥65 8,250 (99.8) 19 (0.2) 3.46 (1.94–6.15) 0.000 Congregate settings Schools 65,740 (99.9) 41 (0.1) Reference Reference Workplaces 14,313 (99.8) 23 (0.2) 2.57 (1.54–4.29) 0.000 1.66 (0.96–2.85) 0.067 Healthcare facilities 18,364 (99.8) 36 (0.2) 3.14 (2.00–4.92) 0.000 2.56 (1.62–4.06) 0.000 Social welfare facilities 9,779 (99.9) 12 (0.1) 1.96 (1.03–3.74) 0.039 1.69 (0.88–3.26) 0.113 Others 8,427 (99.9) 7 (0.1) 1.33 (0.59–2.97) 0.483 1.09 (0.48–2.44) 0.829 Type of contact Close contact 51,225 (99.9) 61 (0.1) 1.40 (0.97–2.02) 0.073 Casual contact 62,375 (99.9) 53 (0.1) Reference Unknown 3,023 (99.8) 5 (0.2) 1.94 (0.77–4.87) 0.155 Table 3. Risk factors for co-prevalent active tuberculosis among 116,742 contacts of tuberculosis cases in congregate settings. TB, tuberculosis; CI, confidence interval. aContact diagnosed within 90 days after diagnosis of the index patient. Table 3. Risk factors for co-prevalent active tuberculosis among 116,742 contacts of tuberculosis cases in congregate settings. TB, tuberculosis; CI, confidence interval. aContact diagnosed within 90 days after diagnosis of the index patient. Positive sputum smear results, identification of cavitary lung lesions on radiographs, and the presence of symptoms consistent with TB infection are known risk factors for TB in contacts26,27. In the present study, con- tacts of TB cases with cavitary lung lesions on chest radiographs exhibited greater risks of incident TB. Scientific Reports \| (2020) 10:1306 \| https://doi.org/10.1038/s41598-020-57697-1 Discussion Although the contact investigation undertaken for this study indicated that the presence of cavitary lung lesions on chest radiographs predicted the length of infectivity of index cases, radiographic examination is not currently included among the criteria for conducting investigations. Therefore, we suggest that it would be advisable to include radi- ography in future criteria. g p y In is generally accepted that 10% individuals who acquire LTBI will develop active TB in the absence of pre- ventive therapy, with 50% developing the disease within 2 years after exposure4,16. A Dutch study28 reported that the incidence of TB per 100,000 person-years was 187 in individuals treated for LTBI, 436 in those who did not complete treatment, and 355 in those who were not treated, similar to our findings. In the present study, among individuals aged ≥36 years, the TB incidence per 100,000 person-years was 159 in those who completed chemoprophylaxis and 333 in those who were not treated. Although the incidence was lower in those who completed chemoprophylaxis, the difference was not significant. Because the 2015 ROK national guidelines on TB management only recommend treatment for LTBI in contacts aged <36 years, testing and treatment for LTBI were not as active in contacts aged ≥36 years, which may have influenced the results. More detailed analyses on the effects of treatment for LTBI are required. www.nature.com/scientificreports/ Contacts Without Contacts With Unadjusted Adjusted Incident TB, No. (%) Incident TBa, No. Discussion LTBI, latent tuberculosis infection; TB, tuberculosis; CI, confidence interval. aContact diagnosed within 9 days after diagnosis of the index patient. bIncluded cases with unknown LTBI findings (n = 79). Table 4. Risk factors for incident active tuberculosis among 116,623 contacts of tuberculosis cases in congregate settings. LTBI, latent tuberculosis infection; TB, tuberculosis; CI, confidence interval. aContact diagnosed within 90 days after diagnosis of the index patient. bIncluded cases with unknown LTBI findings (n = 79). The present study had the following limitations. First, we lacked information on underlying diseases, HIV infection, diabetes, nutritional status, body mass index, and smoking, which are recognised TB risk factors, for the contacts. Second, the death of participants during the follow-up period could not be confirmed. Third, because molecular epidemiological tests for index cases and contacts were conducted only in some epidemic cases, the results were not included in the analysis. This limited analysis of the exact route of transmission. Finally, the mean follow-up duration was 2.92 years, which may be too short for the evaluation of active TB development. The present study had the following limitations. First, we lacked information on underlying diseases, HIV infection, diabetes, nutritional status, body mass index, and smoking, which are recognised TB risk factors, for the contacts. Second, the death of participants during the follow-up period could not be confirmed. Third, because molecular epidemiological tests for index cases and contacts were conducted only in some epidemic cases, the results were not included in the analysis. This limited analysis of the exact route of transmission. Finally, the mean follow-up duration was 2.92 years, which may be too short for the evaluation of active TB development. In conclusion, the present study utilised complete enumeration data obtained from contact investigations conducted in the ROK in 2015 to confirm the development of TB in recent contacts within 2 years and evaluated the effects of the current contact investigations conducted in the ROK. The findings suggest that the completion of chemoprophylaxis for LTBI can lower the risk of TB development in contacts. In future, efforts to increase the contact investigation rates and improve the treatment rates for LTBI are necessary. Furthermore, long-term assessment of the effects of contact investigations is also necessary. Scientific Reports \| (2020) 10:1306 \| https://doi.org/10.1038/s41598-020-57697-1 Discussion (%) Hazard Ratio (95% CI) p–value Hazard Ratio (95% CI) p–value Total 116,243 380 Index factors Sex Male 70,468 (99.6) 276 (0.4) 1.70 (1.36–2.13) 0.000 1.48 (1.15–1.90) 0.002 Female 45,775 (99.8) 104 (0.2) Reference Reference Age, years 0–18 40,942 (99.8) 65 (0.2) Reference Reference 19–35 36,055 (99.7) 92 (0.3) 1.62 (1.18–2.23) 0.003 0.74 (0.46–1.18) 0.209 36–64 23,676 (99.4) 139 (0.6) 3.81 (2.83–5.11) 0.000 0.83 (0.51–1.36) 0.480 ≥65 15,570 (99.5) 84 (0.5) 3.46 (2.50–4.79) 0.000 0.51 (0.29–0.90) 0.020 Sputum smear status Smear-positive 62,007 (99.6) 258 (0.4) 1.90 (1.53–2.36) 0.000 Smear-negative 53,879 (99.8) 120 (0.2) Reference Unknown 357 (99.4) 2 (0.6) 2.61 (0.64–10.56) 0.178 Cavities on chest radiograph Cavities 29,106 (99.5) 158 (0.5) 2.11 (1.72–2.60) 0.000 1.97 (1.59–2.45) 0.000 No Cavities 80,759 (99.7) 207 (0.3) Reference Reference Unknown 6,378 (99.8) 15 (0.2) 0.87 (0.51–1.48) 0.624 0.93 (0.54–1.62) 0.816 Contact factors Sex Male 62,159 (99.6) 243 (0.4) 1.53 (1.24–1.89) 0.000 1.54 (1.22–1.95) 0.000 Female 54,084 (99.7) 137 (0.3) Reference Reference Age, years 0–18 45,038 (99.9) 65 (0.1) Reference Reference 19–35 33,999 (99.7) 88 (0.3) 1.81 (1.31–2.49) 0.000 1.14 (0.73–1.79) 0.552 36–64 29,045 (99.5) 138 (0.5) 3.36 (2.48–4.48) 0.000 0.85 (0.53–1.37) 0.513 ≥65 8,161 (98.9) 89 (1.1) 7.72 (5.61–10.63) 0.000 1.96 (1.14–3.35) 0.014 Congregate settings Schools 65,631 (99.8) 109 (0.2) Reference Reference Workplaces 14,265 (99.7) 48 (0.3) 2.06 (1.46–2.89) 0.000 1.03 (0.65–1.63) 0.869 Healthcare facilities 18,217 (99.2) 147 (0.8) 4.90 (3.84–6.31) 0.000 3.34 (2.18–5.13) 0.000 Social welfare facilities 9,714 (99.3) 65 (0.7) 4.03 (2.96–5.48) 0.000 3.14 (1.90–5.18) 0.000 Others 8,416 (99.9) 11 (0.1) 0.79 (0.42–1.46) 0.456 0.60 (0.30–1.20) 0.149 Type of contact Close contact 51,019 (99.6) 206 (0.4) 1.57 (1.28–1.93) 0.000 1.76 (1.41–2.20) 0.000 Casual contact 62,215 (99.7) 160 (0.3) Reference Reference Unknown 3,009 (99.5) 14 (0.5) 1.82 (1.05–3.14) 0.031 1.64 (0.94–2.86) 0.078 LTBI status LTBI and therapy started 5,446 (99.3) 37 (0.7) 0.24 (0.14–0.42) 0.000 0.50 (0.33–0.74) 0.001 LTBI and no therapy 5,987 (98.7) 81 (1.3) Reference Reference NO LTBI 68,779 (99.9) 81 (0.1) 0.08 (0.06–0.11) 0.000 0.12 (0.08–0.17) 0.000 Not testedb 36,037 (99.5) 181 (0.5) 0.35 (0.28–0.45) 0.000 0.40 (0.30–0.53) 0.000 Table 4. Risk factors for incident active tuberculosis among 116,623 contacts of tuberculosis cases in congregate settings. LTBI, latent tuberculosis infection; TB, tuberculosis; CI, confidence interval. aContact diagnosed within 90 days after diagnosis of the index patient bIncluded cases with unknown LTBI findings (n=79) Table 4. Risk factors for incident active tuberculosis among 116,623 contacts of tuberculosis cases in congregate settings. Discussion In conclusion, the present study utilised complete enumeration data obtained from contact investigations conducted in the ROK in 2015 to confirm the development of TB in recent contacts within 2 years and evaluated the effects of the current contact investigations conducted in the ROK. The findings suggest that the completion of chemoprophylaxis for LTBI can lower the risk of TB development in contacts. In future, efforts to increase the contact investigation rates and improve the treatment rates for LTBI are necessary. Furthermore, long-term assessment of the effects of contact investigations is also necessary. Scientific Reports \| (2020) 10:1306 \| https://doi.org/10.1038/s41598-020-57697-1 www.nature.com/scientificreports/ Contacts Without Contacts With Unadjusted Adjusted Incident TB, No. (%) Incident TBa, No. Discussion (%) Hazard Ratio (95% CI) p–value Hazard Ratio (95% CI) p–value Total 11,433 118 Index factors Sex Male 7,687 (98.8) 95 (1.2) 1.99 (1.26–3.15) 0.003 1.91 (1.19–3.07) 0.007 Female 3,746 (99.4) 23 (0.6) Reference Reference Age, years 0–18 2,425 (98.8) 30 (1.2) Reference 19–35 3,491 (99.1) 32 (0.9) 0.75 (0.45–1.23) 0.261 36–64 3,916 (98.9) 45 (1.1) 0.95 (0.59–1.51) 0.832 ≥65 1,601 (99.3) 11 (0.7) 0.56 (0.28–1.13) 0.108 Sputum smear status Smear-positive 7,508 (98.8) 89 (1.2) 1.61 (1.05–2.44) 0.026 Smear-negative 3,902 (99.3) 29 (0.7) Reference Unknown 23 (100.0) 0 (0.0) … Cavities on chest radiograph Cavities 3,648 (98.2) 67 (1.8) 2.64 (1.82–3.81) 0.000 2.58 (1.76–3.77) 0.000 No Cavities 7,128 (99.3) 49 (0.7) Reference Reference Unknown 657 (99.7) 2 (0.3) 0.43 (0.10–1.79) 0.250 0.65 (0.15–2.76) 0.566 Contact factors Sex Male 6,887 (98.9) 78 (1.1) 1.28 (0.87–1.87) 0.200 Female 4,546 (99.1) 40 (0.9) Reference Age, years 0–18 2,336 (98.8) 28 (1.2) Reference Reference 19–35 2,498 (98.8) 31 (1.2) 1.04 (0.62–1.74) 0.859 0.76 (0.41–1.40) 0.382 36–64 6,176 (99.3) 45 (0.7) 0.61 (0.38–0.99) 0.046 0.29 (0.15–0.57) 0.000 ≥65 423 (96.8) 14 (3.2) 2.77 (1.46–5.27) 0.002 1.22 (0.52–2.88) 0.643 Congregate settings Schools 3,912 (98.8) 48 (1.2) Reference Reference Workplaces 2,849 (99.1) 25 (0.9) 0.72 (0.44–1.17) 0.195 0.57 (0.32–1.03) 0.067 Healthcare facilities 2,430 (98.6) 35 (1.4) 1.18 (0.76–1.83) 0.441 1.19 (0.66–2.16) 0.551 Social welfare facilities 1,291 (99.4) 8 (0.6) 0.50 (0.24–1.07) 0.076 0.78 (0.33–1.83) 0.571 Others 951 (99.8) 2 (0.2) 0.17 (0.04–0.71) 0.015 0.17 (0.04–0.75) 0.020 Type of contact Close contact 6,563 (98.8) 78 (1.2) 1.48 (0.99–2.20) 0.053 Casual contact 4,337 (99.2) 35 (0.8) Reference Unknown 533 (99.1) 5 (0.9) 1.17 (0.46–3.00) 0.735 Preventive therapy Did not start 5,987 (98.7) 81 (1.3) Reference Reference Started, did not complete 1,209 (98.2) 22 (1.8) 1.33 (0.83–2.13) 0.231 0.97 (0.60–1.58) 0.920 Completed 4,237 (99.6) 15 (0.4) 0.23 (0.15–0.45) 0.000 0.16 (0.08–0.29) 0.000 Table 5. Risk factors for incident tuberculosis among 11,551 contacts with LTBI in congregate settings. LTBI, latent tuberculosis infection; TB, tuberculosis; CI, confidence interval. aContact diagnosed within 90 days after diagnosis of the index patient. Table 5. Risk factors for incident tuberculosis among 11,551 contacts with LTBI in congregate settings. LTBI, latent tuberculosis infection; TB, tuberculosis; CI, confidence interval. aContact diagnosed within 90 days after diagnosis of the index patient. www.nature.com/scientificreports/ www.nature.com/scientificreports/ PT completed PT Started, did not complete Did not start PT TB casesa Person-Years of Observation, No. Incidence Rate/100,000b (95% CI) TB casesa Person-Years of Observation, No. Incidence Rate/ 100,000b (95% CI) TB casesa Person-Years of Observation, No. Incidence Rate/ 100,000b (95% CI) Total 15 12,577 119 (58.9–179.6) 22 3,594 612 (356.4–867.9) 81 17,594 460 (360.1–560.6) Age, years 0–35 8 8,187 98 (30.0–165.4) 14 1,827 766 (364.9–1,168.0) 37 4,395 842 (570.6–1,113.0) ≥36 7 4,390 159 (41.3–277.6) 8 1,767 453 (139.1–766.5) 44 13,199 333 (234.9–431.9) Table 6. Incidence rate of tuberculosis among contacts treated by preventive treatment for LTBI in congregate settings. PT, preventive therapy; LTBI, latent tuberculosis infection; TB, tuberculosis; CI, confidence interval. aIncident TB = TB diagnosed within 90 days after diagnosis of the index patient. bIncidence rate in cases per 100,000 person-years observation. Table 6. Incidence rate of tuberculosis among contacts treated by preventive treatment for LTBI in congrega settings. PT, preventive therapy; LTBI, latent tuberculosis infection; TB, tuberculosis; CI, confidence interval aIncident TB = TB diagnosed within 90 days after diagnosis of the index patient. bIncidence rate in cases per 100,000 person-years observation. Table 6. Incidence rate of tuberculosis among contacts treated by preventive treatment for LTBI in congregate settings. PT, preventive therapy; LTBI, latent tuberculosis infection; TB, tuberculosis; CI, confidence interval. aIncident TB = TB diagnosed within 90 days after diagnosis of the index patient. bIncidence rate in cases per 100,000 person-years observation. Contact investigation. According to the Tuberculosis Prevention Act in the ROK, physicians who diagnose TB in private healthcare institutions or public health centres should report this to the public health authority through the KNTSS. TB management staff at public health centres and TB management nurses at private health- care institutions then conduct case investigations to confirm whether the patients are associated with congregate settings and conduct contact investigation for such cases. KTNESS operates on a web - basis and consists of a patient reporting database, database for TB epidemiological investigation in congregate settings, and households contact investigation database In accordance with the 2015 national guidelines for TB management, contact investigation was conducted when respiratory TB cases with a positive acid-fast bacilli smear test or culture test were reported to be associated with congregate settings, or when more than two TB cases were reported in the same congregate setting within 6 months, regardless of the findings in respiratory specimens20. www.nature.com/scientificreports/ On the basis of on-site investigations, contacts were classified as close or casual contacts; chest radiography and LTBI screening were prioritised for close contacts. Close contacts were those who used the same closed indoor spaces and had direct contact with the index cases for prolonged periods of times. LTBI screening involved a tuberculin skin test (TST) or an interferon-gamma release assay (IGRA). A positive TST result is defined as an induration of ≥10 mm (≥5 mm in newborns that had not yet received the BCG vaccine). The IGRA was performed using the QuantiFERON-TB Gold In-tube (QTF) test, and a value of 0.35 international units or more was deemed positive20. Definitions and risk factors. ‘Index patient’ was defined as the first person with confirmed TB within the congregate settings. When contacts were reported to have developed TB within 90 days after the initial report of TB in the index case, the disease was considered to be ‘co-prevalent’. When contacts were reported to have developed TB after 90 days of the report of active TB in the index case, the disease was considered ‘incident’29. Risk factors related to the index cases included sex, age, sputum smear microscopy results, and chest radiograph results, while those for contacts included sex, age, the congregate setting, the type of contact, the presence of LTBI, and treatment for LTBI. Congregate settings were categorized into school, workplace, healthcare facility, social welfare facility, and others. Study population. Overall, 133,423 contacts of 2,763 index cases were extracted from the KNTSS. The following participants were excluded: 15,271 contacts of 152 index cases whose contact investigations initiated before 2015 and continued into 2015, 1,178 contact cases previously diagnosed with TB, 232 household contacts confirmed during the study period, 2 index cases and 2 contacts with duplicate registration. The final sample included 116, 742 contacts of 2,609 index cases (Fig. 1). Of the index cases, 2,594 (99.4%) were pulmonary tuberculosis. Ethics approval and consent to participate. The need for written informed consent from participants was waived, based on the Korean Infectious Disease Control and Prevention Act (No.4). This study was con- ducted in accordance with Korean Infectious Disease Control and Prevention Act and Tuberculosis Prevention Act with permission of KCDC. The study design was approved by the Institutional Review Board (IRB) of KCDC. www.nature.com/scientificreports/ As this is a retrospective study on the existing data on subjects and based on the Korean Infectious Disease Control and Prevention Act (No.4), we received confirmation of the written consent exemption from IRB. Statistical analyses. The Kaplan–Meier method and log-rank tests were used to compare TB incidences according to the contacts’ age. Demographic, laboratory, and clinical determinants (both index cases and contact-related cases) of co-prevalent TB were identified using logistic regression, while those of incident TB were analysed using Cox proportional hazards regression. Multivariate analyses with backward elimination were conducted for the sex and age of index cases and contacts as well as variables with a p-value of <0.05 in univariate analyses. Statistical significance was identified with a 95% confidence interval and a P-value < 0.05. All statistical analyses were performed using SAS version 9.4 (SAS Institute Inc) and Epi-infotm (CDC, Atlanta, GA, USA). Study design and participants. Th d b Study design and participants. This retrospective cohort study was conducted for an average 2.9 years form January 2015 to May 2018. The database for TB epidemiological investigation in congregate settings of the Korean National Tuberculosis Surveillance System (KNTSS), operated by the Korea Centers for Disease Control and Prevention, and the database for reported TB cases were used in this study. Of TB cases in congregate settings, reported to the KNTSS between January 2015 and December 2015, 2,763 cases for whom contact investigations were conducted according to the national guidelines on TB management were identified. To address potential sources of bias, contacts previously diagnosed with TB and contact investigations initiated before 2015 that continued into 2015 were excluded. The household contacts who were confirmed during the study period were also excluded. The timing of TB in contacts was defined as the time between the date on which the index TB case, present in a congregate setting, was reported and the date on which a contact of this individual was reported as an active TB case. Follow-up was continued until active TB was reported in contacts, and until May 31, 2018, in contacts who did not develop TB. Scientific Reports \| (2020) 10:1306 \| https://doi.org/10.1038/s41598-020-57697-1 References Achievements in and challenges of tuberculosis control in South Korea. Emerg. Infect. Dis. 21, 1913–1920 (2015). 22. Korea Centers for Diseases Control and Prevention. 7th Korea National Health and Nutrition Examination Survey 1st year (2016) tuberculin survey support and quality control. KCDC, 2017.h 22. Korea Centers for Diseases Control and Prevention. 7th Korea National Health and Nutrition Examination Survey 1st year (2 tuberculin survey support and quality control KCDC 2017 tuberculin survey support and quality control. KCDC, 2017.h y pp q y 23. Noh, J. The diabetes epidemic in Korea. Endocrinol. Metab. 31, 349–353 (2016). h 24. Park, J. I., Baek, H. & Jung, H. H. Prevalence of chronic kidney disease in Korea: The Korean National Health and Nutritional Examination Survey 2011–2013. J. Korean Med. Sci. 31, 915–923 (2016). Examination Survey 2011–2013. J. Korean Med. Sci. 31, 915–923 f l h d lf l l h y 25. Korean Ministry of Health and Welfare. Korea National Health and Nutrition Examination Survey. Seoul: Korean Ministry of H and Welfare, 2017. 6. Bailey, W. C. et al. Predictive model to identify positive tuberculosis skin test results during contact investigations. JAMA. 287 996–1002 (2002). 27. National Tuberculosis Controllers Association; Centers for Disease Control and Prevention (CDC). Guidelines for the investigation of contacts of persons with infectious tuberculosis. Recommendations from the National Tuberculosis Controllers Association and CDC. MMWR Recomm. Rep., 54, 1–47 (2005). p 8. Erkens, C. G. et al. Risk of developing tuberculosis disease among persons diagnosed with latent tuberculosis infection in the Netherlands. Eur. Respir. J. 48, 1420–1428 (2016). p 9. Fox, G. J., Barry, S. E., Britton, W. J. & Marks, G. B. Contact investigation for tuberculosis: a systematic review and meta-analysis. Eur Respir. J. 41, 140–156 (2013). Acknowledgementsh The authors would like to thank, So Jung Kim, Jieun Kim, Eun Jung Lee, Yun Choi, Yejin Kim, Hye Sung Kim, Ji Eun Park, Ji-eun Yun, Jin A. Park, Soo Jin Park, Yoonsun Kim, Sanghee Kang, Eunyoung Kim, Yoon Jung Seo, Aejeong Lee,Young Sun Choi, Seo Yeon Hong, Daseul Kim, Eunbi Kim, Hee Sun Ahn, Seongeun Kim,, Jiyoung Kim, Cheonggi Hong, Yujin Jang, Youngseok Park for administrative and technical support. Special thanks to Prof. Ji-Hyuk Park, Prof. Yoon Soo Park, Prof. Moon-Hyun Chung, Prof. Byung-Chul Chun, Dr. Ok Park, Dr. Sangwon Lee for thoughtful consultancy. Data availability All data extracted in this study are included in this article. Received: 9 September 2019; Accepted: 24 December 2019; Published: xx xx xxxx Scientific Reports \| (2020) 10:1306 \| https://doi.org/10.1038/s41598-020-57697-1 www.nature.com/scientificreports/ Competing interestsh p g The authors declare no competing interests. References 1. World Health Organization. Global tuberculosis report 2018. Available at: http://www.who.int/tb/publications/global_report/en/ Accessed February 28, 2019. 1. World Health Organization. Global tuberculosis report 2018. Available at: http://www.who.int/tb/publications/global_report/en/. Accessed February 28, 2019.h y 2. Houben, R. M. & Dodd, P. J. The global burden of latent tuberculosis infection: a re-estimation using mathematical modelling. PLoS medicine 13, e1002152 (2016).i ( ) 3. Korea Centers for Diseases Control and Prevention. 2018 Annual report on the notified tuberculosis in Korea 2018. KCDC, 2019 Available at: http://tbzero cdc go kr/tbzero/board/boardList do?leftMenuId=48&paramMenuId=109&menuIdx1=0&menuIdx2=0 3. Korea Centers for Diseases Control and Prevention. 2018 Annual report on the notified tuberculosis in Korea 2018. KCDC, 2019. Available at: http://tbzero.cdc.go.kr/tbzero/board/boardList.do?leftMenuId=48&paramMenuId=109&menuIdx1=0&menuIdx2=0 4. World Health Organization. Updated and consolidated guidelines for programmatic management 2018. Available at: https://www. who.int/tb/publications/2018/latent-tuberculosis-infection/en/. gt 4. World Health Organization. Updated and consolidated guidelines for programmatic management 2018. Available at: https://www. who.int/tb/publications/2018/latent-tuberculosis-infection/en/. t 4. World Health Organization. Updated and consolidated guidelines for programmatic management 2018. Available at: https://www who.int/tb/publications/2018/latent-tuberculosis-infection/en/. who.int/tb/publications/2018/latent-tuberculosis-infection/en/. p 5. Centers for Disease Control and Prevention. Targeted tuberculin testing and treatment of latent tuberculosis infection. MMWR. 49 1–51 (2000).h 6. Styblo, K. The relationship between the risk of tuberculous infection and the risk of developing infectious tuberculosis. Bull. Int Union Tuberc. Lung Dis. 60, 117–119 (1985). g 7. Getahun, H. et al. Management of latent Mycobacterium tuberculosis infection: WHO guidelines for low tuberculosis burden countries. Eur. Respir. J. 44, 1563–1576 (2015). p 8. Korea Centers for Diseases Control and Prevention. Report on the contact investigations at congregated settings 2017. KCDC, 2018 Available at: https://www.cdc.go.kr/board.es?mid=a20501000000&bid=0015 p 8. Korea Centers for Diseases Control and Prevention. Report on p 8. Korea Centers for Diseases Control and Prevention. Report on the contact investigations at Available at: https://www.cdc.go.kr/board.es?mid=a20501000000&bid=0015 p g 9. Lee, M. R. et al. Tuberculosis contact investigation in an intermediate burden setting: implications from a large tuberculosis contact cohort in Taiwan. Eur. Respir. J. 50, pii: 1700851 (2017). cohort in Taiwan. Eur. Respir. J. 50, pii: 1700851 (2017). p p 10. Lee, M. et al. Early and late tuberculosis risks among close contacts in Hong Kong. Int. J. Tuberc. Lung Dis. 12, 281–287 (2008) h l l d k f f f f h ld h h ld h d l h p uberculosis risks among close contacts in Hong Kong. Int. J. Tuberc. Lung Dis. 12, 281–287 (2008). 10. Lee, M. et al. Early and late tuberculosis risks among clo 11. References Singh, M. et al. Prevalence and risk factors for transmission of infection among children in household contact with adults ha pulmonary tuberculosis. Arch. Dis. Child 90, 624–628 (2005).i 12. Anger, H. A. et al. Active case finding and prevention of tuberculosis among a cohort of contacts exposed to infectious tubercu cases in New York City. Clin. Infect. Dis. 54, 1287–1295 (2012).fft 12. Anger, H. A. et al. Active case finding and prevention of tuber i cases in New York City. Clin. Infect. Dis. 54, 1287–1295 (2012)ff y f ( ) 3. Sloot, R., Schim van der Loeff, M. F., Kouw, P. M. & Borgdorff, M. W. Risk of tuberculosis after recent exposure. A 10-year follow-up study of contacts in Amsterdam. Am. J. Respir. Crit. Care Med. 190, 1044–1052 (2014). 13. Sloot, R., Schim van der Loeff, M. F., Kouw, P. M. & Borgdorff, M. W. Risk of tuberculosis after r study of contacts in Amsterdam. Am. J. Respir. Crit. Care Med. 190, 1044–1052 (2014). y p 14. Reichler, M. R. et al. Risk and timing of tuberculosis among close contacts of persons with infectious tuberculosis. J. Infect. Dis. 218, 1000–1008 (2018). 15. Kwon, Y. et al. Results of tuberculosis contact investigation in congregate settings in Korea, 2013. Osong Public Health Res. Perspect. 5, S30–S36 (2014). ( ) 16. Moran-Mendoza, O., Marion, S., Elwood, K., Patrick, D. & FitzGerald, J. Risk factors for developing tuberculosis: a 12-year fo up of contacts of tuberculosis cases. Int. J. Tuberc. Lung Dis. 14, 1112–1119 (2010). Moran-Mendoza, O., Marion, S., Elwood, K., Patrick, D. & FitzGerald, J. Risk factors for developing tuberculosis: a 12-year follow- p of contacts of tuberculosis cases. Int. J. Tuberc. Lung Dis. 14, 1112–1119 (2010). Feng, J. Y. et al. Is 1-year follow-up adequate for adult tuberculosis c 17. Feng, J. Y. et al. Is 1-year follow-up adequate for adult tuberculosis contacts? Eur. Respir. J. 45, 1501–1504 (2015). 18. Guo, J. et al. High incidence and low case detection rate among contacts of tuberculosis cases in Shanghai, China. BMC Infec 19, 320 (2019). ( ) 19. Ferebee, S. H. Controlled chemoprophylaxis trials in tuberculosis. A general review. Adv. Tuberc. Res. 17, 29–106 (1969). rebee, S. H. Controlled chemoprophylaxis trials in tuberculosis. A g 20. Korea Centers for Diseases Control and Prevention. National Tuberculosis Management Guidelines 2015. KCDC, 2015. A at: https://www.cdc.go.kr/board.es?mid=a20507020000&bid=0019 p g 1. Kim, J. H. & Yim, J. J. Author contributions S.Y.P. contributed to the study design, manuscript writing and data analysis. S.H., Y.K., J.K., S.L. and J.Y. helped with data analysis and preparation of the tables. U.N.K. and M.S.P. equally contribute in study design and critically reviewed manuscript. All authors reviewed and approved the manuscript for submission. Additional information Supplementary information is available for this paper at https://doi.org/10.1038/s41598-020-57697-1. Scientific Reports \| (2020) 10:1306 \| https://doi.org/10.1038/s41598-020-57697-1 www.nature.com/scientificreports/ www.nature.com/scientificreports/ Correspondence and requests for materials should be addressed to U.-N.K. or M.-s.P. Reprints and permissions information is available at www.nature.com/reprints. Reprints and permissions information is available at www.nature.com/reprints. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Cre- ative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not per- mitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. © The Author(s) 2020 © The Author(s) 2020 Scientific Reports \| (2020) 10:1306 \| https://doi.org/10.1038/s41598-020-57697-1 10 10
https://openalex.org/W4387948737	https://bmcplantbiol.biomedcentral.com/counter/pdf/10.1186/s12870-023-04519-x	English	null	QTL mapping reveals key factors related to the isoflavone contents and agronomic traits of soybean (Glycine max)	BMC plant biology	2,023	cc-by	14,210	© The Author(s) 2023. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecom‑ mons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. QTL mapping reveals key factors related to the isoflavone contents and agronomic traits of soybean (Glycine max) Jung Min Kim1, Ji Su Seo1,2, Jeong Woo Lee1,2, Jae Il Lyu3, Jaihyunk Ryu1, Seok Hyun Eom4, Bo‑Keun Ha2* and Soon‑Jae Kwon1* BMC Plant Biology BMC Plant Biology Kim et al. BMC Plant Biology (2023) 23:517 https://doi.org/10.1186/s12870-023-04519-x Open Access Abstract Keywords Soybean, Isoflavone, Single nucleotide polymorphism (SNP), Radiation breeding, Quantitative trait locus (QTL) Correspondence: Bo‑Keun Ha bkha@chonnam.ac.kr Soon‑Jae Kwon soonjaekwon@kaeri.re.kr Full list of author information is available at the end of the article Correspondence: Bo‑Keun Ha bkha@chonnam.ac.kr Soon‑Jae Kwon soonjaekwon@kaeri.re.kr Full list of author information is available at the end of the article Bo‑Keun Ha bkha@chonnam.ac.kr Soon‑Jae Kwon soonjaekwon@kaeri.re.kr Full list of author information is available at the end of the article Abstract Background Soybean is a valuable source of edible protein and oil, as well as secondary metabolites that can be used in food products, cosmetics, and medicines. However, because soybean isoflavone content is a quantitative trait influenced by polygenes and environmental interactions, its genetic basis remains unclear. Results This study was conducted to identify causal quantitative trait loci (QTLs) associated with soybean isoflavone contents. A mutant-based F2 population (190 individuals) was created by crossing the Korean cultivar Hwanggeum with low isoflavone contents (1,558 µg g−1) and the soybean mutant DB-088 with high isoflavone contents (6,393 µg g−1). A linkage map (3,049 cM) with an average chromosome length of 152 cM was constructed using the 180K AXIOM® SoyaSNP array. Thirteen QTLs related to agronomic traits were mapped to chromosomes 2, 3, 11, 13, 19, and 20, whereas 29 QTLs associated with isoflavone contents were mapped to chromosomes 1, 3, 8, 11, 14, 15, and 17. Notably, the qMGLI11, qMGNI11, qADZI11, and qTI11, which located Gm11_9877690 to Gm11_9955924 interval on chromosome 11, contributed to the high isoflavone contents and explained 11.9% to 20.1% of the phenotypic var‑ iation. This QTL region included four candidate genes, encoding β-glucosidases 13, 14, 17–1, and 17–2. We observed significant differences in the expression levels of these genes at various seed developmental stages. Candidate genes within the causal QTLs were functionally characterized based on enriched GO terms and KEGG pathways, as well as the results of a co-expression network analysis. A correlation analysis indicated that certain agronomic traits (e.g., days to flowering, days to maturity, and plant height) are positively correlated with isoflavone content. Conclusions Herein, we reported that the major QTL associated with isoflavone contents was located in the inter‑ val from Gm11_9877690 to Gm11_9955924 (78 kb) on chromosome 11. Four β-glucosidase genes were identified that may be involved in high isoflavone contents of soybean DB-088. Thus, the mutant alleles from soybean DB-088 may be useful for marker-assisted selection in developing soybean lines with high isoflavone contents and superior agronomic traits. Background secondary metabolite contents are generally less herit- able than the contents of other major compounds, but the high heritability of specific isoflavone contents has been reported [20–24]. Accordingly, QTL analyses may be applicable for these isoflavone forms in soybean seeds. Soybean contains functional compounds, such as pri- mary (protein and oil) and secondary (tocopherol, sapo- nin, and isoflavone) metabolites, and is an important cash crop because it is used for the production of food and feed, cosmetics, and biofuels [1]. Isoflavones, which are produced exclusively by species in the family Fabaceae (e.g., soybean), contribute to plant defense systems by serving as precursors of glyceollins and major phyto- alexins as well as signaling molecules mediating plant– microbe interactions in the rhizosphere [2, 3]. Most isoflavones (80%–90%) are localized in cotyledons (700– 5,300 µg g−1), among which malonyl-glucoside isoforms account for the largest proportion (60%–80%) [4–8]. The following 12 isoflavone forms have been identified in soy- bean seeds: daidzein, glycitein, genistein, daidzin, glyci- tin, genistin, 6-O-malonyl daidzin, 6-O-malonyl glycinin, 6-O-malonyl genistin, 6-O-acetyl daidzin, 6-O-acetyl glycinin, and 6-O-acetyl genistin [9]. Although almost all legume crops contain isoflavones, the isoflavone con- tent in soybean is more than 100-fold greater than that in other legumes, including chickpea, faba bean, len- til, and peanut [10]. Isoflavone biosynthesis is initiated through the phenylpropanoid pathway, with the resulting compounds subsequently converted to their conjugated forms (i.e., glucosyl- and malonyl-glucosides) and then stored in vacuoles [11]. Isoflavones form a subgroup of phytoestrogens that help to prevent breast cancer, osteo- porosis, heart disease, and menopausal symptoms [12]. Because of their positive effects on human health due to their antioxidant and anti-tumorigenic activities, there is increasing interest in soybean isoflavones worldwide [13]. l During the past decade, molecular markers have been used to construct linkage maps for QTL analyses. Many researchers have tried to reveal the genetic factors under- lying complex traits using various molecular markers, including restriction fragment length polymorphisms, simple sequence repeats (SSRs), and single nucleotide polymorphisms (SNPs) [25–27]. The first linkage map- ping study for isoflavones, which was conducted by Njiti et al. [28], involved the RIL mapping population of ‘Essex’ × ‘Forrest’ and 201 SSR markers. Meksem et al. [29] constructed a linkage map (2,823 cM) with an aver- age marker spacing of 26 cM using 133 SSR markers and the RIL population used by Njiti et al. [28]. Background They detected major QTLs on linkage groups A1, B1, and N. Primomo et al. [30] detected 17 isoflavone-related QTLs on 11 link- age groups using 207 RILs from ‘AC756’ × ‘RCAT Angora’ and 99 SSR markers. Zeng et al. [31] constructed a link- age map (2,020 cM) using 130 RILs and 125 SSR markers from multiple environments and detected novel QTLs (QDZF_1, QGTF_1, and QTIF_1) on linkage group F in Satt144. Gutierrez-Gonzalez et al. [32] created a linkage map (2,151 cM) comprising five QTLs using 446 SNP markers and five SSR markers and 188 RILs derived from ‘Magellan’ × ‘PI 437654’. In the past, low-density linkage maps were constructed because of insufficient polymorphic markers and tech- nological limitations, resulting in large gaps between markers. However, recent advances in next-generation sequencing technologies, including genotyping-by- sequencing, restriction-site-associated DNA sequenc- ing (RAD-seq), and specific locus amplified fragment sequencing (SLAF-seq), have facilitated the identifica- tion of many SNPs relevant for the high-density geno- typing of crops [33–36]. On the basis of SLAF-seq and RAD-seq analyses, Li et al. [37], Pei et al. [38], and Cai et al. [39] detected novel QTLs (qIF20-2, aG8, qMD19, qMG18, qTIF19, and qIF5-1) that explained 4.2%– 59.95% of the phenotypic variation (PV) and included isoflavone-related genes and transcription factor genes. The complete soybean genome sequence was success- fully assembled in 2010 [40], which accelerated the development of high-throughput SNP chip platforms, including SoySNP6K Illumina Infinium BeadChip [41], SoySNP50K Illumina Infinium BeadChip [42], 180K AXIOM® SoyaSNP array [43], NAJU 355K SoySNP array [44], and SoySNP618K array [45]. Kim et al. [46] identi- fied a novel GmSACPD-C allele using the RIL popula- tion derived from a high stearic acid soybean mutant, g yl Despite their importance, determining the genetic basis of isoflavone biosynthesis and accumulation has been difficult, with previous studies having limited suc- cess in accurately identifying the associated quantitative trait loci (QTLs). The isoflavone content is considered to be a quantitative trait primarily influenced by minor QTLs. Earlier research revealed the isoflavone content may vary by up to 3-fold depending on the effects of the genotype, environment, and genotype × environment interaction [14–16]. Linkage mapping is a typical method used to identify genomic loci related to complex quali- tative/quantitative traits, with the recombination fre- quency of bi-parental populations (e.g., F2, DH, RIL, and NIL populations) used to identify causal alleles [17, 18]. [30] detected 17 isoflavone-related QTLs on 11 link- age groups using 207 RILs from ‘AC756’ × ‘RCAT Angora’ and 99 SSR markers. Zeng et al. [31] constructed a link- age map (2,020 cM) using 130 RILs and 125 SSR markers from multiple environments and detected novel QTLs (QDZF_1, QGTF_1, and QTIF_1) on linkage group F in Satt144. Gutierrez-Gonzalez et al. [32] created a linkage map (2,151 cM) comprising five QTLs using 446 SNP markers and five SSR markers and 188 RILs derived from ‘Magellan’×‘PI 437654’ © The Author(s) 2023. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecom‑ mons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Kim et al. BMC Plant Biology (2023) 23:517 Kim et al. BMC Plant Biology (2023) 23:517 Page 2 of 18 Page 2 of 18 Background Soybean contains functional compounds, such as pri- mary (protein and oil) and secondary (tocopherol, sapo- nin, and isoflavone) metabolites, and is an important cash crop because it is used for the production of food and feed, cosmetics, and biofuels [1]. Isoflavones, which are produced exclusively by species in the family Fabaceae (e.g., soybean), contribute to plant defense systems by serving as precursors of glyceollins and major phyto- alexins as well as signaling molecules mediating plant– microbe interactions in the rhizosphere [2, 3]. Most isoflavones (80%–90%) are localized in cotyledons (700– 5,300 µg g−1), among which malonyl-glucoside isoforms account for the largest proportion (60%–80%) [4–8]. The following 12 isoflavone forms have been identified in soy- bean seeds: daidzein, glycitein, genistein, daidzin, glyci- tin, genistin, 6-O-malonyl daidzin, 6-O-malonyl glycinin, 6-O-malonyl genistin, 6-O-acetyl daidzin, 6-O-acetyl glycinin, and 6-O-acetyl genistin [9]. Although almost all legume crops contain isoflavones, the isoflavone con- tent in soybean is more than 100-fold greater than that in other legumes, including chickpea, faba bean, len- til, and peanut [10]. Isoflavone biosynthesis is initiated through the phenylpropanoid pathway, with the resulting compounds subsequently converted to their conjugated forms (i.e., glucosyl- and malonyl-glucosides) and then stored in vacuoles [11]. Isoflavones form a subgroup of phytoestrogens that help to prevent breast cancer, osteo- porosis, heart disease, and menopausal symptoms [12]. Because of their positive effects on human health due to their antioxidant and anti-tumorigenic activities, there is increasing interest in soybean isoflavones worldwide [13]. Despite their importance, determining the genetic basis of isoflavone biosynthesis and accumulation has been difficult, with previous studies having limited suc- cess in accurately identifying the associated quantitative trait loci (QTLs). The isoflavone content is considered to be a quantitative trait primarily influenced by minor QTLs. Earlier research revealed the isoflavone content may vary by up to 3-fold depending on the effects of the genotype, environment, and genotype × environment interaction [14–16]. Linkage mapping is a typical method used to identify genomic loci related to complex quali- tative/quantitative traits, with the recombination fre- quency of bi-parental populations (e.g., F2, DH, RIL, and NIL populations) used to identify causal alleles [17, 18]. Although the F2 generation undergoes only a single mei- otic cycle, which makes linkage mapping challenging, it is useful for quickly and easily generating populations [19]. Additionally, QTL mapping using an F2 population is an efficient way to analyze qualitative traits, such as disease resistance, as well as quantitative traits, including metab- olite contents (e.g., protein, oil, and sucrose). However, secondary metabolite contents are generally less herit- able than the contents of other major compounds, but the high heritability of specific isoflavone contents has been reported [20–24]. Accordingly, QTL analyses may be applicable for these isoflavone forms in soybean seeds. During the past decade, molecular markers have been used to construct linkage maps for QTL analyses. Many researchers have tried to reveal the genetic factors under- lying complex traits using various molecular markers, including restriction fragment length polymorphisms, simple sequence repeats (SSRs), and single nucleotide polymorphisms (SNPs) [25–27]. The first linkage map- ping study for isoflavones, which was conducted by Njiti et al. [28], involved the RIL mapping population of ‘Essex’ × ‘Forrest’ and 201 SSR markers. Meksem et al. [29] constructed a linkage map (2,823 cM) with an aver- age marker spacing of 26 cM using 133 SSR markers and the RIL population used by Njiti et al. [28]. They detected major QTLs on linkage groups A1, B1, and N. Primomo et al. Mapping populationh The soybean mutant DB-088, which has high isoflavone contents, was selected from the MDP consisting of 208 soybean mutants at the Radiation Breeding Farm of the Korea Atomic Energy Research Institute (KAERI; 35.5699°N 126.9772°E, Jeongeup, Jeollabuk, Republic of Korea) [52]. Briefly, 1,000 seeds were collected from seven soybean cultivars and irradiated with 60Co gamma rays (250 Gy) at KAERI in 2008. We subsequently propa- gated 1,695 mutants from the M1 generation to the M5 generation via single seed descent. In the M5 generation, we selected mutants with distinct morphological char- acteristics and agronomic traits (e.g., high yield, growth type, and climate adaptability). We then constructed a mutant population consisting of 208 genotypes without redundant phenotypes and propagated it until the M10 generation; the population was designated as the MDP. The isoflavone contents were measured in the M10 gener- ation as well as in the M11 and M12 generations to deter- mine the year-to-year variation over a 2-year period. Finally, DB-088 was selected following the gamma irra- diation of Danbaek [55]. g y Mutation breeding is an effective strategy for develop- ing new cultivars with ideal agronomic traits and func- tional compounds. Gamma-rays are one of the most commonly used physical mutagens (60%) for muta- tion breeding [50]. More than 3,402 mutant varieties of 210 species are registered in the International Atomic Energy Agency/Mutant Variety Database (http://www. iaea.org/). We previously created a mutant diversity pool (MDP) comprising 208 mutant soybean lines from seven wild-type cultivars that were irradiated with 60Co gamma rays (250 Gy). We subsequently performed target region amplification polymorphism (TRAP) and trans- posable element (TE)-TRAP marker analyses to clar- ify genetic associations as well as gene expression and GWAS analyses of agronomic traits and phytochemical contents using our MDP and selected a soybean mutant with high isoflavone contents (average of 7,351 μg g−1 over a 3-year period) [51–53]. Because there has been relatively little breeding for isoflavone contents, a QTL analysis was conducted using only a few varieties (e.g., Angora, Luheidou2, Wayao, and Zhongdou 27) with iso- flavone contents of 2,246–3,791 μg g−1 as the parental plants [30, 37, 39, 54]. To construct the mutant-based mapping popula- tion, F1 plants were generated by crossing DB-088 with Hwanggeum in the soybean breeding field of Chon- nam National University (36°17′N, 126°39′E, Gwangju, Republic of Korea). Background Although the F2 generation undergoes only a single mei- otic cycle, which makes linkage mapping challenging, it is useful for quickly and easily generating populations [19]. Additionally, QTL mapping using an F2 population is an efficient way to analyze qualitative traits, such as disease resistance, as well as quantitative traits, including metab- olite contents (e.g., protein, oil, and sucrose). However, Kim et al. BMC Plant Biology (2023) 23:517 Kim et al. BMC Plant Biology (2023) 23:517 Page 3 of 18 mutant-based F2 population was phenotyped and geno- typed using the 180K AXIOM® SoyaSNP array for the fine-mapping and QTL analyses. while also completing the genotyping for QTL map- ping using the SoySNP6K platform and a KASP assay. Vuong et al. [47] genotyped 553 soybean plants using SoySNP50K and revealed 14 QTLs containing 60 SNPs associated with the resistance to the soybean cyst nem- atode through a GWAS. Lee et al. [48] genotyped 430 soybean lines from a core collection using the 180K AXIOM® SoyaSNP array and determined that 14 novel SNPs responsible for soya-saponin biosynthesis are closely related to Glyma.07g524600 through a GWAS. Hu et al. [49] identified qPDH1 as a pod dehiscence- related QTL that includes Glyma09g06290 (bHLH tran- scription factor) on chromosome 16 using 211 soybean accessions and the NAJU 355K SoySNP array. Li et al. [48] performed a GWAS and combined the results with QTL mapping data and identified candidate genes (E2, GmPRR3b, and GmVIP5) associated with the flowering time and the circadian clock in 265 soybean landraces using SoySNP618K. Mapping populationh Next, the SNPs were filtered by screening for a lack of identical loci and genetic distor- tion according to locus similarity and the Chi-squared test, respectively, for the linkage mapping analysis. The genetic position was estimated on the basis of the recom- bination frequency using Kosambi’s mapping function. Using MapQTL v6.0 (Kyazma), the map file, genotype file, and phenotype file based on the maximum logarithm of the odds (LOD) scores were determined to identify the significant QTLs using 1,000 permutations with a significance level of 0.05. On the basis of the permuta- tions, the LOD score was set as a fixed value to detect the presence of QTLs across all chromosomes. The ini- tial interval mapping (IM) and multiple QTL mapping (MQM) methods were used for QTL mapping. Using a map file, including genotypes and genetic positions, and the genetic linkage map with LOD peaks, the SNP posi- tions in QTLs were reorganized with MapChart v2.2. Markers were named using the following format: q (i.e., QTL) + chromosome number + trait name + order. Infor- mation regarding the QTLs was obtained from SoyBase (https://www.soybase.org/) and Phytozome (https:// phytozome-next.jgi.doe.gov/). Construction of a genetic linkage map and QTL mappingi The contents of each isoflavone and the total isoflavone (TI) content were quantified as previously described [56]. Ten randomly selected seeds were collected for each soybean plant to measure the contents of the following twelve isoflavone forms: daidzein (DZE), glycitein (GLE), genistein (GNE), daidzin (DZI), glycitin (GLI), genistin (GNI), malonyl daidzin (MDZI), malonyl glycitin (MGLI), malonyl genistin (MGNI), acetyl daidzin (ADGI), acetyl glycitin (AGLI), and acetyl genistin (AGNI). Briefly, the soybean seeds were ground to a powder, after which 1 ml 80% (v/v) methanol was added to 20 mg ground material for the 24-h extraction at room temperature in a shaking incubator (150 rpm). The extracts were filtered using 0.22 µm PTFE syringe filters prior to the ultra-performance liquid chromatography (UPLC) analysis. Twelve isofla- vone standards were purchased from LC Laboratories (Woburn, MA, USA) and Nacalai Tesque (Kyoto, Japan) to identify and quantify the extracted isoflavones. Descriptive statistics and visualization of phenotypic data Descriptive statistics including minimum, maximum, mean, standard deviation (SD), skewness, kurtosis, and coefficient of variation were calculated using Microsoft EXCEL 2016. Mapping populationh A total of 190 F2 plants were grown in plastic pots (15 cm diameter) filled with commercial soil (Hungnong, Pyeongtaek, Republic of Korea) at the Radia- tion Breeding Research Farm of KAERI for the produc- tion and harvest of mature seeds. Examination of agronomic traits and isoflavone contents The agronomic traits of the 190 F2 plants in the map- ping population were evaluated at the Radiation Breed- ing Research Greenhouse of KAERI. More specifically, the following 13 agronomic traits were characterized according to the Korea Seeds & Variety Service standard guidelines (http://www.seed.go.kr/seed/192/subview.do): growth habit (GH; D: determinate, SD: semi-determinate, and ID: indeterminate), plant type (PT; SU: semi-upright, M: medium, and H: horizontal), leaflet shape (LS; O: ovoid and HS: heart-shaped), pod color (PC; BR: brown, DBR: dark brown, and B: black), seed color (SC; YL: yel- low and GY: greenish yellow), hilum color (HC; YL: yel- low and DBR: dark brown), days to flowering (DF), days to maturity (DM), plant height (PH; cm), leaf width (LW; The objective of this study was to address the above- mentioned problems by (1) narrowing the large confi- dence interval of QTLs underlying isoflavone contents; (2) eliminating the genetic bottleneck by using a mutant- based mapping population; and (3) increasing the low isoflavone contents of the Korean soybean cultivar Hwanggeum, which has superior agronomic traits. An F2 mapping population consisting of 190 plants was con- structed by crossing the paternal parent DB-088 with the maternal parent Hwanggeum. To reveal the genetic basis of the ideal agronomic traits and isoflavone contents, the Kim et al. BMC Plant Biology (2023) 23:517 Page 4 of 18 cm), leaf length (LL; cm), number of nodes (NN), and hundred-seed weight (HSW; g). Construction of a genetic linkage map and QTL mapping A total of 26,760 SNP markers were filtered on the basis of a lack of heterozygous alleles and non-polymorphism between the parental plants. To construct the linkage map, polymorphic SNP markers were modified to the following input formats according to the JoinMap v4.1 (Kyazma, Wageningen, Netherlands) instructions: a, maternal allele; h, hetero allele; and b, paternal allele. The JoinMap v4.1 software was used to filter and identify the significant SNPs and genetic linkage according to physi- cal positions. All SNPs that were selected had a segrega- tion ratio that fit the expected 1:2:1 ratio as determined by the Chi-squared test. Mapping populationh The Pearson’s correlation coefficient (PCC) and analysis of variance (ANOVA) were estimated in R as follows: corr_coef (data, verbose = TRUE) function and aov (formula, data = NULL, projections = FASLE, qr = TRUE, contrasts = NULL) function in METAN pack- age, respectively. The broad-sense heritability (H2) was calculated through the variability package as follows: H2 = Vg/Vp, where Vg is the genotypic variance and Vp is the phenotypic variance. A histogram for the phenotypes was generated using the hist function in R. RNA extraction and quantitative real‑time PCR Immature and mature Hwanggeum and DB-088 seeds were sampled at different developmental stages (R5, R6, R6.5, R7, and R8), with three biological replicates per sample. The seeds were immediately frozen and ground to a fine powder using liquid nitrogen and a mortar and pestle. Total RNA was extracted from the ground mate- rial using the TRIzol reagent (Invitrogen, Carlsbad, CA, USA). The RNA quantity and quality were determined using a NanoDrop ND-1000 spectrophotometer (Thermo Fisher Scientific). First-strand cDNA was synthesized from the high-quality RNA using the SuperScript III First-Strand Synthesis SuperMix (Invitrogen). Gene expression levels were analyzed by quantitative real-time PCR, which was completed using the CFX96 Real-time PCR system (Bio-Rad, Hercules, CA, USA) and the iTaq Universal SYBR Green Supermix (Bio-Rad). All steps were performed as described by the manufacturer. Rela- tive expression levels were calculated using the 2−ΔΔCt method [59], with Fbox serving as the reference gene. The primer sets (Table S1) were designed using Primer- 3plus (https://www.primer3plus.com/). l 2 p p In contrast to Danbaek, DB-088 is characterized by indeterminate growth, resulting in delayed flowering and maturation, which makes cultivation and breeding challenging. Therefore, we attempted to increase the iso- flavone contents of Hwanggeum, which has large seeds, matures relatively early, and exhibits determinate growth. Compared with DB-088 during the growth, maturity, and harvesting stages, Hwanggeum had superior agronomic traits, including FT, DM, GH, and HSW (Table S3). Seven quantitative traits (DF, DM, PH, LW, LL, NN, and HSW) differed significantly between the parental plants (Figure S1). In the F2 population, 13 agronomic traits were evalu- ated (Figure S2), of which only HC fit the 3:1 segregation ratio (χ2 = 0.3439 and P = 0.5576). DNA extraction and SNP genotyping Fully expanded trifoliate leaves from 190 F2 plants and two parental plants were placed in 2 ml tubes and fro- zen immediately using liquid nitrogen. The leaves were ground to a fine powder using TissueLyser II (Qiagen, Valencia, CA, USA). Genomic DNA was extracted from 100 mg ground material using the DNeasy Plant Mini Kit (Qiagen). The genomic DNA quality and concentration were determined using the NanoDrop ND-1000 spec- trophotometer (Thermo Fisher Scientific, Waltham, MA, USA). In addition, 190 F2 plants and two parental plants (Hwanggeum and DB-088) were genotyped using the 180K AXIOM® SoyaSNP array (180,961 SNP markers) according to the manufacturer instructions at DNALINK Inc. (Seoul, Republic of Korea). The candidate genes within the causal QTLs were func- tionally annotated according to the Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) databases and a gene co-expression network analysis was performed using the ClueGO [57] and STRING [58] Cytoscape plug-ins. To detect the enriched GO terms and KEGG pathways, the strengths of the relationships between terms and pathways were determined according to chance-corrected kappa statistics and visualized using ClueGO v.2.5.9 and CluePedia v.1.5.9. The term–term associations represent clusters of enriched functions and functional diversity based on a predefined kappa score. The load marker and network specificity were Kim et al. BMC Plant Biology (2023) 23:517 Page 5 of 18 adjusted for Glycine max (Wm82.a2.v1) and the Global option, respectively. Gene–gene interaction networks were constructed using StringApp v.2.0.1. A confidence cutoff value of 0.65 was applied to simplify networks and eliminate duplications. Information regarding the iden- tifiers and gene descriptions were retrieved from the String database. Data were analyzed and visualized using Cytoscape 3.9.1. [MDZI (82.01–1,580.17 μg g−1), MGLI (29.61–260.68 μg g−1), and MGNI (507.4–3,154.96 μg g−1)], acetyl-gluco- sides [ADZI (42.67–294.08 μg g−1) and AGNI (0.54–6.58 μg g−1)], and TI (898.29–5,980.82 μg g−1). The contents of all isoflavone forms were significantly influenced by gen- otypic effects (Table 1). The analysis of the broad-sense heritability of isoflavone contents revealed H2 ranged from 0.52 (GLE) to 0.98 (GNE), indicating that genetic factors have critical effects on soybean seed isoflavone contents. Overall, the observed transgressive segregation suggested that the alleles from both parental plants con- tributed to the isoflavone contents in the F2 population. Correlation between the isoflavone contents and agronomic traits We performed a Pearson correlation analysis to evalu- ate the correlations among agronomic traits and between agronomic traits and isoflavone contents in the F2 popu- lation (Fig. 2). Among the agronomic traits, there were strong positive correlations between PH and NN (r = 0.86, P < 0.001), GH and PH (r = 0.78, P < 0.001), GH and NN (r = 0.75, P < 0.001), DF and DM (r = 0.50, P < 0.001), and DF and GH (r = 0.50, P < 0.001), whereas there were negative correlations between PH and LS (r = − 0.30, P < 0.001), LS and NN (r = − 0.30, P < 0.001), and HSW and SC (r = − 0.37, P < 0.001). The analysis of the iso- flavone contents indicated the TI content was highly positively correlated with the DZI (r = 0.89, P < 0.001), MDZI (r = 0.90, P < 0.001), GNI (r = 0.92, P < 0.001), MGNI (r = 0.97, P < 0.001), and ADZI (r = 0.97, P < 0.001) contents, which was in contrast to the negative cor- relation between the TI content and the AGNI content (r = − 0.23, P < 0.001). Accordingly, the DZI, GNI, MDZI, MGNI, and ADZI contents contributed substantially to the TI content. The contents of three aglycones were highly correlated with the contents of the corresponding malonyl-glucosides (r = 0.86–0.98, P < 0.001). Notably, the AGNI content was negatively correlated with the con- tents of the other isoflavone forms, except for aglycones. Isoflavone contents and agronomic traits of the parents and the F2 populationhf The UPLC analysis revealed obvious differences between the isoflavone contents of Danbaek (original cultivar) and its mutant DB-088. The TI content was 4.04-fold higher for DB-088 (6,393 μg g−1) than for Danbaek (1,581 μg g−1). The DZI, GNI, and MGNI contents were more than 5-fold higher for DB-088 than for Danbaek, whereas there were no significant differences in the GLI and AGNI contents (Table S2). The individual isoflavone and TI contents in the F2 population were measured (Fig. 1). The TI contents were approximately 4.1-fold higher for DB-088 (6,393 μg g−1) than for Hwanggeum (1,558 μg g−1) (Table 1). With the exception of AGNI, the con- tents of each isoflavone form increased significantly in DB-088. In the F2 population, the isoflavone contents were as follows: aglycones [DZE (1.06–23.65 μg g−1), GLE (0.66–7.86 μg g−1), and GNE (3.65–72.78 μg g−1)], gluco- sides [DZI (11.07–430.49 μg g−1), GLI (39.92–120.36 μg g−1), and GNI (149.64–929.4 μg g−1)], malonyl-glucosides Kim et al. BMC Plant Biology (2023) 23:517 Page 6 of 18 Page 6 of 18 Kim et al. BMC Plant Biology (2023) 23:517 Fig 1 Distribution of isoflavone contents in the F population The red and blue dotted lines indicate Hwanggeum and DB 088 respectively The Fig. 1 Distribution of isoflavone contents in the F2 population. The red and blue dotted lines indicate Hwanggeum and DB-088, respectively. The green line indicates the average for the F2 mapping population Kim et al. Isoflavone contents and agronomic traits of the parents and the F2 populationhf BMC Plant Biology (2023) 23:517 Page 7 of 18 Page 7 of 18 Table 1 Descriptive statistics, ANOVA analysis, and broad-sense heritability of isoflavone contents in parental plants and F2 population DZE Daidzein, GLE Glycitein, GNE Genistein, DZI Daidzin, GLI Glycitin, GNI Genistin, MDZI Malonyl daidzin, MGLI Malonyl glycitin, MGNI Malonyl genistin, ADZI Acetyl daidzin, AGLI Acetyl glycitin, AGNI Ccetyl genistin, TI Total isoflavone, ns Not significant, Min Minimum, Max Maximum, SD Standard deviation, g Genotype effect, b Block effect, CV Coefficient of variation, H2 Broad-sense heritability a µg g−1 b The letters adjacent to average ± standard deviation indicates the result of Fisher’s LSD test at the 5% level (n = 3) * Indicates significant at p < 0.001 Traita Parental plants F2 population Effect Skewness Kurtosis CV H2 Hwangguem DB-088 Min Max Mean SD g Aglycones DZE 5.4 ± 0.52bb 17.8 ± 0.69a 1.06 23.65 7.99 4.66 * 0.89 0.81 0.59 0.91 GLE 0.8 ± 0.29b 3.2 ± 0.28a 0.66 7.86 2.63 1.28 * 1.24 1.6 0.60 0.52 GNE 15.7 ± 0.74b 31.6 ± 0.67a 3.65 72.78 21.03 11.97 * 1.21 1.83 0.56 0.98 Glycosides DZI 69.8 ± 5.63b 446.7 ± 20.55a 11.07 430.49 182.85 78.23 * 0.61 0.61 0.44 0.94 GLI 51.9 ± 3.46b 97.8 ± 9.24a 39.92 120.36 73.3 16.36 * 0.13 -0.41 0.25 0.70 GNI 181.7 ± 9b 786.4 ± 38.17a 149.64 929.4 425.69 135.03 * 0.60 0.66 0.33 0.93 Malonyl glycosides MDZI 305.4 ± 7.98b 1550.9 ± 43.21a 82.01 1580.17 688.89 273.34 * 0.40 0.27 0.40 0.97 MGLI 94.5 ± 10.72b 224.5 ± 24.8a 29.61 260.68 146.63 41.11 * -0.12 0.18 0.30 0.82 MGNI 759.1 ± 1.98b 2956.3 ± 87.19a 507.4 3154.96 1663.75 497.4 * 0.34 0.02 0.30 0.97 Acetyl glycosides ADZI 71.9 ± 2.26b 275 ± 6.84a 42.67 294.08 153.77 44.62 * 0.34 0.14 0.30 0.95 AGNI 1.8 ± 1.34a 3.1 ± 0.23a 0.54 6.58 2.2 1 * 1.11 2.44 0.50 0.72 Total isoflavone 1558 ± 26.89b 6393.4 ± 189.87a 898.29 5980.82 3368.71 992.85 * 0.18 -0.04 0.30 0.96 DZE Daidzein, GLE Glycitein, GNE Genistein, DZI Daidzin, GLI Glycitin, GNI Genistin, MDZI Malonyl daidzin, MGLI Malonyl glycitin, MGNI Malonyl genistin, ADZI Acetyl daidzin, AGLI Acetyl glycitin, AGNI Ccetyl genistin, TI Total isoflavone, ns Not significant, Min Minimum, Max Maximum, SD Standard deviation, g Genotype effect, b Block effect, CV Coefficient of variation, H2 Broad-sense heritability 1 chromosome ranged from 114 (chromosomes 11 and 12) to 479 (chromosome 2), with an average of 269 per chromosome. Isoflavone contents and agronomic traits of the parents and the F2 populationhf The relatively even distribution of the SNP markers among the 20 chromosomes was appro- priate for constructing a linkage map. The chromosome length ranged from 87.81 cM (chromosome 4, 142 SNP markers) to 241 cM (chromosome 8, 377 SNP markers), with an average length of 152.46 cM. The total length of the linkage map was 3,049.14 cM. The average interval between adjacent markers ranged from 0.38 cM (chro- mosome 2) to 1.58 cM (chromosome 11), whereas the largest gap between SNP markers varied from 5.80 cM (chromosome 8) to 31.26 cM (chromosome 11). On aver- age, 98.14% of the SNP markers on each chromosome were separated by < 5 cM [range: 92.11% (chromosome 11) to 99.74% (chromosome 3)], implying the mapped SNP markers were generally clustered together on the chromosomes. Moreover, the GNI and MGNI contents were highly posi- tively correlated with the DZI (r = 0.98, P < 0.001) and MDZI (r = 1, P < 0.001) contents, respectively. Moreover, the GNI and MGNI contents were highly posi- tively correlated with the DZI (r = 0.98, P < 0.001) and MDZI (r = 1, P < 0.001) contents, respectively. The analysis of the correlations between the agronomic traits and 12 isoflavone forms indicated DF and DM were positively correlated with the isoflavone contents, with the exception of the aglycone and AGNI contents. Fur- thermore, compared with the other agronomic traits, DF (r = 0.34, P < 0.001), DM (r = 0.50, P < 0.001), and LW (r = 0.34, P < 0.001) had stronger positive correlations with the TI content. The average TI, DF, DM, and LW values for the bottom/top 20% of the F2 population were 1,655.72/5,279.66 μg g−1, 46.4/51.95 days, 127.25/151.05 days, and 7.29/8.65 cm, respectively. The differences were significant (P < 0.001). Genetic linkage map for the F2 population using a high‑density SNP chip array To construct the linkage map, the raw data (180,961 SNPs) were filtered using the following criteria: no miss- ing data and monomorphic regions in the two parental plants and the F2 population (Table 2 and Figure S3). Of the 26,760 SNPs (14.79%) that were retained, 5,382 (20.11%) with a segregation ratio of 1:2:1 were mapped to 20 chromosomes. The number of SNPs mapped to a b The letters adjacent to average ± standard deviation indicates the result of Fisher’s LSD test at the 5% level (n = 3) * Indicates significant at p < 0.001 QTLs for the isoflavone contents and agronomic traits l A QTL mapping analysis was also performed to reveal the causal factors associated with individual isoflavone and TI contents (Table 3 and Fig. 3). A total of 29 QTLs for the isoflavone contents (except for GLE) were mapped to chromosomes 1, 3, 8, 11, 14, 15, and 17. The QTL Page 8 of 18 Kim et al. BMC Plant Biology (2023) 23:517 Fig. 2 Correlations between agronomic traits and isoflavone contents. Pearson’s correlation coefficient (− 1 to + 1) indicates how strongly two variables are linearly related. , P < 0.05; , P < 0.01; *, P < 0.001; and ns, not significant Fig. 2 Correlations between agronomic traits and isoflavone contents. Pearson’s correlation coefficient (− 1 to + 1) indicates how strongly two variables are linearly related. , P < 0.05; , P < 0.01; , P < 0.001; and ns, not significant Fig. 2 Correlations between agronomic traits and isoflavone contents. Pearson’s correlation coefficient (− 1 to + 1) indicates how strongly two variables are linearly related. , P < 0.05; , P < 0.01; , P < 0.001; and ns, not significant had LOD scores of 4.36 and 8.47 and explained 8% and 16.3% of the PV, respectively. The MDZI-related QTLs qMDZI08, qMDZI11, and qMDZI17, which were located on chromosomes 8, 11, and 17, respectively, had LOD scores of 4.56–11.46 and explained 7.2%–19.2% of the PV. The MGLI-related QTLs qMGLI11 and qMGLI15 were located on chromosomes 11 and 15, had LOD scores of 6.73 and 6.14, and explained 11.9% and 10.8% of the PV, respectively. The four MGNI-related QTLs qMGNI01, qMGNI08, qMGNI11, and qMGNI15 were located on chromosomes 1, 8, 11, and 15, respectively, had LOD scores of 4.53–11.01, and explained 6.4%–16.7% of the PV. The three QTLs (qADZI01, qADZI11, and qADZI15) for DZE (qDZE01) on chromosome 1 had a LOD score of 6.61 and explained 14.8% of the PV. The three GNE- related QTLs qGNE01, qGNE11, and qGNE15 were located on chromosomes 1, 11, and 15, respectively. They had LOD scores of 5.2–8.06 and explained 8.7%–14% of the PV. The three DZI-related QTLs qDZI08, qDZI11, and qDZI17 respectively located on chromosomes 8, 11, and 17 had LOD scores of 4.68–7.48 and explained 7.8%– 12.9% of the PV. QTLs for the isoflavone contents and agronomic traits The two GLI-related QTLs qGLI15 and qGLI17 were located on chromosomes 15 and 17, had LOD scores of 4.87 and 4.58, and explained 10% and 9.4% of the PV, respectively. The two GNI-related QTLs qGNI08 and qGNI15 on chromosomes 8 and 15 Kim et al. BMC Plant Biology (2023) 23:517 Page 9 of 18 Table 2 Information of linkage map in F2 population No. of mapped marker, number of mapped markers Chr Chromosome Chr No. of mapped markers Length (cM) Average spacing (cM) Max gap (cM) Gap < 5 cM Chr 1 313 165.01 0.53 10.79 99.68% Chr 2 479 181.68 0.38 8.58 99.58% Chr 3 380 180.07 0.47 7.65 99.74% Chr 4 142 87.81 0.62 9.72 97.18% Chr 5 242 107.81 0.45 9.10 98.76% Chr 6 178 128.06 0.72 13.14 97.19% Chr 7 308 181.38 0.59 11.15 97.40% Chr 8 377 241.83 0.64 5.80 99.47% Chr 9 271 152.60 0.56 17.86 98.15% Chr 10 195 123.94 0.64 12.51 97.44% Chr 11 114 180.34 1.58 31.26 92.11% Chr 12 114 107.92 0.95 17.48 96.49% Chr 13 282 189.20 0.67 19.81 98.23% Chr 14 250 153.13 0.61 25.31 98.40% Chr 15 308 140.89 0.46 8.36 99.03% Chr 16 397 176.55 0.44 11.81 99.24% Chr 17 324 180.10 0.56 13.34 98.46% Chr 18 319 148.46 0.47 19.03 98.75% Chr 19 180 88.95 0.49 7.37 99.44% Chr 20 209 133.42 0.64 8.33 98.06% Total 5,382 3,049.14 Mean 269.10 152.46 0.62 13.42 98.14% Table 2 Information of linkage map in F2 population trait due to the Hwanggeum alleles [e.g., DF (− 3.01 days) and DM (− 10.48 days)]. Hence, these alleles may be use- ful for increasing the soybean seed isoflavone contents and improving the undesirable traits of DB-088 (e.g., late flowering, late maturity, and indeterminate growth). Expression patterns of the candidate genes affecting the isoflavone contents and agronomic traitsh The Hwanggeum and DB-088 seed TI contents were measured at five seed developmental stages (Figure S5). The increase in the TI content varied among the exam- ined stages. In addition, the TI content was signifi- cantly higher (P < 0.001) in the DB-088 seeds than in the Hwanggeum seeds at all developmental stages. Moreover, the Hwanggeum and DB-088 seed TI contents increased rapidly in the R6.5, but the TI contents in R6.5 and R7 were 3.12-fold and 3.7-fold higher in the DB-088 seeds than in the Hwanggeum seeds, respectively.h The expression levels of four candidate genes within the major interval Gm11_9877690–Gm11_9955924 were analyzed in different seed developmental stages (Fig. 4). Among the four BGLU (β-glucosidase) genes, the BGLU13 (Glyma.11g129800) and BGLU17-1 (Glyma.11g129900) expression levels in Hwanggeum increased by more than 7-fold in R7, whereas the BGLU14 (Glyma.11g130000) and BGLU17 (Glyma.11g130100) expression levels in DB-088 tended to be highest during the early stages (R5, R6, and R6.5). The candidate gene expression levels dif- fered significantly between the parental plants. The inter- actions between genes with similar expression patterns, which were revealed by the co-expression analysis, sug- gest the genes may contribute to the regulation of isofla- vone contents and agronomic traits. for ADZI located on chromosomes 1, 11, and 15, respec- tively, had LOD scores of 5.01–11.7 and explained 7.9%–20.1% of the PV. The three AGNI-related QTLs qAGNI03, qAGNI11, and qAGNI14 on chromosomes 3, 11, and 14, respectively, had LOD scores of 4.51–17.65 and explained 6.4%–29.6% of the PV. The three QTLs (qTI01, qTI08, and qTI11) for the TI content on chromo- somes 1, 8, and 11, respectively, had LOD scores of 4.37– 10.40 and explained 7.2%–18.6% of the PV. Co‑expression networks for the candidate genes responsible for isoflavone contents and agronomi responsible for isoflavone contents and agronomic traits The 267 candidate genes within the QTLs related to iso- flavone contents were annotated with 102 GO terms, resulting in 298 connected gene nodes and 596 edges. These genes were categorized into 58 biological pro- cesses, 19 cellular components, and 28 molecular func- tions (Fig. 5). A total of 128 functionally characterized genes were annotated with 15 GO terms, including cata- lytic activity, nucleic acid binding, and cellular response to chemical stimulus. Many genes were annotated with multiple GO terms. The 174 candidate genes within the QTLs related to agronomic traits were used to construct the co-expression networks on the basis of gene–gene interactions. According to the enriched GO terms and KEGG pathways, the candidate genes were categorized into 49 biological processes, 17 cellular components, and 13 molecular functions (Figure S6). Detailed regarding these GO terms are provided in Tables S5 and S6. p Eight intervals (Gm01_50161351–Gm01_50361209, Gm08_443145–Gm08_672177, Gm11_9334773– Gm11_9492454, Gm11_9492454–Gm11_9625710, Gm11_9877690–Gm11_9955924, Gm15_42297136– Gm15_42406919, Gm15_45646568–Gm15_45951454, and Gm17_9608605–Gm17_10719220) were detected in at least two QTLs for isoflavone contents. The intervals on chromosome 11 were associated with the negative additive effect on the TI content (− 613.6 μg g−1) due to the Hwanggeum allele, which explained 20.1% of the PV. The three agronomic trait-associated intervals within Gm11_9955924–Gm11_11922317 overlapped the isofla- vone-related intervals (Table S4 and Figure S4). A nega- tive additive effect reflected a decrease in the agronomic Kim et al. BMC Plant Biology (2023) 23:517 Page 10 of 18 Table 3 QTL analysis for isoflavone contents in the F2 population DZE Daidzein, GNE Genistein, DZI Daidzin, GLI Glycitin, GNI Genistin, MDZI Malonyl daidzin, MGLI Malonyl glycitin, MGNI Malonyl genistin, ADZI Acetyl daidzin, AGNI Acetyl genistin, TI Total isoflavone, QTL Quantitative trait locus, Chr Chromosome; 1.5-LOD interval Indicates the significant location of a QTL, Bin no. Number of bin markers, GD Genetic distance, PD Physical distance, Peak Major peak in interval, LOD Logarithm of odds, and the ratio of the probability that a QTL is present, R2 The ratio (%) of the QTL explains for the phenotypic variation, AE The additive effect, and negative value means that it is reduced by allele of HG, Ref Reference Trait QTL Chr 1.5-LOD interval GD (cM) PD (Kb) LOD R2 AE Ref DZE qDZE01 1 Gm01_49284469 – Gm01_49344756 106.48–107.03 60.29 6.61 14.8 2.68 Gutierrez-Gonzalez. et al. Co‑expression networks for the candidate genes responsible for isoflavone contents and agronomi Number of bin markers, GD Genetic distance, PD Physical distance, Peak Major peak in interval, LOD Logarithm of odds, and the ratio of the probability that a QTL is present, R2 The ratio (%) of the QTL explains for the phenotypic variation, AE The additive effect, and negative value means that it is reduced by allele of HG, Ref Reference isoflavone contents, an interaction was detected between two BGLU genes, Glyma.11g129800 (BGLU13) and Glyma.11g129900 (BGLU17), which were associated with MGLI, MGNI, ADZI, and TI (Fig. 6). The co-expression network analysis indicated 40 genes were either directly or indirectly linked through one or more interactions with the isoflavone contents, reflect- ing the similarity in their functions (Fig. 6). Additionally, within the causal QTLs, one gene related to agronomic traits and four genes associated with isoflavone con- tents were involved in gene–gene interactions. Specifi- cally, within the major interval for agronomic traits on chromosome 11, Glyma.11g129900 (β-glucosidase 17, BGLU17) interacted with four BGLU genes, namely Glyma.11g129500 (BGLU13), Glyma.11g129700 (BGLU17), Glyma.11g129300 (BGLU15), and Glyma.11g129600 (BGLU17), which were responsi- ble for DF, GH, PT, and LL (Figure S7). In terms of the Co‑expression networks for the candidate genes responsible for isoflavone contents and agronomi [60] GNE qGNE01 1 Gm01_50161351 – Gm01_50361209 114.05–116.75 199.86 5.20 8.7 5.22 qGNE11 11 Gm11_9492454 – Gm11_9625710 76.88–77.95 133.26 8.06 14 6.23 qGNE15 15 Gm15_42297136 – Gm15_42406919 15.63–15.90 109.78 6.08 10.3 -5.54 DZI qDZI08 8 Gm08_773460 – Gm08_837769 4.28–4.54 64.31 4.68 7.8 -29.96 qDZI11 11 Gm11_9334773 – Gm11_9492454 75.82–76.88 157.68 7.48 12.9 -38.43 qDZI17 17 Gm17_9608605 – Gm17_10719220 73.60–84.05 1110.62 5.27 8.9 -32.47 Yoshikawa et al. [61] GLI qGLI15 15 Gm15_45646568 – Gm15_45951454 19.70–19.86 304.89 4.87 10 7.463 qGLI17 17 Gm17_9608605 – Gm17_10719220 73.60–84.05 1110.62 4.58 9.4 -6.59 Yoshikawa et al. [61] GNI qGNI08 8 Gm08_136941 4.36 8 -55.80 qGNI15 15 Gm15_31169446 – Gm15_34246376 11.36–11.66 3076.93 8.47 16.3 -81.42 Han et al. [62] MDZI qMDZI08 8 Gm08_443145 – Gm08_672177 1.62–2.41 229.03 4.65 7.2 -103.67 qMDZI11 11 Gm11_9334773 – Gm11_9492454 79.29–93.82 157.68 11.46 19.2 -164.88 qMDZI17 17 Gm17_9608605 – Gm17_10719220 73.60–84.05 1110.62 4.82 7.4 -103.66 Yoshikawa et al. [61] MGLI qMGLI11 11 Gm11_9877690 – Gm11_9955924 79.29–79.82 78.23 6.73 11.9 -20.19 qMGLI15 15 Gm15_45951454 – Gm15_45646568 19.70–19.86 304.89 6.14 10.8 20.25 MGNI qMGNI01 1 Gm01_50161351 – Gm01_50361209 114.05–116.75 199.86 5.15 7.3 190.65 qMGNI08 8 Gm08_443145 – Gm08_672177 1.62–2.41 229.03 4.53 6.4 -181.30 qMGNI11 11 Gm11_9877690 – Gm11_9955924 79.29–79.82 78.23 11.01 16.7 -286.52 qMGNI15 15 Gm15_42297136 – Gm15_42406919 15.63–15.90 109.78 8.54 12.6 -266.30 ADZI qADZI01 1 Gm01_50161351 – Gm01_50361209 114.05–116.75 199.85 5.01 7.9 17.30 qADZI11 11 Gm11_9877690 – Gm11_9955924 79.29–79.82 78.23 11.7 20.1 -28.08 qADZI15 15 Gm15_42297136 – Gm15_42406919 15.63–15.90 109.78 7.52 12.2 -22.79 AGNI qAGNI03 3 Gm03_10738489 – Gm03_23292786 45.73 12,554.30 4.51 6.4 -0.24 Kassem et al. [63] qAGNI11 11 Gm11_9492454 – Gm11_9625710 76.88–77.95 133.26 6.25 9.1 0.39 qAGNI14 14 Gm14_2393929 – Gm14_2410449 13.77–14.03 16.52 17.65 29.6 -0.77 TI qTI01 1 Gm01_50161351 – Gm01_50361209 114.05–116.75 199.86 4.37 7.2 365.35 qTI08 8 Gm08_1864255 – Gm08_2234603 11.26–13.41 370.35 5.26 8.8 -418.24 qTI11 11 Gm11_9877690 – Gm11_9955924 79.29–79.82 78.23 10.40 18.6 -613.60 Table 3 QTL analysis for isoflavone contents in the F2 population DZE Daidzein, GNE Genistein, DZI Daidzin, GLI Glycitin, GNI Genistin, MDZI Malonyl daidzin, MGLI Malonyl glycitin, MGNI Malonyl genistin, ADZI Acetyl daidzin, AGNI Acetyl genistin, TI Total isoflavone, QTL Quantitative trait locus, Chr Chromosome; 1.5-LOD interval Indicates the significant location of a QTL, Bin no. Discussionl Isoflavones are polyphenolic compounds that are highly abundant in soybean [10]. Soybean isoflavones are mul- tifunctional compounds important for legume–microbe interactions and the estrogenic effects of 17B-estradiol [3, 12]. However, the practical breeding strategies for developing soybean varieties with high isoflavone con- tents are inefficient and inadequate. We previously developed the DB-088 mutant with high isoflavone Kim et al. BMC Plant Biology (2023) 23:517 Kim et al. BMC Plant Biology Page 11 of 18 Fig. 3 Analysis of QTLs related to isoflavone contents using a high-density linkage map of 20 chromosomes. The SNP position and genetic distance are provided on the right and left sides, respectively. The QTLs are positioned to the right of the chromosome. The QTLs in the same color are in identical positions on the chromosome Fig. 3 Analysis of QTLs related to isoflavone contents using a high-density linkage map of 20 chromosomes. The SNP position and genetic distance are provided on the right and left sides, respectively. The QTLs are positioned to the right of the chromosome. The QTLs in the same color are in identical positions on the chromosome analyses were conducted using homozygous populations, but maintaining large populations can be cumbersome, time-consuming, and costly [38, 39, 67]. This compelled us to conduct a QTL mapping analysis of isoflavone con- tents and agronomic traits using a heterozygous popu- lation (F2 generation). By identifying causal QTLs with potential candidate genes, we revealed the potential appli- cability of these traits. The heritability of the contents of 12 isoflavones varied, from the relatively low heritability of the aglycone contents (35%–72%) to the higher herit- ability of the glucoside (63%–98%), malonyl-glucoside (72%–96%), and TI (44%–99%) contents [5, 30, 68, 69]. Additionally, according to our data, the heritability of individual isoflavone contents varied from 52% (GLE) to 98% (GNE), with an average of 86.4%. The similarity in the heritability of isoflavone contents between the het- erozygous and homozygous populations suggests that the isoflavone contents will be maintained at similar levels between generations. The correlation analysis showed that the TI content was highly positively correlated with the DZI, GNI, MDZI, MGNI, and ADZI contents (r > 0.89). Moreover, the contents of individual isoflavones and the TI content were positively correlated with the flowering contents via gamma irradiation [52]. Discussionl Compared with the corresponding contents in Danbaek and Hwanggeum, there was an approximately 4-fold increase in the TI content (1,580.51/1,558.04 μg g−1, respectively) in DB-088 (aglycones: 52.61 μg g−1 and glucosides: 6,340.75 μg g−1); the contents of the aglycones and conjugated forms were 3.7-fold/2.3-fold (14.34/21.98 μg g−1, respec- tively) and 4-fold/4.1-fold (1,566.17/1,536.06 μg g−1, respectively) higher in DB-088, respectively (Table S2). Isoflavone conjugates, which are the main constituents in seeds, are relatively stable and soluble, making them appropriate for storage in vacuoles [64]. Additionally, isoflavone conjugates can be converted into aglycones via fermentation, thereby increasing their bioavailability for the production of pharmaceuticals as well as mois- turizing and anti-aging cosmetics [65, 66]. In the current study, our objective was to identify QTLs associated with high isoflavone contents in DB-088. The SoyBase database contains relatively few QTLs related to secondary metabolites. For example, it includes 297, 81, and 31 QTLs related to the isoflavone, tocopherol, and saponin contents, respectively, whereas there are 1,032 QTLs associated with the oil content. Most earlier QTL Page 12 of 18 Kim et al. BMC Plant Biology (2023) 23:517 Fig. 4 Expression levels of four candidate genes in the major interval in different seed developmental stages. The expression level was normalized to the reference gene Fbox. Error bars represent the standard error. , , and * indicate significance at P < 0.05, 0.01, and 0.001, respectively Fig. 4 Expression levels of four candidate genes in the major interval in different seed developmental stages. The expression level was normalized to the reference gene Fbox. Error bars represent the standard error. , , and ** indicate significance at P < 0.05, 0.01, and 0.001, respectively can increase by up to 60.41% depending on the maturity time [70]. Moreover, Wang et al. [71] detected a positive correlation between the isoflavone content and maturity time, with isoflavone levels increasing during the seed filling stage under cool conditions. However, the GNI (r = − 0.27), DZI (r = − 0.17), and GNE (r = − 0.33) con- tents were negatively correlated with DM. time and maturity time. Azam et al. [4] also reported that MDZI (r = 0.92), MGNI (r = 0.90), DZI (r = 0.81), and GNI (r = 0.89) contents are positively correlated with the TI content, but their analysis detected a negative correlation between the ADZI and TI contents. (See figure on next page.) Fig. 5 Enriched GO terms and KEGG pathways among the candidate genes associated with isoflavone contents. Nodes indicate a specific term and are linked on the basis of term–term similarity. Different colors reflect the degree of enrichment, with only the most significant term presented. The bar and node colors match the color of the corresponding term. The node size reflects significance and group colors were arbitrarily selected. The percentage of genes per term and terms per group are indicated in the bar graph and pie chart, respectively. This figure was adapted with permission from the copyright holder (Kyoto Encyclopedia of Genes and Genomes) Discussionl This interval was also related to agronomic traits, including DF, DM, GH, PT, LL, and NN. In fact, the variation in the isoflavone contents differed significantly depending on DF and DM, which is consistent with the results of our correlation analysis. Ultimately, four puta- tive genes encoding a β-glucosidase (BGLU) were iden- tified within this 78-kb interval on chromosome 11. The additive effect of this region on the TI content (613.6 μg g−1) was due to the mutant alleles in the paternal plant. seeds. Of these QTLs, six related to isoflavone contents were also identified in earlier studies. Major QTLs for isoflavone contents were previously localized to chromo- some 5. More specifically, gen-A1, dai-A1, and tot-A1, which are related to the GNE, DZE, and TI contents, are located in a 7–13.9 cM interval between Satt236 and Sat_271 and account for 5.8%–9.2% of the PV [72], whereas qGEN5, qDAI5, and TOT5 (Satt236–Sat_271 and Satt174–Sat_236) explain 5.5%–8.3% of the PV and have a positive additive effect (50.3–161.7) [60]. Gutier- rez-Gonzalez et al. [32] identified major QTLs (qGNE5, qDAI5, and qTOT5) at the same position on chromo- some 5 (BARC-042999–08498–BARC-016279–02316) using a different set of RILs. Yang et al. [73] also reported that the major QTL for the DZE, GNE, and TI contents is located near Sat_217 on chromosome 5. Furthermore, Cai et al. [39] confirmed that qIF5-1 is a major isofla- vone-related QTL (611.4 kb) that contains 81 candidate genes, including genes encoding transcription factors and ubiquitin-related proteins. However, in the present study, although no significant QTLs were detected on chromo- some 5, eight QTLs affecting more than two types of iso- flavones were mapped to chromosome 11; these QTLs explained up to 20.1% of the PV. Kassem et al. [63, 74] identified Satt197, Satt197, Satt251, and Sct26 on chro- mosome 11 as the major QTLs for the seed GLE content across multiple environments. These QTLs are local- ized to a 2.9-cM interval and explain 21%–50.2% of the PV. Additionally, qGEN11 (15.6 cM), qDAI11 (13.6 cM), and qTOT (3.7 cM) are located between BARC-04299– 08498 and BARC-016279–02316 (623.4 kb) and account for 7% of the PV, whereas qGLY11 (responsible for 3% of the PV) was identified in an interval between BARC- 054421–12081 and BARC-050069–09363 that includes a gene encoding 4-coumarate:CoA ligase (4CL) [32]. Discussionl In the present study, a delay in DF and DM by approximately 5 and 24 days, respectively, in the mapping population resulted in a 3.2- fold increase in the isoflavone content. However, agly- cone and AGNI contents were negatively correlated with almost all of the examined agronomic traits, including DF [r = − 0.12 (DZE) to − 0.27 (GLE)] and DM [r = − 0.16 (DZE) to − 0.37(GLE)]. Our findings are consistent with those of a previous study that showed the TI content We verified 29 QTLs for isoflavone contents using a mutant-based F2 mapping population and 180,961 SNP markers. The QTLs mapped to chromosomes 1, 3, 8, 11, 14, 15, and 17 had LOD scores of 4.36–17.7 and explained 6.4%–29.6% of the PV, suggestive of the presence of a major QTL underlying the isoflavone contents of soybean Kim et al. BMC Plant Biology (2023) 23:517 Page 13 of 18 (See legend on previous page.) Fig. 5 (See legend on previous page.) Fig. 5 (See legend on previous page.) Kim et al. BMC Plant Biology (2023) 23:517 Page 14 of 18 Fig. 6 Gene–gene interaction network comprising the candidate genes associated with isoflavone contents. Edges (gray line) indicate gene–gene interactions. Node (gene) colors were arbitrarily selected. High confidence scores are indicated by thick and dark lines Fig. 6 Gene–gene interaction network comprising the candidate genes associated with isoflavone contents. Edges (gray interactions. Node (gene) colors were arbitrarily selected. High confidence scores are indicated by thick and dark lines eraction network comprising the candidate genes associated with isoflavone contents. Edges (gray line) indicate gene–gene ne) colors were arbitrarily selected. High confidence scores are indicated by thick and dark lines were located on chromosomes 1, 8, 11, 15, and 17. The QTLs in the intervals had LOD scores of 4.37–11.7 and explained 6.4%–20.1% of the PV, suggestive of the broad contributions of both parental plants and the presence of additive effects on various types of isoflavones. Some of the identified genes encode proteins involved in the phenylpropanoid pathway (e.g., flavonoid and isoflavone biosynthesis). For example, the qGLI15 and qMGLI15 QTLs appear to include a gene (Glyma.15g240700) encoding MAP kinase 19. Genes encoding glycosyl hydrolase family members and acyl-CoA synthetase/ ligases were identified in qGNE01, qMGNI01, qMGLI11, qMGNI11, qADZI11, and qTI11. The Gm11_9877690– Gm11_9955924 interval related to the MGLI, MGNI, ADZI, and TI contents explained 11.9%–20.1% of the PV (Table 3). Discussionl In the current study, 22 QTLs associated with more than two isoflavone forms were revealed as closely related and seeds. Of these QTLs, six related to isoflavone contents were also identified in earlier studies. Major QTLs for isoflavone contents were previously localized to chromo- some 5. More specifically, gen-A1, dai-A1, and tot-A1, which are related to the GNE, DZE, and TI contents, are located in a 7–13.9 cM interval between Satt236 and Sat_271 and account for 5.8%–9.2% of the PV [72], whereas qGEN5, qDAI5, and TOT5 (Satt236–Sat_271 and Satt174–Sat_236) explain 5.5%–8.3% of the PV and have a positive additive effect (50.3–161.7) [60]. Gutier- rez-Gonzalez et al. [32] identified major QTLs (qGNE5, qDAI5, and qTOT5) at the same position on chromo- some 5 (BARC-042999–08498–BARC-016279–02316) using a different set of RILs. Yang et al. [73] also reported that the major QTL for the DZE, GNE, and TI contents is located near Sat_217 on chromosome 5. Furthermore, Cai et al. [39] confirmed that qIF5-1 is a major isofla- vone-related QTL (611.4 kb) that contains 81 candidate genes, including genes encoding transcription factors and ubiquitin-related proteins. However, in the present study, although no significant QTLs were detected on chromo- some 5, eight QTLs affecting more than two types of iso- flavones were mapped to chromosome 11; these QTLs explained up to 20.1% of the PV. Kassem et al. [63, 74] identified Satt197, Satt197, Satt251, and Sct26 on chro- mosome 11 as the major QTLs for the seed GLE content across multiple environments. These QTLs are local- ized to a 2.9-cM interval and explain 21%–50.2% of the PV. Additionally, qGEN11 (15.6 cM), qDAI11 (13.6 cM), and qTOT (3.7 cM) are located between BARC-04299– 08498 and BARC-016279–02316 (623.4 kb) and account for 7% of the PV, whereas qGLY11 (responsible for 3% of the PV) was identified in an interval between BARC- 054421–12081 and BARC-050069–09363 that includes a gene encoding 4-coumarate:CoA ligase (4CL) [32]. In the current study, 22 QTLs associated with more than two isoflavone forms were revealed as closely related and In Hwanggeum, BGLU13 (Glyma.11g129800) and BGLU17 (Glyma.11g129900) were expressed after R6.5, which marks the onset of significant isoflavone synthe- sis. This suggests the enhanced hydrolysis of glucosides (deglycosylation) leads to decreased isoflavone accumula- tion in seeds [Hwanggeum (1,711 μg g−1)/DB-088 (6,462 μg g−1)] (Fig. 4 and Figure S5). Conclusion We mapped the causal QTLs underlying agronomic traits and isoflavone contents using an F2 mapping pop- ulation derived from the soybean mutant DB-088 and a high-density SNP chip array. The QTLs identified in this study included novel QTLs as well as previously reported QTLs. In the major interval on chromosome 11, we identified the following candidate genes associated with favorable agronomic traits and high TI contents: BGLU13 (Glyma.11g129800), BGLU14 (Glyma.11g130000), and BGLU17 (Glyma.11g129900 and Glyma.11g130100). The study results may be useful for future research aimed at developing molecular markers applicable for the breed- ing of new soybean lines with optimized isoflavone con- tents and agronomic performance. fl v g . To date, there has been insufficient research on the effects of BGLU genes on isoflavone accumulation in soybean seeds. The BGLU13 and BGLU14 amino acid sequences are respectively 98% and 87% similar to that of BGLU17, which belongs to a group of hydrolases that specifically target flavonoid and isoflavonoid conjugates [75, 76]. Previous studies characterized a β-glucosidase (β-D-glucoside glucohydrolase; EC 3.2.1.21) as a glyco- syl hydrolase contributing to the plant defense system (e.g., against biotic and abiotic stresses and herbivores), lignification, and cell wall remodeling. This enzyme can catalyze the hydrolysis of glucosides to their aglycones in two steps (glycosylation and deglycosylation), resulting in structural and functional diversity [77]. It cleaves the β-glucosidic linkages of glucose conjugates in microor- ganisms, animals, and plants [78, 79]. A total of 47, 51, and 64 BGLU genes have been identified in Arabidop- sis thaliana, Medicago truncatula, and Brassica rapa, respectively [76, 77, 80]. Ishihara et al. [81] determined that BGLU6 is an acyl-glucose-dependent glucosyltrans- ferase homolog that is co-expressed with phenylpro- panoid biosynthetic genes; the mutation of this gene in the transgenic bglu6 mutant line decreases the flavonol 3-O-gentiobioside 7-O-rhamnoside (F3GG7R) content in accordance with the lack of F3GG7R associated with Ler BGLU6 genes. Our study findings were consistent with the results of previous studies showing that the total isoflavone content is approximately 4-fold higher in fully mature seeds (R7) than in immature seeds (R5) and green full-sized seeds (R6), with more than 80% of the TI con- tent comprising glycosides and malonyl-glucosides. The glycosylation of isoflavones (glycosylation and malonyla- tion) leads to increased solubility and stability, which is conducive to transport and storage in vacuoles. Conclusion Thus, the overexpression of BGLU13 (Glyma.11g129800) and BGLU17 (Glyma.11g129900) in Hwanggeum likely inhib- its the formation and accumulation of isoflavone-gluco- sides in the late seed developmental stage because of the deglycosylation. Conversely, these genes are expressed at extremely low levels in DB-088, which enables glucosides to accumulate in the seeds. Supplementary Information The online version contains supplementary material available at https://doi. org/10.1186/s12870-023-04519-x. Additional file 1: Figure S1. Morphology of plants and their seeds. From left to right: Danbaek, DB-088, and Hwangguem. (A) Plants. (B) Seeds. Additional file 2: Figure S2. Distribution of agronomic traits in the F2 population. Growth habit (D, determinate; SD, semi-determinate; ID, indeterminate); Plant type (SU, semi-upright; M, middle; H, horizontal); Leaf shape (O, ovoid; HS, heart-shaped); Pod color (BR, brown; DBR, dark brown; BK, black); Seed color (YL, yellow; GY, greenish yellow); Hilum color (YL, yellow; DBR, dark brown). The red and blue dotted lines indicate Hwangguem and DB-088, respectively. The green line indicates the aver‑ age for the F2 mapping population. Additional file 3: Figure S3. Construction of a high-density linkage map in 20 chromosomes. The SNP position and genetic distance are provided on the right and left sides, respectively. Additional file 4: Figure S4. Analysis of QTLs related to agronomic traits using a high-density linkage map of 20 chromosomes. The SNP position and genetic distance are provided on the right and left sides, respectively. The QTLs are positioned to the right of the chromosome. The QTLs in the same color are in identical positions on the chromosome. Additional file 5: Figure S5. Total isoflavone content in seeds of Hwanggeum and DB-088 according to developmental stages. , , and ** indicate significance at p < 0.05, 0.01, and 0.001, respectively. Additional file 6: Figure S6. Enriched GO term and KEGG pathway analy‑ ses of candidate genes associated with agronomic traits. Nodes indicated a specific term and are functionally linked based on term-term similarity. Each different color indicates the term enrichment, and only the most significant term is shown. The colors of the bars and nodes match accord‑ ing to respective terms. The node size means significance and group color are arbitrary. The percentage of genes per term and terms per group are shown in bars label and pie chart, respectively. Figure adapted with copy‑ right permission from the Kyoto Encyclopedia of Genes and Genomes. Additional file 7: Figure S7. Gene-gene interactions network of candi‑ date genes associated with agronomic traits. Edges (gray line) indicate gene-gene interactions and node (gene) colors are arbitrary. Higher confidence scores are represented as thicker and darker lines. Additional file 8: Table S1. List of qRT-PCR primers. Additional file 8: Table S1. List of qRT-PCR primers. Additional file 9: Table S2. Discussionl In contrast, BGLU14 (Glyma.11g130000) and BGLU17-2 (Glyma.11g130100) tended to be highly expressed in the early seed devel- opmental stages in DB-088, suggesting the encoded enzymes may function as glycosyltransferases (glycosyla- tion) involved in synthesizing isoflavone conjugates. They are also likely related to the early flowering (by 10 days) Kim et al. BMC Plant Biology (2023) 23:517 Kim et al. BMC Plant Biology (2023) 23:517 Kim et al. BMC Plant Biology (2023) 23:517 Page 15 of 18 Page 15 of 18 and early maturation (by 20 days) of Hwanggeum. The positive correlation between specific traits (DF and DM) and the TI content (0.34 and 0.50, respectively) suggests the traits are associated with isoflavone accumulation. The co-expression analysis indicated that the gene–gene interaction was consistent with BGLU expression pat- terns (Glyma.11g129800 and Glyma.11g129900) (Fig. 4). Specifically, BGLU17 (Glyma.11g129900) was associated with agronomic traits (DF, PT, GH, and LL) and isofla- vone contents (MGLI, MGNI, ADZI, and TI). These find- ings suggest that these genes may have similar regulatory effects on isoflavone contents and agronomic traits. Author details 16. Lozovaya VV, Lygin AV, Ulanov AV, Nelson RL, Daydé J, Widholm JM. Effect of temperature and soil moisture status during seed development on soybean seed isoflavone concentration and composition. Crop Sci. 2005;45(5):1934–40. https://doi.org/10.2135/cropsci2004.0567. 1 Advanced Radiation Technology Institute, Korea Atomic Energy Research Institute, Jeongeup 56212, Republic of Korea. 2 Department of Applied Plant Science, College of Agriculture and Life Sciences, Chonnam National University, Gwangju 61186, Republic of Korea. 3 Department of Agricultural Biotechnology, National Institute of Agricultural Sciences, RDA, Jeonju 54874, Republic of Korea. 4 Department of Smart Farm Science, College of Life Sci‑ ences, Kyung Hee University, Yongin 17104, Republic of Korea. 17. Collard BC, Jahufer M, Brouwer J, Pang ECK. An introduction to markers, quantitative trait loci (QTL) mapping and marker-assisted selection for crop improvement: the basic concepts. Euphytica. 2005;142(1):169–96. https://doi.org/10.1007/s10681-005-1681-5. 18. Young N. QTL mapping and quantitative disease resistance in plants. Annu Rev Phytopathol. 1996;34(1):479–501. https://doi.org/10.1146/ annurev.phyto.34.1.479. Received: 25 July 2023 Accepted: 10 October 2023 19. Singh B, Singh A. Mapping populations. In: Marker-assisted plant breed‑ ing: principles and practices. Springer; 2015. p. 125–50. https://doi.org/10. 1007/978-81-322-2316-0_5. Availability of data and materials The information of sequence data and positions of QTLs in this study can be obtained in SoyBase (https://www.soybase.org/) and Phytozome (https:// phytozome-next.jgi.doe.gov/). Other data and materials generated during this study are included in this research article [and its supplementary information files]. The datasets about phenotype and genotype of the repository can be found below: https://github.com/jmkim0803/QTL-mapping. 10. Setchell KD, Radd S. Soy and other legumes:‘bean’around a long time but are they the ‘superfoods’ of the millennium and what are the safety issues for their constituent phytoestrogens? Asia Pac J Clin Nutr. 2000;9(S1):S13– 22. https://doi.org/10.1046/j.1440-6047.2000.009ss13.x. 11. Yu O, Shi J, Hession AO, Maxwell CA, McGonigle B, Odell JT. Metabolic engi‑ neering to increase isoflavone biosynthesis in soybean seed. Phytochemis‑ try. 2003;63(7):753–63. https://doi.org/10.1016/s0031-9422(03)00345-5. Funding h k This work was supported by the research program of Korea Atomic Energy Research Institute (Project No. 523320–23). 9. Kudou S, Fleury Y, Welti D, Magnolato D, Uchida T, Kitamura K, et al. Malo‑ nyl isoflavone glycosides in soybean seeds (Glycinemax Merrill). Agric Biol Chem. 1991;55(9):2227–33. https://doi.org/10.1271/bbb1961.55.2227. Declarations y g 12. Wei H, Bowen R, Cai Q, Barnes S, Wang Y. Antioxidant and antipromo‑ tional effects of the soybean isoflavone genistein. Proc Soc Exp Biol Med. 1995;208(1):124–30. https://doi.org/10.3181/00379727-208-43844. Authors’ contributions SJK and BKH conceived and revised the manuscript. SJK developed a mutant soybean population. BKH and JMK developed the mapping population. JMK and SJK designed the experiment and drafted the manuscript. SJS, JWL, and SHE assisted in fieldwork and phenotyping. JIL assisted in genotyping. JR analyzed the data. All authors reviewed the manuscript. 7. Wang HJ, Murphy PA. Isoflavone composition of American and Japanese soybeans in Iowa: effects of variety, crop year, and location. J Agric Food Chem. 1994;42(8):1674–7. https://doi.org/10.1021/jf00044a017. 8. Yoon H, Yi J, taye Desta K, Shin MJ, Lee Y, Lee S, et al. Yearly variation of isoflavone composition and yield-related traits of 35 Korean soybean germplasm. Korean Soc Breed Sci. 2021;53(4):411–23. https://doi.org/10. 9787/kjbs.2021.53.4.411. Acknowledgements h h h k h The authors thank the editor and reviewers for constructive review and sug‑ gestions to improve this manuscript. 6. Tsukamoto C, Shimada S, Igita K, Kudou S, Kokubun M, Okubo K, et al. Factors affecting isoflavone content in soybean seeds: changes in isofla‑ vones, saponins, and composition of fatty acids at different temperatures during seed development. J Agric Food Chem. 1995;43(5):1184–92. https://doi.org/10.1021/jf00053a012.l Supplementary Information Total and individual isoflavone contents in seeds of Danbaek, Hwangguem, and DB-088. Additional file 10: Table S3. Descriptive statistics of agronomic traits in parental plants and F2 population. Kim et al. BMC Plant Biology (2023) 23:517 Page 16 of 18 Page 16 of 18 3. Sreevidya V, Srinivasa Rao C, Sullia S, Ladha JK, Reddy PM. Metabolic engineering of rice with soybean isoflavone synthase for promoting nodulation gene expression in rhizobia. J Exp Bot. 2006;57(9):1957–69. https://doi.org/10.1093/jxb/erj143. Additional file 11: Table S4. QTL analysis for agronomic traits in the F2 population. Additional file 12: Table S5. Information of GO terms for agronomic tratis. Additional file 13: Table S6. Information of GO terms for isoflavone contents. Additional file 12: Table S5. Information of GO terms for agronomic tratis. Additional file 12: Table S5. Information of GO terms for agronomic tratis. 4. Azam M, Zhang S, Abdelghany AM, Shaibu AS, Feng Y, Li Y, et al. Seed isoflavone profiling of 1168 soybean accessions from major growing ecoregions in China. Food Res Int. 2020;130:108957. https://doi.org/10. 1016/j.foodres.2019.108957. Additional file 13: Table S6. Information of GO terms for isoflavone contents. 5. Kim HS, Kang BK, Seo JH, Ha TJ, Kim HT, Shin SO, et al. Quantitative vari‑ ation of total seed isoflavone and its compositions in Korean soybean cultivars (Glycinemax (L.) Merr.). Korean J Crop Sci. 2019;64(2):89–101. https://doi.org/10.7740/kjcs.2019.64.2.089. Ethics approval and consent to participate We ensure that all plant materials used in the current study were developed through our previous research [53] and collected from the Radiation Breed‑ ing Farm of the Korea Atomic Energy Research Institute (KAERI; 35.5699°N 126.9772°E, Jeongeup, Jeollabuk, Republic of Korea), following relevant guidelines and regulations. 13. Rimbach G, De Pascual-Teresa S, Ewins B, Matsugo S, Uchida Y, Minihane A-M, et al. Antioxidant and free radical scavenging activity of isoflavone metabolites. Xenobiotica. 2003;33(9):913–25. https://doi.org/10.1080/ 0049825031000150444. 14. Caldwell CR, Britz SJ, Mirecki RM. Effect of temperature, elevated carbon dioxide, and drought during seed development on the isoflavone content of dwarf soybean [Glycinemax (L.) Merrill] grown in controlled environments. J Agric Food Chem. 2005;53(4):1125–9. https://doi.org/10. 1021/jf0355351. 1. Goldsmith PD. Economics of soybean production, marketing, and utiliza‑ tion. In: Soybeans. Elsevier; 2008. p. 117–50. https://doi.org/10.1016/ b978-1-893997-64-6.50008-1. Competing interests j 15. Eldridge AC, Kwolek WF. Soybean isoflavones: effect of environment and variety on composition. J Agric Food Chem. 1983;31(2):394–6. https://doi. org/10.1021/jf00116a052. 2. Lygin AV, Hill CB, Zernova OV, Crull L, Widholm JM, Hartman GL, et al. Response of soybean pathogens to glyceollin. Phytopathology. 2010;100(9):897–903. https://doi.org/10.1094/phyto-100-9-0897. References Large-scale identification, mapping, and genotyping of single-nucleotide polymor‑ phisms in the human genome. Sci. 1998;280(5366):1077–82. https://doi. org/10.1126/science.280.5366.1077. 46. Kim SH, Ryu J, Kim WJ, Kang R, Seo E, Kim G, et al. Identification of a new GmSACPD-C allele in high stearic acid mutant Hfa180 derived from gamma-ray irradiation. Mol Breed. 2019;39(2):1–12. https://doi.org/10. 1007/s11032-019-0928-0. 28. Njiti VN, Meksem K, Yuan J, Lightfoot DA, Banz WJ, Winters TA. DNA markers associated with loci underlying seed phytoestrogen content in soybeans. J Med Food. 1999;2(3–4):185–7. https://doi.org/10.1089/jmf. 1999.2.185. 47. Vuong T, Sonah H, Meinhardt C, Deshmukh R, Kadam S, Nelson R, et al. Genetic architecture of cyst nematode resistance revealed by genome- wide association study in soybean. BMC Genomics. 2015;16(1):1–13. https://doi.org/10.1186/s12864-015-1811-y. 29. Meksem K, Njiti V, Banz W, Iqbal M, Kassem MM, Hyten DL, et al. Genomic regions that underlie soybean seed isoflavone content. J Biomed Bio‑ technol. 2001;1(1):38–44. https://doi.org/10.1155/s1110724301000110. 48. Lee SB, Lee KS, Kim HY, Kim DY, Seo MS, Jeong SC, et al. The discovery of novel SNPs associated with group A soyasaponin biosynthesis from Korea soybean core collection. Genomics. 2022;114(4):110432. https://doi.org/ 10.1016/j.ygeno.2022.110432. g 30. Primomo VS, Poysa V, Ablett GR, Jackson CJ, Gijzen M, Rajcan I. Mapping QTL for individual and total isoflavone content in soybean seeds. Crop Sci. 2005;45(6):2454–64. https://doi.org/10.2135/cropsci2004.0672.i 49. Hu D, Kan G, Hu W, Li Y, Hao D, Li X, et al. Identification of loci and can‑ didate genes responsible for pod dehiscence in soybean via genome- wide association analysis across multiple environments. Front Plant Sci. 2019;10:811. https://doi.org/10.3389/fpls.2019.00811. 31. Zeng G, Li D, Han Y, Teng W, Wang J, Qiu L, et al. Identification of QTL underlying isoflavone contents in soybean seeds among multiple envi‑ ronments. Theor Appl Genet. 2009;118(8):1455–63. https://doi.org/10. 1007/s00122-009-0994-5. 50. Kharkwal M, Pandey R, Pawar S. Mutation breeding for crop improve‑ ment. In: Plant breeding. Springer; 2004. p. 601–45. https://doi.org/10. 1007/978-94-007-1040-5_26 32. Gutierrez-Gonzalez JJ, Vuong TD, Zhong R, Yu O, Lee JD, Shannon G, et al. Major locus and other novel additive and epistatic loci involved in modu‑ lation of isoflavone concentration in soybean seeds. Theor Appl Genet. 2011;123(8):1375–85. https://doi.org/10.1007/s00122-011-1673-x. 51. Hung NN, Kim D-G, Lyu JI, Park K-C, Kim JM, Kim J-B, et al. Detecting genetic mobility using a transposon-based marker system in gamma-ray irradiated soybean mutants. Plants. 2021;10(2):373. https://doi.org/10. 3390/plants10020373. 33. Elshire RJ, Glaubitz JC, Sun Q, Poland JA, Kawamoto K, Buckler ES, et al. References A robust, simple genotyping-by-sequencing (GBS) approach for high diversity species. PLoS ONE. 2011;6(5):e19379. https://doi.org/10.1371/ journal.pone.0019379. 52. Kim D-G, Lyu JI, Lee M-K, Kim JM, Hung NN, Hong MJ, et al. Construction of soybean mutant diversity pool (MDP) lines and an analysis of their genetic relationships and associations using TRAP markers. Agronomy. 2020;10(2):253. https://doi.org/10.3390/agronomy10020253. 2020;10(2):253. https://doi.org/10.3390/agronomy10020253. 34. Miller MR, Dunham JP, Amores A, Cresko WA, Johnson EA. Rapid and cost-effective polymorphism identification and genotyping using restric‑ tion site associated DNA (RAD) markers. Genome Res. 2007;17(2):240–8. https://doi.org/10.1101/gr.5681207. 53. Kim DG, Lyu JI, Lim YJ, Kim JM, Hung NN, Eom SH, et al. Differential gene expression associated with altered isoflavone and fatty acid contents in soybean mutant diversity pool. Plants. 2021;10(6):1037. 35. Sun X, Liu D, Zhang X, Li W, Liu H, Hong W, et al. SLAF-seq: an efficient method of large-scale de novo SNP discovery and genotyping using high-throughput sequencing. PLoS ONE. 2013;8(3):e58700. https://doi. org/10.1371/journal.pone.0058700. 54. Wu D, Li D, Zhao X, Zhan Y, Teng W, Qiu L, et al. Identification of a candi‑ date gene associated with isoflavone content in soybean seeds using genome-wide association and linkage mapping. Plant J. 2020;104(4):950– 63. https://doi.org/10.1111/tpj.14972. 55. Kim S, Hong E, Kim Y, Lee S, Park K, Kim H, et al. A new high protein and good seed quality soybean variet “Danbaegkong.” RDA J Agric Sci (Upland & Industrial Crops). 1996;38(1):228–32. 36. Thomson MJ. High-throughput SNP genotyping to accelerate crop improvement. Plant Breed Biotech. 2014;2(3):195–212. https://doi.org/10. 9787/pbb.2014.2.3.195. 37. Li B, Tian L, Zhang J, Huang L, Han F, Yan S, et al. Construction of a high- density genetic map based on large-scale markers developed by specific length amplified fragment sequencing (SLAF-seq) and its application to QTL analysis for isoflavone content in Glycinemax. BMC Genomics. 2014;15(1):1–16. https://doi.org/10.1186/1471-2164-15-1086. 56. Kim JM, Lyu JI, Kim DG, Hung NN, Seo JS, Ahn JW, et al. Genome wide association study to detect genetic regions related to isoflavone content in a mutant soybean population derived from radiation breeding. Front Plant Sci. 2022:2987. https://doi.org/10.3389/fpls.2022.968466. 57. Bindea G, Mlecnik B, Hackl H, Charoentong P, Tosolini M, Kirilovsky A, et al. ClueGO: a Cytoscape plug-in to decipher functionally grouped gene ontology and pathway annotation networks. Bioinformatics. 2009;25(8):1091–3. https://doi.org/10.1093/bioinformatics/btp101. 38. Pei R, Zhang J, Tian L, Zhang S, Han F, Yan S, et al. Identification of novel QTL associated with soybean isoflavone content. Crop J. 2018;6(3):244– 52. https://doi.org/10.1016/j.cj.2017.10.004. 39. References 20. Chapman A, Pantalone V, Ustun A, Allen F, Landau-Ellis D, Trigiano R, et al. Quantitative trait loci for agronomic and seed quality traits in an F2 and F4: 6 soybean population. Euphytica. 2003;129(3):387–93. https://doi.org/ 10.1023/A:1022282726117. 2. Lygin AV, Hill CB, Zernova OV, Crull L, Widholm JM, Hartman GL, et al. Response of soybean pathogens to glyceollin. Phytopathology. 2010;100(9):897–903. https://doi.org/10.1094/phyto-100-9-0897. 21. Chiari L, Piovesan ND, Naoe LK, José IC, Viana JMS, Moreira MA, et al. Genetic parameters relating isoflavone and protein content in soybean seeds. Euphytica. 2004;138(1):55–60. https://doi.org/10.1023/b:euph. 0000047060.03101.4a. Kim et al. BMC Plant Biology (2023) 23:517 Page 17 of 18 Kim et al. BMC Plant Biology (2023) 23:517 22. Fasoula VA, Harris DK, Boerma HR. Validation and designation of quan‑ titative trait loci for seed protein, seed oil, and seed weight from two soybean populations. Crop Sci. 2004;44(4):1218–25. https://doi.org/10. 2135/cropsci2004.1218. 42. Song Q, Hyten DL, Jia G, Quigley CV, Fickus EW, Nelson RL, et al. Develop‑ ment and evaluation of SoySNP50K, a high-density genotyping array for soybean. PLoS ONE. 2013;8(1):e54985. https://doi.org/10.1371/journal. pone.0054985. 23. Maughan P, Maroof M, Buss G. Identification of quantitative trait loci controlling sucrose content in soybean (Glycinemax). Mol Breed. 2000;6(1):105–11. https://doi.org/10.1023/A:1009628614988. 43. Lee YG, Jeong N, Kim JH, Lee K, Kim KH, Pirani A, et al. Development, validation and genetic analysis of a large soybean SNP genotyping array. Plant J. 2015;81(4):625–36. https://doi.org/10.1111/tpj.12755. 24. Wang D, Diers B, Arelli P, Shoemaker R. Loci underlying resistance to race 3 of soybean cyst nematode in Glycinesoja plant introduction 468916. Theor Appl Genet. 2001;103(4):561–6. https://doi.org/10.1007/pl00002910. 44. Wang J, Chu S, Zhang H, Zhu Y, Cheng H, Yu D. Development and application of a novel genome-wide SNP array reveals domestication history in soybean. Sci Rep. 2016;6(1):1–10. https://doi.org/10.1038/ srep20728. pp p g p 25. Botstein D, White RL, Skolnick M, Davis RW. Construction of a genetic link‑ age map in man using restriction fragment length polymorphisms. Amer J Hum Genet. 1980;32(3):314. 6247908. 45. Li YF, Li YH, Su SS, Reif JC, Qi ZM, Wang XB, et al. SoySNP618K array: a high-resolution single nucleotide polymorphism platform as a valuable genomic resource for soybean genetics and breeding. J Integr Plant Biol. 2022;64(3):632–48. https://doi.org/10.1111/jipb.13202. 26. Jeffreys AJ, Wilson V, Thein SL. Hypervariable ‘minisatellite’regions in human DNA. Nature. 1985;314(6006):67–73. https://doi.org/10.1038/314067a0. 27. Wang DG, Fan J-B, Siao C-J, Berno A, Young P, Sapolsky R, et al. Publisher’s Note 64. Farag MA, Huhman DV, Dixon RA, Sumner LW. Metabolomics reveals novel pathways and differential mechanistic and elicitor-specific responses in phenylpropanoid and isoflavonoid biosynthesis in Medica- gotruncatula cell cultures. Plant Physiol. 2008;146(2):387. https://doi.org/ 10.1104/pp.107.108431. Springer Nature remains neutral with regard to jurisdictional claims in pub‑ lished maps and institutional affiliations. Springer Nature remains neutral with regard to jurisdictional claims in pub‑ lished maps and institutional affiliations. 65. Iovine B, Iannella ML, Gasparri F, Giannini V, Monfrecola G, Bevilacqua MA. A comparative analysis of the photo-protective effects of soy isoflavones in their aglycone and glucoside forms. Int J Mol Sci. 2012;13(12):16444– 56. https://doi.org/10.3390/ijms131216444. p g j 66. Vitale DC, Piazza C, Melilli B, Drago F, Salomone S. Isoflavones: estrogenic activity, biological effect and bioavailability. Eur J Drug Metab Pharma‑ cokinet. 2013;38:15–25. https://doi.org/10.1007/s13318-012-0112-y. 67. Knizia D, Yuan J, Bellaloui N, Vuong T, Usovsky M, Song Q, et al. The soy‑ bean high density ‘forrest’by ‘williams 82’snp-based genetic linkage map identifies QTL and candidate genes for seed isoflavone content. Plants. 2021;10(10):2029. https://doi.org/10.3390/plants10102029. 68. Morrison M, Cober E, Saleem M, McLaughlin N, Frégeau-Reid J, Ma B, et al. Changes in isoflavone concentration with 58 years of genetic improvement of short-season soybean cultivars in Canada. Crop Sci. 2008;48(6):2201–8. https://doi.org/10.2135/cropsci2008.01.0023. 69. Zhao Q, Qin J, Li X, Liu B, Liu Y, Yang Q, et al. Coordinate inheritance of seed isoflavone and protein in soybean. Agriculture. 2022;12(8):1178. https://doi.org/10.3390/agriculture12081178. g g 70. Zhang J, Ge Y, Han F, Li B, Yan S, Sun J, et al. Isoflavone content of soybean cultivars from maturity group 0 to VI grown in northern and southern China. J Am Oil Chem Soc. 2014;91(6):1019–28. https://doi.org/10.1007/ s11746-014-2440-3. 71. Wang C, Sherrard M, Pagadala S, Wixon R, Scott RA. Isoflavone con‑ tent among maturity group 0 to II soybeans. J Am Oil Chem Soc. 2000;77(5):483–7. https://doi.org/10.1007/s11746-000-0077-6. 72. Gutierrez-Gonzalez JJ, Wu X, Zhang J, Lee J-D, Ellersieck M, Shannon JG, et al. Genetic control of soybean seed isoflavone content: importance of statistical model and epistasis in complex traits. Theor Appl Genet. 2009;119(6):1069–83. https://doi.org/10.1007/s00122-009-1109-z. 73. Yang K, Moon J-K, Jeong N, Chun H-K, Kang S-T, Back K, et al. Novel major quantitative trait loci regulating the content of isoflavone in soybean seeds. Genes Genom. 2011;33(6):685–92. https://doi.org/10.1007/ s13258-011-0043-z. 74. Kassem MA, Meksem K, Iqbal M, Njiti V, Banz W, Winters T, et al. Definition of soybean genomic regions that control seed phytoestrogen amounts. J Biomed Biotechnol. 2004;2004(1):52–60. References Cai Z, Cheng Y, Ma Z, Liu X, Ma Q, Xia Q, et al. Fine-mapping of QTLs for individual and total isoflavone content in soybean (Glycinemax L.) using a high-density genetic map. Theor Appl Genet. 2018;131(3):555–68. https://doi.org/10.1007/s00122-017-3018-x. 58. Doncheva NT, Morris JH, Gorodkin J, Jensen LJ. Cytoscape StringApp: network analysis and visualization of proteomics data. J Proteome Res. 2018;18(2):623–32. https://doi.org/10.1021/acs.jproteome.8b00702. 59. Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real-time quantitative PCR and the 2−ΔΔCT method. Methods. 2001;25(4):402–8. https://doi.org/10.1006/meth.2001.1262. 40. Schmutz J, Cannon SB, Schlueter J, Ma J, Mitros T, Nelson W, et al. Genome sequence of the palaeopolyploid soybean. Nature. 2010;463(7278):178– 83. https://doi.org/10.1038/nature08670. 60. Gutierrez-Gonzalez JJ, Wu X, Gillman JD, Lee JD, Zhong R, Yu O, et al. Intricate environment-modulated genetic networks control isoflavone accumulation in soybean seeds. BMC Plant Biol. 2010;10(1):1–16. https:// doi.org/10.1186/1471-2229-10-105. 41. Akond M, Liu S, Schoener L, Anderson JA, Kantartzi SK, Meksem K, et al. A SNP-based genetic linkage map of soybean using the SoySNP6K Illumina Infinium BeadChip genotyping array. Plant Genet Genom Biotech. 2013;1(3):80–9. https://doi.org/10.5147/pggb.v1i3.154. Kim et al. BMC Plant Biology (2023) 23:517 Kim et al. BMC Plant Biology (2023) 23:517 Page 18 of 18 Os4bglu12 β-glucosidase. BMC Plant Biol. 2006;6(1):1–19. https://doi.org/ 10.1186/1471-2229-6-33. 61. Yoshikawa T, Okumoto Y, Ogata D, Sayama T, Teraishi M, Terai M, et al. Transgressive segregation of isoflavone contents under the control of four QTLs in a cross between distantly related soybean varieties. Breeding Science. 2010;60(3):243–54. https://doi.org/10.1270/jsbbs.60.243. 80. Dong X, Jiang Y, Hur Y. Genome-wide analysis of glycoside hydrolase fam‑ ily 1 β-glucosidase genes in Brassicarapa and their potential role in pollen development. Int J Mol Sci. 2019;20(7):1663. https://doi.org/10.3390/ ijms20071663. 62. Han Y, Teng W, Wang Y, Zhao X, Wu L, Li D, et al. Unconditional and conditional QTL underlying the genetic interrelationships between soybean seed isoflavone, and protein or oil contents. Plant Breeding. 2015;134(3):300–9. https://doi.org/10.1111/pbr.12259. 81. Ishihara H, Tohge T, Viehöver P, Fernie AR, Weisshaar B, Stracke R. Natural variation in flavonol accumulation in Arabidopsis is determined by the flavonol glucosyltransferase BGLU6. J Exp Bot. 2016;67(5):1505–17. https://doi.org/10.1093/jxb/erv546. 63. Kassem M, Shultz J, Meksem K, Cho Y, Wood A, Iqbal M, et al. An updated ‘Essex’by ‘Forrest’linkage map and first composite interval map of QTL underlying six soybean traits. Theor Appl Genet. 2006;113(6):1015–26. https://doi.org/10.1007/s00122-006-0361-8. Publisher’s Note https://doi.org/10.1155/s1110 724304304018. 75. Roepke J, Bozzo GG. Arabidopsisthaliana β-glucosidase BGLU15 attacks flavonol 3-O-β-glucoside-7-O-α-rhamnosides. Phytochemistry. 2015;109:14–24. https://doi.org/10.1016/j.phytochem.2014.10.028. • fast, convenient online submission • thorough peer review by experienced researchers in your ﬁeld • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year • At BMC, research is always in progress. Learn more biomedcentral.com/submissions Ready to submit your research Ready to submit your research ? Choose BMC and benefit from: ? Choose BMC and benefit from: • fast, convenient online submission • thorough peer review by experienced researchers in your ﬁeld • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year • At BMC, research is always in progress. Learn more biomedcentral.com/submissions Ready to submit your research Ready to submit your research ? Choose BMC and benefit from: ? Choose BMC and benefit from: • fast, convenient online submission • thorough peer review by experienced researchers in your ﬁeld • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year • At BMC, research is always in progress. Learn more biomedcentral.com/submissions Ready to submit your research Ready to submit your research ? Choose BMC and benefit from: ? Choose BMC and benefit from: 76. Xu Z, Escamilla-Treviño L, Zeng L, Lalgondar M, Bevan D, Winkel B, et al. Functional genomic analysis of Arabidopsisthaliana glycoside hydrolase family 1. Plant Mol Biol. 2004;55(3):343–67. https://doi.org/10.1007/ s11103-004-0790-1. 77. Yang J, Ma L, Jiang W, Yao Y, Tang Y, Pang Y. Comprehensive identification and characterization of abiotic stress and hormone responsive glycosyl hydrolase family 1 genes in Medicagotruncatula. Plant Physiol Biochem. 2021;158:21–33. https://doi.org/10.1016/j.plaphy.2020.11.046. 78. Hsieh M-C, Graham TL. Partial purification and characterization of a soybean β-glucosidase with high specific activity towards isoflavone con‑ jugates. Phytochemistry. 2001;58(7):995–1005. https://doi.org/10.1016/ s0031-9422(01)00380-6. 79. Opassiri R, Pomthong B, Onkoksoong T, Akiyama T, Esen A, Ketudat Cairns JR. Analysis of rice glycosyl hydrolase family 1 and expression of 79. • fast, convenient online submission • thorough peer review by experienced researchers in your ﬁeld • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year • At BMC, research is always in progress. Learn more biomedcentral.com/submissions Ready to submit your research Ready to submit your research ? Choose BMC and benefit from: ? Choose BMC and benefit from: Publisher’s Note Opassiri R, Pomthong B, Onkoksoong T, Akiyama T, Esen A, Ketudat Cairns JR. Analysis of rice glycosyl hydrolase family 1 and expression of
https://openalex.org/W1600166721	https://authors.library.caltech.edu/65788/1/562.full.pdf	English	null	CD9, a major platelet cell surface glycoprotein, is a ROCA antigen and is expressed in the nervous system	The journal of neuroscience/The Journal of neuroscience	1,995	cc-by	11,104	CD9, a Major Platelet Cell Surface Glycoprotein, Is a ROCA Antigen and Is Expressed in the Nervous System Zaven Kaprielian, Kyung-Ok Cho, Michael Hadjiargyrou, and Paul H. Patterson Biology Division, California Institute of Technology, Pasadena, California 91125 We previously generated a monoclonal antibody (mAb), ROCA 1, which binds preferentially to rostral versus caudal sympathetic ganglia and intercostal nerves. Two other mAbs, ROCA2 and B2C11, bind to the same structures but not in rostrocaudal gradients. All three mAbs recognize a 26 kDa cell surface protein. Amino acid sequence data obtained from the affinity purified 26 kDa protein showed some ho- mology with human CD9, a tetraspan protein implicated in intercellular signaling in hematopoietic cells. Using the PCR, we obtained cDNA clones representing the entire rat CD9 coding sequence from sciatic nerve and sympathetic gan- glia. ROCA 1, ROCA2, and B2C11 each immunoprecipitate a 26 kDa protein from CHO cells stably transfected with one of the clones, demonstrating that the ROCA cell surface antigen is indeed rat CD9. We find that CD9 mRNA is widely expressed, with particularly high levels present in a number of neural tissues. In situ hybridization demonstrates that peripheral neurons and Schwann cells, as well as adrenal chromaffin cells express CD9 mRNA. Consistent with im- munoblot analyses showing that, unlike the ROCA 1 epitope, the 26 kDa protein is not expressed in a rostrocaudal gra- dient, we find similar levels of rat CD9 mRNA in rostral and caudal intercostal nerves. In developing postnatal rat sciatic nerve, CD9 mRNA levels are coordinately regulated with the expression of myelin genes. These results provide another example of a cell surface protein expressed by both he- matopoietic and neural cells, and suggest a role for CD9 in intercellular signaling in the nervous system. the migration of neurons and glia, as well as axon outgrowth, guidance, and fasciculation. Many of these proteins belong to four major cell adhesion receptor families: cadherins, lg super·· family, integrins, and selectins (Hynes and Lander, 1992). Mosr of these proteins possess single transmembrane domains or are linked to the membrane via a phosphoinositol anchor, and they contain one or more lg domains, an EGF or fibronectin repeat, and a cysteine-rich repeat or a kinase domain. Hematopoietic and neural cells express members of each family, with the ex- ception of the selectins, which have thus far been identified only in hematopoietic cells. CD9, a Major Platelet Cell Surface Glycoprotein, Is a ROCA Antigen and Is Expressed in the Nervous System A new superfamily of cell surface proteins known as the te- traspan (Gil et al., 1992) or transmembrane 4 superfamily (TM4SF; Wright et al., 1993) has recently been identified. Mem- bers of this family include (1) the leucocyte proteins CD9 (Boucheix et al., 1991; Lanza et al., 1991; Martin-Alonso et al., 1992; Mitamuraet al., 1992; Rubinstein eta!., I 993a,b), MRP-1 (identical to human CD9; Miyake et al., 1991), CD37 (Classon et al., 1989), CD53 (Amiot, 1990; Angelisova et al., 1990), OX- 44 antigen (rat homolog of human CD53; Bellacosa et al., 1991 ), CD63/ME 491 (Hotta et al., 1988; Metzelaar et al., 1991 ), ADI (rat homolog of human CD63/ME 491; Nishikata et al., 1992), R2/C33/IA4 (Gaugitsch et al., 1991; Imai et al., 1992; Gil et al., 1992, respectively), TAPA-1 (Oren et al., 1990; Andria et al., 1991 ), and A 15 (Emi et al., 1993); (2) the tumor-associated antigens C0-029 (Szala et al., 1990) and L6 (Marken et al., 1992); (3) the lung epithelial protein TI-I (Kallin et al., 1991); and (4) the schistosome surface proteins Sm23 (Wright et al., 1990) and Sj23 (Davern et al., 1991 ). The current model for the structure of superfamily members consists of four transmem- brane domains divided by two extracellular loops, with cyto- plasmic N- and C-termini (Horejsi and Vlcek, 1991; Levy et al., 1991; Reynolds et al., 1992; Tomlinson et al., 1993). While many of these proteins are known to be expressed in hemato- poietic and a variety of other cells, little information is available on their distribution in the nervous system (Kemshead et al., 1982; Komada et al., 1983; Ross et al., 1986; Von dem Borne et al., 1989; Oren et al., 1990). Although the in vivo function of these proteins has yet to be elucidated, antibody perturbation experiments implicate certain tetraspan family members in neu- ral cell adhesion, motility, and growth regulation (Anton et al., 1994; Hadjiargyrou and Patterson, 1994). [Key words: CD9, tetraspan family, ROCA antigen, neu- rospecification, Schwann cell, platelets, adrenal chromaffin cells, myelin, cell surface] Cell surface proteins play fundamental roles in cell-cell inter- actions in the nervous and hematopoietic systems. In the he- matopoietic system, some of these roles include platelet acti- vation and aggregation, lymphocyte trafficking, and tumor cell motility. In the nervous system, cell surface proteins play key roles in neuron-neuron and neuron-glia interactions, and direct We previously demonstrated that the ROCA! Received Mar. 29, 1994; revised June 13, 1994; accepted July 12, 1994. We thank Dr. Richard Akeson for mAb B2C 11; Doreen McDowell for assistance in platelet and media preparation; Li Ching Lo for assistance in in situ hybrid- ization; Lisa Banner, Ming Ji Fann, and Kai Zinn for help with primer selection and PCR analysis; and Joshua Sanes, Karen Allendorfer, and Shilpi Banerjee for critically reading the manuscript. This work was supported by an individual NRSA to Z.K., an American Heart Association Research Training fellowship to M.H., and a NINOS grant to P.H.P. Correspondence should be addressed to Dr. Zaven Kaprielian, Division of Biology, 216-76, California Institute of Technology, Pasadena, CA 91125. Copyright© 1995 Society for Neuroscience 0270-6474/95/150562-12$05.00/0 The Journal of Neuroscience, January 1995, 15(1 ): 562-573 The Journal of Neuroscience, January 1995, 15(1 ): 562-573 Copyright© 1995 Society for Neuroscience 0270-6474/95/150562-12$05.00/0 Correspondence should be addressed to Dr. Zaven Kaprielian, Division of Biology, 216-76, California Institute of Technology, Pasadena, CA 91125. Molecular cloning of rat CD9 Single strand cDNA was prepared from adult rat SCG and ScN RNA using Maloney murine leukemia virus reverse transcriptase (Promega) and oligo-dT primers. Two primers, NT! 7 and CTI 9, representing portions of the N- and C-terminus of human/bovine CD9, respectively, were synthesized as follows: NT17, 5'-ATGCCGGTCAAAGGAGG- 3'; CT19, 5'-AAGCTTGACTCTAGACCAT-3'. Rat CD9 DNA was amplified by adding these two primers and Taq DNA polymerase (Pro- mega) to SCG or ScN cDNA, and carrying out the following polymerase chain reaction (PCR): 40 cycles of 95°C for 1 min, 52°C for 1.5 min, 72°C for 1.5 min, and a final incubation at 72°C for 10 min. The PCR products amplified from both the SCG and ScN templates each con- tained a major band of the expected size (678 bp) on 2% agarose gels. These products were isolated and inserted into the Smal site of the pBluescript II SK+ (Stratagene) vector. ( , ) lmmunoprecipitation. Stably transfected CHO cell lines were grown to confluency in 100 mm tissue culture dishes containing DME-10% FBS. For metabolic labeling, the medium was replaced with 5 ml of DME (without cysteine or methionine or serum) to which 500 µCi (final concentration= I 00 µCi/ml) 35S-methionine-translabel (ICN) had been added, and the cells were then incubated for 3 hr at 3 7°C. Following the labeling period, cells were washed 3 x with 5 ml of phosphate- buffered saline (PBS; 0.9% NaCl, 100 mM NaP04 ) and lysed with 2 ml of protease-inhibitor containing low salt homogenization buffer (see above) containing I% CHAPS (Boehringer Mannheim). After multiple rounds of trituration and vortexing, the cell extract was centrifuged at I 00,000 x g for I hr in a Beckman Type 50 rotor. The resulting su- pernatant (2 ml) was mixed with 150 µI of Protein G-conjugated se- pharose beads (Pharmacia) and rotated end-over-end for 3 hr at room temperature. Then, 500 µI of the preabsorbed extract was mixed with 500 µl of ROCA!, ROCA2, or B2Cl I hybridoma supernatant and rotated end-over-end for 12-16 hr at 4°C. After this incubation, 30 µI of Protein G-conjugated sepharose beads were added to the mixture and the incubation was continued for another 3 hr at room temperature. The beads were then centrifuged in a microfuge, the supernatant re- moved, and the beads washed three times with PBS containing 0.05% Tween-20 by a series ofresuspensions and centrifugations. Protein preparations and immunoblotting NP40 extracts ofadult rat peripheral nerve membrane/cytoskeletal frac- tions were prepared as previously described (Kaprielian and Patterson, 1993). Platelet extracts were derived from membrane/cytoskeletal frac- tions of platelet rich plasma, prepared according to Boucheix et al., 1983. In brief, blood from freshly killed adult rats was first collected into 1/9 volume of0.13 M sodium citrate, and then centrifuged at 1000 rpm ( 150 x g) for 10 min in a Sorvall SS 34 rotor. The resulting pellet was discarded, while the supernatant, which contains the platelet-rich plasma (PRP), was then centrifuged at 100,000 x gfor I hrin a Beckman 70 Ti rotor. The pellets were resuspended in protease inhibitor-con- taining low-salt homogenization buffer (Kaprielian and Patterson, 1993). This material, which will be referred to as the platelet membrane/cy- toskeletal fraction, was either used immediately or frozen at -80°C until further use. Detergent extracts were obtained by solubilizing the mem- brane/cytoskeletal fraction with 1% NP40 as previously described (Ka- prielian and Patterson, 1993). Protein concentration was determined by the method of Lowry et al. ( 19 51 ), and immunoblots were performed as previously described (Kaprielian and Patterson, 1993). y Fusion protein analysis. The MBP-CD9 fusion protein was generated by in-frame fusion to a maltose binding protein (New England Biolabs). Briefly, to obtain a cDNA encoding the large, putative extracellular domain of rat CD9, a PCR was carried out using a primer corresponding to amino acids 111 to 116 (ATGGATCCCACAAGGACGAGGTGAT- T AA), and another corresponding to amino acids 184-190 (ATAAGCTTCACTTGCTGTGGAAGACCTC), along with Full ScN #13 as a template. To facilitate the cloning, a BamHI site was incor- porated into the 5' end of the forward primer(underlined), and a HindIII site was incorporated into the 5' end of the backward primer (under- lined). The PCR product was extracted with phenol and chloroform, digested with BamHI and HindIII, and ligated with pMal-c2 vector that had been digested with BamHI and HindIII. The expression of this fusion protein was induced by IPTG (isopropyl-~-o-thiogalactopyran oside), and sonicated E. coli extracts were subjected to SOS-PAGE (Smith and Johnson, 1988). Molecular cloning of rat CD9 After a final wash in PBS without detergent, the beads were placed in l x nonreducing sample buffer for 5 min to elute bound antigens (Kaprielian and Pat- terson, 1993). Proteins were subjected to one-dimensional SOS-PAGE, and stained and destained as previously described (Kaprielian and Pat- terson, 1993). Immediately prior to drying, the gels were incubated in Antibodies p g p q y Transfection. CHO cells were transfected by lipofectin (BRL) (Feigner et al., 1987). Briefly, cells were plated at 2.5 x 10'/100 mm dish and grown overnight at 37°C in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS, Hyclone) and penicillin-streptomycin (50 U/ml and 50 µglml, respectively). Li- pofectin (25 µg) was mixed with 5 µg rat CD9 cDNA (construct) in 200 µl of serum-free DMEM and added to the cells in 5 ml of serum-free medium. The cells were then incubated for 5 hr after which the lipo- fectin-DNA medium was replaced by 10 ml of serum-containing DMEM. Cells were finally incubated at 37°C for 48 hr prior to the addition of 400 µg!ml G4 l 8. Twelve days later colonies were isolated using cloning cylinders. ROCA! and ROCA2 are both IgG2b mouse mAbs that have been previously characterized (Suzue et al., 1990; Kaprielian and Patterson, 1993; Tole and Patterson, 1993). B2Cl I is an IgG2a mouse mAb that has also been previously characterized (Akeson and Warren, 1984). CD9, a Major Platelet Cell Surface Glycoprotein, Is a ROCA Antigen and Is Expressed in the Nervous System mAb recog- nizes two distinct proteins (60 kDa and 26 kDa) in membrane/ cytoskeletal fractions of peripheral nerves and ganglia. The 60 kDa protein is expressed at higher levels in rostral than in caudal intercostal nerves, and has been identified as peripherin (Ka- prielian and Patterson, 1993). In contrast, it is the ROCA! The Journal of Neuroscience, January 1995, 15(1) 563 epitope on the 26 kDa protein, and not the protein itself, that is preferentially visualized immunohistochemically in rostral nerves and ganglia (Suzue et al., 1990; Kaprielian and Patterson, 1993). We now have three mAbs against the 26 kDa protein, ROCA!, ROCA2 (Kaprielian and Patterson, 1993; Tole and Patterson, 1993), and B2Cl I (Akeson and Warren, l 984f. Ami- no acid sequence data obtained from the affinity purified 26 kDa protein raised the possibility that it could be the rat ho- molog of human CD9 (Kaprielian and Patterson, 1993). Here we describe the isolation of cDNA clones encoding rat CD9, confirm that it is the ROCA surface antigen, and define the regions of CD9 that contain the epitope for each mAb. In ad- dition, we report the tissue, cell type, positional, and develop- mental expression of rat CD9 mRNA. identical except that at each of four positions one of the clones contained a nucleotide that was different from the one present in the other three. This analysis yielded a consensus nucleotide sequence for rat CD9 (see Fig. 2). Eight additional PCR-isolated cDNAs (four each from ScN and SCG RNA) were sequenced in one direction using the primers listed above. Six of these proved to be identical to the consensus rat CD9 sequence, while two contained single nucleotide differences. We attrib- ute all of the single nucleotide changes to PCR errors. Searches for related sequences were performed using the BLAST program. Expression of rat CD9 in CHO cells Subcloning of CD9 cDNA. The Full ScN #13 cDNA was inserted into the pBluescript II SK+ vector (described above). Digestion with BamHI and Xhol yielded a fragment containing all of the CD9 cDNA and approximately 50 bp of vector. This fragment was then subcloned into the BamHl/XhoI linearized eukaryotic expression vector, pcDNAI/neo (Invitrogen). After transformation of bacteria, a plasmid with the intact 5' -3' CD9-cDNA insert was selected. Large amounts of this DNA were purified (Qiagen plasmid kit) and subsequently used for transfection. Sc.N. PLAT. Sc.N. PLAT. Sc.N. PLAT. Figure J. ROCA 1, ROCA2, and B2CI 1 each bind a 26 kDa protein in extracts of adult rat sciatic nerves and platelets. One percent NP-40 extracts of adult rat sciatic nerve and platelet membrane/cytoskeletal fractions were subjected to SOS-PAGE and immunoblot analysis as described in Materials and Methods. Each mAb binds a 26 kDa protein in both sciatic nerve (Sc.N.) and platelet (PLAT.) extracts. Primary antibody binding was visualized with a peroxidase-conjugated goat anti-mouse IgG secondary antibody. Approximately 40 µg of protein was loaded in each lane. phor screens with a Phosphorimager 400S (Molecular Dynamics) and computations performed with lMAOEQUANT software. Amplify (Amersham) for 30 min at room temperature. For autora- diography, dried gels were exposed to Kodak X-OMAT AR film for 12-24 hr at - 80"C. phor screens with a Phosphorimager 400S (Molecular Dynamics) and computations performed with lMAOEQUANT software. p p In situ hybridization. The Full ScN # 13 cDNA was linearized with BamHI or EcoRI and transcribed with T7 or T3 RNA polymerase (MEGAscript Transcription Kit, Ambion) to generate antisense and sense rat CD9 RNA probes, respectively. Nonradioactive labeling was achieved by incorporating digoxygenin-11-UTP (Boehringer Mann- heim) into the RNA probes during the transcription reaction. Tissue sections (15 µm thick) on glass slides were hybridized in plastic slide mailers with digoxygenin-labeled probes by a modification (Birren et al., 1993) of the procedure of Harland (1991) with the following excep- tions. All tissue sections were placed on untreated Super frost/Plus microscope slides (Fisher Scientific). Prehybridization and hybridiza- tion steps were carried out at 50°C (instead of 60"C) for those sections containing sciatic or intercostal nerves. The alkaline phosphatase-«>n- jugated anti-digoxygenin Fab fragments were preabsorbed with extracts of tissue parts dissected from developing or adult rats, depending on the age of the tissue to be examined. Immunocytochemistry. Transfected CHO cells were harvested, re- plated on sterile glass coverslips in serum-containing DMEM and grown for another 48 hr. Surface labeling of the cells with ROCA2 and B2Cl 1 hybridoma supematants was carried out as described in Hadjiargyrou and Patterson, 1994. Cells were photographed using a Nikon fluores- cence microscope (Diaphot 300). p ( p ) RNA Jsolaiion and Northern analysis. Total RNA was prepared from various tissues as described by (Chomczynski and Sacchi 1987. RNA was separated on 1.5% agarose-formaldehyde gels, transferred to Hy- bond nylon membrane (Amersham), and UV cross-linked (Stratagene). Sc.N. PLAT. Sc.N. PLAT. Sc.N. PLAT. The PCR was used to generate CD9, MBP and P0 DNA by using the cloned cDNAs as templates, along with universal and reverse PCR primers. DNA for the 3' end of 28S rRNA was amplified using rat genomic DNA as the template and the following two PCR primers: forward, TGAAGCAGAA TTCACCAAGC; backward, GTCGAGGG- CTGACTTTCAA T (Choi, 1985). Each of the DNA probes were labeled with 32P by the random priming method (Feinberg and Vogelstein, 1983). Blots were hybridized and washed according to Cho et al., 1991 . The same blot was reused multiple times by placing the filter in boiling water for 1 min to strip previously bound radioactivity. Blots were exposed to either Kodak X-OMAT AR film or Molecular Dynamics Phosphor screens. Radioactivity was quantitated by scanning the Phos- DNA sequence analysis Two, independent subclones derived from each of the original PCR products named [Full ScN #13 and Full ScN #42 (from ScN cDNA), and Full SCG #3 and Full SCG #8 (from SCG cDNA)] were sequenced by the dideoxy chain termination method (Sanger et al., 1977) using double-stranded DNA as template and T7 DNA polymerase (Sequen- ase, U.S. Biochemicals). Both strands of each subclone were sequenced not only with the T3, SK, KS, and T7 primers (Stratagene), but also with the four internal primers: RCD9 189, 5'-GCTGGGGCCCTCAT- GATG-3'; RCD9 255, 5'-CTGGGATTGTTCTTCGGA-3'; RCD9 490, 5'-GAACAAGGATGAGCCCAG-3'; RCD9 550, 5'-GACATCTG- CCCCAAAAAG-3'. The sequences of these subclones proved to be 564 Kaprielian et al. • Rat CD9 Is a ROCA Antigen and Is Expressed in Neural Cells ROCA1 ROCA2 B2C11 -~~~ +-26kD Sc.N. PLAT. Sc.N. PLAT. Sc.N. PLAT. B2C11 Results Expression of the ROCA surface antigen in platelets To determine whether the ROCA mAbs recognize a 26 kDa protein in hematopoietic cells we performed immunoblots on Expression of the ROCA surface antigen in platelets To determine whether the ROCA mAbs recognize a 26 kDa protein in hematopoietic cells we performed immunoblots on Figure 2. Top, Amino acid sequence alignment of RAT, MOUSE, HUMAN, MONKEY, and BOVINE CD9. The putative transmembrane domains (TMJ- IV) are shaded. Identical residues between sequences are indicated by dashes. Bottom, Proposed topological model ofCD9. Open circles represent residues that are identical across species. Solid circles identify interspecies amino acid differences. The two lightly shaded circles identify the positions of the two amino acids that are missing in rat CD9 (see top). The asterisk denotes the asparagine residue that is a possible site for N-linked glycosylation. RAT MOUSE HUMAN MONKEY BOVINE RAT MOUSE HUMAN MONKEY BOVINE RAT MOUSE HUMAN MONKEY BOVINE RAT MOUSE HUMAN MONKEY BOVINE RAT MOUSE HUMAN MONKEY BOVINE 1 MPVKGGSKCI ------s--- ------T--- ------T--- ------T--- 51 NHS NNH- NNN- NND- 101 151 NCCGIAGGVE D---IA-PL- N---LA-GV- D---LA-GV- D---LT-VP- TMI TMII 50 TKSIFEQET -------- --------TN ---- ---- TN -------- TMIII 100 150 ·:'~HI<DEVIKE LQEFYKDTYQ KLRNKDEPQR ETLKAIHMAL --D----- L-E--K---Q --RS------ - ------M-- ·--o----- V-E--K---N --KT------ -------Y-- --D----- V-E--K---N --KT------ -------Y- - --E----- V-K--E---N --KN------ - ---- --I-- 200 QFISDICPKK QVLESFQVKS CPDAIDEVFH - -IS-T--K- QLLESFQVKP --E--S-V-N --IS-I--K- DVLETFTVKS --D--K-V-D N--- --IS-I--K- DVLETFTIKS --D--K-V-D N- -- --LT-T--P- NLIDSLKTRP - -E--D-I-R s--- 201 TMIV 228 -''''~"'' RRSREMV --S-E-- --N-E-- --N-E-- --N-D-- Cytoplasm RAT MOUSE HUMAN MONKEY BOVINE RAT MOUSE HUMAN MONKEY BOVINE RAT MOUSE HUMAN MONKEY BOVINE RAT MOUSE HUMAN MONKEY BOVINE RAT MOUSE HUMAN MONKEY BOVINE 1 MPVKGGSKCI ------s--- ------T--- ------T--- ------T--- 51 NHS NNH- NNN- NND- 101 151 NCCGIAGGVE D---IA-PL- N---LA-GV- D---LA-GV- D---LT-VP- TMI TMII 50 TKSIFEQET -------- --------TN ---- ---- TN -------- TMIII 100 150 ·:'~HI<DEVIKE LQEFYKDTYQ KLRNKDEPQR ETLKAIHMAL --D----- L-E--K---Q --RS------ - ------M-- ·--o----- V-E--K---N --KT------ -------Y-- --D----- V-E--K---N --KT------ -------Y- - --E----- V-K--E---N --KN------ - ---- --I-- 200 QFISDICPKK QVLESFQVKS CPDAIDEVFH - -IS-T--K- QLLESFQVKP --E--S-V-N --IS-I--K- DVLETFTVKS --D--K-V-D N--- --IS-I--K- DVLETFTIKS --D--K-V-D N- -- --LT-T--P- NLIDSLKTRP - -E--D-I-R s--- 201 TMIV 228 -''''~"'' RRSREMV --S-E-- --N-E-- --N-E-- --N-D-- Cytoplas TMII RAT MOUSE HUMAN MONKEY BOVINE 201 TMIV 228 -''''~"'' RRSREMV --S-E-- --N-E-- --N-E-- --N-D-- Cytoplasm Cytoplasm 566 Kaprielian et al. Results • Rat CD9 Is a ROCA Antigen and Is Expressed in Neural Cells 566 Kaprielian et al. • Rat CD9 Is a ROCA Antigen and Is Expressed in Neural Cells 566 Kaprielian et al. • Rat CD9 Is a ROCA Antigen and Is Expressed in Neural Cells A. 1 2 3 .,_CD9 B. RC>CA."1 RC>CA.2 B2C"1 "1 c. R.OC.A."1 RC>CA.2 B2C1 "1 Figure 3. The 26 kDa ROCA cell surface antigen is rat CD9. A, ROCAI (lane I), ROCA2 (lane 2), and B2Cl l (lane J)each immunoprecipitat a 26 kDa protein from a 1% CHAPS extract of" S-Met labeled CHO cells that had been stably transfected with the Full ScN. #13 cDNA, encodin rat CD9. The stably transfected CHO cell clone used in this particular experiment was CD9-y. The arrow indicates the position ofCD9 (26 kDa) B, ROCA2 and B2Cl I, but not ROCAI label the surfaces ofCD9'Y cells. Live CD9'Y cells were sequentially incubated with the various mAbs and a goat anti-mouse lgG ftuorescein-conjugated secondary antibody. The top portion of each panel is a phase micrograph, while the bottom portion is a fluorescence micrograph of the same field. C, ROCAI, ROCA2, and B2Cl 1 do not label the surfaces ofnontransfectcd CHO cells. Live CHO cells were sequentially incubated with the various mAbs and a goat anti-mouse lgG ftuorescein-conjugated secondary antibody. The top portion of each panel is a phase micrograph, while the bottom portion is a fluorescence micrograph of the same field. B. RC>CA."1 RC>CA.2 B2C"1 "1 RC>CA.2 B. B2C"1 "1 RC>CA."1 c. R.OC.A."1 RC>CA.2 B2C1 "1 Figure 3. The 26 kDa ROCA cell surface antigen is rat CD9. A, ROCAI (lane I), ROCA2 (lane 2), and B2Cl l (lane J)each immunoprecipitate a 26 kDa protein from a 1% CHAPS extract of" S-Met labeled CHO cells that had been stably transfected with the Full ScN. #13 cDNA, encoding rat CD9. The stably transfected CHO cell clone used in this particular experiment was CD9-y. The arrow indicates the position ofCD9 (26 kDa). B, ROCA2 and B2Cl I, but not ROCAI label the surfaces ofCD9'Y cells. Live CD9'Y cells were sequentially incubated with the various mAbs and a goat anti-mouse lgG ftuorescein-conjugated secondary antibody. The top portion of each panel is a phase micrograph, while the bottom portion is a fluorescence micrograph of the same field. C, ROCAI, ROCA2, and B2Cl 1 do not label the surfaces ofnontransfectcd CHO cells. Results Live CHO cells were sequentially incubated with the various mAbs and a goat anti-mouse lgG ftuorescein-conjugated secondary antibody. The top portion of each panel is a phase micrograph, while the bottom portion is a fluorescence micrograph of the same field. c. R.OC.A."1 RC>CA.2 B2C1 "1 B2C1 "1 c. Figure 3. The 26 kDa ROCA cell surface antigen is rat CD9. A, ROCAI (lane I), ROCA2 (lane 2), and B2Cl l (lane J)each immunoprecipitate a 26 kDa protein from a 1% CHAPS extract of" S-Met labeled CHO cells that had been stably transfected with the Full ScN. #13 cDNA, encoding rat CD9. The stably transfected CHO cell clone used in this particular experiment was CD9-y. The arrow indicates the position ofCD9 (26 kDa). B, ROCA2 and B2Cl I, but not ROCAI label the surfaces ofCD9'Y cells. Live CD9'Y cells were sequentially incubated with the various mAbs and a goat anti-mouse lgG ftuorescein-conjugated secondary antibody. The top portion of each panel is a phase micrograph, while the bottom portion is a fluorescence micrograph of the same field. C, ROCAI, ROCA2, and B2Cl 1 do not label the surfaces ofnontransfectcd CHO cells. Live CHO cells were sequentially incubated with the various mAbs and a goat anti-mouse lgG ftuorescein-conjugated secondary antibody. The top portion of each panel is a phase micrograph, while the bottom portion is a fluorescence micrograph of the same field. The Joumal of Neuroscience, January 1995, 15(1) 567 + + + - ROCA1 ROCA2 B2C11 + CONTROL + + + - ROCA1 ROCA2 B2C11 + CONTROL Figure 4. R OCA2 and B2C 11 bind the large, putative extracellular domain of rat CD9. Proteins within extracts gen- erated from E. coli expressing the MBP- CD9 fusion protein ( +) or MBP (-) were subjected to SDS-PAGE and im- munoblot analysis. ROCA2 and B2Cl 1 but not ROCA I, bind a 49 kDa protein present only in extracts containing the MBP-CD9 fusion protein. Primary an- tibody binding was visualized with a peroxidase-conjugated goat anti-mouse lgG secondary antibody. The control reflects the binding of the secondary an- tibody alone. Figure 4. R OCA2 and B2C 11 bind the large, putative extracellular domain of rat CD9. Proteins within extracts gen- erated from E. coli expressing the MBP- CD9 fusion protein ( +) or MBP (-) were subjected to SDS-PAGE and im- munoblot analysis. Isolation of rat CD9 cDNA clones Isolation of rat CD9 cDNA clones To clone the cDNA encoding rat CD9, we used the PCR with nondegenerate oligonucleotide primers corresponding to the NH2-terminal and COOH-terminal sequences ofhuman/bovine CD9. Amplifications were performed using cDNA templates derived from adult rat sciatic nerve or superior cervical sym- pathetic ganglia mRNA. In both cases, 690 bp products were obtained (data not shown). This corresponds to the expected size for a cDNA clone encoding the entire rat CD9 protein based upon the nucleotide sequences of human and bovine CD9. Two cDNAs obtained using the sciatic nerve (ScN) template, full ScN# 13 and #42, as well as two cDNAs obtained using the superior cervical sympathetic ganglia (SCG) template, full SCG#3 and #8, were sequenced in both directions (see Materials and Methods), and found to be nearly identical (see Materials and Methods). In addition, we have isolated and characterized three full length cDNAs encoding rat CD9 from a postnatal day 9 rat CONTROL detergent-solubilized membrane/cytoskeletal fractions of adult rat platelets using ROCA! and ROCA2. We also used mAb B2Cl I (Akeson and Warren, 1984) after finding that this mAb binds strongly to the affinity-purified 26 kDa ROCAl/2 antigen from adult rat peripheral nerves (data not shown; mAb B2Cl I is further characterized in the accompanying article; Hadjiar- gyrou and Patterson, 1994). Each mAb recognizes a 26 kDa protein in the platelet extract that co migrates with the previously described peripheral nerve protein (Fig. I). In addition, ROCA2 and B2C 11 each bind larger, glycosylated forms of the 26 kDa protein (Kaprielian and Patterson, 1993) in both platelet and nerve extracts. These results are consistent with the possibility that the 26 kDa ROCA antigen may be rat CD9. sciatic nerve cDNA library. The sequences of these clones pre- cisely match those of the PCR products (data not shown). The rat protein is 95% identical to mouse CD9, and 93%, 92%, and 83% identical to the human, monkey, and bovine proteins, respectively (Fig. 2). Rat, mouse, and bovine CD9 each consist of 226 amino acid residues, while human and monkey CD9 are comprised of 228 amino acids. Only rat and mouse CD9 contain a string of seven amino acids between positions 120 and 126, ELQEFYK, which is identical to the sequence of a tryptic peptide generated from the affinity-purified 26 kDa protein (Fig. 2 and Kaprielian and Patterson, 1993). At position 50 in the rat sequence there exists one putative, N-linked gly- cosylation site (NHS). A similar site is present in the mouse, human, monkey and bovine sequences; each of these sequences, however, possesses additional asparagine residues immediately before the consensus asparagine. Mouse and bovine CD9 con- tain one, while human and monkey CD9 contain two additional asparagines (Fig. 2). Consistent with previous observations (Ru- binstein et al., I 993b), the majority of interspecies amino acid differences in the CD9 protein are present in the large extra- cellular domain (Fig. 2). Results ROCA2 and B2Cl 1, but not ROCA I, bind a 49 kDa protein present only in extracts containing the MBP-CD9 fusion protein. Primary an- tibody binding was visualized with a peroxidase-conjugated goat anti-mouse lgG secondary antibody. The control reflects the binding of the secondary an- tibody alone. B2C11 B2C11 CONTROL Expression of rat CD9 mRNA Tissue distribution. Northern blot analysis was performed using total RNA from various adult rat tissues and the full ScN #13 rat CD9 cDNA probe. This probe identifies a 1.3 kb RNA band in a wide range of tissues (Fig. 5A). For each tissue, the ratio of CD9 mRNA to 28S rRNA was calculated (Fig. 5B). Based upon this analysis, adult rat platelets are the most abundant source of CD9 mRNA. This is consistent with the known distribution ofCD9 in other species. Superior cervical ganglia (SCG), sciatic nerve (ScN) and spinal cord all express high levels of CD9 mRNA. Low, but clearly detectable levels are present in dorsal root ganglia (DRG) and cerebellum, but very little signal is detected in the cerebrum. While significant levels ofCD9 mRNA are present in a number of non-neural tissues such as esophagus, heart, lung, adrenal, spleen, and kidney, no hybridization was observed to liver or skeletal muscle RNA. The rat CD9 probe also hybridizes to a minor, 3.2 kb species in superior cervical ganglia and sciatic nerve RNA (Fig. 5A; see Discussion). B '1 320 20 Cf) GO C\I -- 0) c () 10 0 U) :::l t O> Cl1 Cl1 Q) -aI 0 U) w l B B '1 320 20 Cf) GO C\I -- 0) c () 10 0 U) :::l t O> Cl1 Cl1 Q) -aI 0 U) w l Figure 5. Rat CD9 mRNA is expressed by a wide range ofneural and non-neural tissues. A, Northern blot analysis was performed using 2 µg of total RNA derived from various adult rat tissues and 32P-labeled Full ScN #13 (rat CD9) and 28S cDNA probes. The top panel shows the hybridization of the rat CD9 probe to a 1.3 kb species present in many tissues, and the bottom panel shows the hybridization of the 28S probe to all tissues. The apparent size difference between platelet CD9 mRNA and CD9 mRNAs in other tissues is due to frowning of this particular gel and the placement of the platelet lane at the extreme edge of the gel. B, The radioactivity in all lanes was quantified using a Phosphorimager, and plotted as the ratio ofCD9 mRNA to 28S rRNA versus tissue type. l l Cell type specific expression. In situ hybridization studies were performed to unambiguously determine whether neurons as well as glia express CD9 mRNA. Expression of rat CD9 mRNA Digoxygenin-labeled antisense and sense RNA probes derived from the full ScN #13 cDNA were hybridized to sections of selected neural and non-neural tissues. The cytoplasm of adult rat SCG and DRG neurons is clearly labeled by the antisense probe (Fig. 6a,c). While most of the neurons present within SCG sections are labeled by the rat CD9 probe, only a small subset of DRG neurons detectably express CD9 mRNA. This finding is consistent with a significantly lower level of CD9 mRNA in the DRG as compared to the SCG, as determined by Northern analysis. Schwann cells present within sections of adult rat peripheral nerve also clearly express CD9 mRNA (Fig. 6b). The identification of these cells as Schwann cells is based upon (1) labeling of very similar appearing cells using Po and MBP RNA probes (data not shown), (2) the bi- polar, longitudinam morphology of the labelled cells, and (3) the large number ofCD9+ cells. In addition, the cytoplasm of chro- maffin cells in the adrenal medulla (Fig. 6d) and epithelial cells in the esophagus (data not shown) strongly label with the an- tisense CD9 probe. In all cases, no hybridization was detected with the sense probe (data not shown). Figure 5. Rat CD9 mRNA is expressed by a wide range ofneural and non-neural tissues. A, Northern blot analysis was performed using 2 µg of total RNA derived from various adult rat tissues and 32P-labeled Full ScN #13 (rat CD9) and 28S cDNA probes. The top panel shows the hybridization of the rat CD9 probe to a 1.3 kb species present in many tissues, and the bottom panel shows the hybridization of the 28S probe to all tissues. The apparent size difference between platelet CD9 mRNA and CD9 mRNAs in other tissues is due to frowning of this particular gel and the placement of the platelet lane at the extreme edge of the gel. B, The radioactivity in all lanes was quantified using a Phosphorimager, and plotted as the ratio ofCD9 mRNA to 28S rRNA versus tissue type. Figure 5. Rat CD9 mRNA is expressed by a wide range ofneural and non-neural tissues. A, Northern blot analysis was performed using 2 µg of total RNA derived from various adult rat tissues and 32P-labeled Full ScN #13 (rat CD9) and 28S cDNA probes. The 26 kDa ROCA antigen is rat CD9 The 26 kDa ROCA antigen is rat CD9 To confirm the identity of the 26 kDa ROCA antigen, the rat CD9 cDNA clone, full ScN # 13, was subcloned into the pc- DNAI/neo expression vector (InVitrogen) and introduced into CHO cells. Multiple, stably transfected, clonal CHO cell lines were isolated through 0418 selection. ROCAl, ROCA2, and B2CI I each immunoprecipitate a 26 kDa protein from deter- gent extracts of such cell lines (Fig. 3A). Isotype-matched mAbs do not immunoprecipitate CD9 (data not shown). In addition, ROCA2 and B2C I I, but not ROCA!, strongly and uniformly 568 Kaprielian et al. • Rat CD9 Is a ROCA Antigen and Is Expressed In Neural Cells 568 Kaprielian et al. • Rat CD9 Is a ROCA Antigen and Is Expressed In Neural Cells 568 Kaprielian et al. • Rat CD9 Is a ROCA Antigen and Is Expressed In Neural Ce A 3.2 Kb ->- CD9 1.3Kb ~ 28S B '1 320 20 Cf) GO C\I -- 0) c () 10 0 U) :::l t O> Cl1 Cl1 Q) -aI 0 U) w l Figure 5. Rat CD9 mRNA is expressed by a wide range ofneural and non neural tissues A Northern blot analysis was performed using 2 µg A 3.2 Kb ->- CD9 1.3Kb ~ 28S label the surfaces of these same cells (Fig. 38). None of our anti- rat CD9 mAbs bind the surfaces of (Fig. 3C), or immunopre- cipitate any protein from (data not shown), the mock-transfected CHO cells (Fig. 3C). A To prove that ROCA2 and B2Cl I recognize the large, pu- tative extracellular domain ofCD9, we generated a fusion pro- tein (see Materials and Methods) containing residues 111-190 of rat CD9. Consistent with the fact that ROCA2 and B2CJ 1 bind living cells expressing the entire rat CD9 protein, ROCA2 and B2CI I, but not ROCA I, bind this fusion protein (Fig. 4). Expression of rat CD9 mRNA Expression of rat CD9 mRNA The left portion of the adrenal gland section (d) contains strongly labeled chromaffin cells of the adrenal medulla, while the right portion shows the unlabeled cortex. cific for Po and MBP, two myelin genes whose expression is known to be developmentally regulated in these nerves (Wiggins et al., 1975; Lees and Brostoff, 1984; Lemke and Axel, 1985; Stahl et al., 1990). Northern analysis shows that the mRNA expression of all three genes dramatically rises from a low, but clearly detectable level in newborn sciatic nerves, to maximum levels by postnatal day 14 (Fig. 8A). By day 60, CD9, MBP, and Po mRNA expression is downregulated to adult levels. Jn situ hybridization performed with the full ScN #13 CD9 antisense probe confirms these results. While only a small number of scattered Schwann cells express CD9 mRNA in the newborn ScN, large numbers are strongly labeled by postnatal days 7 and 14 (Fig. 8B). By day 60, the number oflabeled Schwann cells is reduced to adult levels. Similar results were obtained with a MBP antisense probe (data not shown). a.. m ~ -- 0> c () T2 T4 T6 T8 T10 T12 lntercostal Nerves a.. m ~ -- 0> c () T2 T4 T6 T8 T10 T12 lntercostal Nerves Figure 7. The expression of rat CD9 mRNA is not graded along the rostrocaudal axis. Northern blot analysis was performed using 2 µ.g of total RNA derived from adult rat intercostal nerves taken from various rostrocaudal levels, and 32P-labeled Full ScN # 13 (rat CD9) and MBP cDNA probes. The identical transfer was probed with each cDNA in succession after removal of previously bound radioactivity. The radio- activity representing the hybridization of each probe was quantified using a Phosphorimager and the ratio of CD9 mRNA to MBP mRNA plotted versus rostrocaudal position. Expression of rat CD9 mRNA The top panel shows the hybridization of the rat CD9 probe to a 1.3 kb species present in many tissues, and the bottom panel shows the hybridization of the 28S probe to all tissues. The apparent size difference between platelet CD9 mRNA and CD9 mRNAs in other tissues is due to frowning of this particular gel and the placement of the platelet lane at the extreme edge of the gel. B, The radioactivity in all lanes was quantified using a Phosphorimager, and plotted as the ratio ofCD9 mRNA to 28S rRNA versus tissue type. Positional and developmental expression. We also examined the rostrocaudal distribution ofCD9 mRNA. Northern analysis detects essentially equivalent levels of the 1.3 kb CD9 mRNA (normalized to the amount of myelin basic protein (MBP) mRNA) in adult rat intercostal nerves taken from various ros- trocaudal levels (Fig. 7). Similarly, the number ofSchwann cells expressing rat CD9 mRNA does not appear to vary significantly between rostral and caudal nerves (data not shown). Northern and in situ analyses were also used to determine if CD9 mRNA expression is developmentally regulated in post- natal rat ScNs. We performed similar analyses with probes spe- The Journal of Neuroscience, January 1995, 15(1) H9 a b - --- c t 0 . O· Figure 6. Rat CD9 mRNA is expressed by peripheral neurons, Schwann cells and chromaffin cells. Digoxygenin-labeled antisense RNA probes derived from the Full ScN # 13 cDNA were hybridized to cryosections of adult rat SCG (a), sciatic nerve (b), DRG (c), and adrenal gland (d). The cytoplasm of neurons present within the SCG and DRG sections (a, c) and bipolar Schwann cells in the sciatic nerve section (b) are all brightly labeled. The left portion of the adrenal gland section (d) contains strongly labeled chromaffin cells of the adrenal medulla, while the right portion shows the unlabeled cortex. a b - --- b c Figure 6. Rat CD9 mRNA is expressed by peripheral neurons, Schwann cells and chromaffin cells. Digoxygenin-labeled antisense RNA probes derived from the Full ScN # 13 cDNA were hybridized to cryosections of adult rat SCG (a), sciatic nerve (b), DRG (c), and adrenal gland (d). The cytoplasm of neurons present within the SCG and DRG sections (a, c) and bipolar Schwann cells in the sciatic nerve section (b) are all brightly labeled. The ROCA cell surface 26 kDa protein is rat CD9 The identical transfer was probed with each cDNA in succession after removal of previously-bound ra- dioactivity. The radioactivity representing the hybridization of each probe was quantified using a Phosphorimager and the ratio of the various mRNAs to 28S rRNA plotted versus developmental age. B, Digoxy- genin-labeled antisense RNA probes derived from the Full ScN #13 cDNA were hybridized to cryosections of developing postnatal rat sci- atic nerves. The number of Schwann cells expressing rat CD9 signifi- cantly increases by postnatal day 7, and then decreases to adult levels by postnatal day 60. Figure 8. The expression of rat CD9 mRNA is developmentally reg- ulated in postnatal sciatic nerves. A, Northern blot analysis was per- formed using 2 µg of total RNA derived from sciatic nerves taken from postnatal rats of various ages, and 32P-labeled Full ScN #13 (rat CD9), MBP, Po, and 28S cDNA probes. The identical transfer was probed with each cDNA in succession after removal of previously-bound ra- dioactivity. The radioactivity representing the hybridization of each probe was quantified using a Phosphorimager and the ratio of the various mRNAs to 28S rRNA plotted versus developmental age. B, Digoxy- genin-labeled antisense RNA probes derived from the Full ScN #13 cDNA were hybridized to cryosections of developing postnatal rat sci- atic nerves. The number of Schwann cells expressing rat CD9 signifi- cantly increases by postnatal day 7, and then decreases to adult levels by postnatal day 60. ) The deduced rat CD9 protein sequence contains the same string of seven amino acids, ELQEFYK, as one of the tryptic peptides generated from the affinity purified 26 kDa protein (Kaprielian and Patterson, 1993). The N-terminal 14 amino acids (deduced from the cDNA) of the rat protein, PVKGGSKC/KYLLF, are similar to those obtained by se- quencing the N-terminus of the affinity-purified ScN protein, XVKGGIDEVFYLLF (Kaprielian and Patterson, 1993). All 12 of the rat CD9 cDNAs, contain the amino acid string, SKCIK, not IDEVF, between positions 6 and 10 within this string. This discrepancy may be due to protein sequencing errors or to the existence of two forms of rat CD9 with different N-terminal sequences. Support for the protein sequencing error explanation comes from previous attempts to sequence the N-terminus of human CD9; three different groups failed to unambiguously determine the identity of all five residues (Higashara et al., 1990; Boucheix et al., 1991; Lanza et al., 1991). The ROCA cell surface 26 kDa protein is rat CD9 in vivo observations (Kaprielian and Patterson, 1993), these results suggest that the ROCA! epitope may not always exist in an accessible state on the cell surface. In this regard, the transfected CHO cells may represent a model system for further study of epitope masking on CD9. Discussion We previously generated the mAb, ROCAl , which recognizes an epitope on a 26 kDa cell surface protein that is preferentially accessible in rostral intercostal nerves and rostral sympathetic ganglia (Suzue et al., 1990; Kaprielian and Patterson, 1993). Two other mAbs, ROCA2 and B2Cl l, also recognize the 26 kDa protein, but do not detect a rostrocaudal, immunohisto- chemical gradient. Amino acid sequences obtained from the affinity-purified protein suggested that it could be the rat ho- molog of CD9 (Kaprielian and Patterson, 1993). We present here the isolation of cDNA clones encoding rat CD9 from pe- ripheral nerves and ganglia, and demonstrate that the ROCA cell surface protein is, in fact, rat CD9. We also show that rat CD9 mRNA is expressed at high levels in peripheral neurons T10 T12 T12 T2 T4 T6 T8 T1 lntercostal Nerves lntercostal Nerves Figure 7. The expression of rat CD9 mRNA is not graded along the rostrocaudal axis. Northern blot analysis was performed using 2 µ.g of total RNA derived from adult rat intercostal nerves taken from various rostrocaudal levels, and 32P-labeled Full ScN # 13 (rat CD9) and MBP cDNA probes. The identical transfer was probed with each cDNA in succession after removal of previously bound radioactivity. The radio- activity representing the hybridization of each probe was quantified using a Phosphorimager and the ratio of CD9 mRNA to MBP mRNA plotted versus rostrocaudal position. 570 Kaprielian et al. • Rat CD9 Is a ROCA Antigen and Is Expressed in Neural Cells s A 9 <( z a: 6 .... en 00 C\I CD9 -.... <( z a: 3 E Po MBP NB P7 P14 P30 P60 P165 AGE Figure 8. The expression of rat CD9 mRNA is developmentally reg- ulated in postnatal sciatic nerves. A, Northern blot analysis was per- formed using 2 µg of total RNA derived from sciatic nerves taken from postnatal rats of various ages, and 32P-labeled Full ScN #13 (rat CD9), MBP, Po, and 28S cDNA probes. The identical transfer was probed with each cDNA in succession after removal of previously-bound ra- dioactivity. The radioactivity representing the hybridization of each probe was quantified using a Phosphorimager and the ratio of the various mRNAs to 28S rRNA plotted versus developmental age. B, Digoxy- genin-labeled antisense RNA probes derived from the Full ScN #13 cDNA were hybridized to cryosections of developing postnatal rat sci- atic nerves. Discussion The number of Schwann cells expressing rat CD9 signifi- cantly increases by postnatal day 7, and then decreases to adult levels by postnatal day 60. A 9 <( z a: 6 .... en 00 C\I CD9 -.... <( z a: 3 E Po MBP NB P7 P14 P30 P60 P165 AGE and Schwann cells, as well as in certain parts of the central nervous system. In addition, we demonstrate that while CD9 mRNA is not positionally regulated in intercostal nerves, it is coordinately regulated with myelin genes during postnatal ScN development. The ROCA cell surface 26 kDa protein is rat CD9 The ROCA cell surface 26 kDa protein is rat CD9 Using the PCR, as well as direct cDNA library screening, we obtained cDNA clones representing the probable entire coding region of rat CD9 from SCG and ScN. The high degree ofamino acid identity between the deduced protein sequence of the rat form and the mouse (95%), human (93%), monkey (92%), and bovine (83%) forms strongly supports the identification of these cDNAs as rat CD9. Our CD9 mAbs bind the cell surface and fail to cross-react across species. The larger, putative extracel- lular loop that contains the majority of interspecies amino acid differences (Martin-Alonso et al., 1992; Reynolds et al., 1992; Wright et al., 1993), also contains the epitopes recognized by ROCA2 and B2Cl 1, as shown by positive binding to immu- noblots of a fusion protein containing residues 111-190 of the rat protein. Interestingly, ROCAl does not bind this fusion protein, even though it binds the full-length protein very well (Fig. 3A). This result, coupled with the finding that ROCAl can bind to the complete fusion protein on immunoblots and to intact primary cells, suggests that the position-dependent ROCA 1 epitope resides on the smaller extracellular loop (residues 36- 55) ofCD9. Figure 8. The expression of rat CD9 mRNA is developmentally reg- ulated in postnatal sciatic nerves. A, Northern blot analysis was per- formed using 2 µg of total RNA derived from sciatic nerves taken from postnatal rats of various ages, and 32P-labeled Full ScN #13 (rat CD9), MBP, Po, and 28S cDNA probes. The identical transfer was probed with each cDNA in succession after removal of previously-bound ra- dioactivity. The radioactivity representing the hybridization of each probe was quantified using a Phosphorimager and the ratio of the various mRNAs to 28S rRNA plotted versus developmental age. B, Digoxy- genin-labeled antisense RNA probes derived from the Full ScN #13 cDNA were hybridized to cryosections of developing postnatal rat sci- atic nerves. The number of Schwann cells expressing rat CD9 signifi- cantly increases by postnatal day 7, and then decreases to adult levels by postnatal day 60. Figure 8. The expression of rat CD9 mRNA is developmentally reg- ulated in postnatal sciatic nerves. A, Northern blot analysis was per- formed using 2 µg of total RNA derived from sciatic nerves taken from postnatal rats of various ages, and 32P-labeled Full ScN #13 (rat CD9), MBP, Po, and 28S cDNA probes. Expression of rat CD9 mRNA -.-- "'lt. - ~ ,..... -- .,,.. Figure 8. continued. P30 ~ .. ---· -~ - ·-- - - ... P60 P165 - - - > - -- --' . ..... -- -"" - .. . NB .... .. NB -.. - , P60 ~P14 .. ... -.-- "'lt. - ~ ,..... -- .,,.. Figure 8. continued. Figure 8. continued. is maintained into adulthood. We have previously shown that satellite cells in sections of adult SCG are DRG are labeled by ROCAl and 2 (Suzue et al., 1990; Kaprielian and Patterson, 1993; Tole and Patterson, 1993), as well as B2Cl l (unpublished observations). We cannot, however, unambiguously detect label in the cytoplasm of ganglionic satellite cells in our in situ anal- ysis. This may reflect the presence oflower levels ofCD9 mRNA in satellite cells than in neurons. Taken together, the northern blot and in situ hybridization results confirm and extend pre- vious immunoblot and immunohistochemical analyses dem- onstrating that rat peripheral neurons, Schwann cells, and chro- maffin cells all express CD9 (Kaprielian and Patterson, 1993; Tole and Patterson, 1993). To our knowledge, this is the first study utilizing in situ hybridization to localize CD9 mRNA in any tissue. is maintained into adulthood. We have previously shown that satellite cells in sections of adult SCG are DRG are labeled by ROCAl and 2 (Suzue et al., 1990; Kaprielian and Patterson, 1993; Tole and Patterson, 1993), as well as B2Cl l (unpublished observations). We cannot, however, unambiguously detect label in the cytoplasm of ganglionic satellite cells in our in situ anal- ysis. This may reflect the presence oflower levels ofCD9 mRNA in satellite cells than in neurons. Taken together, the northern blot and in situ hybridization results confirm and extend pre- vious immunoblot and immunohistochemical analyses dem- onstrating that rat peripheral neurons, Schwann cells, and chro- maffin cells all express CD9 (Kaprielian and Patterson, 1993; Tole and Patterson, 1993). To our knowledge, this is the first Positional and developmenta:l expression. Northern blot anal- ysis shows that equivalent levels ofCD9 mRNA are present in rostral and caudal intercostal nerves. This result is consistent with previous immunoblot analyses demonstrating that both ROCAI and ROCA2, as well as B2Cl I (Z. Kaprielian, unpub- lished observations), detect uniform levels of the 26 kDa, rat CD9 protein in rostral and caudal intercostal nerves (Kaprielian and Patterson, 1993). Expression of rat CD9 mRNA Tissue and cell type specific expression. We detect a major 1.3 kb band of rat CD9 mRNA in many, but not all adult tissues. Of the tissues examined, only liver and skeletal muscle contain no detectable rat CD9 mRNA. The significant levels of CD9 mRNA present in certain neural tissues, including adult rat peripheral nerves and ganglia, spinal cord, and cerebellum, are consistent with the isolation ofrat CD9 cDNAs from adult SCG and ScN DNA templates. Other studies have identified a similar size mRNA band in Northern blots using CD9 cDNA probes (Boucheix et al., 1991; Martin-Alonso et al., 1992). However, an extensive analysis ofCD9 mRNA tissue distribution has not previously been described. The rat CD9 probe also hybridized weakly to a 3.2 kb mRNA band. While the nature of this species is not known, this band is present in a wide variety of tissues and cell lines (data not shown). We performed additional PCR reactions to search for a form of CD9 containing the amino acid string IDEVF from amino acid position 7 to 11. Sequence analysis of 23 distinct clones identified either the same form of rat CD9 that we report here (N-terminus: MPVKGGSKCIK) or nonsense sequences (data not shown). Despite these results, however, we cannot rule out the existence of another, perhaps alternatively spliced, form of rat CD9. Immunoprecipitation ofa 26 kDa protein by ROCA I, ROCA2, and B2Cl I from detergent extracts of CHO cells stably trans- fected with a rat CD9 cDNA confirms the identity of the ROCA cell surface antigen as CD9. Interestingly, ROCA!, unfike ROCA2 and B2C 11, does not label the surfaces of the trans- fected cells. ROCA! does not bind the surface of a variety of other cell types that are strongly labeled with ROCA2 and B2Cl 1 (Hadjiargyrou and Patterson, 1994, and unpublished observa- tions), but does bind to the surfaces of primary Schwann cells (Tole, unpublished observations). Together with our previous In situ hybridization shows that peripheral neurons, glia and chromaffin cells express rat CD9 mRNA. In particular, neurons in SCG and DRG, Schwann cells in ScNs, and chromaffin cells in the adrenal medulla are all strongly labeled by the CD9 an- tisense probe. Neuronal labeling is clearly detectable in early postnatal sympathetic and sensory ganglia (data not shown) and The Journal of Neoroscience, January 1995, 15(1) 571 B B NB .... .. -.. - , ~P14 .. ... Expression of rat CD9 mRNA These findings all support the notion that ROCA I binding reflects the presence of a positionally graded epitope, rather than a rostrocaudal gradient ofCD9 protein. Of course, the formal possibility remains, that ROCAI recognizes 572 Kaprielian et al. • Rat CD9 Is a ROCA Antigen and Is Expressed in Neural Cells an as yet unidentified third antigen (in addition to CD9 and peripherin) whose distribution may directly account for the im- munohistochemical gradient. kemia antigen p 24). Inhibition of aggregation by ALB6Fab. FEBS Lett 161:289-295. Lett 161:289-295. Boucheix C, Benoit P, Frachet P, Billard M, Worthington RE, Gagnon Boucheix C, Benoit P, Frachet P, Billard M, Worthington RE, Gagnon J, Uzan G (1991) Molecular cloning of the CD9 antigen. J Biol Chem 266:117-122. Northern blot and in situ hybridization analyses shows that the expression ofCD9 mRNA is significantly upregulated during early postnatal SCN development, and downregulated in the adult. This time course suggests that CD9 and peripheral myelin gene expression (Wiggins et al., 197 5; Lees and Brostoff, 1984; Lemke and Axel, 1985; Stahl et al., 1990) are coordinately reg- ulated during postnatal development. Together with the ex- pression of the CD9 protein on neuronal as well as Schwann cell surfaces, these developmental data further suggest a role for CD9 in mediating Schwann cell-axon interactions during my- elination. Moreover, the results of antibody perturbation ex- periments strongly suggest a role for CD9 in neural cell adhesion (Hadjiargyrou and Patterson, 1994). Cho KO, Wall JB, Pugh PC, Ito M, Mueller SA, Kennedy MB (1991) The a subunit of Type II Ca2+ /Calmodulin-dependent protein kinase is highly conserved in Drosophila. Neuron 7:439-450. Choi YC (1985) Structural organization of ribosomal RNAs from Novikoff hepatoma II: characterization of possible binding sites of SS rRNA and 5.8S rRNa to 28S rRNA. J Biol Chem 260:12773- 12779. Chomczynski P, Sacchi N (1987) Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal Biochem 162:156-159. Classon BJ, Williams AF, Willis AC, Seed B, Stamenkovic I ( 1989) The primary structure of the human leukocyte antigen CD3 7, a species homologue of the rat MRC OX-44 antigen. J Exp Med 169: 1497- 1502. Davern KM, Wright MD, Herrmann YR, Mitchell GF (1991) Further characterization of the Schistosomajaponicum protein Sj23 that is a target of an immunodiagnostic monoclonal antibody. Mo! Biochem Parasitol 48:67-76. References Akeson R, Warren SL (1984) Detection ofa cell surface antigen found on rat peripheral nervous system neurons and multiple glia: astro- cytes, oligodendrocytes, and Schwann cells. J Neurosci Res 12:41- 57. Griffith L, Slupsky J, Seehafer J, Boshkov L, Shaw ARE (1991) Platelet activation by immobilized monoclonal antibody: evidence for a CD9 proximal signal. Blood 78: 1753-1759. Amiot M (1990) Identification and analysis ofcDNA clones encoding CD53. A pan-leucocyte antigen related to membrane transport pro- teins. J Immunol 145:4322-4325. Hadjiargyrou M, Patterson PH (1995) An anti-CD9 monoclonal an- tibody promotes adhesion and induces proliferation ofSchwann cells in vitro. J Neurosci 15:574-583. Andria ML, Hsieh C, Oren R, Francke U, Levy S (1991) Genomic organization and chromosomal localization of the TAPA-1 gene. J Immunol 147:1030-1036. Harland RM (1991) In situ hybridization: an improved whole mount method forXenopus embryos. In: Methods in cell biology (Kay BK, Peng HJ, eds), pp 675-685. New York: Academic. Angelisova P, Vlcek C, Stefanova I, Lipoldova M, Horejsi V (1990) The leukocyte surface antigen CD53 is a protein structurally similar to the CD37 and MRC OX-44 antigens. lmmunogenetics 32:281- 285. Higashara M, Takahata K, Yatomi Y, Nakahara K, Kurokawa K (1990) Purification and partial characterization of CD9 antigen of human platelets. FEBS Lett 264:270-274. Anton ES, Hadjiargyrou M, Patterson PH, Matthew WD (1995) CD9 plays a role in Schwann cell migration in vitro. J Neurosci 15:584- 595. p Horejsi V, Vlcek C (1991) Novel structurally distinct family of leu- cocyte surface glycoproteins including CD9, CD37, CD53 and CD63. FEBS Lett 288:1-4. Bellacosa A, Lazo PA, Bear SA, Tsichlis PN ( 1991) The rat leukocyte antigen MRC OX-44 is a member of a new family of cell surface proteins which appear to be involved in growth regulation. Mo! Cell Biol 11:2864-2872. Hotta H, Ross AH, Huebner K, Isobe M, Wendenborn S, Chao MY, Ricciardi RP, Tsujimoto Y, Croce CM, Koprowski H (1988) Mo- lecular cloning and characterization of an antigen associated with early stages of melanoma tumor progression. Cancer Res 48:2955-2962. Bergoffen J, Scherer SS, Wang S, Oronzi Scott M, Bone LJ, Paul DL, Chen K, Lensch MW, Chance PF, Fischbeck KH (1993) Connexin mutations in X-linked Charcot-Marie-Tooth disease. Science 262: 2039-2042. Hynes RO, Lander AD (1992) Contact and adhesive specificities in the associations, migrations, and targeting of cells and axons. Cell 68: 303-322. Expression of rat CD9 mRNA Interestingly, the developmental expression ofCD9 also par- allels that of another peripheral myelin gene PMP-22 (Welcher et al., 1991; Snipes et al., 1992). PMP-22 is the target of mu- tations that result in the trembler phenotype in mice and in Charcot-Marie-Tooth disease ( CMT) Type 1 a in humans (Snipes et al., 1993). Although PMP-22 bears no sequence homology with members of the CD9 tetraspan family, this protein has also been implicated in growth regulation and possesses four putative membrane-spanning domains (Lemke, 1993; Snipes et al., 1993; Suter et al., 1993). In this regard, it is interesting to note that another four transmembrane spanning domain mol- ecule, connexin 32 (Cx32), has recently been shown, like MBP, PO, and PMP-22, to be downregulated in transected sciatic nerves (S. Scherer, personal communication). Moreover, patients with X-linked CMT possess mutations in the gene encoding Cx32 (Bergoffen et al., 1993). Thus, it seems possible that proteins containing the four transmembrane domain structural motif may play important roles in peripheral nerve myelination. Emi N, Kitaori K, Seto M, Ueda R, Saito H, Takahashi T (1993) Isolation ofa novel cDNA clone showing marked similarity to ME491/ CD63 superfamily. Immunogenetics 37: 193-198. Feinberg AP, Vogelstein B (1983) A technique for radiolabeling DNA restriction endonuclease fragments to high specific activity. Anal Bio- chem 132:6-13. Feigner PL, Gader TR, Holm M, Roman R, Chan HW, Wenz M, Northrop JP, Ringold GM, Danielsen M ( 1987) Lipofection: a high- ly efficient lipid-mediated DNA-transfection procedure. Proc Natl Acad Sci USA 84:7413-7417. Forsyth KD (1991) Anti"CD9 antibodies augment neutrophil adher- ence to endothelium. Immunology 72:292-296. Gaugitsch HW, Hofer E, Huber NE, Schnabl E, Baumruker T (1991) A new superfamily of lymphoid and melanoma cell proteins with extensive homology to Schistosoma mansoni antigen Sm23. Eur J Immunol 21:377-383. Gil ML, Vita N, Lebel-Binay S, Miloux B, Chalon P, Kaghad M, Mar- chiol-Fournigault C, Conjeaud H, Caput D, Ferrara P, Fradelizi D, Quillet-Mary A (1992) A member of the tetraspans transmembrane protein superfamily is recognized by a monoclonal antibody raised against an HLA class I-deficient, lymphokine-activated killer-suscep- tible, B lymphocyte line. Cloning and preliminary functional studies. J Immunol 148:2826-2833. References Ikeyama S, Koyama M, Yamaoko M, Sasada R, Miyake M (1993) Suppression of cell motility and metastasis by transfection with hu- man motility-related protein (MRP-1/CD9) DNA. J Exp Med 177: 1231-1237. Birren SJ, Lo LC, Anderson DJ (1993) Sympathetic neurons undergo a developmental switch in trophic dependence. Development 119: 597-610. Imai T, Fukudome K, Takagi S, Nagira M, Furuse M, Fukuhara N, Nishimura M, Hinuma Y, Yoshie 0 (1992) C33 antigen recognized by monoclonal antibodies inhibitory to human T cell leukemia virus Boucheix C, Soria C, Mirshahi M, Soria J, Perrot JY, Fournier N, Billard M, Rosenfeld C (1983) Characteristics of platelet aggregation in- duced by the monoclonal antibody ALB6 (acute lymphoblastic leu- The Journal of Neuroscience, January 1995, 15(1) 573 type I-induced syncytium formation is a member ofa new family of transmembrane proteins including CD9, CD37, CD53 and CD63. J Immunol 149:2879-2886. type I-induced syncytium formation is a member ofa new family of transmembrane proteins including CD9, CD37, CD53 and CD63. J Immunol 149:2879-2886. Reynolds SR, Shoemaker CB, Ham DA (1992) T and B cell epitope mapping of SM23, an integral membrane protein of Schistosoma mansoni. J Immunol 149:3995--4001. Jennings LK, Fox CF, Kouns WC, McKay CP, Ballou LR, Schultz HE (1990) The activation of human platelets mediated by anti-human platelet p24/CD9 monoclonal antibodies. J Biol Chem 265:3815- 3822. Ross AH, Pleasure D, Sonnenfeld K, Atkinson B, Kreider B, Jackson DM, Taff I, Scarpini E, Lisak RP, Koprowski H (1986) Expression of melanoma-associated antigens by normal and neurofibroma schwann cells. Cancer Res 46:5887-5892. Kallin B, de Martin R, Etzold T, Sorrentino V, Philipson L (1991) Cloning of a growth arrest-specific and transforming growth factor b- regulated gene, TI I, from a epithelial cell line. Mo! Cell Biol 10: 5338-5345. Rubinstein E, Benoit P, Billard M, Plaisance S, Prenant M, Uzan G, Boucheix C (l 993a) Organization of the human CD9 gene. Gen- omics 16:132-138. Rubinstein E, Billard M, Plaisance S, Prenant M, Boucheix C (l 993b) Molecular cloning of the mouse equivalent of CD9 antigen. Thromb Res 71:377-383. Kaprielian Z, Patterson PH (1993) Surface and cytoskeletal markers of rostrocaudal position in the mammalian nervous system. J Neu- rosci 13:2495-2508. Sanger F, Nicklen S, Coulson AR ( 1977) DNA sequencing with chain- terminating inhibitors. Proc Natl Acad Sci USA 74:5463-5467. References Suzue T, Kaprielian Z, Patterson PH (1990) A monoclonal antibody that defines rostrocaudal gradients in the mammalian nervous system. Neuron 5:421--431. Martin-Alonso JM, Hernando N, Ghosh S, Coca-Prados M (1992) Molecular cloning of the bovine CD9 antigen from ocular ciliary epithelial cells. J Biochem 112:63-67. Szala S, Kasai Y, Steplewski Z, Rodeck U, Koprowski H, Linnebach AJ (1990) Molecular cloning of cDNA for the human tumor-as- sociated antigen C0-029 and identification ofrelated transmembrane antigens. Proc Natl Acad Sci USA 87:6833-6837. Masselis-Smith A, Jensen GS, Seehater JG, Slupsky JR, Shaw ARE ( 1990) Anti-CD9 monoclonal antibodies induce homotypic adhesion ofpre-B cell lines by a novel mechanism. J Immunol 144: 1607-1613. g Tole S, Patterson PH (1993) Distribution of CD9 in the developing and mature rat nervous system. Dev Dynamics 197:94-106. Metzelaar MJ, Wijingaard PLJ, Peters LJ, Sixma JJ, Nieuwenhuis HK, Clevers HC (1991) CD63 antigen. A novel lysosomal membrane glycoprotein, cloned by a screening procedure for intracellular anti- gens in eukaryotic cells. J Biol Chem 266:3239-3245. Tomlinson MG, Williams AF, Wright MD (1993) Epitope mapping of anti-rat CD53 monoclonal antibodies. Implications for the mem- brane orientation of the transmembrane 4 superfamily. Eur J Im- munol 23: 136-140. Von dem Borne AEGK, Modderman PW, Admiraal LG, Nieuwenhuis HK (1989) Platelet antibodies, the overall results. In: Leukocyte typing, white cell differentiation antigens, 4:951-966. Mitamura T, Iwamoto R, Umata T, Torno Y, Urabe I, Tsuneoka M, Mekada E (1992) The 27-kD diphtheria toxin receptor-associated protein (DRAP27) from Vero cells is the monkey homologue of hu- man CD9 antigen: expression of DRAP27 elevates the number of diphtheria toxin receptors on toxin-sensitive cells. J Cell Biol 118: 1389-1399. Welcher AA, Suter U, Leon MD, Snipes GJ, Shooter EM (1991) A myelin protein is encoded by the homologue of a growth arrest-specific gene. Proc Natl Acad Sci USA 88:7195-7199. g Wiggins RC, Benjamins JA, Morell P (1975) Appearance of myelin proteins in rat sciatic nerve during development. Brain Res 89:99- 106. Miyake M, Koyama M, Seno M, lkeyama S (1991) Identification of the motility-related protein (MRP-1), recognized by monoclonal an- tibody M31-15, which inhibits cell motility. J Exp Med 174:1347- 1354. Wright MD, Henkle KJ, Mitchell GF (1990) An immunogenic Mr 23,000 integral membrane protein of Schistosoma mansoni worms that closely resembles a human tumor-associated antigen. J Immunol 144:3195-3200. References Kemshead JT, Fritschy J, Asser U, Sutherland R, Greaves MF (1982) Monoclonal antibodies defining markers with apparent selectivity for particular haematopoietic cell types may also detect antigens on cells of neural crest origin. Hybridoma 1:109-123. Slupsky JR, Seehafer JG, Tang SC, Masselis-Smith A, Shaw ARE (1989) Evidence that monoclonal antibodies against CD9 antigen induce specific association between CD9 and the platelet glycoprotein Ilb- llla complex. J Biol Chem 264:12289-12293. Komada Y, Peiper SC, Melvin SL, Metzgar DW, Tarnowski BH, Green AA (1983) A monoclonal antibody (SJ-9A4) to p24 present on alls, neuroblastomas and platelets. I. Characterization and development of a unique radioimmunometric assay. Leuk Res 7:487--498. Smith DB, Johnson KS (1988) Single-step purification of polypeptides expressed in Escherichia coli as fusions with glutathione S-transferase. Gene 67:31--40. Snipes GJ, Suter U, Welcher AA, Shooter EM (1992) Characterization of a novel peripheral nervous system myelin protein (PMP-22/SR 13). J Cell Biol 117:225-238. Lanza F, WolfD, Fox CF, Kieffer N, Seyer JM, Fried VA, Coughlin SR, Phillips DR, Jennings LK ( 1991) CDNA cloning and expression of platelet p24/CD9. J Biol Chem 266:10638-10645. Snipes GJ, Suter U, Shooter EM (1993) The genetics of myelin. Curr Opin Neurobiol 3:694-702. Lees MB, Brostoff SW ( 1984) Proteins of myelin. In: Myelin (Morell P, ed), pp 197-224. New York: Plenum. pp Lemke G (1993) The molecular genetics of myelination: an update. Glia 7:263-271. Stahl N, Harry J, Popko B (1990) Quantitative analysis of myelin protein gene expression during development in the rat sciatic nerve. Mot Brain Res 8:209-212. Lemke G, Axel R (1985) Isolation and sequence ofa cDNA encoding the major structural protein of peripheral myelin. Cell 40: 501-508. Steinman L ( 1993) Connections between the immune and the nervous system. Proc Natl Acad Sci USA 90:7912-7914. Levy S, Nguyen VQ, Andria ML, Takahashi S ( 1991) Structure and membrane topology ofTAPA-1. J Biol Chem 266:14597-14602. Suter U, Welcher AA, Snipes GJ (1993) Progress in the molecular understanding of hereditary peripheral neuropathies reveals new in- sights into the biology of the peripheral nervous system. Trends Neu- rosci 16:50-56. Lowry OH, Rosebrough NJ, Farr AL, Randall RJ (1951) Protein measurement with the Folin phenol reagent. J Biol Chem 193:265- 275. Marken JS, Schieven GL, Hellstrom I, Hellstrom KE, Aruffo A (1992) Cloning and expression of the tumor-associated antigen L6. Proc Natl Acad Sci USA 89:3503-3507. References Nishikata H, Oliver C, Mergenhagen SE, Siraganian RP (1992) The rat mast cell antigen AD I (homologue to human CD63 or melanoma antigen ME49 l) is expressed in other cells in culture. J Immunol 149: 862-870. Wright MD, Rochelle JM, Tomlinson MG, Seldin MF, Williams AF (1993) Gene structure, chromosomal localization, and protein se- quence of mouse CD53 (Cd53): evidence that the transmembrane 4 superfamily arose by gene duplication. Int Immunol 5:209-216. Oren R, Takahashi S, Doss C, Levy R, Levy S (1990) TAPA-1, the target of an antiproliferative antibody, defines a new family oftrans- membrane proteins. Mo! Cell Biol 10:4007--4015.
https://openalex.org/W1985032382	http://fulir.irb.hr/2047/1/journal.pone.0093781.pdf	English	null	Astral Microtubule Pivoting Promotes Their Search for Cortical Anchor Sites during Mitosis in Budding Yeast	PloS one	2,014	cc-by	8,728	Introduction the linker protein Kar9 [10], which in turn binds to a type-V myosin Myo2 [11,12]. Myosin walks along cortical actin filaments, thereby moving the microtubule plus end towards the neck. As the plus end moves along the cortex, the whole astral microtubule pivots around the SPB, ending up oriented towards the neck [13]. This reorientation or angular movement of microtubules requires actin, Myo2 and Kar9 [14]. Kar9 is preferentially localized at the daughter-bound SPB and its astral microtubules, thus only the microtubules extending from the daughter-bound SPB become oriented towards the neck [14]. Once at the neck, the microtubule plus end is captured by Bud6, a protein that binds actin and formin [15], which stabilizes the position of the microtubule and of the spindle near the neck. Proper segregation of genetic material between the two daughter cells requires the mitotic spindle to become properly positioned with respect to the cell division plane. In addition to controlling chromosome segregation, the position of the spindle determines the fate of the daughter cells during development of metazoan organisms. Forces that orient and position the spindle are typically generated at the sites where astral microtubules, which grow from the centrosomes, interact with the cell cortex. A large body of knowledge about the general principles governing spindle positioning comes from studies in the budding yeast Saccharomyces cerevisiae (reviewed in [1,2]). As the name suggests, budding yeast divides by budding, which means that a daughter cell grows as a bud on the mother cell. The two cells are connected by a small opening called neck, through which one end of the spindle has to be moved during mitosis. In the dynein pathway, the plus end of a growing microtubule accumulates dynein in a Bik1/CLIP-170- and Pac1/LIS1- dependent manner. Dynein reaches the plus end by being transported along the microtubule by the kinesin Kip2 or by directly binding from the cytoplasm to the plus end [16,17,18,19], but dynein may also diffuse along the microtubule, as in fission yeast [20,21]. When the plus end brings dynein close to the cortical anchor protein Num1, dynein binds to the anchor in a process termed off-loading [19,22], and may detach from the anchor in response to load forces, as shown in fission yeast [23]. Upon binding to the anchor, dynein starts to walk towards the minus end of the microtubule, thereby pulling on the microtubule and moving the spindle. Abstract * E-mail: tolic@mpi-cbg.de Competing Interests: The authors have declared that no competing interests exist. * E-mail: tolic@mpi-cbg.de Competing Interests: The authors have declared that no competing interests exist. * E-mail: tolic@mpi-cbg.de ¤ Current address: Institute of Molecular Biology, Mainz, Germany ¤ Current address: Institute of Molecular Biology, Mainz, Germany ¤ Current address: Institute of Molecular Biology, Mainz, Germany Stephan Baumga¨rtner1¤, Iva M. Tolic´1,2* of Molecular Cell Biology and Genetics, Dresden, Germany, 2 Division of Molecular Biology, Rud–er Bosˇkovic´ Institute, Zagreb, Croatia Max Planck Institute of Molecular Cell Biology and Genetics, Dresden, Germany, 2 Division of Molecular Biology, Rud–er Bosˇkovic´ Inst Astral Microtubule Pivoting Promotes Their Search for Cortical Anchor Sites during Mitosis in Budding Yeast Stephan Baumga¨rtner1¤, Iva M. Tolic´1,2* Abstract Positioning of the mitotic spindle is crucial for proper cell division. In the budding yeast Saccharomyces cerevisiae, two mechanisms contribute to spindle positioning. In the Kar9 pathway, astral microtubules emanating from the daughter- bound spindle pole body interact via the linker protein Kar9 with the myosin Myo2, which moves the microtubule along the actin cables towards the neck. In the dynein pathway, astral microtubules off-load dynein onto the cortical anchor protein Num1, which is followed by dynein pulling on the spindle. Yet, the mechanism by which microtubules target cortical anchor sites is unknown. Here we quantify the pivoting motion of astral microtubules around the spindle pole bodies, which occurs during spindle translocation towards the neck and through the neck. We show that this pivoting is largely driven by the Kar9 pathway. The microtubules emanating from the daughter-bound spindle pole body pivot faster than those at the mother-bound spindle pole body. The Kar9 pathway reduces the time needed for an astral microtubule inside the daughter cell to start pulling on the spindle. Thus, we propose a new role for microtubule pivoting: By pivoting around the spindle pole body, microtubules explore the space laterally, which helps them search for cortical anchor sites in the context of spindle positioning in budding yeast. Citation: Baumga¨rtner S, Tolic´ IM (2014) Astral Microtubule Pivoting Promotes Their Search for Cortical Anchor Sites during Mitosis in Budding Yeast. PLoS ONE 9(4): e93781. doi:10.1371/journal.pone.0093781 Editor: Attila Csika´sz-Nagy, Fondazione Edmund Mach, Research and Innovation Centre, Italy Received December 19, 2013; Accepted March 10, 2014; Published April 10, 2014 Received December 19, 2013; Accepted March 10, 2014; Published April 10, 2014 Received December 19, 2013; Accepted March 10, 2014; Published April 10, 2014 mga¨rtner, Tolic´. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits tion, and reproduction in any medium, provided the original author and source are credited. Copyright: 2014 Baumga¨rtner, Tolic´. This is an open-access article distributed under the terms of the Creative Commons Attribu unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: This work was funded by the Max Planck Society, http://www.mpg.de/en. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. April 2014 \| Volume 9 \| Issue 4 \| e93781 Introduction However, for dynein to exert force to translocate the spindle, the microtubule, which carries dynein, Spindle positioning and movement is driven through interac- tions between the plus end of astral microtubules extending from the spindle pole body (SPB, a centrosome equivalent) and the cell cortex. Two pathways have been described: (i) the Kar9 pathway is responsible for positioning the spindle close to the neck prior to anaphase [3], and (ii) the dynein pathway generates forces to pull the spindle through the neck into the daughter cell during anaphase [4,5,6,7] and contributes to spindle elongation [8,9]. In the Kar9 pathway, astral microtubule plus ends connect to cortical actin via the plus-end binding protein Bim1/EB-1, which binds to April 2014 \| Volume 9 \| Issue 4 \| e93781 April 2014 \| Volume 9 \| Issue 4 \| e93781 PLOS ONE \| www.plosone.org 1 Astral Microtubule Pivoting as a Search Strategy Figure 1. Astral microtubules pivot around the spindle pole body. (A) Image of the spindle and astral microtubules (green) in a budding yeast cell expressing tubulin-GFP (bright-field image of the cell is overlaid in magenta); scale bar is 2 mm. In the corresponding scheme below, the gray rod is the spindle and the dashed white lines represent different positions of astral microtubules. (B) Time-lapse images of the region marked by the white rectangle in panel A; scale bar is 2 mm. The schemes below show the position of the spindle and astral microtubules in each image; time is given in seconds. Note that the astral microtubules perform angular motion around the SPB, see also Movie S1. (C) Angle of the astral microtubule extending from the SPB in the daughter cell (black) and of the astral microtubule extending from the SPB in the mother cell (gray) as a function of time, for the cell shown in panels A and B. The angles were calculated with respect to the spindle. doi:10 1371/journal pone 0093781 g001 must find anchor proteins at the cortex to off-load dynein onto the anchor. The mechanism by which microtubules target cortical anchor sites has remained unknown so far. We have recently shown that during mitosis in fission yeast microtubules pivot around the SPB, which accelerates their search for kinetochores [24]. The pivoting motion allows microtubules to explore space laterally, as they search for targets such as kinetochores. Here we quantify the pivoting of astral microtubules in budding yeast. Astral Microtubules Pivot around the Spindle Pole Live-cell imaging of tubulin-labeled cells revealed that astral microtubules change their orientation with respect to the cell and to the spindle, where one end of the microtubule is attached to the SPB and the other end moves in the cytoplasm or along the cell cortex, thereby sweeping through the cell (Fig. 1, A and B; Movie S1) [5,13]. To quantify the process of microtubule pivoting, we investigated time series of the angle between the astral microtubule and the spindle in wild-type cells (Fig. 1, B and C). To distinguish whether microtubule pivoting is directed or random, we calculated the mean squared angular displacement (MSAD) [26], and found that it scales roughly linearly with time (Fig. 2, A and B; see also [24]). Such a linear relationship is a feature of random movement, and from the slope we calculated the corresponding effective angular diffusion coefficient of astral microtubules (Table 1). doi:10.1371/journal.pone.0093781.g001 length [24], in the analysis of the MSAD we used only those microtubules that were 0.75–1.25 mm long. length [24], in the analysis of the MSAD we used only those microtubules that were 0.75–1.25 mm long. We first compare the microtubules originating from the daughter-bound SPB in the 3 strains. The angular diffusion coefficient was 4 times lower in kar9D cells than in wild-type cells (Fig. 2A; Table 1), implying that the Kar9 pathway increases the speed of pivoting of the microtubules extending from the daughter-bound SPB. In num1D cells, the angular diffusion coefficient was 30% lower than in wild-type cells, suggesting that the dynein pathway also affects microtubule pivoting, though to a smaller extent. At the mother-bound SPB, the diffusion coefficient in kar9D and num1D cells was roughly 2 times lower than in wild-type cells (Fig. 2B; Table 1), suggesting that both the Kar9 and the dynein pathway affect microtubule pivoting in the mother cell. The effect of the Kar9 pathway at the mother-bound SPB is most likely due to the residual amount of Kar9 at this SPB [14]. Introduction We propose that, similarly to the search for kinetochores, microtubule pivoting helps them to search for cortical anchor sites in order to move the spindle into the bud. Results We set out to study microtubule and spindle movements by imaging budding yeast cells expressing a-tubulin-GFP with high time resolution (0.4–0.6 s). The presence of a small number of microtubules in these cells enabled us to observe the dynamics and movement of each astral microtubule during its lifetime. In order to test the role of the Kar9 pathway on these movements, we used a strain lacking Kar9, in which microtubules do not interact with actin and myosin. Likewise, to test the role of the dynein pathway, we used a strain lacking the anchor protein Num1, in which microtubules do not off-load dynein to the cortex because the cortical anchor proteins are missing. Cells lacking Kar9 (kar9D) did not show a defect in spindle translocation into the bud, whereas the cells lacking the dynein pathway (num1D) did (Table S1), suggesting that the spindle translocation into the bud is mainly driven by dynein, consistent with previous work [4]. Astral microtubules in kar9D cells lived for significantly longer time and grew to a larger length than those in wild type, whereas the effect of deleting Num1 on microtubule dynamics was smaller (Table S2). These data suggest that Kar9 has a major effect on microtubule dynamics, consistent with a previous study [25]. Figure 1. Astral microtubules pivot around the spindle pole body. (A) Image of the spindle and astral microtubules (green) in a budding yeast cell expressing tubulin-GFP (bright-field image of the cell is overlaid in magenta); scale bar is 2 mm. In the corresponding scheme below, the gray rod is the spindle and the dashed white lines represent different positions of astral microtubules. (B) Time-lapse images of the region marked by the white rectangle in panel A; scale bar is 2 mm. The schemes below show the position of the spindle and astral microtubules in each image; time is given in seconds. Note that the astral microtubules perform angular motion around the SPB, see also Movie S1. (C) Angle of the astral microtubule extending from the SPB in the daughter cell (black) and of the astral microtubule extending from the SPB in the mother cell (gray) as a function of time, for the cell shown in panels A and B. The angles were calculated with respect to the spindle. Kar9 Pathway Increases the Speed of Microtubule Pivoting In A and B, microtubules of length 0.75–1.25 mm were included, whereas in C and D, microtubules of length 1.25–1.75 mm were included in the analysis. Wild-type cells are shown in black, kar9D in magenta, and num1D in green. One-minute-long time series of the angle were used, error bars represent s.e.m., n denotes the number of microtubules. The number of cells was 9 for wild type, 10 for kar9D and 10 for num1D. Linear fits with weights 1/s.e.m., MSAD = 2DMTDt+offset are also shown (lines) and the resulting angular diffusion coefficients DMT are given in Table 1. Fitting a parabola of a form MSAD = v2(Dt)2+2DMTDt+offset to the data resulted in negative values for v2 for all three strains in A and B and thus did not yield a meaningful result. doi:10.1371/journal.pone.0093781.g002 We conclude that the Kar9 pathway has a major effect on the speed of microtubule pivoting. The dynein pathway also influences the speed of pivoting, though to a smaller extent. We conclude that the Kar9 pathway has a major effect on the speed of microtubule pivoting. The dynein pathway also influences the speed of pivoting, though to a smaller extent. results are consistent with the finding that Kar9 is preferentially loaded to the daughter-bound SPB and the microtubules emanating from it [14]. Similar results were obtained for longer microtubules (1.25–1.75 mm long; Fig. 2, C and D; Table 1), suggesting that the observed differences in the angular diffusion coefficient are not due to a lack of interaction between the microtubules and the cell cortex because short microtubules were chosen for analysis. Kar9 Pathway Increases the Speed of Microtubule Pivoting To test the role of the Kar9 pathway and the dynein pathway in microtubule pivoting, we analyzed microtubule angles in kar9D and num1D cells (examples are shown in Movie S2 and Movie S3, respectively). As in wild type, MSAD of microtubules in kar9D and num1D cells scaled linearly with time (Fig. 2, A and B). However, the slope of this relationship, i.e., the angular diffusion coefficient of the microtubules, differed between the strains and between the microtubules extending from the two SPBs. Because the angular diffusion coefficient has been shown to depend on microtubule Finally, we compare the microtubules extending from the two SPBs. In wild-type and num1D cells, the angular diffusion coefficient of the microtubules originating from the daughter- bound SPB was higher than that of the microtubules extending from the mother-bound SPB, whereas in kar9D cells such a difference was not observed (Fig. 2, A and B; Table 1). These April 2014 \| Volume 9 \| Issue 4 \| e93781 PLOS ONE \| www.plosone.org 2 Astral Microtubule Pivoting as a Search Strategy Figure 2. Microtubules pivot faster in cells with an intact Kar9 pathway. Mean squared angular displacement (MSAD) of astral microtubules extending from the SPB destined to the daughter (A and C) and from the SPB destined to the mother cell (B and D). In A and B, microtubules of length 0.75–1.25 mm were included, whereas in C and D, microtubules of length 1.25–1.75 mm were included in the analysis. Wild-type cells are shown in black, kar9D in magenta, and num1D in green. One-minute-long time series of the angle were used, error bars represent s.e.m., n denotes the number of microtubules. The number of cells was 9 for wild type, 10 for kar9D and 10 for num1D. Linear fits with weights 1/s.e.m., MSAD = 2DMTDt+offset are also shown (lines) and the resulting angular diffusion coefficients DMT are given in Table 1. Fitting a parabola of a form MSAD = v2(Dt)2+2DMTDt+offset to the data resulted in negative values for v2 for all three strains in A and B and thus did not yield a meaningful result. doi:10.1371/journal.pone.0093781.g002 Figure 2. Microtubules pivot faster in cells with an intact Kar9 pathway. Mean squared angular displacement (MSAD) of astral microtubules extending from the SPB destined to the daughter (A and C) and from the SPB destined to the mother cell (B and D). Kar9 Pathway Promotes Frequent Spindle Movements towards the Bud the angular diffusion coefficient of the microtubule decreases with increasing microtubule length [27,28,29]. An astral microtubule can be described as a thin stiff rod with one end connected to a fixed point through a free joint, while the other end is free to move [24]. This allows the rod to perform angular movement (scheme in Fig. 1A, dashed lines). Thermally driven angular diffusion of such a rod is described by D (degrees2/s) = (3?1802 ln(L/d) kBT)/(4 p3 L3 g), where L is the length and d the diameter of the rod, kB is the Boltzmann constant, T is absolute temperature, and g is the viscosity of the medium [27,28,29]. This equation is a good approximation for L..d [27,29]. The transition of one spindle pole from the mother into the daughter cell does not occur in a smooth and continuous manner but exhibits phases of rapid movements towards the daughter cell, which are sometimes followed by similar retractions (Fig. 4, A–C; Movie S4). These events can be translation or rotation or a combination of both. We hypothesized that the pivoting of the astral microtubules around the spindle pole, which results in a sweeping movement of the microtubule inside the daughter cell, helps them find cortical anchors for dynein. To test this hypothesis, we measured the time elapsed from the entry of an astral microtubule into the daughter cell until the beginning of a pulling event, in which the spindle moves towards the daughter cell. We refer to this time as reaction time. When a microtubule enters the daughter cell and the spindle subsequently exhibits an event of rapid motion towards the daughter cell, this suggests that the microtubule has found a cortical anchor site for dynein, which then pulls on the microtubule. We determined the relationship between the length of the astral microtubules and their effective angular diffusion coefficient, which was calculated from the MSAD of the microtubules. We found that the diffusion coefficient decreases with increasing microtubule length in wild-type, kar9D and num1D cells, both at the daughter-bound and the mother-bound SPB (Fig. 3, A and B). However, only kar9D cells showed a decrease consistent with the theoretical prediction for thermally driven motion of a thin rod. The corresponding fit with the effective viscosity of the cytoplasm as a single free parameter suggests that the cytoplasm is roughly 500 times more viscous than water (Fig. 3, magenta lines). Kar9 Pathway Promotes Frequent Spindle Movements towards the Bud This value is of the same order of magnitude as the one previously measured for the cytoplasm of fission yeast cells [30]. We found that in kar9D cells astral microtubules often remained in the daughter cell for a long time, without an accompanying movement of the spindle towards the daughter (Fig. 4, D–F; Movie S5). Consistently, the mean reaction time of the spindle was longer in kar9D cells than in wild-type cells (Fig. 4G). In both strains, the reaction time was correlated neither with the distance between the daughter-bound SPB and the neck (r2 = 20.17 and 0.001 for wild type and kar9D, respectively), nor with the length of the astral microtubules (r2 = 0.33 and 20.08 for wild type and kar9D, respectively). These results argue against the possibility that the longer reaction time in kar9D cells is a consequence of a larger average distance between the daughter-bound SPB and the neck or longer microtubules than in wild type. Although we cannot exclude the possibility that kar9D cells have a longer reaction time because of other defects in this mutant, we favor the model in which the microtubules in kar9D cells show this delay because they pivot less than in wild type. We propose that the Kar9 pathway, by increasing the speed of microtubule pivoting, helps the microtu- bules to explore the space laterally. This angular movement of the microtubules improves their search for cortical anchor sites for dynein in the daughter cell, which results in pulling of the spindle from the mother to the daughter cell. The diffusion coefficient in wild-type and num1D cells did not show a decrease consistent with the theoretical prediction for thermally driven motion of a thin rod. The deviation from this prediction was more pronounced at the daughter-bound than at the mother-bound pole (Fig. 3, A and B). Based on the analysis of the effective angular diffusion coefficient as a function of microtubule length, we conclude that the pivoting of astral microtubules is to a large extent driven by active components, which belong to the Kar9 pathway. Figure 3. Microtubule pivoting is driven by active processes. The angular diffusion coefficient DMT as a function of microtubule length L, for microtubules extending from the SPB destined to the daughter cell (A) and from the SPB destined to the mother cell (B). Microtubule Pivoting in Wild-type Cells is an Active Process, Whereas that in Kar9D Cells is Consistent with Thermal Motion Microtubule pivoting may be driven by active components such as motor proteins, or it may be thermally driven. In the latter case, Table 1. Effective angular diffusion coefficients DMT of astral microtubules. Effective angular diffusion coefficient, DMT (degrees2/s) wild type kar9D num1D MTs from the daughter SPB L = 0.75–1.25 mm 27.361.2 (n = 28) 6.860.2 (n = 21) 19.862.3 (n = 30) L = 1.25–1.75 mm 28.461.6 (n = 8) 6.260.3 (n = 20) 19.861.7 (n = 22) MTs from the mother SPB L = 0.75–1.25 mm 13.561.6 (n = 13) 7.560.7 (n = 14) 6.560.6 (n = 10) L = 1.25–1.75 mm 5.261.2 (n = 2) 4.460.4 (n = 10) 7.761.0 (n = 9) DMT was calculated separately for the microtubules extending from the SPB destined to the daughter cell and from the SPB destined to the mother cell, in each of the 3 strains (wild type, kar9D, num1D). DMT was obtained by fitting the equation MSAD = 2DMTDt+offset to the mean squared angular displacement (MSAD) data shown in Figure 2, with weights 1/s.e.m. The errors are the square root of the diagonal elements of the covariance matrix from the linear fit, n denotes the number of microtubules. The number of cells was 9 for wild type, 10 for kar9D and 10 for num1D. doi:10.1371/journal.pone.0093781.t001 Table 1. Effective angular diffusion coefficients DMT of astral microtubules. DMT was calculated separately for the microtubules extending from the SPB destined to the daughter cell and from the SPB destined to the mother cell, in each of the 3 strains (wild type, kar9D, num1D). DMT was obtained by fitting the equation MSAD = 2DMTDt+offset to the mean squared angular displacement (MSAD) data shown in Figure 2, with weights 1/s.e.m. The errors are the square root of the diagonal elements of the covariance matrix from the linear fit, n denotes the number of microtubules. The number of cells was 9 for wild type, 10 for kar9D and 10 for num1D. doi:10.1371/journal.pone.0093781.t001 April 2014 \| Volume 9 \| Issue 4 \| e93781 3 PLOS ONE \| www.plosone.org 3 Astral Microtubule Pivoting as a Search Strategy The Mechanism of Microtubule Pivoting The Mechanism of Microtubule Pivoting During mitosis in budding yeast, the spindle forms inside the mother cell and has to move through a narrow neck into the daughter cell or the bud. Two types of motor proteins are involved in the generation of force on astral microtubules and thus on the spindle: myosin V, which binds to the plus end of the microtubule via the adaptor protein Kar9, and cytoplasmic dynein [6,7,10,11]. We have shown that in the absence of Kar9, microtubules pivot substantially slower than in wild type. Thus, the pivoting motion of astral microtubules is most likely driven by the movement of myosin along cortical actin filaments. As the plus end of the microtubule is being moved by myosin, while the minus end is bound to the SPB, which does not move as much as the plus end, the whole astral microtubule pivots around the SPB. Our observation that the microtubules extending from the daughter- bound SPB pivot faster than those at the mother-bound SPB is consistent with the fact that Kar9 is predominantly found at the daughter-bound SPB and the associated microtubules [14]. Figure 3. Microtubule pivoting is driven by active processes. Figure 3. Microtubule pivoting is driven by active processes. The angular diffusion coefficient DMT as a function of microtubule length L, for microtubules extending from the SPB destined to the daughter cell (A) and from the SPB destined to the mother cell (B). Data from wild-type cells are shown in black, from kar9D cells in magenta, and from num1D cells in green. The angular diffusion coefficient was calculated as DMT = MSAD/(2Dt), where Dt = 360.5 s. MSAD and the corresponding mean microtubule length L were calculated from 20- second-long time series of the microtubule angle and length, respectively. The number of data points in each bin was 5–130, with more points for shorter microtubules; error bars represent s.e.m. For kar9D cells, a single-parameter fit of the equation DMT (degrees2/s) = (3?1802 ln(L/d) kBT)/(4 p3 L3 g) to the data yielded the viscosity g = 475 cP in panel A and g = 384 cP in panel B (magenta lines). Here, d = 25 nm, kB = 1.3810223 J/K, and T = 298.15 K was used. The gray area marks the diffusion coefficient values calculated for a viscosity of 300 cP (upper bound) to 700 cP (lower bound). Kar9 Pathway Promotes Frequent Spindle Movements towards the Bud Data from wild-type cells are shown in black, from kar9D cells in magenta, and from num1D cells in green. The angular diffusion coefficient was calculated as DMT = MSAD/(2Dt), where Dt = 360.5 s. MSAD and the corresponding mean microtubule length L were calculated from 20- second-long time series of the microtubule angle and length, respectively. The number of data points in each bin was 5–130, with more points for shorter microtubules; error bars represent s.e.m. For kar9D cells, a single-parameter fit of the equation DMT (degrees2/s) = (3?1802 ln(L/d) kBT)/(4 p3 L3 g) to the data yielded the viscosity g = 475 cP in panel A and g = 384 cP in panel B (magenta lines). Here, d = 25 nm, kB = 1.3810223 J/K, and T = 298.15 K was used. The gray area marks the diffusion coefficient values calculated for a viscosity of 300 cP (upper bound) to 700 cP (lower bound). doi:10.1371/journal.pone.0093781.g003 April 2014 \| Volume 9 \| Issue 4 \| e93781 The Mechanism of Microtubule Pivoting The schemes below show the position of the spindle and astral microtubule in each image; time is given in minutes:seconds. See also Movie S4. (C) Position of the SPB destined to the bud (black) and of the tip of the astral microtubule extending from that SPB (blue) as a function of time, for the cell shown in panels A and B. Position 0 denotes the bud neck. (D)–(F) Data from a kar9D cell; the legend is equivalent to that for panels (A)–(C), respectively. Time is given in seconds. See also Movie S5. (G) Reaction time for wild-type and kar9D cells. Number of cells is 16 in wild type and 15 in kar9D; number of astral microtubules in the daughter cell is 50 in wild type and 17 in kar9D; number of events is 29 for wild type and 12 for kar9D; total observation time is 350 min for wild type and 239 min for kar9D. Error bars represent s.e.m.; p = 0.05 from a two-sided t-test comparing the reaction time in wild-type and kar9D cells. doi:10.1371/journal.pone.0093781.g004 The Mechanism of Microtubule Pivoting doi:10.1371/journal.pone.0093781.g003 Similarly to myosin, cortically anchored dynein in the daughter and the mother cell pulls on astral microtubules and thus on the April 2014 \| Volume 9 \| Issue 4 \| e93781 April 2014 \| Volume 9 \| Issue 4 \| e93781 4 PLOS ONE \| www.plosone.org Astral Microtubule Pivoting as a Search Strategy Figure 4. Events of rapid spindle movement are more frequent in cells with an intact Kar9 pathway. (A) Image of the spindle and astral microtubules (green) in a wild-type cell expressing tubulin-GFP (bright-field image of the cell is overlaid in magenta); scale bar is 2 mm. The corresponding scheme is shown below. (B) Time-lapse images of the region marked with the white rectangle in panel A; scale bar is 2 mm. The schemes below show the position of the spindle and astral microtubule in each image; time is given in minutes:seconds. See also Movie S4. (C) Position of the SPB destined to the bud (black) and of the tip of the astral microtubule extending from that SPB (blue) as a function of time, for the cell shown in panels A and B. Position 0 denotes the bud neck. (D)–(F) Data from a kar9D cell; the legend is equivalent to that for panels (A)–(C), respectively. Time is given in seconds. See also Movie S5. (G) Reaction time for wild-type and kar9D cells. Number of cells is 16 in wild type and 15 in kar9D; number of astral microtubules in the daughter cell is 50 in wild type and 17 in kar9D; number of events is 29 for wild type and 12 for kar9D; total observation time is 350 min for wild type and 239 min for kar9D. Error bars represent s.e.m.; p = 0.05 from a two-sided t-test comparing the reaction time in wild-type and kar9D cells. doi:10.1371/journal.pone.0093781.g004 Figure 4. Events of rapid spindle movement are more frequent in cells with an intact Kar9 pathway. (A Figure 4. Events of rapid spindle movement are more frequent in cells with an intact Kar9 pathway. (A) Image of the spindle and astral microtubules (green) in a wild-type cell expressing tubulin-GFP (bright-field image of the cell is overlaid in magenta); scale bar is 2 mm. The corresponding scheme is shown below. (B) Time-lapse images of the region marked with the white rectangle in panel A; scale bar is 2 mm. April 2014 \| Volume 9 \| Issue 4 \| e93781 Sample Preparation The strains used in this work are listed in Table S3. Cells were grown on agar plates with appropriate supplements and stored in a cold room at 4uC. For experiments, cells were transferred to liquid medium and stored overnight in a shaker at 25uC in order to prepare log-phase cultures. Samples for experiments were prepared by using 35 mm MatTek plastic dishes (MatTek Corporation, Ashland, MA, USA). First, a glass coverslip was glued from the inside of the dish onto the hole on the bottom by using silicon grease. The dish was turned upside down and the hole was filled with melted agar and covered with a second coverslip. After the agar solidified the second coverslip was removed and 5–8 ml of the liquid cell culture was pipetted onto the thin agar patch on the bottom of the dish. When the liquid medium had dried out the sample was closed with a new high precision 0.1760.005 mm coverslip (Menzel Glaeser, Glasbear- beitungswerk GmbH & Co. KG, Braunschweig, Germany) and sealed with silicon grease to prevent drying out. Pivoting as a Search Strategy We found that at short time-scales, microtubules pivot around the spindle pole in a random manner, thereby exploring the space through which they move. We propose that the pivoting movement of microtubules facilitates their search for cortical anchors. This search mechanism is similar to the one used by microtubules in order to find kinetochores during mitosis in fission yeast, which is also based on microtubule pivoting around the SPB [24,31]. However, while the pivoting motion of the polar microtubules inside the fission yeast nucleus is driven by thermal forces [24], the pivoting motion of astral microtubules in the budding yeast cytoplasm is driven largely by myosin motors. In both cases, irrespective of whether the pivoting is a passive or an active process, this motion allows microtubules to explore space laterally, as they search for targets such as kinetochores or cortical anchors. Microscopy Microscopy was performed on an Andor spinning disc microscope (Andor Technology, Belfast UK). The scan head was a Yokagawa CSU-X1 unit with a pinhole size of 24 mm and a pinhole spacing of 240 mm. An Olympus UplanSApo 100x/1.4 N.A. oil objective (Olympus, Tokyo, Japan) and an Andor iXon EM+ DU-897 BV cooled and back illuminated Electron Multiplying Charge Coupled Device (EMCCD) camera with a physical pixel size of 16 mm (Andor Technology plc., Belfast, UK) were used. The resulting xy-pixel size in the images was 77.4 nm. For excitation, a Sapphire 488 nm solid-state laser was used (75 mW; Coherent, Inc., Santa Clara, CA, USA). The exposure time was 40–70 ms. The microscope was equipped with a motorized Prior ProScanIII xy scanning stage (Prior Scientific Inc., Rock- land, MA, USA) together with a Prior NanoScanZ piezo inset to control the z-position. Figure 5. The pivoting motion of astral microtubules helps them to find cortical anchors. When astral microtubules enter the daughter cell, they exhibit pivoting motion around the SPB. This motion may serve as a search mechanism for cortical anchors for dynein. Once the microtubule finds an anchor, dynein can be off-loaded and the spindle is pulled towards the daughter cell. doi:10.1371/journal.pone.0093781.g005 To image the fast angular motion of the astral microtubules, z- stacks of images were acquired with a time resolution of 0.4–0.6 s. This was achieved by reducing the imaging area to the observed cell only. To ensure that the entire spindle and astral microtubules were captured, z-stacks of 4–6 optical slices with a z-spacing of 300–600 nm were acquired. A bright-field image of the cell was acquired immediately before the start of the fluorescence imaging in order to visualize the outline of the cell. A single bright-field image was sufficient because the cells did not move. Materials and Methods Sample Preparation microtubules has two roles. First, it helps to orient the spindle towards the neck inside the mother cell. Second, pivoting facilitates cortical capture of microtubules inside the daughter cell, which is needed for dynein to become off-loaded from the microtubule to the cortical anchors in order to exert force on the microtubule (Fig. 5). Dual Role of Microtubule Pivoting in Spindle Orientation and Movement spindle, thereby inducing their angular motion. In kar9D cells, the observed less extensive pivoting of astral microtubules is most likely due to a combination of dynein-driven motion and thermal motion of microtubules that are not in contact with the cell cortex. While the spindle is in the mother cell, myosin moves the plus end of astral microtubules along actin cables, which are polarized towards the neck, thereby directing the overall pivoting of astral microtubules and the movement of the spindle towards the neck. When the spindle reaches the neck, astral microtubules enter the daughter cell. To pull the spindle through the neck, dynein that is on the microtubules needs to be off-loaded to the cortical anchor protein Num1 [22]. In addition to the forces acting on the plus end of the microtubule, the events at the minus end are important for the pivoting motion. In principle, pivoting is possible if one end of the microtubule is freely jointed to a fixed point, which allows the microtubule to rotate around this point. In the context of budding yeast mitosis, the minus ends of astral microtubules are freely jointed to the SPB. It will be interesting to unravel the molecular basis and the physical properties of this free joint that allows for pivoting of the microtubules. We have shown here that astral microtubule pivoting, which is driven mainly by the Kar9 pathway, reduces the time needed for an astral microtubule inside the daughter cell to generate pulling of the spindle towards or into the daughter cell. We interpret these pulling events as successful microtubule explorations of the cortex in search for the anchor sites, followed by dynein off-loading to the anchor and pulling on the microtubule. Thus, pivoting of PLOS ONE \| www.plosone.org April 2014 \| Volume 9 \| Issue 4 \| e93781 5 Astral Microtubule Pivoting as a Search Strategy Table S2 Astral microtubule lifetime and length. (DOC) Table S2 Astral microtubule lifetime and length. (DOC) Table S3 Strains used in this study. (DOC) Movie S1 Dynamics of the spindle and astral microtu- bules in a wild-type cell. Live-cell microscopy of a budding yeast cell during mitosis. Microtubules are labeled with GFP- TUB1. The spindle is visible as a thick bright line in the center and astral microtubules grow outwards. Note that astral microtubules perform pivoting motion around the SPB. At the beginning of the movie, the astral microtubule extending from the daughter-bound (upper) SPB and the one at the mother-bound (lower) SPB perform extensive rotations around the SPB, while later in the movie the microtubule at the daughter-bound SPB rotates more than the one at the mother-bound SPB. Images were acquired at 0.4-second intervals, represent maximum-intensity projections of 5 slices per time frame, and are displayed at 20 frames per second. The scale bar represents 2 mm. Time is given in minutes:seconds. The movie corresponds to Figure 1B. (MOV) Due to the fast angular movement, growth and shrinkage of astral microtubules, in combination with their dim signal, it was not possible to extend the tracking algorithm described above to astral microtubules. These microtubules were therefore tracked manually by clicking on their tip in the maximum-intensity projection. The measured coordinates were obtained by using the manual tracking plugin for Fiji and stored in a text file. In addition to the coordinates of the spindle poles and astral microtubules, the position of the mother-daughter neck and the mostdistant point of the daughter cell were measured in a single bright-field image of each cell. Movie S2 Dynamics of the spindle and astral microtu- bules in a kar9D cell. Astral microtubules in this cell are significantly longer than in wild type and show less pivoting. In the beginning of the movie, the spindle is roughly perpendicular to the mother-bud axis. Microtubules are labeled with GFP-TUB1. Images were acquired as in Movie S1. The scale bar represents 2 mm. Time is given in minutes:seconds. (MOV) Movie S2 Dynamics of the spindle and astral microtu- bules in a kar9D cell. Astral microtubules in this cell are significantly longer than in wild type and show less pivoting. In the beginning of the movie, the spindle is roughly perpendicular to the mother-bud axis. Microtubules are labeled with GFP-TUB1. Images were acquired as in Movie S1. The scale bar represents 2 mm. Time is given in minutes:seconds. (MOV) Image Processing Thus, the spindle and the astral microtubules are more often found at small than at large angles with respect to the imaging plane, which implies that the measured lengths and angles are closer to the real ones than in the estimate mentioned above. In addition, this systematic error affects the angular diffusion coefficient of the astral microtubules emanating from the daughter-bound and the mother-bound SPB in a similar man Proper ner. Based on these arguments, we conclude that the observed differences in the measured angular diffusion coefficients are unlikely due to the two-dimensional measurements of microtubule length and angle. Detection of Events of Rapid Spindle Pole Movement Events of rapid spindle movement were detected manually in the following way. Cases where the spindle is completely in the mother cell and astral microtubules have not entered the daughter cell yet were selected. In these cases, the time of the microtubule entry into the daughter cell was recorded, as well as the time of the subsequent reaction of the spindle, which was either rapid translation towards the neck or rotation towards the neck. The time difference between microtubule entry and the spindle reaction was denoted as reaction time. g g To automatize the tracking of the mitotic spindle, a new custom-made Fiji plugin was developed and implemented together with Johannes Schindelin (Image Processing Facility, MPI-CBG). When the spindle is in the mother cell, it is usually short (,3 mm long) and straight, thus it was modeled as a straight line. As initialization of the tracking, a straight line was manually drawn along the spindle. In order to reduce the noise, the line width was set to 4 pixels. Before automatically adjusting the drawn line to the spindle, the average intensity of the image was subtracted as background and the image was smoothed by a Gaussian blur of 1– 2 pixels radius. The adjustment algorithm involves two steps: (i) The manually drawn initialization line is rotated until it is aligned to the intensity profile of the spindle. For this purpose line profiles perpendicular to the initial line at its endpoints were taken. To those profiles a Gaussian function was fitted to determine the central point of the intensity profile of the spindle. The central points of the two line profiles were used to calculate the line of the spindle. (ii) The ends of this line were adjusted to the spindle ends by inspecting the intensity profile along the line. The endpoints were defined where the intensity drops to half the maximum value. This two-step algorithm was repeated until convergence and the resulting line was used as initialization for the next frame. The resulting lines corresponding to the spindle position were visually inspected and their position was adjusted manually in cases of error. The coordinates of the spindle poles were exported as a text file for further analysis. Image Processing Image processing was performed with the open-source Fiji software package [32]. Maximum-intensity projections were created from each z-stack and used for the measurements of the position of the spindle and astral microtubules. We did not measure positions along the z-axis because the corresponding point-spread function of the microscope is about 800 nm, which is more than half of the length of a typical microtubule. We have previously estimated the systematic error resulting from two- dimensional measurements of microtubule length in maximum- intensity projections, compared to three-dimensional measure- ments in z-stacks, see Supplementary Note 3 in Ref. [24]. If one assumes that microtubules extend from the SPB isotropically, the real microtubule length is roughly 20% larger than the average measured microtubule length, whereas the real angular diffusion coefficient of the microtubules is roughly 30% smaller than the measured one [24]. In the current work, the corrections are most Figure 5. The pivoting motion of astral microtubules helps them to find cortical anchors. When astral microtubules enter the daughter cell, they exhibit pivoting motion around the SPB. This motion may serve as a search mechanism for cortical anchors for dynein. Once the microtubule finds an anchor, dynein can be off-loaded and the spindle is pulled towards the daughter cell. doi:10.1371/journal.pone.0093781.g005 April 2014 \| Volume 9 \| Issue 4 \| e93781 April 2014 \| Volume 9 \| Issue 4 \| e93781 6 PLOS ONE \| www.plosone.org Astral Microtubule Pivoting as a Search Strategy Astral Microtubule Pivoting as a Search Strategy the cell axis towards the daughter cell. Thus, a lies within the range [2180u, +180u]. likely smaller because the mother-bud axis of the imaged cells lies roughly parallel to the imaging plane. Thus, the spindle and the astral microtubules are more often found at small than at large angles with respect to the imaging plane, which implies that the measured lengths and angles are closer to the real ones than in the estimate mentioned above. In addition, this systematic error affects the angular diffusion coefficient of the astral microtubules emanating from the daughter-bound and the mother-bound SPB in a similar man Proper ner. Based on these arguments, we conclude that the observed differences in the measured angular diffusion coefficients are unlikely due to the two-dimensional measurements of microtubule length and angle. likely smaller because the mother-bud axis of the imaged cells lies roughly parallel to the imaging plane. PLOS ONE \| www.plosone.org Table S1 Spindle translocation efficiency, cell and spindle size. Table S2 Astral microtubule lifetime and length. (DOC) Angles and Distances Movie S3 Dynamics of the spindle and astral microtu- bules in a num1D cell. Microtubules show pivoting motion as in wild type, but the spindle does not enter the bud. Microtubules are labeled with GFP-TUB1. Images were acquired as in Movie S1. The scale bar represents 2 mm. Time is given in minute- s:seconds. (MOV) Data analysis and plotting was done using the open-source R programming language [33]. Coordinates of the spindle poles and microtubule tips were tracked over time in the maximum intensity projections as described above. The length of the spindle and astral microtubules was calculated from the tracked coordinates of the spindle poles and astral microtubule tips. Furthermore, the distance from the spindle poles and the astral microtubule tips associated with the daughter spindle pole to the neck were calculated. The distance from an object to the neck was denoted negative for objects in the mother cell and positive for those in the daughter cell. References 17. Markus SM, Punch JJ, Lee WL (2009) Motor- and tail-dependent targeting of dynein to microtubule plus ends and the cell cortex. Curr Biol 19: 196–205. 1. Markus SM, Kalutkiewicz KA, Lee WL (2012) Astral microtubule asymmetry provides directional cues for spindle positioning in budding yeast. Exp Cell Res 318: 1400–1406. 18. Sheeman B, Carvalho P, Sagot I, Geiser J, Kho D, et al. (2003) Determinants of S. cerevisiae Dynein Localization and Activation. Implications for the Mechanism of Spindle Positioning. Curr Biol 13: 364–372. 2. Segal M, Bloom K (2001) Control of spindle polarity and orientation in Saccharomyces cerevisiae. Trends Cell Biol 11: 160–166. 3. Beach DL, Thibodeaux J, Maddox P, Yeh E, Bloom K (2000) The role of the proteins Kar9 and Myo2 in orienting the mitotic spindle of budding yeast. Curr Biol 10: 1497–1506. 19. Lee WL, Oberle JR, Cooper JA (2003) The role of the lissencephaly protein Pac1 during nuclear migration in budding yeast. J Cell Biol 160: 355–364. 20. Ananthanarayanan V, Schattat M, Vogel SK, Krull A, Pavin N, et al. (2013) Dynein motion switches from diffusive to directed upon cortical anchoring. Cell 153: 1526–1536. 4. Yeh E, Skibbens RV, Cheng JW, Salmon ED, Bloom K (1995) Spindle dynamics and cell cycle regulation of dynein in the budding yeast, Saccharomyces cerevisiae. J Cell Biol 130: 687–700. 21. Pavin N, Tolic-Norrelykke IM (2013) Dynein, microtubule and cargo: a menage a trois. Biochem Soc Trans 41: 1731–1735. y 5. Carminati JL, Stearns T (1997) Microtubules orient the mitotic spindle in yeast through dynein-dependent interactions with the cell cortex. J Cell Biol 138: 629– 641. 22. Lee WL, Kaiser MA, Cooper JA (2005) The offloading model for dynein function: differential function of motor subunits. J Cell Biol 168: 201–207. 6. Eshel D, Urrestarazu LA, Vissers S, Jauniaux JC, van Vliet-Reedijk JC, et al. (1993) Cytoplasmic dynein is required for normal nuclear segregation in yeast. Proc Natl Acad Sci U S A 90: 11172–11176. 23. Vogel SK, Pavin N, Maghelli N, Julicher F, Tolic-Norrelykke IM (2009) Self- organization of dynein motors generates meiotic nuclear oscillations. PLoS Biol 7: e1000087. 7. Li YY, Yeh E, Hays T, Bloom K (1993) Disruption of mitotic spindle orientation in a yeast dynein mutant. Proc Natl Acad Sci U S A 90: 10096–10100. 24. Kalinina I, Nandi A, Delivani P, Chacon MR, Klemm AH, et al. References (2013) Pivoting of microtubules around the spindle pole accelerates kinetochore capture. Nat Cell Biol 15: 82–87. 8. Gerson-Gurwitz A, Movshovich N, Avunie R, Fridman V, Moyal K, et al. (2009) Mid-anaphase arrest in S. cerevisiae cells eliminated for the function of Cin8 and dynein. Cell Mol Life Sci 66: 301–313. 25. Yeh E, Yang C, Chin E, Maddox P, Salmon ED, et al. (2000) Dynamic positioning of mitotic spindles in yeast: role of microtubule motors and cortical determinants. Mol Biol Cell 11: 3949–3961. 9. Saunders WS, Koshland D, Eshel D, Gibbons IR, Hoyt MA (1995) Saccharomyces cerevisiae kinesin- and dynein-related proteins required for anaphase chromosome segregation. J Cell Biol 128: 617–624. 26. Berg HC (1993) Random walks in biology. Princeton, N.J.: Princeton University Press. 10. Miller RK, Rose MD (1998) Kar9p is a novel cortical protein required for cytoplasmic microtubule orientation in yeast. J Cell Biol 140: 377–390. 27. Broersma S (1960) Rotational diffusion constant of a cylindrical particle. J Chem Phys 32: 1626–1631. 11. Yin H, Pruyne D, Huffaker TC, Bretscher A (2000) Myosin V orientates the mitotic spindle in yeast. Nature 406: 1013–1015. 28. Hunt AJ, Gittes F, Howard J (1994) The force exerted by a single kinesin molecule against a viscous load. Biophys J 67: 766–781. p y 12. Lee L, Tirnauer JS, Li J, Schuyler SC, Liu JY, et al. (2000) Positioning of the mitotic spindle by a cortical-microtubule capture mechanism. Science 287: 2260–2262. 29. Tirado MM, de la Torre JG (1979) Translational friction coefficients of rigid, symmetric top macromolecules. Application to circular cylinders. J Chem Phys 71: 2581–2587. 13. Adames NR, Cooper JA (2000) Microtubule interactions with the cell cortex causing nuclear movements in Saccharomyces cerevisiae. J Cell Biol 149: 863– 874. 30. Tolic-Norrelykke IM, Munteanu EL, Thon G, Oddershede L, Berg-Sorensen K (2004) Anomalous diffusion in living yeast cells. Phys Rev Lett 93: 078102. 14. Liakopoulos D, Kusch J, Grava S, Vogel J, Barral Y (2003) Asymmetric loading of Kar9 onto spindle poles and microtubules ensures proper spindle alignment. Cell 112: 561–574. 31. Pavin N, Tolic´-Nørrelykke IM (2014) Swinging a sword: How microtubules search for their targets. Syst Synth Biol: published online 16 February 2014. 32. Schindelin J, Arganda-Carreras I, Frise E, Kaynig V, Longair M, et al. (2012) Fiji: an open-source platform for biological-image analysis. Nat Methods 9: 676– 682. 15. Moseley JB, Sagot I, Manning AL, Xu Y, Eck MJ, et al. Astral Microtubule Pivoting as a Search Strategy Astral Microtubule Pivoting as a Search Strategy Schindelin from the Image Processing Facility of MPI-CBG for help with the development and implementation of the spindle tracker, Daniel James White and Britta Schroth-Diez from the Light Microscopy Facility of MPI- CBG for help with microscopy, Ivana Sˇaric´ for the drawings, Nenad Pavin and the Tolic´ lab for discussions. Movie S5 A kar9D cell without spindle translocation into the bud. Microtubules are labeled with GFP-TUB1. Images were acquired as in Movie S1. The scale bar represents 2 mm. Time is given in minutes:seconds. The movie corresponds to Figure 4E. Acknowledgments Conceived and designed the experiments: SB IMT. Performed the experiments: SB. Analyzed the data: SB. Wrote the paper: SB IMT. Conceived and designed the experiments: SB IMT. Performed the experiments: SB. Analyzed the data: SB. Wrote the paper: SB IMT. We thank Jeffrey Moore, John Cooper, Elaine Yeh and Kerry Bloom for strains, Hannes Weisse for performing pilot experiments, Johannes We thank Jeffrey Moore, John Cooper, Elaine Yeh and Kerry Bloom for strains, Hannes Weisse for performing pilot experiments, Johannes We thank Jeffrey Moore, John Cooper, Elaine Yeh and Kerry Bloom for strains, Hannes Weisse for performing pilot experiments, Johannes (MOV) Movie S4 Microtubule pivoting and spindle transloca- tion in a wild-type cell. Spindle translocation into the bud occurs between 1:40 and 2:30 minutes. Microtubules are labeled with GFP-TUB1. Images were acquired as in Movie S1. The scale bar represents 2 mm. Time is given in minutes:seconds. The movie corresponds to Figure 4B. (MOV) The angles a of the astral microtubules were calculated with respect to the spindle axis, and were defined in direction away from the spindle. The spindle axis was calculated with respect to April 2014 \| Volume 9 \| Issue 4 \| e93781 PLOS ONE \| www.plosone.org 7 References (2004) A conserved mechanism for Bni1- and mDia1-induced actin assembly and dual regulation of Bni1 by Bud6 and profilin. Mol Biol Cell 15: 896–907. 33. R Core Team (2013) R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. Available: http:// www.R-project.org/. 16. Carvalho P, Gupta ML Jr, Hoyt MA, Pellman D (2004) Cell cycle control of kinesin-mediated transport of Bik1 (CLIP-170) regulates microtubule stability and dynein activation. Dev Cell 6: 815–829. April 2014 \| Volume 9 \| Issue 4 \| e93781 PLOS ONE \| www.plosone.org 8
https://openalex.org/W2614310520	https://europepmc.org/articles/pmc5446845?pdf=render	English	null	Endovascular Crossing of Chronic Total Occlusions Using an Impulse: An Explorative Design Study	Cardiovascular engineering and technology	2,017	cc-by	16,037	INTRODUCTION Abstract—In this study we investigated whether exerting an impulse on a Chronic Total Occlusion (CTO) improves the success rate of CTO crossing as compared to the currently used method of statically pushing the guidewire against the CTO. A prototype (Ø2 mm) was developed that generates translational momentum using a spring-loaded indenter and converts it to an impulse during impact. Mechanical perfor- mance was evaluated by measuring the peak force and momentum for different spring compressions and strike distances in air and blood-mimicking ﬂuid. Puncture perfor- mance, in terms of number of punctures, number of strikes to puncture, and energy transfer from the indenter to the CTO, was assessed for six tip shapes (stamp, wedge, spherical, pointed, hollow spherical, and ringed) on three CTO models with different weight percentages of gelatin and calcium. As a control, a Ø0.4 mm rigid rod was tested. A maximum indenter momentum of 1.3 mNs (velocity of 3.4 m/s), a peak force of 19.2 N (vs. 1.5 N reported in literature and 2.7 N for the control), and CTO displacement of 1.4 mm (vs. 2.7 mm for the control) were measured. The spherical and ringed tips were most effective, with on average 2.3 strikes to puncture the most calciﬁed CTO model. The prototype generated sufﬁcient peak forces to puncture highly calciﬁed CTO models, which are considered most difﬁcult to cross during PCI. Furthermore, CTO displacement was minimized, result- ing in a more effective procedure. In future, a smaller, faster, and ﬂexible clinical prototype will be developed. During Percutaneous Coronary Interventions (PCI), a small guidewire (Ø0.36 mm) is gradually dri- ven through the vasculature from an incision point in the groin or wrist towards an occlusion in the coro- naries. Once arrived at the occlusion, a static (axial) load is applied on the guidewire by the interventional cardiologist from outside the body to puncture and cross the occlusion, after which the occlusion is reo- pened using a balloon catheter. Address correspondence to Aime´ e Sakes, Department of Biomechanical Engineering, Faculty of Mechanical, Maritime and Materials Engineering, Delft University of Technology, Mekelweg 2, 2628 CD Delft, The Netherlands. Electronic mail: a.sakes@tudelft.nl Cardiovascular Engineering and Technology, Vol. 8, No. 2, June 2017 ( 2017) pp. 145–163 DOI: 10.1007/s13239-017-0306-1 Endovascular Crossing of Chronic Total Occlusions Using an Impulse: An Explorative Design Study AIME´ E SAKES , MARLEEN VAN DER WIEL, DIMITRA DODOU , and PAUL BREEDVELD Department of Biomechanical Engineering, Faculty of Mechanical, Maritime and Materials Engineering, Delft University of Technology, Mekelweg 2, 2628 CD Delft, The Netherlands (Received 29 April 2016; accepted 28 April 2017; published online 17 May 2017) Associate Editor Ajit P. Yoganathan oversaw the review of this article. INTRODUCTION Coronary Chronic Total Occlusions (CTOs), deﬁned as heavily calciﬁed, complete coronary occlusions of over 3 months old, represent the most challenging lesion type to be crossed during PCI, requiring a high skill level of the inter- ventional cardiologists.1 The development and use of several dedicated guidewires, such as the Conﬁanza Pro (Asahi Intecc, Nagoya, Japan) and the Progress 200T (Abbott Vascular, Abbott Park, IL), dedicated cross- ing and support catheters, such as the Tornus (Asahi Intecc, Nagoya, Japan) and the Crossboss (Boston Scientiﬁc, Natick, MA), crossing tools, such as the Frontrunner XP (Ø0.76–1 mm; Cordis Corporation, Miami, FL), the Crosser Catheter (Ø0.6–1.5 mm; BARD Peripheral Vascular Inc., Tempe, AZ), and the Truepath (Ø0.43 mm; Boston Scientiﬁc, Natich, MA), and crossing strategies (see Sakes et al.2 for a com- prehensive overview) have contributed to a steady in- crease in the technical success rate of PCI in CTOs (i.e., the ability to cross the CTO and to successfully reopen the artery), as well as in the overall procedural success rate of these interventions (i.e., the proportion of procedures with no nosocomial major adverse cardiac events).3 However, the overall procedural success rate is still undesirably low. Whereas experienced operators can achieve success rates of up to 90%, success rates of experienced operators not specialized in CTO PCI are Keywords—Chronic Total Occlusions (CTO), Percutaneous Coronary Interventions (PCI), Medical device design, Cap puncture, Crossing. 145 1869-408X/17/0600-0145/0 2017 The Author(s). This article is an open access publication SAKES et al. 146 lower than 55%.4 The success rate is further depended on the characteristics of the CTO, of which age, increasing occlusion length, tortuosity, and cap ambi- guity are historical predictors for technical and pro- cedural failure, and to a lesser extent the used crossing technique (i.e., antegrade, retrograde, or a combined technique).3,5 Furthermore, the critical buckling load of the crossing guidewire under dynamic impulse loading condition increases with decreasing load duration.9 The goal of this study was to explore the use of dynamic impulse loading for puncturing heavily cal- ciﬁed coronary CTOs. For this purpose, a prototype puncture tool able to deliver an impulse onto a CTO was designed and experimentally tested. Tip Design To deﬁne the prototype diameter required for atraumatic navigation through the vascular system, the diameters of the arteries that potentially need to be crossed during PCI from the incision point to the lesion site were analyzed based on angiographic and ultrasound data in Refs. 10–16. Coronary CTOs are found in the Right Coronary Artery (RCA), Left Anterior Descending (LAD), Left Circumﬂex (LCx), and Left Main Trunk (LMT).5 In order to reach the coronary CTO, two main approaches can be taken17: (1) the radial approach, in which the CTO is reached via the radial artery in the wrist, and (2) the femoral approach, in which the CTO is reached via the common femoral artery. The smallest artery that needs to be crossed to reach the CTO is the RCA, which has a minimum diameter of 2.2 mm.12 Accordingly, the maximum prototype diameter was set to 2 mm. Guidewire buckling can be prevented by increasing the critical load of the guidewire, by decreasing the penetration load of the CTO or by bypassing the CTO using so-called dissection re-entry techniques. In the latter case, the CTO is crossed via the much softer blood vessel wall. Because this method damages the blood vessel wall and re-entry is often challenging,8 we will focus on intraluminal crossing methods instead. Most clinically applied methods of preventing guide- wire buckling aim to increase the critical load of the guidewire. This can be achieved by encompassing the guidewire with a support catheter or by employing a second device as support, such as a balloon catheter (balloon anchoring technique) or a guidewire (wire anchoring technique).5 Despite these measures, cross- ing difﬁculties (mainly due to guidewire buckling) are still observed.5 The prototype tip consists of three components: (1) an actuation component that transfers the transla- tional momentum through the instrument shaft to- wards the tip, (2) an indenter that applies an impulse onto the CTO, and (3) a reload mechanism that re- loads and releases the actuation component. To increase the buckling resistance of the guidewire and prevent the energy dissipation associated with CTO displacement, a possible solution could be to use a dynamic loading method, in which a moving guide- wire or dedicated CTO device would collide with a stationary CTO, converting translational momentum of the guidewire to an impulse exerted on the CTO. INTRODUCTION q ) Failing to cross heavily calciﬁed CTOs intralumi- nally (i.e., crossing through the original coronary lu- men and CTO body) with a guidewire accounts for approximately two-third of PCI failures and is mainly the result of guidewire buckling.6 Guidewire buckling occurs because the required puncture force of the CTO often exceeds the maximum load that the guidewire tip can sustain. Furthermore, the static load applied on the CTO may lead to displacement of the CTO up to 4 mm (estimated based on Thind et al.7). Such a large displacement not only leads to energy dissipation, thereby reducing the maximum force that can be delivered on the CTO, but also increases the chance of blood vessel wall damage due to stretching. Reload Mechanism To allow for multiple impacts onto the CTO, a reload mechanism was implemented in the prototype. To re- load the device, we used a compliant lock mechanism consisting of a gripper with two ﬂexible barbed plates (Fig. 2, yellow) and an inner movable core (Fig. 2, blue). By translating the gripper forward into the hollow sec- tion of the indenter stop, the plates of the gripper deform inwards and passively lock into place. Subsequently, the core is pushed forward into the hollow section to actively lock the grip. We opted for a mechanism that works without control feedback, since most of the haptic feedback will be lost during operation. FIGURE 2. Tip design. The tip consists of ﬁve main parts: an outer shaft (grey), an indenter (green) with indenter stop, a compression spring (purple), a position block (white), and a reload mechanism (gripper: yellow and locking core: blue). Row indications: (1) Neutral position, showing the spring (purple) in rest (free length). (2) Connected position, showing the gripper inserted into the indenter stop. (3) Locked posi- tion, showing the gripper locked into place into the indenter stop by moving the locking core forward. (4) Spring com- pression by pulling the load mechanism backwards. (5) Loa- ded position, showing the spring being compressed by translating the load mechanism with the indenter stop back- wards. (6) Ultimate position, showing the indenter being re- leased and accelerated forward by the compression spring until the indenter stop reaches the position block. Indenter The indenter receives the momentum generated by the compression spring and subsequently applies an impulse onto the CTO. The indenter consists of two parts: the indenter tip and the indenter stop. Indenter Stop In order to make sure that the indenter has a control- lable reach, a stop mechanism was integrated into the design. The stop mechanism is connected to the indenter by means of a rod (Ø1 mm) and is positioned proximal to the position block (Fig. 2). During spring loading, the indenter stop, and thus the indenter, is pulled backwards (i.e., towards the handle of the instrument) until maxi- mum spring compression is reached. Subsequently, the indenter stop is released and accelerated forward by the spring force until it reaches the position block. Indenter Tip To explore the potential of dynamic impulse for puncturing a CTO and to allow for testing multiple tip shapes, the indenter tip was designed to be inter- changeable. For the tip shapes, inspiration was drawn from commonly used indenter shapes in the ﬁeld of rock fracturing and needle interventions. In total, six tip shapes were chosen (Fig. 3): two blunt solid shapes (a right-angled stamp indenter and a spherical indenter with the tip radius equal to the indenter radius), two sharp solid shapes (a wedge indenter and a pointed Endovascular Crossing of Chronic Total Occlusions using an Impulse Endovascular Crossing of Chronic Total Occlusions using an Impulse 147 indenter, both with an edge radius of 0.2 mm), and two hollow shapes (a hollow spherical indenter and a ringed indenter). The hollow-shaped indenters allow for guiding another system through the shaft and tip, such as a guidewire or balloon catheter, in a future (hollow) prototype, which could be advantageous for clinical use. The ringed indenter was designed with a maxi- mized inner passage diameter: the indenter diameter minus four times the edge radius. The hollow spherical indenter was designed with a spherical tip and a Ø0.4 mm lumen, through which the thinnest (coro- nary) guidewires can pass. Actuation Component The actuation component consists of a compression spring located at the distal tip of the instrument. Translational momentum is generated by compressing the spring with a load mechanism and a position block (Fig. 2). By varying the spring compression distance, the momentum (p) of the indenter can be adjusted. Tip Design Here, momentum is deﬁned as mass (m [kg]) 9 velocity (v [m/s]) of the guidewire (p½Ns ¼ mv), and impulse is deﬁned as the integral of the impact peak force (Fpeak [N]) over the time interval dt [s] during which it acts (J½Ns ¼ Fpeak dtÞ. Due to the (high) velocity of the guidewire, the damping of the blood vessel wall and the inertia of the CTO act as a ‘‘counterforce’’ to the impulse, and as such restrict the displacement of the CTO and the associated energy dissipation (Fig. 1). FIGURE 1. Impulse crossing method. The guidewire or crossing tool (blue arrow) collides with the CTO (orange), converting translational momentum to an impulse during im- pact, which in turn punctures the CTO (red jagged line). Due to the high velocity of the crossing tool, the inertia of the CTO and the damping of the blood vessel wall deliver the majority of the reaction force (purple arrows) and thus minimize CTO displacement and energy dissipation to the blood vessel wall (indicated by the spring and damper combination). FIGURE 1. Impulse crossing method. The guidewire or crossing tool (blue arrow) collides with the CTO (orange), converting translational momentum to an impulse during im- pact, which in turn punctures the CTO (red jagged line). Due to the high velocity of the crossing tool, the inertia of the CTO and the damping of the blood vessel wall deliver the majority of the reaction force (purple arrows) and thus minimize CTO displacement and energy dissipation to the blood vessel wall (indicated by the spring and damper combination). FIGURE 1. Impulse crossing method. The guidewire or crossing tool (blue arrow) collides with the CTO (orange), converting translational momentum to an impulse during im- pact, which in turn punctures the CTO (red jagged line). Due to the high velocity of the crossing tool, the inertia of the CTO and the damping of the blood vessel wall deliver the majority of the reaction force (purple arrows) and thus minimize CTO displacement and energy dissipation to the blood vessel wall (indicated by the spring and damper combination). Prototype Tip Design A functional prototype was manufactured based on the ﬁnal tip design (see Figs. 4, 5). We decided to design a rigid prototype in which the shaft is scalable to dif- ferent lengths and allows for future redesign into a ﬂexible prototype. The outer shaft of the prototype tip is subdivided into three parts to improve manufactura- bility and consists of two standardized Ø2.0 9 0.1 mm (Ø 9 wall thickness) stainless steel (SS) capillary tubes (C and E in Fig. 4) that ﬁt around both ends of the position block (containing a Ø1.2 mm central lumen; D in Fig. 4; SS). The compression spring (B in Fig. 5; C00170003016, Associated Spring SPEC, Evasham, FIGURE 2. Tip design. The tip consists of ﬁve main parts: an outer shaft (grey), an indenter (green) with indenter stop, a compression spring (purple), a position block (white), and a reload mechanism (gripper: yellow and locking core: blue). Row indications: (1) Neutral position, showing the spring (purple) in rest (free length). (2) Connected position, showing the gripper inserted into the indenter stop. (3) Locked posi- tion, showing the gripper locked into place into the indenter stop by moving the locking core forward. (4) Spring com- pression by pulling the load mechanism backwards. (5) Loa- ded position, showing the spring being compressed by translating the load mechanism with the indenter stop back- wards. (6) Ultimate position, showing the indenter being re- leased and accelerated forward by the compression spring until the indenter stop reaches the position block. FIGURE 3. Indenter tip shapes. From left to right: (a) stamp indenter, (b) spherical indenter, (c) wedge indenter, (d) pointed indenter, (e), hollow spherical indenter with guidewire passage, and (f) ringed indenter. SAKES et al. 148 SAKES et al. 148 FIGURE 3. Indenter tip shapes. From left to right: (a) stamp indenter, (b) spherical indenter, (c) wedge indenter, (d) pointed indenter, (e), hollow spherical indenter with guidewire passage, and (f) ringed indenter. FIGURE 5. Final design. Letter indications: (A) Indenter, (B) Compression spring, (C) Capillary tube front, (D) Position block, (E) Capillary tube back, (F) Indenter stop, (G1) Gripper reload mechanism connected to (G2) Square sliding bearing, (H) Handgrip gripper connected to G2 (and thus the gripper G1) by means of a screw thread, (I1) Core reload mechanism connected to (I2) Handgrip core, (J) Support capillary tube, (K) Lid, and (L) Enveloping box. Prototype Tip Design G1, G2, and H are inserted in the rectangular cutout of the enveloping box L. UK; Øouter = 1.7 mm; Øwire = 0.32 mm, L = 16 mm; K = 1.3 N/mm; Lmax compression = 3.5 mm) is placed distally to the position block within the capillary tube. The indenter consists of two interconnected (screwable) parts: (1) the indenter tip (A in Fig. 4; Ø1.8 mm; L = 6 mm; SS) with variable tip shapes, and (2) a Ø1.0 mm rod connected to the indenter stop (F in Fig. 4; Ø1.8 mm; SS). The tip of the gripper (Ø1.8 mm; Ø0.5 mm inner lumen; SS; G1 in Fig. 5) ﬁts within the hollow section (containing of a narrowed opening and a 45 slanted surface) of the indenter stop (distal end of F in Fig. 4), whereas the other (proximal) end of the gripper is connected to a square sliding bearing (G2 in Fig. 4; brass) that is designed to ﬁt in the handle. The compliant movement of the gripper, necessary to allow the gripper move through the narrowed opening of the hollow part, is possible by the axial slit (w = 0.35 mm). The locking core (I1 in Fig. 4; Ø0.5 mm; tempered steel) runs through the gripper towards the handle where it is connected to a square sliding bearing (G2 in Fig. 4). part (K in Figs. 5, 6), both made of aluminum, and functions as the handle of the device, allowing for loading, locking, and surmounting of the prototype in an experimental setup (Fig. 6). An axial slot was carved in the bottom of the square sliding bearing (connected to the gripper; G2 in Fig. 5), which, to- gether with a screw in the enveloping box, limits the motion range of the gripper. The gripper is controlled by the square sliding bearing (G2 in Fig. 5) that is connected to a handgrip (H in Fig. 5; brass) by means of a M6 screw thread. The core is connected to a sec- ond handgrip (I2 in Fig. 5; SS) that is guided through the brass handgrip connected to the gripper. Finally, a support part (J in Figs. 5, 6; brass) is connected to the proximal end of the box and guided around the cap- illary tube to prevent buckling and bending of this structure. Prototype Tip Design In the bottom of part L, a slot is present through which a screw is inserted (not indicated) and connected to G2 to limit the reach of the gripper (G1; Fig. 4). G1, G2, and H are inserted in the rectangular cutout of the enveloping box L. FIGURE 4. Prototype tip design and control interface. Letter indications: (A) Indenter, (B) Compression spring, (C) Capil- lary tube front, (D) Position block, (E) Capillary tube back, (F) Indenter stop, (G1) Gripper reload mechanism, (G2) Square sliding bearing connected to the gripper, (H) Handgrip grip- per, (I1) Core reload mechanism, and (I2) Handgrip core. The match is shown for scale purposes. FIGURE 4. Prototype tip design and control interface. Letter indications: (A) Indenter, (B) Compression spring, (C) Capil- lary tube front, (D) Position block, (E) Capillary tube back, (F) Indenter stop, (G1) Gripper reload mechanism, (G2) Square sliding bearing connected to the gripper, (H) Handgrip grip- per, (I1) Core reload mechanism, and (I2) Handgrip core. The match is shown for scale purposes. FIGURE 5. Final design. Letter indications: (A) Indenter, (B) Compression spring, (C) Capillary tube front, (D) Position block, (E) Capillary tube back, (F) Indenter stop, (G1) Gripper reload mechanism connected to (G2) Square sliding bearing, (H) Handgrip gripper connected to G2 (and thus the gripper G1) by means of a screw thread, (I1) Core reload mechanism connected to (I2) Handgrip core, (J) Support capillary tube, (K) Lid, and (L) Enveloping box. In the bottom of part L, a slot is present through which a screw is inserted (not indicated) and connected to G2 to limit the reach of the gripper (G1; Fig. 4). G1, G2, and H are inserted in the rectangular cutout of the enveloping box L. FIGURE 5. Final design. Letter indications: (A) Indenter, (B) Compression spring, (C) Capillary tube front, (D) Position block, (E) Capillary tube back, (F) Indenter stop, (G1) Gripper reload mechanism connected to (G2) Square sliding bearing, (H) Handgrip gripper connected to G2 (and thus the gripper G1) by means of a screw thread, (I1) Core reload mechanism connected to (I2) Handgrip core, (J) Support capillary tube, (K) Lid, and (L) Enveloping box. In the bottom of part L, a slot is present through which a screw is inserted (not indicated) and connected to G2 to limit the reach of the gripper (G1; Fig. 4). Independent Variables 1. Strike distance To evaluate the effect of the distance between the prototype tip and the CTO (e.g., due to incorrect positioning) on the indenter momentum and impact peak force, two strike distances (i.e., the distance from the indenter in neutral position [see Fig. 2] to the object onto which the impulse is delivered) were tested: 1 mm and 3 mm. The former was selected as the minimum distance between the indenter and the target object, as this allowed the indenter to bounce back after the collision. By bouncing back, static loading of the CTO was prevented. The 3 mm distance was just under the maximum indenter reach (designed at 4 mm). FIGURE 7. Working principle of the handle. (0) The gripper is decoupled from the system (neutral position). (1) The control input gripper (red: H), and thus the gripper (yellow: G1), is translated forward to couple the gripper (yellow: G1) with the indenter (A). (2) The gripper (yellow: G1) is locked in place by translating the control input core (blue: I2), and thus the core (I1), forward. (3) The control input gripper (red: H), and with it the control input core (blue: I2), is pulled backwards to load the prototype (loaded position). (4) If desired, the prototype can be locked in the loaded position by rotating the control input gripper (red: H). (5) The prototype is released by pulling the locking core (blue: I2) further backwards. 2. Spring compression distance To evaluate the effect of the actuation force on the indenter momentum, three spring compression distances were tested: 3.5 mm (corresponding to the maximum compression distance of the spring), 2.8 mm, and 2.0 mm. The 2.8 mm and 2.0 mm spring compression distances were realized by twisting (and thus unscrewing) the brass handgrip H (Figs. 4, 5) 360o and 720o, respec- 2. Spring compression distance To evaluate the effect of the actuation force on the indenter momentum, three spring compression distances were tested: 3.5 mm (corresponding to the maximum compression distance of the spring), 2.8 mm, and 2.0 mm. The 2.8 mm and 2.0 mm spring compression distances were realized by twisting (and thus unscrewing) the brass handgrip H (Figs. 4, 5) 360o and 720o, respec- positioned inside the hollow insert of the indenter (connected position in Fig. 2). (2) The handgrip of the core (I2 in Fig. Dependent Variables 1. The indenter momentum (p) generated by the compression spring. p was calculated by mul- tiplying the indenter mass (m) with the indenter velocity (v). 2. The impact peak force (Fpeak) delivered by the indenter. Measurement Variables Measurement Variables Measurement Variables FIGURE 7. Working principle of the handle. (0) The gripper is decoupled from the system (neutral position). (1) The control input gripper (red: H), and thus the gripper (yellow: G1), is translated forward to couple the gripper (yellow: G1) with the indenter (A). (2) The gripper (yellow: G1) is locked in place by translating the control input core (blue: I2), and thus the core (I1), forward. (3) The control input gripper (red: H), and with it the control input core (blue: I2), is pulled backwards to load the prototype (loaded position). (4) If desired, the prototype can be locked in the loaded position by rotating the control input gripper (red: H). (5) The prototype is released by pulling the locking core (blue: I2) further backwards. Handle Design An overview of the steps to load, trigger, and release the indenter is given in Fig. 7. (0) Before loading the spring, the control input core (I2 in Fig. 5) is pulled approximately 10 mm out of the handgrip. (1) The handgrip is pushed into the handle until the gripper is The handle (Figs. 5, 6) mechanically controls the gripper and core at precise distances and allows for locking the mechanism in the ‘‘loaded’’ position. The main enveloping box [26 mm 9 41 mm 9 25 mm (l 9 w 9 h)] consists of a bottom (L in Figs. 5, 6) and a top 149 Endovascular Crossing of Chronic Total Occlusions using an Impulse FIGURE 6. The complete assembled prototype. Letter indi- cations: for (A–I2) see Fig. 4, (J) Support capillary tube, (K) Lid, and (L) Enveloping box. Part G (not indicated, see Fig. 4) is placed inside part L. In the bottom of part L a screw is inserted (not indicated) to limit the reach of the gripper (G1; Fig. 4). At this stage, the gripper and indenter can be locked in the ‘‘loaded’’ position (or the spring compression dis- tance can be altered) by twisting the handgrip (H in Figs. 5, 6). (5) To release the spring, the core handgrip (I2 in Figs. 4, 5) is pulled out of the handle, releasing the gripper and allowing the indenter to accelerate (towards the ultimate position in Fig. 2). Experiment 1: Mechanical Performance of the Prototype FIGURE 6. The complete assembled prototype. Letter indi- cations: for (A–I2) see Fig. 4, (J) Support capillary tube, (K) Lid, and (L) Enveloping box. Part G (not indicated, see Fig. 4) is placed inside part L. In the bottom of part L a screw is inserted (not indicated) to limit the reach of the gripper (G1; Fig. 4). FIGURE 6. The complete assembled prototype. Letter indi- cations: for (A–I2) see Fig. 4, (J) Support capillary tube, (K) Lid, and (L) Enveloping box. Part G (not indicated, see Fig. 4) is placed inside part L. In the bottom of part L a screw is inserted (not indicated) to limit the reach of the gripper (G1; Fig. 4). In Experiment 1, the indenter momentum and the impact peak force were measured as a function of the strike distance, the spring compression distance, and the medium through which the indenter was launched. Data Analysis Per measurement, the indenter velocity was deter- mined by dividing the travelled distance of the indenter tip (from neutral position to impact) with the elapsed time (0.1 ms per image) using the associated software package Photron FASTCAM Viewer v3.5.3. Subse- quently, the indenter moment per measurement was determined by multiplying the indenter velocity with Measurement Protocol The impact peak force was measured for all 12 conditions deﬁned by the three independent variables: 2 strike distances 9 3 spring compression distances 9 2 surrounding media. The indenter velocity was mea- sured only in air, thus for 6 conditions (2 strike dis- tances 9 3 spring compression distances), as it was not possible to video capture the motion in BMF. Each condition was tested ﬁve times. In Cheng et al.,21 an exponential formula was presented to calculate the viscosity of a glyc- erol-water mixture. The formula was numeri- cally derived based on experimental data from earlier studies and was used to develop and validate a correlation model that estimates the viscosity of glycerol-water mixtures. The for- mula applies for all glycerol concentration (0 and 100 wt%) and temperatures between 0 and 100 C. According to this formula, the viscos- ity of blood (2.5–2.8 mPas at 36–40 C) at an ambient temperature of 18 C can be approx- imated with a mixture of 25 wt% glycerine and 75 wt% clear water. Independent Variables The BMF was made of 25% weight percentage (wt) glycerine and 75 wt% clear water, a mixture widely used to simulate blood (sometimes with added element such as nylon particles and sodium iodide to improve the optical properties for Doppler ﬂow and particle image velocimetry measurements).18–21 21 3. Surrounding medium The mechanical perfor- mance was evaluated in air and in Blood- Mimicking Fluid (BMF), with the latter approximating the clinical environment more closely than air. The BMF was made of 25% weight percentage (wt) glycerine and 75 wt% clear water, a mixture widely used to simulate blood (sometimes with added element such as nylon particles and sodium iodide to improve the optical properties for Doppler ﬂow and particle image velocimetry measurements).18–21 21 3. Surrounding medium The mechanical perfor- mance was evaluated in air and in Blood- Mimicking Fluid (BMF), with the latter approximating the clinical environment more closely than air. The BMF was made of 25% weight percentage (wt) glycerine and 75 wt% clear water, a mixture widely used to simulate blood (sometimes with added element such as nylon particles and sodium iodide to improve the optical properties for Doppler ﬂow and particle image velocimetry measurements).18–21 In Cheng et al.,21 an exponential formula was presented to calculate the viscosity of a glyc- erol-water mixture. The formula was numeri- cally derived based on experimental data from earlier studies and was used to develop and validate a correlation model that estimates the viscosity of glycerol-water mixtures. The for- mula applies for all glycerol concentration (0 and 100 wt%) and temperatures between 0 and 100 C. According to this formula, the viscos- ity of blood (2.5–2.8 mPas at 36–40 C) at an ambient temperature of 18 C can be approx- imated with a mixture of 25 wt% glycerine and 75 wt% clear water. Measurement Setup The prototype was suspended vertically, with its tip pointing downwards, in a construction of Thorlabs optomechanics (Thorlabs, Inc., Newton, NJ), consisting of a solid aluminum breadboard (MB3030/M), a con- struction rail (XE25L375/M), and a single-axis transla- tion stage (PT1/M) (Fig. 8). The prototype was fastened on the translation stage using an intermediate plate construction. The indenter strike was received by a small aluminum table (for the measurements in air) or bucket (in which the prototype tip was immersed for the mea- surements in BMF) fastened on top of a miniature load cell (LSB200, FSH00103, FUTEK Advanced Sensor Technology, Inc., Irvine, CA; see Fig. 8). The effect of gravity in this experiment was considered negligible, since the estimated gravity force is three orders of magnitude smaller than the maximum spring force. FIGURE 8. Measurement setup for Experiment 1. The mea- surement setup consisted of a breadboard, (25 mm con- struction) rail, and a single-axis translation stage to which the prototype was connected. The prototype tip was pointing downwards onto a miniature load cell, on top of which a small aluminum table or bucket was fastened to receive the impact peak force in air and blood mimicking ﬂuid, respectively. For the velocity measurement, a HSV-camera was placed next to the breadboard. FIGURE 8. Measurement setup for Experiment 1. The mea- surement setup consisted of a breadboard, (25 mm con- struction) rail, and a single-axis translation stage to which the prototype was connected. The prototype tip was pointing downwards onto a miniature load cell, on top of which a small aluminum table or bucket was fastened to receive the impact peak force in air and blood mimicking ﬂuid, respectively. For the velocity measurement, a HSV-camera was placed next to the breadboard. For the determination of the indenter momentum, the indenter velocity was measured from High Speed Videos (HSVs) made with a Photron Fastcam APX-RS (Photron, Inc., San Diego, CA) at a frame rate of 10 kHz, and the indenter mass was measured with a high precision balance (Mettler PJ360 DeltaRange, Mettler-Toledo International Inc., Columbus, OH). FIGURE 8. Measurement setup for Experiment 1. The mea- surement setup consisted of a breadboard, (25 mm con- struction) rail, and a single-axis translation stage to which the prototype was connected. Independent Variables 5, 6) is pushed into the handle to lock the gripper in place (locked position in Fig. 2). (3) Both the handgrip and control input core are pulled out of the handle until the connection between the handgrip and the square sliding bearing becomes visi- ble (G2 and H in Fig. 5; loaded position in Fig. 5). (4) 150 SAKES et al. tively. The 2.0 mm spring compression dis- tance corresponded to the minimum distance required for the indenter to reach an object at 3 mm distance. tively. The 2.0 mm spring compression dis- tance corresponded to the minimum distance required for the indenter to reach an object at 3 mm distance. The impact peak force was measured with the miniature load cell, connected to an analogue signal conditioner (CPJ RAIL, SCAIME, Annemasse, France) and a data acquisition system with a sampling rate set to 50 kHz (NI USB-6211, National Instru- ments Corporation, Austin, TX). This system was controlled through LabVIEW 2014 (National Instru- ments Corporation, Austin, TX). For each of the im- pact peak force measurements, 1 s of data before and 1 s after the indenter strike were sampled. 3. Surrounding medium The mechanical perfor- mance was evaluated in air and in Blood- Mimicking Fluid (BMF), with the latter approximating the clinical environment more closely than air. The BMF was made of 25% weight percentage (wt) glycerine and 75 wt% clear water, a mixture widely used to simulate blood (sometimes with added element such as nylon particles and sodium iodide to improve the optical properties for Doppler ﬂow and particle image velocimetry measurements).18–21 In Cheng et al.,21 an exponential formula was presented to calculate the viscosity of a glyc- erol-water mixture. The formula was numeri- cally derived based on experimental data from earlier studies and was used to develop and validate a correlation model that estimates the viscosity of glycerol-water mixtures. The for- mula applies for all glycerol concentration (0 and 100 wt%) and temperatures between 0 and 100 C. According to this formula, the viscos- ity of blood (2.5–2.8 mPas at 36–40 C) at an ambient temperature of 18 C can be approx- imated with a mixture of 25 wt% glycerine and 75 wt% clear water. 3. Surrounding medium The mechanical perfor- mance was evaluated in air and in Blood- Mimicking Fluid (BMF), with the latter approximating the clinical environment more closely than air. Dependent Variables 2. Number of strikes for puncture This measure was used as an indication of the efﬁciency of the puncture (scale classiﬁcation). A CTO is a heterogeneous mix of materials, including intracellular and extracellular lipids, smooth muscle cells, a collagen-rich extracellular matrix, cholesterol, dense collagen, and calcium.1,17 Collagen forms the major structural component of the extracellular matrix of the CTO. CTOs can be soft, hard, or contain both hard and soft regions.1 Soft CTOs are usually younger than hard CTOs and mainly consist of fat-laden cells and loose ﬁbrous tissue, whereas hard CTOs are characterized by dense ﬁbrous tissue and contain cal- ciﬁed regions.1 This age-related increase in calcium and collagen content of CTOs substantiates the difﬁculty of crossing older occlusions. p ( ) 3. Energy transfer from the indenter to the CTO model The main working principle of the im- pact method is the collision between a moving body (the indenter) and an initially non-mov- ing body (the CTO). During impact, kinetic energy of the indenter can be transferred into kinetic energy of the CTO or dissipated into heat, deformation, or fracture. There are two main types of collisions that can occur: (1) an elastic collision and (2) an inelastic collision. An elastic collision is featured by conservation of kinetic energy, whereas in the case of an inelastic collision kinetic energy dissipates into other forms of energy. For our application, the elastic collision is drawn as the ‘‘worst case scenario’’, as both the indenter momentum and kinetic energy are conserved and transferred to the CTO, leading to CTO displacement. The ‘‘ideal scenario’’, on the other hand, is a perfect inelastic collision, where the two colliding bodies stick together after collision, and the maximum amount of kinetic energy is ab- sorbed by the CTO, potentially resulting in p ( ) 3. Energy transfer from the indenter to the CTO model The main working principle of the im- pact method is the collision between a moving body (the indenter) and an initially non-mov- ing body (the CTO). During impact, kinetic energy of the indenter can be transferred into kinetic energy of the CTO or dissipated into heat, deformation, or fracture. There are two main types of collisions that can occur: (1) an elastic collision and (2) an inelastic collision. Measurement Setup The prototype tip was pointing downwards onto a miniature load cell, on top of which a small aluminum table or bucket was fastened to receive the impact peak force in air and blood mimicking ﬂuid, respectively. For the velocity measurement, a HSV-camera was placed next to the breadboard. Endovascular Crossing of Chronic Total Occlusions using an Impulse 151 which will eventually develop into HA crystals and thus calciﬁcation.21 the indenter mass. Finally, per condition, the mean indenter velocity and mean indenter momentum with their associating standard deviations were determined across the ﬁve repetitions. Along the length of the CTO, three regions can be distinguished, with distinct material properties: (1) the proximal cap, (2) the main body (core), and (3) the distal cap of the CTO.17 The proximal cap is a thick- ened ﬁbrous structure at the proximal end (i.e., the ﬁrst side encountered by the blood ﬂow) of the CTO. It is the hardest part of the lesion and contains particularly densely-packed collagen and calciﬁed tissue. The distal cap is also a thickened structure but somewhat thinner and softer than the proximal cap. The core between the two caps is softer and consists mostly of organized thrombus and lipids.1 The data from the miniature load cell was processed with MATLAB 2013b (The Mathworks, Inc., Natick, MA) to identify the impact peak force (highest value measured) for each measurement and the mean impact peak force (i.e., the mean of the ﬁve impact peak forces measured) per condition. Furthermore, to determine the eﬀect of the surrounding medium on the impact peak force, the ratio between the mean peak force in BMF was compared to that in air (Fpeak BMF/Fpeak Air [-]) for each condition. In order to evaluate the performance of the proto- type to cross a CTO, we decided to focus on punc- turing the proximal cap of heavily calciﬁed CTOs, as this is the hardest, most diﬃcult part to penetrate during PCI. Puncture performance was evaluated in terms of eﬃciency and eﬃcacy. Speciﬁcally, the fol- lowing three measures were deﬁned: Experiment 2: Puncture Performance of the Prototype Experiment 2: Puncture Performance of the Prototype In Experiment 2, the puncture performance of the prototype was evaluated on a CTO model by mea- suring the number of punctures, the puncture success rate, the number of indenter strikes necessary to achieve a puncture, and the distribution of energy dissipation between the indenter and the CTO model, as a function of the indenter tip shape and the hardness of the CTO model. 1. Number of punctures This measure was used as an indication of the efﬁcacy of the puncture (binary classiﬁcation: puncture vs. no punc- ture). In our experiment, puncture is deﬁned as complete breakage of the proximal cap. The maximum number of strikes allowed per trial was set to 10. Measurement Variables Dependent Variables An elastic collision is featured by conservation of kinetic energy, whereas in the case of an inelastic collision kinetic energy dissipates into other forms of energy. For our application, the elastic collision is drawn as the ‘‘worst case scenario’’, as both the indenter momentum and kinetic energy are conserved and transferred to the CTO, leading to CTO displacement. The ‘‘ideal scenario’’, on the other hand, is a perfect inelastic collision, where the two colliding bodies stick together after collision, and the maximum amount of kinetic energy is ab- sorbed by the CTO, potentially resulting in 3. Energy transfer from the indenter to the CTO model The main working principle of the im- pact method is the collision between a moving body (the indenter) and an initially non-mov- ing body (the CTO). During impact, kinetic energy of the indenter can be transferred into kinetic energy of the CTO or dissipated into heat, deformation, or fracture. There are two main types of collisions that can occur: (1) an elastic collision and (2) an inelastic collision. The calciﬁcation process is similar to bone devel- opment and generally consists of two main mecha- nisms.17 (1) High phosphate levels in the CTO cause smooth muscle cells to differentiate into osteoblasts, eventually resulting in the formation of HydroxyA- patite (HA; a calcium phosphate usually denoted as Ca10(OH)2(PO4)6)) crystals and thus calciﬁed regions.21 (2) Cytokines (i.e., proteins that can signal other cells) signalize osteoblast and osteoblast-like cells to the CTO from circulating stem cells, smooth muscle cells, or pericytes (i.e., contractile cells that wrap around the endothelium of the blood vessel wall), 152 SAKES et al. implanting hydroxyapatite-coated bio-ab- sorbable polymer sponges in the coronaries. Even though the use of animal models is pre- ferred for device evaluation, these are expen- sive and do not allow for consecutive tests and evaluation under constant test conditions. Therefore, it was decided to build an artiﬁcial CTO model to determine the feasibility of technology. The proximal cap model was made from calcium sulfate (CaSO4; SHERAALPIN Hartgips hellblau, SHERA Werkstoff-Tech- nologie GmbH & Co. KG) and a gelatin mix- ture (sheet gelatin, Dr. Oetker, Bielefeld, Germany). These materials were selected to mimic the high calcium and collagen concen- trations found in the proximal caps of heavily calciﬁed CTOs. Calcium sulfate is a mineral that can be found in the human body and has similar mechanical properties as organic HA crystals. Dependent Variables Speciﬁcally, the compressive strength of CaSO4 is 44 MPa (after 1 h) and that of (carbonated) HA formed at body temperature approximately 30–70 MPa.23 Higher compres- sive strengths are found in calcium-deﬁcient HA (Ca9HOH(PO4)6 (approximately 170 MPa) and sintered HA (up to 500 MPa).23 However, the HA crystals found in CTOs are most often re- ferred to as carbonated HA; it is unclear whe- ther calcium-deﬁcient HA is present in CTOs and whether sintered HA is formed in a differ- ent way than natural HA (as it is manufactured using signiﬁcantly higher temperatures than body temperature). Therefore, the lower com- pressive strength values found for carbonated HA seem more likely to represent naturally found HA in CTOs. CTO puncture. Whether the occurring collision is elastic or inelastic depends on the charac- teristics of the CTO and its environment, such as the CTO mass, damping coefﬁcient, and spring constant, as well as on the velocity and mass of the indenter. To determine the type of collision, the following two variables were measured: – Indenter bounce velocity To estimate the amount of energy absorbed by the environ- ment and the CTO combined and to com- pare this energy absorption in the cases of puncture versus no puncture, the average velocity of the indenter after bouncing away from the proximal cap was measured. – CTO displacement To determine the type of collision (i.e., elastic or inelastic) and esti- mate the amount of energy absorbed by the environment and CTO separately, the CTO displacement was measured. All measurements in Experiment 2 were conducted with the maximized impact condition as identiﬁed from Experiment 1 (i.e., strike distance of 1 mm and spring compression distance of 3.5 mm). A pilot study we conducted indicated that testing with lower impact forces would only provide trivial evidence of a lower eﬃcacy and eﬃciency than the maximized condition. Measurement Setup and 50 wt% gelatin mixture in a Ø10 mm cir- cular slice and represented the mean calcium percentage in CTOs in the PCI-failure group in Cho et al.24 The reference model was created with 75 wt% calcium powder and 25 wt% liquid gelatin mixture. Whereas reported average concentrations of calcium content within calci- ﬁed CTOs are generally lower than 75%,24 CTO caps are heterogeneous and thus local calcium concentrations as high as 75% are expected. Next to the reference model, a brittle model (77 wt% calcium powder and 23 wt% clear water) was also prepared to represent the most heavily calciﬁed CTOs that can be encountered. For Experiment 2, the prototype was suspended in the same way as described in Experiment 1 (see Fig. 10). The sole difference was that, instead of load cell, a CTO model was placed directly under the pro- totype (Fig. 9). Measurement Protocol The number of punctures and the number of strikes were measured for all 18 conditions deﬁned by the two independent variables (6 indenter tip shapes x 3 degrees of model hardness). Additionally, to evaluate the overall eﬀectiveness of the impact method as a function of the CTO model hardness, the results from all six tip shapes were grouped together per CTO model, result- ing in a data analysis of 3 groups within the 18 per- formed trials. Each of the ductile, reference, and brittle models were prepared in a single batch to limit the variability in the CTO models, and all measurements were conducted within 4 h, to limit the effect of CTO model hardening over time. Three measurements were conducted per condition. Proximal caps are on average 0.5 mm thick.25 Due to molding difﬁculties to produce homo- geneous 0.5 mm models, a model thickness of 1 mm was set. Unlike acute occlusions, CTOs are strongly con- nected to the blood vessel wall, which in turn is con- nected to cardiac muscle tissue of the heart. In order to simulate the environment of the CTO while allowing for consecutive measurement of the prototype under stable test conditions, it was chosen to simulate the environment of the CTO using a mixture of 25 wt% gelatin and 75 wt% water. When set, this mixture has an estimated Young’s modulus between 100 and 130 kPa, resembling the Young’s modulus of cardiac muscle tissue.26 Furthermore, to emulate the strong blood vessel wall connection, a Ø10 mm notch with approximately 7 mm depth was created in the center of the gelatin mixture (Ø50 mm, 26 mL; Fig. 9) in which the proximal cap model was placed. This notch also allowed for simultaneous stretching and compressing of the gelatin, similar to the clinical situation. For a complete overview of the steps followed to manufac- ture the CTO models, see the Appendix. The measurements of the energy transfer from the indenter to the CTO model were conducted with the spherical tip shape. One energy transfer measurement was conducted for the ductile and reference models and two for the brittle model. Independent Variables The puncture performance of the prototype was eval- uated as a function of the following two variables: 1. Indenter tip shape The tip shape that is the most efﬁcient and effective in puncturing a CTO depends on the material properties of the CTO. Fracture of brittle materials leads to material pulverization and chip formation and is usually conducted with blunt tip shapes. Fracture of ductile materials, on the other hand, associates with material tearing and is commonly achieved by using sharp tip shapes, such as needles. Due to the uncertainty of the material properties of the CTO and its envi- ronment, the previously described tip shapes were tested: stamp, spherical, wedge, pointed, hollow spherical with guidewire passage, and ringed. To mimic the densely-packed collagen and overall heterogeneous morphology of the CTO at the proximal cap, the calcium sulfate was mixed with gelatin, which is the product of structural and chemical degradation of collagen. Transluminal calcium ‡50% (as assessed by Multi-Detector Computed Tomography (MDCT)) is a strong predictor for PCI failure. In a study of Cho et al.,24 the calcium burden of CTOs was determined using MDCT. In this study a mean calcium percentage (determined as the percentage of calcium cross-sectional area divided by vessel area) of 53.9 ± 20.3% was found in the PCI-failure cases. As we focus on crossing the most calciﬁed CTOs, proximal caps with three degrees of hardness were tested (called henceforth ductile, reference, and brittle) with a calcium percentage over 50%. The ductile model was created with 50 wt% calcium powder 2. Proximal cap hardness The development of a representative CTO animal model has been proven difﬁcult mainly due to the lack of spontaneous atherosclerosis in animals.1 Re- cently, Suzuki et al.22 succeeded in developing an animal model (pig and rabbit) that was histologically similar to human CTOs by Endovascular Crossing of Chronic Total Occlusions using an Impulse 153 Data Analysis Per indenter tip shape and CTO model, the number of strikes to achieve puncture and the number of punctures were counted. Furthermore, per CTO model (tip shapes grouped together), the number of punc- tures, the success rate (deﬁned as the percentage of FIGURE 10. Measurement setup for Experiment 2. The mea- surement setup consisted of the same Thorlabs facility as in Experiment 1 (see Fig. 10), with the only difference that in- stead of the load cell, the CTO model (see Fig. 11) was placed directly under the prototype tip. FIGURE 9. The CTO model. The CTO model consisted of a small container (Ø50 mm, 26 mL) ﬁlled with gelatin repre- senting the surrounding cardiac tissue (environment model), in which a Ø1 mm, 1 mm thick slice of stiffened plaster cast mixed with gelatin/water was placed. FIGURE 9. The CTO model. The CTO model consisted of a small container (Ø50 mm, 26 mL) ﬁlled with gelatin repre- senting the surrounding cardiac tissue (environment model), in which a Ø1 mm, 1 mm thick slice of stiffened plaster cast mixed with gelatin/water was placed. FIGURE 9. The CTO model. The CTO model consisted of a small container (Ø50 mm, 26 mL) ﬁlled with gelatin repre- senting the surrounding cardiac tissue (environment model), in which a Ø1 mm, 1 mm thick slice of stiffened plaster cast mixed with gelatin/water was placed. FIGURE 10. Measurement setup for Experiment 2. The mea- surement setup consisted of the same Thorlabs facility as in Experiment 1 (see Fig. 10), with the only difference that in- stead of the load cell, the CTO model (see Fig. 11) was placed directly under the prototype tip. SAKES et al. 154 punctures across the 18 conditions), as well as the minimum, maximum, and mean number of strikes across the 18 conditions to achieve puncture were determined. mean indenter momentum of 1.33 mNs (for m = 0.39 grams). Decreasing the compression spring distance and increasing the strike distance was associated with a decrease in indenter velocity. In the condition of min- imal spring compression distance (2 mm), the indenter did not reach the ultimate position (3 mm). Impact Peak Force in BMF Figure 13 shows the impact peak force measurements in BMF. Data were approximately normal. The impact peak force in BMF was approximately a factor 0.49– 0.11 lower than that of the impact peak force in air (Table 2). The strike distance was of greater inﬂuence on the impact peak forces in BMF than in air, which is illustrated by a lower impact peak force in BMF in Control Experiment Strike distance/spring compression distance Indenter velocity (v; m/s) Indenter momentum (p; mNs) 1 mm/3.5 mm 3.4 ± 0.11 1.33 ± 0.04 3 mm/3.5 mm 3.0 ± 0.16 1.17 ± 0.06 1 mm/2.8 mm 2.2 ± 0.07 0.86 ± 0.03 3 mm/2.8 mm 1.3 ± 0.20 0.51 ± 0.08 1 mm/2.0 mm 1.0 ± 0.05 0.39 ± 0.02 3 mm/2.0 mma – – aFor this condition, the velocity of the indenter could not be determined with accuracy due to the minimal travelled distance of the indenter out of the outer shaft. TABLE 1. Velocity (mean 6 standard deviation, n 5 5) and mean indenter momentum (p 5 mv) for the six conditions. Control Experiment In order to compare the performance of the newly developed device with that of currently used guidewires and to assess the validity of the used CTO model, a separate control experiment was performed. In this experiment, a rigid Ø0.4 mm stainless steel rod was used to model the ﬂexible guidewire tip. The rigid rod model was driven into the ductile and reference CTO model (equal to the previously described models) by a linear stage. The force required to puncture the CTO model as well as the resulting displacement of the model were measured. The velocity of the insertion ranged between 0.1, 0.5, 5, 50, 500, and 1000 mm/s, to simulate low velocity and high velocity insertion. Currently used insertion velocities of the guidewire are not described in literature. Therefore, a wide velocity range was tested, with the higher insertion velocities chosen to determine the eﬀect of high-speed insertion. FIGURE 11. Example of an HSV image sequence of an indenter strike with an object distance of 3 mm and a spring compression distance of 3.5 mm. An average indenter veloc- ity of 3.0 m/s can be derived. FIGURE 11. Example of an HSV image sequence of an indenter strike with an object distance of 3 mm and a spring compression distance of 3.5 mm. An average indenter veloc- ity of 3.0 m/s can be derived. FIGURE 11. Example of an HSV image sequence of an indenter strike with an object distance of 3 mm and a spring compression distance of 3.5 mm. An average indenter veloc- ity of 3.0 m/s can be derived. FIGURE 11. Example of an HSV image sequence of an indenter strike with an object distance of 3 mm and a spring compression distance of 3.5 mm. An average indenter veloc- ity of 3.0 m/s can be derived. TABLE 1. Velocity (mean 6 standard deviation, n 5 5) and mean indenter momentum (p 5 mv) for the six conditions. Data Analysis To estimate energy transfer from the indenter to the CTO, the indenter momentum and kinetic energy (Ekinetic ¼ 1 2 mv2, with m = mass of the indenter [kg] and v = the velocity of the indenter [m/s]) 0.1 ms be- fore and 0.1 ms after impact was calculated using the indenter velocity derived from the HSV images using the method previously described. Furthermore, the maximum displacement of the CTO model per mea- surement was manually derived from the captured HSV images, and the energy absorbed by the (gelatine) environment was calculated as U ¼ 1 2 VEe2, with V = volume of the gelatine (26 mL), E = Young’s modulus of the gelatine (115 kPa), and e = strain of the gelatine during impact (approximated as the dis- placement of the proximal cap model divided by the original location). Impact Peak Force Impact Peak Force in Air Impact Peak Force in Air Figure 12 shows the impact peak force measurements in air. The data were approximately normal (based on a one-sample Kolmogorov–Smirnov test). The impact peak forces (mean ± standard deviation, n = 5) are presented in Table 2. The maximum mean impact peak force was approximately 19.2 N, generated with max- imized actuation force and independent of the strike distance. The impact peak force decreased with decreasing spring compression distance. aFor this condition, the velocity of the indenter could not be determined with accuracy due to the minimal travelled distance of the indenter out of the outer shaft. Energy Transfer from the Indenter to the CTO Model The average velocity of the (spherical) indenter after bouncing from the proximal cap model was approxi- mately 0.5 m/s in the cases in which the proximal cap model was not punctured (Table 5). No differences FIGURE 12. Impact peak force of the indenter in air. Each color (n 5 5 strikes) represents a different strike distance x spring compression distance combination. For each indenter strike, represented by one line, time 0 corresponds to the maximum impact peak force. FIGURE 12. Impact peak force of the indenter in air. Each color (n 5 5 strikes) represents a different strike distance x spring compression distance combination. For each indenter strike, represented by one line, time 0 corresponds to the maximum impact peak force. TABLE 2. Impact peak force Fpeak (mean 6 standard deviation, n 5 5) in air and in blood mimicking ﬂuid (BMF) for the six conditions, and ratio of the impact peak force in BMF and air (Fpeak BMF/Fpeak air). Strike distance/spring compression distance Fpeak air (N) Fpeak BMF (N) Fpeak BMF/Fpeak air (-) 1 /3.5 mm 18.9 ± 0.55 9.2 ± 0.08 0.49 3 /3.5 mm 19.2 ± 0.65 7.9 ± 0.75 0.41 1 /2.8 mm 13.0 ± 0.20 5.8 ± 0.29 0.45 3 /2.8 mm 12.9 ± 0.58 4.1 ± 0.74 0.32 1 /2.0 mm 6.4 ± 0.13 2.4 ± 0.20 0.38 3 /2.0 mm 6.4 ± 0.41 0.7 ± 0.68 0.11 FIGURE 13. Impact peak force of the indenter in BMF. Each color (n 5 5 strikes) represents a different strike distance x spring compression distance combination. For each indenter strike, represented by one line, time 0 corresponds to the maximum impact peak force. FIGURE 13. Impact peak force of the indenter in BMF. Each color (n 5 5 strikes) represents a different strike distance x spring compression distance combination. For each indenter strike, represented by one line, time 0 corresponds to the maximum impact peak force. FIGURE 13. Impact peak force of the indenter in BMF. Each color (n 5 5 strikes) represents a different strike distance x spring compression distance combination. For each indenter strike, represented by one line, time 0 corresponds to the maximum impact peak force. FIGURE 12. Impact peak force of the indenter in air. Each color (n 5 5 strikes) represents a different strike distance x spring compression distance combination. Endovascular Crossing of Chronic Total Occlusions using an Impulse Endovascular Crossing of Chronic Total Occlusions using an Impulse 155 passage (2.7 strikes on average), and the ringed (2 strikes on average) tip shape. The pointed tip shape was generally less effective and efﬁcient to puncture the brittle model, as compared to all other tip shapes. Speciﬁcally, the only case in which the brittle model was not punctured occurred with the pointed tip shape. Moreover, in its two successful cases, this indenter needed 8 strikes to puncture the model, which is much higher than the average of 2.5 strikes required with the other ﬁve tip shapes (n = 15). comparison to air and Fpeak BMF/Fpeak air ratio for a strike distance of 3 mm. comparison to air and Fpeak BMF/Fpeak air ratio for a strike distance of 3 mm. Experiment 2: Puncture Performance of the Prototype Number of Punctures, Success Rate, and Number of Strikes Table 3 shows the puncture performance for the different indenter tip shapes (see also Fig. 14 for a typical result of multiple impact strikes on the three CTO models with the spherical indenter). For the ductile model, the indenter left an imprint of plastic material deformation, but did not manage to penetrate the model within 10 strikes. The measurements on the reference model did not show convincing results to indicate which tip shapes were the most effective and efﬁcient. However, the wedge tip shape was the only tip shape able to puncture the model during all three tests (6 strikes on average). Finally, for the brittle model the most efﬁcient and effective tip shapes were the spheri- cal (2.7 strikes on average), the spherical with guidewire Table 4 shows the number of punctures, success rate, and the number of strikes for puncture, for the ductile, reference, and brittle models. It can be seen that the number of punctures (efﬁcacy) and the success rate increased, and number of strikes for puncture (efﬁ- ciency) decreased with increasing model hardness. Indenter Momentum Indenter Momentum In Fig. 11, a typical HSV image sequence of an indenter strike is illustrated. The calculated momenta for each of the six tested conditions (2 strike distances 9 3 spring compression distances) are illustrated in Table 1. The mean velocity generated with maximized actuation force was 3.4 m/s, which translates to a Energy Transfer from the Indenter to the CTO Model For each indenter strike, represented by one line, time 0 corresponds to the maximum impact peak force. TABLE 2. Impact peak force Fpeak (mean 6 standard deviation, n 5 5) in air and in blood mimicking ﬂuid (BMF) for the six conditions, and ratio of the impact peak force in BMF and air (Fpeak BMF/Fpeak air). Strike distance/spring compression distance Fpeak air (N) Fpeak BMF (N) Fpeak BMF/Fpeak air (-) 1 /3.5 mm 18.9 ± 0.55 9.2 ± 0.08 0.49 3 /3.5 mm 19.2 ± 0.65 7.9 ± 0.75 0.41 1 /2.8 mm 13.0 ± 0.20 5.8 ± 0.29 0.45 3 /2.8 mm 12.9 ± 0.58 4.1 ± 0.74 0.32 1 /2.0 mm 6.4 ± 0.13 2.4 ± 0.20 0.38 3 /2.0 mm 6.4 ± 0.41 0.7 ± 0.68 0.11 TABLE 2. Impact peak force Fpeak (mean 6 standard deviation, n 5 5) in air and in blood mimicking ﬂuid (BMF) for the six conditions, and ratio of the impact peak force in BMF and air (Fpeak BMF/Fpeak air). Strike distance/spring compression distance Fpeak air (N) Fpeak BMF (N) Fpeak BMF/Fpeak air (-) 1 /3.5 mm 18.9 ± 0.55 9.2 ± 0.08 0.49 3 /3.5 mm 19.2 ± 0.65 7.9 ± 0.75 0.41 1 /2.8 mm 13.0 ± 0.20 5.8 ± 0.29 0.45 3 /2.8 mm 12.9 ± 0.58 4.1 ± 0.74 0.32 1 /2.0 mm 6.4 ± 0.13 2.4 ± 0.20 0.38 3 /2.0 mm 6.4 ± 0.41 0.7 ± 0.68 0.11 force Fpeak (mean 6 standard deviation, n 5 5) in air and in blood mimicking ﬂuid (BMF) for the six onditions, and ratio of the impact peak force in BMF and air (Fpeak BMF/Fpeak air). TABLE 2. Impact peak force Fpeak (mean 6 standard deviation, n 5 5) in air and in blood mimicking conditions, and ratio of the impact peak force in BMF and air (Fpeak BMF/Fpeak air SAKES et al. 156 TABLE 3. Number of strikes to puncture the proximal cap models, using indenters with different tip shapes. The tip shapes are presented in the top row. The CTO model type and test number (n 5 3) are given in the left two columns. TABLE 3. Number of strikes to puncture the proximal cap models, using indenters with different tip shapes. The tip shapes are presented in the top row. Energy Transfer from the Indenter to the CTO Model The CTO model type and test number (n 5 3) are given in the left two columns. Type of model # Stamp Spherical Wedge Pointed Spherical GWa Ringed Ductile model I No puncture No puncture No puncture No puncture No puncture No puncture II No puncture No puncture No puncture No puncture No puncture No puncture III No puncture No puncture No puncture No puncture No puncture No puncture Reference model I 9 4 6 No puncture No puncture 5 II No puncture 5 7 7 9 10 III 6 No puncture 5 No puncture 10 9 Brittle model I 2 3 3 8 3 2 II 3 3 3 8 2 2 III 5 2 3 No puncture 3 2 The maximum number of strikes was set to 10, and # indicates the test number (I, II, or III). aSpherical indenter with guidewire passage. The maximum number of strikes was set to 10, and # indicates the test number (I, II, or III). aSpherical indenter with guidewire passage. The maximum number of strikes was set to 10, and # indicates the test number (I, II, or III). aSpherical indenter with guidewire passage. FIGURE 14. Typical proximal cap models after striking the indenter multiple times onto them. The illustrated proximal cap models were struck with the spherical indenter tip shape, on (A) the ductile model (10 strikes on all of the models), (B) the reference model (from left to right: 4, 5, and 10 strikes), and (C) the brittle model (from left to right: 3, 3, and 2 strikes). FIGURE 14. Typical proximal cap models after striking the indenter multiple times onto them. The illustrated proximal cap models were struck with the spherical indenter tip shape, on (A) the ductile model (10 strikes on all of the models), (B) the reference model (from left to right: 4, 5, and 10 strikes), and (C) the brittle model (from left to right: 3, 3, and 2 strikes). FIGURE 14. Typical proximal cap models after striking the indenter multiple times onto them. The illustrated proximal cap models were struck with the spherical indenter tip shape, on (A) the ductile model (10 strikes on all of the models), (B) the reference model (from left to right: 4, 5, and 10 strikes), and (C) the brittle model (from left to right: 3, 3, and 2 strikes). TABLE 4. Indenter Momentum Indenter Momentum The theoretical maximum velocity of the indenter (frictional eﬀects neglected) was calculated as 6.6 m/s at the moment the indenter protrudes from the pro- totype, using Eqs. (1–4): F ¼ K x; ð1Þ a ¼ F m ; ð2Þ F ¼ K x; ð1Þ a ¼ F m ; ð2Þ v ¼ v0 þ a t ; and ð3Þ s ¼ s0 þ v t; ð4Þ ð1Þ Energy Transfer from the Indenter to the CTO Model Number of punctures (all tip shapes combined) for the three CTO models. Type of model Number of punctures Success rate Mean number of strikes (range) Ductile model 0/18 0% – Reference model 13/18 72% 7.3 (4–8) Brittle model 17/18 94% 3.4 (2–8) TABLE 4. Number of punctures (all tip shapes combined) for the three CTO models. ment of the CTO model ranged between 1 and 2.7 mm, with the lowest value measured for the highest pene- tration velocity. were observed in the bouncing velocities for the dif- ferent proximal cap models. A slightly higher bounce velocity of 0.8 m/s was observed when the brittle proximal cap model was punctured (Table 5). The maximum displacement of the proximal cap models in case no puncture occurred was between 0.8 and 0.9 mm (Fig. 15; Table 5). A larger displacement was observed for the punctured brittle model in com- parison to the models in which no puncture was achieved, with a maximum of 1.4 mm (Table 5). Mechanical Performance Based on this data, the energy loss of the indenter is estimated to be between 2.34 and 2.11 J (~95–98% energy loss, see Table 5). Most of this energy was ab- sorbed by the proximal cap of the CTO model (about 95–99%), whereas the remaining 1–5% was absorbed by the gelatin. Summary of Main Findings Mechanical Performance Control Experiment ð2Þ The mean force to penetrate through the CTO model was 0.24 ± 0.09 N for the ductile model and 2.7 ± 1.1 N for the reference model. No signiﬁcant diﬀerences were found in penetration force for the diﬀerent insertion velocities for each CTO model as determined by two separate ANOVAs. The displace- v ¼ v0 þ a t ; and ð3Þ ð4Þ s ¼ s0 þ v t; Endovascular Crossing of Chronic Total Occlusions using an Impulse 157 FIGURE 15. A visualization of the analysis of HSV images to derive the maximum CTO displacement of a non-punctured brittle model using the spherical indenter. FIGURE 15. A visualization of the analysis of HSV images to derive the maximum CTO displacement of a non-punctured brittle model using the spherical indenter. TABLE 5. Energy transfer from the indenter to the CTO model. Type of model Initial velocity indenter (m/s) Initial momentum (mkgm/s) Initial kinetic energy (mJ) Bounce velocity indenter (m/s) Bounce momentum (mkgm/s) Bounce kinetic energy (mJ) Loss kinetic energy indenter (mJ) CTO displace- ment (mm) Energy absorption gelatin (mJ) Energy absorption CTO (mJ) Ductile model (no puncture) 3.3 1.29 2.16 0.5 0.20 0.05 2.11 (97.74%) 0.8 0.020 (1%) 2.09 (99.0% Reference model (no puncture) 3.5 1.37 2.39 0.5 0.20 0.05 2.34 (97.96%) 0.8 0.020 (0.9%) 2.32 (99.1% Brittle model (no puncture) 3.3 1.29 2.16 0.5 0.20 0.05 2.11 (97.74%) 0.9 0.025 (1.2%) 2.09 (98.8% Brittle model (punctured) 3.5 1.37 2.39 0.8 0.31 0.12 2.26 (94.78%) 1.4 0.060 (5.4%) 2.20 (94.6% FIGURE 15. A visualization of the analysis of HSV images to derive the maximum CTO displacement of a non-punctured brittle model using the spherical indenter. with a = acceleration [m/s2] of the spring and thus the indenter, F = actuation force of the spring [N] of 4.55 N, K = spring constant [N/mm] of 1.3 N/mm, m = mass of the indenter [grams] of 0.39 grams, x = maximum spring compression distance [mm] of 3.5 mm, s = distance travelled [mm], s0 = distance travelled at time t0 [mm], t = time step [s], v = velocity of the indenter [m/s] at t1 = t0 + t, and v0 = velocity of the indenter [m/s] at time t0. TABLE 5. Energy transfer from the indenter to the CTO model. Control Experiment Nonetheless, it can be suggested that the pointed indenter tip shape may be the least appropriate for fracturing CTOs, as these are usually heavily calciﬁed and thus brittle. model in Thind et al.,7 which could have been caused by the larger diameter of the rigid rod we used (Ø0.4 mm vs. Ø0.36 mm), the thickness of the proxi- mal cap (1 mm), and the consistency of the model. Nonetheless, the 0.2–2.7 N penetration force found in the artiﬁcial CTO model is comparable to the values found in the animal model of Thind et al.7 and Roy et al.27 The maximum tip load (i.e., the load in grams that a guidewire can withstand before buckling) of dedicated coronary CTO guidewires ranges between 0.8 and 26.7 grams, which is equal to 0.008 and 0.26 N.5 If we compare the maximum tip load of dedicated guide- wires (i.e., ~ 0.26 N),5 with the required penetration force of the reference CTO model (2.7 N), it becomes clear that buckling is likely when using dedicated coronary CTO guidewires. It can, therefore, be argued that a dedicated CTO guidewire is unable to penetrate the reference model. During puncture some minor chips of the CTO proximal cap model were formed, which are poten- tially harmful, as they can lead to stroke in the smaller blood vessels of the brain, for example. In order to minimize this risk, the device should only be used to penetrate the proximal cap of the CTO; crossing of the CTO body should be executed using a dedicated guidewire. Furthermore, it is recommended that the device is used in combination with a proximal emboli ﬁlter or active aspiration. In the prototype experiments, maximum mean im- pact peak forces of 19.2 N and 9.2 N were measured in air and BMF, respectively, which in theory is suﬃcient to puncture the CTO. Moreover, even though impact peak forces were 50 to 90% lower in the BMF than in air, they were still above the values in Thind et al.7 It must be noted, however, that since no experiments have been performed on real CTOs, there is still uncertainty about the required impact peak force to puncture real in vivo CTOs. Control Experiment From the indenter mass and theoretical maximum velocity of the indenter, a theoretical maximum indenter momentum of 2.34 mNs can be deducted. The actual (measured) indenter momentum of 1.40 mNs (with an associated velocity of 3.6 m/s) is approxi- mately 40% lower than the theoretical value. It can thus be concluded that friction between the indenter, compression spring, and outer shaft plays an impor- tant role within the mechanism. Potentially, a higher eﬃciency may be realizable by reducing frictional ef- fects. This could, for example, be achieved by a dif- ferent material choice for the moving parts (the indenter and spring) and the surrounding catheter tube, or by lubricating the moving parts. Impact Peak Force In Thind et al.,7 the puncture force of 2, 6, 12, and 15 weeks old rabbit femoral CTOs was measured using a probe similar in proﬁle to a Ø0.36 mm guidewire. These authors found that the puncture force is signiﬁcantly lower in CTOs younger than 6 weeks of age (0.61 N and 0.78 N in CTOs of 2 and 6 weeks old, respectively) than in those older than 12 weeks (1.21 N and 1.52 N in CTOs of 12 and 15 weeks old, respectively). These values are in line with Roy et al.,27 in which puncture forces of excised peripheral CTOs between 0.3 and 1.7 N for soft (containing loose ﬁbrous tissue, fat, thrombus, or micro-blood vessels) and hard (containing collagen and speckled calcium) CTOs, respectively, were mea- sured. In the control experiment, it was also found that the less calciﬁed CTO models required signiﬁcantly lower penetration forces. For the brittle model slightly higher penetration forces were found compared to the animal SAKES et al. 158 made regarding their efﬁciency or effectiveness on fracturing models; all performed similarly well in fracturing the brittle material models and were able to puncture reference models, with more ductile charac- teristics. Furthermore, the ‘‘open’’ shapes (hollow spherical and ringed) performed comparably to the ‘‘closed’’ shapes (stamp, spherical, and wedge), while being advantageous for future clinical prototypes in which a guidewire may need to be guided through the crossing instrument. Due to the uncertainty of the material characteristics of real CTOs, it is difﬁcult to provide a recommendation about which tip shape is the most suitable for CTO crossing. Control Experiment Energy Transfer from the Indenter to the CTO Model The observed maximum displacements of the CTO models (1.4 mm) were lower than the displacements of over 4 mm reported for guidewires statically pu- shed against a CTO until puncture is achieved (derived from Thind et al.7). Furthermore, in the control experiment, the average displacement of proximal cap models induced by an almost static (0.1 mm/s) guidewire force was 2.1 mm (n = 4). It may, therefore, be argued that the impulse method shows merit to minimize stretch of the blood vessel wall during PCI of CTOs. However, other negative effects due to the impulse should be investigated, such as the risk of blood vessel wall damage as a result of a direct hit or damage due to device failure, in which, for example, the indenter or spring disconnect from the device. Puncture Performance Number of Punctures & Number of Strikes The highest puncture performance and eﬃciency was observed for the brittle CTO models. The eﬀectiveness of the impact method to puncture the more ductile CTO models was limited, which can be explained by the fact that ductile materials absorb energy from the impact loads by deformation (elastic and plastic). This ﬁnding is consistent with the fact that the impulse method is more widely used to fracture hard and brittle materials, rather than tough and/or elastic materials. Puncture performance of the prototype was assessed using the six indenter tip shapes on the three proximal cap models. For the ductile models, it was expected to see improved puncture performance in terms of num- ber of punctures and number of strikes for puncture with the sharp indenter tip shapes. Even though no puncture was achieved, the sharpest tools (pointed and wedge indenter) did leave a clearly visible imprint. For the brittle models, the ringed tip shape was the most effective, whereas the pointed indenter tip shape showed the least effective results. For the non-pointed indenter tip shapes tested, no distinction could be In our experiments, the amount of kinetic energy lost during the indenter strike was large (kinetic energy before and after collision were determined at 2.2 mJ and 0.1 mJ, respectively), indicating that the collisions with the CTO models were highly inelastic, albeit not purely inelastic, since up to 5.4% of the energy dissi- pated in the environment. To limit the displacement of the CTO and associated energy dissipation, the veloc- ity of the indenter should be increased, allowing the inertia and damping of the CTO and its environment to provide a higher counterforce to the applied impulse Endovascular Crossing of Chronic Total Occlusions using an Impulse 159 that of the loss of the indenter caused by failure of the connection with the indenter stop. By adding a second stop at the distal end of the catheter, the risk of indenter loss may be minimized. Finally, to allow for hand-held operation of the prototype, the enveloping box needs to be redesigned. The handle can, in theory, be 3D-printed in any arbitrary shape to meet the needs and preferences of the operator, as long as it contains an insert for the loading, locking, and trigger mechanism. than in the current design. Puncture Performance This could be achieved by diminishing the eﬀect of dry and viscous friction between the moving parts, by increasing the actuation force, and by minimizing the indenter mass. Further- more, in a future experiment, the energy dissipation to the environment should be estimated by measuring the energy absorbed by the CTO proximal cap model on a ﬁxed surface. The higher bounce velocity of the indenter and CTO model displacement observed in the cases of successful CTO model puncture compared to the cases when no puncture was achieved can be ex- plained by two possible phenomena. First, it is likely that in the cases of successful puncture, tilting of the two parts of the proximal cap model after fracturing has led to a more concentrated distribution of the impact force on the gelatin, leading to a larger dis- placement as compared to the cases of no puncture. Second, in the case of puncture less energy is likely to be lost in the proximal cap model, compared to a non-puncture case. The prototype has proven eﬀective in puncturing the CTO models. However, it is recommended that the indenter velocity is increased to further decrease the CTO displacement and increase the puncture eﬀectiveness. For this purpose, the eﬀect of the BMF viscosity should be minimized by, for example, add- ing a sealing that prevents blood from entering the prototype. Furthermore, stiﬀer spring designs should be investigated, and the eﬀect of friction within the device tip should be minimized by, for example, using low friction coeﬃcient material combinations. Finally, diﬀerent types of output characteristics need to be investigated, including diﬀerent vibrating mo- tions, to determine the most optimal output charac- teristic for achieving puncture of the diﬀerent CTO models. Design Recommendations The most important redesign step of the prototype concerns the transformation of the rigid design into a ﬂexible clinical instrument. In order to allow for navigating towards the lesion site, a ﬂexible shaft will need to be placed between the tip section and the handle. The ﬂexible shaft should be axially incom- pressible or force neutral (such as a Bowden cable) to prevent axial compression and movement during activation of the tip section. Furthermore, the inner locking core needs to be replaced by a ﬂexible cable that runs from the distal tip of the device, through the ﬂexible shaft, towards the handle. To actuate the gripper, a ﬂexible axially stiﬀtube or cables should also run through the shaft of the device. Additionally, the rigid tip parts such as the outer shaft, gripper, position block, and indenter need to be redesigned to allow for bending motion. For this purpose, the outer shaft can be replaced by a ﬂexible shaft similar to a catheter, a ﬂexible cable can replace the connection between the indenter tip and stop, and the position block and the gripper can contain joints to allow for bending motions while restricting axial compression or extension. Moreover, compression of the spring during actuation may straighten the previously bend catheter shaft, potentially resulting in undesired pressure on, and thus damage of, the arterial wall. The addition of steering cables can possibly prevent or counteract this eﬀect by retaining the bend shape of the catheter during loading and unloading. An- other important issue that needs to be addressed is To allow for atraumatic navigating through the vasculature and account for diameter diﬀerences between patients, the prototype should be further miniaturized. Current clinically available dedicated CTO devices have a diameter of in between Ø0.43 and Ø1.5 mm, so miniaturization to at least the Ø1.5 mm is preferred. Miniaturization of the prototype is relatively easy to achieve due to the small number of parts (14) and low complexity of the prototype. The main chal- lenge lies in the miniaturization of the gripper and its counterpart, the indenter stop. CONCLUSION The following steps were followed to prepare the environment models: The low success rates of endovascular revasculariza- tion of CTOs can be mainly attributed to the inability to cross the proximal caps of heavily calciﬁed CTOs. In an eﬀort to pursue improved crossing ability, we investi- gated a new crossing method in which an impulse is applied onto the CTO. Using an impulse to penetrate the proximal cap of heavily calciﬁed CTOs is proven to be advantageous as it increases the buckling resistance of the tool, minimizes movement of the CTO, and can potentially decrease the penetration load. A proof-of- principle prototype was developed that uses a spring- loaded indenter with interchangeable tip shapes and a compliant reload mechanism. The prototype was eval- uated in terms of its mechanical performance and puncture eﬀectiveness on CTO models made of gelatin and calcium. From this experiment it became clear that the proposed impulse prototype outperforms currently available dedicated CTO guidewires when it came to puncture performance. A maximum mean indenter velocity of 3.6 m/s, translating to an indenter momen- tum of 1.33 mNs, and a maximum mean impact peak force of 9.2 N were measured in BMF. This impact peak force is well over what can be delivered by dedicated CTO guidewires, as well as the 1.52 N puncture force previously measured in rabbit femoral CTOs of 15 weeks of age, and should thus be suﬃcient to pene- trate real CTOs.7 Furthermore, the displacement of the CTO model was signiﬁcantly less with the proposed prototype than in a control experiment with a rigid rod and in the animal CTO model of Thind et al.7 In contrast to current CTO devices and guidewires, the prototype was most effective and efﬁcient on the brittle CTO models using the spherical, hollow spherical with guide- wire passage, and ringed tip shape. Future developments will be focused on developing a smaller (Ø1 mm), faster, and ﬂexible clinical prototype that allows for penetrat- ing both soft and hard CTOs with minimal tissue deformation and energy dissipation. This prototype will be guided over a guidewire through the vasculature to- wards the occlusion in one of the coronaries where it can aid in penetrating the proximal caps of the most heavily calciﬁed CTOs. 1. Weigh a speciﬁc amount (depending on the quantity you need) of gelatin (Sheet gelatin, Dr. Oetker, Bielefeld, Germany). 2. CONCLUSION Put the gelatin sheets in a bath of cold fresh water (~15 C), and leave to soak for 5 min. 3. Take the gelatin sheets from the bath and wring them meticulously before putting them in a plastic container. 4. Weigh the new (soaked) mass of the gelatin sheets, and add an amount of cold clear water to create the 25 wt% of gelatin. 5. Put the gelatin-ﬁlled container in a bath of warm water (~50 C) and stir gently until the con- tent is completely liquid. 6. Take the container out of the bath and leave it to rest for approximately 15 min, during which all bubbles will ﬂoat to the surface and can be removed from the mixture. 7. Pour the gelatin mixture in a small container that is placed in the mold illustrated in Fig. 16. 8. Close the mold with the lid to create a notch within the gelatin volume. 9. Leave the gelatin-ﬁlled containers for approxi- mately 1 h in a refrigerator (~2 C), before gently removing the lid of the mold from the gelatin (see Fig. 17). 10. Close the lid of container and leave the gelatin mixture to harden in the refrigerator for an- other 24 h. FIGURE 16. The mold for creating the environment models. The mold consists of laser-cut disks in between which 10 circular containers can be placed. The lid of the mold, in which several rods are placed, creates the notches. FIGURE 16. The mold for creating the environment models. The mold consists of laser-cut disks in between which 10 circular containers can be placed. The lid of the mold, in which several rods are placed, creates the notches. Limitations of this Study The development of an accurate and reproducible coronary CTO model is a complex undertaking and, up to today, has not produced a representative replica of a CTO. This is mainly because simulating calciﬁ- cation and the inﬂammatory component is diﬃcult. In order to evaluate the device performance, it was chosen to use an artiﬁcial CTO model. This CTO model allowed us to test the device under controlled condi- tions, which is not possible using an animal CTO model. However, image analysis of the CTO models to further characterize them is needed. This will deter- mine how well the artiﬁcial CTO models mimic real CTOs. In future research it is recommended that the device is tested in an animal CTO model, such as the 160 SAKES et al. surrounding cardiac tissue) and proximal cap models is described. one described by Thind et al.,7 to further determine its effectiveness. ACKNOWLEDGMENTS We would like to thank Paul Henselmans and David de Jager for their contribution in the develop- ment and manufacturing of the prototype, as well as the department of Animal Sciences of Wageningen University, in particular Johan van Leeuwen and Remco Pieters, for the opportunity and assistance to capture HSVs of the prototype. 2. Weigh a speciﬁc amount (depending on the quantity you need) of calcium sulfate (SHERAALPIN Hartgips hellblau, SHERA Werkstoff-Technologie GmbH & Co. KG). 3. Put the powder in a small plastic cup and add a correct amount of gelatin mixture into the cup. – For the ductile proximal cap model: 50 wt% calcium sulfate powder and 50 wt% gelatin mixture; Proximal Cap Models The following steps were followed to create the proximal cap models: 1. Prepare the gelatin mixture as described in step 1–6 from the previous procedure (surrounding tissue model). APPENDIX: CTO MODEL PREPARATION In this appendix, the procedure that was followed to prepare the environment models (representing the Endovascular Crossing of Chronic Total Occlusions using an Impulse Endovascular Crossing of Chronic Total Occlusions using an Impulse 161 FIGURE 18. The proximal cap model. Left: Mold for the proximal cap models, to create the desired 1 mm material thickness, ﬁlled with the gelatin/plaster cast mixture Right: the complete molded slice of material (left), thin-walled tube to press the circles (top right), and ﬁnal proximal cap models (lower right). FIGURE 17. Environment models with notches in the center. FIGURE 18. The proximal cap model. Left: Mold for the proximal cap models, to create the desired 1 mm material thickness, ﬁlled with the gelatin/plaster cast mixture Right: the complete molded slice of material (left), thin-walled tube to press the circles (top right), and ﬁnal proximal cap models (lower right). 5. Pour the mixture onto the cap-model mold, and close with force to get a 1 mm thick slice of proximal cap models (see Fig. 18). 6. Leave the ﬁlled cap-model mold for 20 min in a refrigerator (~2 C) to stiffen. 7. Remove the cap-model mold from the refrig- erator and open it to take out the formed slice of (at this moment still ﬂexible) material. FIGURE 17. Environment models with notches in the center. 8. Press the small circles from the material with the help of a thin-walled tube, as shown in Fig. 18. Note: To ensure constant tissue stiﬀness, the tissue model is to be used directly from the refrigerator (<1 h). 9. Pack the models into cling ﬁlm, and leave them for another 24 h in the refrigerator. Note: To ensure constant material stiﬀness, the model is to be used directly from the refrigerator (< 1 h). Note: To ensure constant material stiﬀness, the model is to be used directly from the refrigerator (< 1 h). Proximal Cap Models REFERENCES 1Stone, G. W., N. J. Reifart, I. Moussa, A. Hoye, D. A. Cox, A. Colombo, et al. Percutaneous recanalization of chronically occluded coronary arteries. A consensus doc- ument—part I. Circulation 112(15):2364–2372, 2005. doi: 10.1161/CIRCULATIONAHA.104.481283. 15Sandgren, T., B. Sonesson, A˚ . R. Ahlgren, and T. La¨ nne. The diameter of the common femoral artery in healthy human: inﬂuence of sex, age, and body size. J. Vasc. Surg. 29(3):503–510, 1999. 16 2Sakes, A., E. Regar, J. Dankelman, and P. Breedveld. Crossing total occlusions: navigating towards recanaliza- tion. Cardiovasc. Eng. Technol. 7(2):103–117, 2016. doi: 10.1007/s13239-016-0255-0. 16Yoo, B. S., J. Yoon, J. Y. Ko, J. Y. Kim, S. H. Lee, S. O. Hwang, et al. Anatomical consideration of the radial artery for transradial coronary procedures: arterial diameter, branching anomaly and vessel tortuosity. Int. J. Cardiol. 101(3):421–427, 2005. doi:10.1016/j.ijcard.2004.03.061. 17 3Stone, G. W., N. J. Reifart, I. Moussa, A. Hoye, D. A. Cox, A. Colombo, et al. Percutaneous recanalization of chronically occluded coronary arteries—a consensus doc- ument—part II. Circulation 112(16):2530–2537, 2005. doi: 10.1161/CIRCULATIONAHA.105.583716. 17Yalonetsky, S., A. B. Osherov, B. H. Strauss, A. R. Ga- lassi, S. Tomasello, and H. Khamis. Chronic Total Occlu- sions. A Guide to Recanalization (2nd ed.). Hoboken: Wiley-Blackwell, 2013. 4Galassi, A., S. Tomasello, N. Reifart, G. S. Werner, G. Sianos, H. Bonnier, et al. In-hospital outcomes of percu- taneous coronary intervention in patients with chronic total occlusion: insights from the ERCTO (European Registry of Chronic Total Occlusion) registry. EuroInter- vention 7(4):472–479, 2011. doi:10.4244/EIJV7I4A77. 5 y , 18Ramnarine, K. V., D. K. Nassiri, P. R. Hoskins, and J. Lubbers. Validation of a new blood-mimicking ﬂuid for use in doppler ﬂow test objects. Ultrasound Med. Biol. 24(3):451–459, 1998. doi:10.1016/S0301-5629(97)00277-9. 19 ( ) , / ( ) 19Samavat, H., and J. Evans. An ideal blood mimicking ﬂuid for doppler ultrasound phantoms. J. Med. Phys. 31(4):275– 278, 2006. doi:10.4103/0971-6203.29198. 20 5Sianos, G., G. S. Werner, A. Galassi, M. I. Papafaklis, J. Escaned, D. Hildick-Smith, et al. Recanalisation of chronic total coronary occlusions: 2012 consensus document from the EuroCTO Club. EuroIntervention 8(1):139–145, 2012. doi:10.4244/EIJV8I1A21. / 20Yousif, M. Y., D. W. Holdsworth, and T. L. Poepping. A blood-mimicking ﬂuid for particle image velocimetry with silicone vascular models. Exp. Fluids 50(3):769–774, 2011. doi:10.4103/0971-6203.29198. 21 6Kinoshita, I., O. Katoh, J. Nariyama, S. Otsuji, H. Ta- teyama, T. Kobayashi, et al. STATEMENT OF ANIMAL STUDIES 9Lock, M. A study of buckling and snapping under dynamic load. Technical Report No. TR-058 (340-30)-3, El Se- gundo: Aerospace Corporation, 1967. 10 No animal studies were carried out by the authors for this article. 10Boutouyrie, P., C. Bussy, P. Lacolley, X. Girerd, B. La- loux, and S. Laurent. Association between local pulse pressure, mean blood pressure, and large-artery remodel- ing. Circulation 100(13):1387–1393, 1999. 11 OPEN ACCESS g ( ) 11Dodge, J., B. G. Brown, E. L. Bolson, and H. T. Dodge. Lumen diameter of normal human coronary arteries. Inﬂuence of age, sex, anatomic variation, and left ventric- ular hypertrophy or dilation. Circulation 86(1):232–246, 1992. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. 12Kahraman, H., M. Ozaydin, E. Varol, S. M. Aslan, A. Dogan, A. Altinbas, et al. The diameters of the aorta and its major branches in patients with isolated coronary artery ectasia. Tex. Heart I. J. 33(4):463–468, 2006. 13 13Kotowycz, M. A., and V. Dzˇ avı´k. Radial artery patency after transradial catheterization. Circulation 5(1):127–133, 2012. doi:10.1161/CIRCINTERVENTIONS.111.965871. 14 / 14Saito, S., H. Ikei, G. Hosokawa, and S. Tanaka. Inﬂuence of the ratio between radial artery inner diameter and sheath outer diameter on radial artery ﬂow after transradial coronary intervention. Catheter. Cardiovasc. Interv. 46(2):173–178, 1999. doi:10.1002/(SICI)1522726X(199902) 46:2<173::AID-CCD12>3.0.CO;2-4. ˚ STATEMENT OF HUMAN STUDIES STATEMENT OF HUMAN STUDIES No human studies were carried out by the authors for this article. 8Stone, G. W., A. Colombo, P. S. Teirstein, J. W. Moses, M. B. Leon, N. J. Reifart, et al. Percutaneous recanaliza- tion of chronically occluded coronary arteries: Procedural techniques, devices, and results. Catheter. Cardiovasc. In- terv. 66(2):217–236, 2005. doi:10.1002/ccd.20489. 9 CONFLICT OF INTEREST This work is part of the research program Image Guided Interventional Treatment (IGIT) of Coronary Chronic Total Occlusions within the research program interactive Multi-Interventional Tools (iMIT), which is funded by the Dutch Technology Foundation (Toe- gepaste en Technische Wetenschappen, TTW; Re- search Grant Number 12710), which is part of the Netherlands Organisation for Scientiﬁc Research (NWO). STW did not have an active role in the design nor this paper. – For the reference proximal cap model: 75 wt% calcium sulfate powder and 25 wt% gelatin mixture; – For the brittle proximal cap model: 77 wt% calcium sulfate powder and 23 wt% clear water. 4. Stir the mixture gently, while holding the cup in a bath of warm water (~50 C), until a homo- geneous mixture has formed. 162 SAKES et al. EuroIntervention 6(8):997–1002, 2011. doi:10.4244/EIJV6I8 A172. EuroIntervention 6(8):997–1002, 2011. doi:10.4244/EIJV6I8 A172. 8 REFERENCES Coronary angioplasty of chronic total occlusions with bridging collateral vessels: immediate and follow-up outcome from a large single- center experience. J. Am. Col. Cardiol. 26(2):409–415, 1995. doi:10.1016/0735-1097(95)80015-9. / 21Cheng, N. S. Formula for the viscosity of a glycerol-water mixture. Ind. Eng. Chem. Res. 47(9):3285–3288, 2008. doi: 10.1021/ie071349z. 22 22Suzuki, Y., A. Oyane, F. Ikeno, J. K. Lyons, and A. C. Yeung. Development of animal model for calciﬁed chronic total occlusion. Catheter. Cardiovasc. Interv. 74(3):468–475, 2009. doi:10.1002/ccd.22024. 23 / ( ) 7Thind, A., B. Strauss, A. Teitelbaum, R. Karshaﬁan, M. Ladouceur, C. Whyne, et al. A novel method for the measurement of proximal ﬁbrous cap puncture force in chronic total occlusions: the effect of increasing age. / 23Martin, R., and P. Brown. Mechanical properties of hydroxyapatite formed at physiological temperature. J. Mat. Sci: Mat. Med. 6(3):138–143, 1995. doi:10.1007/ BF00120289. 24Cho, J. R., Y. J. Kim, C. M. Ahn, J. Y. Moon, J. S. Kim, H. S. Kim, et al. Quantiﬁcation of regional calcium burden in chronic total occlusion by 64-slice multi-detector com- puted tomography and procedural outcomes of percuta- neous coronary intervention. Int. J. Cardiol. 145(1):9–14, 2010. doi:10.1016/j.ijcard.2009.05.006. 25Barrett, S., M. Sutcliffe, S. Howarth, Z. Y. Li, and J. Gillard. Experimental measurement of the mechanical 24Cho, J. R., Y. J. Kim, C. M. Ahn, J. Y. Moon, J. S. Kim, H. S. Kim, et al. Quantiﬁcation of regional calcium burden in chronic total occlusion by 64-slice multi-detector com- puted tomography and procedural outcomes of percuta- neous coronary intervention. Int. J. Cardiol. 145(1):9–14, 2010. doi:10.1016/j.ijcard.2009.05.006. /j j 25Barrett, S., M. Sutcliffe, S. Howarth, Z. Y. Li, and J. Gillard. Experimental measurement of the mechanical Endovascular Crossing of Chronic Total Occlusions using an Impulse Endovascular Crossing of Chronic Total Occlusions using an Impulse 163 properties of carotid atherothrombotic plaque ﬁbrous cap. J. Biomech. 42(11):1650–1655, 2009. doi:10.1016/j.jbiomech. 2009.04.025. Mat. Sci: Mat. Med. 6(3):138–143, 1995. doi:10.1007/ BF00120289. properties of carotid atherothrombotic plaque ﬁbrous cap. J. Biomech. 42(11):1650–1655, 2009. doi:10.1016/j.jbiomech. 2009.04.025. 24Cho, J. R., Y. J. Kim, C. M. Ahn, J. Y. Moon, J. S. Kim, H. S. Kim, et al. Quantiﬁcation of regional calcium burden in chronic total occlusion by 64-slice multi-detector com- puted tomography and procedural outcomes of percuta- neous coronary intervention. Int. J. Cardiol. 145(1):9–14, 2010. doi:10.1016/j.ijcard.2009.05.006. 24Cho, J. R., Y. J. Kim, C. M. Ahn, J. Y. Moon, J. S. Kim, H. S. Kim, et al. Quantiﬁcation of regional calcium burden in chronic total occlusion by 64-slice multi-detector com- puted tomography and procedural outcomes of percuta- neous coronary intervention. Int. J. Cardiol. 145(1):9–14, 2010. doi:10.1016/j.ijcard.2009.05.006. 26Houwink, R., and H. K. De Decker. Elasticity, plasticity and structure of matter. Cambridge: Cambridge University Press, 1971. 27Roy, T., G. Liu, N. Shaikh, A. D. Dueck, and G. A. Wright. Puncturing plaques: relating MRI characteristics of periph- eral artery lesions to guidewire puncture forces. J. Endovasc. Ther. 24(1):35–46, 2016. doi:10.1177/1526602816671135. /j j 25Barrett, S., M. Sutcliffe, S. Howarth, Z. Y. Li, and J. Gillard. Experimental measurement of the mechanical
https://openalex.org/W2154692504	https://journals.library.brocku.ca/brocked/index.php/home/article/download/239/177	English	null	Songs in Their Heads: Music and Its Meaning in Children’s Lives (2<sup>nd</sup> ed.).	Brock education	2,011	cc-by	1,360	BOOK REVIEW BOOK REVIEW Title: Songs in Their Heads: Music and Its Meaning in Children’s Lives (2nd ed.). Author: Patricia Shehan Campbell Publisher: New York: Oxford University Press. Year of Publication 2010 Vanessa Mio (Ph.D. student in the Joint PhD in Educational Studies at University of Windsor. Master of Music, Bachelor of Music and Performance Diploma from Indiana University Jacobs School of Music. Violin Instructor at the Algoma Conservatory of Music.) vanessamio5@gmail.com BOOK REVIEW Title: Songs in Their Heads: Music and Its Meaning in Children’s Lives (2nd ed.). Author: Patricia Shehan Campbell Publisher: New York: Oxford University Press. Year of Publication 2010 BOOK REVIEW Title: Songs in Their Heads: Music and Its Meaning in Children’s Lives (2nd ed.). Author: Patricia Shehan Campbell Publisher: New York: Oxford University Press. Year of Publication 2010 Reviewed by: Yet here was a preschool child who had Brock Education, 21(1), 98-100 98 demonstrated the concept with ease and nonchalance, and with considerable feeling” (p. 35). Campbell also describes a young boy tapping his milk carton on the table in a “syncopated pattern” (p. 37), and a fifth grader who struck a metal dustpan against the garbage can in a notated repeated pattern (p. 41). The children described here appear to be inadvertently creating sounds in their activities which Campbell over-zealously attributes to an understanding of rhythmic behaviour. What makes her argument questionable, is the distinct difference between rhythm with intent and simple sounds. Part 2, “On Music: Conversations with Children”, presents twenty candid interviews, followed directly by Campbell’s reflections and commentary. Through the interviews, Campbell determines common threads amongst the diverse backgrounds, experiences, and circumstances of the children. She also provides an ethnomusicological perspective of the musical genres children hear at home, engage in with friends, the way they view music in their schools, and how they view themselves as singers or instrumentalists. The interview format in this portion of the study is effective. As Campbell captures the uninhibited perspectives of the children, it allows the reader to form interpretations before being exposed to the opinions of the author. Through the commentary, the reader can then naturally engage in critical analysis as their opinions are either reinforced or challenged. Part 3, “For Children: Prospects for Their Music Education”, serves as an additional forum for commentary and reflections, intended to “examine the collage that they form together” (p. 215). Campbell explains that recent influences of technology and mass media on children serve as the inspiration behind publishing the second edition. Campbell is undoubtedly passionate about this issue, but she presents a deterministic claim in stating that “today’s children are born digital” (p. 219). Although technology is undeniably influential in Western culture, Campbell could have stated that one cannot be born digital but rather one adjusts to a digital age. Is it one’s interaction with technology that is influential, or the fact that it exists? Through her statement, Campbell seems to be homogenizing all children, when in reality, access to certain technologies is still part of a privileged class. In addition to media, Campbell also addresses the place and weight of music education in the curriculum and offers interdisciplinary (Klein, 2004) insights for music educators to consider. Reviewed by: Vanessa Mio (Ph.D. student in the Joint PhD in Educational Studies at University of Windsor. Master of Music, Bachelor of Music and Performance Diploma from Indiana University Jacobs School of Music. Violin Instructor at the Algoma Conservatory of Music.) vanessamio5@gmail.com Looking back on my childhood, it is difficult to remember a time when I was not in formal music training. As a music educator, I often ponder the circumstances that sparked my desire to play musical instruments. Did I always have melodies running through my head? Did I naturally engage in musical activities or sing songs to myself? In her book, Songs in Their Heads: Music and Its Meaning in Children’s Lives, Patricia Campbell addresses these questions while exploring how children of diverse ages, classes, and cultures use music in their daily lives, how ethnomusicology, the study of music through cultural influences (Aubert, 2007), shape their musical experiences, and how recent advancements in technology and the influence of popular culture affect their musical choices. Campbell (2010) introduces her inspiration for the studies through a story of a young boy in a schoolyard saying that every morning he wakes up with songs in his head. This seemingly simple statement triggered numerous questions for Campbell (2010) as she then wondered if these were "songs of the musically gifted? Or do all children carry within them musical gifts- songs, musical ‘urges,’ and thoughts about music- at the edge of their consciousness?” (p. 3). Part 1, “In Music: Children at Musical Play”, documents the natural musical engagements of children through descriptive ethnographic field-based observations and graphic musical notations. The musical examples embedded within the text support the observations, and allow the reader to interpret various transcriptions of rhythms, vocalizations, songs, and chants that Campbell audio/video taped. Campbell concludes that, through “musicking,” (Elliott, 1995) children “kinesthetically expel their energy … and make their own music to fit their movements and their imaginations” (Campbell, 2010, p. 95). Through substantial descriptions, Campbell supports her claims regarding intrinsic “musical doodlings” (Kartomi, 1991) in children. However, she ventures too far in her analysis of rhythmic behaviours. In one example Campbell (2010) describes a young girl who changed the meter unknowingly while singing a song. Campbell states that metric shifts in music “may be perceived by musicians who perform complex works …. Reviewed by: While Campbell believes that intellectual growth can occur when “songs, rhythms, and listening experiences [are] integrated into lessons on social studies, the language arts, mathematics, and the sciences” (p. 272), she does not address a standard for quality of music presented in the classroom. Successful integration occurs when all subject areas are equally cultivated, and although this method may foster learning, it does not necessarily foster musical development. As a result, the success of integration is dependent on the abilities of the teachers, and their willingness to maintain a high standard of pitch and rhythmic proficiency in their students. Although Campbell attests to the positive effects of integration in general, she neglects essential aspects of music education, which undermine her philosophies surrounding musical development in children. In this section, Campbell also addresses issues of musical choice and warns parents of the poor musical quality present in music specifically made for young children. She suggests consulting the Billboard charts to see what is popular, and advocates exposing students to all of the various genres. However, from a musician’s perspective it is apparent that many artists in popular culture utilize technologically altered voices and studio effects rather than organic talent. Therefore, suggesting that popular music translates as good quality is a severely overstated argument and should be seen as equivalent to Campbell’s opinion regarding children’s music and the adverse effects of poor quality. 99 Brock Education, 21(1), 98-100 Although Campbell's credible intentions in providing perspective, knowledge, and insight into the inherently musical lives of children are apparent, she expresses deterministic ideas that could benefit from further massaging. Extra support from external sources may have added more credibility. Additionally, a biography listing Campbell’s musical and educational background would have been advantageous to understand the context that contributed to the development of her philosophies In essence, although Campbell presents bold arguments, she is successful in capturing children’s intrinsic musical perspectives and behaviours. Brock Education, 21(1), 98-100 References Aubert, L. (2007). The music of the other: New challenges for ethnomusicology in a global age. Burlington, VT: Ashgate Publishing Company. Elliott, D. J. (1995). Music matters. New York: Oxford University Press. Elliott, D. J. (1995). Music matters. New York: Oxford University Press. Elliott, D. J. (1995). Music matters. New York: Oxford University Press. Kartomi, M. (1991). Musical improvisations by children at play. The World of Music, 33(3), 53-65. Kartomi, M. (1991). Musical improvisations by children at play. The World of Music, 33(3), 53-65. Klein, J. T. (2004). Interdisciplinarity and complexity: An evolving relationship. Emergence: Complexity & Organization, 6(1-2), 2-10. Klein, J. T. (2004). Interdisciplinarity and complexity: An evolving relationship. Emergence: Complexity & Organization, 6(1-2), 2-10. 100 Brock Education, 21(1), 98-100
https://openalex.org/W2901193065	https://www.frontiersin.org/articles/10.3389/fimmu.2018.02677/pdf	English	null	Bovine Lactoferrin Enhances TLR7-Mediated Responses in Plasmacytoid Dendritic Cells in Elderly Women: Results From a Nutritional Intervention Study With Bovine Lactoferrin, GOS and Vitamin D	Frontiers in immunology	2,018	cc-by	9,491	CLINICAL TRIAL published: 20 November 2018 doi: 10.3389/ﬁmmu.2018.02677 Bovine Lactoferrin Enhances TLR7-Mediated Responses in Plasmacytoid Dendritic Cells in Elderly Women: Results From a Nutritional Intervention Study With Bovine Lactoferrin, GOS and Vitamin D Marloes van Splunter 1, Olaf Perdijk 1, Henriëtte Fick-Brinkhof 2, Anouk L. Feitsma 3, Esther G. Floris-Vollenbroek 4, Ben Meijer 1, Sylvia Brugman 1, Huub F. J. Savelkoul 1, Els van Hoffen 4 and R. J. Joost van Neerven 1,3* Keywords: aging, TLR stimulation, pDCs, mDCs, bovine lactoferrin, GOS, vitamin D, inﬂammation Edited by: Edited by: Haruki Kitazawa, Tohoku University, Japan Reviewed by: Md. Aminul Islam, Universität Bonn, Germany Claudio Nicoletti, Università degli Studi di Firenze, Italy Correspondence: R. J. Joost van Neerven joost.vanneerven@wur.nl Reviewed by: Md. Aminul Islam, Universität Bonn, Germany Claudio Nicoletti, Università degli Studi di Firenze, Italy During aging the immune system is dysregulated. Especially plasmacytoid dendritic cells (pDCs) and myeloid DCs (mDCs) have reduced Toll like receptor (TLR)-mediated responses resulting in increased susceptibility to infections. Consumption of bovine lactoferrin (bLF) has been shown to reduce infections with viruses. Galacto- oligosacharides (GOS) and vitamin D are associated with reduced pro-inﬂammatory cytokine levels in serum, and increased TLR7/8 responses, respectively. A double-blind placebo-controlled nutritional intervention study in elderly women was performed, to investigate the potential of bLF, GOS, and vitamin D to restore TLR responsiveness of pDCs and mDCs and to reduce inﬂammatory markers in serum. The nutritional intervention group (n = 15) received bLF for 3 weeks, followed by 3 weeks of bLF + GOS, and subsequently 3 weeks of bLF + GOS + vitamin D. The placebo group (n = 15) received maltodextrin for 9 weeks. Every 3 weeks, blood was collected and TLR responses of pDCs and mDCs, and inﬂammation-related markers in serum were measured. After 3 weeks of bLF supplementation, increased TLR7/8 and TLR1/2 responses were observed in pDCs of the nutritional intervention group compared to the placebo group. When the effects of the entire nutritional intervention were investigated, increased TLR1/2 mediated responses in mDCs were observed, and in serum sVCAM tended to decrease. Finally, based on the RAND-36 questionnaire physical function tended to improve in the intervention group. Since especially TLR7-mediated responses in pDCs were enhanced after bLF supplementation compared to placebo, this suggests that bLF may contribute to antiviral responses mediated by pDC in elderly women. Clinical trial registry number: NCT03026244, clinicaltrials.gov: Correspondence: R. J. Joost van Neerven joost.vanneerven@wur.nl Specialty section: This article was submitted to Nutritional Immunology, a section of the journal Frontiers in Immunology Specialty section: This article was submitted to Nutritional Immunology, a section of the journal Frontiers in Immunology Received: 30 August 2018 Accepted: 30 October 2018 Published: 20 November 2018 Received: 30 August 2018 Accepted: 30 October 2018 Published: 20 November 2018 1 Cell Biology and Immunology, Wageningen University, Wageningen, Netherlands, 2 Human Nutrition, Wageningen University, Wageningen, Netherlands, 3 FrieslandCampina, Amersfoort, Netherlands, 4 NIZO Food Research, Ede, Netherlands Edited by: Haruki Kitazawa, Tohoku University, Japan Abbreviations: bLF, bovine Lactoferrin; COMP, cartilage oligomeric matrix protein (COMP); GOS, Galacto-oligosacharides; mDC, myeloid dendritic cell; OA, osteoarthritis; Pam, PAM3CSK4 or (S)-[2,3-Bis(palmitoyloxy)-(2-RS)-propyl]-N- palmitoyl-(R)-Cys-(S)-Ser-(S)-Lys4-OH; pDC, plasmacytoid dendritic cell; RM- (M)ANOVA, repeated measures (M)ANOVA; TLR, Toll-like receptors. Citation: van Splunter M, Perdijk O, Fick-Brinkhof H, Feitsma AL, Floris-Vollenbroek EG, Meijer B, Brugman S, Savelkoul HFJ, van Hoffen E and van Neerven RJJ (2018) Bovine Lactoferrin Enhances TLR7-Mediated Responses in Plasmacytoid Dendritic Cells in Elderly Women: Results From a Nutritional Intervention Study With Bovine Lactoferrin, GOS and Vitamin D. Front. Immunol. 9:2677. doi: 10.3389/ﬁmmu.2018.02677 November 2018 \| Volume 9 \| Article 2677 Frontiers in Immunology \| www.frontiersin.org 1 Lactoferrin Enhances TLR7 Responses in Elderly van Splunter et al. INTRODUCTION The study was set up as a double-blind placebo-controlled nutritional intervention study, to investigate the potential of bLF, GOS, and vitamin D supplementation to restore TLR responsiveness of pDCs and mDCs and to reduce inﬂammatory cytokines in serum. During aging, the immune system becomes dysregulated, as indicated by two phenomena: immunosenescence and inﬂammaging. In immunosenescence, both the innate and adaptive immune system are dysregulated. Dysregulation of the immune system seems to involve, among others, changes in the number and function of lymphocytes and innate immune cells, as well as altered expression of Toll-like receptors (TLRs) (1–3). Because of these compromised innate and adaptive immune responses, elderly people have a decreased ability to respond to infection and vaccination (2–5). Furthermore, many age-related health disorders, such as osteoarthritis, metabolic diseases, cognitive decline, onset of frailty, and cardiovascular diseases are associated with inﬂammation, often referred to as inﬂammaging (4, 6–9). Inﬂammaging is associated with increased serum concentrations of pro-inﬂammatory cytokines (3), acute-phase proteins and soluble adhesion markers (10). The age-related reduced response to TLR stimulation is best described for myeloid dendritic cells (mDCs) and plasmacytoid dendritic cells (pDCs) (8, 11–14). Interestingly, in relation to anti-viral immune responses, pDCs of elderly people have been shown to produce lower concentrations of antiviral IFN-α and pro-inﬂammatory cytokines upon TLR7 and TLR9 stimulation, resulting in lowered antiviral immunity (15, 16). Study Set-Up The eﬀect of bLF in combination with galacto-oligosacharides (GOS) and vitamin D in elderly women (65–85 years) was studied in a double-blind placebo-controlled nutritional intervention study. The protocol was approved by the Medical Ethics Committee of Wageningen University, the Netherlands (protocol no. NL57345.081.16), and registered at clinicaltrials.gov (identiﬁer NCT03026244). As TLR7 expression is X-linked (31), the study was performed in women only. Female subjects (65–85 years) were recruited. After providing informed consent, subjects were screened and were included when they were generally healthy, having a BMI 20–30, good mental status, and non-smoking. Subjects with chronic inﬂammatory, autoimmune, or gastrointestinal diseases or immune-compromised individuals were excluded from participation. Subjects using hormone replacement therapy, anti-inﬂammatory drugs (>1 × week) or immunosuppressive drugs were excluded. Furthermore, subjects were not allowed to use light therapy or go on holiday to a sunny destination. An overview of subject characteristics of the two study groups is given in Table 1. Because of the seasonal eﬀect on vitamin D status, the study was executed in the winter period (January until March 2017). Elderly people are more susceptible to severe inﬂuenza and respiratory syncytial virus (RSV) infection leading more often to hospitalization compared to adults (17) and are less responsive to inﬂuenza vaccination (8). Inﬂuenza and RSV are both single stranded RNA viruses, and the innate immune response to such viruses is mainly TLR7 mediated (14, 18). Bovine Lactoferrin (bLF) is linked to reduced number of infections by rhinoviruses and hepatitis C which are also recognized by TLR7 (19–21). bLF is an antimicrobial protein that is known to prevent sepsis, fungal infections, and enterocolitis in premature infants (22–24). In addition, bLF has been described to have anti-inﬂammatory eﬀects (25, 26). Therefore, bLF might be a nutrient from milk that is able to restore TLR7 responses of pDC to viruses. Next to this, elderly have reduced serum concentrations of vitamin D in winter, which has been shown to correlate with reduced expression and responsiveness of TLR7 and TLR8 on monocytes (27). This suggests that vitamin D may impact TLR7/8 responsiveness. Study Design y g Stratiﬁcation and randomization was performed by a non- blinded person not involved in the study, all investigators were blinded until all data were collected. Women were stratiﬁed according to age, BMI, reported arthrosis, use of vitamin D supplements preceding the study, and use of medication for blood pressure or cholesterol. Subjects were randomly assigned to treatment or placebo using a random number generator. Women (n = 15) in the nutritional intervention group were supplemented with only bLF for 3 weeks, followed by 3 weeks bLF + GOS, followed by 3 weeks bLF + GOS + vitamin D, see Figure 1. In parallel, the placebo group (n = 15) received maltodextrin as placebo for bLF and GOS, and capsules ﬁlled with maltodextrin as placebo for vitamin D. Subjects were instructed to maintain their habitual diet, but to stop any vitamin D or prebiotic supplementation during the study, starting from 2 weeks before study start. Sample collection was done every 3 weeks at Wageningen University at the clinical studies facility. The primary outcome of this study was IFN-α and IL-6 production by PBMCs upon ex vivo TLR7 stimulation. Secondary outcomes were TNF-α production by PBMCs upon ex vivo TLR7 stimulation, and the percentage of IFN-α, IL-6, or TNF- α-producing pDCs in PBMCs upon ex vivo TLR7 stimulation. In addition to direct eﬀects on immune function, aging is also associated with microbiota changes in the gastrointestinal tract. Prebiotic oligosaccharides, such as galacto-oligosaccharides (GOS), have been shown to increase the concentrations of beneﬁcial Biﬁdobacteria in the gut of elderly in several studies (28–30). Interestingly, consumption of GOS also reduced the concentrations of circulating pro-inﬂammatory cytokines (29). November 2018 \| Volume 9 \| Article 2677 Frontiers in Immunology \| www.frontiersin.org 2 Lactoferrin Enhances TLR7 Responses in Elderly van Splunter et al. TABLE 1 \| Characteristics of study participants. bLF and GOS have an FDA-approved GRAS status. From week 6–9, subjects received bLF + GOS + vitamin D. Vitamin D capsules (Supra D Forte Supradyn, Berlin, Germany) contained 20 µg cholecalciferol (= 800 IE) per capsule. Maltodextrin was used as placebo product, as well as carrier for the nutritional intervention products. As placebo for vitamin D capsules, empty capsules were purchased and ﬁlled with ∼250 mg maize based maltodextrin (Glucidex, IT19 premium, Roquette, Nord-Pas-de- Calais, France). bLF and GOS were mixed in bulk amount, with maltodextrin as carrier. Rand-36 Questionnaire At the beginning and end of the intervention study participants ﬁlled in the RAND 36-item short form survey instrument (RAND-36) questionnaire (32). The RAND-36 questionnaire is a validated questionnaire for the perception of health and well- being. The RAND-36 was analyzed according to the protocol on the subjects physical functioning, limitations in physical health, pain and general health (https://www.rand.org/health/ surveys_tools/mos/36-item-short-form/scoring.html). Statistics was done by rank-transforming the data or logit transformation (only general health). A generalized linear model repeated measures ANOVA was performed. Sample Size Calculation Since information on the impact of diet on innate immune responses in the elderly is lacking, several assumptions were made based on studies using similar outcomes as anticipated for the NOBLE study. Furthermore, it was assumed that a potential eﬀect may be detected at any time point after intervention. Therefore, dependent and independent t-tests for 2 means were used as basis for sample size calculation, rather than ANOVA. Selected primary outcome of this study was IFN-α and IL-6 production by PBMCs upon TLR7 stimulation. Therefore, for sample size calculation, we used a study in which the percentage of ex vivo IFN-α producing pDCs was measured after TLR7 stimulation upon hormonal replacement therapy (12). In this study, the eﬀect was an increase of 11.2% of IFN-α producing pDCs, with a pooled SD of 9.9 and an assumed rho for paired samples 0.3. Based on two-sided statistical testing (dependent samples, t-test), α = 0.05 (chance of type-I error) and β = 0.20 (chance of type-II error), it was calculated using the software package Statistica (2013) that at least 12 subjects per group were needed for detection of this eﬀect at any time point after supplementation. Assuming that the placebo group would not show a change in the % IFN-α producing pDCs, 14 subjects would be needed to conﬁrm this by comparing the intervention and placebo group with an independent t-test. Because of potential dropouts during the study, 1 extra subject per group was included, resulting in a group size of 15 subjects, and a total study size of 30 subjects. Verum (“intervention”) and placebo powders were aliquoted in blinded and coded jars in a food-grade environment, in adequate amounts to allow dosing over at least 21 days. After careful instruction, participants dosed the powders at home. One dose consisted of a smoothly streaked plastic scoop that was provided, which was dissolved in water and consumed after the evening meal. Vitamin D or placebo capsules were provided in separate blinded and coded jars, and were swallowed together with the dissolved powder. Compliance to the intake of study product was monitored by weighing of powders and counting of capsules at start and end of each intervention period. Subjects also recorded the intake of study product daily in a diary. Study Design The exact content of each component in the ﬁnal powder to allow proper dosing was checked and conﬁrmed. Powders were analyzed for potential microbiological contamination (i.e., Bacillus cereus, Enterobacteriaceae, yeast and fungi, total plate count, coliforms, and E. coli). All values were below detection, conﬁrming the microbiological food safety of the powder batches. FIGURE 1 \| Nutritional intervention study design. The intervention group (n = 15) received bovine lactoferrin (bLF) for 3 weeks (days 0–21), followed by 3 weeks of bLF + Galacto-oligosacharides (GOS) (days 21–42) and 3 weeks of bLF + GOS + vitamin D. The placebo group (n = 15) received maltodextrin for 9 weeks (days 0–63). At study days 0, 21, 42, and 63, blood was collected. Study Products In the intervention group, subjects received 3 weeks of bLF (1.026 g/day Vivinal Lactoferrin powder, containing an active dose of bLF of 1 g/day; FrieslandCampina Domo, Amersfoort, The Netherlands). This was followed by 3 weeks of supplementation with bLF + GOS (1 g/day bLF as above; 3.67 g/day Vivinal GOS powder, containing an active dose of GOS of 2.64 g/day; FrieslandCampina Domo, Amersfoort, The Netherlands). Both Statistical Analysis Statistical analysis was performed by using IBM SPSS Statistics version 23. Data were tested for normal distribution using Shapiro-Wilk test. The intracellular cytokine production (% of all pDCs or all mDC) of pDC and mDCs per stimulation (RPMI, Pam, LPS, R848, and CpG) were analyzed by repeated measures MANOVA (RM-MANOVA). To obtain normally distributed data, percentages were logit-transformed or rank-transformed. For TLR expression the median MFI of TLR 2, 4, 7, and 9 expression was analyzed per cell type (pDC or mDC). The median MFI was 10 log transformed to obtain normally distributed data. TLR expression was analyzed by RM-ANOVA per TLR per cell type (pDC or mDC). Serum concentrations of pro-inﬂammatory markers (pg/ml) in serum were 10 log transformed to obtain normally distributed data. As the value 0 cannot be 10 log transformed, this value was artiﬁcially put on 0.001 to obtain a value after transformation. No RM-MANOVA could be performed on pro-inﬂammatory markers in serum as too many donors would be excluded, therefore an RM-ANOVA per marker was performed. After data transformation, outliers (>2SD) were removed. The transformed data were used for analysis. RM-MANOVA and RM-ANOVA were performed with additional analysis of contrasts (diﬀerence and repeated) of time∗treatment. All statistical results are therefore a diﬀerence over time between the nutritional intervention group and placebo group. Statistically signiﬁcant results are indicated in the ﬁgures as ∗(p < 0.05) or ∗∗(p < 0.01). As this was a nutrition intervention study with a relatively low number of individuals (n = 15 per group), we were also interested in statistical trends Flow cytometry data analysis was performed by using FlowJo software (version 10 TreeStar, Inc.) and gating was performed as is shown in Figure S1, in line with Panda et al. (8) PDCs were gated as Lineage2−HLA-DR+CD11c−CD123+ and mDCs were gated as Lineage2−HLA-DR+CD11c+CD123−. Data were exported as median ﬂuorescent intensity for either all pDCs or mDCs per TLR. TLR Expression in pDCs and mDCs p p Isolated PBMCs were stained with a TLR antibody panel (Table 2) to measure the expression of TLR2, 4, 7, and 9 ex vivo. To measure expression of TLRs, 2 × 106 freshly isolated PBMCs/donor were stained in a 96 well plate (NUNC PP Sigma- Aldrich 7116). Per well 200 µl FACS buﬀer (PBS (Lonza BE17- 516Q/12) + 2mM EDTA (Merck CBI 108418); 0.5% BSA (Roche 10735086001); 0.01% NaN3 (Merck CBI 822335) was added to wash the cells. Washing was done by centrifuging the plate at 400 × g for 3 min at 4◦C. First, extracellular surface markers (Table 2) including 5 µl Fc block (564220, BD Pharmingen) were stained for 30 min on ice covered in aluminum foil and washed twice with cold PBS. Cells were stained with Fixable Viability Dye FVD520 (65086718, Ebioscience) in PBS and incubated for 20 min in the fridge, followed by washing the cells FACS buﬀer. Afterwards cells were permeabilized by adding IC ﬁxation buﬀer (00-8222-49, Ebioscience) to each well and incubated for 30 min at room temperature, followed by washing twice in Perm buﬀer (00-8333-56, Ebioscience). The intracellular antibody mix (Table 2) in Perm buﬀer was incubated for 20 min in the fridge, followed by washing the cells twice in Perm buﬀer. Cells were resuspended in 300 ul FACS buﬀer and measured for 240 s on the FACS CANTO II at medium ﬂow rate, threshold 45,000. For the panels ﬂuorescent minus one (FMO) controls and isotype controls were included. Registration of Adverse Events Any adverse events (AE) were registered at each visit, with subjects being asked for any health complaints since the last visit. Furthermore, participants recorded AEs in a diary. The AEs were evaluated by the study physician, and judged as being or not November 2018 \| Volume 9 \| Article 2677 Frontiers in Immunology \| www.frontiersin.org 3 Lactoferrin Enhances TLR7 Responses in Elderly van Splunter et al. being related to the study product (i.e., dietary intervention) or to the study procedures (i.e., blood sampling). Sigma-Aldrich) in RPMI-1640 with 5% human AB serum (H4522, Sigma Aldrich). Afterwards, cells were harvested by pipetting and stained with Lineage2, HLA-DR, CD11c, CD123, IFN-α, TNF-α, and IL-6, as described for the TLR staining. Cells were resuspended in 250 ul FACS buﬀer and measured for 200 s with FACS Canto II. Flow cytometry data analysis was performed by using FlowJo software (version 10 TreeStar, Inc.) and gating was performed as is shown in Figure S1. Data were exported as % cytokine-positive pDCs or mDC as % of all pDCs or mDCs. Cytokine and Pro-inﬂammatory Marker Measurements in Serum Measurements in Serum In serum, IL-1β (558279, BD Pharmingen); TNF-α (560112, BD Pharmingen), IL-6 (558276, BD Pharmingen), sCD106 (sVCAM-1; 560427, BD Pharmingen), sCD54 (ICAM-1; 560269, BD Pharmingen), and IL-10 (558274, BD Pharmingen) were measured by cytometric bead array, according to manufacturer’s protocol. Beads were measured for 50 s at high speed using a FACS CANTO II. Furthermore, IL-1RA (CHC1183, Thermo Fisher) and cartilage oligomeric matrix protein (COMP) (DY3134, R&D systems) were measured by ELISA according to manufacturer’s protocol. CRP was measured with a immunoturbidemetric assay using the c802 module of Cobas 8000 from Roche. 25-OH-vitamine D was measured using chemiluminescent immunoassay using Liaison XL from Diasorin. Blood Sampling Blood was collected at study days 0, 21, 42, and 63 for serum storage (10 mL tubes; cat.no. 367895, BD) or for PBMC isolation (K2-EDTA; 4 × 10mL; cat.no. 367525, BD). Serum tubes were left at room temperature for at least 30 min before centrifugation at 2000 × g 10 min at room temperature. Serum was aliquoted and stored at -80◦C. PBMCs were isolated within 6 h using 50 ml Leucosep tubes (227290, Greiner Bio-One) ﬁlled with Ficoll plaque plus (17-1440-02, GE Healthcare Life Sciences) according to manufacturer’s protocol. Intracellular Cytokine Measurement in pDCs and mDCs In order to measure intracellular cytokines, 2 x106 PBMCs were stimulated in a 12-well plate (CLS3513-50ea, Sigma-Aldrich) (total volume 1 ml) for 3 h in the absence or presence of PAM3CSK4 (Pam) 10 µg/ml (L2000, EMC microcollections), Ultra-pure LPS 0.1 µg/ml (3pelps, Invivogen), R848 3 µg/ml (TLRL-R848-5, Invivogen) or CpG 3 µg/ml (TLRL-2216-1 (class “A”), Invivogen) in the presence of Brefeldin A (B7651, November 2018 \| Volume 9 \| Article 2677 Frontiers in Immunology \| www.frontiersin.org 4 Lactoferrin Enhances TLR7 Responses in Elderly van Splunter et al. TABLE 2 \| Antibodies used for TLR expression and intracellular cytokine measurements. Antibody Fluorochrome Host Isotype Light chain Company Catalog number Panel Extra/intracellular Lineage 2 FITC mouse IgG1 κ BD 643397 TLR and cytokine Extra HLA-DR APC-Cy7 mouse IgG2b κ Ebioscience 47-9956-42 TLR and cytokine Extra CD123 PE-Cy5 mouse IgG1 κ Ebioscience 15-1239-42 TLR and cytokine Extra CD11c PE-Cy7 mouse IgG1 κ Ebioscience 25-0116-42 TLR and cytokine Extra TLR2 biotin mouse IgG2a κ Ebioscience 13992282 TLR Extra TLR2 ic biotin mouse IgG2a κ Ebioscience 13472785 TLR Extra streptavidin BV510 BD 563261 TLR Extra TLR4 BV421 mouse IgG1 κ BD 564401 TLR Extra TLR4 ic BV421 mouse IgG1 κ BD 562438 TLR Extra FVD 520 eﬂuor520 Ebioscience 65-0867-18 TLR and cytokine Extra TLR7 lc PE mouse IgG2a R&D Systems IC5875P TLR Intra TLR7 ic PE mouse IgG2a R&D Systems IC003P TLR Intra TLR9 APC rat IgG2a κ Ebioscience 17909982 TLR Intra TLR9 ic APC rat IgG2a κ Ebioscience 17-4321-81 TLR Intra CD16 BV510 mouse IgG1 κ BD 740203 Cytokine Extra IL-6 PE rat IgG1 κ Ebioscience 12706982 Cytokine Intra IL-6 ic PE rat IgG1 κ Ebioscience 12430183 cytokine Intra IFN-alpha V450 mouse IgG1 κ BD 561382 Cytokine Intra IFN-alpha ic V450 mouse IgG1 κ BD Horizon 561504 cytokine Intra TNF-alpha APC mouse IgG1 κ Ebioscience 17734982 Cytokine Intra TNF-alpha ic APC mouse IgG1 κ Ebioscience 17-4714-41 Cytokine Intra TABLE 2 \| Antibodies used for TLR expression and intracellular cytokine measurements. The Effect of Nutritional Intervention on Intracellular Cytokine Production in pDCs Upon TLR Stimulation and not only in statistically signiﬁcant diﬀerences. We considered trends relevant if 0.05 ≤p ≤0.10 and indicated this in ﬁgures with a # symbol. and not only in statistically signiﬁcant diﬀerences. We considered trends relevant if 0.05 ≤p ≤0.10 and indicated this in ﬁgures with a # symbol. To determine if the nutritional intervention impacted the antiviral response in pDCs, intracellular IFN-α, IL-6, and TNF- α production was measured after stimulation with TLR7/8 ligand R848. The percentage of IL-6+ pDCs increased signiﬁcantly (p = 0.005) and IFNα+ pDCs tended to increase (p = 0.09) at day 21 compared to day 0 in the nutritional intervention group compared to the placebo group (Figure 2). As the nutritional intervention group only consumed bLF during the ﬁrst 21 days, this indicates that lactoferrin supplementation increased the response of pDCs to TLR7/8 stimulation. TNF-α+ pDCs tended to increase at day 42 compared to day 0 in the nutritional intervention group after TLR7/8 stimulation, while the placebo group increased from days 21 to 42. Safety and Tolerability Based on the number of participants needed for the study and the inclusion and exclusion criteria, 30 elderly women were selected to participate in the study. In both the intervention group (n = 15) and placebo group (n = 15) no subjects dropped out during the study period. The intervention with bLF, GOS, and vitamin D was generally well-tolerated and safe, as only a few mild study- related AE were reported (e.g., ﬂatulence and change in bowel habit), mainly in the intervention group. No moderately severe AE related to the study were reported. One non-study related serious adverse event was reported in the placebo group. At day 63, the number of TNF-α+, IL-6+, and IFN-α+ pDCs was the same in the placebo group and the nutritional intervention group after TLR7/8 stimulation. Overall, it can be concluded that bLF speciﬁcally increased IL-6 and tended to increase IFN-α production in pDCs (days 0 to 21) upon stimulation with R848, but that subsequent supplementation of GOS and vitamin D did not further enhance this eﬀect compared to the placebo group. This increased cytokine production was not caused by an increase of TLR7 expression in pDCs of the intervention group compared to the placebo group (Figure 2B). Percentages of pDCs and mDCs To determine if the nutritional intervention aﬀected the numbers of circulating pDCs and mDCs, unstimulated PBMCs were stained with Lineage 2, HLA-DR, CD11c, and CD123. Using the gating strategy as depicted in Figure S1, the percentage of pDCs and mDCs were determined at each time point. Table 3 shows that, although there were some ﬂuctuations in pDC and mDC percentages over time, no signiﬁcant changes between the nutritional intervention and placebo group were found. In order to determine which subset of cytokine producing pDCs was responsible for the increase of IL-6+ and IFN-α+ pDCs, a boolean gating strategy was performed, resulting in November 2018 \| Volume 9 \| Article 2677 Frontiers in Immunology \| www.frontiersin.org 5 Lactoferrin Enhances TLR7 Responses in Elderly van Splunter et al. mainly due to IFN-α+TNF-α+ pDCs, and for IL-6 the majority of positive pDCs are IL-6+TNF-α+. This is in contrast to TNF-α, where the majority of TNF-α production is derived from single TNF-α+ pDCs. all eight combinations of IL-6, IFN-α, and TNF-α production from single to triple positive pDCs. Figure 2C depicts the four combinations of cytokines that were observed to be produced by pDCs. Figure 2C shows that the increase in IFN-α production is In addition to stimulation through TLR7/8, stimulations of pDCs were performed through TLR1/2 (Pam), TLR4 (LPS), and TLR9 (CpG). In Table S1, the percentages of pDCs containing intracellular IL-6, TNF-α, and IFN-α after stimulation with Pam, LPS, and CpG are shown. TABLE 3 \| The percentage of pDCs and mDC (%pDC or %mDC of all DCs) in unstimulated PBMCs (median + range) for the nutritional intervention and placebo group at study days 0, 21, 42, and 63. Treatment Day 0 Day 21 Day 42 Day 63 %pDCs Intervention 11.4% (3.1–45.0%) 12.6% (4.5–40.5%) 14.4% (6.0–34.2%) 13.3% (4.8–30.3%) Placebo 14.3% (3.3–27.7%) 13.7% (5.5–31.4%) 13.5% (4.9–31.0%) 15.9% (4.8–22.6%) %mDCs Intervention 68.2% (26.9–86.0%) 76.8% (50.3–87.7%) 73.2% (55.4–88.3%) 74.1% (36.1–85.8%) Placebo 67.6% (28.0–88.9%) 76.2% (58.9–89.8%) 77.2% (49.5–90.0%) 72.9% (37.7–86.9%) TABLE 3 \| The percentage of pDCs and mDC (%pDC or %mDC of all DCs) in unstimulated PBMCs (median + range) for the nutritional intervention and placebo group at study days 0, 21, 42, and 63. unstimulated PBMCs (median + range) for the nutritional intervention and placebo group at study days 0, 21, 42, and 63. Percentages of pDCs and mDCs November 2018 \| Volume 9 \| Article 2677 6 Frontiers in Immunology \| www.frontiersin.org Lactoferrin Enhances TLR7 Responses in Elderly van Splunter et al. tended to increase in the intervention group, both from days 0 to 42 (p = 0.088), as well as, from days 21 to 42 (p = 0.085) (data not shown). Overall, these results show that enhanced cytokine production upon TLR7/8 and to a lesser extent TLR1/2 stimulation mainly occurs after 21 days in pDCs and is thus primarily the result of ingestion of bLF. most likely due to a combined eﬀect of bLF, GOS, and vitamin D. IL-6, TNF-α, sICAM, IL-1β, CRP, IL-1Ra, and IL-10 did not change during the study. Cartilage oligomeric matrix protein (COMP) as potential prognostic marker for osteoarthritis (33) was also measured in serum, as were vitamin D levels (Figure S3). Vitamin D levels did not change during the study. Although the concentration of soluble ICAM and COMP seemed to decrease gradually over time in the nutritional intervention group, this was not signiﬁcant at any of the timepoints. RAND-36 Questionnaire Table 4 shows the outcomes of analysis of the RAND-36 questionnaire in the nutritional intervention and placebo group. The questionnaire was only completed before and after the study, and thus compares health status at day 0 vs. day 63. A score of 100 is considered very healthy, a score of 0 very unhealthy. Physical function tended to improve in the nutritional intervention group (p = 0.09), but no eﬀects were observed on pain, general health and limiting physical health. The results showed no major change when individuals with osteoarthritis were excluded from this analysis, indicating that the main eﬀect of supplementation accounted for the whole group and not speciﬁcally for individuals with osteoarthritis (not shown). Likewise, TLR9 stimulation with CpG induced more IFN- α+ mDCs in the nutritional intervention group only when comparing days 63 to 0 (p = 0.029). At the same time, in the nutritional intervention group the percentage IL-6+ mDCs after TLR9 stimulation at day 63 tended to be increased in the nutritional intervention group compared to the placebo group, as IL-6+ mDCs in the placobo group decreases more over time. It should be mentioned however, that the IFN-α+ and IL-6+ mDC percentages after TLR9 stimulation are very low, and may not be clinically relevant. In contrast to the results obtained in pDCs, no diﬀerences between nutritional intervention and placebo group were observed when comparing cytokine production between days 0 and 21 (bLF treatment alone), but rather from days 0 to 63 (after the complete intervention period). Percentages of pDCs and mDCs Treatment Day 0 Day 21 Day 42 Day 63 %pDCs Intervention 11.4% (3.1–45.0%) 12.6% (4.5–40.5%) 14.4% (6.0–34.2%) 13.3% (4.8–30.3%) Placebo 14.3% (3.3–27.7%) 13.7% (5.5–31.4%) 13.5% (4.9–31.0%) 15.9% (4.8–22.6%) %mDCs Intervention 68.2% (26.9–86.0%) 76.8% (50.3–87.7%) 73.2% (55.4–88.3%) 74.1% (36.1–85.8%) Placebo 67.6% (28.0–88.9%) 76.2% (58.9–89.8%) 77.2% (49.5–90.0%) 72.9% (37.7–86.9%) In all stimulations, a higher percentage of pDCs was positive for TNF-α compared to IL-6 and IFN-α. Even though the percentage of positive cells was low, a signiﬁcant increase of IL-6 (p = 0.021) was observed in the nutritional intervention group compared to the placebo group after Pam stimulation at day 21, after 3 weeks of bLF supplementation. At day 21, no eﬀects of the nutritional intervention were observed when pDCs were stimulated with LPS and CpG. At other time-points, no eﬀects of intervention were detected in pDC, with the exception of the percentage of TNF-α+ pDC in response to CpG stimulation, that FIGURE 2 \| Intracellular cytokine production in pDCs upon R848 stimulation. (A) TNF-α, IL-6, and IFN-α positive pDCs (% positive for cytokine of all pDCs) at study days 0, 21, 42, and 63, shown as mean +/−SEM. (B) TLR7 expression on pDCs at study days 0, 21, 42, and 63. (C) Boolean gating of produced cytokine combinations only by pDCs over time. Statistical analysis was done on ranks based on logit transformed data using repeated measures MANOVA. Outliers (>2SD) were removed. p < 0.05; p < 0.01; # is given for trends, all values indicate a difference between Intervention and placebo group over time. # for IFN-α+ pDCs is obtained by analyzing days 0 and 21 only. FIGURE 2 \| Intracellular cytokine production in pDCs upon R848 stimulation. (A) TNF-α, IL-6, and IFN-α positive pDCs (% positive for cytokine of all pDCs) at study days 0, 21, 42, and 63, shown as mean +/−SEM. (B) TLR7 expression on pDCs at study days 0, 21, 42, and 63. (C) Boolean gating of produced cytokine combinations only by pDCs over time. Statistical analysis was done on ranks based on logit transformed data using repeated measures MANOVA. Outliers (>2SD) were removed. p < 0.05; *p < 0.01; # is given for trends, all values indicate a difference between Intervention and placebo group over time. # for IFN-α+ pDCs is obtained by analyzing days 0 and 21 only. Intracellular Cytokine Production in mDCs Upon TLR Stimulation None of the markers was signiﬁcantly changed at day 21, indicating that 3 weeks of bLF supplementation alone did not have clear eﬀects on serum markers involved in inﬂammation. The only tendency to an eﬀect on inﬂammatory markers was observed at day 63 after the addition of GOS and vitamin D to bLF supplementation (sVCAM). Next, the eﬀect of the nutritional intervention on mDC activation was studied. For all TLR stimulations, a higher percentage of mDCs produced TNF-α, compared to IL-6 and IFN-α, see Table S2. Upon TLR1/2 stimulation, the percentage of TNF- α+ mDCs from the nutritional intervention-treated subjects increased compared to the placebo treated subjects at day 63 compared to day 0 (p = 0.03), due to a lower percentage of TNF-α+ mDCs at t = 0. No Effect of Nutritional Intervention on TLR Expression in pDCs and mDCs This study demonstrates that 3 weeks of bLF supplementation increased intracellular cytokine production in pDCs in response to TLR7/8 and TLR1/2 activation. Increased responses to TLR activation were only seen in mDC after completion of the intervention study at day 63. The intervention did not show clear changes in inﬂammatory markers, as only sVCAM tended to decrease in the intervention group. In order to analyze whether the changes in intracellular cytokine production upon TLR stimulation was due to modulation of expression of any of the TLRs, the TLR expression levels on pDCs and mDCs were measured in unstimulated PBMCs ex vivo. No signiﬁcant diﬀerences in expression of TLR2, 4, 7, and 9 were observed for pDCs and mDCs when comparing the nutritional intervention with the placebo group over time (Figure S2), although the expression levels varied over time in both groups. This suggests that the changes in intracellular cytokines upon TLR stimulation or the diﬀerences between the intervention and placebo group were not due to diﬀerences in TLR expression levels. The rationale to study postmenopausal women aged >65 was based on several observations. TLR7 is X-linked (31) and women are found to have higher IFN-α production by pDCs compared to men (34). Furthermore, TLR7 function is reduced in postmenopausal women compared pre-menopausal women and can be improved by hormone replacement therapy, indicating that the reduced TLR7 responsiveness in elderly women can be restored (12). rum Markers Involved in Inﬂammation pDCs are the primary producers of IFN-α upon inﬂuenza infection, which is TLR7-mediated (14). As bLF is linked to reduced number of infections by rhinoviruses and hepatitis C that are also recognized by TLR7 (19–21), we hypothesized that nutritional intervention with bLF might increase the production of the antiviral cytokine IFN-α by pDCs. Indeed, bLF supplementation alone (days 0 to 21) enhanced the production of IL-6 and tended to increase the production of IFN-α in response to TLR7/8 stimulation in pDCs. Subsequent supplementation of To explore if the nutritional intervention aﬀected serum concentrations of markers involved in inﬂammation, a wide range of aging-associated inﬂammatory markers was measured. Figure 3 shows that the only inﬂammation related marker that tended to decrease after the entire invention period was soluble VCAM, which decreased at day 63 compared to day 0 (p = 0.07). As there is no clear decrease of sVCAM levels from days 42 to 63 in the intervention group, the observed decrease is November 2018 \| Volume 9 \| Article 2677 Frontiers in Immunology \| www.frontiersin.org 7 Lactoferrin Enhances TLR7 Responses in Elderly van Splunter et al. FIGURE 3 \| Serum markers involved in inﬂammation. Concentrations of IL-6, TNF-α, IL-1β, CRP, sVCAM, sICAM, IL-1RA, and IL-10 at study days 0, 21, 42, and 63 Data shown as scatter plots with median value. Analysis was done on the 10 log transformated data or ranks (IL-1β and TNF-α) by repeated measures ANOVA,. Outliers (>2 SD )were removed. All statistical differences are differences over time between the nutritional intervention group and the placebo group. p < 0.05; or # i given for trends. OS and vitamin D did not further enhance cytokine production pDCs in response to any TLR stimuli used compared to acebo. In contrast, supplementation of bLF during the ﬁrst 3 weeks did not result in enhanced responses of mDCs the same stimuli, nor aﬀected serum markers involved inﬂammation. FIGURE 3 \| Serum markers involved in inﬂammation. Concentrations of IL-6, TNF-α, IL-1β, CRP, sVCAM, sICAM, IL-1RA, and IL-10 at study days 0, 21, 42, and 63. Data shown as scatter plots with median value. Analysis was done on the 10 log transformated data or ranks (IL-1β and TNF-α) by repeated measures ANOVA,. Outliers (>2 SD )were removed. All statistical differences are differences over time between the nutritional intervention group and the placebo group. rum Markers Involved in Inﬂammation *p < 0.05; or # is given for trends 3 weeks did not result in enhanced responses of mDCs to the same stimuli, nor aﬀected serum markers involved in inﬂammation. GOS and vitamin D did not further enhance cytokine production of pDCs in response to any TLR stimuli used compared to placebo. In contrast, supplementation of bLF during the ﬁrst November 2018 \| Volume 9 \| Article 2677 Frontiers in Immunology \| www.frontiersin.org 8 Lactoferrin Enhances TLR7 Responses in Elderly van Splunter et al. TABLE 4 \| Mean scores and ranks for placebo and nutritional intervention group based on the RAND-36 questionaire. Timepoint Placebo Nutritional intervention P-values (all individuals including OA) P-values (healthy individuals without OA) Mean Rank Mean Rank Time x treatment n = 30 Time x treatment n = 24 Phys Func Day 0 95 18.1 85 12.9 0.092 0.087 Day 63 92 15.5 89 15.5 Lim. Phys Health Day 0 95 16 90 15 0.27 0.27 Day 63 85 14.5 97 16.5 Pain Day 0 93 16.7 87 14.3 0.41 0.35 Day 63 88 15.3 86 15.7 General Health Day 0 77 14.9 77 16.1 0.95 0.72 Day 63 77 15.1 76 15.9 Phys Func, physical functioning; Lim. Phys health, role limitations due to physical health problems. pro-inﬂammatory cytokines in this study. This might be because our intervention period was shorter, or because the concentration of GOS was lower compared to other studies (29, 30). Another, explanation is that our group size was too small to detect signiﬁcant changes in these markers, see below. In many parameters measured in this study, the placebo group shows similar kinetics as the intervention group, which seems a time-related eﬀect and may result in underestimation of the eﬀects of the nutritional intervention. We hypothesized that these time-related eﬀects were due to sunlight induced vitamin D. However, vitamin D did not signiﬁcantly change in the placebo or the nutritional intervention group during the study (Figure S3). Hence, it remains elusive which time-related factor contributed to this eﬀect in all subjects during the study period, and it stresses the need of taking along a placebo group. Serum concentrations of vitamin D correlate with increased IL-1β, TNF-α, and IL-6 production by monocytes upon TLR7 stimulation, while TLR7 expression is inversely correlated with vitamin D levels in serum (27). rum Markers Involved in Inﬂammation In this study, we did not observe any increase of TNF-α, IL-6, or IFN-α in stimulated pDC or mDC or changes in TLR7 expression when comparing days 42 with 63. A possible explanation is that 3 weeks of vitamin D supplementation were not suﬃcient to signiﬁcantly enhance 25-OH vitamin D concentration in serum (Figure S3). In elderly, the expression of TLR1 and TLR7 on mDCs and TLR7 on pDCs was reported to be reduced compared to young adults, while the expression of TLR2 on mDCs and TLR9 on pDCs is unchanged (8, 16). Besides, a change in signaling events downstream of TLR activation occurs upon aging that leads to reduced cytokine secretion upon stimulation (4, 7). The defective TLR function in aging is illustrated by decreased cytokine production of monocytes and pDCs, as well as, mDCs of elderly people in response to ligation of TLR1/2 (35), TLR4 (36), TLR7, and TLR9 (8). In our study, the observed changes in cytokine production by pDCs after bLF supplementation were not due to increased TLR expression levels, suggesting that bLF supplementation may have improved downstream signaling of TLRs. bLF is 69% homologous to human lactoferrin on the protein level (37) and can be taken up by human cells via lactoferrin-receptor (intelectin) (38). We hypothesize that bLF exerts its eﬀect by binding to intelectin or one of the other bLF receptors that are expressed by immune cells. These receptors are low-density lipoprotein receptor-related protein-1 (LRP-1 or CD91) (39, 40), CD14 on monocytes in complex with LPS (41), TLR4 (42), and CXCR4 (40). It is currently not known whether bLF is taken up and ends up in the blood as the whole protein or that bLF is partially digested and active peptides end up in blood, which subsequently exert an eﬀect on pDCs. Interestingly, two recent publications have also reported eﬀects of dietary intervention in elderly people on the innate immune response after TLR activation (43, 44). In the ﬁrst study, a mediterranean diet did not have an eﬀect on circulating mDC and pDC numbers, but reduced the cytokine responses of PBMC in response to TLR-mediated stimulation (44). In the second study, a mediterranean diet with additional vitamin D3 supplementation was shown to have gender-speciﬁc eﬀects on TLR-mediated activation of PBMC in women, that in analogy with the data presented here for pDC showed increased induction of CD40+CD86+ cells upon stimulation (43). Frontiers in Immunology \| www.frontiersin.org SUPPLEMENTARY MATERIAL The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/ﬁmmu. 2018.02677/full#supplementary-material rum Markers Involved in Inﬂammation Recent hypotheses have suggested that the decreased response to stimuli through TLR in elderly people and the enhanced steady state production of cytokines by blood cells of elderly people may have the same underlying cause, being miRNAs that regulate activation of myeloid cells downstream of TLRs (45). It should be mentioned that the group size in this study was based on power calculations related to the primary and secondary outcomes of the study (IFN-α, TNF-α, and IL-6 production by pDCs upon ex vivo stimulation). For statistically signiﬁcant eﬀects on inﬂammatory serum markers, larger group sizes are needed (26). In addition, the population of elderly women in this study was in general healthy and mobile, with only a few subjects with Osteoarthritis (OA) (6/30). Besides, the average age was relatively low compared to other studies (46, 47). Inclusion of GOS has been shown to exert an eﬀect on inﬂammatory serum markers via the increase of Biﬁdobacteria levels, which are important for the production of short chain fatty acids (SCFAs), and to reduce pro-inﬂammatory cytokine concentrations in serum (29). However, we did not observe a reduction in November 2018 \| Volume 9 \| Article 2677 Frontiers in Immunology \| www.frontiersin.org 9 Lactoferrin Enhances TLR7 Responses in Elderly van Splunter et al. larger study groups including less mobile elderly women or more (non-hospitalized) women with chronic inﬂammatory diseases might be recommended in future studies to demonstrate eﬀects on inﬂammatory markers with these ingredients. larger study groups including less mobile elderly women or more (non-hospitalized) women with chronic inﬂammatory diseases might be recommended in future studies to demonstrate eﬀects on inﬂammatory markers with these ingredients. to conﬁrm and extend our ﬁndings on DCs and inﬂammatory markers in elderly women. AUTHOR CONTRIBUTIONS Study design RvN, EvH, HS, SB, and MvS. EF-V, EvH, MvS, BM, HF-B, OP, and AF execution and logistics of the study. MvS, BM, and AF data analysis. MvS, OP, SB, HS, EvH, and RvN manuscript writing and editing. Study design RvN, EvH, HS, SB, and MvS. EF-V, EvH, MvS, BM, HF-B, OP, and AF execution and logistics of the study. MvS, BM, and AF data analysis. MvS, OP, SB, HS, EvH, and RvN manuscript writing and editing. ACKNOWLEDGMENTS We would like to thank Johanna Wessel, Erik van den Brink, Jantien Takens, Diana Emmen, Ineke Klöpping, Evelien Kramer, Joyce Schloesser, Maartje van den Belt, Marja Kanning, and Guido Staring for helping with the execution of the NOBLE study. In conclusion, nutritional supplementation with bLF, GOS, and vitamin D is safe and enhances responses to TLR stimuli in elderly women in both pDCs and mDCs, while no clear eﬀects on pro-inﬂammatory markers in serum were observed, possibly due to the study group size. Especially TLR7-mediated responses in pDCs were enhanced after bLF supplementation compared to placebo, suggesting that bLF may contribute to protection against viral infections in elderly women. The outcomes of this nutritional intervention study warrants future studies, that should be powered on the basis of the outcomes described here FUNDING MvS, OP, EF-V, and EvH received research funding from FrieslandCampina to execute the study. OP is supported by the Netherlands Organization of Scientiﬁc Research (NWO) as part of the technology foundation STW (project number 13017). MvS, OP, EF-V, and EvH received research funding from FrieslandCampina to execute the study. OP is supported by the Netherlands Organization of Scientiﬁc Research (NWO) as part of the technology foundation STW (project number 13017). MvS, OP, EF-V, and EvH received research funding from FrieslandCampina to execute the study. OP is supported by the Netherlands Organization of Scientiﬁc Research (NWO) as part of the technology foundation STW (project number 13017). ETHICS STATEMENT In addition to the immune parameters described, the RAND- 36 questionnaire was used to assess the health status of the subjects. The RAND-36 questionnaire results showed a trend toward improved physical function in the intervention group, as compared to the placebo. bLF has been shown to prevent arthritis in experimental animal models (48). COMP is a physiological parameter that is predictive for development of osteoarthritis (33). The tendency toward improved physical function together with the non-signiﬁcant decrease in COMP concentrations in the intervention group, might suggest that nutritional supplementation with bLF, possibly in combination with GOS and vitamin D, could be relevant for prevention of osteoarthritis (OA). For practical reasons it was not possible to investigate the eﬀect of bLF, GOS, and vitamin D in a parallel study setup with all interventions separately and combined in four active groups and one placebo group. Therefore, a longitudinal set-up with a staged introduction of bLF, GOS, and Vitamin D was chosen. However, in this study setup it is only possible to assess the eﬀect of bLF after 3 weeks, and of the combined intervention of bLF, GOS, and vitamin D after completion of the study. As a result, we cannot conclude if the eﬀect on physical function is the result of bLF for a prolonged period or of the GOS and vitamin D that was introduced later. However, larger studies with separate study groups including higher numbers of elderly people with OA will be needed to investigate this nutritional intervention eﬀect further. This study was carried out in accordance with the recommendations of Medical Ethics Committee of Wageningen University, the Netherlands with written informed consent from all subjects. All subjects gave written informed consent in accordance with the Declaration of Helsinki. The protocol was approved by the Medical Ethics Committee of Wageningen University, the Netherlands. This study was carried out in accordance with the recommendations of Medical Ethics Committee of Wageningen University, the Netherlands with written informed consent from all subjects. All subjects gave written informed consent in accordance with the Declaration of Helsinki. The protocol was approved by the Medical Ethics Committee of Wageningen University, the Netherlands. REFERENCES Age and low levels of circulating vitamin D are associated with impaired innate immune function. J Leukoc Biol. (2012) 91:829–38. doi: 10.1189/jlb.1011523 11. Garbe K, Bratke K, Wagner S, Virchow JC, Lommatzsch M. Plasmacytoid dendritic cells and their Toll-like receptor 9 expression selectively decrease with age. Hum Immunol. (2012) 73:493–7. doi: 10.1016/j.humimm.2012.02.007 28. Vulevic J, Juric A, Walton GE, Claus SP, Tzortzis G, Toward RE, et al. Inﬂuence of galacto-oligosaccharide mixture (B-GOS) on gut microbiota, immune parameters and metabonomics in elderly persons. Br J Nutr. (2015) 114:586–95. doi: 10.1017/S00071145150 01889 12. Seillet C, Laﬀont S, Trémollières F, Rouquié N, Ribot C, Arnal JF, et al. The TLR-mediated response of plasmacytoid dendritic cells is positively regulated by estradiol in vivo through cell-intrinsic estrogen receptor α signaling. Blood (2012) 119:454–64. doi: 10.1182/blood-2011-08- 371831 29. Vulevic J, Drakoularakou A, Yaqoob P, Tzortzis G, Gibson GR. Modulation of the fecal microﬂora proﬁle and immune function by a novel trans-galactooligosaccharide mixture (B-GOS) in healthy elderly volunteers. Am J Clin Nutr. (2008) 88:1438–46. doi: 10.3945/ajcn.2008. 26242 13. Agrawal A, Agrawal S, Cao JN, Su H, Osann K, Gupta S. Altered innate immune functioning of dendritic cells in elderly humans: a role of phosphoinositide 3-kinase-signaling pathway. J Immunol. (2007) 178:6912– 22. doi: 10.4049/jimmunol.178.11.6912 30. Canfora EE, van der Beek CM, Hermes GDA, Goossens GH, Jocken JWE, Holst JJ, et al. Supplementation of diet with galacto-oligosaccharides increases biﬁdobacteria, but not insulin sensitivity, in obese prediabetic individuals. Gastroenterology (2017) 153:87–97.e3. doi: 10.1053/j.gastro.2017. 03.051 14. Canaday DH, Amponsay NA, Jones L, Tisch DJ, Hornick TR, Ramachandra L. Inﬂuenza-induced production of interferon-alpha is defective in geriatric individuals. J Clin Immunol. (2010) 30:373–83. doi: 10.1007/s10875-010-9374-9 31. Fish EN. The X-ﬁles in immunity: sex-based diﬀerences predispose immune responses. Nat Rev Immunol. (2008) 8:737–44. doi: 10.1038/n ri2394 15. Metcalf TU, Cubas RA, Ghneim K, Cartwright MJ, van Grevenynghe J, Richner JM, et al. Global analyses revealed age-related alterations in innate immune responses after stimulation of pathogen recognition receptors. Aging Cell (2015) 14:421–32. doi: 10.1111/acel.12320 32. Ware JE, Sherbourne CD. The MOS 36-item short-form health survey ( SF-36 ): i. conceptual framework and item selection. Med Care (1992) 30:473–83. 33. Verma P, Dalal K. Serum cartilage oligomeric matrix protein (COMP) in knee osteoarthritis: a novel diagnostic and prognostic biomarker. J Orthop Res. (2013) 31:999–1006. doi: 10.1002/jor.22324 16. Jing Y, Shaheen E, Drake RR, Chen N, Gravenstein S, Deng Y. REFERENCES 6. Michaud M, Balardy L, Moulis G, Gaudin C, Peyrot C, Vellas B, et al. Proinﬂammatory cytokines, aging, and age-related diseases. J Am Med Dir Assoc. (2013) 14:877–82. doi: 10.1016/j.jamda.2013. 05.009 1. Shaw AC, Panda A, Joshi SR, Qian F, Allore HG, Montgomery RR. Dysregulation of human Toll-like receptor function in aging. Ageing Res Rev. (2011) 10:346–53. doi: 10.1016/j.arr.2010.10.007 7. Kollmann TR, Levy O, Montgomery RR, Goriely S. Innate immune function by Toll-like receptors: distinct responses in newborns and the elderly. Immunity (2012) 37:771–83. doi: 10.1016/j.immuni.2012. 10.014 7. Kollmann TR, Levy O, Montgomery RR, Goriely S. Innate immune function by Toll-like receptors: distinct responses in newborns and the elderly. Immunity (2012) 37:771–83. doi: 10.1016/j.immuni.2012. 10.014 2. Pawelec G. Age and immunity: what is “immunosenescence“? Exp Gerontol. (2018) 105:4–9. doi: 10.1016/j.exger.2017.10.024 3. Pawelec G, Goldeck D, Derhovanessian E. Inﬂammation, ageing and chronic disease. Curr Opin Immunol. (2014) 29:23–8. doi: 10.1016/j.coi.2014.03.007 8. Panda A, Qian F, Mohanty S, van Duin D, Newman FK, Zhang L, et al. Age-associated decrease in TLR function in primary human dendritic cells predicts inﬂuenza vaccine response. J Immunol. (2010) 184:2518–27. doi: 10.4049/jimmunol.0901022 8. Panda A, Qian F, Mohanty S, van Duin D, Newman FK, Zhang L, et al. Age-associated decrease in TLR function in primary human dendritic cells predicts inﬂuenza vaccine response. J Immunol. (2010) 184:2518–27. doi: 10.4049/jimmunol.0901022 4. Shaw AC, Goldstein DR, Montgomery RR. Age-dependent dysregulation of innate immunity. Nat Rev Immunol. (2013) 13:875–87. doi: 10.1038/nri3547 5. Pera A, Campos C, López N, Hassouneh F, Alonso C, Tarazona R, et al. Immunosenescence: implications for response to infection and vaccination in older people. Maturitas (2015) 82:50–55. doi: 10.1016/j.maturitas.2015. 05.004 5. Pera A, Campos C, López N, Hassouneh F, Alonso C, Tarazona R, et al. Immunosenescence: implications for response to infection and vaccination in older people. Maturitas (2015) 82:50–55. doi: 10.1016/j.maturitas.2015. 05.004 9. Poulose SM, Miller MG, Scott T, Shukitt-Hale B. Nutritional factors aﬀecting adult neurogenesis and cognitive function. Adv Nutr. (2017) 8:804–11. doi: 10.3945/an.117.016261 November 2018 \| Volume 9 \| Article 2677 Frontiers in Immunology \| www.frontiersin.org 10 Lactoferrin Enhances TLR7 Responses in Elderly van Splunter et al. 10. Calder PC, Ahluwalia N, Albers R, Bosco N. A Consideration of biomarkers to be used for evaluation of inﬂammation in human nutritional studies. Br J Nutr. (2013) 109 (Suppl 1):S1–34. doi: 10.1017/S00071145120 05119 27. Alvarez-Rodriguez L, Lopez-Hoyos M, Garcia-Unzueta M, Amado JA, Cacho PM, Martinez-Taboada VM. REFERENCES Aging is associated with a numerical and functional decline in plasmacytoid dendritic cells, whereas myeloid dendritic cells are relatively unaltered in human peripheral blood. Hum Immunol. (2009) 70:777–84. doi: 10.1016/j.humimm.2009.07.005 34. Berghöfer B, Frommer T, Haley G, Fink L, Bein G, Hackstein H. TLR7 ligands induce higher IFN-a production in females. J Immunol Ref. (2006) 177:2088–96. doi: 10.4049/jimmunol.177.4.2088 17. Thompson WW, Shay DK, Weintraub E, Brammer L, Cox N, Anderson LJ, et al. Mortality associated with inﬂuenza and respiratory syncytial virus in the United States. JAMA (2003) 289:179–86. doi: 10.1001/jama.289.2.179 35. van Duin D, Mohanty S, Thomas V, Ginter S, Montgomery RR, Fikrig E, et al. Age-associated defect in human TLR-1/2 function. J Immunol. (2007) 178:970–5. doi: 10.4049/jimmunol.178.2.970 18. Wang JP, Bowen GN, Padden C, Cerny A, Finberg RW, Newburger PE, et al. Toll-like receptor – mediated activation of neutrophils by inﬂuenza A virus. Blood (2008) 112:2028–34. doi: 10.1182/blood-2008-01- 132860 36. Van Den Biggelaar AH, Huizinga TW, De Craen AJ, Gussekloo J, Heijmans BT, Frölich M, et al. Impaired innate immunity predicts frailty in old age. The Leiden 85-plus study. Exp Gerontol. (2004) 39:1407–14. doi: 10.1016/j.exger.2004.06.009 19. Vitetta L, Coulson S, Beck SL, Gramotnev H, Du S, Lewis S. The clinical eﬃcacy of a bovine lactoferrin/whey protein Ig-rich fraction (Lf/IgF) for the common cold: a double blind randomized study. Complement Ther Med. (2013) 21:164–71. doi: 10.1016/j.ctim.2012.12.006 37. Liao Y, Jiang R, Lonnerdal B. Biochemical and molecular impacts of lactoferrin on small intestinal growth and development during early life. Biochem Cell Biol. (2012) 90:476–84. doi: 10.1139/o11-075 38. Lönnerdal B, Jiang R, Du X. Bovine lactoferrin can be taken up by the human intestinal lactoferrin receptor and exert bioactivities. J Pediatr Gastroenterol Nutr. (2011) 53:606–14. doi: 10.1097/MPG.0b013e31823 0a419 20. Triantaﬁlou K, Vakakis E, Richer EA, Evans GL, Villiers JP, Triantaﬁlou M. Human rhinovirus recognition in non-immune cells is mediated by Toll-like receptors and MDA-5, which trigger a synergetic pro-inﬂammatory immune response. Virulence (2011) 2:22–9. doi: 10.4161/viru.2.1.13807 39. Suzuki YA, Lopez V, Lönnerdal B. Mammalian lactoferrin receptors: structure and function. Cell Mol Life Sci. (2005) 62:2560–75. doi: 10.1007/s00018-005-5371-1 21. Okada S, Tanaka K, Sato T, Ueno H, Saito S, Okusaka T, et al. Dose-response trial of lactoferrin in patients with chronic hepatitis C. Jpn J Cancer Res. (2002) 93:1063–9. doi: 10.1111/j.1349-7006.2002.tb02484.x 40. Takayama Y, Aoki R, Uchida R, Tajima A, Aoki-Yoshida A. Role of CXC chemokine receptor type 4 as a lactoferrin receptor. REFERENCES Biochem Cell Biol. (2017) 95:57–63. doi: 10.1139/bcb-2016-0039 22. Manzoni P, Rinaldi M, Cattani S, Pugni L, Romeo MG, Messner H, et al. Bovine lactoferrin supplementation for prevention of late-onset sepsis in very low-birth-weight neonates a randomized trial. JAMA (2009) 302:1421–8. doi: 10.1001/jama.2009.1403 41. Birgens HS, Hansen NE, Karle H, Kristensen LØ. Receptor binding of lactoferrin by human monocytes. Br J Haematol. (1983) 54:383–91. doi: 10.1111/j.1365-2141.1983.tb02113.x 23. Manzoni P, StolﬁI, Messner H, Cattani S, Laforgia N, Romeo MG, et al. Bovine lactoferrin prevents invasive fungal infections in very low birth weight infants: a randomized controlled trial. Pediatrics (2012) 129:116–23. doi: 10.1542/peds.2011-0279 42. Ando K, Hasegawa K, Shindo KI, Furusawa T, Fujino T, Kikugawa K, et al. Human lactoferrin activates NF-κB through the Toll-like receptor 4 pathway while it interferes with the lipopolysaccharide-stimulated TLR4 signaling. FEBS J. (2010) 277:2051–66. doi: 10.1111/j.1742-4658.2010.0 7620.x 24. Pammi M. Abrams SA. Oral lactoferrin for the prevention of sepsis and necrotizing enterocolitis in preterm infants. Cochrane Database Syst Rev. (2015) 2:CD007137. doi: 10.1002/14651858.CD007137.pub4 43. Maijo M, Ivory K, Clements SJ, Dainty JR, Jennings A, Gillings R, et al. One-year consumption of a mediterranean-like dietary pattern with vitamin D3 supplements induced small scale but extensive changes of immune cell phenotype, co-receptor expression and innate immune responses in healthy elderly subjects: results from the United Kingdom arm of the NU-AGE trial. Front Physiol. (2018) 9:997. doi: 10.3389/fphys.2018. 00997 25. Håversen L, Ohlsson BG, Hahn-Zoric M, Hanson LÅ, Mattsby-Baltzer I. Lactoferrin down-regulates the LPS-induced cytokine production in monocytic cells via NF-κB. Cell Immunol. (2002) 220:83–95. doi: 10.1016/S0008-8749(03)00006-6 26. Bharadwaj S, Naidu TA, Betageri GV, Prasadarao NV, Naidu AS. Inﬂammatory responses improve with milk ribonuclease-enriched lactoferrin supplementation in postmenopausal women. Inﬂamm Res. (2010) 59:971–8. doi: 10.1007/s00011-010-0211-7 44. Clements SJ, Maijo M, Ivory K, Nicoletti C, Carding SR. Age- associated decline in dendritic cell function and the impact of November 2018 \| Volume 9 \| Article 2677 Frontiers in Immunology \| www.frontiersin.org 11 Lactoferrin Enhances TLR7 Responses in Elderly van Splunter et al. mediterranean diet intervention in elderly subjects. Front Nutr. (2017) 4:65. doi: 10.3389/fnut.2017.00065 Conﬂict of Interest Statement: MvS, OP, EF-V, EvH received research funding from FrieslandCampina to execute the intervention study. RvN is an employee of FrieslandCampina. mediterranean diet intervention in elderly subjects. Front Nutr. (2017) 4:65. doi: 10.3389/fnut.2017.00065 45. Olivieri F, Rippo MR, Monsurrò V, Salvioli S, Capri M, Procopio AD, et al. Frontiers in Immunology \| www.frontiersin.org November 2018 \| Volume 9 \| Article 2677 REFERENCES MicroRNAs linking inﬂamm-aging, cellular senescence and cancer. Ageing Res Rev. (2013) 12:1056–68. doi: 10.1016/j.arr.2013.05.001 The remaining authors declare that the research was conducted in the absence of any commercial or ﬁnancial relationships that could be construed as a potential conﬂict of interest. 46. Della Bella S, Bierti L, Presicce P, Arienti R, Valenti M, Saresella M, et al. Peripheral blood dendritic cells and monocytes are diﬀerently regulated in the elderly. Clin Immunol. (2007) 122:220–8. doi: 10.1016/j.clim.2006. 09.012 Copyright © 2018 van Splunter, Perdijk, Fick-Brinkhof, Feitsma, Floris-Vollenbroek, Meijer, Brugman, Savelkoul, van Hoﬀen and van Neerven. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. 47. Forsey RJ, Thompson JM, Ernerudh J, Hurst TL, Strindhall J, Johansson B, et al. Plasma cytokine proﬁles in elderly humans. Mech Ageing Dev. (2003) 124:487–93. doi: 10.1016/S0047-6374(03)00025-3 48. Samarasinghe RM, Kanwar RK, Kanwar JR. The eﬀect of oral administration of iron saturated-bovine lactoferrin encapsulated chitosan-nanocarriers on osteoarthritis. Biomaterials (2014) 35:7522–34. doi: 10.1016/j.biomaterials.2014.04.109 November 2018 \| Volume 9 \| Article 2677 Frontiers in Immunology \| www.frontiersin.org 12
https://openalex.org/W1028711795	https://www.annalsofgeophysics.eu/index.php/annals/article/download/3368/3414	English	null	Recent changes in rainfalland air temperature at Agnone(Molise - Central Italy)	Annals of geophysics	2,009	cc-by	6,113	ANNALS OF GEOPHYSICS, VOL. 47, N. 6, December 2004 ANNALS OF GEOPHYSICS, VOL. 47, N. 6, December 2004 Mailing address: Dr. Pietro Aucelli, Dipartimento di Scienze e Tecnologie per l’Ambiente e il Territorio, Uni- versità degli Studi del Molise, 86170 Isernia, Italy; e-mail: aucelli@unimol.it Recent changes in rainfall and air temperature at Agnone (Molise – Central Italy) Michela Izzo (1), Pietro P.C. Aucelli (1) and Adriano Mazzarella (2) (1) Dipartimento di Scienze e Tecnologie per l’Ambiente e il Territorio, Università degli Studi del Molise, Isernia, Italy (2) Dipartimento di Geofisica e Vulcanologia, Università degli Studi di Napoli «Federico II», Napoli, Italy Abstract An exhaustive daily rainfall and extreme air temperature series (1883-2000) was reconstructed for Agnone, a small town in Molise (Central Italy). Long-term analysis identified an increasing trend of 1.3 ± 0.4°C per 100 years, statistically confident at the 95% level, only for minimum air temperature, and of a seasonal march, reason- ably stationary along the entire investigated interval, explaining more than 50% of the corresponding monthly vari- ance, with maxima in November and July for rainfall and air temperature, respectively. Daily clustering analysis evidenced scale-invariant properties, largely dependent on the threshold value, for all the investigated parameters. exists about their quantification (Ney, 1959; Kelly and Wigley, 1992; Halgh, 1994; Stouffer et al., 1994; Lassen and Friis-Christensen, 1995; Mitchell et al., 1995; Santer et al., 1995; Friis- Christensen and Svensmark, 1997; Hegerl et al., 1997; Lean and Rind, 1999; Rozelot, 2001). Key words Molise Region – climatic change – long- term instrumental series – fractal analysis – the Can- tor dust method Key words Molise Region – climatic change – long- term instrumental series – fractal analysis – the Can- tor dust method 2.1. Historical outline of the Molise climate 2.1. Historical outline of the Molise climate This observatory continued its activity with- out interruption until 1977 (fig. 1), when it closed down and the task of carrying out precip- itation and air temperature measurements was taken over by the State Forest Department, on whose premises at Agnone a pluviometer and a thermometer were set up. This station works to this day, even after the old instruments were re- placed by electronic ones in March 2002. In Molise, there is a substantial lack of cli- matic studies and this is the first in such a direc- tion. In Molise, the earliest instrumental obser- vations date back to the end of the 19th century, when in some areas of the region the first mete- orological measurements began under the direc- tion of private citizens who had a natural aptitude for recording such kinds of data, which were then sent to the Central Office of Meteorology (COM), as it was then known. The closing down of the de Horatiis station resulted in a huge loss, because it assured not only the collection of precipitation and air tem- perature, but also of other meteorological pa- rameters (atmospheric pressure, wind speed and direction, state of the sky). Most of the doc- umentation concerning the station and many of its instruments donated by the de Horatiis fam- ily can now be found in the Agnone municipal library. The longest daily rainfall and air temperature series concern the following towns: Agnone, where instrumental meteorological measurements were started by the de Horatiis family in 1875; Ve- nafro, where Ludovico Giannini founded a mete- orological observatory in 1883; Isernia, where Camillo D’Apollonio began meteorological ob- servations in 1898; Campobasso, the main town of the region, where precipitation measurements began in 1886, and temperature observations in 1918, both carried out by Francesco Pesce. The observatory was directed by its founder for 68 years, until 1942, when management passed first into the hands of his son Cesare, un- til 1957, and then of his nephew Pier Francesco junior, who was in charge of the station until 1977. This great continuity in station manage- ment, with well documented changes, makes the quality of the Agnone series data really appre- ciable. There were various reasons behind our deci- sion to choose the daily Agnone series. 2.2. The Agnone climatic station In Agnone, meteorological observations were initiated by Pier Francesco Rufino de Ho- ratiis, who founded an observatory in 1875, which was soon to become part of the national network of the State Central Office of Meteorol- ogy and Geodynamics (its name at that time). 1. Introduction Naturally, analysis which is carried out on a series of this kind requires great effort not only in collecting data, but also in verifying their homo- geneity (Barriendos et al., 2002; Camuffo, 2002a,b; Camuffo and Jones, 2002; Cocheo and Camuffo, 2002; Demarée et al., 2002; Maugeri et al., 2002a,b). This operation ensures that the time variations observed in the meteorological parameters can be identified as climatic signals and it is thus extremely important to collect the so-called metadata, i.e. all the information about the history of the station where measurements have been taken. The reconstruction and analysis of past cli- mate is important not only in basic research, but also in land use management. Moreover, instru- mental meteorological data constitute the refer- ence basis for setting up indirect techniques in such fields as palynology and dendrochronolo- gy, which are commonly used to investigate the history of past climate. y p It is widely held that a combined effect of natural and human causes could underlie recent climatic changes, even if great uncertainty still This study presents the results of the recon- struction and analysis of the daily precipitation and temperature series at Agnone, an upland town (800 m a.s.l.) in Molise, a small region in Central Italy. In analysing air temperature, only 1689 Michela Izzo, Pietro P.C. Aucelli and Adriano Mazzarella the daily maximum and minimum values were considered. Extreme air temperature is more suit- able than daily mean temperature since: a) due to the quasi-sinusoidal nature of the diurnal temper- ature cycle, temperatures close to extreme values occur more frequently than those around the mean daily value; b) international methods of recording data across the world have only recent- ly been standardized; therefore, when a historical series is examined, it must be ascertained that the data used for deriving mean values were observed at the same local times throughout. This problem obviously does not hold for extreme temperatures (Palumbo and Mazzarella, 1984). rienced a significant change in its urban nucle- us, nor has there been human activity which could significantly affect precipitation and tem- perature. From this point of view, it is reason- able to state that the Agnone series does not suf- fer the so-called heat-island effect which com- monly masks the meteorological variables measured in urban areas (Palumbo and Maz- zarella, 1984). 3. Methodology In the analysis of the daily Agnone rainfall and air temperature series, problems arise when the original registers were missing. In their ab- sence we analysed the Daily Meteoric Bulletin of the COM that attributed the record of the dai- ly rainfall and of maximum air temperature to the previous day. Hence, considering the years for which mean monthly and annual data were 2.1. Historical outline of the Molise climate First of all, it goes the furthest back in time but, what is much more important, it shows a good com- pleteness of data, a significant continuity in terms of station management and substantial constancy of the environment where measure- ments were taken. Furthermore, it should be re- called that over the years Agnone has not expe- We consulted the following sources for the acquisition of daily rainfall and air temperature data: 1690 Recent changes in rainfall and air temperature at Agnone (Molise – Central Italy) Fig. 1. Locations of the Agnone climatic station since 1883. Fig. 1. Locations of the Agnone climatic station since 1883. – For the 1883-1921 period, original hand- written registers available about UCEA and, in their absence, the Daily Meteoric Bulletins of the COM. obtained from this source, the daily rainfall and maximum air temperature series has been shift- ed back by one day. The accuracy of the daily rainfall data and extreme air temperature is estimated to be equal to 0.1 mm and 0.1°C, respectively (De Horatiis, 1967). – For the 1922-2000 period, the Annals of the National Hydrographic Service (NHS) (at first in the department of Chieti and then in the department of Pescara). As regards the period after 1977, when the old station was closed down, no significant in- fluence was exerted on precipitation and tem- perature measurements, so the whole series was analysed without differentiating between the period prior to closure and the period after it. According to a posthumous publication by de Horatiis (1967), the observations made in the 1875-1882 period were not included in those designated for publication, as they were carried out with old instruments, in different locations and in the early stages of the observer’s experi- ence, who was therefore not particularly confi- dent about their quality. 3.2. Seasonal analysis Seasonal analysis is performed here accord- ing to Fourier analysis, that provides the values of mean amplitude (A), phase (σ) of the 12 and 6 months harmonics. The confidence level of each harmonic is computed according to 1−exp⋅ ⋅[−(0.833 A/vpe)2], where vpe is the vector probable error, i.e. the scattering of individual harmonics from the mean A (Mazzarella, 1998). The harmonic is found to be confident at 95% (99%) level when A≥2.08 vpe (A≥3.00 vpe). To ascertain the stationarity of the annual har- monic over the entire investigated interval, the same seasonal analysis is here applied to two non-overlapping subintervals (1883-1940, 1941- 2000) and then the results compared. log log D N N t t i i i i 1 1 = = + + / . log log 2 3 0 6309 = = _ _ i i (3.2) (3.2) where Ni is the number of events occurring in the interval of size ti. Different Cantor-like sets with the other fractal dimension D can be obtained simply by starting from different values of Ni and ti. To compute the fractal time clustering of a cata- logue, divide the time interval t0, over which the N events occur, into a series of n smaller in- tervals of length t=t0/n, with n =2,3,4... and compute the fraction R = N/n of intervals of length t occupied by events. If the distribution of events has a fractal structure then 3.1. Long-term analysis To construct a Cantor set, start with the closed interval [0,1]; remove the middle third leaving two closed intervals each with a length 1/3; con- tinue by deleting the middle third of these two intervals, leaving four closed intervals of length 1/9. The process of removing the open middle third of the remaining closed intervals is repeat- ed an infinite number of times and defines the Cantor set. This set shows a scale-invariant clus- tering quantified by the fractal dimension D 3.1. Long-term analysis Long-term analysis is performed here ac- cording to the least square regression line be- tween the N available yearly values of each in- vestigated parameter and time, only when it is found to be confident at a level not less than 1691 Michela Izzo, Pietro P.C. Aucelli and Adriano Mazzarella 95% (than 99%), i.e. when the following rela- tionship is verified (Mazzarella, 1998): 95% (than 99%), i.e. when the following rela- tionship is verified (Mazzarella, 1998): ent scales in the same manner, the major variable being the fractal dimension which is used as a measure of the nature of the phenomenon. The available daily rainfall and extreme air tempera- ture events measured in Agnone over a period of 80 years provide a wide range of available time- scales. This is why the fractal approach to time clustering of daily rainfall and extreme air tem- perature catalogue is here based upon a box- counting algorithm, like the Cantor dust method. To construct a Cantor set, start with the closed interval [0,1]; remove the middle third leaving two closed intervals each with a length 1/3; con- tinue by deleting the middle third of these two intervals, leaving four closed intervals of length 1/9. The process of removing the open middle third of the remaining closed intervals is repeat- ed an infinite number of times and defines the Cantor set. This set shows a scale-invariant clus- tering quantified by the fractal dimension D ent scales in the same manner, the major variable being the fractal dimension which is used as a measure of the nature of the phenomenon. The available daily rainfall and extreme air tempera- ture events measured in Agnone over a period of 80 years provide a wide range of available time- scales. This is why the fractal approach to time clustering of daily rainfall and extreme air tem- perature catalogue is here based upon a box- counting algorithm, like the Cantor dust method. ( ) ( ) . ( . ) ln N r r 3 2 1 1 1 96 2 58 $ $ - + - (3.1) where r is the correlation coefficient between the series of N annual values of each meteoro- logical series versus time. 3.3. Fractal daily clustering analysis Analysis performed on 120 years of data showed the existence of dishomogeneity be- tween daily values for the period 1883-1921. We therefore decided to postpone the daily analysis of this time interval and limit fractal analysis on- ly to the period 1922-2000. Fractals (Mazzarella, 1998) are sets that are not topological. For sets that are topological the Hausdoff-Besicovitch di- mension D is an integer (0 for points, 1 for any curve, 2 for any surface). For sets that are fractals D is not an integer but a real number in which the value of the whole number describes the topo- logical nature of the data set under consideration and the size of the decimal fraction represents the irregularity exhibited within the data. The es- sential feature of fractal sets is their scale invari- ance. The basic concept of a fractal distribution is that a phenomenon will be repeated on differ- ) D - R Ct (1 = (3.3) (3.3) or, equivalently, on a log-log scaled plane ( ) ( ) ( ) log log R C D t 1 = + - (3.4) (3.4) where C is a constant of proportionality (Maz- zarella, 1998, 1999). The power law (3.4) is further verified to be significantly different from the relative uniform distribution in which the same number N of events is equally spaced in time (no time clustering). As a mathematical representation, the relation (3.4) 1692 Recent changes in rainfall and air temperature at Agnone (Molise – Central Italy) could be valid over an infinite range; however, for physical applications, there will be upper and low- er limits on the applicability of the fractal distri- bution (scaling region). The fractal dimension D=1−s is estimated from the slope s that pro- vides the best fitting of the least square regression line of log(R) on log(t) when it is found to be con- fident at a level not less than 99%, i.e., when the relationship (3.1) is verified, with r representing the correlation coefficient between log(R) and log(t). It is worth noting that an increasingly clus- tered distribution of events causes an increase of D up to the limiting value D=1 that is the topo- logical dimension of a line; on the other hand, a concentration of events at a definite time corre- sponds to the limit D=0, i.e., to the dimension of a point. 3.3. Fractal daily clustering analysis The fractal dimension characterises the data distribution globally and, in general, does not reflect the situation in the different time subsets. Thus, to investigate the change in D with time, one can calculate D for catalogue subsets with a fixed number of events or choose a fixed-length time window and calculate consecutive values of D shifting the window through the catalogue. Here both options are adopted using windows containing no less than 100 events, the minimum for sound fractal analysis (Smith, 1988). variability (fig. 2) and the least square regres- sion analysis on time shows no significant time trend. Seasonal analysis of the available month- ly values (table I) shows that the annual and semiannual harmonics are determined at a level greater than 99%, with maxima occurring at the beginning of January for the annual wave and at the beginning of March for the semiannual wave. Analysis performed on the two non-over- lapping subintervals (1883-1940 and 1941- 2000) shows a small increase in the amplitude of annual harmonic. To compute the daily clustering of rainfall according to the Cantor dust method, fraction R (the fraction of time intervals including a rainy day) is computed as a function of the interval t, on a log-log scaled plane, for different intensi- ty thresholds T. The smallest time interval is chosen to be 1 day, gradually increased by a factor of 2. Fig. 2. Interannual variation of rainfall measured at Agnone in the 1883-2000 interval. 4.1. Rainfall 4.1. Rainfall The yearly values of rainfall measured in Agnone (1883-2000) show a wide interannual Fig. 2. Interannual variation of rainfall measured at Agnone in the 1883-2000 interval. Table I. Seasonal analysis of monthly rainfall measured in Agnone for the 1883-2000 interval and for the two different non-overlapping subintervals (1883-1940, 1941-2000). A, vpe and σ represent the mean amplitude, the vector probable error and the phase of the harmonic, respectively. Table I. Seasonal analysis of monthly rainfall measured in Agnone for the 1883-2000 interval and for the two different non-overlapping subintervals (1883-1940, 1941-2000). A, vpe and σ represent the mean amplitude, the vector probable error and the phase of the harmonic, respectively. Table I. Seasonal analysis of monthly rainfall measured in Agnone for the 1883-2000 interval and for the two different non-overlapping subintervals (1883-1940, 1941-2000). A, vpe and σ represent the mean amplitude, the vector probable error and the phase of the harmonic, respectively. Interval Total mean monthly value Period A vpe σ 1883-2000 75.7 mm 12 months 23.6 mm 2.6 81.8° 6 months 12.6 mm 2.2 177.0° 1883-1940 75.1 mm 12 months 21.1 mm 4.0 78.2° 6 months 14.1 mm 3.0 189.9° 1941-2000 76.2 mm 12 months 26.0 mm 3.5 84.6° 6 months 12.0 mm 3.1 162.1° 1693 Michela Izzo, Pietro P.C. Aucelli and Adriano Mazzarella Fig. 3. The logarithm of the fraction R of time intervals t (in days) including the rainfall events for Agnone Ob- servatory, as a function of log(t) (interval: 1922-2000); the values relative to daily rainfall events equal to or greater than 0.2 mm, 5 mm, 10 mm, 20 mm, and 30 mm, respectively, are plotted from the top of the graph down- wards. Fig. 3. The logarithm of the fraction R of time intervals t (in days) including the rainfall events for Agnone Ob- servatory, as a function of log(t) (interval: 1922-2000); the values relative to daily rainfall events equal to or greater than 0.2 mm, 5 mm, 10 mm, 20 mm, and 30 mm, respectively, are plotted from the top of the graph down- wards. Table II. Fractal dimension and scaling region val- ues of the rainfall recorded in Agnone in the 1922- 2000 interval, found to be confident at a level not less than 95% for different intensity threshold values. tensity-dependent dimensions. 4.1. Rainfall Some typical plots of log(R) on log(t) for different intensity thresholds (T ≥0.2 mm, T ≥5 mm, T ≥10 mm, T ≥20 mm, T ≥30 mm) are reported in fig. 3. Table II shows the relative values of fractal di- mension and of specific scaling region, found to be confident at a level not lower than 95% and significantly different from the relative uniform distribution. The time variation of D, computed according to both different time variable win- dows enclosing successive daily events and to a time fixed windows, show substantial stability over the period 1922-2000 with no significant time trend. Intensity Fractal Scaling threshold dimension region ≥0.2 mm 0.5 ±0.1 1-8 days ≥5 mm 0.31±0.08 1-8 days ≥10 mm 0.25±0.07 1-16 days ≥20 mm 0.16 ±0.05 1-32 days ≥30 mm 0.10 ±0.03 1-64 days The fractal dimension and the specific scal- ing region of the daily rainy events is found to be strongly dependent on the chosen intensity threshold: the higher this value, the lower the fractal dimension and the wider the relative scaling region. As a result, the rainfall process is characterised by an infinite hierarchy of in- 4.2. Air temperature The yearly values of minimum and maxi- mum air temperature measured at Agnone (1883-2000) show wide interannual variability and least square regression analysis on time 1694 1694 Recent changes in rainfall and air temperature at Agnone (Molise – Central Italy) Recent changes in rainfall and air temperature at Agnone (Molise – Central Italy) Recent changes in rainfall and air temperature at Agnone (Molise – Central Italy) shows an increasing time trend equal to 1.3 ± 0.4°C per 100 years, confident at least at 95% level, only for Tmin (fig. 4). Seasonal analysis of the monthly values of Tmax and Tmin (table IIIa,b) shows that the annual and semian- nual harmonics are determined at a level greater than 99%, albeit with a large dominance of the annual wave whose maximum occurs at the end of July. Analysis performed on the two non- overlapping subintervals (1883-1940 and 1941- 2000) shows no significant differences. Fig. 4. Interannual variation of minimum air tem- perature measured at Agnone in the 1883-2000 inter- val. The line represents the least square regression line of Tmin on time, calculated at 95% level of confi- dence. The continuous daily series of Tmin and Tmax measured at Agnone has been transformed into a daily series of events simply considering dai- ly values of Tmin lower than 5°C, 0°C and −5°C and daily values of Tmax higher than 25°C and 30°C. Hence, the same fractal methodology ap- plied to rainy days is applied to lowest and highest daily minimum and maximum air tem- perature, respectively. As regards minimum temperature, the lower this value, the lower the fractal dimension; for maximum temperature, the higher the threshold value, the lower the fractal dimension and the wider the relative scaling region (fig. 5). Fig. 4. Interannual variation of minimum air tem- perature measured at Agnone in the 1883-2000 inter- val. The line represents the least square regression line of Tmin on time, calculated at 95% level of confi- dence. g g g Table IV shows the relative fractal dimen- sion and scaling region values obtained at a lev- el of confidence not lower than 99% and signif- icantly different from the relative uniform dis- Table IIIa,b. Seasonal analysis of monthly maximum (Tmax) (a) and minimum (Tmin) (b) air temperature meas- ured in Agnone for the 1883-2000 interval and for the two different non-overlapping subintervals (1883-1940, 1941-2000). 4.2. Air temperature A, vpe and σ represent the mean amplitude, the vector probable error and the phase of the harmon- ic, respectively. Table IIIa,b. Seasonal analysis of monthly maximum (Tmax) (a) and minimum (Tmin) (b) air temperature meas- ured in Agnone for the 1883-2000 interval and for the two different non-overlapping subintervals (1883-1940, 1941-2000). A, vpe and σ represent the mean amplitude, the vector probable error and the phase of the harmon- ic, respectively. 1941 2000). A, vpe and σ represent the mean amplitude, the vector probable error and the phase of the harmon ic, respectively. Interval Total mean monthly value Period A vpe σ 1883-2000 15.7 °C 12 months 9.5 °C 0.1 −113.9° 6 months 0.9°C 0.1 −36.4° 1883-1940 15.7°C 12 months 9.4°C 0.2 −114.1° 6 months 0.9°C 0.1 −37.2° 1941-2000 15.6°C 12 months 9.5°C 0.2 −113.7° 6 months 0.9°C 0.1 −35.6° 1883-2000 7.9°C 12 months 7.6°C 0.1 −117.4° 6 months 0.4°C 0.1 −29.7° 1883-1940 7.3°C 12 months 7.5°C 0.1 −118.2° 6 months 0.4°C 0.1 −30.9° 1941-2000 8.4°C 12 months 7.6°C 0.1 −116.7° 6 months 0.5°C 0.1 −28.8° a b b 1695 Michela Izzo, Pietro P.C. Aucelli and Adriano Mazzarella a b Fig. 5a,b. The logarithm of the fraction R on t (in days) (interval: 1922-2000) including: a) the daily minimum temperature events for Agnone Observatory as a function of log(t); the values relative to daily events of Tmin low- er or equal to 5°C, 0°C and −5°C respectively, are plotted from the bottom of the graph upwards; b) the daily maximum temperature events for Agnone Observatory as a function of log(t); the values relative to daily events with Tmax greater than or equal to 25°C and 30°C respectively, are plotted from the bottom of the graph down- wards. a b b a Fig. 5a,b. The logarithm of the fraction R on t (in days) (interval: 1922-2000) including: a) the daily minimum temperature events for Agnone Observatory as a function of log(t); the values relative to daily events of Tmin low- er or equal to 5°C, 0°C and −5°C respectively, are plotted from the bottom of the graph upwards; b) the daily maximum temperature events for Agnone Observatory as a function of log(t); the values relative to daily events with Tmax greater than or equal to 25°C and 30°C respectively, are plotted from the bottom of the graph down- wards. 6. Conclusions Reconstruction of the instrumental climatic series constitutes a fundamental step in under- standing climatic evolution. The importance of reconstructing a historical series such as that of Agnone lies in the fact that 120 years of climat- ic history constitutes a valid starting point for comparisons with results from the analysis of other stations, both within and outside the re- gion, and for setting up other analytical tech- niques which allow us to investigate the history of climatic trends more deeply. The time variation of the fractal dimension of daily rainfall does not reveal any significant trend for any intensity threshold, which sug- gests that Agnone exhibits no trend towards the climatic tropicalization which seems to be characteristic of the coastal areas (Mazzarella, 1999; Brunetti et al., 2000). The time variation of the fractal dimension shows a decrease, confident at the 99% level, on- ly in minimum air temperature lower than −5°C. This means that in Agnone there has been a sig- nificant reduction in the occurrence of lowest temperatures. This seems to be perfectly com- patible with a world which is heating up increas- ingly. According to various estimates, there has been a global increase in surface mean tempera- ture between 0.3 and 0.6°C in the last century (data from NCDC; Jones et al., 1991). In this sense, the Agnone series is also an important point of reference for further analysis which will involve other long-term stations in Molise: it could be useful to identify the heat- island effect in the Campobasso series, as Cam- pobasso, more than other stations in Molise, should allow us to clearly underline and quanti- fy it. Analysis of the Agnone series produced important results particularly if they are insert- ed into a wider context, enabling us to discover similarities and differences with other places. The asymmetric trend of the minimum and maximum temperatures, underlined, among oth- er thing by the decreasing trend in daily thermal excursion, may have a major effect on local ecosystems, given the physiological effects which the minimum temperature has on living beings. Consequently, it would be well worth in- vestigating the contemporary evolution of vege- tation at Agnone, in order to discover any changes that may be related to climatic evolution. Application of fractal analysis confirmed the validity of this method in describing precipitation events and temperature extremes and enhancing our knowledge of them. Recent changes in rainfall and air temperature at Agnone (Molise – Central Italy) As regards precipitation and maximum air temperature, the higher the intensity threshold, the lower the fractal dimension and longer the relative scaling region. The opposite holds for minimum temperature: the lower the threshold value, the lower the fractal dimension. Thus the different intensity thresholds reveal different properties of the physical process: when using a higher threshold, the low-intensity parts are «cut-off» and only the properties of the high-in- tensity events (the opposite occurs for Tmin) are taken into account. This gives evidence of the so-called multifractality in which a multiplica- tive cascade process is responsible for the con- centration of energy fluxes into successively smaller parts of the atmosphere (Lovejoy and Schertzer, 1990; Olsson et al., 1993). A preliminary analysis conducted on other stations in Molise, shows a general tendency to- wards an increase in temperature, particularly in the minima. However, the situation through- out Molise is by no means uniform due to the fact that, besides stations which experienced substantial stability in the period examined, such as Chiauci, there are others, such as Cam- pobasso (the Region’s capital) and Guardiare- gia which show positive trends, even more so than Agnone, both in mean temperature and minimum temperature. 5. Discussion Table IV. Fractal dimension and scaling region val- ues of daily extreme air temperatures recorded in Agnone from 1922 to 2000, found to be confident at a level not less than 99% for different threshold values. Long-term analysis shows no time trend for rainfall and maximum air temperature and a sta- tistically significant increase in yearly values of minimum air temperature equal to 1.3 ± 0.4°C per 100 years. Such values are more accurate than rainfall and maximum air temperature as they are measured at the end of the night, in the absence of convective motions. Monthly rainfall and extreme air temperature show high seasonal variation that remains substantially unchanged over the entire interval of observation. Fractal analysis gave results which agree with those ob- tained in other studies (Mazzarella, 1999). level not less than 99% for different threshold values. Intensity Fractal Scaling threshold dimension region ≤−5°C 0.34±0.05 1-128 days ≤0°C 0.64±0.03 1-128 days ≤5°C 0.84±0.01 1-64 days ≥25°C 0.74±0.03 1-64 days ≥30°C 0.46±0.02 1-128 days Variability in D with the threshold may pro- vide insights into the nature of rain and the ther- mal field over Agnone, as the observed catalogue of rain and extreme air temperature also reflects the intermittency of the field, i.e. its three-di- mensional structure organised in eddies and sub- eddies and changeable in time for the same points. Research in intermittency (Berndtsson and Niemczynowicz, 1988; Fraedrich and Larn- der, 1993; Lovejoy and Schertzer, 1999; Maz- zarella, 1999) shows that cascade processes, where the large scale multiplicatively modulates the small, when carried out over wide enough ranges of scales with a repeating (scale invariant) mechanism, generally tend to multifractal struc- tures. tribution. Investigation of the time variation of D shows, over the entire 1922-2000 interval, no significant time trends for Tmax with a substan- tial stability of the value of D and a decreasing time trend for Tmin, representative of an ever rar- er occurrence of lowest values of minimum air temperature. The different trend of the minimum tempera- ture in comparison with the maximum is also confirmed by the time analysis of the daily ther- mal excursion (the difference between the daily maximum temperature and the daily minimum temperature) averaged over the period of a year. It shows a decreasing rate of 1.3 ±0.7°C per 100 years, confident at a level just below 99%. 5. Discussion 1696 Recent changes in rainfall and air temperature at Agnone (Molise – Central Italy) Recent changes in rainfall and air temperature at Agnone (Molise – Central Italy) REFERENCES pp mos. Sol.-Terr. Phys., 57 (8), 835-845. BARRIENDOS, M., J. MARTÍN-VIDE, J.C. PEÑA and R. RO- DRIGUEZ (2002): Daily meteorological observations in Cádiz – San Fernando. Analysis of the documentary sources and the instrumental data content (1786-1996), Climatic Change, 53, 151-170. y LEAN, J. and D. RIND (1999): Evaluating sun-climate rela- tionships since the Little Ice Age, J. Atmos. Sol.-Terr. Phys., 61 (1/2), 25-36. y LOVEJOY, S. and D. SCHERTZER (1990): Multifractals, univer- sality classes and satellite and radar measurements of clouds and rain fields, J. Geophys. Res., 95, 2021-2034. g BERNDTSSON, R. and J. NIEMCZYNOWICZ (1988): Spatial and temporal scales in rainfall analysis – Some aspects and future perspectives, J. Hydrol., 100, 293-313. MAUGERI, M., L. BUFFONI and F. CHLISTOVSKY (2002a): Daily Milan temperature and pressure series (1763- 1998): history of the observations and data and meta- data recovery, Climatic Change, 53, 101-117. BRUNETTI, M., M. MAUGERI and T. NANNI (2000): Variations of temperature and precipitation in Italy from 1866 to 1995, Theor. Appl. Climatol., 65, 165-174. 1995, Theor. Appl. Climatol., 65, 165-174. MAUGERI, M., L. BUFFONI, B. DELMONTE and A. FASSINA (2002b): Daily Milan temperature and pressure series (1763-1998): completing and homogenising the data, Climatic Change, 53, 119-149. CAMUFFO, D. (2002a): Calibration and instrumental errors in early measurements of air temperature, Climatic Change, 53, 297-329. in early measurements of air temperature, Climatic Change, 53, 297-329. g CAMUFFO, D. (2002b): Errors in early temperature series arising from changes in style of measuring time, sam- pling schedule and number of observations, Climatic Change, 53, 331-352. MAZZARELLA, A. (1998): The time clustering of floodings in Venice and the Cantor dust method, Theor. Appl. Cli- matol., 59, 147-150. MAZZARELLA, A. (1999): Multifractal dynamic rainfall processes in Italy, Theor. Appl. Climatol., 63, 73-78. g CAMUFFO, D. and P. JONES (2002): Improved understanding of past climatic variability from early daily European instrumental sources, Climatic Change, 53, 1-4. MITCHELL, J.F.B., T.C. JOHNS, J.M. GREGORY and S.F.B. TETT (1995): Climate response to increasing levels of greenhouse gases and sulphate aerosols, Nature, 376, 501-506. COCHEO, C. and D. CAMUFFO (2002): Corrections of sys- tematic errors and data homogenisation in the daily temperature Padova series (1725-1998), Climatic Change, 53, 77-100. NEY, E.P. (1959): Cosmic radiation and weather, Nature, 183, 451-452. DE HORATIIS, P.F.R. (1967): Osservazioni Meteoriche in Agnone del Sannio, 1881-1942 (Stabilimento Ti- pografico già Civelli, Firenze). OLSSON, J., J. NIEMCZYNOWICZ and R. 6. Conclusions Hence this study is a valid basis for further analysis, with the aim, amongst others, of in-depth investigation. By extending the analysis to other places, within and outside Molise, it will be possible to establish whether or not there is a zonal distri- bution of the change and, in the event of a pos- itive answer, try to determine the causes of the specific pattern. Furthermore, one of the main Furthermore, given the influence on evapo- ration and evapotranspiration rates, it cannot be excluded that in future this asymmetry in temperature trends could also affect precipita- tion. 1697 Michela Izzo, Pietro P.C. Aucelli and Adriano Mazzarella look at recent trends, in Greenhouse-Gas-Induced Cli- mate Change, edited by M.E. SCHLESINGER (Elsevier Press, Amsterdam), pp. 615. objectives is to go increasingly further back in time, in order to establish whether climatic change has always occurred in the same way and with the same structure. KELLY, P.M. and T.M.L. WIGLEY (1992): Solar cycle length, greenhouse forcing and global climate, Nature, 360, 328-330. LASSEN, K. and E. FRIIS-CHRISTENSEN (1995): Variability of the solar cycle length during the past five centuries and the apparent association with terrestrial climate, J. At- mos. Sol.-Terr. Phys., 57 (8), 835-845. REFERENCES BERNDTSSON (1993): Fractal analysis of high-resolution of rainfall time se- ries, J. Geophys. Res., 98, 23265-23274. PALUMBO, A. and A. MAZZARELLA (1984): Local recent changes in extreme air temperature, Climate Change, 6, 303-309. DEMARÉE, G.R., P.-J. LACHAERT, T. VERHOEVE and E. THOEN (2002): The long-term daily central Belgium temperature (CBT) series (1767-1998) and early in- strumental meteorological observations in Belgium, Climatic Change, 53, 269-293. ROZELOT, J.P. (2001): Possible links between the solar ra- dius variations and the Earth’s climate evolution over the past four centuries, J. Atmos. Sol.-Terr. Phys., 63, 375-386. g FRAEDRICH, K. and C. LARNDER (1993): Scaling regimes of composite rainfall time series, Tellus, 45A, 289-298. FRIIS-CHRISTENSEN, E. and H. SVENSMARK (1997): What do we really know about the sun-climate connection?, Adv. Space Res., 20 (4/5), 913-921. SANTER, B.D., K.E. TAYLOR, T.M.L. WIGLEY, J.E. PENNER, P.D. JONES and U. CUBASH (1995): Towards the detec- tion and attribution of an anthropogenic effect on cli- mate, Climate Dyn., 12, 77-100. p HALGH, J.D. (1994): The role of stratospheric ozone in modulating the solar radiative forcing of climate, Na- ture, 370, 544-546. SMITH, L.A. (1988): Intrinsic limits on dimension calcula- tion, Phys. Lett. A, 133, 283-288. HEGERL, G.C., K. HASSELMANN, U. CUBASCH, J.F.B. MITCHELL, E. ROECKNER, R. VOSS and J. WASZKEWITZ (1997): Multi-fingerprint detection and attribution analy- sis of greenhouse gas, greenhouse gas-plus-aerosol and solar forced climate change, Clim. Dyn., 13, 613-634. STOUFFER, R.J., S. MANABE K. and YA. VINNIKOV (1994): Model assessment of the role of natural variability in recent global warming, Nature, 367, 634-636. solar forced climate change, Clim. Dyn., 13, 613-634. (received May 13, 2003; accepted July 19, 2004) (received May 13, 2003; accepted July 19, 2004) g y JONES, P. A., T.M.L. WIGLEY and G. FARMER (1991): Marine and land temperature data sets: a comparison and a 1698
https://openalex.org/W4230221772	https://zenodo.org/records/1638567/files/article.pdf	English	null	Separate dwellings	Public health	1,900	public-domain	5,799	Read before the ~Ietropolitan ]3ranch of the Incorporated Society of Medical Officers of Health, December, 1899. i b , , p ( p , ) 9 ~ For the history of events leading to this enactment, see "Tenementcd Buildings and Duties on Inhabited Houses," Jot~rnal of the Sanitary Institltte, part i., eel. xix., 1898. Separate Dwellings Separate Dwellings 309 r.b~a-'V, ~ Read before the ~Ietropolitan ]3ranch of the Incorporated Society of Medical Officers of Health, December, 1899. ~" See PUnLIC HE.~LTH, xi., p. 822 (September, 1899). 9 ~ For the history of events leading to this enactment, see "Tenementcd Buildings and Duties on Inhabited Houses," Jot~rnal of the Sanitary Institltte, part i., eel. xix., 1898. , , " See PUnLIC HE.~LTH, xi., p. 822 (Septemb Read before the ~Ietropolitan ]3ranch of the Incorporated Society of Medical Officers of Health, December, 1899. ~" See PUnLIC HE.~LTH, xi., p. 822 (September, 1899). 9 ~ For the history of events leading to this enactment, see "Tenementcd Buildingsand Dutieson Inhabited Houses " Jot~rnal of the Sanitary Institltte , , ~" See PUnLIC HE.~LTH, xi., p. 822 (September, 1899). 9 F h hi f l di hi BY JOHN F. J. SYKES, D.Sc., ~I.D., Medical Officer of IIealth, St. rancras, and Lecturer on Public IIealth, Guy's Hospital. Each floor contains a separate sitting-room, bedroom, kitchen, water-closet, and larder, and there is a separate coal cellar for the occupiers of each floor in the yard at the back of the house. The occupier of each floor has the use on specified days of a common scullolT or ~ash-houso on the ground-floor and of a common front entrance and hall. The tenants of the upper floors also have common use of the staircase, and all the rooms in each house open direct on to the landing or floor upon which they are situated, and none of the rooms communicate with one another. The Attorney-General, for tlle respondent, the Surveyor of Taxes oil behalf of the Crown, said it must be remembered that the exemption in the modern Acts was engrafted on the terms of I~ule 6 of Schedule B of 48 George III., cap. 55, which provides that "where any house shall be lot in different storys, tenements, lodgings, or landings, and shall be inhabited by two or more persons or families, the same shall nevertheless be subject to and shall in like manner be charged to the said duties as if such house or tenement was inhabited, by one person or family only, and the landlord or owner shall be deemed the occupier of such dwelling- house, and shall be charged to the said duties." The short point was that a structurally separate dwelling" was necessary, as distin- guished from separate floors. The test was the calJability of each grolq~ of rooms to be shut o[]" from the rest. Ilere each room~not each set of rooms~communicatcd with a common staircase. The intentidn with which they were erected "only becomes important after it has been found that the dwellings are separate in fact. Here the Commissioners found as a fact that the dwellings were not separate. The Attorney-General could not have put the c~so more cle,~rly, and the 3ustices on appeal could not have rejected the conten~mn more decidedly, when they both expressed the opinion tliat Rule 6, Schedule B, ,t8 George III., cap. 55, must be set aside, and did no~ govelll the question of assessing to inhabited house duty a house built or adapted and used for providing separate dwellings at rents not exceeding .s a year if the medical officer of health give a certificate. BY JOHN F. J. SYKES, D.Sc., ~I.D., Medical Officer of IIealth, St. rancras, and Lecturer on Public IIealth, Guy's Hospital. ThE case of Scamal~ v. Lce,t heard in the Queen's Bench Division on l~Iarch 23rd, 1899, before ~t:. Justice Grantham and ~Ir. Justice Kennedy, again directs our attention to the liousing of the working classes, and especially to the interpretation to be placed upon the words "separate dwellings." In this case the owner of certain houses claimed exemption from the inhabited house duty under the Customs and Inland I~evenue Act, 1890 (53 and 54 u cap. 8), section 26, subsection 2, and the Act of 1891 (54 and 55 Vie., cap. 25), section 4, the former of which enacts ~ that "the assessment to inhabited house duty of any house. originally built or adapted by additions or alterations and used for the sole purpose of providing separate dwellings, where the annual value of each dwelling shall not amount to s shall be discharged by the Commissioners for executing the Acts relating to the inhabited house duties, provided Shat a certificate of the medical officer of health for the district in which the hot'so is situate . . . shall be produced to them to the effect that the hotistr is so constructed as to afford suitable accommodation for each of the families or persons inhabiting it, and that due provision is made for theii" sanitary requirements...." The medical officer of health certified that the houses were (at the date of his examination) "constructed so as to afford suitable accommodation for each of the families or persons who were repre- sented to me as then inhabiting the said houses respectively, and that due provision was made for their sanitary requirements." There was no definition of what was meant by these words, but there was added in a note: "The above certificate is not to be taken as implying any opinion on the par~ of the medical officer of health that the houses in question are 'built or adapted by addi- tions or alterations and Used for the sole purpose of providing separate dwellings,' his' examination having led him to a contrary opinion." Separate Dwellings 3 l 0 ,P.bu~ a~ath The construction of the houses is as follows : Each house has a front-door, kept closed as in an ordinary house, but separate bolls are provided for the t'enants on the first and second floors re- spectively, and a knocker for the use of the tenant on the ground- floor. BY JOHN F. J. SYKES, D.Sc., ~I.D., Medical Officer of IIealth, St. rancras, and Lecturer on Public IIealth, Guy's Hospital. If the certificate of the medical officer of health be not produced to the CoInmissioners, then 1Rule 6 does apply, ancl the sep,~rate dwellings in the house, instead of being assessed separately and escaping house duty or paying a reduced duty, are assessed Separate Dwellings 311 rob~'-y, l~j rob~'-y, l~j together, and full duty is payable on the whole house, unless the dwellings, on aecount of being situated on a common stair ~ outside the main building or open to the outer air, and of being each entered separately by one entrance, or,front-door, are regarded by the Commissioners as separate houses, and assessed separately. p , p y The word "separately" occurs in sections 71 and 72, the Public tlealth Act, 1875, and in section 96, the Public Health (London) Act, 1891, making it unlawful to let or occupy scl~arately as a dwell- i~g any underground room as defihed in the sections. This is held to mean separately from that part of the house which is 11ot under- ground, and subsection 6 of section 96 of the Public Healtll (London) Act, 1891, also expressly enacts tha~ if two or more underground rooms are occupied together, each of them shall be deemed to be separately occupied, also meaning separately occupied from the rest of the house. This is interesting, as indicating the meaning of separately with regard to the letting or occupationr-- that is, the usage of that particular part of the house, namely, the basement floor. Subsection 8 of the same section characterizes a dwelling as a sleeping place. It must be observed that as there may be one or more dwellings separately occupied on the basement floor, therefore the "floor" of the house does not constitute the separation ; in other words, it is a question of separate usage, and not of separate construction, that constitutes the separation. For th~ economy of space effected by building common stair dwcll~gs, see lock Dwellings, Sectmn of Arolntecture, vol. w., T.ransachons of trite r- national C~o~gres~ of Hyglcne, 1891. BY JOHN F. J. SYKES, D.Sc., ~I.D., Medical Officer of IIealth, St. rancras, and Lecturer on Public IIealth, Guy's Hospital. Let us examine more closely the words of the Act in accordance with which the medical officer of health has to certify ; they are : "That the house is so constructed as to afford suitable accom- modation for each of the families or persons inhabiting it, and that due provision is made for their sanitary requirements." This is a question of construction and provision for separate families or persons in the same house, and the duty is cast upon the medical officer of health of deciding what is suitable and due in the con- strueli,~ of a hous8 so used. Let us also examine more closely the words describing the kind of house to which such a certificate applies--" any house originally built or adapted by additions or alterations and used for the sole purpose of providing separate dwellings" at a certain annual value. Here again we have placed first the construction, and, secondly, the usage in accordance with the construction. Now, the'High Cot~rt has decided, as a question of law, that a separate dwelling merely means a dwelling separately occupied ; in Separate Dwellings 312 [r~buo ae~ta 'other words, that separate ztsagc constitutes the separation of dwell- ings in a house, and therefore that a "house let in separate dwell- ings" is synonymous with a "dwelling-house let in separate tene- ments." Consequently, this decision throws upon medical officers of health the responsibility of deciding the question of the construc- tion of dwelling-houses used in separate tenements, and whether a dwelling-house in which the separate tenements are not struc- turally severed, as well as those in which they are structurally severed, is or is not so constructed as to afford suitable accommoda- tion and provide due sanitary requirements, and if not, what accom- modation and sanitary requirements should be provided. y q p If it is allowed that a house so constructed that every room opens separately on to the staircase, as in an ordinary dwelling-house constructed for one family, is a house affording suitable accom- modation for two or more families, then any such house could be adapted or altered, and it would only remain to decide what addi- tions or alterations would entitle such a house to a certificate. BY JOHN F. J. SYKES, D.Sc., ~I.D., Medical Officer of IIealth, St. rancras, and Lecturer on Public IIealth, Guy's Hospital. In order to be clear, we must define what we are dealing with, and we must have in mind that the definition of erections may be (a) according to structure, and (b) according to usage. According to structure, we have the terms " building," "house," ~ room." According to usage, we have the terms "tenement," "h~01)ita- tion," "dwelling," "lodging." B~dldi~g.--Any structure, provided it has some sort of attach- ment to the ground, and some sort of roof, appears to be a building in law. I:[oztse.--I.n law "a house is an enclosed space, with walls, a roof, and an entrance capable of being closed" (1881, Esdaile v. ~tIetropolitan, etc., Railway Compalty, Lands Clauses Act, 18,15, pcr Cave, ;l.; see 46 J.P., 103). According to the instructions issued for the guidance of enumerators at the Census of 1891, "all the space within the external and party walls of a building was to be con- sidered a separate house, by however many families, living in distinct tenements or apartments, it might be occupied." This last definition of a house is on the lines of Rule 6 of Schedule B of 48 George II L, cap. 55, above quoted, and under which the case of 41ttorltcy-Ge~teral v. Jlfutual 'l'ol~til~e lVestmil~ster Chambers Associa- ti,m, Limited, was decided. The recent, decision in the case of Rogers v. Hosegood, reported in the Times of November 8th, 1899, throws another light upon what is a house. Mr. Justice Farwell, in delivering judgment, said, "... I cannot see any. substantial difference, for the purposes of a covenant Separate Dwellings 313 r,bru,,rr. ~I r,bru,,rr. ~I of this nature, between a terrace of adjoining residences, separated from one another vertically, and a pile of residences separated from one another horizontally. If the building of tile defendant was carried no higher than the ground-floor, it would, in my opinion, be impossible to contend that each of the residences opening on to the inner area was not a separate house, and I cannot see that the superposition of seven other rows of residences can make any difference. This is a mere question of construction, on which a rating case such as that cited (Attorncy-Gci~cral v. ~'l[~lttml 7'ontinc lVcst- minster Chambers Association, Limited, 1 Ex. BY JOHN F. J. SYKES, D.Sc., ~I.D., Medical Officer of IIealth, St. rancras, and Lecturer on Public IIealth, Guy's Hospital. This view prevails at the revision of the lists of Parliamentary voters who claim upon a dwelling, residence, or abode; and in the Public Health (London) Act of 1891, as well as in the Public Health Act of 1875, an underground room in which a person passes the night is deemed to be occupied as a dwelling. Lodging.--It would appear that whereas a tenant, in consideration of the payment of'rent, may by agreement or lease obtain exclusive possession of another's property, and can maintain trespass, a lodger, although paying reng in consideration, has only the license to use, but not to exclusively occupy, and cannot maintain trespass ; so that a lodging ~ppcars to be dependent upon the absence,of an agreement or lease. In addition to these single words, certain terms compounded of these words are used~dwolling-room, dwelling-house, inhabited building, inhabited house, inhabited room, tenemen~ building, tenement house, tenement dwelling-house. The last is often abbreviated into tenement-house, meaning a dwelling-house let in separate tenements, and it may be assumed that the word in-. habited, or habitation,' does not necessarily imply dwelling. I~ would also appearthat a "house let in lodgings" includes a house let in separate dwellings, whethe]- self-contained, self.complete, or neither, in which the several occupiers dwell without agreement or lease to define their respective liabilities ; hence probably the reason for bye-laws for houses let in lodgings, as in any class of dwelling- house let in separate dwellings some part of the house is used in common by the inmates. The legal definition of a dwelling-house is most important. For the purpose of rating to the poor, the presence of a door shutting off the house from the street, and common to the use of two tenants, together with a common s~air, was adjudged not to substantially affect the occupation, which was for all practical purposes distinct, and the Court was of opinion that the appellant, who was non-resident, was entitled to have flats separately assessed for rating purposes (see Saunders v. St. Mary, Lambeth, Times, February 12th, 1891, County of London Sessions, before Sir P. H, Edlin, Q.C., Chairman, sitting at Clerkenwell). Also for the purpose of rating to the relief of the poor, a "dwelling- house" includes par~ of a house occupied as a separate dwelling which need not be structurally separated from the rest, and the occupants using a common w.e. (see Allchurch and Anotl~erv. BY JOHN F. J. SYKES, D.Sc., ~I.D., Medical Officer of IIealth, St. rancras, and Lecturer on Public IIealth, Guy's Hospital. D., 469) can have little bearing; but I would point out that in that case the Court held that the whole block was a house, but pointed out that each set of chambers might be a house also. The case of the domus mansionalis, cited at p. 477, has some bearing on the point .... But even if the whole structure could be regarded as a private residence only, it is difficult to see how thirty or forty different families and estab- lishments can find place therein, unless the owner of such entire messuage is carrying on the trade of letting apartments." This last paragraph of the decision is important as bearing upon those piles of artizans' dwellings constructed in separate and self-contained dwellings with outs!de staircases, which, in accordance with the judgment in the case ot ~ the Model Lodging House, Streatham Street, Bloomsbury, decided in 1853, are regarded as constructed in separate houses, and are each assessed sepai'ately by the Com- missioners of Inland I~evenue, and in consequence of which some doub~ is thrown upon the power of a sanitary authority to register such houses under the bye-laws for houses let in lodgings. The last paragraph of Mr. Justice Farwell's decision tends towards removing this doubt. This question is one that this Society should seriously consider in all its bearings, but it is not desirable to pursue it further here. Room.--Any apartment in a nouse, or any division separated from the rest by a partition, may be called a room. The elasticity of this word is considerable. Now as to usage: Tcnement.~In law a tenement is any species of permanent property that may be held, and may be the whole or part of a premises, building, or lmuse. In popular parlance, it has come to mean a room or rooms let to one family or person as a tenement in a hoffse let out in separate dwellings. IIabitation.--For our purpose this means httman habitation, and a place occupied for human habitation would appear to have a more extended meaning th~n a dwelling-place. 14 S D Separate Dwellings 314 ~buc He~azh ~buc He~azh Dwelling.~In law ~ho essential par~ of dwelling appears to be sleeping or passing the night. BY JOHN F. J. SYKES, D.Sc., ~I.D., Medical Officer of IIealth, St. rancras, and Lecturer on Public IIealth, Guy's Hospital. "The ~lssessment Committee of the Ilemlon Union, before Mr. Justice Separate Dwellings 315 Fem~,'y. 1~oj Fem~,'y. 1~oj A. L. Smith and Mr. Justice Grantham, reported in the Estates Gazette of April 18th, 1891, and an appeal before the Master of the I~olls and Lords Justices Bowen and Kay, reported in the Estates Gazette of July llth, 1891). I mention these rating cases as bearing upon the point, because originally a house was only taxed separately to the window-tax (which subsequently became inhabited house duty)when it was a separa~ dwelling-house, as defined by Rule 6 in the year 1808. This definition has had, or ts having, gradually lopped off from it, for one purpose or another, the private street-door and private staircase (Sa~nzders v. St. MaTy, Lambeth), the private w.c. (Allchz~rch v. Itelzdoll U~Liol~), the private wash-house (Seaman v. Lee), until nothing will be left but the single room. What will happen to the assessment of an old elght-roomed house, occupied by eigh~ families, one in each room, when it comes to be treated in the same manner as the new house containing eight dwellings of three or four rooms each, the tes~ in each case being confined to the single entry to each dwelling ? since local rating authorities no longer appear to take into consideration the outside or open stair, although the Commissioners of Taxes do so still. In one respect the two classes of authorities appear to be diverging in the same line--namely, in regarding a dwelling with but one,entrance or f~ont-door as "structurally severed," with less regard than formerly to what the dwelling con- rains, although in this respect in different localities the standard differs. But, as I know dwellings that have associated sinks and taps on balconies, and associated w.e.'s, as well as associated wash- houses, all of ~-hich dwellings have been exempted, and whose owners have not applied to me for certificates, I can only assume that no regardis paid as to whether a dwelling is self-contained, or self-complete, or associated. p Practically what happens is that, at the quinquennial assessment, the rating authority assesses separately all structurally severed dwellings, whether self-contained, self-complete, Or associated, and the Commissioners of Taxes lump together those in each compound house with an inside staircase. BY JOHN F. J. SYKES, D.Sc., ~I.D., Medical Officer of IIealth, St. rancras, and Lecturer on Public IIealth, Guy's Hospital. If the owner objects'to be taxed on the lump sum, he is referred to the medical officer of health, and the Board of Inland Revenue consider a fresh certificate of the medical officer of health indispensable at each quinquennial assess- ment; so that the medical officer of health's certificate is practically supplied for the purpose of overriding the open staircase if he is of opinion that it is a better sanitary arrangement to have it closed. If we analyze dwelling-houses, we shall see more clearly in. a 3 16 Separate Dwellings trabUc ae~th trabUc ae~th table of classification the relationship of usage and construction of tlle several kinds of dwellings to each other: table of classification the relationship of usage and construction of tlle several kinds of dwellings to each other: For an analysis of the essential requirements of a dwelling, see "DweIling Accommodation in Large Cities," Journal of the Sa~zitary Institute, p~rt i., vol. xx., 1899. CLASSIFICATION OF DWEI~LINC;-IIOUSFS. The simple dwelling-house (used as one dwelling). The compound dwelling-house (used as two or more separate dwellings). A. The dwellings structurally severed, and the staircase either (i.) ouCside, or (it.) inside, and if inside, either (1) per- manently open to the outer air, or (2) closed by windows or skylights, or not directly lighted. (a) The severed dwellings self-contained (i.e., each dwelling having only one entrance, and contain- ing withil\| its precincts, at the least, room for sleeping, living, cooking, draw-tap and sink, w.e., and wash-house or scullery that can be improvised as a wash-house)2: (b) The severed dwellings not self-contained, but self- complete (i.c., each scullery, w.c., or wash-house, although outside the dwelling, being reserved for use by tl~e occupiers of one dwelling only, and 9 kept under lock and key by tim tenant). (c) The dwellings not self-contained or self-complete, but associated. 1. With associated sculleries, or taps and sinks. 2. With associated w.c.'s. 3. With associated wash-houses, or coppers or troughs. g B. The dwellings not structurally severed, and tim staircases not outside or permanently open to the ou~r air. 1. With common intercommunication, or landings upon which each of the rooms open separately~e.g., made-down houses. 9,. With common landings and common living-rooms-- e.!/., (a) hotels, boarding-houses, residential clubs; (b) with separate cubicles in common dormitories or corridors--e.g., i~owton lodging-houses. 3. ~Vith common landings, common living-rooms, and common sleeping-rooms--e.g., common lodging- houses.- Separate Dwellings 317 rob~,~,, l~0j The Commissioners of Taxes require an outside or open stair to tile building, and a single entrance to each dwelling, as a condition of separate assessment of each.dwelling, in order to exempt from inhabited house duty--that is, a premium is allowed for such methdd of construction and adaptation, which sanitarily, as far as it goes, is a good condition. But the onus of certifying that such condition is unnecessary is thrown upon the medical officer of health--in short, he is placed in the invidious position of certifying that a lower sanitary standard should be accepted in this respect by the Commissioners than that which has hitherto guided them. CLASSIFICATION OF DWEI~LINC;-IIOUSFS. j When, added to all this, we are clamouring for open air, per- flation, and through-ventilation as the most efficicnb protection against tubercular diseases, especially consumption of the lungs, then, on the grounds of health, it is impossible to maintain that suitable accommodation and duo provision is made in a house let in separate dwellings unless the common stair is amply and permanently ventilated to the open air, preferably at.every floor. Section 69 of the London Building Act, 1894, provides that (1) in every b~ilding constructed or adapted to be occupied in separate tenements by more than two families, the principal stair- case used by the several families in common shall be ventilated upon every story abo{'o the ground-story by means of windows or skylights opening directly into the external air, or shall be other- wise adequately ventilated; (2) the principal staircase in every building being a dwelling-house, and not subject to the provisions of subsection 1 of this section, shall be ventilated by means of a window or skylight opening directly into the external air. This section is an attempt in the right direction, but does not meet the case sufficiently, since a closed window or skylight, over which everybody, and therefore nobody, has control, may be a means of lighting, but is an utterly unreliable means of continuous through.ventilation or perfiation, as such windows are invariably closed. Continuous perflation by means of permanent openings is the only method of construction of permanent value. In order to provide against back-to-back dwellings I must again insist, as I have done on several occasions before, that one of the conditions of a certificate should be that every dwelling of two or three rooms should have one of the rooms situated Jn" the opposite side of the building on which the other room or rooms are situated, and every dwelling of four rooms two of such rooms so situated, and that between each pair of front and back rooms there should be a door of communication for purposes of access and perflation, and no other means of entry to the other room be provided, other- wise by making the access from the common stair the dwellings may still be constructed as back-to-back dwellings, or simply con- verged into them by locking the door of intercommunication between th0 front and back rooms. CLASSIFICATION OF DWEI~LINC;-IIOUSFS. Referring to the table of classification, practically all those dwellings classed under "A "--that is, those dwellings structurally severed, or that have only one e~rance, provided they have an outside or open staircase--are exempted from inhabited house duty; but if they have an inside or enclosed staircase, they are referred to the medical officer of health. That this should be the only reason, as it actually is in fact, for referring the method of con- struction to medical officers of health, is sufficiently obvious to show the anomalous position of these officers, especially when I assure you that I know era block of absolutely back-to-back dwellings of two, three, and four rooms rewarded for their method of construc- tion by the remission of inhabited house, duty, no certificate being required, because the common stair is open to the air, and each dwelling has but one entrance. g Is a medical officer of health justified in giving a certificate to enable an owner to enclose an outside staircase ? or is he justified in giving a certificate that a house let in separate dwellings is properly constructed if the common staircase and landings being inside are not permanently open to perflation ? There are a number of reasons for requiring that the common stair should be through-ventilated. The fundamental objection to a closed ah'-shaft formed by the staircase is the common usage of the same air by a number of different families ; the effect of this is the spread of infectious diseases, and especially of those that attack children, and that are more often nursed at home. Another objec- tion is that the chimney of a dwelling-room, "~'hen a fire is burning in the grate, exerts an enormous suctional power, and every burning grate is an extractor, the air to which in a closed house with an enclosed staircase is supplied by the down-draughts of the chimneys of the unlighted fires. Hence in made-down tenemented houses, let out in one- and two-room dwellings, the doors of the rooms are frequently ajar, and the tenants nearly always complain of smoky chimneys. What else could be expected in an enclosed house with Separate Dwellings 318 f~bar ~,o~th all the fires burning at the same time? When, furthermore, tlle w.c.'s open on to the common staircase, so that air-supply to the house is drawn through them, the condition becomes still more unjustifiable. CLASSIFICATION OF DWEI~LINC;-IIOUSFS. The following diagram illustrates the three stages A, B, and C in dwellings of 2, 8, and 4 rooms : ~h~'V. z~l 319 .~ t Separate Dwellings /3. C. 1"cab '~'t .~ 319 ~h~'V. z~l To explain this diagram it must be premised that in a building two rooms deep with dwellings let in single roofias it would be im- possible to have the window of a single room opening on to both fronts of the building, and so the principle of through ventilation for single or odd rooms cannot be made to apply; the diagram therefore shows two (II.), three (III.), and four (IIII.) room dwell- ings (horizontally) in three stages, A, B, and C, (vertically). The "A" column shows a two, three, and four-roomed dwelling constructed respectively with only one entrance and back-to-back. The "B" column shows the application of the requirement that "every dwelling of two or three rooms should have one of ~ne rooms situated on the opposite front of the building on which the other room or rooms are situated, and every dwelling of four rooms two of such rooms, and that between each pair of front and back rooms there should be a door of communication," but it also shows bow, if two entrances are allowed, perflation or through ventilat!on may be defeated by locking the doorway between the front and back room. The "C" column shows the effect of requiring also that "no other means of entry to the other room be provided," since Separate Dwellings 320 [Publlo aealth on every occasion of entry and when the doorway is open more or less between the front and back room and the window is loose, ajar, or open, the oppprtunity for perflation is constantly recurring. There is also another advantage in these requirements in buildings when facing north and south : some of the rooms are then situated on bothsides of the building instead of being like those in the "A" column, the left-hand rooms of which when facing north are alto- gether deprived 05 sunlight as well as moving air. g p g g On February 12th, 1897, this Branch of the Society adopted certain recommendations with regard to certificates under the Customs and Inland l~ovenue Acts, 1890 and 1891, which recommehdations were published in PubLic HEALTH for March, 1897, vol. ix., p. 183. Feb~L~'. l~J Feb~L~'. l~J 1Referring now again to the table of classification, it will be seen that the High Court decision in the case of Seaman v. Lee covers, in addition to those dwellings contained under "A.," structurally severed, whether self-contained, self-complete, or associated, also those dwellings contained under "]3. 1," dwellings not structurally severed, and the stairs not permanently open to the air, such as many houses let in lodgings and made-down houses, or houses original!y constructed for'one family and occupied by many. Is it possible to ameliorate the condition of these made-down tenemented houses ? There are some of which it can be Said that it is neither possible nor desirable, others that are doubtful, and others, again, of which there may be possibly some hope that by additions and alterations an improved condition might be brought about. These houses have been constructed originally for one family--that is, vertically, and not horizontally as for more families than one. Therefore the problem would be: (1)the structural conversion of a vertical dwelling-house into a house of horizontal dwellings; (2) the cbst of such conversion; and (3) the recoupment towards the cost from the remission of inhabited house duty. The structural conversion, assuming the rooms to .have proper height and area, would involve the provision (1) of ample through- ventilation of the common stair; (9,) of ~hrough-ventilation to the dwellings from one front to the other; (8) of w.c. with access lobby on each floor (one w.c. for every twelve persons is already required in all houses by the L.C.C. bye-laws); (4) draw-tap and sink on each floor; (5) drainage in accordance with bye-laws and regulations; and (6)clothes-washing accommodation. The two last are frequently found already provided. 'Now, it/is obvious that by providing the first somewhat more completely and making the common stair permanently open to the air, and providing the second by making the dwellings to open only by one entrance on each floor, the owner might claim exemp- tion from inhabited house duty on th e ground that the dwellings are now separate houses, and render it unnecessary to go to the expense of any further provision with the object .of obtaining a certificate. CLASSIFICATION OF DWEI~LINC;-IIOUSFS. p p 9 In those recommendations no provision was made for continuous perflation or permanent through-ventilation of the common stair- case, landings, and passages. No provision was made for .perflation or through ventilation of the rooms, nor against back-to-back dwellings. No provision was made for cutting off a~rially the water-closets from the dwellings and from the common stair. No provision was made for the area of sleeping and living rooms, and an error was made in stating this to be defined by the London Building Act, 1804. In section,70, as to habitable rooms, there is no such provision, although there is a provision that every habit- able room except rooms in the roof shall be at least 8 feet 6 idches in height. For these reasons the recommendations need revision, and the remaining recommendations were : That there should be a suffici0nt and available supply of Water on every floor ; That there should be at least one water-closet, properly supplied with water, for every twelve occupants or less on every floor ; That the drainage of the premises should be in accordance with the regulations recognised by the authority in whoso jurisdiction the house is situated ; and That accommodation for clothes-washing should be provided, sufficient for the number of persons inhabiting the house. That accommodation for clothes-washing should be provided, sufficient for the number of persons inhabiting the house. It must not be assumed that these minimum requirements for a certificate cover the ground for buildings of the latest erection or to be hereafter erected. This ground is better covered for new 9 building~ by the London Building Acts, 1894 and 1898, and the special requiremon~z therein for artisans' dwelling, and the build- ing regulations for this class of buildings approved by the Lbndon County Council on December 8rd, 1889. 321 Separate Dwellings Feb~L~'. l~J That this possible course has not been adopted by owners tends to show that the expense of even the lesser structural conversion is greater than the recoupment of a few shillings or perhaps pounds annually from the remission of-inhabited house duty would financially justify.
https://openalex.org/W3209240662	https://europepmc.org/articles/pmc8575311?pdf=render	English	null	Schistosomiasis messaging in endemic communities: Lessons and implications for interventions from rural Uganda, a rapid ethnographic assessment study	PLoS neglected tropical diseases	2,021	cc-by	9,014	Editor: Anne W. Rimoin, University of California, Los Angeles, UNITED STATES Editor: Anne W. Rimoin, University of California, Los Angeles, UNITED STATES Editor: Anne W. Rimoin, University of California, Los Angeles, UNITED STATES Received: April 30, 2021 Accepted: October 10, 2021 Published: October 27, 2021 Editor: Anne W. Rimoin, University of California, Los Angeles, UNITED STATES Received: April 30, 2021 Accepted: October 10, 2021 Published: October 27, 2021 Agnes SsaliID1, Lucy PickeringID2, Edith Nalwadda1, Lazaaro Mujumbusi1,2, Janet SeeleyID1,3, Poppy H. L. LambertonID4,5 Agnes SsaliID1, Lucy PickeringID2, Edith Nalwadda1, Lazaaro Mujumbusi1,2, Janet SeeleyID1,3, Poppy H. L. LambertonID4,5 1 MRC/UVRI & LSHTM Uganda Research Unit, Entebbe, Uganda, 2 Institute of Health and Wellbeing, University of Glasgow, Glasgow, United Kingdom, 3 London School of Hygiene and Tropical Medicine, London, United Kingdom, 4 Institute of Biodiversity, Animal Health & Comparative Medicine, University of Glasgow, Glasgow, United Kingdom, 5 Wellcome Centre for Integrative Parasitology, University of Glasgow Glasgow, United Kingdom a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 * Agnes.Ssali@mrcuganda.org (AS); Poppy.Lamberton@glasgow.ac.uk (PHLL) OPEN ACCESS Citation: Ssali A, Pickering L, Nalwadda E, Mujumbusi L, Seeley J, Lamberton PHL (2021) Schistosomiasis messaging in endemic communities: Lessons and implications for interventions from rural Uganda, a rapid ethnographic assessment study. PLoS Negl Trop Dis 15(10): e0009893. https://doi.org/10.1371/ journal.pntd.0009893 Background Over 240 million people are infected with schistosomiasis, the majority in sub-Saharan Africa. In Uganda, high infection rates exist in communities on the shores of Lake Victoria. Praziquantel mass drug administration (MDA) delivered by village health teams is the main- stay of schistosomiasis control. However, treatment uptake remains suboptimal, with many people unaware of treatment or thinking it is only for children. Furthermore, people are often rapidly reinfected post-treatment due to continued exposure. In three Schistosoma mansoni high endemicity lake-shore communities in Mayuge district, Eastern Uganda, we investi- gated the sources of schistosomiasis information, remembered content of information, and the perception of information and related practices towards the control of schistosomiasis. PLOS NEGLECTED TROPICAL DISEASES PLOS NEGLECTED TROPICAL DISEASES PLOS NEGLECTED TROPICAL DISEASES Schistosomiasis messaging: Lessons and implications from rural Uganda Author summary Schistosomiasis is a global-health concern causing severe disease, particularly in commu- nities in tropical areas such as Uganda. The parasite is spread in areas with inadequate sanitation and a lack of a safe water supply. Government control efforts focus on mass drug administration for people living in affected areas, with most treatments administered to school-aged children. However, drug uptake is low, and people are rapidly reinfected. In three heavily affected communities on the shores of Lake Victoria, we explored the sources of schistosomiasis information, how messages were relayed to community mem- bers, the remembered content of these messages and the way messages were perceived. Common sources of information were health workers at government health facilities, trained village health team members, teachers, and radio programmes. Our findings show that the information shared from the different sources is not consistent and, in some cases, this has caused confusion and prompted a reluctance to engage with treatment or preventative efforts. We propose a framework where there is dialogue between commu- nity member representatives, health workers based in the community, and government technical staff to come up with clear, concise, and consistent messages. Competing interests: The authors have declared that no competing interests exist. Methods and principal findings Peer Review History: PLOS recognizes the benefits of transparency in the peer review process; therefore, we enable the publication of all of the content of peer review and author responses alongside final, published articles. The editorial history of this article is available here: https://doi.org/10.1371/journal.pntd.0009893 Data were collected from September 2017 to March 2018 using a rapid ethnographic assessment that included transect walks, observations, individual in-depth interviews and focus group discussions. Data were analysed thematically using iterative categorisation. We found that the main sources of schistosomiasis information included health workers at government facilities, village health teams, teachers, and radio programmes produced by the Ministry of Health. These messages described the symptoms of schistosomiasis, but did not mention the side effects of praziquantel treatment. Despite this messaging, the main cause of the disease and transmission was unclear to most participants. The translation of schistosomiasis on the radio into the local language ‘ekidada’—meaning swollen stomach— increased, rather than reduced, confusion about the cause(s) of schistosomiasis, due to believed links between ekidada and witchcraft, and prompted a reluctance to engage with treatment or preventative efforts. Copyright: © 2021 Ssali et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability Statement: Due to the nature of the data and the informed consent obtained, in 1 / 15 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0009893 October 27, 2021 PLOS NEGLECTED TROPICAL DISEASES Conclusion and significance which participants were assured of anonymity, we cannot include all of the raw data, as some information may jeopardise the anonymity of the participant. Therefore, the data upon which this paper is based have been included in the Supplementary information and with all reference to names, household location, more specific job specifications identifiers removed. which participants were assured of anonymity, we cannot include all of the raw data, as some information may jeopardise the anonymity of the participant. Therefore, the data upon which this paper is based have been included in the Supplementary information and with all reference to names, household location, more specific job specifications identifiers removed. This study highlights gaps in schistosomiasis messaging. We recommend MDA is comple- mented by effective, evidence-based messaging on schistosomiasis transmission, preven- tion, and treatment, that is sensitive to local language and context issues, resulting in clear, concise, and consistent messages, to increase effectiveness. Funding: This study was funded by a Medical Research Council (https://mrc.ukri.org/) Global Challenges Research Fund Award (MR/P025447/1) to primary investigator PHLL, and co-investigators LP and JS. PHLL is also funded by a European Research Council (ERC) (https://erc.europa.eu/) Starting Grant (SCHISTO_PERSIST 680088), the Engineering and Physical Sciences Research Council (https://epsrc.ukri.org/) (EP/R01437X/1 and EP/T003618/1) and the Wellcome Trust (https://wellcome.org/) [204820/Z/16/Z]. The funders played no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript. Introduction Schistosomiasis, commonly known in East Africa as Bilharzia, is a neglected tropical disease (NTD) caused by parasitic flatworms of the genus Schistosoma [1]. In this paper we focus on intestinal schistosomiasis cause by Schistosoma mansoni, which is spread by parasite eggs excreted in stool from infected humans. When eggs reach fresh water, they hatch into a larval stage that infects snails. The parasite then reproduces asexually in the snails resulting in thou- sands of larvae being released into the water, which burrow into people in contact with con- taminated water. Transmission occurs in areas with inadequate access to, or use of, sanitation and safe water supplies. The ongoing World Health Organization strategy for populations in endemic areas is to reduce morbidity, prevalence, and transmission through mass drug admin- istration (MDA), with the anthelmintic praziquantel. Over 235 million people required prazi- quantel treatment in 2019, with 90% of those living in Africa [2]. The majority of MDA is administered at schools and through health facilities [3,4]. Research relating to schistosomiasis messaging across East Africa has shown that messages aimed as school-age children can improve biomedical understandings of transmission [5,6], but that messaging to caregivers and other adults also remains important [7,8]. This can be particularly effective when messages are targeted at key life-stage points, such as infant testing for mothers of young children [9], and when designed in relation to existing language and con- cepts [10], common misconceptions and misinformation [8,11,12] and understandings of, and fears around, treatment [12]. Not only has it been shown that messaging should address the 2 / 15 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0009893 October 27, 2021 PLOS NEGLECTED TROPICAL DISEASES Schistosomiasis messaging: Lessons and implications from rural Uganda primary concerns of communities [7], and local and biomedical understandings of health and common misconceptions [8,10–12], but it should also be combined with practical recommen- dations [13]. Messaging alone can provide challenges where people are advised against contact with infected water without having affordable access to improved water or other means of pro- tecting themselves, leading some to argue that messaging is most effective when situated along- side other strategies [5,14]. In Uganda, S. mansoni is common, found in 82 of the 134 districts in the country [15], caus- ing a serious public-health burden in at least 38 of the districts [9,13,16,17]. Introduction Several studies conducted in heavily endemic communities in Uganda link the continued burden to a lack of access to improved water and sanitation, open defaecation practices, and limited uptake of MDA, particularly by adults [9,15,18–21]. Uptake of MDA in Uganda has been compromised by a lack of knowledge about schistosomiasis transmission and prevention [5,22], a lack of knowledge about the MDA programme itself [18], and assumptions that MDA is only for chil- dren [18]. There is a call to rethink the top-down approach to schistosomiasis control, and making sure that people are correctly informed about the disease, its transmission, and the treatment available for it, will be key to giving people ownership over the problem [23,24]. Despite control efforts in endemic areas in Uganda, little is known about which schistosomia- sis messages are presented, and to whom, and how they are perceived by community members. Improved knowledge on the current use of messaging, and a better understanding of which messaging methods are effective, is needed. The results presented here are from a rapid ethnographic assessment undertaken to explore specific reasons why the schistosomiasis burden has remained high in three lakeshore commu- nities, despite long-term MDA. This exploratory method enables complex understandings, priorities, and proposed solutions of community members to be identified and explored in a rich and detailed way. A key theme among the suggestions made by community members was the need for improved messaging, which is the focus of this paper. Our aim was to identify the sources of schistosomiasis information, the nature of the con- tent shared, and how information was perceived by lake-shore communities in Mayuge Dis- trict, Uganda. These results can assist in the identification of additional interventions against schistosomiasis in endemic lake shore communities in East Africa, and inform how best to engage the people affected with current, and any new, interventions. Ethics statement The study had ethical clearance from the University of Glasgow College of Social Science Eth- ics Committee (CSSEC: 400160134), which adheres to UK Economic and Social Research Council standards. Further approvals were obtained from the Uganda Virus Research Institute Research Ethics Committee (GC/127/17/06/601) and the national regulatory body, Uganda National Council for Science and Technology (UNCST) (SS4241). All adults recruited pro- vided informed written consent. The parents or legal guardians of the children were asked for consent before their children were included in the discussion groups, providing informed written consent. The children were asked for their assent and verbal assent obtained prior to being recruited. Study area The study was conducted in three S. mansoni endemic communities, Bugoto, Bwondha and Musubi, on the shores of Lake Victoria in Mayuge District, Eastern Uganda (Fig 1). No S. hae- matobium is found in this area. This area was selected because there is a unique, highly 3 / 15 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0009893 October 27, 2021 oi.org/10.1371/journal.pntd.0009893 October 27, 2021 PLOS NEGLECTED TROPICAL DISEASES Schistosomiasis messaging: Lessons and implications from rural Uganda Fig 1. (A) Map of Uganda showing location of Mayuge District. (B) Map of Mayuge District showing the location of the three study site villages: Bugoto, Bwondha and Musubi. Created using QGIS 3.14 Software (QGIS Association. QGIS Geographic Information System. 3.14 ed2020) using a base layer from Natural Earth (http://www. naturalearthdata.com) for the reference maps, with district boundaries created using Uganda Bureau of Statistics (Uganda Bureau of Statistics (UBOS). Uganda Administrative Boundaries GIS Database. Kampala, Uganda: Government of Uganda, UBOS; 2006). https://doi.org/10.1371/journal.pntd.0009893.g001 Fig 1. (A) Map of Uganda showing location of Mayuge District. (B) Map of Mayuge District showing the location of the three study site villages: Bugoto, Bwondha and Musubi. Created using QGIS 3.14 Software (QGIS Association. QGIS Geographic Information System. 3.14 ed2020) using a base layer from Natural Earth (http://www. naturalearthdata.com) for the reference maps, with district boundaries created using Uganda Bureau of Statistics (Uganda Bureau of Statistics (UBOS). Uganda Administrative Boundaries GIS Database. Kampala, Uganda: Government of Uganda, UBOS; 2006). https://doi.org/10.1371/journal.pntd.0009893.g001 Fig 1. (A) Map of Uganda showing location of Mayuge District. (B) Map of Mayuge District showing the location of the three study site villages: Bugoto, Bwondha and Musubi. Created using QGIS 3.14 Software (QGIS Association. QGIS Geographic Information System. 3.14 ed2020) using a base layer from Natural Earth (http://www. naturalearthdata.com) for the reference maps, with district boundaries created using Uganda Bureau of Statistics (Uganda Bureau of Statistics (UBOS). Uganda Administrative Boundaries GIS Database. Kampala, Uganda: Government of Uganda, UBOS; 2006). https://doi.org/10.1371/journal.pntd.0009893.g001 https://doi.org/10.1371/journal.pntd.0009893.g001 https://doi.org/10.1371/journal.pntd.0009893.g001 detailed, longitudinal dataset from 2004 onwards from the three main municipal primary schools in these communities, showing that S. mansoni infection prevalence, intensity and associated morbidity in these communities had not decreased despite over a decade of MDA [25,26]. Study area There is one health centre in Bugoto and one in Bwondha. Musubi community does not have a health centre, however they access the one in Bugoto which is about 10 kilometres away. These facilities mainly offer outpatient services to community members. The terms health facility or health centre, as used in this paper, refer to government health centres (owned by the government, and the health workers there are paid by the government). Study area The main social and economic activities for the populations of Bwondha (population: ~60,000), Bugoto (population: ~6,000), and Musubi (population: ~2,000) include fishing, fish processing and trading, some farming and small businesses including shops, market kiosks and small alcohol drinking places, and restaurants. Bwondha is also a transport hub to the islands, with porters carrying goods and people from the shore to boats travelling to the islands. Fishing, farming (particularly rice farming), water fetching, and portering to boats moored offshore are all occupations with high levels of water contact. All three communities are close to the lake, with some houses just 30 metres from the shore. In all communities, fetch- ing and using water for cooking, bathing, washing clothes and general household use, from infected sites such as Lake Victoria and nearby swamps, increase the risk of S. mansoni expo- sure. There is one health centre in Bugoto and one in Bwondha. Musubi community does not have a health centre, however they access the one in Bugoto which is about 10 kilometres away. These facilities mainly offer outpatient services to community members. The terms health facility or health centre, as used in this paper, refer to government health centres (owned by the government, and the health workers there are paid by the government). detailed, longitudinal dataset from 2004 onwards from the three main municipal primary schools in these communities, showing that S. mansoni infection prevalence, intensity and associated morbidity in these communities had not decreased despite over a decade of MDA [25,26]. The main social and economic activities for the populations of Bwondha (population: ~60,000), Bugoto (population: ~6,000), and Musubi (population: ~2,000) include fishing, fish processing and trading, some farming and small businesses including shops, market kiosks and small alcohol drinking places, and restaurants. Bwondha is also a transport hub to the islands, with porters carrying goods and people from the shore to boats travelling to the islands. Fishing, farming (particularly rice farming), water fetching, and portering to boats moored offshore are all occupations with high levels of water contact. All three communities are close to the lake, with some houses just 30 metres from the shore. In all communities, fetch- ing and using water for cooking, bathing, washing clothes and general household use, from infected sites such as Lake Victoria and nearby swamps, increase the risk of S. mansoni expo- sure. Study design The study used a rapid, but in-depth, assessment procedure called rapid ethnographic assess- ment [27–29], to explore socially constructed and transmitted ideas about schistosomiasis dis- ease, transmission, and control in the study communities. The study aimed to provide a strong, locally-informed, behaviourally-consistent foundation for the design and feasibility of complex interventions to reduce transmission of schistosomiasis in these endemic areas. Two social science trained research assistants, one female and one male, who both spoke the local language, spent six weeks in each community (18 weeks in total). The research assis- tants were experienced in recruitment and qualitative data collection, having previously worked on other ethnographic research projects that targeted both children and adults. The assessment included transect walks in each community, and structured and unstructured observations of everyday forms of interaction with water and hygiene (Table 1). In addition, the researchers conducted a total of 60 individual IDIs and 19 FGDs (Table 1). During the observations and transect walks, several people were identified as key informants, often because of their occupation. Their inclusion in the IDIs depended on their interest in the study, schistosomiasis and/or health research and what they were contributing to the commu- nity. The fisherfolk were also considered leaders of their different small groups, or were named key persons because of their age and duration of stay in the community and involvement in fishing activities. The FGDs participants were identified through the local council leadership, and men’s and women’s leaders assisted the study by identifying relevant people to include in the study to help us gain a wide range of perspectives. In Table 1 we provide an overview of the data collection methods. Table 1. The rapid ethnographic assessment approach components. Activity Involvement Introductory visits to community gate keepers Visits to key leaders in the community, who included district political and technical leaders, village health team leaders, community local council leaders and professions including health workers and teachers, and religious and cultural leaders. These included both men and women. Transect walk Walks guided by community members with an understanding of the study aims and the local community; these were often VHTs. These walks focused on providing research assistants with an overview of the community, coupled with particular attention to sites related to water access (improved and unimproved), water contact (fishing, bathing etc.) and sanitation, including latrines and open defaecation sites. Study population The target population was individuals living in the three endemic communities. Participants were purposively selected from a range of community categories including NTD technical staff, health workers, teachers, members of the village health teams (VHTs), community lead- ers, fisherfolk, parents and children, to provide community-wide insights into schistosomiasis, transmission, control and related-messaging. The participants were recruited from their places of work or study, which included: the lake shore, fish drying grounds, gardens, schools, or in the community while selling fish, making nets, cleaning nets, repairing boats, in their shops, amongst other activities. One hundred and ninety-two adults participated in either in-depth 4 / 15 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0009893 October 27, 2021 PLOS NEGLECTED TROPICAL DISEASES Schistosomiasis messaging: Lessons and implications from rural Uganda interviews (IDIs) or focus group discussions (FGDs). The FGDs also included 40 school-age children aged 8–14 years, who were both in and out of education at the time of the study. Data collection The VHTs, or other community members guided the research assistants during the conduct of 26 transect walks across the communities, where they highlighted key water, sanitation, and hygiene (WASH) sites. Thirty-four structured observations were undertaken across these WASH sites. The research assistants participated in everyday social and economic activities, such as attending prayers in mosques and churches, attending a funeral ceremony and market day activities, to become immersed in what community members live through, appreciate their circumstances, and observe everyday activities and conversation in relation to schistoso- miasis, water, and sanitation. The duration of each observation varied between one and three hours. In each of the three communities, an introductory FGD was held with between 8–10 key informants (with community leaders) using a semi-structured topic guide. The topic guide (S1 Text) included questions focused on the knowledge that people had about schistosomiasis and what the main sources of that information were, how the messages were presented, and what people understood in terms of how schistosomiasis is transmitted and prevented in their communities. After the introductory FGD with community leaders in each community, four further FGDs were held with different groups of adults in each of the communities, all addressing the same issues, using the same topic guide. These included older men (35 years), younger men (18–34 years), older women (35 years), and younger women (18–34 years). Data from all the FGDs, including the introductory ones, were included in the analyses. Within each of the three communities 20 individual IDIs were conducted to deepen understanding of topics raised during the FGDs. Four FGDs were conducted with school-aged children, both school attending and non-attending, in two of the communities, Bugoto and Musubi. In each com- munity, one FGD was held with children aged 8–11 years and the other was for children aged 12–14 years. We did not hold FGDs with children from Bwondha community due to the impassable roads which occurred during the rainy season at the end of the data collection period. All FGDs were performed in Lusoga, the language commonly used in the region. The research assistants wrote up observation notes in daily diaries. In-depth interviews and FGDs were recorded on digital recorders and each of the participants gave verbal consent for the recording prior to being enrolled and participating. Information sheets, consent forms and topic guides were all translated into Lusoga. Data collection As many respondents were not able to read in either Lusoga or English, the study information documents were read by the research assis- tants to each respondent, and each gave their written informed consent using a signature or thumb print before agreeing to take part in the study. The parents or legal guardians of the children were asked for consent, and each gave their written informed consent using a signa- ture or thumb print, before their children were included in the discussion groups. The children were asked for their assent before being included. Study design These walks also provided a means for research assistants to become visible and introduced to the community, and for guides to identify particular individuals or groups whom they considered may be of particular interest for the individual interviews and group discussions in the study. Structured observations Having identified key sites through transect walks, research assistants undertook structured observations, noting types of water contact, duration, approximate age and gender of those contacting water, water contact for mixed use (e.g. water fetching and swimming) at different time points (morning, afternoon and evening) and on different days of the week; no observations were undertaken of open defaecation sites to maintain community members’ dignity and privacy. Focus group discussions (FGDs) Focus group discussions were first undertaken with community leaders (8–10 people), which helped raise the profile of the study in the community, as well as providing perspectives from individuals with influence in different domains. These leaders included both men and women. Further FGDs were then undertaken with older men (35), older women (35), younger men (18–34 years), younger women (18–34 years), and children (8–11 years and 12–14 years) in order to capture both common attitudes and concerns by gender and age and the range of experiences and views within them. The same FGD guide was used in the introductory FGDs as for all the other FGDs. Individual in-depth interviews (IDIs) Research-assistant-led semi-structured interviews about schistosomiasis. We targeted key individuals who shared their knowledge and experiences. Participant observation Underpinning all other components was participant observation. Research assistants participated in day-to-day activities such as playing football, attending prayers at church and mosque, and being invited to attend a funeral. https://doi.org/10.1371/journal.pntd.0009893.t001 . The rapid ethnographic assessment approach components PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0009893 October 27, 2021 5 / 15 PLOS NEGLECTED TROPICAL DISEASES Schistosomiasis messaging: Lessons and implications from rural Uganda Results A total of 192 adults and 40 children participated in FGDs and 60 adults took part in IDIs. The study findings are grouped into three main themes: 1) sources of information; 2) content shared; and 3) perception of information and practice towards the control of schistosomiasis. We then set out suggestions made by respondents in relation to schistosomiasis messaging. Data management and analysis Data were uploaded to secure, password-protected, computers each day. Notebooks were stored in a locked cupboard throughout data collection field trips. At the mid- and end-point of each period of data collection in each community, data were transferred to the secure insti- tution server at the Uganda Virus Research Institute. All audio files were transcribed and translated into English by the research assistants. Data were anonymised before being shared with the wider research team, maintaining only details of the age, gender, community, and FGD/IDI for transcribed information. The transcripts were uploaded into NVivo 12 to man- age the coding for thematic analysis. Coding was both deductive following the research 6 / 15 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0009893 October 27, 2021 PLOS NEGLECTED TROPICAL DISEASES Schistosomiasis messaging: Lessons and implications from rural Uganda questions and inductive where new codes were added to reflect emergent themes from the data as coding progressed, resulting in a total of six overarching group codes and 30 sub-group codes (S1 Table). Coded data were analysed using iterative categorisation [30], enabling us to identify major and minor themes in relation to the central research questions. Raw anon- ymised data are available in S1 Data. Sources of schistosomiasis information Research participants mentioned several channels through which they received information on schistosomiasis. The most frequently mentioned source of information was the health workers at the government health facilities. The VHTs, who worked in the community, also sensitised people about schistosomiasis in general. When describing these sources, some par- ticipants clearly distinguished VHTs from health workers at the government health facilities, but in most cases they referred to health workers more generally in the vernacular language to include staff at the health facilities and VHTs, and sometimes national and international researchers were also referred to as health workers. Information about schistosomiasis was usually received through the VHTs when they moved in the community to offer treatment. A 20 year old man was asked how a person learnt about schistosomiasis from health centres, to which he replied: ‘No I have never, I only get such information when VHTs move to give us [praziquantel] tablets’ (Bugoto IDI). Another man, aged 42, from the same community also mentioned of VHTs that ‘they sensitise us about any illness they want’ (Bugoto IDI). Some participants said that they had read or seen health charts about schistosomiasis on the walls of the government health facilities in their communities. A 67 year old woman from Bugoto (IDI) commented: ‘I have heard a lot about it [Bilharzia] and I have seen pictures of it in those health centres’. Some parents and caretakers mentioned schools, teachers, and children as information dis- tribution points, particularly during the seasons when the children received MDA at their schools or when researchers were collecting samples and treating them. One parent shared that: The health workers I have talked about went to schools and taught children about Bilharzia, they took off their stool and blood for check-up and when [my] children came back, they told me that they were checked for Bilharzia. (Female, 53 years, Fish trader, Bugoto IDI) Another parent noted: We also get the information from children because health workers go to schools and check them, they ask us to consent for the children to be treated. Sources of schistosomiasis information (Male, 53 years, Fisherman, Bugoto IDI) (Male, 53 years, Fisherman, Bugoto IDI) 7 / 15 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0009893 October 27, 2021 PLOS NEGLECTED TROPICAL DISEASES Schistosomiasis messaging: Lessons and implications from rural Uganda Other participants, particularly in the 18–34 years FGDs, mentioned their sources of infor- mation being the school when they were still studying from these primary schools in the three communities. Only a small number of adult participants mentioned watching videos and films about schistosomiasis transmission; these had been viewed in school when they were children. How- ever, those who had watched videos or films were better able to describe how the parasite may enter the body than those who had not. Radio programmes (media), were mentioned as another source of information. MDA was well known in the community and most respondents were aware of it and some had been recipients of praziquantel, or their children, or a relative. These people also often mentioned the side effects of treatment: I have been hearing announcements [media], health workers also come to this community and tell us about it, they also give us tablets to swallow. And what shows that we have Bilhar- zia when we swallow the tablets we feel bad, after swallowing you find someone lying down, due to diarrhoea and vomiting. (Female, 27 years, Casual worker, Musubi IDI) As well as these formal channels of information such as radio messaging, health workers, VHTs, MDA programmes and research participation described above, a number of partici- pants mentioned developing their understanding of schistosomiasis through seeing friends and neighbours becoming sick and interpreting those symptoms as schistosomiasis. One 53 year old man mentioned seeing friends with ‘swollen stomachs’. This interpretation of symp- toms in others’ bodies is notable because there are many other causes of the characteristic swollen stomach of late-stage schistosomiasis, discussed below. Schistosomiasis content shared The VHTs emphasised to community members the importance of not defaecating in the lake, they talked about the parasites/worms that transmit schistosomiasis and some participants reported the need to use gum boots while in the water. Some of the common messages deliv- ered by the VHTs are illustrated below: The VHTs sensitise them not to drink that water. They also tell them that if one defaecates in the lake when he has Bilharzia, any other person who accesses the lake or pond stands high chances of catching Bilharzia. They teach them that snails also produce parasites that cause Bilharzia. (Community leaders, Bugoto FGD) VHT taught us that Bilharzia are parasites found in the lake, they make the stomach swell. They say that if you go to fetch water, you shouldn’t spend a lot of time in the lake because the worm enters through the hair follicles. (Male, 55 years, Fisherman, Musubi IDI) Some people, when referring to information provided by the VHTs, described symptoms of other illnesses in response to questions about schistosomiasis symptoms: 8 / 15 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0009893 October 27, 2021 oi.org/10.1371/journal.pntd.0009893 October 27, 2021 PLOS NEGLECTED TROPICAL DISEASES Schistosomiasis messaging: Lessons and implications from rural Uganda And they tell us that those children suffer from kwashiorkor [acute malnutrition], and kwash- iorkor is in English we don’t understand what it is. Health workers tell us it is kwashiorkor and not Bilharzia yet for us we know kwashiorkor also makes someone to swell like that. (Community leaders Bugoto FGD) And they tell us that those children suffer from kwashiorkor [acute malnutrition], and kwash- iorkor is in English we don’t understand what it is. Health workers tell us it is kwashiorkor and not Bilharzia yet for us we know kwashiorkor also makes someone to swell like that. One respondent mentioned content in relation to what they witnessed in the community: I have also seen someone that is suffering from it [Bilharzia] and I looked keenly at them and the situation they were going through was indeed terrible, they may fatten the legs and the stomach, their respiratory will be bad, the way they walk too, so by the time God calls them, they die a very disturbing death with so much pain, the foot will have grown fat and so shall the stomach. Schistosomiasis content shared (Female, 67 years, Bugoto IDI) Messages delivered through radio programmes were often about signs of schistosomiasis, as well as information to discourage defaecation at the lake to prevent schistosomiasis: They usually advise us over the radios to leave water under the sun for 24 hours before use and they also encourage farmers and fishermen to wear gumboots. (Male, 34 years, Farmer, Musubi IDI) However, many respondents mentioned that messages aired on the radio were not always clear and they found the messages confusing: They confused us by saying there is no ekidada (bilharzia) that is bewitched. And they said there is ekidada found in water and lake. We hear on radio. (Male, 34 years, Farmer, Musubi IDI) The information shared with school children through the teachers and researchers was also deemed insufficient for them to share with their parents as one young man explained: Actually, for me when I came back from school and told my father that we were given Bilhar- zia tablets, he asked what Bilharzia is, but I failed to answer him. Actually, for me when I came back from school and told my father that we were given Bilhar- zia tablets, he asked what Bilharzia is, but I failed to answer him. Participants noted that when they did receive information, they did not always act on this information in the ways intended as explained below: People have not yet understood the issue of leaving water in the sun. That announcement is always on the radio about leaving water in the sun for 24 hours before use, for them after fetching water if they are going to bath a baby or cook food they just use the water. (Male, 53 years, fisherman, Bugoto IDI) Thus, while messages may have been shared, the content was not always understood or acted upon. 9 / 15 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0009893 October 27, 2021 PLOS NEGLECTED TROPICAL DISEASES Schistosomiasis messaging: Lessons and implications from rural Uganda Perceptions towards messaging Participants discussed being exposed to information, some of which was perceived as interest- ing and some confusing. However, adverts with schistosomiasis information were presented in a way that listeners found amusing and therefore memorable: The advert goes, where do you defaecate? I defaecate in the lake, where do you fetch water? I fetch from the lake; eeh you are the ones who have caused Bilharzia (laughs). As mentioned above, radio messages sometimes aired controversial information according to some participants, that led to confusion among the listeners due to beliefs in supernatural forces or witchcraft: I heard someone on radio saying that Bilharzia is ekidada but we got confused because we have people who have suffered from ekidada, there is one we even took to Jinja referral hospi- tal, she was bewitched [through] ekidada when she was pregnant. (Male, 34 years, Farmer, Musubi IDI) Concerns about and experience of side effects after taking the praziquantel by some commu- nity members were often shared within the community and these affected responses to MDA: So for Bilharzia if one person takes drugs and feels bad he will spoil other 10 people that they shouldn’t take praziquantel because it is bad, therefore if Bilharzia is also given radio talk shows you never know people will get to know the benefits of praziquantel. (Participant-observation, Bugoto) There were various rumours about the MDA treatment which were reported in the com- munity and these in some cases resulted in the distortion of messages passed on from the dif- ferent sources: Hmmm, some women say that when you swallow Bilharzia tablets you stop giving birth, they say I cannot swallow them, others say that those tablets are for Illuminati [a satanic organiza- tion or cult, present in the area, according to the community members]. Everyone speaks in his or her own way. (Female, 40 years, sells silver fish, Bugoto IDI) (Female, 40 years, sells silver fish, Bugoto IDI) In this study we found that while information had been shared through different health workers and media, there was considerable uncertainty about the meaning of the messages and how schistosomiasis could be tackled in the study area. Discussion This study was conducted in three high S. mansoni endemicity communities, on the shores of Lake Victoria. The study identified the different sources of schistosomiasis information sur- rounding the disease, transmission and control, the content of these messages, and how they were perceived. We found that information was shared with community members by professional health workers and village health teams (VHTs) through meetings and at the health facilities. Radio PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0009893 October 27, 2021 10 / 15 PLOS NEGLECTED TROPICAL DISEASES Schistosomiasis messaging: Lessons and implications from rural Uganda programs about schistosomiasis and information from teachers and researchers to children and then on to their guardians after MDA were noted as important channels of information dissemination. Indeed, people in the 18–34 years old FGDs still reported their primary schools during MDA as their main source of information on schistosomiasis. These people were often still heavily infected, but treatment coverage in these age groups is much lower with many thinking it is only for school-aged children [18], so it is important that accurate information reaches all age groups to improve coverage in groups outside of the school years [31]. Our study results have shown that there is some awareness of schistosomiasis in all the stud- ied communities and the level of information about schistosomiasis in the community is com- parable to that noted in earlier studies [18,32]. With respect to disease symptoms, there was a general accurate understanding surrounding the swollen stomach, and the disease was often mentioned in relation to other correct signs, symptoms and side effects such as diarrhoea and vomiting. A lack of correct knowledge of schistosomiasis has been shown to limit the control of schistosomiasis in endemic communities among caregivers of children aged 2–4 years [9]. Despite some correct knowledge in the communities we studied, the often-reported confusion is likely to have a similar potential negative effect on MDA. A particularly important aspect that requires follow up when relaying schistosomiasis information in this study area is what was referred to as ekidada (literally translated: swollen stomach) and which was mentioned by many participants to be caused by witchcraft [6]. In Uganda witchcraft has previously been named as one of the causes of infection, and supports our findings here, and people’s under- standing of the situation may hinder treatment efforts [9]. PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0009893 October 27, 2021 Discussion 11 / 15 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0009893 October 27, 2021 PLOS NEGLECTED TROPICAL DISEASES Schistosomiasis messaging: Lessons and implications from rural Uganda Our results reveal that as information about schistosomiasis moved between people it changed and it is affected by both a lack of understanding of the biomedical facts and the already existing local ideas about causality and treatment. This may mean that the messages are not ingrained as facts by some people, but rather as information they have heard. If change is to happen, all community members need to be a part of, and own, shared understandings of, and thus control of, the disease. One advantage of radio programme messaging is that each individual who hears the infor- mation directly receives the same message which may potentially reduce the loss of details as information flows from one person to another. However, to strengthen the intervention, the radio programs need to be consistent, clear, and accurate, and can be coupled with improved training for teachers and the VHTs so that they can answer questions in a better way, with con- fidence and accuracy which would potentially reduce loss of detail as information is transmitted. We recommend improved guidance for schistosomiasis control programmes at a national and local level, with clear short messages regarding the disease, symptoms, treatment, and side effects, but avoiding the language of ekidada. It is critical to engage different stakeholders, including government technical control programme staff, teachers, VHTs, local community and political leaders, and targeted research community members, to agree on the best terms in the vernacular to use during the process of preparing messages for disease prevention and control. Study limitations This study has some limitations. Time and weather constraints limited the number of people we could talk to, particularly the exclusion of children’s voices from Bwondha due to impass- able road and physical access to the community. Further time in the field sites would inevitably have led to the researchers being able to undertake participant-observation over a more extended period, creating further opportunities to see how Bilharzia is discussed in everyday, spontaneous speech with a wider range of people, as community members became increas- ingly familiar with their presence. However, the rapid ethnographic assessment method used does enable rich and complex qualitative data and improved understandings to be gained, that can complement ongoing quantitative studies, and findings are useful to help inform, adjust, and improve the current control policies. The pragmatic constraints of necessarily rapid research meant that being able to access the community through the NTD technical persons was a significant advantage, but also means that this route of access and the rapid time frame could have influenced responses of some of the participants. Similarly, depending on teachers to identify children to take part could have resulted in some bias regarding which children par- ticipated. Finally, we purposively selected study participants with a view to gaining a wide range of views and with an ultimate research goal of identifying potential solutions, and as a result less engaged members of the community, whose understandings and practices are also incredibly important, may have been under-represented, despite the work undertaken by the research assistants to actively recruit diversely across the communities. Discussion There was mention of rumours that the praziquantel drugs may affect fertility as well as concerns about other side effects of drugs, as has been noted elsewhere [33]. As side effects are common [14], but also reported to be positively associated with infection intensity [34,35], it is an imperative that these are not ignored in schistosomiasis messaging. Indeed, they should be more widely disseminated, to forewarn individuals, but also to explain that symptoms will pass and may be indicative of needing treatment and of a successful treatment. This is particularly important as messaging about side effects commonly occurs in these communities, but between less-informed community members via informal routes, rather than from VHTs for example. As information flows through communities, it is important that those who experi- ence side effects are able to explain what these side effects mean, especially as our study found that description of side effects alone can be a barrier to treatment uptake. A study conducted in Nebbi district in Northwest Uganda, highlighted the need for biomedical messages to be disseminated in ways that can convince, especially the adults, to take drugs [32]. They focused on the association of some symptoms with witchcraft showing the value of clear messaging relating to symptoms and treatment to address, together, fears relating to causation and treatment. School attending children are the most common recipients of MDA, however our results have shown children do not always accurately relay to their parents and guardians why they are receiving treatment, or how to stop themselves and their families from getting reinfected. This lack of information may be caused by either the children not being fully informed why they are receiving MDA, or messages being forgotten or mixed up whilst being relayed. The lack of accurate information may reduce the possibility for changes in behaviour for the chil- dren and their parents/guardians. A study conducted in Kenya showed that parents are used to sending children to fetch water from the lake as a normal practice. This commonly saves money, as they would otherwise need to pay for clean water, and also usually enables collection of water from closer to their homes [5]. Improved messaging about the risks of these practices and the need for treatment might help to reduce this parent-driven risk factor for children. Acknowledgments We wish to acknowledge and thank all the respondents in this study, in particular the children, parents, and different community members. We thank the district and community leaders, and the community health teams who supported the research assistants during data collection. We thank the community as a whole, for allowing us to study the everyday activities they undertake as they go about their days. We thank the whole research team and investigators from MRC/UVRI & LSHTM Uganda Research Unit and the University of Glasgow for the col- laboration and support to conduct this study, especially our driver Hassan Ssenyonga. We thank Suzan Trienekens for creating the maps. Conclusions While people in the three communities studied are aware of schistosomiasis, perceptions and implementing positive control measures for individuals and the community, as a whole, are hampered by mixed messages. A clear, concise, and consistent message needs to be created through an inclusive stakeholder dialogue comprising of the media, technical health workers, teachers, local and cultural community leaders and lay community members who represent 12 / 15 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0009893 October 27, 2021 PLOS NEGLECTED TROPICAL DISEASES Schistosomiasis messaging: Lessons and implications from rural Uganda men and women. This message then needs to be transmitted to community members in a clear and consistent way. Dialoguing would lead to a partnership to develop appropriate edu- cational intervention messages and dissemination approaches to help control, and eventually eliminate, schistosomiasis in these heavily affected lakeshore communities. S1 Table. Coding framework showing the group and sub-group codes used for data analy- ses. (DOCX) S1 Table. Coding framework showing the group and sub-group codes used for data analy- ses. (DOCX) S1 Text. Topic guide for the focus group discussions and in-depth interviews. (DOCX) S1 Text. Topic guide for the focus group discussions and in-depth interviews. (DOCX) References 1. King C.H., The evolving schistosomiasis agenda 2007–2017—Why we are moving beyond morbidity control toward elimination of transmission. PLoS Neglected Tropical Diseases, 2017. 11(4): p. e0005517. https://doi.org/10.1371/journal.pntd.0005517 PMID: 28426653 2. World Health Organization. Schistosomiasis. 2019; https://www.who.int/news-room/fact-sheets/detail/ schistosomiasis [accessed 10/09/21]. 3. World Health Organization, Health in 2015: from MDGs, millennium development goals to SDGs, sus- tainable development goals. 2015, Geneva: World Health Organization. 4. World Health Organization, Integrating neglected tropical diseases into global health and development: fourth WHO report on neglected tropical diseases. 2017, Geneva: World Health Organization. 5. Takeuchi R., et al., Is there a gap between health education content and practice toward schistosomiasis prevention among schoolchildren along the shores of Lake Victoria in Kenya? PLoS Neglected Tropical Diseases, 2019. 13(8): p. e0007572. https://doi.org/10.1371/journal.pntd.0007572 PMID: 31425499 6. Muhumuza S., et al., Understanding low uptake of mass treatment for intestinal schistosomiasis among school children: a qualitative study in Jinja district, Uganda. Journal of Biosocial Science, 2015. 47(4): p. 505–20. https://doi.org/10.1017/S002193201400011X PMID: 24735860 7. Odhiambo G.O., et al., Low levels of awareness despite high prevalence of schistosomiasis among communities in Nyalenda informal settlement, Kisumu City, Western Kenya. PLoS Neglected Tropical Diseases, 2014. 8(4): p. e2784. https://doi.org/10.1371/journal.pntd.0002784 PMID: 24699502 8. Sacolo H., Chimbari M., and Kalinda C., Knowledge, attitudes and practices on Schistosomiasis in sub- Saharan Africa: a systematic review. BMC Infectious Diseases, 2018. 18(1): p. 46. https://doi.org/10. 1186/s12879-017-2923-6 PMID: 29347919 9. Kibira S.P.S., et al., Schistosomiasis infection in pre-school aged children in Uganda: a qualitative descriptive study to identify routes of exposure. BMC Infectious Diseases, 2019. 19(1): p. 165. https:// doi.org/10.1186/s12879-019-3803-z PMID: 30764781 10. Ojwang B.O., Linguistic Conceptualizations of Disease Among the Luo of Kenya. Qualitative Health Research, 2018. 28(3): p. 433–445. https://doi.org/10.1177/1049732317747875 PMID: 29313465 11. Fleming F., et al., Process evaluation of schistosomiasis control in Uganda, 2003 to 2006: perceptions, attitudes and constraints of a national programme. Parasitology, 2009. 136(13): p. 1759–1769. https:// doi.org/10.1017/S0031182009990709 PMID: 19695107 12. Adeneye A., et al., Sociocultural aspects of mass delivery of praziquantel in schistosomiasis control: The Abeokuta experience. Research in Social and Administrative Pharmacy, 2007. 3(2): p. 183–198. https://doi.org/10.1016/j.sapharm.2006.07.001 PMID: 17561219 13. Kabatereine N.B., et al., Integrated prevalence mapping of schistosomiasis, soil-transmitted helminthia- sis and malaria in lakeside and island communities in Lake Victoria, Uganda. Parasites & Vectors, 2011. 4(1): p. 232. https://doi.org/10.1186/1756-3305-4-232 14. Author Contributions Conceptualization: Lucy Pickering, Janet Seeley, Poppy H. L. Lamberton. Conceptualization: Lucy Pickering, Janet Seeley, Poppy H. L. Lamberton. Data curation: Agnes Ssali, Lucy Pickering, Edith Nalwadda, Lazaaro Mujumbusi, Poppy H. L. Lamberton. Formal analysis: Agnes Ssali, Lucy Pickering, Edith Nalwadda, Lazaaro Mujumbusi. Funding acquisition: Lucy Pickering, Janet Seeley, Poppy H. L. Lamberton. Investigation: Edith Nalwadda, Lazaaro Mujumbusi. Methodology: Agnes Ssali, Lucy Pickering, Edith Nalwadda, Lazaaro Mujumbusi, Janet Seeley, Poppy H. L. Lamberton. Project administration: Agnes Ssali, Lucy Pickering, Edith Nalwadda, Lazaaro Mujumbusi, Janet Seeley, Poppy H. L. Lamberton. Resources: Janet Seeley, Poppy H. L. Lamberton. Supervision: Agnes Ssali, Lucy Pickering, Janet Seeley, Poppy H. L. Lamberton. Validation: Agnes Ssali, Lucy Pickering, Edith Nalwadda, Lazaaro Mujumbusi, Janet Seeley, Poppy H. L. Lamberton. Visualization: Poppy H. L. Lamberton. Visualization: Poppy H. L. Lamberton. Writing – original draft: Agnes Ssali. Writing – original draft: Agnes Ssali. 13 / 15 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0009893 October 27, 2021 PLOS NEGLECTED TROPICAL DISEASES Schistosomiasis messaging: Lessons and implications from rural Uganda Writing – review & editing: Agnes Ssali, Lucy Pickering, Edith Nalwadda, Lazaaro Mujum- busi, Janet Seeley, Poppy H. L. Lamberton. Writing – review & editing: Agnes Ssali, Lucy Pickering, Edith Nalwadda, Lazaaro Mujum- busi, Janet Seeley, Poppy H. L. Lamberton. 19. Kabatereine N., et al., Community perceptions, attitude, practices and treatment seeking behaviour for schistosomiasis in L. Victoria islands in Uganda. BMC Research Notes, 2014. 7(1): p. 900. https://doi. org/10.1186/1756-0500-7-900 References Muhumuza S., et al., Uptake of preventive treatment for intestinal schistosomiasis among school chil- dren in Jinja district, Uganda: a cross sectional study. PloS One, 2013. 8(5): p. e63438. https://doi.org/ 10.1371/journal.pone.0063438 PMID: 23667617 15. Exum N.G., et al., The prevalence of schistosomiasis in Uganda: A nationally representative population estimate to inform control programs and water and sanitation interventions. PLoS Neglected Tropical Diseases, 2019. 13(8): p. e0007617. https://doi.org/10.1371/journal.pntd.0007617 PMID: 31412023 16. Kabatereine N.B., et al., Progress towards countrywide control of schistosomiasis and soil-transmitted helminthiasis in Uganda. Transactions of the Royal Society of Tropical Medicine and Hygiene, 2006. 100(3): p. 208–215. https://doi.org/10.1016/j.trstmh.2005.03.015 PMID: 16378628 17. Nalugwa A., et al., Intestinal schistosomiasis among preschool children along the shores of Lake Victo- ria in Uganda. Acta Tropica, 2015. 142: p. 115–121. https://doi.org/10.1016/j.actatropica.2014.11.014 PMID: 25454166 18. Adriko M., et al., Low praziquantel treatment coverage for Schistosoma mansoni in Mayuge District, Uganda, due to the absence of treatment opportunities, rather than systematic non-compliance. Tropi- cal Medicine and Infectious Disease, 2018. 3(4): p. 111. https://doi.org/10.3390/tropicalmed3040111 PMID: 30297642 19. Kabatereine N., et al., Community perceptions, attitude, practices and treatment seeking behaviour for schistosomiasis in L. Victoria islands in Uganda. BMC Research Notes, 2014. 7(1): p. 900. https://doi. org/10.1186/1756-0500-7-900 14 / 15 PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0009893 October 27, 2021 PLOS NEGLECTED TROPICAL DISEASES Schistosomiasis messaging: Lessons and implications from rural Uganda 20. Sanya R.E., et al., Perceptions about interventions to control schistosomiasis among the Lake Victoria island communities of Koome, Uganda. PLoS Neglected Tropical Diseases, 2017. 11(10): p. e0005982. https://doi.org/10.1371/journal.pntd.0005982 PMID: 28968470 21. Tukahebwa E.M., et al., Comparative efficacy of one versus two doses of praziquantel on cure rate of Schistosoma mansoni infection and re-infection in Mayuge District, Uganda. Transactions of the Royal Society of Tropical Medicine and Hygiene, 2013. 107(6): p. 397–404. https://doi.org/10.1093/trstmh/ trt024 PMID: 23596262 22. Tuhebwe D., et al., Uptake of mass drug administration programme for schistosomiasis control in Koome Islands, Central Uganda. PLoS One, 2015. 10(4): p. e0123673. https://doi.org/10.1371/journal. pone.0123673 PMID: 25830917 23. Onasanya A., et al., Rethinking the top-down approach to schistosomiasis control and elimination in sub-Saharan Africa. Frontiers in Public Health, 2021. 9(99). https://doi.org/10.3389/fpubh.2021. 622809 PMID: 33681133 24. Krauth S.J., et al., A call for systems epidemiology to tackle the complexity of schistosomiasis, its con- trol, and its elimination. Tropical Medicine and Infectious Disease, 2019. 4(1): p. 21. https://doi.org/10. 3390/tropicalmed4010021 PMID: 30699922 25. PLOS Neglected Tropical Diseases \| https://doi.org/10.1371/journal.pntd.0009893 October 27, 2021 References Lamberton PHL, et al., Sensitivity and specificity of multiple Kato-Katz thick smears and a Circulating Cathodic Antigen test for Schistosoma mansoni diagnosis pre- and post-repeated-praziquantel treat- ment. PLoS Neglected Tropical Diseases, 2014. 8(9): p. e3139. https://doi.org/10.1371/journal.pntd. 0003139 PMID: 25211217 26. Crellen T., et al., Reduced efficacy of praziquantel against Schistosoma mansoni is associated with multiple rounds of mass drug administration. Clinical Infectious Diseases, 2016. 63(9): p. 1151–1159. https://doi.org/10.1093/cid/ciw506 PMID: 27470241 27. Mugisha J., et al., A study of knowledge, experience, and beliefs about hepatitis B virus (HBV) infection in south western Uganda. Frontiers in Public Health, 2019. 7: p. 304. https://doi.org/10.3389/fpubh. 2019.00304 PMID: 31709217 28. Manderson L. and Aaby P., Can rapid anthropological procedures be applied to tropical diseases? Health Policy and Planning, 1992. 7(1): p. 46–55. https://doi.org/10.1093/heapol/7.1.46 29. Bond V., et al., Value and limitations of broad brush surveys used in community-randomized trials in Southern Africa. Qualitative Health Research, 2019. 29(5): p. 700–718. https://doi.org/10.1177/ 1049732318809940 PMID: 30556470 30. Neale J., Iterative categorization (IC): a systematic technique for analysing qualitative data. Addiction, 2016. 111(6): p. 1096–1106. https://doi.org/10.1111/add.13314 PMID: 26806155 31. Faust C.L., et al., Schistosomiasis control: Leave no age group behind. Trends in Parasitology, 2020. 36(7): p. 582–591. https://doi.org/10.1016/j.pt.2020.04.012 PMID: 32430274 32. Parker M., Allen T., and Hastings J., Resisting control of neglected tropical diseases: dilemmas in the mass treatment of schistosomiasis and soil-transmitted helminths in north-west Uganda. Journal of Bio- social Science, 2008. 40(2): p. 161–181. https://doi.org/10.1017/S0021932007002301 PMID: 17761005 33. Hastings J., Rumours, riots and the rejection of mass drug administration for the treatment of Schistoso- miasis in Morogoro, Tanzania. Journal of Biosocial Science, 2016. 48(S1): p. S16–S39. https://doi.org/ 10.1017/S0021932016000018 PMID: 27428064 34. Erko B., et al., Efficacy and side effects of praziquantel in the treatment of Schistosomiasis mansoni in schoolchildren in Shesha Kekele Elementary School, Wondo Genet, Southern Ethiopia. Asian Pacific Journal of Tropical Biomedicine, 2012. 2(3): p. 235–239. https://doi.org/10.1016/S2221-1691(12) 60049-5 PMID: 23569905 35. Raso G., et al., Efficacy and side effects of praziquantel against Schistosoma mansoni in a community of western Coˆte d’Ivoire. Transactions of The Royal Society of Tropical Medicine and Hygiene, 2004. 98(1): p. 18–27. https://doi.org/10.1016/s0035-9203(03)00003-8 PMID: 14702835 15 / 15
https://openalex.org/W4380628781	https://www.frontiersin.org/articles/10.3389/fcvm.2023.1180792/pdf	English	null	A bibliometric analysis of myocardial ischemia/reperfusion injury from 2000 to 2023	Frontiers in cardiovascular medicine	2,023	cc-by	13,966	The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Conclusion: Research on MIRI is ﬂourishing. It is necessary to conduct an in-depth investigation of the interaction between different mechanisms and multi-target therapy will be the focus and hotspot of MIRI research in the future. myocardial ischemia/reperfusion injury, reperfusion therapy, myocardial infarction, mechanism, multi-target therapy, bibliometric analysis EDITED BY Tommaso Gori, Johannes Gutenberg University Mainz, Germany REVIEWED BY Peng Zhu, Southern Medical University, China Yang Yang, First Afﬁliated Hospital of Zhengzhou University, China CORRESPONDENCE Xiaochang Ma maxiaochang@x263.net Mi Liu liumi177@qq.com RECEIVED 06 March 2023 ACCEPTED 22 May 2023 PUBLISHED 13 June 2023 CITATION Wang Y, Guo L, Zhang Z, Fu S, Huang P, Wang A, Liu M and Ma X (2023) A bibliometric analysis of myocardial ischemia/reperfusion injury from 2000 to 2023. Front. Cardiovasc. Med. 10:1180792. doi: 10.3389/fcvm.2023.1180792 COPYRIGHT © 2023 Wang, Guo, Zhang, Fu, Huang, Wang, Liu and Ma. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms 1Xiyuan Hospital, China Academy of Chinese Medical Sciences, Beijing, China, 2Graduate School, Beijing University of Chinese Medicine, Beijing, China, 3National Clinical Research Center for Chinese Medicine Cardiology, Beijing, China Background: Myocardial ischemia/reperfusion injury (MIRI) refers to the more severe damage that occurs in the previously ischemic myocardium after a short-term interruption of myocardial blood supply followed by restoration of blood ﬂow within a certain period of time. MIRI has become a major challenge affecting the therapeutic efﬁcacy of cardiovascular surgery. Methods: A scientiﬁc literature search on MIRI-related papers published from 2000 to 2023 in the Web of Science Core Collection database was conducted. VOSviewer was used for bibliometric analysis to understand the scientiﬁc development and research hotspots in this ﬁeld. Results: A total of 5,595 papers from 81 countries/regions, 3,840 research institutions, and 26,202 authors were included. China published the most papers, but the United States had the most signiﬁcant inﬂuence. Harvard University was the leading research institution, and inﬂuential authors included Lefer David J., Hausenloy Derek J., Yellon Derek M., and others. All keywords can be divided into four different directions: risk factors, poor prognosis, mechanisms and cardioprotection. Front. Cardiovasc. Med. 10:1180792. doi: 10.3389/fcvm.2023.1180792 COPYRIGHT © 2023 Wang, Guo, Zhang, Fu, Huang, Wang, Liu and Ma. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). TYPE Review PUBLISHED 13 June 2023 DOI 10.3389/fcvm.2023.1180792 Frontiers in Cardiovascular Medicine A bibliometric analysis of myocardial ischemia/reperfusion injury from 2000 to 2023 EDITED BY Tommaso Gori, Johannes Gutenberg University Mainz, Germany REVIEWED BY Peng Zhu, Southern Medical University, China Yang Yang, First Afﬁliated Hospital of Zhengzhou University, China CORRESPONDENCE Xiaochang Ma maxiaochang@x263.net Mi Liu liumi177@qq.com RECEIVED 06 March 2023 ACCEPTED 22 May 2023 PUBLISHED 13 June 2023 CITATION Wang Y, Guo L, Zhang Z, Fu S, Huang P, Wang A, Liu M and Ma X (2023) A bibliometric analysis of myocardial ischemia/reperfusion injury from 2000 to 2023. Front. Cardiovasc. Med. 10:1180792. doi: 10.3389/fcvm.2023.1180792 COPYRIGHT © 2023 Wang, Guo, Zhang, Fu, Huang, Wang, Liu and Ma. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Introduction The heart is a vital organ in maintaining the body’s circulatory system, and the heart muscle requires an adequate supply of blood and oxygen to maintain its function. Myocardial infarction (MI) is a consequence of coronary artery occlusion, resulting in irreversible damage to the myocardium due to ischemia and hypoxia, and poses a severe threat to human health, with high rates of disability and mortality worldwide (1–3). As early as the 1970s, Ginks et al. (4) performed myocardial ischemia-reperfusion mapping in dogs and found that reperfusion therapy was effective in restoring blood ﬂow and reducing myocardial injury after MI. After years of research and observation, reperfusion therapy such as percutaneous coronary intervention (PCI) and coronary artery bypass grafting (CABG) have become the ﬁrst-line treatment strategy for MI (5, 6). However, it has also been found that these therapies may induce myocardial, vascular, or electrophysiological dysfunction, leading to worsened cardiac function, and is responsible for up to 50% of the ﬁnal infarct size. This phenomenon is known as myocardial Frontiers in Cardiovascular Medicine 01 frontiersin.org Wang et al. 10.3389/fcvm.2023.1180792 10.3389/fcvm.2023.1180792 ischemia/reperfusion injury (MIRI), which reduces the efﬁcacy of myocardial reperfusion therapy (7–10). injur”). The search was conducted from January 1, 2000 to January 7, 2023, with article and review types selected, and English language limited. A total of 5,595 papers were obtained. The results were exported in txt. format as “Full Record and Cited References”. To prevent data deviation due to database updates, the data search and export were completed on January 7, 2023. Although more effective reperfusion techniques and drugs that improve MIRI have emerged, the incidence of secondary myocardial damage after blood ﬂow restoration remains high due to narrow intervention windows and individual differences in susceptibility to reperfusion injury (11, 12). Therefore, the pathogenesis and prevention of MIRI remain a research hotspot in the cardiovascular ﬁeld. Results The data for this study was obtained from WoSCC. In order to more accurately capture the topic, we conducted subject searches in SCI-Expanded, SSCI, CCR-Expanded, IC, and ESCI. The search formula for this study was set as follows: TS = (myocardial NEAR/ 1 “reperfusion injur”) OR TS = (Cardiac NEAR/1 “reperfusion Analysis method Bibliometrics presents the knowledge structure and frontier trends of a research ﬁeld by modern techniques to visualize countries, institutions, authors, journals, documents and keywords (13, 14). Therefore, we reviewed the literature on Myocardial Ischemia-Reperfusion in the Web of Science Core Collection (WoSCC) database to provide a reference for future research on Myocardial Ischemia-Reperfusion. We primarily employed VOSviewer for data visualization, in conjunction with Excel, CiteSpace 6.1.R6, and Pajek 5.16 (15, 16). Firsty import the data into the CiteSpace software and check the duplicate data. Second, Synonyms are modiﬁed and merged before each visualization to show results more accurately. Third, we selected an appropriate number of nodes and set corresponding “Layout” parameters in VOSviewer, while leaving other options at their default values. We select the appropriate number of nodes for data visualization. Contribution of institutions A total of 3,840 institutions participated in research on MIRI after merging and eliminating meaningless nodes. Statistics and visual analysis were performed on 134 institutions with a publication volume of more than 15. Table 2 shows the top 10 institutions by publication volume, of which 9 are from China, but their centrality is low (<0.1). Harvard University is the only research institution that exceeds 0.1 and has the highest citation volume (7,372 times). It is worth noting that although University College London has only 59 publications, it ranks third in terms of citation volume (6,295 times). According to Figure 4A and Figure 4B, it can be seen that European and American countries, represented by Harvard University with relatively stable publication volume, began research in this area earlier; while Chinese research institutions have shown a signiﬁcant ﬂuctuating growth trend in the past decade. Analysis of annual publications distribution According to the search results, a total of 5,595 papers related to MIRI were collected by WoSCC from January 1st, 2000 to FIGURE 1 Flowchart of literature selection. Frontiers in Cardiovascular Medicine 02 frontiersin.org Wang et al. 10.3389/fcvm.2023.1180792 January 7th, 2023, as shown in Figure 1. Among them, there were 4,732 articles and 863 reviews, with a total citation count of 160,289 (excluding self-citations), an average citation frequency of 33.51 times per article, and an h-index of 165. Figure 2 shows the annual publication and citation volume. Overall, research related to MIRI showed an increasing trend, with the highest number of publications (589) and citations (24,492) in 2021. the degree of cooperation, which are represented by different colors. The connections between countries/regions are mainly focused on the cooperation between the US and other countries, including China, Germany and Japan. January 7th, 2023, as shown in Figure 1. Among them, there were 4,732 articles and 863 reviews, with a total citation count of 160,289 (excluding self-citations), an average citation frequency of 33.51 times per article, and an h-index of 165. Figure 2 shows the annual publication and citation volume. Overall, research related to MIRI showed an increasing trend, with the highest number of publications (589) and citations (24,492) in 2021. TABLE 1 Top 10 productive countries/regions in the ﬁeld of MIRI. TABLE 2 Top 10 productive institutions in the ﬁeld of MIRI. TABLE 1 Top 10 productive countries/regions in the ﬁeld of MIRI. TABLE 1 Top 10 productive countries/regions in the ﬁeld of MIRI. Rank Country/Region Publications Citations Centrality 1 CHINA 2,845 56,807 0.19 2 USA 1,325 75,508 0.31 3 GERMANY 305 15,967 0.13 4 JAPAN 278 13,349 0.01 5 ENGLAND 233 18,535 0.19 6 ITALY 182 9,002 0.16 7 CANADA 135 8,834 0.03 8 NETHERLANDS 117 6,749 0.11 9 FRANCE 109 4,486 0.1 10 AUSTRALIA 108 3,337 0.02 USA, the United States of America. FIGURE 3 Distribution of countries/regions involved in MIRI. (A) The annual number of papers published by the top 10 countries. (B) Cooccurrence diagram of countries/regions with more than 5 publications. The size of the nodes corresponds to the publication volume of countries/regions. The lines represent the connections between countries/regions, and their thickness represents the link strength. Different colors represent different clusters. TABLE 2 Top 10 productive institutions in the ﬁeld of MIRI. Rank Institution Publications Citations Centrality 1 Fourth Mil Med Univ 153 6,427 0.04 2 Wuhan Univ 127 2,969 0.05 3 Chinese Acad Med Sci 115 2,279 0.09 4 Harvard Univ 108 7,372 0.28 5 Nanjing Med Univ 98 2,145 0.03 6 Capital Med Univ 96 1,729 0.04 7 Fudan Univ 91 1,971 0.08 8 Huazhong Univ Sci & Technol 91 1,840 0.06 9 Shanghai Jiao Tong Univ 89 2,461 0.03 10 Cent S Univ 84 1,640 0.02 Frontiers in Cardiovascular Medicine 04 frontiersin.org Rank Institution Publications Citations Centrality 1 Fourth Mil Med Univ 153 6,427 0.04 2 Wuhan Univ 127 2,969 0.05 3 Chinese Acad Med Sci 115 2,279 0.09 4 Harvard Univ 108 7,372 0.28 5 Nanjing Med Univ 98 2,145 0.03 6 Capital Med Univ 96 1,729 0.04 7 Fudan Univ 91 1,971 0.08 8 Huazhong Univ Sci & Technol 91 1,840 0.06 9 Shanghai Jiao Tong Univ 89 2,461 0.03 10 Cent S Univ 84 1,640 0.02 Rank Country/Region Publications Citations Centrality 1 CHINA 2,845 56,807 0.19 2 USA 1,325 75,508 0.31 3 GERMANY 305 15,967 0.13 4 JAPAN 278 13,349 0.01 5 ENGLAND 233 18,535 0.19 6 ITALY 182 9,002 0.16 7 CANADA 135 8,834 0.03 8 NETHERLANDS 117 6,749 0.11 9 FRANCE 109 4,486 0.1 10 AUSTRALIA 108 3,337 0.02 USA the United States of America FIGURE 3 Distribution of countries/regions involved in MIRI. Contribution of countries/regions A total of 81 countries/regions participated in research on MIRI. Table 1 and Figure 3A show the top 10 countries/regions in terms of publication volume, as well as their citation counts, centrality, and annual publication volume. According to the statistics, China (2,845 papers) surpassed the United States (1,325 papers) in publication volume since 2011, followed by Germany (305 papers), Japan (278 papers), and England (233 papers). The United States had the highest citation count (75,508 times), followed by China (56,807 times), England (18,535 times), Germany (15,967 times), and Japan (13,349 times). The citation frequency of other countries was less than 10,000. Centrality of countries/regions is an important indicator of their importance. From the perspective of centrality, the United States, China, England, Italy, and Germany have high centrality and play important roles in this ﬁeld. In this ﬁeld, the cooperation between institutions is relatively close and international exchanges are frequent. For example, the Fourth Military Medical University has cooperated with other institutions 80 times, especially with Temple University (14 times). In addition, two academic groups have been formed, one represented by Fudan Univ, Chinese Acad Med Sci and Nanjing Med Univ, and the other by Wuhan Univ and Huazhong Univ Sci & Technol, as shown in Figure 4C. Figure 3B shows a visualization analysis of 52 countries/ regions with publication volume exceeding 5 papers. The size of the nodes corresponds to the publication volume, and the lines represent the connections between countries/regions. The countries/regions are roughly divided into six clusters based on FIGURE 2 Trends in the growth of publications and the number of cited papers worldwide from 2000 to 2023. The data for 2023 are not complete. Frontiers in Cardiovascular Medicine 03 frontiersin.org Wang et al. 10.3389/fcvm.2023.1180792 TABLE 1 Top 10 productive countries/regions in the ﬁeld of MIRI. Rank Country/Region Publications Citations Centrality 1 CHINA 2,845 56,807 0.19 2 USA 1,325 75,508 0.31 3 GERMANY 305 15,967 0.13 4 JAPAN 278 13,349 0.01 5 ENGLAND 233 18,535 0.19 6 ITALY 182 9,002 0.16 7 CANADA 135 8,834 0.03 8 NETHERLANDS 117 6,749 0.11 9 FRANCE 109 4,486 0.1 10 AUSTRALIA 108 3,337 0.02 USA, the United States of America. TABLE 2 Top 10 productive institutions in the ﬁeld of MIRI. Contribution of countries/regions Rank Institution Publications Citations Centrality 1 Fourth Mil Med Univ 153 6,427 0.04 2 Wuhan Univ 127 2,969 0.05 3 Chinese Acad Med Sci 115 2,279 0.09 4 Harvard Univ 108 7,372 0.28 5 Nanjing Med Univ 98 2,145 0.03 6 Capital Med Univ 96 1,729 0.04 7 Fudan Univ 91 1,971 0.08 8 Huazhong Univ Sci & Technol 91 1,840 0.06 9 Shanghai Jiao Tong Univ 89 2,461 0.03 10 Cent S Univ 84 1,640 0.02 TABLE 1 Top 10 productive countries/regions in the ﬁeld of MIRI. The lines represent the connections between institutions, and their thickness represents the link strength. Different colors represent different clusters. (continued) Yellon, Derek M. (6,774), followed by Hausenloy, Derek J. (6,601), as shown in Table 3. Frontiers in Cardiovascular Medicine TABLE 1 Top 10 productive countries/regions in the ﬁeld of MIRI. (A) The annual number of papers published by the top 10 countries. (B) Cooccurrence diagram of countries/regions with more than 5 publications. The size of the nodes corresponds to the publication volume of countries/regions. The lines represent the connections between countries/regions, and their thickness represents the link strength. Different colors represent different clusters. FIGURE 3 Distribution of countries/regions involved in MIRI. (A) The annual number of papers published by the top 10 countries. (B) Cooccurrence diagram of countries/regions with more than 5 publications. The size of the nodes corresponds to the publication volume of countries/regions. The lines represent the connections between countries/regions, and their thickness represents the link strength. Different colors represent different clusters. Frontiers in Cardiovascular Medicine 04 frontiersin.org Wang et al. 10.3389/fcvm.2023.1180792 FIGURE 4 Distribution of institutions involved in MIRI. (A) The annual number of papers published by the top 10 institutions. (B) The time-overlay map of the cooperation network among the institutions. The color of an item is determined by the average year, where colors range from blue (2000y) to green to yellow (2023y). (C) Cooccurrence diagram of institutions with more than 15 publications. The size of the nodes corresponds to the publication volume of institutions. The lines represent the connections between institutions, and their thickness represents the link strength. Different colors represent different clusters. (continued) FIGURE 4 Distribution of institutions involved in MIRI. (A) The annual number of papers published by the top 10 institutions. (B) The time-overlay map of the cooperation network among the institutions. The color of an item is determined by the average year, where colors range from blue (2000y) to green to yellow (2023y). (C) Cooccurrence diagram of institutions with more than 15 publications. The size of the nodes corresponds to the publication volume of institutions. The lines represent the connections between institutions, and their thickness represents the link strength. Different colors represent different clusters. (continued) FIGURE 4 Distribution of institutions involved in MIRI. (A) The annual number of papers published by the top 10 institutions. (B) The time-overlay map of the cooperation network among the institutions. The color of an item is determined by the average year, where colors range from blue (2000y) to green to yellow (2023y). (C) Cooccurrence diagram of institutions with more than 15 publications. The size of the nodes corresponds to the publication volume of institutions. frontiersin.org Yellon, Derek M. (6,774), followed by Hausenloy, Derek J. (6,601), as shown in Table 3. Over the past 20 years, a total of 26,202 authors have participated in research related to MIRI. Among them, 168 authors with more than 10 publications were selected for visualization analysis. The largest number of papers was published by Lefer, David J. and Xia, Zhengyuan (45), followed by Gao, Erhe and Yang, Jian (42). The most co-cited author was The collaboration among the authors of MIRI-related literature was displayed in VOSviewer. The same cluster often represents close collaboration and provides information for ﬁnding research partners. Several academic groups with relatively ﬁxed collaborations have emerged in this ﬁeld, as shown in Figure 5A. Furthermore, we can see that academic groups represented by Frontiers in Cardiovascular Medicine 05 frontiersin.org Wang et al. 10.3389/fcvm.2023.1180792 FIGURE 4 Continued. FIGURE 4 Xia Zhengyuan, Gao Erhe, Yang Jian have more frequently with the outside world, while academic groups represented by Lefer David J. and Schulz Rainer are relatively isolated. into 4 clusters: Hausenloy Derek J, Heusch Gerd etc. (blue); Bolli Roberto, Zhao ZQ etc. (green); Zhou Hao, Zhang Y etc. (red); Halestrap AP, Murry CE etc. (yellow). When two or more authors are cited by the same article, a co- citation relationship exists. The size of the node in the map represents the total frequency of co-citation. The larger the node, the more frequently it is co-cited, indicating greater inﬂuence in the ﬁeld. Figure 5B shows that the research hotspots of the authors are highly homogeneous. The authors are mainly divided Frontiers in Cardiovascular Medicine Highly co-cited references Co-citation analysis is a dynamic process that changes over time. It is used to study the internal connections between TABLE 3 Top 10 productive authors and co-cited authors in the ﬁeld of MIRI. Rank Author Publication Country Institution Co-cited Author Citation Country Institution 1 Lefer David J. 45 USA Louisiana State University Health Sciences Center New Orleans Hausenloy, Derek J. 2,100 ENGLAND University of London 2 Xia Zhengyuan 45 Hong Kong, China University of Hong Kong Heusch, Gerd 1,180 Germany University of Duisburg Essen 3 Gao Erhe 42 USA Pennsylvania Commonwealth System of Higher Education Yellon, Derek M. 944 ENGLAND University of London 4 Yang Jian 42 China China Three Gorges University Bolli Roberto 592 USA Univ Louisville 5 Yang Yang 38 China Fourth Mil Med Univ Zhao Zq 536 USA Emory Univ 6 Schulz Rainer 33 Germany Univ Giessen Zhou, Hao 482 China Chinese Peoples Liberat Army Gen Hosp 7 Hausenloy Derek J. 31 ENGLAND University of London Kloner Ra 468 USA University of Southern California 8 Zhang Jing 30 China Fourth Mil Med Univ Frangogiannis Ng 459 USA Albert Einstein Coll Med 9 Yellon Derek M. 29 ENGLAND University of London Zhang Y 445 China Peking Univ 10 Ma Xinliang 28 USA Jefferson University Ferdinandy, Peter 393 Hungary Univ Szeged TABLE 3 Top 10 productive authors and co-cited authors in the ﬁeld of MIRI. Frontiers in Cardiovascular Medicine 06 frontiersin.org frontiersin.org Wang et al. 10.3389/fcvm.2023.1180792 GURE 5 stribution of authors involved in MIRI. (A) Cooccurrence diagram of authors. with more than 10 publications. The size of the nodes corresponds to the blications by the author, and the connections between the nodes reﬂect the collaboration relationship. Different colors represent different clusters. (B) ooccurrence diagram of co-cited authors. The size of the node in the map represents the total frequency of co-citation. The larger the node, the more quently it is co-cited. FIGURE 5 Distribution of authors involved in MIRI. (A) Cooccurrence diagram of authors. with more than 10 publications. The size of the nodes corresponds to the publications by the author, and the connections between the nodes reﬂect the collaboration relationship. Different colors represent different clusters. (B) Cooccurrence diagram of co-cited authors. The size of the node in the map represents the total frequency of co-citation. The larger the node, the more frequently it is co-cited. The papers were divided into three clusters. Frontiers in Cardiovascular Medicine Highly co-cited references The green cluster was led by “Mechanisms of disease: Myocardial reperfusion injury” (Yellon DM, 2007), with the highest number of co- citations (631 times). It mainly described four types of cardiac dysfunction caused by reperfusion injury, and summarizes the reasons for the discrepancies in outcomes of single-target literature and depict the dynamic structure of scientiﬁc development. The top 10 co-cited papers, totaling 11 papers, are listed in Table 4. In addition, 94 papers with co-citations exceeding 60 times were subjected to visual analysis, where the size of the nodes was proportional to the number of co-citations (Figure 6). 07 frontiersin.org Wang et al. 10.3389/fcvm.2023.1180792 Rank Title Author Year Journal IF 1 Mechanisms of disease: myocardial reperfusion injury Yellon, Derek M. 2007 NEW ENGLAND JOURNAL OF MEDICINE 176.082 2 Myocardial ischemia-reperfusion injury: a neglected therapeutic target Hausenloy, Derek J. 2013 JOURNAL OF CLINICAL INVESTIGATION 19.477 3 Preconditioning with ischemia: a delay of lethal cell injury in ischemic myocardium Murry, Charles E. 1986 CIRCULATION 39.922 4 Inhibition of myocardial injury by ischemic postconditioning during reperfusion: comparison with ischemic preconditioning Zhao, ZQ 2003 AMERICAN JOURNAL OF PHYSIOLOGY- HEART AND CIRCULATORY PHYSIOLOGY 5.125 5 Evolving Therapies for Myocardial Ischemia/Reperfusion Injury Ibanez, Borja 2015 JOURNAL OF THE AMERICAN COLLEGE OF CARDIOLOGY 27.206 6 Mechanisms underlying acute protection from cardiac ischemia- reperfusion injury Murphy, Elizabeth 2008 PHYSIOLOGICAL REVIEWS 46.513 7 Ischemia and reperfusion-from mechanism to translation Eltzschig, Holger K. 2011 NATURE MEDICINE 87.244 8 Myocardial reperfusion: a double-edged sword? Braunwald, Eugene 1985 JOURNAL OF CLINICAL INVESTIGATION 19.477 9 New directions for protecting the heart against ischaemia-reperfusion injury: targeting the Reperfusion Injury Salvage Kinase (RISK)-pathway Hausenloy, Derek J. 2004 CARDIOVASCULAR RESEARCH 14.239 10 Distinct roles of autophagy in the heart during ischemia and reperfusion— roles of AMP-activatedprotein kinase and Beclin 1 in mediatingautophagy Matsui, Yutaka 2007 CIRCULATION RESEARCH 23.218 10 Effect of cyclosporine on reperfusion injury in acute myocardial infarction Piot, Christophe 2008 NEW ENGLAND JOURNAL OF MEDICINE 176.082 microvascular obstruction (MVO), and lethal myocardial reperfusion injury. It discussed in detail pathological mechanisms such as oxidative stress, calcium overload, pH value correction, mPTP, inﬂammation, and cell apoptosis, as well as new therapies. However, the article did not afﬁrm the view that MI area would increase with prolonged reperfusion time. In addition, the study conﬁrmed the position of cardiac magnetic resonance (CMR) imaging in the diagnosis and efﬁcacy evaluation of MIRI (17). Highly co-cited references interventions for MIRI in animal models and clinical studies. The study also conﬁrmed new strategies to prevent lethal reperfusion injury by reperfusion injury salvage kinase (RISK) pathway and mitochondrial permeability transition pore (mPTP) (1). The red cluster was led by “Myocardial ischemia-reperfusion injury: a neglected therapeutic target” (Hausenloy DJ, 2013), with the highest number of co-citations (529 times). The article identiﬁed four recognized forms of MIRI, namely, reperfusion-induced arrhythmias, myocardial stunning, FIGURE 6 Cooccurrence diagram of references with co-citations exceeding 60 times. The size of the node indicates the cooccurrence frequencies of references, and the link reﬂects the cooccurrence relationship. The color of the node represents the respective cluster. FIGURE 6 Cooccurrence diagram of references with co-citations exceeding 60 times. The size of the node indicates the cooccurrence frequencies of references, and the link reﬂects the cooccurrence relationship. The color of the node represents the respective cluster. Frontiers in Cardiovascular Medicine 08 frontiersin.org Wang et al. Wang et al. 10.3389/fcvm.2023.1180792 Analysis of research hotspots The blue cluster was led by “Preconditioning with ischemia: a delay of lethal cell injury in ischemic myocardium” (Murry CE, 1986), followed by “Inhibition of myocardial injury by ischemic postconditioning during reperfusion: comparison with ischemic preconditioning” (Zhao ZQ, 2003) with 322 and 238 co-citations, respectively. The former was proposed by Murry et al. (18) using a canine model, which found that multiple brief ischemic episodes over a period of time could protect the heart from subsequent sustained ischemic injury, thereby introducing the concept of ischemic preconditioning. The latter, proposed by Zhi-Qing Zhao et al. (19), compared the effects of ischemic postconditioning and ischemic preconditioning and demonstrated that both were equally effective in reducing infarct size and protecting endothelial function. The keywords summarize the research topic of a paper and can be used to analyze the research hotspots and directions in the ﬁeld of MIRI. Before visualization, synonyms (e.g., salvia miltiorrhiza and danshen), different spellings (e.g., ischemia, ischemic, and ischaemic), abbreviations (eg, IL-6 and interleukin-6), and singular/ plural forms (e.g., arrhythmia and arrhythmias) should be merged. In VOSviewer, the keyword threshold was set at 15, resulting in a total of 142 keywords. The most frequent keyword was “MIRI” (2,918 times), followed by “myocardial ischemia” (734 times) and “apoptosis” (700 times), as shown in Table 5. These keywords can be divided into four different directions: (1) keywords related to risk factors are diabetes mellitus, hyperlipidemia, hypertension, and aging; (2) keywords related to poor prognosis, such as arrhythmia, myocardial stunning, cardiac function, and heart failure; (3) the study of pathological and physiological mechanisms mainly revolves around cell death, oxidative stress, inﬂammation, endoplasmic reticulum and mitochondria, non-coding RNAs (miRNA, lncRNA), and biomarkers, involving hot signaling pathways such as the PI3K/ Distribution of journals The dual-map overlay of journals reveal the relatives position of the topic of study to the main research disciplines. Each point on the map represents a journal, with the citation graph on the left and the cited graph on the right. The curve represents the validation line, with different colors representing different citation relationships. TABLE 5 Top 20 keywords in the ﬁeld of MIRI. TABLE 5 Top 20 keywords in the ﬁeld of MIRI. Rank Keywords n Rank Keywords n 1 MIRI 2,918 11 heart 205 2 myocardial ischemia 734 12 ROS 203 3 apoptosis 700 13 diabetes mellitus 162 4 MI 612 14 PI3K/AKT 158 5 cardioprotection 431 15 NO 150 6 oxidative stress 395 16 H/R injury 120 7 inﬂammation 336 17 antioxidants 112 8 cardiac myocytes 261 18 cardiovascular diseases 105 9 mitochondria 224 19 heart failure 102 10 autophagy 212 20 ER stress 91 Figure 7 identiﬁes three main paths, indicating that papers published in the “4 Molecular Biology, Biology, and Immunology” journal primarily reference papers in the ﬁelds of “8 Molecular Biology, Biology, and Genetics” and “5 Health, Nursing, and Medicine”. In addition, papers published in journals such as “8 Molecular Biology, Biology, and Genetics” are also commonly cited in papers in the “2 Medicine, Medicine, and Clinical” ﬁeld. Currently, research on MIRI is mainly focused on clinical and molecular biology aspects. FIGURE 7 The dual-map overlay of journals. Each node on the map represents a journal, with the citation graph on the left and the cited graph on the right. The curve represents the validation line, with different colors representing different citation relationships. The dual-map overlay of journals. Each node on the map represents a journal, with the citation graph on the left and the cited graph on the right. The curve represents the validation line, with different colors representing different citation relationships. Frontiers in Cardiovascular Medicine Frontiers in Cardiovascular Medicine 09 frontiersin.org Wang et al. 10.3389/fcvm.2023.1180792 FIGURE 8 VOSviewer overlay visualization of keywords. Each column represents a cluster, and the color represents the average time. The closer the blue item is, the earlier it appears. AKT pathway, Nf-κB pathway, and TLRs signaling pathway; (4) the main treatment-related keywords are cardiac protection, ischemic preconditioning, ischemic postconditioning, melatonin, dexmedetomidine, resveratrol, and Danshen, among others. since 2003. Distribution of journals Since 2021, it has further increased to over 500 papers per year, indicating that research on MIRI remains a major focus in the cardiovascular ﬁeld. Among them, the total number of papers published in China is more than 2,850, accounting for more than half of the total publication output. Especially since 2013, the number of publications has rapidly increased, indicating that research on MIRI has received increasing attention from the Chinese in the past decade. However, there is still a problem of insufﬁcient inﬂuence. As one of the earliest institutions to begin research on myocardial ischemia reperfusion injury, the United States is another country with more than 1,000 publications, with the highest centrality. Among the top 10 most co-cited authors, 4 are from the United States. In the overlay visualization of keyword time series (Figure 8), each column represents a cluster, and the color represents the average time. The closer the color is to blue, the more frequently the keyword appeared in the early stage, and the closer to yellow, the more frequently the keyword appeared in recent years, which can reﬂect the research hotspot in a ﬁeld to some extent. Keywords such as ferroptosis (2021.472), pyroptosis (2021.095), NLRP3 (2020.4), lncRNA (2020.25), mitochondrial dynamics (2020.2353), exosomes (2019.9643), mitophagy (2019.804), sirtuins (2019.7273), and necroptosis (2019.6522) have been frequently appearing in recent years, indicating that they are hotspots in the ﬁeld of MIRI in recent years. It is worth noting that traditional Chinese medicine (TCM) (2019.2727) has gradually attracted attention worldwide for its role in treating MIRI. Representative drugs and effective ingredients include danshen, berberine, ﬂavonoids and so on. In terms of cooperative relationships, frequent collaborations exist between countries/regions and institutions. For example, European and American academic institutions represented by Harvard University and University College London, and Chinese academic institutions represented by Fudan University and the Chinese Academy of Medical Sciences, all maintaining frequent collaborative relationships with other institutions. Frontiers in Cardiovascular Medicine On the prognosis On the prognosis, research indicates that reperfusion injury accounts for up to 50% of the ﬁnal myocardial damage in acute MI (37). Common sequelae of reperfusion injury include heart failure, remodeling, arrhythmias, and myocardial stunning, which are consistent with our survey results. Acute ST-segment elevation myocardial infarction (STEMI) patients who undergo primary PCI are prone to develop ventricular arrhythmias following ischemia-reperfusion injury, which are usually easily managed or self-terminated (38). The underlying cause may be related to the instability of action potential resulting from the inability of mitochondria to recover or maintain their inner membrane potential after prolonged ischemia (39–41). Moreover, this phenomenon is more common and severe in elderly female rats (42), which may be associated with the decline of estrogen receptors and antioxidant activity in their myocardium, rather than the decrease in serum estrogen levels (43). Myocardial stunning refers to systolic and diastolic dysfunction in patients with acute myocardial ischemia after reperfusion, and the severity is proportional to the duration of ischemia (17). The mechanism of myocardial stunning is relatively mature. It mainly attributes to the massive formation of reactive oxygen species and calcium overload in cardiomyocytes and microvascular endothelial cells after reperfusion, which leads to endothelial dysfunction and decreased responsiveness to calcium in the excitation-contraction coupling mechanism. However this process is entirely reversible and generally lasts for several hours or days (44–46). Acute ST-segment elevation myocardial infarction (STEMI) patients who undergo primary PCI are prone to develop ventricular arrhythmias following ischemia-reperfusion injury, which are usually easily managed or self-terminated (38). The underlying cause may be related to the instability of action potential resulting from the inability of mitochondria to recover or maintain their inner membrane potential after prolonged ischemia (39–41). Moreover, this phenomenon is more common and severe in elderly female rats (42), which may be associated with the decline of estrogen receptors and antioxidant activity in their myocardium, rather than the decrease in serum estrogen levels (43). Myocardial stunning refers to systolic and diastolic dysfunction in patients with acute myocardial ischemia after reperfusion, and the severity is proportional to the duration of ischemia (17). The mechanism of myocardial stunning is relatively mature. It mainly attributes to the massive formation of reactive oxygen species and calcium overload in cardiomyocytes and microvascular endothelial cells after reperfusion, which leads to endothelial dysfunction and decreased responsiveness to calcium in the excitation-contraction coupling mechanism. Regarding the mechanisms Regarding the mechanisms, we found that the research ﬁelds with higher output focus on regulated cell death, oxidative stress, inﬂammatory response, non-coding RNA, as well as mitochondrial and endoplasmic reticulum stress. They form an intertwined association between different pathways that affect MIRI by regulating common pathway molecules. As the main type of cell death during reperfusion, regulated cell death plays an important role in the pathogenesis of MIRI (54). In our VOSviewer visualization, apoptosis has received the most attention. Nevertheless, non-apoptotic forms of regulated cell death, such as ferroptosis, pyroptosis, necroptosis, and autophagy have increasingly received the attention of researchers in recent years. They can operate alone or coexist with other forms of cell death, thereby playing a role in the enlargement of infarct size and the deterioration of heart function caused by ischemia- reperfusion (26, 55). The generation of reactive oxygen species (ROS) is the central pathogenic mechanism of MIRI. Ferroptosis is an iron overload and iron-dependent ROS accumulation process, and its regulatory mechanisms involve multiple signaling pathways and metabolic pathways, especially the glutathione peroxidase 4 (GPX4) axis (10, 26, 56). During myocardial ischemia-reperfusion, the increase in intracellular free iron levels and the decrease in GPX4 activity lead to the massive release of ROS within myocardial cells, causing lipid peroxidation (57–60). In addition, under the chemotactic effect of ROS, neutrophils accumulate in infarcted myocardial tissue several hours after reperfusion, and ferroptosis also promotes this process through Toll-like receptor 4-dependent signaling pathways, triggering harmful inﬂammatory responses and ultimately leading to cell death (17, 61). Studies have shown that this process mainly occurs during the reperfusion phase of MIRI rather than the ischemic phase (62, 63). It has been found that ferroptosis can be effectively inhibited by ferrostatin 1, liproxstatin 1, iron chelators, and antioxidants during reperfusion, which can protect against myocardial injury, reduce infarct size, and improve cardiac function in acute or chronic MIRI (62, 64–69). In addition, MVO as a potentially preexisting risk factor worsens within minutes after reperfusion and persists for at least 1 week (31), resulting in myocardial damage due to inadequate perfusion, known as the “no-reﬂow” phenomenon (32–34). Approximately 50% of acute MI patients experience ischemia- reperfusion injury in cardiac microvascular endothelial cells (CMECs), which is the main factor leading to the ﬁnal infarct size and adverse cardiovascular outcomes (35, 36). In risk factors In risk factors, the most common keywords related to MIRI are diabetes mellitus, hyperlipidemias, hypertension, and aging, all of which have been reported to be associated with MIRI (20–23). In terms of publications, the quantity of papers on myocardial ischemia- reperfusion injury has surpassed 100 papers per year Frontiers in Cardiovascular Medicine 10 frontiersin.org Wang et al. 10.3389/fcvm.2023.1180792 Diabetes mellitus is a common comorbidity in cardiovascular disease patients and increases the risk of cardiovascular disease by 2–4 times compared to non-diabetic patients (24, 25). However, the controversy remains as to whether it increases the susceptibility of the heart to ischemia-reperfusion injury (26). In the co-morbidity study of MIRI and diabetes mellitus, a popular target is AMPK, which has been found to improve cardiovascular complications related to diabetes mellitus by stimulating AMPK. The mechanism mainly involves the downregulation of AMPK in the heart tissue of animals and humans with type 2 diabetes mellitus or metabolic syndrome, leading to energy metabolism disorder, aggravated inﬂammation, and intensiﬁed cellular autophagy, apoptosis, ferroptosis, and necroptosis pathways (27–30). developing heart failure or death remains high (51, 52). A questionnaire survey involving 850 patients showed that 2 out of every 5 patients had heart failure-related quality of life impairment after MI, causing a signiﬁcant social and economic burden (53). Frontiers in Cardiovascular Medicine Cardioprotection Multiple studies prove autophagy is primarily a pro-survival mechanism during short-term ischemia and hypoxia (9, 85). When the supply of oxygen and nutrients to cardiomyocytes is reduced and ATP is depleted quickly, the AMPK/mTOR pathway is activated (86, 87). At this time, cardiomyocytes utilize autophagy to degrade excessive or potentially dangerous cytosolic entities, such as damaged organelles or misfolded proteins, and acquire metabolic substrates to increase ATP production (88, 89). At the same time, an appropriate level of autophagy can reduce ROS production, decrease NLRP3-related inﬂammatory responses, and decrease other types of cell death, including necroptosis and apoptosis (90). However, research indicates that autophagy plays a dual role in MIRI, depending on the degree of its activation (91, 92). Unlike the ischemic phase, ROS accumulation is deemed the primary factor affecting autophagic ﬂux during the reperfusion phase (93, 94). Elevated levels of ROS during reperfusion cause the opening of the mitochondrial permeability transition pore (MPTP), which promotes ROS release, activates Bnip3 (95–97), and induces the expression of the autophagy-related protein Beclin1 (93, 98). In normal conditions, the anti-apoptotic protein Bcl-2 binds to Beclin1, preventing autophagy. However, this balance could be disrupted by Bnip3, promoting autophagosome formation and increasing the autophagy rate (99, 100), eventually leading to cell death caused by excessive degradation of cellular components (54). Nevertheless, some perspectives propose that ischemia-reperfusion injury is associated with deﬁciencies in autophagosome-lysosome fusion (88, 101), which lead to cell death by impaired clearance of autophagosomes. MI commonly causes two processes of myocardial injury, the ﬁrst occurring during ischemia and the second possibly after reperfusion. The following treatment strategies can be summarized in light of these two processes. The ﬁrst is mechanical ischemic conditioning, including brief ischemia- reperfusion cycles in the heart or tissues away from the heart, which can be achieved by ischemic preconditioning or postconditioning methods (138–141). The second strategy involves drug therapy proven to protect the myocardium. For patients with pre-existing coronary artery disease, long-term and standardized treatment has been shown to effectively prevent major adverse cardiovascular events (MACE). Such as aspirin and ticagrelor, which can prevent reperfusion injury when given before reperfusion and effectively limit the area of MI (142, 143). The protective effect of simvastatin on contractile function in acute MIRI models may be related to the inhibition of the RhoA/ ROCK pathway. On the prognosis However this process is entirely reversible and generally lasts for several hours or days (44–46). Pyroptosis is a highly inﬂammatory form of cell death contributing to ischemia-reperfusion injury when overactivated. After reperfusion, the increased levels of calcium ions and ROS lead to the formation of NLRP3 inﬂammasomes that activate caspases, initiating the pyroptotic pathway (26, 70). On the hand, activated caspase-1/11 can activate Gasdermin D (GSDMD), a pore-forming protein that mediates cell death, increasing cell permeability and resulting in cell lysis typically (71–73). On the other hand, activated caspases cleave IL-1β and IL-18, releasing them outside the cell through GSDMD membrane pores, further triggering inﬂammation (26, 74). However, Shi et al. (75) found that caspase-11 may be the only pathway to trigger pyroptosis in cardiac myocytes. They also demonstrated that knocking out the Infarct size is the main determinant of patient prognosis, and MIRI may further increase the infarct size (47, 48). There are two recognized forms of irreversible MIRI: MVO and lethal myocardial reperfusion injury (17, 49, 50). In the early stages of MI, ventricular compensatory mechanisms are activated, and cardiac function remains normal or slightly reduced. As the infarct size expands and non-infarcted areas remodel, the risk of Frontiers in Cardiovascular Medicine 11 frontiersin.org Wang et al. 10.3389/fcvm.2023.1180792 inducing mRNA degradation or inhibiting their translation (104, 105). Previous studies have shown that during MIRI, miR-29c and miR-125a are signiﬁcantly downregulated (106, 107), while miR- 135b-3p is upregulated (108), playing roles in promoting autophagy and ferroptosis, respectively. In vivo/in vitro experiments showed that modulation of miR-1, miR-126, miR-140-3p, miR-214-5p, miR-125b and miR-24 could exert anti-apoptotic effects (109–114). Regulation of miR-133a, miR-15 exert anti-apoptotic effects (115– 117). Moreover, miRNAs can bind to several mRNA molecules, allowing them to play multiple cellular functions. For example, miR-29b binds to PTEN. Its overexpression can reduce PTEN expression level and increased the protein levels of p-Akt/Akt and p-eNOS/eNOS, thereby exerting Anti-oxidative stress, Anti- inﬂammatory and Anti-apoptosis effects (118). MiR-125a-5p targets KLF13, TGFBR1, and DAAM1, promoting M2 macrophage polarization, inhibiting ﬁbroblast proliferation and activation, and promoting angiogenesis, subsequently improving myocardial cell apoptosis and inﬂammation (107). Table 6 provides a summary of some important miRNAs. Table 6 provides a summary of some important miRNAs. GSDMD gene signiﬁcantly reduced the levels of LDH and IL-18 after hypoxia/reoxygenation, and reduced the area of MI induced by ischemia-reperfusion in mice. On the prognosis Necrosis has long been considered an uncontrolled form of cell death, but it has been found to occur in a regulated manner as well, known as necroptosis. Necroptosis involves the activation of the RIPK1/RIPK3/MLKL pathway and is recognized as another major programmed cell death type in MIRI now (76, 77). The process involves ischemia and oxidative stress-induced cardiac injury, as follows: the classical necroptotic pathway is usually initiated by the phosphorylation of RIPK1, which further phosphorylates RIPK3. The complex formed by RIPK1 and RIPK3 can induce MLKL oligomerization and translocation to the plasma membrane, leading to Ca2+ or Na+ ion inﬂux and directly forming a pore, releasing damage-associated molecular patterns and causing membrane rupture (78, 79). However, RIPK3-induced myocardial necrosis can also occur independently of RIPK1 (80). When MIRI occurs, RIPK3 can be directly activated. Then through the RIPK3- CaMKII or RIPK3-PGAM5-CypD cascade, promoting the opening of the mPTP and participating in multiple signaling pathways that induce myocardial death (26, 81–84). Frontiers in Cardiovascular Medicine Cardioprotection Research shows SGLT2 inhibitors can alleviate the damage of MI in diabetic and non-diabetic hearts (144–146), reduce MIRI by inhibiting cardiomyocyte autophagy and protecting mitochondrial function, and reduce cardiovascular mortality and heart failure (HF) rehospitalization rate of patients after myocardial ischemia-reperfusion by targeting multiple pathways (26, 146–148). Nicorandil is widely used in the treatment of coronary heart disease and has a dual effect as a nitric oxide (NO) donor and increases cell membrane permeability to potassium ions. It has been shown to alleviate oxidative stress, inﬂammation, and apoptosis induced by ischemia-reperfusion (149, 150). A recent randomized double-blind controlled trial showed that nicorandil administered before primary PCI could improve the myocardial perfusion grade and increase the ejection fraction, and reduce myocardial infarct size in patients with ST- segment elevation MI (151). In addition, the soluble guanylate cyclase (sGC) stimulator vericiguat has been shown to reduce MIRI by improving microcirculation (152). Moreover, there are As an entrance to molecular regulators, non-coding RNAs (including miRNAs, lncRNAs, circRNAs) affect cellular function through targeting various molecules in signalling pathways and have been widely studied in cardiovascular disease (102, 103). Among them, miRNAs are the most widely studied ncRNAs. Elevated levels of ROS can cause DNA damage and regulate miRNAs, which can negatively regulate gene expression by Frontiers in Cardiovascular Medicine 12 frontiersin.org Wang et al. 10.3389/fcvm.2023.1180792 TABLE 6 The regulatory role of MicroRNAs in MIRI. MicroRNAs Species Expression Targeted genes Mechanism Refs. Cardioprotection miR-1 rat ↓ Hsp90aa1 Pro-apoptosis (109) miR-125a rat/H9c2 ↓ DRAM2 Anti-oxidative stress, Anti-autophagy (107) miR-125a-5p mice ↓ KLF13, TGFBR1, DAAM1 Anti-inﬂammatory, Anti-apoptosis (119) miR-125b rat ↓ SIRT7 Anti-apoptosis (113) miR-126 rat ↑ ERRFI1 Anti-apoptosis (110) miR-128 mice/H9c2 ↓ Plk2 Anti-apoptosis (120) miR-128 rat/H9c2 ↓ TXNIP Anti-oxidative stress, Anti-apoptosis (121) miR-128-1-5p rat/H9c2 ↓ Gadd45g Anti-apoptosis (122) miR-129 rat/H9c2 ↓ PTEN Anti-apoptosis (123) miR-129 cardiomyocytes ↓ TLR4 Anti-inﬂammatory (124) miR-129-5p H9c2 ↓ TRPM7 Anti-inﬂammatory, Anti-apoptosis (125) miR-129-5p rat ↑ HMGB1 Anti-apoptosis (126) miR-133a rat/H9c2 ↓ ELAVL1 Anti-pyroptosis (116) miR-133a rat/H9c2 ↓ IGF1R Anti-apoptosis (115) miR-135b-3p rat/H9c2 ↑ GPX4 Pro-Ferroptosis (108) miR-138 mice ↓ EGR1 Anti-inﬂammatory (127) miR-138-5p mice ↓ Ltb4r1 Anti-inﬂammatory (128) miR-140 mice ↓ YES1 Anti-apoptosis (129) miR-140-3p H9c2 ↓ PTEN Anti-oxidative stress, Anti-apoptosis (111) miR-15b-5p rat/H9c2 ↑ Sirt3 Anti-pyroptosis (117) miR-155-5p mice ↑ JAK2/STAT1 Pro-inﬂammation (130) miR-155-5p mice ↑ NEDD4 Pro-apoptosis (131) miR-181a-5p cardiomyocytes ↑ ADCY1 Pro-pyroptosis (132) miR-182-5p rat/H9c2 ↑ STK17A Pro-oxidative stress (133) miR-21 mice ↓ SPP1 Anti-oxidative stress, Anti-inﬂammatory, Anti-apoptosis (134) miR-214-5p mice ↓ FASLG Anti-apoptosis (112) miR-24 rat/H9c2 ↑ Keap1 Anti-apoptosis (114) miR-29b rat/H9c2 ↓ PTEN Anti-oxidative stress, Anti-inﬂammatory, Anti-apoptosis (118) miR-29b-3p rat/H9c2 ↓ HMCN1 Anti-oxidative stress, Anti-ﬁbrosis, Anti-apoptosis (135) miR-29c mice ↓ PTEN Anti-autophagy (106) miR-30b mice ↓ CypD Anti-necrosis (136) miR-486 mice ↓ PTEN, FoxO1 Anti-apoptosis (137) reperfusion injury, as it signiﬁcantly improves postinjury cardiac function and increases drug utilization compared to other delivery methods (170). MicroRNAs use exosomes as a carrier to enable cell-to-cell communication (171). As previously mentioned, MIRI is often accompanied by abnormal expression of miRNAs. Enrichment of speciﬁc miRNAs by mesenchymal stem cell- secreted extracellular vesicles has shown promising results in regulating miRNA levels in cardiomyocytes in various preclinical experiments, making it a potential therapeutic approach (102, 172). It is worth noting that some miRNAs are dysregulated in multiple cardiovascular diseases. Therefore, selecting miRNAs that are dysregulated throughout the entire disease process leading to MIRI may have a stronger therapeutic effect. At the same time, it is essential to consider the safe and effective translation of preclinical experiments to clinical practice. increasing studies and reports on the use of single or compound traditional Chinese medicine in the prevention and treatment of MIRI (153, 154). Resveratrol can reduce oxidative stress levels, Fe2 + content and inhibit ferroptosis induced by ischemia-reperfusion (155). Yang et al. Cardioprotection (156) ﬁrst reported that neocryptotanshinone can promote autophagolysosome clearance of protein aggregates via the ERK1/2-Nrf2-LAMP2 pathway, exerting therapeutic advantages for MIRI. Other Chinese medicines and effective ingredients, such as Madder, Calenduloside E, and Tubeimoside I, can also reduce infarct size and alleviate MIRI through different mechanisms, such as reducing inﬂammation, oxidative stress, or inhibiting cell death (157–159). Thirdly, with the advancement of technology, research on novel therapeutic methods such as nanomedicine has gradually been carried out. Compared to free drugs, nanomedicine has better therapeutic effects and safety, attributed to its multifunctional carrier selection, targeted and controlled drug release, and improved bioavailability (160). Currently, common nanocarriers include liposomes (161–163), polymer nanoparticles (164, 165), inorganic nanoparticles (166) and extracellular vesicles (167–169). Carvedilol, a nonselective β-blocker, was encapsulated into platelet membrane vesicles (PMVs). Targeted administration of PMVs@Carvedilol may be a promising treatment for myocardial frontiersin.org Frontiers in Cardiovascular Medicine References 1. Yellon DM, Hausenloy DJ. Mechanisms of disease: myocardial reperfusion injury. N Engl J Med. (2007) 357(11):1121–35. doi: 10.1056/NEJMra071667 1. Yellon DM, Hausenloy DJ. Mechanisms of disease: myocardial reperfusion injury. N Engl J Med. (2007) 357(11):1121–35. doi: 10.1056/NEJMra071667 11. Fordyce CB, Gersh BJ, Stone GW, Granger CB. Novel therapeutics in myocardial infarction: targeting microvascular dysfunction and reperfusion injury. Trends Pharmacol Sci. (2015) 36:605–16. doi: 10.1016/j.tips.2015.06.004 11. Fordyce CB, Gersh BJ, Stone GW, Granger CB. Novel therapeutics in myocardial infarction: targeting microvascular dysfunction and reperfusion injury. Trends Pharmacol Sci. (2015) 36:605–16. doi: 10.1016/j.tips.2015.06.004 2. Thygesen K, Alpert JS, Jaffe AS, Chaitman BR, Bax JJ, Morrow DA, et al. Fourth universal deﬁnition of myocardial infarction (2018). Circulation. (2018) 138(20): e618–51. doi: 10.1161/CIR.0000000000000617 2. Thygesen K, Alpert JS, Jaffe AS, Chaitman BR, Bax JJ, Morrow DA, et al. Fourth universal deﬁnition of myocardial infarction (2018). Circulation. (2018) 138(20): e618–51. doi: 10.1161/CIR.0000000000000617 12. Saﬁri S, Karamzad N, Singh K, Carson-Chahhoud K, Adams C, Nejadghaderi SA, et al. Burden of ischemic heart disease and its attributable risk factors in 204 countries and territories, 1990-2019. Eur J Prev Cardiol. (2022) 29:420–31. doi: 10. 1093/eurjpc/zwab213 3. Roth GA, Mensah GA, Johnson CO, Addolorato G, Ammirati E, Baddour LM, et al. Global burden of cardiovascular diseases and risk factors, 1990-2019: update from the GBD 2019 study. J Am Coll Cardiol. (2020) 76(25):2982–3021. doi: 10. 1016/j.jacc.2020.11.010 13. Broadus RN. Toward a deﬁnition of “bibliometrics”. Scientometrics. (1987) 12(5- 6):373–9. doi: 10.1007/BF02016680 14. Ninkov A, Frank JR, Maggio LA. Bibliometrics: methods for studying academic publishing. Perspect Med Educ. (2022) 11(3):173–6. doi: 10.1007/s40037- 021-00695-4 4. Ginks WR, Sybers HD, Maroko PR, Covell JW, Sobel BE, Ross J. Coronary artery reperfusion. II. Reduction of myocardial infarct size at 1 week after the coronary occlusion. J Clin Invest. (1972) 51(10):2717–23. doi: 10.1172/JCI107091 15. Moral-Munoz JA, Herrera-Viedma E, Santisteban-Espejo A, Cobo MJ. Software tools for conducting bibliometric analysis in science: an up-to-date review. Prof Inf. (2020) 29(1):e290103. doi: 10.3145/epi.2020.ene.03 5. Jernberg T, Johanson P, Held C, Svennblad B, Lindbäck J, Wallentin L, et al. Association between adoption of evidence-based treatment and survival for patients with ST-elevation myocardial infarction. J Am Med Assoc. (2011) 305(16):1677–84. doi: 10.1001/jama.2011.522 16. van Eck NJ, Waltman L. Software survey: VOSviewer, a computer program for bibliometric mapping. Scientometrics. (2010) 84(2):523–38. doi: 10.1007/s11192-009- 0146-3 6. Hausenloy DJ, Botker HE, Engstrom T, Erlinge D, Heusch G, Ibanez B, et al. Author contributions YW and LG designed the study and co-wrote the manuscript. ZZ and PH collected the data. SF and AW re-examined the data. Conﬂict of interest Using information visualization techniques, we have attempted to elucidate the research progress, hot topics, and frontiers in MIRI over the past two decades. Although the annual publication output in China has far exceeded that in the United States in recent years, its academic inﬂuence is far behind. In addition, we have identiﬁed scholars, institutions, and representative literature that play important roles in this ﬁeld. Keyword analysis shows that the main research direction is the pathogenesis of MIRI and corresponding protective strategies, with ferroptosis and pyroptosis as the latest hot topics. The authors declare that the research was conducted in the absence of any commercial or ﬁnancial relationships that could be construed as a potential conﬂict of interest. Funding This study was supported by the National Key Research and Development Program of China (No. 2018YFC1707401). Publisher’s note All claims expressed in this article are solely those of the authors and do not necessarily represent those of their afﬁliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher. Limitations Firstly, this study’s bibliometric analysis only includes papers in the WoSCC database, while other databases such as PubMed, Cochrane library, and Google Scholar are excluded. However, it should be noted that WoSCC is widely recognized as one of the Frontiers in Cardiovascular Medicine 13 frontiersin.org Wang et al. 10.3389/fcvm.2023.1180792 YW and ZZ analyzed the data. XM and ML reviewed and revised the manuscript. All authors contributed to the article and approved the submitted version. YW and ZZ analyzed the data. XM and ML reviewed and revised the manuscript. All authors contributed to the article and approved the submitted version. most authoritative scientiﬁc literature search platforms, covering the vast majority of research on MIRI and still maintaining a certain degree of representativeness. Secondly, papers published in recent years are rarely cited, which may lead to the omission of some recently published papers with signiﬁcant contributions when analyzing highly co-cited papers, indicating the necessity of updating future research. In addition, changing job positions or using different names within the same institution during an author’s career can also have a signiﬁcant impact on research results. References Trends Cardiovasc Med. (2018) 28 (4):263–71. doi: 10.1016/j.tcm.2017.11.005 23. Lin Z, Zhang YH. Ischemic reperfusion injury in hypertensive rat heart. J Hypertens. (2022) 40(Suppl 1):e114. doi: 10.1097/01.hjh.0000836412.18071.b8 46. Heusch G. Myocardial stunning and hibernation revisited. Nat Rev Cardiol. (2021) 18(7):522–36. doi: 10.1038/s41569-021-00506-7 24. Lejay A, Fang F, John R, Van JAD, Barr M, Thaveau F, et al. Ischemia reperfusion injury, ischemic conditioning and diabetes mellitus. J Mol Cell Cardiol. (2016) 91:11–22. doi: 10.1016/j.yjmcc.2015.12.020 47. Stone GW, Selker HP, Thiele H, Patel MR, Udelson JE, Ohman EM, et al. Relationship between infarct size and outcomes following primary PCI: patient-level analysis from 10 randomized trials. J Am Coll Cardiol. (2016) 67(14):1674–83. doi: 10.1016/j.jacc.2016.01.069 25. Verma S, Pandey A, Bhatt DL. Forecasting heart failure risk in diabetes. J Am Coll Cardiol. (2022) 79(23):2294–7. doi: 10.1016/j.jacc.2022.04.011 48. Liu T, Howarth AG, Chen Y, Nair AR, Yang HJ, Ren D, et al. Intramyocardial hemorrhage and the “wave front” of reperfusion injury compromising myocardial salvage. J Am Coll Cardiol. (2022) 79(1):35–8. doi: 10.1016/j.jacc.2021.10.034 26. Adameova A, Horvath C, Abdul-Ghani S, Varga ZV, Saadeh Suleiman M, Dhalla NS. Interplay of oxidative stress and necrosis-like cell death in cardiac ischemia/ reperfusion injury: a focus on necroptosis. Biomedicines. (2022) 10(1):127. doi: 10. 3390/biomedicines10010127 49. Ibanez B, Fuster V, Jimenez-Borreguero J, Badimon JJ. Lethal myocardial reperfusion injury: a necessary evil? Int J Cardiol. (2011) 151(1):3–11. doi: 10.1016/ j.ijcard.2010.10.056 27. Gan L, Xie D, Liu J, Lau WB, Christopher TA, Lopez B, et al. Small extracellular microvesicles mediated pathological communications between dysfunctional adipocytes and cardiomyocytes as a novel mechanism exacerbating ischemia/ reperfusion injury in diabetic mice. Circulation. (2020) 141(12):968–83. doi: 10. 1161/CIRCULATIONAHA.119.042640 50. de Waha S, Patel MR, Granger CB, Ohman EM, Maehara A, Eitel I, et al. Relationship between microvascular obstruction and adverse events following primary percutaneous coronary intervention for ST-segment elevation myocardial infarction: an individual patient data pooled analysis from seven randomized trials. Eur Heart J. (2017) 38(47):3502–10. doi: 10.1093/eurheartj/ehx414 28. Paskeh MDA, Asadi A, Mirzaei S, Hashemi M, Entezari M, Raesi R, et al. Targeting AMPK signaling in ischemic/reperfusion injury: from molecular mechanism to pharmacological interventions. Cell Signal. (2022) 94:110323. doi: 10. 1016/j.cellsig.2022.110323 51. Del Buono MG, Moroni F, Montone RA, Azzalini L, Sanna T, Abbate A. Ischemic cardiomyopathy and heart failure after acute myocardial infarction. Curr Cardiol Rep. (2022) 24(10):1505–15. doi: 10.1007/s11886-022-01766-6 29. Kandula N, Kumar S, Mandlem VKK, Siddabathuni A, Singh S, Kosuru R. References Role of AMPK in myocardial ischemia-reperfusion injury-induced cell death in the presence and absence of diabetes. Oxid Med Cell Longev. (2022) 2022:7346699. doi: 10.1155/2022/7346699 52. Heidenreich PA, Bozkurt B, Aguilar D, Allen LA, Byun JJ, et al. 2022 AHA/ ACC/HFSA guideline for the management of heart failure: a report of the American college of cardiology/American heart association joint committee on clinical practice guidelines. J Am Coll Cardiol. (2022) 79(17):e263–421. doi: 10.1016/ j.jacc.2021.12.012 30. Yao M, Wang Z, Jiang L, Wang L, Yang Y, Wang Q, et al. Oxytocin ameliorates high glucose- and ischemia/reperfusion-induced myocardial injury by suppressing pyroptosis via AMPK signaling pathway. Biomed Pharmacother. (2022) 153:113498. doi: 10.1016/j.biopha.2022.113498 53. Wohlfahrt P, Jenca D, Stehlik J, Melenovský V, Mrázková J, Stanek V, et al. Heart failure-related quality-of-life impairment after myocardial infarction. Clin Res Cardiol. (2023) 112(1):39–48. doi: 10.1007/s00392-022-02008-z 31. Ibáñez B, Heusch G, Ovize M, Van de Werf F. Evolving therapies for myocardial ischemia/reperfusion injury. J Am Coll Cardiol. (2015) 65(14):1454–71. doi: 10.1016/j. jacc.2015.02.032 54. Del Re DP, Amgalan D, Linkermann A, Liu Q, Kitsis RN. Fundamental mechanisms of regulated cell death and implications for heart disease. Physiol Rev. (2019) 99(4):1765–817. doi: 10.1152/physrev.00022.2018 32. Kloner RA, Ganote CE, Jennings RB. The “no-reﬂow” phenomenon after temporary coronary occlusion in the dog. J Clin Invest. (1974) 54(6):1496–508. doi: 10.1172/JCI107898 55. Lu L-Q, Tian J, Luo X-J, Peng J. Targeting the pathways of regulated necrosis: a potential strategy for alleviation of cardio-cerebrovascular injury. Cell Mol Life Sci. (2021) 78(1):63–78. doi: 10.1007/s00018-020-03587-8 33. Rezkalla SH, Kloner RA. No-reﬂow phenomenon. Circulation. (2002) 105 (5):656–62. doi: 10.1161/hc0502.102867 56. Li N, Jiang W, Wang W, Xiong R, Wu X, Geng Q. Ferroptosis and its emerging roles in cardiovascular diseases. Pharmacol Res. (2021) 166:105466. doi: 10.1016/j. phrs.2021.105466 34. Allencherril J, Jneid H, Atar D, Alam M, Levine G, Kloner RA, et al. Pathophysiology, diagnosis, and management of the no-reﬂow phenomenon. Cardiovasc Drugs Ther. (2019) 33(5):589–97. doi: 10.1007/s10557-019-06901-0 57. Granger DN, Kvietys PR. Reperfusion injury and reactive oxygen species: the evolution of a concept. Redox Biol. (2015) 6:524–51. doi: 10.1016/j.redox.2015.08.020 35. Niccoli G, Burzotta F, Galiuto L, Crea F. Myocardial no-reﬂow in humans. J Am Coll Cardiol. (2009) 54(4):281–92. doi: 10.1016/j.jacc.2009.03.054 58. Park E, Chung SW. ROS-mediated autophagy increases intracellular iron levels and ferroptosis by ferritin and transferrin receptor regulation. Cell Death Dis. (2019) 10(11):822. doi: 10.1038/s41419-019-2064-5 36. References Targeting reperfusion injury in patients with ST-segment elevation myocardial infarction: trials and tribulations. Eur Heart J. (2017) 38(13):935–41. doi: 10.1093/ eurheartj/ehw145 17. Hausenloy DJ, Yellon DM. Myocardial ischemia-reperfusion injury: a neglected therapeutic target. J Clin Invest. (2013) 123(1):92–100. doi: 10.1172/JCI62874 7. Maron DJ, Hochman JS, Reynolds HR, Bangalore S, O’Brien SM, Boden WE, et al. Initial invasive or conservative strategy for stable coronary disease. N Engl J Med. (2020) 382(15):1395–407. doi: 10.1056/NEJMoa1915922 18. Murry CE, Jennings RB, Reimer KA. Preconditioning with ischemia: a delay of lethal cell injury in ischemic myocardium. Circulation. (1986) 74(5):1124–36. doi: 10. 1161/01.cir.74.5.1124 8. Braunwald E, Kloner RA. Myocardial reperfusion: a double-edged sword? J Clin Invest. (1985) 76(5):1713–9. doi: 10.1172/JCI112160 19. Zhao ZQ, Corvera JS, Halkos ME, Kerendi F, Wang NP, Guyton RA, et al. Inhibition of myocardial injury by ischemic postconditioning during reperfusion: comparison with ischemic preconditioning. Am J Physiol Heart Circ Physiol. (2003) 285(2):H579–88. doi: 10.1152/ajpheart.01064.2002 9. Heusch G. Myocardial ischaemia-reperfusion injury and cardioprotection in perspective. Nat Rev Cardiol. (2020) 17(12):773–89. doi: 10.1038/s41569-020-0403-y 20. Ferdinandy P, Schulz R, Baxter GF. Interaction of cardiovascular risk factors with myocardial ischemia/reperfusion injury, preconditioning, and postconditioning. Pharmacol Rev. (2007) 59(4):418–58. doi: 10.1124/pr.107.06002 10. Han X, Zhang J, Liu J, Wang H, Du F, Zeng X, et al. Targeting ferroptosis: a novel insight against myocardial infarction and ischemia-reperfusion injuries. Apoptosis. (2022) 6:108–23. doi: 10.1007/s10495-022-01785-2 20. Ferdinandy P, Schulz R, Baxter GF. Interaction of cardiovascular risk factors with myocardial ischemia/reperfusion injury, preconditioning, and postconditioning. Pharmacol Rev. (2007) 59(4):418–58. doi: 10.1124/pr.107.06002 Frontiers in Cardiovascular Medicine 14 14 frontiersin.org Wang et al. 10.3389/fcvm.2023.1180792 Wang et al. responsible for increasing ischemia- and reperfusion-induced ventricular arrhythmia in older female rats. Life Sci. (2021) 271:119190. doi: 10.1016/j.lfs.2021.119190 21. Penna C, Andreadou I, Aragno M, Beauloye C, Bertrand L, Lazou A, et al. Effect of hyperglycaemia and diabetes on acute myocardial ischaemia-reperfusion injury and cardioprotection by ischaemic conditioning protocols. Br J Pharmacol. (2020) 177 (23):5312–35. doi: 10.1111/bph.14993 44. Verma S, Fedak PWM, Weisel RD, Butany J, Rao V, Maitland A, et al. Fundamentals of reperfusion injury for the clinical cardiologist. Circulation. (2002) 105(20):2332–6. doi: 10.1161/01.cir.0000016602.96363.36 22. Andreadou I, Schulz R, Badimon L, Adameová A, Kleinbongard P, Lecour S, et al. Hyperlipidaemia and cardioprotection: animal models for translational studies. Br J Pharmacol. (2020) 177(23):5287–311. doi: 10.1111/bph.14931 45. Guaricci AI, Bulzis G, Pontone G, Scicchitano P, Carbonara R, Rabbat M, et al. Current interpretation of myocardial stunning. References Zhang B, Liu G, Huang B, Liu H, Jiang H, Hu Z, et al. KDM3A attenuates myocardial ischemic and reperfusion injury by ameliorating cardiac microvascular endothelial cell pyroptosis. Oxid Med Cell Longev. (2022) 2022:4622520. doi: 10. 1155/2022/4622520 59. Li X, Ma N, Xu J, Zhang Y, Yang P, Su X, et al. Targeting ferroptosis: pathological mechanism and treatment of ischemia-reperfusion injury. Oxid Med Cell Longev. (2021) 2021:1587922. doi: 10.1155/2021/1587922 37. Hausenloy DJ, Bøtker HE, Condorelli G, Ferdinandy P, Garcia-Dorado D, Heusch G, et al. Translating cardioprotection for patient beneﬁt: position paper from the working group of cellular biology of the heart of the European society of cardiology. Cardiovasc Res. (2013) 98(1):7–27. doi: 10.1093/cvr/cvt004 60. Miyamoto HD, Ikeda M, Ide T, Tadokoro T, Furusawa S, Abe K, et al. Iron overload via heme degradation in the endoplasmic Reticulum triggers ferroptosis in myocardial ischemia-reperfusion injury. JACC Basic Transl Sci. (2022) 7(8):800–19. doi: 10.1016/j.jacbts.2022.03.012 38. Neri M, Riezzo I, Pascale N, Pomara C, Turillazzi E. Ischemia/reperfusion injury following acute myocardial infarction: a critical issue for clinicians and forensic pathologists. Mediators Inﬂamm. (2017) 2017:7018393. doi: 10.1155/2017/ 7018393 61. Li W, Feng G, Gauthier JM, Lokshina I, Higashikubo R, Evans S, et al. Ferroptotic cell death and TLR4/trif signaling initiate neutrophil recruitment after heart transplantation. J Clin Invest. (2019) 129(6):2293–304. doi: 10.1172/JCI126428 39. Akar FG, Aon MA, Tomaselli GF, O’Rourke B. The mitochondrial origin of postischemic arrhythmias. J Clin Invest. (2005) 115(12):3527–35. doi: 10.1172/ JCI25371 62. Tang L-J, Luo X-J, Tu H, Chen H, Xiong XM, Li NS, et al. Ferroptosis occurs in phase of reperfusion but not ischemia in rat heart following ischemia or ischemia/ reperfusion. Naunyn Schmiedebergs Arch Pharmacol. (2021) 394(2):401–10. doi: 10. 1007/s00210-020-01932-z 40. Consolini AE, Ragone MI, Bonazzola P, Colareda GA. Mitochondrial bioenergetics during ischemia and reperfusion. Adv Exp Med Biol. (2017) 982:141–67. doi: 10.1007/978-3-319-55330-6_8 63. Zhou L, Han S, Guo J, Qiu T, Zhou J, Shen L. Ferroptosis-A new dawn in the treatment of organ ischemia-reperfusion injury. Cells. (2022) 11(22):3653. doi: 10. 3390/cells11223653 41. Solhjoo S, O’Rourke B. Mitochondrial instability during regional ischemia- reperfusion underlies arrhythmias in monolayers of cardiomyocytes. J Mol Cell Cardiol. (2015) 78:90–9. doi: 10.1016/j.yjmcc.2014.09.024 64. Gao M, Monian P, Quadri N, Ramasamy R, Jiang X. Glutaminolysis and transferrin regulate ferroptosis. Mol Cell. (2015) 59(2):298–308. doi: 10.1016/j. molcel.2015.06.011 42. Bozdogan Ö, Kaya ST, Yasar S, Orallar H. References Bcl-2 antiapoptotic proteins inhibit beclin 1-dependent autophagy. Cell. (2005) 122 (6):927–39. doi: 10.1016/j.cell.2005.07.002 78. Chen H, Tang LJ, Tu H, Zhou YJ, Li NS, Luo XJ, et al. Arctiin protects rat heart against ischemia/reperfusion injury via a mechanism involving reduction of necroptosis. Eur J Pharmacol. (2020) 875:173053. doi: 10.1016/j.ejphar.2020.173053 101. Ma X, Liu H, Foyil SR, Godar RJ, Weinheimer CJ, Hill JA, et al. Impaired autophagosome clearance contributes to cardiomyocyte death in ischemia/ reperfusion injury. Circulation. (2012) 125(25):3170–81. doi: 10.1161/ CIRCULATIONAHA.111.041814 79. Li L, Lin L, Lei S, Shi S, Chen C, Xia Z. Maslinic acid inhibits myocardial ischemia-reperfusion injury-induced apoptosis and necroptosis via promoting autophagic ﬂux. DNA Cell Biol. (2022) 41(5):487–97. doi: 10.1089/dna.2021. 0918 102. Marinescu M-C, Lazar A-L, Marta MM, Cozma A, Catana C-S. Non-coding RNAs: prevention, diagnosis, and treatment in myocardial ischemia-reperfusion injury. Int J Mol Sci. (2022) 23(5):2728. doi: 10.3390/ijms23052728 80. Luedde M, Lutz M, Carter N, Sosna J, Jacoby C, Vucur M, et al. RIP3, A kinase promoting necroptotic cell death, mediates adverse remodelling after myocardial infarction. Cardiovasc Res. (2014) 103(2):206–16. doi: 10.1093/cvr/cvu146 103. Poller W, Dimmeler S, Heymans S, Zeller T, Haas J, Karakas M, et al. Non- coding RNAs in cardiovascular diseases: diagnostic and therapeutic perspectives. Eur Heart J. (2018) 39(29):2704–16. doi: 10.1093/eurheartj/ehx165 81. Zhang T, Zhang Y, Cui M, Jin L, Wang Y, Lv F, et al. CaMKII is a RIP3 substrate mediating ischemia- and oxidative stress-induced myocardial necroptosis. Nat Med. (2016) 22(2):175–82. doi: 10.1038/nm.4017 104. Kabekkodu SP, Shukla V, Varghese VK, D’Souza J, Chakrabarty S, Satyamoorthy K. Clustered miRNAs and their role in biological functions and diseases. Biol Rev Camb Philos Soc. (2018) 93(4):1955–86. doi: 10.1111/brv.12428 82. Zhou H, Li D, Zhu P, Ma Q, Toan S, Wang J, et al. Inhibitory effect of melatonin on necroptosis via repressing the Ripk3-PGAM5-CypD-mPTP pathway attenuates cardiac microvascular ischemia-reperfusion injury. J Pineal Res. (2018) 65(3): e12503. doi: 10.1111/jpi.12503 105. Bartel DP. MicroRNAs: genomics, biogenesis, mechanism, and function. Cell. (2004) 116(2):281–97. doi: 10.1016/s0092-8674(04)00045-5 106. Li T, Gu J, Yang O, Wang J, Wang Y, Kong J. Bone marrow mesenchymal stem cell-derived exosomal miRNA-29c decreases cardiac ischemia/reperfusion injury through inhibition of excessive autophagy via the PTEN/akt/mTOR signaling pathway. Circ J. (2020) 84(8):1304–11. doi: 10.1253/circj.CJ-19-1060 83. Zhu P, Hu S, Jin Q, Li D, Tian F, Toan S, et al. Ripk3 promotes ER stress- induced necroptosis in cardiac IR injury: a mechanism involving calcium overload/ XO/ROS/mPTP pathway. References The gasdermins, a protein family executing cell death and inﬂammation. Nat Rev Immunol. (2020) 20(3):143–57. doi: 10.1038/s41577-019- 0228-2 94. Popov SV, Mukhomedzyanov AV, Voronkov NS, Derkachev IA, Boshchenko AA, Fu F, et al. Regulation of autophagy of the heart in ischemia and reperfusion. Apoptosis. (2023) 28(1-2):55–80. doi: 10.1007/s10495-022-01786-1 72. Chen X, He WT, Hu L, Li J, Fang Y, Wang X, et al. Pyroptosis is driven by non- selective gasdermin-D pore and its morphology is different from MLKL channel- mediated necroptosis. Cell Res. (2016) 26(9):1007–20. doi: 10.1038/cr.2016.100 95. Gottlieb RA, Carreira RS. Autophagy in health and disease. 5. Mitophagy as a way of life. Am J Physiol Cell Physiol. (2010) 299(2):C203–210. doi: 10.1152/ajpcell. 00097.2010 73. Galluzzi L, Vitale I, Aaronson SA, Abrams JM, Adam D, Agostinis P, et al. Molecular mechanisms of cell death: recommendations of the nomenclature committee on cell death 2018. Cell Death Differ. (2018) 25(3):486–541. doi: 10. 1038/s41418-017-0012-4 96. Zhao J, Zhang J, Liu Q, Wang Y, Jin Y, Yang Y, et al. Hongjingtian injection protects against myocardial ischemia reperfusion-induced apoptosis by blocking ROS induced autophagic- ﬂux. Biomed Pharmacother. (2021) 135:111205. doi: 10. 1016/j.biopha.2020.111205 74. Zhang X, Qu H, Yang T, Kong X, Zhou H. Regulation and functions of NLRP3 inﬂammasome in cardiac ﬁbrosis: current knowledge and clinical signiﬁcance. Biomed Pharmacother. (2021) 143:112219. doi: 10.1016/j.biopha.2021.112219 97. Cadenas S. ROS and redox signaling in myocardial ischemia-reperfusion injury and cardioprotection. Free Radic Biol Med. (2018) 117:76–89. doi: 10.1016/j. freeradbiomed.2018.01.024 75. Shi H, Gao Y, Dong Z, Yang J, Gao R, Li X, et al. GSDMD-mediated cardiomyocyte pyroptosis promotes myocardial I/R injury. Circ Res. (2021) 129 (3):383–96. doi: 10.1161/CIRCRESAHA.120.318629 98. Hariharan N, Zhai P, Sadoshima J. Oxidative stress stimulates autophagic ﬂux during ischemia/reperfusion. Antioxid Redox Signal. (2011) 14(11):2179–90. doi: 10. 1089/ars.2010.3488 76. Degterev A, Huang Z, Boyce M, Li Y, Jagtap P, Mizushima N, et al. Chemical inhibitor of nonapoptotic cell death with therapeutic potential for ischemic brain injury. Nat Chem Biol. (2005) 1(2):112–9. doi: 10.1038/nchembio711 99. Mokhtari B, Badalzadeh R. Protective and deleterious effects of autophagy in the setting of myocardial ischemia/reperfusion injury: an overview. Mol Biol Rep. (2022) 49(11):11081–99. doi: 10.1007/s11033-022-07837-9 77. Vandenabeele P, Galluzzi L, Berghe TV, Kroemer G. Molecular mechanisms of necroptosis: an ordered cellular explosion. Nat Rev Mol Cell Biol. (2010) 11 (10):700–14. doi: 10.1038/nrm2970 100. Pattingre S, Tassa A, Qu X, Garuti R, Liang XH, Mizushima N, et al. References Effect of ATP-dependent channel modulators on ischemia-induced arrhythmia change depending on age and gender. Exp Biol Med. (2013) 238(10):1170–79. doi: 10.1177/ 1535370213498980 65. Wu X, Li Y, Zhang S, Zhou X. Ferroptosis as a novel therapeutic target for cardiovascular disease. Theranostics. (2021) 11(7):3052–9. doi: 10.7150/thno.54113 66. Fang X, Ardehali H, Min J, Wang F. The molecular and metabolic landscape of iron and ferroptosis in cardiovascular disease. Nat Rev Cardiol. (2023) 20(1):7–23. doi: 10.1038/s41569-022-00735-4 43. Bozdogan O, Bozcaarmutlu A, Kaya ST, Sapmaz C, Ozarslan TO, Eksioglu D, et al. Decreasing myocardial estrogen receptors and antioxidant activity may be Frontiers in Cardiovascular Medicine 15 frontiersin.org Wang et al. Wang et al. 10.3389/fcvm.2023.1180792 10.3389/fcvm.2023.1180792 67. Fang X, Wang H, Han D, Xie E, Yang X, Wei J, et al. Ferroptosis as a target for protection against cardiomyopathy. Proc Natl Acad Sci U S A. (2019) 116(7):2672–80. doi: 10.1073/pnas.1821022116 90. Mehto S, Jena KK, Nath P, Chauhan S, Kolapalli SP, Das SK, et al. The crohn’s disease risk factor IRGM limits NLRP3 inﬂammasome activation by impeding its assembly and by mediating its selective autophagy. Mol Cell. (2019) 73 (3):429–445.e7. doi: 10.1016/j.molcel.2018.11.018 68. Zhang Y, Xin L, Xiang M, Shang C, Wang Y, Wang Y, et al. The molecular mechanisms of ferroptosis and its role in cardiovascular disease. Biomed Pharmacother. (2022) 145:112423. doi: 10.1016/j.biopha.2021.112423 91. Ma S, Wang Y, Chen Y, Cao F. The role of the autophagy in myocardial ischemia/reperfusion injury. Biochim Biophys Acta. (2015) 1852(2):271–6. doi: 10. 1016/j.bbadis.2014.05.010 69. Feng Y, Madungwe NB, Imam Aliagan AD, Tombo N, Bopassa JC. Liproxstatin- 1 protects the mouse myocardium against ischemia/reperfusion injury by decreasing VDAC1 levels and restoring GPX4 levels. Biochem Biophys Res Commun. (2019) 520(3):606–11. doi: 10.1016/j.bbrc.2019.10.006 92. Yu Y-W, Que J-Q, Liu S, Huang KY, Qian L, Weng YB, et al. Sodium-glucose co- transporter-2 inhibitor of dapagliﬂozin attenuates myocardial ischemia/reperfusion injury by limiting NLRP3 inﬂammasome activation and modulating autophagy. Front Cardiovasc Med. (2022) 8:768214. doi: 10.3389/fcvm.2021.768214 70. Lu F, Lan Z, Xin Z, He C, Guo Z, Xia X, et al. Emerging insights into molecular mechanisms underlying pyroptosis and functions of inﬂammasomes in diseases. J Cell Physiol. (2020) 235(4):3207–21. doi: 10.1002/jcp.29268 93. Xiang M, Lu Y, Xin L, Gao J, Shang C, Jiang Z, et al. Role of oxidative stress in reperfusion following myocardial ischemia and its treatments. Oxid Med Cell Longev. (2021) 2021:6614009. doi: 10.1155/2021/6614009 71. Broz P, Pelegrín P, Shao F. References Redox Biol. (2018) 16:157–68. doi: 10.1016/j.redox.2018. 02.019 107. Wu Q, Shang Y, Bai Y, Wu Y, Wang H, Shen T. Sufentanil preconditioning protects against myocardial ischemia/reperfusion injury via miR-125a/DRAM2 axis. Cell Cycle. (2021) 20(4):383–91. doi: 10.1080/15384101.2021.1875668 84. Ying L, Benjanuwattra J, Chattipakorn SC, Chattipakorn N. The role of RIPK3- regulated cell death pathways and necroptosis in the pathogenesis of cardiac ischaemia-reperfusion injury. Acta Physiol. (2021) 231(2):e13541. doi: 10.1111/apha. 13541 108. Sun W, Shi R, Guo J, Wang H, Shen L, Shi H, et al. miR-135b-3p promotes cardiomyocyte ferroptosis by targeting GPX4 and aggravates myocardial ischemia/ reperfusion injury. Front Cardiovasc Med. (2021) 8:663832. doi: 10.3389/fcvm.2021. 663832 85. Levine B, Yuan J. Autophagy in cell death: an innocent convict? J Clin Invest. (2005) 115(10):2679–88. doi: 10.1172/JCI26390 109. Zhu WS, Guo W, Zhu JN, Tang CM, Fu YH, Lin QX, et al. Hsp90aa1: a novel target gene of miR-1 in cardiac ischemia/reperfusion injury. Sci Rep. (2016) 6 (1):24498. doi: 10.1038/srep24498 86. Egan D, Kim J, Shaw RJ, Guan K-L. The autophagy initiating kinase ULK1 is regulated via opposing phosphorylation by AMPK and mTOR. Autophagy. (2011) 7 (6):643–4. doi: 10.4161/auto.7.6.15123 87. Matsui Y, Takagi H, Qu X, Abdellatif M, Sakoda H, Asano T, et al. Distinct roles of autophagy in the heart during ischemia and reperfusion: roles of AMP-activated protein kinase and beclin 1 in mediating autophagy. Circ Res. (2007) 100(6):914–22. doi: 10.1161/01.RES.0000261924.76669.36 110. Wang W, Zheng Y, Wang M, Yan M, Jiang J, Li Z. Exosomes derived miR-126 attenuates oxidative stress and apoptosis from ischemia and reperfusion injury by targeting ERRFI1. Gene. (2019) 690:75–80. doi: 10.1016/j.gene.2018.12.044 111. Yang D, Wang M, Hu Z, Ma Y, Shi Y, Cao X, et al. Extracorporeal cardiac shock wave-induced exosome derived from endothelial colony-forming cells carrying miR-140-3p alleviate cardiomyocyte hypoxia/reoxygenation injury via the PTEN/PI3K/AKT pathway. Front Cell Dev Biol. (2021) 9:779936. doi: 10.3389/fcell. 2021.779936 88. Mokhtari B, Badalzadeh R. The potentials of distinct functions of autophagy to be targeted for attenuation of myocardial ischemia/reperfusion injury in preclinical studies: an up-to-date review. J Physiol Biochem. (2021) 77(3):377–404. doi: 10. 1007/s13105-021-00824-x 112. Lu Y, Xi J, Zhang Y, Li C, Chen W, Hu X, et al. MicroRNA-214-5p protects against myocardial ischemia reperfusion injury through targeting the FAS ligand. Arch Med Sci. (2020) 16(5):1119–29. doi: 10.5114/aoms.2019.85405 112. Lu Y, Xi J, Zhang Y, Li C, Chen W, Hu X, et al. Frontiers in Cardiovascular Medicine References miR-15b-5p regulates the NLRP3 inﬂammasome signal through targeting SIRT3 to regulate hypoxia/reoxygenation-induced cardiomyocyte pyroptosis process. Shock. (2022) 58(2):147–57. doi: 10.1097/SHK.0000000000001961 139. Ma L-L, Kong F-J, Dong Z, Xin K-Y, Wang X-X, Sun A-J, et al. Hypertrophic preconditioning attenuates myocardial ischaemia-reperfusion injury by modulating SIRT3-SOD2-mROS-dependent autophagy. Cell Prolif. (2021) 54(7):e13051. doi: 10. 1111/cpr.13051 118. Li K, Zhou P, Li S, Zheng S, Wang D. MicroRNA-29b reduces myocardial ischemia-reperfusion injury in rats via down-regulating PTEN and activating the akt/eNOS signaling pathway. J Thromb Thrombolysis. (2022) 53(1):123–35. doi: 10. 1007/s11239-021-02535-y 140. Stiermaier T, Jensen J-O, Rommel K-P, de Waha-Thiele S, Fuernau G, Desch S, et al. Combined intrahospital remote ischemic perconditioning and postconditioning improves clinical outcome in ST-elevation myocardial infarction. Circ Res. (2019) 124 (10):1482–91. doi: 10.1161/CIRCRESAHA.118.314500 119. Gao L, Qiu F, Cao H, Li H, Dai G, Ma T, et al. Therapeutic delivery of microRNA-125a-5p oligonucleotides improves recovery from myocardial ischemia/ reperfusion injury in mice and swine. Theranostics. (2023) 13(2):685–703. doi: 10. 7150/thno.73568 120. Zhao D, Shun E, Ling F, Liu Q, Warsi A, Wang B, et al. Plk2 regulated by miR- 128 induces ischemia-reperfusion injury in cardiac cells. Mol Ther Nucleic Acids. (2020) 19:458–67. doi: 10.1016/j.omtn.2019.11.029 141. Zhao Z, Shi Q, Guo Q, Peng L, Li X, Rao L, et al. Remote ischemic preconditioning can extend the tolerance to extended drug-coated balloon inﬂation time by reducing myocardial damage during percutaneous coronary intervention. Int J Cardiol. (2022) 353:3–8. doi: 10.1016/j.ijcard.2022.01.049 121. Chen B, Zheng L, Zhu T, Jiao K. LncRNA FOXD3-AS1 aggravates myocardial ischemia/reperfusion injury by inactivating the Redd1/AKT/GSK3β/Nrf2 signaling pathway via the miR-128/TXNIP axis. J Biochem Mol Toxicol. (2022) 36(12):e23218. doi: 10.1002/jbt.23218 142. Basili S, Tanzilli G, Raparelli V, Calvieri C, Pignatelli P, Carnevale R, et al. Aspirin reload before elective percutaneous coronary intervention: impact on serum thromboxane b2 and myocardial reperfusion indexes. Circ Cardiovasc Interv. (2014) 7(4):577–84. doi: 10.1161/CIRCINTERVENTIONS.113.001197 122. Wan X, Yao B, Ma Y, Liu Y, Tang Y, Hu J, et al. MicroRNA-128-1-5p attenuates myocardial ischemia/reperfusion injury by suppressing Gadd45g- mediated apoptotic signaling. Biochem Biophys Res Commun. (2020) 530(1):314–21. doi: 10.1016/j.bbrc.2020.07.009 143. Ye Y, Birnbaum GD, Perez-Polo JR, Nanhwan MK, Nylander S, Birnbaum Y. Ticagrelor protects the heart against reperfusion injury and improves remodeling after myocardial infarction. Arterioscler Thromb Vasc Biol. (2015) 35(8):1805–14. doi: 10. 1161/ATVBAHA.115.305655 123. Dai Z-H, Jiang Z-M, Tu H, Mao L, Song G-L, Yang Z-B, et al. References MicroRNA-214-5p protects against myocardial ischemia reperfusion injury through targeting the FAS ligand. Arch Med Sci. (2020) 16(5):1119–29. doi: 10.5114/aoms.2019.85405 89. Mishra PK, Adameova A, Hill JA, Baines CP, Kang PM, Downey JM, et al. Guidelines for evaluating myocardial cell death. Am J Physiol Heart Circ Physiol. (2019) 317(5):H891–922. doi: 10.1152/ajpheart.00259.2019 Frontiers in Cardiovascular Medicine 16 frontiersin.org Wang et al. 10.3389/fcvm.2023.1180792 10.3389/fcvm.2023.1180792 113. Chen Q, Liu Y, Ding X, Li Q, Qiu F, Wang M, et al. Bone marrow mesenchymal stem cell-secreted exosomes carrying microRNA-125b protect against myocardial ischemia reperfusion injury via targeting SIRT7. Mol Cell Biochem. (2020) 465(1- 2):103–14. doi: 10.1007/s11010-019-03671-z 134. Yuan J, Fu X. MicroRNA-21 mediates the protective role of emulsiﬁed isoﬂurane against myocardial ischemia/reperfusion injury in mice by targeting SPP1. Cell Signal. (2021) 86:110086. doi: 10.1016/j.cellsig.2021.110086 135. Han J, Zhang Z, Zhang Z, Yang S. Artemisinin relieves myocardial ischemia- reperfusion injury via modulating miR-29b-3p and hemicentin 1. Front Pharmacol. (2022) 13:918966. doi: 10.3389/fphar.2022.918966 114. Xiao X, Lu Z, Lin V, May A, Shaw DH, Wang Z, et al. MicroRNA miR-24-3p reduces apoptosis and regulates Keap1-Nrf2 pathway in mouse cardiomyocytes responding to ischemia/reperfusion injury. Oxid Med Cell Longev. (2018) 2018:7042105. doi: 10.1155/2018/7042105 136. Wang K, An T, Zhou L-Y, Liu C-Y, Zhang X-J, Feng C, et al. E2F1-regulated miR-30b suppresses cyclophilin D and protects heart from ischemia/reperfusion injury and necrotic cell death. Cell Death Differ. (2015) 22(5):743–54. doi: 10.1038/ cdd.2014.165 115. Liu X-M, Zhang Z, Zhong J, Li N, Wang T, Wang L, et al. Long non-coding RNA MALAT1 modulates myocardial ischemia-reperfusion injury through the PI3K/akt/eNOS pathway by sponging miRNA-133a-3p to target IGF1R expression. Eur J Pharmacol. (2022) 916:174719. doi: 10.1016/j.ejphar.2021.174719 137. Bei Y, Lu D, Bär C, Chatterjee S, Costa A, Riedel I, et al. miR-486 attenuates cardiac ischemia/reperfusion injury and mediates the beneﬁcial effect of exercise for myocardial protection. Mol Ther. (2022) 30(4):1675–91. doi: 10.1016/j.ymthe.2022. 01.031 116. Liu N, Xie L, Xiao P, Chen X, Kong W, Lou Q, et al. Cardiac ﬁbroblasts secrete exosome microRNA to suppress cardiomyocyte pyroptosis in myocardial ischemia/ reperfusion injury. Mol Cell Biochem. (2022) 477(4):1249–60. doi: 10.1007/s11010- 021-04343-7 138. Bøtker HE, Kharbanda R, Schmidt MR, Bøttcher M, Kaltoft AK, Terkelsen CJ, et al. Remote ischaemic conditioning before hospital admission, as a complement to angioplasty, and effect on myocardial salvage in patients with acute myocardial infarction: a randomised trial. Lancet. (2010) 375(9716):727–34. doi: 10.1016/S0140- 6736(09)62001-8 117. Xu J, Chen X, Nie W. References miR-129 attenuates myocardial ischemia reperfusion injury by regulating the expression of PTEN in rats. Biomed Res Int. (2021) 2021:5535788. doi: 10.1155/2021/5535788 144. Sayour AA, Celeng C, Oláh A, Ruppert M, Merkely B, Radovits T. Sodium- glucose cotransporter 2 inhibitors reduce myocardial infarct size in preclinical animal models of myocardial ischaemia-reperfusion injury: a meta-analysis. Diabetologia. (2021) 64(4):737–48. doi: 10.1007/s00125-020-05359-2 124. Zheng S, Wang L, Ma H, Sun F, Wen F. microRNA-129 overexpression in endothelial cell-derived extracellular vesicle inﬂuences inﬂammatory response caused by myocardial ischemia/reperfusion injury. Cell Biol Int. (2021) 45 (8):1743–56. doi: 10.1002/cbin.11614 145. Lim VG, Bell RM, Arjun S, Kolatsi-Joannou M, Long DA, Yellon DM. SGLT2 inhibitor, canagliﬂozin, attenuates myocardial infarction in the diabetic and nondiabetic heart. JACC Basic Transl Sci. (2019) 4(1):15–26. doi: 10.1016/j.jacbts. 2018.10.002 125. Liu S, Liao Q, Xu W, Zhang Z, Yin M, Cao X. MiR-129-5p protects H9c2 cardiac myoblasts from hypoxia/reoxygenation injury by targeting TRPM7 and inhibiting NLRP3 inﬂammasome activation. J Cardiovasc Pharmacol. (2021) 77 (5):586–93. doi: 10.1097/FJC.0000000000000991 146. Nikolaou PE, Mylonas N, Makridakis M, Makrecka-Kuka M, Iliou A, Zerikiotis S, et al. Cardioprotection by selective SGLT-2 inhibitors in a non-diabetic mouse model of myocardial ischemia/reperfusion injury: a class or a drug effect? Basic Res Cardiol. (2022) 117(1):27. doi: 10.1007/s00395-022-00934-7 126. Chen Z-X, He D, Mo Q-W, Xie L-P, Liang J-R, Liu L, et al. MiR-129-5p protects against myocardial ischemia-reperfusion injury via targeting HMGB1. Eur Rev Med Pharmacol Sci. (2020) 24(8):4440–50. doi: 10.26355/eurrev_202004_21026 147. Andreadou I, Bell RM, Bøtker HE, Zuurbier CJ. SGLT2 Inhibitors reduce infarct size in reperfused ischemic heart and improve cardiac function during ischemic episodes in preclinical models. Biochim Biophys Acta Mol Basis Dis. (2020) 1866(7):165770. doi: 10.1016/j.bbadis.2020.165770 127. Huang C, Qu Y, Feng F, Zhang H, Shu L, Zhu X, et al. Cardioprotective effect of circ_SMG6 knockdown against myocardial ischemia/reperfusion injury correlates with miR-138-5p-mediated EGR1/TLR4/TRIF inactivation. Oxid Med Cell Longev. (2022) 2022:1927260. doi: 10.1155/2022/1927260 148. Zhang F, Cao X, Zhao C, Chen L, Chen X. Empagliﬂozin activates JAK2/STAT3 signaling and protects cardiomyocytes from hypoxia/reoxygenation injury under high glucose conditions. J Thromb Thrombolysis. (2022) 17:116–25. doi: 10.1007/s11239- 022-02719-0 128. Chang Y, Xing L, Zhou W, Zhang W. Up-regulating microRNA-138-5p enhances the protective role of dexmedetomidine on myocardial ischemia- reperfusion injury mice via down-regulating Ltb4r1. Cell Cycle. (2021) 20(4):445–58. doi: 10.1080/15384101.2021.1878330 149. Pearce L, Carr RD, Yellon DM, Davidson SM. Nicorandil - an effective multitarget drug for cardioprotection? Cardiovasc Drugs Ther. (2023) 37(1):5–8. References Am J Chin Med. (2023) 51(1):1–17. doi: 10.1142/S0192415X23500015 164. Huang C, Zhou S, Chen C, Wang X, Ding R, Xu Y, et al. Biodegradable redox- responsive AIEgen-based-covalent organic framework nanocarriers for long-term treatment of myocardial ischemia/reperfusion injury. Small. (2022) 18(47):e2205062. doi: 10.1002/smll.202205062 155. Li T, Tan Y, Ouyang S, He J, Liu L. Resveratrol protects against myocardial ischemia-reperfusion injury via attenuating ferroptosis. Gene. (2022) 808:145968. doi: 10.1016/j.gene.2021.145968 165. Wei Y, Zhu M, Li S, Hong T, Guo X, Li Y, et al. Engineered biomimetic nanoplatform protects the myocardium against ischemia/reperfusion injury by inhibiting pyroptosis. ACS Appl Mater Interfaces. (2021) 13(29):33756–66. doi: 10. 1021/acsami.1c03421 156. Yang Y, Shao M, Yao J, Yang S, Cheng W, Ma L, et al. Neocryptotanshinone protects against myocardial ischemia-reperfusion injury by promoting autolysosome degradation of protein aggregates via the ERK1/2-Nrf2-LAMP2 pathway. Phytomedicine. (2023) 110:154625. doi: 10.1016/j.phymed.2022.154625 166. Sharma AK, Kumar A, Sahu M, Sharma G, Datusalia AK, Rajput SK. Exercise preconditioning and low dose copper nanoparticles exhibits cardioprotection through targeting GSK-3β phosphorylation in ischemia/reperfusion induced myocardial infarction. Microvasc Res. (2018) 120:59–66. doi: 10.1016/j.mvr.2018.06.003 157. Gao J, Wang Z, Ye Z. Madder (Rubia cordifolia L.) alleviates myocardial ischemia-reperfusion injury by protecting endothelial cells from apoptosis and inﬂammation. Mediators Inﬂamm. (2023) 2023:5015039. doi: 10.1155/2023/5015039 167. Gao L, Qiu F, Cao H, Li H, Dai G, Ma T, et al. Therapeutic delivery of microRNA-125a-5p oligonucleotides improves recovery from myocardial ischemia/ reperfusion injury in mice and swine. Theranostics. (2023) 13(2):685–703. doi: 10. 7150/thno.73568 158. Wang R, Wang M, Zhou J, Dai Z, Sun G, Sun X. Calenduloside E suppresses calcium overload by promoting the interaction between L-type calcium channels and Bcl2-associated athanogene 3 to alleviate myocardial ischemia/reperfusion injury. J Adv Res. (2021) 34:173–86. doi: 10.1016/j.jare.2020.10.005 168. Majka M, Kleibert M, Wojciechowska M. Impact of the main cardiovascular risk factors on plasma extracellular vesicles and their inﬂuence on the heart’s vulnerability to ischemia-reperfusion injury. Cells. (2021) 10(12):3331. doi: 10.3390/ cells10123331 159. Lv D, Luo M, Cheng Z, Wang R, Yang X, Guo Y, et al. Tubeimoside I ameliorates myocardial ischemia-reperfusion injury through SIRT3-dependent regulation of oxidative stress and apoptosis. Oxid Med Cell Longev. (2021) 2021:5577019. doi: 10.1155/2021/5577019 169. Yu S, Tang X, Zheng T, Li S, Ren H, Wu H, et al. Plasma-derived extracellular vesicles transfer microRNA-130a-3p to alleviate myocardial ischemia/reperfusion injury by targeting ATG16L1. Cell Tissue Res. (2022) 389(1):99–114. doi: 10.1007/ s00441-022-03605-0 160. References doi: 10.1007/s10557-022-07397-x 129. Yang S, Li H, Chen L. MicroRNA-140 attenuates myocardial ischemia- reperfusion injury through suppressing mitochondria-mediated apoptosis by targeting YES1. J Cell Biochem. (2019) 120(3):3813–21. doi: 10.1002/jcb.27663 150. Peng Y-W, Major T, Deatrick KB, Mohammed A, Jeakle M, Charpie JR. Nicorandil attenuates ventricular dysfunction and organ injury after cardiopulmonary bypass. Int J Cardiol. (2022) 368:62–8. doi: 10.1016/j.ijcard.2022. 08.030 130. Ge X, Meng Q, Wei L, Liu J, Li M, Liang X, et al. Myocardial ischemia- reperfusion induced cardiac extracellular vesicles harbour proinﬂammatory features and aggravate heart injury. J Extracell Vesicles. (2021) 10(4):e12072. doi: 10.1002/ jev2.12072 151. Qian G, Zhang Y, Dong W, Jiang ZC, Li T, Cheng LQ, et al. Effects of nicorandil administration on infarct size in patients with ST-segment-elevation myocardial infarction undergoing primary percutaneous coronary intervention: the CHANGE trial. J Am Heart Assoc. (2022) 11(18):e026232. doi: 10.1161/JAHA.122. 026232 131. Hu C, Liao J, Huang R, Su Q, He L. MicroRNA-155-5p in serum derived- exosomes promotes ischaemia-reperfusion injury by reducing CypD ubiquitination by NEDD4. ESC Heart Fail. (2023) 10(2):1144–57. doi: 10.1002/ehf2.14279 132. Yan X, Huang Y. Mechanism of total glucosides of paeony in hypoxia/ reoxygenation-induced cardiomyocyte pyroptosis. J Bioenerg Biomembr. (2021) 53 (6):643–53. doi: 10.1007/s10863-021-09921-4 152. Cai Y, Zhang B, Shalamu A, Gao T, Ge J. Soluble guanylate cyclase (sGC) stimulator vericiguat alleviates myocardial ischemia-reperfusion injury by improving microcirculation. Ann Transl Med. (2022) 10(12):662. doi: 10.21037/atm- 22-2583 133. Li X, Jin Y. Inhibition of miR-182-5p attenuates ROS and protects against myocardial ischemia-reperfusion injury by targeting STK17A. Cell Cycle. (2022) 21 (15):1639–50. doi: 10.1080/15384101.2022.2060640 133. Li X, Jin Y. Inhibition of miR-182-5p attenuates ROS and protects against myocardial ischemia-reperfusion injury by targeting STK17A. Cell Cycle. (2022) 21 (15):1639–50. doi: 10.1080/15384101.2022.2060640 153. Chen J, Huang Q, Li J, Yao Y, Sun W, Zhang Z, et al. Panax ginseng against myocardial ischemia/reperfusion injury: a review of preclinical evidence and 153. Chen J, Huang Q, Li J, Yao Y, Sun W, Zhang Z, et al. Panax ginseng against myocardial ischemia/reperfusion injury: a review of preclinical evidence and Frontiers in Cardiovascular Medicine 17 frontiersin.org Wang et al. 10.3389/fcvm.2023.1180792 potential mechanisms. J Ethnopharmacol. (2023) 300:115715. doi: 10.1016/j.jep.2022. 115715 potential mechanisms. J Ethnopharmacol. (2023) 300:115715. doi: 10.1016/j.jep.2022. 115715 myocardial ischemia-reperfusion injury. J Nanobiotechnology. (2022) 20(1):454. doi: 10.1186/s12951-022-01652-x 154. Dong L, Shen Z, Chi H, Wang Y, Shi Z, Fang H, et al. Research progress of Chinese medicine in the treatment of myocardial ischemia-reperfusion injury. References Wang J, Liu Y, Liu Y, Huang H, Roy S, Song Z, et al. Recent advances in nanomedicines for imaging and therapy of myocardial ischemia-reperfusion injury. J Controlled Release. (2023) 353:563–90. doi: 10.1016/j.jconrel.2022.11.057 170. Zhou T, Yang X, Wang T, Xu M, Huang Z, Yu R, et al. Platelet-membrane- encapsulated carvedilol with improved targeting ability for relieving myocardial ischemia-reperfusion injury. Membranes. (2022) 12(6):605. doi: 10.3390/ membranes12060605 161. Chen J, Song Y, Wang Q, Li Q, Tan H, Gao J, et al. Targeted neutrophil- mimetic liposomes promote cardiac repair by adsorbing proinﬂammatory cytokines and regulating the immune microenvironment. J Nanobiotechnology. (2022) 20 (1):218. doi: 10.1186/s12951-022-01433-6 162. Mukhamadiyarov RA, Senokosova EA, Krutitsky SS, Voevoda DV, Pyshnaya IA, Ivanov VV, et al. Size-dependent ability of liposomes to accumulate in the ischemic myocardium and protect the heart. J Cardiovasc Pharmacol. (2018) 72 (3):143–52. doi: 10.1097/FJC.0000000000000606 171. Asgarpour K, Shojaei Z, Amiri F, Ai J, Mahjoubin-Tehran M, Ghasemi F, et al. Exosomal microRNAs derived from mesenchymal stem cells: cell-to-cell messages. Cell Commun Signal. (2020) 18(1):149. doi: 10.1186/s12964-020-00650-6 172. Wang D, Niu Z, Wang X. The regulatory role of non-coding RNA in autophagy in myocardial ischemia-reperfusion injury. Front Pharmacol. (2022) 13:822669. doi: 10.3389/fphar.2022.822669 163. Weng X, Tan H, Huang Z, Chen J, Zhang N, Wang Q, et al. Targeted delivery and ROS-responsive release of resolvin D1 by platelet chimeric liposome ameliorates 18 Frontiers in Cardiovascular Medicine 18 frontiersin.org
https://openalex.org/W2905213758	https://bmchealthservres.biomedcentral.com/track/pdf/10.1186/s12913-018-3776-4	English	null	Precision of provider licensure data for mapping member accessibility to Medicaid managed care provider networks	BMC health services research	2,018	cc-by	8,049	© The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Bell et al. BMC Health Services Research (2018) 18:974 https://doi.org/10.1186/s12913-018-3776-4 Bell et al. BMC Health Services Research (2018) 18:974 https://doi.org/10.1186/s12913-018-3776-4 Open Access Precision of provider licensure data for mapping member accessibility to Medicaid managed care provider networks Nathaniel Bell1, Ana Lòpez-DeFede2, Rebecca C. Wilkerson2 and Kathy Mayfield-Smith2 Abstract Background: In July 2018, the Centers for Medicare and Medicaid Services (CMS) updated its Medicaid Managed Care (MMC) regulations that govern network and access standards for enrollees. There have been few published studies of whether there is accurate geographic information on primary care providers to monitor network adequacy. Methods: We analyzed a sample of nurse practitioner (NP) and physician address data registered in the state labor, licensing, and regulation (LLR) boards and the National Provider Index (NPI) using employment location data contained in the patient-centered medical home (PCMH) data file. Our main outcome measures were address discordance (%) at the clinic-level, city, ZIP code, and county spatial extent and the distance, in miles, between employment location and the LLR/NPI address on file. Results: Based on LLR records, address information provided by NPs corresponded to their place of employment in 5% of all cases. NP address information registered in the NPI corresponded to their place of employment in 64% of all cases. Among physicians, the address information provided in the LLR and NPI corresponded to the place of employment in 64 and 72% of all instances. For NPs, the average distance between the PCMH and the LLR address was 21.5 miles. Using the NPI, the distance decreased to 7.4 miles. For physicians, the average distance between the PCMH and the LLR and NPI addresses was 7.2 and 4.3 miles. Conclusions: Publicly available data to forecast state-wide distributions of the NP workforce for MMC members may not be reliable if done using state licensure board data. Meaningful improvements to correspond with MMC policy changes require collecting and releasing information on place of employment. Keywords: Medicaid, Nurse practitioners, Geocoding, Licensure, Patient-centered medical homes Keywords: Medicaid, Nurse practitioners, Geocoding, Licensure, Patient-centered medical homes Background providers and provisions are regulated. Although the new regulations largely make official requirements that were already a staple of state MMC contract practices, it adds greater regulatory oversight and accountability to how states design their managed care and utilize con- tractors [3]. In July 2018, the Centers for Medicare and Medicaid Services (CMS) updated its Medicaid Managed Care (MMC) regulations to ensure adequate network and access standards for enrollees for access to primary care as well as a host of specialist, hospital, and pharmaco- logical services [1]. The rule updated the 2016 federal policy underlying the need that states report on the availability and accessibility of services provided through their MMC networks [2]. In total, 11 different types of Many see MMC network adequacy standards as a means to help beneficiaries more easily navigate and use their coverage [4–6]. In particular, Medicaid recipients have consistently reported less timely access to health care services than other population groups [7–9]. They are also a population group who more often require treatment for complex health conditions, many of which go untreated due to barriers to care access [10, 11]. Correspondence: adefede@mailbox.sc.edu; ADEFEDE@mpr.sc.edu 2Division of Integrated Health and Policy Research, Institute for Families in Society, University of South Carolina, Columbia, SC 29208, USA Full list of author information is available at the end of the article * Correspondence: adefede@mailbox.sc.edu; ADEFEDE@mpr.sc.edu 2Division of Integrated Health and Policy Research, Institute for Families in Society, University of South Carolina, Columbia, SC 29208, USA Full list of author information is available at the end of the article As multiple MCOs partici- pate in state MMC delivery, access to state-wide data on the entire NP workforce allows individual MCOs to ac- tively target network expansion areas based on NP work- force locations. At the same time, NPs are a limited resource for many states owing to migration toward states with fewer practice regulations. Statewide data sources that represent NP workforce distributions there- fore could come to play a potentially pivotal role for advancing alternative MMC models for primary care delivery. One challenge of monitoring current travel time and distance standard calculations is the evolving care team complexity. Historically, states have monitored care capacity through calculating provider-to-population ra- tios of current providers to expected enrollees, or through applying geo-mapping algorithms to calculate distance from enrollees to providers in miles and mi- nutes of drive time. Although geo-mapping algorithms make it feasible to estimate the time and distance MMC populations must travel to obtain healthcare ser- vices, MMCs contract with many different types of providers. In particular, Medicaid programs are increas- ingly relying on nurse practitioners (NPs) for primary care delivery [15–17], particularly for rural and vulner- able populations [18, 19]. This increase is in large part a response to mounting pressures on primary care de- livery, particularly in communities with provider shortages. Another benefit of state licensure data is that it is the only source of information available to link NPs to their precepting/supervising physicians. Twenty states have collaborative practice agreements that require NPs to work under physician supervision [25, 26]. Of these, eight require extended supervision for a period of time (e.g., 2 years), and five limit the number of NPs a phys- ician can supervise (e.g., 2 NPs). At least four of these states also require NPs to practice within the same clinic or specific geographic distance of the physician, thus po- tentially restraining provider penetration into shortage areas. Each of these benefits, however, requires access to accurate information on nursing workforce distributions. At issue is that the accuracy of information on NP practice locations is not easily verifiable. For example, many states currently exclude NPs from workforce as- sessments and forecast projections because they are not universally considered autonomous primary care pro- viders [20]. The same rationale excludes NPs from feder- ally defined health professional shortage areas (HPSA) calculations [21] despite consistently outpacing physi- cians in improving primary care capacity in these areas [22]. Page 2 of 10 Page 2 of 10 Bell et al. BMC Health Services Research (2018) 18:974 Bell et al. BMC Health Services Research (2018) 18:974 Other licensure data sources may also be limited. For instance, all clinicians who elect to participate with CMS are required to have a National Provider Identifier (NPI). However, CMS does not require a clinician to use their personal or professional address when registering their NPI, and no flag is provided that describes the address type provided. Nor is there any requirement that a pro- vider needs to update their address information after switching places of employment. The potential impact of this limitation may not be trivial. One study investigating positional error in address information listed in the na- tional physician masterfile found that nearly 40% of the mailing addresses were over 6 miles from their corre- sponding practice location [24]. CMS emphasizes that such standards will help to protect the long-term health outcomes among beneficiaries by making it more possible to ensure better access to pri- mary and preventive care services [12]. CMS emphasizes that such standards will help to protect the long-term health outcomes among beneficiaries by making it more possible to ensure better access to pri- mary and preventive care services [12]. y p Although many different thresholds underline CMS requirements to ensure adequate access to care, enrollee distance and drive time standards to providers are the most common feature of state contracts [13]. MMC travel and distance standards are allowed to vary by state as well as by provider and service specialty. Maximum distance standards to primary care physicians range from 5 miles in metropolitan areas throughout Arizona to within 60 miles among Frontier areas in New Mexico [14]. Similar thresholds are in place for travel times. For some states, plans are required to demonstrate that a majority of members (e.g., 90%) can access network pro- viders within specific thresholds. The updated federal rule does not change the travel time or distance parame- ters previously established by the state. p g p These challenges aside, a key benefit of licensure data to Managed Care Organizations (MCO) is that they are publicly available, population based, and are released with geographic identifiers. An additional challenge stems from the structure of NP licensure data itself, which often does not differ- entiate whether the registry address reflects where NPs practice or where they reside. For example, in one of the most comprehensive analysis of NP practice distribution to date, the authors could not determine whether the li- censure data from 11 of 12 state workload assessments represented where NPs practiced [23]. The reliance on administrative data to forecast drive times and distances to providers, coupled with growing complexity in care teams who participate in MMCs, un- derscores the need to refocus attention on available data for monitoring the spatial distribution of the NP work- force. Of particular relevance is whether publicly avail- able state licensure board data accurately represent the distribution of the workforce. To our knowledge, little comparable research exists as to whether licensure ad- dress information registered by NPs corresponds with where they work or where they live. As such, the objective of this study was to evaluate the frequency in which geocoded licensure data for the NP workforce Bell et al. BMC Health Services Research (2018) 18:974 Page 3 of 10 Page 3 of 10 contacted by CMS regarding changes/information con- cerning Medicare/Medicaid program. corresponded to their place of employment. To estimate the reliability of licensure data for representing provider practice locations, we “ground truthed” our analysis using practice location data recorded in the National Committee for Quality Assurance (NCQA) patient-centered medical home (PCMH) provider file. The PCMH administrative file is one of the few publicly available sources of informa- tion that can be used to confirm whether the address in- formation obtained by state represents where providers practice. We examined the agreement and discordance (e.g., %) between addresses listed in the PCMH file against two publicly available data sources frequently used to map the NP workforce: state licensure board (LLR) records and the national provider identifier (NPI) file. For comparison purposes, we contrasted our findings against LLR and NPI address information recorded for primary care physicians. The PCMH data feed file is distributed by the NCQA. Our research group receives a monthly data feed file from the NCQA containing practice-level identification numbers, the practice name, recognition level (e.g., Level 3), number of clinicians, certification/expiration year, address, as well as provider-level information for em- ployees, including provider name, their NPI, and creden- tial (e.g. MD, NP). Data sources Our evaluation is based on 2017 South Carolina (SC) NP and physician workforce data. Each year, SC nurses and physicians are required to register or renew their clinical license in order to qualify as a practitioner in the state they wish to practice. Statutory language as to what constitutes a right to practice differs from jurisdiction to jurisdiction (e.g., primary state of residence, demonstra- tion of continuing education, practice hours, etc.). In SC, the public can request an electronic licensee roster from each licensure board for ten dollars. A limited number of data elements are provided in the LLR roster. These include the clinician’s first, middle, and last name; their state licensure number; credential information (e.g. family practice, certified nurse midwife); whether they are board certified for any of their credentials; as well as street-level address information. The LLR does not spe- cify whether the mailing address provided corresponds with a practice location or a personal mailing address. Our analysis employed a composite geocoding meth- odology, which allowed for situs, linear, and area refer- encing. We linked address information (e.g., street name, street suffix, ZIP code) provided in the LLR, NPI, and PCMH databases to street centerline data using the ESRI commercial geocoding software. For the linkages, we used the ESRI Street Map Premium Address File, which is an enhanced version of commercial street reference data from HERE, TomTom, and INCREMENT P. The benefit of purchasing enhanced centerline files is access to more precise and up-to-date address information. Prior to geocoding, we standardized each address file to US Postal Service mailing format to increase the likeli- hood of matching the provider address information between files and with the street centerline file. Standardization was done using ZP4 address correction software. p p g The NPI is a numeric identifier assigned to healthcare providers who elect to provide services to individuals covered under CMS. It is a 10-digit permanent number. Each month, CMS provides an updated NPI release file that is downloadable through the National Plan and Pro- vider Enumeration System (NPPES). The NPI contains elements such as provider first, middle, and last name; a taxonomy description that specifies their credential type; the date in which the provider’s information was last up- dated by CMS; entry space for the provider to list their license number, provider, and state in which they prac- tice; and address-level information pertaining to a mail- ing address. These data elements are also publicly available through the NCQA website [27]. We used the September 2017 LLR, NPI, and NCQA files to ensure currency in the data linkages. Data sources CMS does not require that a clinician use their personal or professional address and no flag is pro- vided that describes the address type provided. The ad- dress field represents where the clinician elects to be Geocoding d Geocoding, also called address matching, is a widely used methodology to map the geographic distribution of health care providers and for identifying neighborhoods or regions where populations are under-served [28–30]. In this approach, electronic databases containing per- sonal identifiers such as address information are spatially linked to situs (i.e., point), linear (i.e., streets), or area (i.e., Census tracts) boundaries and assigned correspond- ing latitude and longitude coordinates. The quality of the address data available for mapping health-related events has long been a notable point of research interest within the geographic, computer science, and mathemat- ical disciplines [31]. This has simply not been the case within network adequacy studies. The lack of research would not necessarily be as significant a problem if the address information contained in provider registries only represented places of employment. Data linkages The address-standardized geocoded data files were linked in SAS using SQL scripting language. Only re- cords that had accurate licensure information across all Page 4 of 10 Bell et al. BMC Health Services Research (2018) 18:974 Page 4 of 10 Page 4 of 10 Bell et al. BMC Health Services Research (2018) 18:974 Bell et al. BMC Health Services Research (2018) 18:974 Primary study variable three registries were included in the analysis. Prior to linking each data file, we used the VLOOKUP function in Microsoft Excel to vet the LLR and NPI licensure numbers listed in the PCMH. We flagged and amended all instances where the PCMH documented the pro- vider’s LLR or NPI incorrectly. As the PCMH file is an administrative file and not used for billing purposes, we presumed the license number contained in the LLR and NPI to be correct. We used provider first, middle, and last names to confirm instances where there was a li- cense mismatch. Our primary study variable was an indicator of accuracy between LLR, NPI and PCMH address fields. Accuracy was measured as a discrete variable to estimate overall agreement as well as a continuous variable, in miles, to measure positional error between the PCMH location and the location of the providers mailing address. Analysis Our study was observational. Discrepancies were evalu- ated using cross tabulations and radar plots. As a pro- vider could be employed at more than one PCMH, we considered an address match to be correct if at least one of the addresses listed in the PCMH file matched the ad- dress information in the LLR or NPI. In these instances, we dropped all additional practice locations from the comparisons in order to avoid deflating the denominator used to assess the overall match rates. All analyses were conducted in ArcGIS, version 10.4 and SAS, version 9.4. We used the SPEDIS procedure in SAS software to identify potential matches that would have been missed due to discrepancies in spelling that may not have been corrected using the ZP4 software (i.e., crossing vs. xing) as well as trailing suffixes (i.e., 100 Main St. vs. 100 Main St. Suite 202B). The SPEDIS function is a form of fuzzy matching; it determines the likelihood of a match be- tween the target characters and returns a score ranging from 0 to 100 [32]. The value 0 signifies a perfect match. We used visual observation of the records to identify a score in which unmatched practice-level addresses were referencing the same location and no false-positive matches were included. We report address matches with and without use of the SPEDIS procedure. Overall provider workforce numbers and data linkage matches In September 2017, a total of 938 physicians and 171 NPs were employed within one or more PCMHs that are recognized as either a family practice/general medi- cine, internal medicine, or pediatric medical home. This workforce distribution represents approximately 15% of the state’s 6387 physicians with a primary specialty in these fields of medicine and 5% of the state’s 3745 NPs who are actively practicing in the state (e.g., having a documented primary physician supervisor). After linking the LLR linkage to the PCMH file, the number of phys- ician and NP address files with matching licensure infor- mation was reduced to 912 and 168 (97 and 98% record retention). After linkage to the NPI file, the number of physician and NP address records with matching LLR, NPI, and PCMH licensure information was reduced to 880 and 162 (94 and 95% record retention). These were the final numbers used for all subsequent evaluations. Prior to all data linkages, we manually edited the licen- sure numbers recorded in the PCMH file for 248 pro- viders. An additional 17 NPI numbers in the PCMH file were also amended. Regional designation After geocoding, we categorized each practice as a rural, suburban, or urban PCMH by spatially assigning the practice location to its corresponding US Census Bureau ZIP code tabulated area (ZCTA). ZCTAs approximate US Postal Service ZIP codes and are defined by the Cen- sus for statistical purposes. In our evaluation, we created ZCTA-level class breaks in order to maximize spatial correlation with county-level classification system based on Census Metropolitan and Micropolitan Statistical Area definitions as well as to highlight variation within counties. Classifications were based on the percentage of the ZCTA’s total population that was urban as per the Census 2010. Urban Areas were defined as ZCTAs with an urban population comprising more than 72.5% of the total population. Suburban ZCTAs were defined from an urban population comprising between 43.0 and 72.5% of the total population. A ZCTA was designed as a rural area if its urban population percentage was less than 43.0%. All 424 of the state’s ZCTAs were designated. Our classification breakdowns for state ZCTAs was in- ternally determined. The objective was to ensure that ZCTAs accurately corresponded to county classifica- tions, while also representing the rural-urban distribu- tion at a smaller aggregate unit (e.g., urban ZCTAs within a rural county would still be urban, even if sur- rounded by rural areas). Accuracy of licensure data based on address concordance Statistics shown in Table 1 correspond to the overall agreement between the LLR, NPI, and PCMH address fields at different geographic scales. For all providers, the match statistics increased in a stepwise fashion as the precision of the analysis decreased (e.g., moving from the clinic location to the county where the PCMH was located). Overall, less than 2% of the LLR records for NPs corresponded to the PCMH where they were Bell et al. Regional designation BMC Health Services Research (2018) 18:974 Page 5 of 10 Table 1 Comparison of address concordance among Nurse Practitioner and Physician LLR and NPI files with employment address locations recorded in state PCMH data file Provider Scale Address matchesa With SPEDIS address matchingb LLR (n, %) NPI (n, %) LLR (n, %) NPI (n, %) Nurse practitioner Practice 3 1.9 72 44.4 8 4.9 103 63.6 City 36 22.2 115 71.0 – – ZIP code 58 35.8 124 76.5 – – County 104 64.2 140 86.4 – – Physician Practice 506 57.5 416 47.3 627 71.3 632 71.8 City 700 79.6 702 79.8 – – ZIP code 752 85.5 776 88.2 – – County 784 89.1 810 92.0 – – aAll percentages derived from an N of 880 for physicians and an N of 162 for nurse practitioners bAll fuzzy-matches using the SPEDIS procedure in SAS on practice-level comparisons were evaluated using a tolerance score of <= 32 Table 1 Comparison of address concordance among Nurse Practitioner and Physician LLR and NPI files with employment address locations recorded in state PCMH data file aAll percentages derived from an N of 880 for physicians and an N of 162 for nurse practitioners bAll fuzzy-matches using the SPEDIS procedure in SAS on practice-level comparisons were evaluated using a tolerance score of <= 3 employed. In comparison, 44% of the NP addresses listed in the NPI matched the address provided in PCMH file. When evaluated at the county scale, 64% of LLR records corresponded with the county of the PCMH. At the same geographic scale, 86% of the NPI records corresponded with the county of the PCMH. Regional designation Table 2 Comparison of address concordance among Physician and NP LLR and NPI licensure files within urban, suburban, and rural PCMHs Provider Data file & scale ZCTA Classification Urban (n, %) Suburban (n, %) Rural (n, %) Nurse Practitioner LLR Practice 7 6.9 0 0.0 1 3.2 City 21 20.8 8 21.6 7 22.6 ZIP code 42 41.6 10 27.0 7 22.6 County 73 72.3 16 43.2 18 58.1 NPI Practice 65 64.4 18 49.6 20 64.5 City 71 70.3 24 64.9 20 64.5 ZIP code 80 79.2 25 67.6 20 64.5 County 91 90.1 28 75.7 22 77.4 Physician LLR Practice 488 70.3 83 66.9 56 57.1 City 545 78.5 87 70.2 68 69.4 ZIP code 597 86.0 91 73.4 65 66.3 County 621 89.5 95 76.6 82 83.7 NPI Practice 481 69.3 91 73.4 60 61.2 City 541 78.0 93 75.0 68 69.4 ZIP code 615 88.6 97 78.2 65 66.3 County 641 92.4 101 81.5 85 86.7 Physician N = 880, Nurse Practitioner N = 162 The SPEDIS procedure improved the practice-level matching for all providers. We found that a SPEDIS score value of ‘<=32’ increased the practice-level matches between the registries without resulting in any false-positive matches. The SPEDIS procedure was not used to improve the overall match rate at the city or county scale because of prior data cleaning. It was not used for ZIP code matches because of meaningful differ- ences in numeric values. Table 2 contrasts address match statistics for NPs and physicians after defining a medical home as an urban, suburban, or rural practice. On average, urban discord- ance was lower in urban areas for NPs than for either suburban or rural providers. Similar differences existed by licensure data source. For example, LLR data for only one NP provider (3.2%) who practiced in a rural area corresponded to the place of employment, whereas 65% of all NPI addresses for rural NPs corresponded to em- ployment location. Similar trends existed for physician licensure data. With the exception of rural physician practice locations using the NPI, the concordance be- tween PCMH and both licensure files was higher among physicians for all geographic types. All ZCTA compari- sons used the SPEDIS match scores for practice level comparisons. The data used to build Table 2 is shown graphically in Fig. 1 using a radar graph. Regional designation For interpretation, greater concentricity (e.g., roundness) in the lines represents more uniform address concordance at each spatial scale Physician N = 880, Nurse Practitioner N = 162 Page 6 of 10 Bell et al. BMC Health Services Research (2018) 18:974 Fig. 1 Radar graphs showing the registry address concordance for all providers at the practice-level, city, ZIP code, and county-level spatial extent across urban, suburban, and rural areas. Scores represent the percentage of all matching address files (e.g., 0 = 0% of addresses matched, 100 = 100% of addresses matched) Fig. 1 Radar graphs showing the registry address concordance for all providers at the practice-level, city, ZIP code, and county-level spatial extent across urban, suburban, and rural areas. Scores represent the percentage of all matching address files (e.g., 0 = 0% of addresses matched, 100 = 100% of addresses matched) For all providers, the positional error depended on whether the address information was derived from the LLR or NPI. Positional error also varied among urban, suburban, and rural practices. In almost every instance, positional error increased when moving from urban to rural locations. For NPs, the greatest amount of pos- itional error in the LLR was among providers in subur- ban and rural areas, with an average difference of 25.3 and 24.7 miles between the medical home where the provider worked and where they received their license. When derived from the NPI, spatial error decreased to 8.1 and 15.9 miles, respectively. There were few discrep- ancies in the positional error among physician addresses based on LLR data across urban, suburban, and rural and for each area type. The further the data lines are to the outer edge of the graph represents greater address concordance between the address information provided in the LLR/NPI and the employment address of the PCMH. Spatial error of licensure data Spatial error across the registries is shown in Table 3. Spatial error was measured as the network distance, in miles, between the location of the PCMH and the geo- coded address registered by the provider in the LLR/ NPI. Statistics shown represent the average, standard de- viation, minimum, and maximum amount of displace- ment between both address locations. Table 3 Positional error (in miles) between the address information provided in registries and to the address location of the PCMH. All statistics are stratified by urban, suburban, and rural PCMHs Data file Statistic Nurse Practitioners Physicians Urban Suburban Rural Urban Suburban Rural LLR Mean spatial error 19.2 25.3 24.7 7.1 7.6 7.5 SD 57.7 36.5 39.1 43.1 25.2 15.0 Min. spatial error 0.0 0.7 0.0 0.0 0.0 0.0 Max. spatial error 538.1 220.3 227.1 735.9 230.4 106.0 NPI Mean spatial error 4.7 8.1 15.9 3.8 4.5 7.9 SD 13.1 13.8 44.2 11.3 13.1 19.6 Min. spatial error 0.0 0.0 0.0 0.0 0.0 0.0 Max. spatial error 118.2 45.1 242.5 188.1 119.5 141.7 miles) between the address information provided in registries and to the address location of the PCMH. urban, suburban, and rural PCMHs Table 3 Positional error (in miles) between the address information provided in registries and to the address All statistics are stratified by urban, suburban, and rural PCMHs Bell et al. BMC Health Services Research (2018) 18:974 Page 7 of 10 Page 7 of 10 actually practice. As shown in this analysis, the potential for incorrect assignment of the NP workforce distribu- tion based on state licensure data is substantial. practices, ranging from 7.1 to 7.6 to 7.5 miles, respect- ively. Based on NPI data, positional error in the phys- ician file decreased to an average of 3.8 and 4.5 miles among urban and suburban providers and increased to 7.9 miles among rural providers. Our evaluation suggests that state licensure data may be a poor source of information for forecasting the location of the entire NP workforce. We found that less than 5 % of all addresses in the LLR corresponded to an NPs place of employment. Even at the county level, nearly 40% of the addresses provided to the LLR do not correspond to the county where NPs practiced. By comparison, the NPI data are more accurate. Sensitivity analysis of address matching concordance We conducted a sensitivity analysis to determine whether the results were potentially biased as a result of providers not updating their address information with CMS. We used the NPI “last update” field to assess the whether there was an inverse association between the duration, in days, since a provider last updated their contact information with CMS and the address matching rate. All evaluations were based on SPEDIS-amended data fields. NPs whose address did not match the address of the PCMH had gone longer since submitting their last up- date to the NPI (1157 days [970] vs. 1022 days [830]), but these differences were not statistically significant (p 0.330). Physicians whose address did not correspond to the PCMH where they worked had updated their profile more recently than those whose address did match the PCMH, but these differences were not statistically sig- nificant (1287 days [1007] vs. 1315 days [1014]; p 0.682). The potential significance of the poor spatial accuracy may be greatest for monitoring network adequacy needs for rural populations. Our analysis showed a 1.3-fold de- crease in accuracy for rural NP LLR addresses and over a 3-fold decrease in accuracy for the same providers based on NPI data. In effect, this distinction could po- tentially create under estimations of care needs if based on provider-to-population ratios, particularly if urban providers are artificially being shifted into a rural county simply due to poor address information. Moving for- ward, systems that choose to use provider-to-enrollee standards for forecasting care needs will need to con- sider that such practices could under or over inflate cap- acity estimates. Spatial error of licensure data The 12-fold increase in practice-level matches for NPs based on the address data registered with CMS suggests that these data are better suited for forecast- ing workforce needs. However, even after address cleaning and standardization we could only obtain practice-level matches for 64% of all NPs using the NPI. This level of accuracy raises concerns over the ability to conduct high-level or granular spatial analyses from the available data. Discussion If the LLR is in fact representing the place of residence of the NP workforce then these data could po- tentially be more useful for looking at the workforce capacity in those areas with respect to scope of practice Page 8 of 10 Page 8 of 10 Bell et al. BMC Health Services Research (2018) 18:974 practice within a recognized PCMH. The transition to PCMHs is a relatively recent phenomenon across the country. As such, the improved match rate among phys- ician addresses could represent the fact that the NP workforce is relatively new addition to integrated models of primary care that were until very recently physician practices. However, the lack of statistical significance in our sensitivity analysis contrasting address matches against the length of time since a provider last updated their information with CMS shows that the available data to test this theory are poor. Continued evaluation of these trends over time could confirm the potential suitability of the PCMH employment file as the de facto standard for measuring positional accuracy in workforce distribution estimates given that these data are publicly available. Lastly, our analysis was state centric. Although SC’s medical home distributions mirror the current “cot- tage industry” practice landscape throughout the country [33], and its geographical distribution of disease and dis- parity trends among racial/ethnic minorities resemble national trends [34–42], further evaluations are war- ranted to confirm if these trends are exhibited regionally as well as nationally. limitations. The subtlety within this context is important given that scope of practice regulations on NPs prac- ticing in SC are some of the most stringent in the coun- try [25]. y Based on these findings, the data suggest that NP workforce projections and distribution estimates are most reflective of the actual workforce location when based on county-level data. One opportunity to improve the granularity of the available data would be to advo- cate for nursing licensure boards to begin recording place of employment information among its members as well as making this information publicly available. From the available literature, this appears to be the practice in the state of Texas, but we are unaware if any other state has taken a similar approach [23]. CMS: Centers for Medicare and Medicaid Services; HPSA: Health Professional Shortage Areas; LLR: Labor, licensing, and regulation; MMC: Medicaid managed care; NCQA: National Committee for Quality Assurance; NP: Nurse practitioner; NPI: National Provider Identifier; NPPES: National Plan and Provider Enumeration System; PCMH: Patient Centered Medical Home; ZCTA: ZIP code tabulated area Conclusion The federal government now formally requires all state Medicaid agencies to establish network and access stan- dards for beneficiaries in effort to ensure adequate ac- cess to primary care and a host of other specialty providers and services. The use of external sources for validating provider practice locations has the potential to add real value to the updated federal rule given that MCOs suffer the sample problem of not requiring pro- viders to update, in a timely manner, changes of employ- ment information. Additionally, when hospitals purchase practices, generic hospital names often replace practice names within data systems, thereby exacerbating the mismatches. As both NPI and PCMH data files are rou- tinely updated and freely available, these data could be used to regularly (e.g. annually) estimate the spatial ac- curacy of primary care provider distributions on a state-wide basis. At the same time, these data are limited for monitoring NP practice locations owing to the lack of distinction over place of practice. Unless there is widespread improvement by state agencies to begin re- cording where NPs practice, there is a strong potential for artificially introducing spatial bias into network ad- equacy evaluations and workforce projections. Limitations The findings from this study do come with four import- ant limitations. First, our findings may be generalizable to the workforce that specializes in family practice, in- ternal medicine, and pediatrics. However, there is no a priori reason to presume that the accuracy of licensure or identification files varies according to a provider’s chosen specialty. Additional evaluations based on Me- dicaid claims data could be used to further distinguish whether these patterns are similar among a larger pro- vider sample. Second, because we could not determine whether each provider was either employed as a full- or part-time clinician we could not estimate whether poor address concordance was due to employment elsewhere or in multiple sites. This is one potential explanation for the poor address match rate based on LLR data. How- ever, the 12-fold increase in address matching among the same provider within the NPI reduces the likelihood that this bias affected our analysis given that a provider can only register a single NPI with CMS. A third and re- lated limitation stems from the use of the PCMH ad- dress file to verify provider employment locations. Approximately 15% of all physicians and 5% of all NPs Discussion To some extent, data- bases such as the NPI as well as the Physician Masterfile avoid this problem by asking clinicians to record whether the registered mailing address represents the lo- cation of a clinical practice. This helps, but there is no mandate that the workplace address must be entered and no flag is distributed in the NPI that specifies whether an employment or residential address was pro- vided. Nor is there any mandate with the NPI that a pro- vider needs to update their address information after switching places of employment. At the same time, NPs themselves may be the best advocates for amending li- censure board data requirements, particularly if the lack of information provided in the registry artificially ex- cludes them from network adequacy forecasts or work- force expansions. Discussion Key network adequacy requirements specified in the 2016 federal regulations for MMC took effect in 2018. Although most states have historically held MCOs to minimum access and accountability standards, the re- cent federal rule underlies the need to ensure adequate access and accessibility for providers who elect to par- ticipate in MMC. The recent federal rule maintains that states continue to evaluate and report access statistics and geographic relationships between enrollees and their healthcare providers. Despite widespread adoption of geographic accessibility standards, and a federal mandate to amend accessibility regulations, to our knowledge, this study is the first study to evaluate the accuracy of publicly available data for monitoring the geographic distribution of available primary care providers who could participate in MMC. Accounting for the geographic detail used in this ana- lysis necessitates a substantial amount of data cleaning and standardization. Our manual inspection of the data found a number of errors in the licensure data entered into the NCQA data file compared to the registries. In addition, the rate of practice-level address agreement improved sub- stantially through using SPEDIS fuzzy-matching tech- niques without generating any false positive matches. The result of these processes warrants consideration as a stand- ard methodological approach for linking disparate data files for mapping workforce distributions. We found little evidence that state licensure data ac- curately reflects the workforce location of NPs. This limitation is significant given the importance of licensure data for mapping the location of primary care providers, assessing geographic relationships between NPs and physician preceptors, as well as forecasting gaps in net- work adequacy. At issue is that the lack of contextual in- formation as to whether the address information in the LLR corresponds to a personal mailing address or a place of employment. This limitation decreases the ac- curacy of any attempt to estimate where providers These discrepancies may also hold some hidden bene- fits. Although we primarily interpret from this analysis that the LLR is a very poor information source for map- ping the NP workforce distribution, another interpret- ation is that it is potentially a very good source of information for mapping the potential workforce distri- bution. Competing interests The authors declare that they have no competing interests. 19. Buerhaus PI, DesRoches CM, Dittus R, Donelan K. Practice characteristics of primary care nurse practitioners and physicians. Nurs Outlook. 2015;63:144–53. Publisher’s Note 20. Morgan P, Strand De Oliveira J, Short NM. Physician assistants and nurse practitioners: a missing component in state workforce assessments. J Interprof Care. 2011;25:252–7. Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. 21. United States General Accounting Office. Health care shortage areas : designations not a useful tool for directing resources to the underserved : report to congressional committees. Washington, D.C.: The Office distributor; 1995. Received: 6 September 2018 Accepted: 28 November 2018 Received: 6 September 2018 Accepted: 28 November 2018 24. McLafferty S, Freeman VL, Barrett RE, Luo L, Shockley A. Spatial error in geocoding physician location data from the AMA physician Masterfile: implications for spatial accessibility analysis. Spat Spatiotemporal Epidemio 2012;3:31–8. Funding CMS, U.S. Department of Health And Human Services; 2017. 13. Murrin S. State standards for access to Care in Medicaid Managed Care: US Department of Health and Human Services Office of Inspector General; 2014 Availability of data and materials 14. Medicaid MCO Access standards. Primary care. 2018. https://www.kff. org/other/state-indicator/medicaid-mco-access-standards-primary-care/ ?currentTimeframe=0&sortModel=%7B%22colId%22: %22Location%22,%22sort%22:%22asc%22%7D. Accessed 14 June 2018. LLR and NPI data files are publicly available or available for a small fee. PCMH data is available from the National Committee for Quality Assurance. Author details 1 1College of Nursing, University of South Carolina, Columbia, SC 29208, USA. 2Division of Integrated Health and Policy Research, Institute for Families in Society, University of South Carolina, Columbia, SC 29208, USA. 22. [homepage]. 2017. https://www.nhsc.hrsa.gov. Accessed 28 Jan 2018. 23. Skillman S, Keppel G, Doescher M, Kaplan L, Andrilla C. Assessing Rural- Urban Nurse Practitioner Supply and Distribution in 12 States Using Available Data Sources. University of Washington, Seattle: WWAMI Rural Health Research Center; 2015. Consent for publication Not Applicable 18. Brom HM, Salsberry PJ, Graham MC. Leveraging health care reform to accelerate nurse practitioner full practice authority. J Am Assoc Nurse Pract. 2018;30:120–30. Authors’ contributions 15. Intrator O, Feng ZL, Mor V, Gifford D, Bourbonniere M, Zinn J. The employment of nurse practitioners and physician assistants in US nursing homes. Gerontologist. 2005;45:486–95. NB, ALD, RCW, and KMS conceptualized the study design. NB and RCW conducted the statistical analysis. NB drafted the initial draft of the article. All authors discussed the results and implications and commented on the manuscript at all stages. All authors read and approved the final manuscript. 16. Lowe RA, Localio AR, Schwarz DF, Williams S, Tuton LW, Maroney S, Nicklin D, Goldfarb N, Vojta DD, Feldman HI. Association between primary care practice characteristics and emergency department use in a medicaid managed care organization. Med Care. 2005;43:792–800. pp p p This study was approved by the University of South Carolina IRB 17. Mobley LR, Subramanian S, Tangka FK, Hoover S, Wang JT, Hall IJ, Singh SD. Breast Cancer screening among women with Medicaid, 2006-2008: a multilevel analysis. J Racial Ethn Health Disparities. 2017;4:446–54. Acknowledgements Not Applicable 12. Lipson DJ, Libersky J, Bradley K, Lewis C, Wishon Siegwarth A, Lester R. Promoting access in Medicaid and CHIP managed care: a toolkit for ensuring provider network adequacy and service availability. Baltimore: Division of Managed Care Plans, Center for Medicaid and CHIP Services, CMS, U.S. Department of Health And Human Services; 2017. References 1. Centers for Medicare and Medicaid Services. Fed. Reg. 27498 (May 6, 2016) (to be codified at 442 C.F.R. Parts 431, 433, 438, 440, 457 and 495): Department of Health and Human Services; 2016. 25. State Practice Environment. 2017. https://www.aanp.org/legislation- regulation/state-legislation/state-practice-environment. Accessed 8 Jan 2017 2. Medicaid and CHIP Managed Care Final Rule (CMS-2390-F). CMS. 2016. https://www.federalregister.gov/documents/2016/05/06/2016-09581/ medicaid-and-childrens-health-insurance-program-chip-programs-medicaid- managed-care-chip-delivered. Accessed 20 June 2018. 2. Medicaid and CHIP Managed Care Final Rule (CMS-2390-F). CMS. 2016. https://www.federalregister.gov/documents/2016/05/06/2016-09581/ medicaid-and-childrens-health-insurance-program-chip-programs-medicaid- managed-care-chip-delivered. Accessed 20 June 2018. 26. State law chart: Nurse Practitioner Practice Authority. 2017. https:// wwwprod.ama-assn.org/sites/default/files/media-browser/specialty%20 group/arc/ama-chart-np-prescriptive-authority.pdf. Accessed 21 Jan 2018. 3. Rosenbaum S, Schmucker S, Beckerman J. Provider Networks and Access in Medicaid Managed Care: A look at federal and state standards. To the point (blog): Commonwealth Fund; 2018. 3. Rosenbaum S, Schmucker S, Beckerman J. Provider Networks and Access in Medicaid Managed Care: A look at federal and state standards. To the point (blog): Commonwealth Fund; 2018. 27. NCQA Report cards: patient centered medical homes. 2017. https:// reportcards.ncqa.org/#/practices/list?recognition=Patient-Centered%20 Medical%20Home. Accessed 5 Nov 2018. 4. Decker SL, Kostova D, Kenney GM, Long SK. Health status, risk factors, and medical conditions among persons enrolled in Medicaid vs uninsured low- income adults potentially eligible for Medicaid under the affordable care act. JAMA. 2013;309:2579–86. 28. Luo W, Wang F. Measures of spatial accessibility to healthcare in a GIS environment: synthesis and a case study in Chicago region. Environ Plan B. 2003;30:865–684. 29. Ricketts T, Randolph R. Urban–rural flows of physicians. J Rural Health. 2007; 23:277–85. 5. Hall MA, Ginsburg PB. A Better approach to regulating provider network adequacy: USC-Brookings Schaeffer initiative for Health Policy. 2017. 30. Peterson L, Bazemore A, Bragg E, Xierali I, Warshaw G. Rural–urban distribution of the U. S. geriatrics physician workforce. J Am Geriatr Soc. 2011;59:699–703. 6. O’Donnell R. Reforming Medicaid for medically complex children. Pediatrics. 2013;131(Suppl 2):S160–2. 6. O’Donnell R. Reforming Medicaid for medically complex children. Pediatrics. 2013;131(Suppl 2):S160–2. 7. Asplin BR, Rhodes KV, Levy H, Lurie N, Crain AL, Carlin BP, Kellermann AL. Insurance status and access to urgent ambulatory care follow-up appointments. JAMA. 2005;294:1248–54. 31. Goldberg D, Cockburn M. The effect of administrative boundaries and geocoding error on cancer rates in California. Spat Spatiotemporal Epidemiol. 2012;3:39–54. 8. Decker SL. Two-thirds of primary care physicians accepted new Medicaid patients in 2011-12: a baseline to measure future acceptance rates. Health Aff (Millwood). 2013;32:1183–7. 8. Decker SL. Abbreviations CMS C f Page 9 of 10 Page 9 of 10 Bell et al. BMC Health Services Research (2018) 18:974 References Two-thirds of primary care physicians accepted new Medicaid patients in 2011-12: a baseline to measure future acceptance rates. Health Aff (Millwood). 2013;32:1183–7. 32. Use of SPEDIS Function in Finding Specific Values. Accessed 5 Nov 2018, at http://www2.sas.com/proceedings/sugi25/25/cc/25p086.pdf. 33. NCQA. Patient Centered Medical Home Data Feed. 2017. 9. Rhodes KV, Kenney GM, Friedman AB, Saloner B, Lawson CC, Chearo D, Wissoker D, Polsky D. Primary care access for new patients on the eve of health care reform. JAMA Intern Med. 2014;174:861–9. 34. Adams SA, Butler WM, Fulton J, Heiney SP, Williams EM, Delage AF, Khang L, Hebert JR. Racial disparities in breast cancer mortality in a multiethnic cohort in the southeast. Cancer. 2012;118:2693–9. health care reform. JAMA Intern Med. 2014;174:861–9. 10. Capp R, Rosenthal MS, Desai MM, Kelley L, Borgstrom C, Cobbs-Lomax DL, Simonette P, Spatz ES. Characteristics of Medicaid enrollees with frequent ED use. Am J Emerg Med. 2013;31:1333–7. 35. Babatunde OA, Adams SA, Eberth JM, Wirth MD, Choi SK, Hebert JR. Racial disparities in endometrial cancer mortality-to-incidence ratios among blacks and whites in South Carolina. Cancer Causes Control. 2016;27:503–11. 11. Thomas LV, Wedel KR, Christopher JE. Access to transportation and health care visits for Medicaid enrollees with diabetes. J Rural Health. 2018;34:162–72. Page 10 of 10 Bell et al. BMC Health Services Research (2018) 18:974 Bell et al. BMC Health Services Research (2018) 18:974 36. Chakraborty H, Weissman S, Duffus WA, Hossain A, Samantapudi AV, Iyer M, Albrecht H. HIV community viral load trends in South Carolina. Int J STD AIDS. 2017;28:265–76. 37. Goldhagen J, Remo R, Bryant T, Wludyka P, Dailey A, Wood D, Watts G, Livingood W. The health status of southern children: a neglected regional disparity. Pediatrics. 2005;116:E746–E53. 38. Hebert JR, Daguise VG, Hurley DM, Wilkerson RC, Mosley CM, Adams SA, Puett R, Burch JB, Steck SE, Bolick-Aldrich SW. Mapping Cancer mortality-to-incidence ratios to illustrate racial and sex disparities in a high-risk population. Cancer. 2009;115:2539–52. 39. Laditka SB, Laditka JN, Probst JC. Racial and ethnic disparities in potentially avoidable delivery complications among pregnant Medicaid beneficiaries in South Carolina. Matern Child Health J. 2006;10:339–50. 40. Levin S, Mayer-Davis EJ, Ainsworth BE, Addy CL, Wheeler FC. Racial/ethnic health disparities in South Carolina and the role of rural locality and educational attainment. South Med J. 2001;94:711–8. 41. Bell et al. BMC Health Services Research (2018) 18:974 36. Chakraborty H, Weissman S, Duffus WA, Hossain A, Samantapudi AV, Iyer M, Albrecht H. HIV community viral load trends in South Carolina. Int J STD AIDS. 2017;28:265–76. 37. Goldhagen J, Remo R, Bryant T, Wludyka P, Dailey A, Wood D, Watts G, Livingood W. The health status of southern children: a neglected regional disparity. Pediatrics. 2005;116:E746–E53. 38. Hebert JR, Daguise VG, Hurley DM, Wilkerson RC, Mosley CM, Adams SA, Puett R, Burch JB, Steck SE, Bolick-Aldrich SW. Mapping Cancer mortality-to-incidence ratios to illustrate racial and sex disparities in a high-risk population. Cancer. 2009;115:2539–52. 39. Laditka SB, Laditka JN, Probst JC. Racial and ethnic disparities in potentially avoidable delivery complications among pregnant Medicaid beneficiaries in South Carolina. Matern Child Health J. 2006;10:339–50. 40. Levin S, Mayer-Davis EJ, Ainsworth BE, Addy CL, Wheeler FC. Racial/ethnic health disparities in South Carolina and the role of rural locality and educational attainment. South Med J. 2001;94:711–8. 41. Pettitt DJ, Talton J, Dabelea D, Divers J, Imperatore G, Lawrence JM, Liese AD, Linder B, Mayer-Davis EJ, Pihoker C, Saydah SH, Standiford DA, Hamman RF, Grp SDYS. Prevalence of diabetes in US youth in 2009: the SEARCH for diabetes in youth study. Diabetes Care. 2014;37:402–8. 42. Probst JC, Moore CG, Glover SH, Samuels ME. Person and place: the compounding effects of race/ethnicity and rurality on health. Am J Public Health. 2004;94:1695–703. References Pettitt DJ, Talton J, Dabelea D, Divers J, Imperatore G, Lawrence JM, Liese AD, Linder B, Mayer-Davis EJ, Pihoker C, Saydah SH, Standiford DA, Hamman RF, Grp SDYS. Prevalence of diabetes in US youth in 2009: the SEARCH for diabetes in youth study. Diabetes Care. 2014;37:402–8. 42. Probst JC, Moore CG, Glover SH, Samuels ME. Person and place: the compounding effects of race/ethnicity and rurality on health. Am J Public Health. 2004;94:1695–703.
https://openalex.org/W2154559038	https://hqlo.biomedcentral.com/counter/pdf/10.1186/1477-7525-10-9	English	null	Validation of the self-management ability scale (SMAS) and development and validation of a shorter scale (SMAS-S) among older patients shortly after hospitalisation	Health and quality of life outcomes	2,012	cc-by	6,136	RESEARCH Open Access Validation of the self-management ability scale (SMAS) and development and validation of a shorter scale (SMAS-S) among older patients shortly after hospitalisation © 2012 Cramm et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Validation of the self-management ability scale (SMAS) and development and validation of a shorter scale (SMAS-S) among older patients shortly after hospitalisation ne M Cramm1, Mathilde MH Strating1, Paul L de Vreede2, Nardi Steverink3 and Anna P Nieboe Abstract Background: The 30-item Self-Management Ability Scale (SMAS) measures self-management abilities (SMA). Objectives of this study were to (1) validate the SMAS among older people shortly after hospitalisation and (2) shorten the SMAS while maintaining adequate validity and reliability. Methods: Our study was conducted among older individuals (≥65) who had recently been discharged from a hospital. Three months after hospital admission, 296/456 patients (65% response) were interviewed in their homes. We tested the instrument by means of structural equation modelling, and examined its validity and reliability. In addition, we tested internal consistency of the SMAS and SMAS-S among a study sample of patients at risk for cardiovascular diseases. Results: After eliminating 12 items, the confirmatory factor analyses revealed good indices of fit with the resulting 18-item SMAS (SMAS-S). To estimate construct validity of the instrument, we looked at correlations between SMAS subscale scores and overall well-being scores as measured by Social Product Function (SPF-IL) and Cantril’s ladder. All SMAS subscales of the original and short version significantly correlated with SPF-IL scores (all at p ≤0.001) and Cantril’s ladder (for the cognitive well-being subscale p ≤0.01; all other subscales at p ≤0.001). The findings indicated validity. Analyses of the SMAS and SMAS-S in the sample of patients at risk for cardiovascular diseases showed that both instruments are reliable. Conclusions: The psychometric properties of both the SMAS and SMAS-S are good. The SMAS-S is a promising alternate instrument to evaluate self-management abilities. Successful aging requires the proactive management of resources in an environment of increasing losses and declining gains [2], and depends on individuals’ abilities to self-manage their lives and aging processes. Although such self-regulation is often related primarily to aspects of physical health, such as physical activity and diet [4-6], the social and psychological aspects of life - social contacts, adaptation, well-being - are equally important to older peoples’ ability to ‘age well’ [7]. Despite acknowledgement of the importance of individuals’ con- tributions to aging successfully and the existence of psy- chosocial theories of successful aging [2,8-12], relatively few suggestions have been made to help older people self-regulate and maintain their well-being [13]. Cramm et al. Health and Quality of Life Outcomes 2012, 10:9 http://www.hqlo.com/content/10/1/9 Background Besides a growing population of older people, a greater proportion live alone and sociological changes have forced them to rely more often on their own resources [1]. They are also hospitalised with increasing frequency as the risk for (multiple) chronic diseases increases with age [3]. They often experience losses in various domains of functioning, which leads to frailty, especially after hospitalisation [2]. Accurately assessing how they man- age their well-being is thus critical. * Correspondence: nieboer@bmg.eur.nl 1Institute of Health Policy & Management (iBMG), Erasmus University, Rotterdam, The Netherlands Full list of author information is available at the end of the article Page 2 of 7 Cramm et al. Health and Quality of Life Outcomes 2012, 10:9 http://www.hqlo.com/content/10/1/9 The self-management of well-being (SMW) theory [13], based on the theory of social production functions (SPF) [14,15], offers guidelines for achieving better self- regulation with regard to well-being. SMW theory posits that successful aging is a life-long process of realizing and sustaining well-being, even in the face of declining resources. Rather than being the process of neutralising losses and discrepancies, successful aging focuses on individuals’ reserve capacities to realize and sustain phy- sical and social well-being using external and internal resources [13]. External resources contribute directly to aspects of well-being, such as food, shelter, fitness (phy- sical well-being) and friends (social well-being). They tend to decline with age. Self-management abilities (internal resources) are needed to manage external resources in such a way that physical and social well- being are maintained or restored when lost [16]. SMW theory incorporates six core abilities to form the compo- site construct of self-management: (1) take initiatives (be instrumental or self-motivating in realizing aspects of well-being); (2) invest in resources for long-term bene- fits; (3) maintain variety in resources (achieve and main- tain various resources for each dimension of well-being); (4) ensure resource multifunctionality (gain and main- tain resources or activities that serve multiple dimen- sions of well-being simultaneously and in a mutually reinforcing way); (5) self-efficaciously manage resources (gain and maintain a belief in personal competence to achieve well-being); and (6) maintain a positive frame of mind. Each of these abilities must be related explicitly to the dimensions of well-being specified in the SPF the- ory: physical well-being (comfort and stimulation) and social well-being (affection, behavioural confirmation, and status) [13-15,17,18]. Analyses Analyses The analyses included the following seven steps. 1. The sample characteristics were analysed using descriptive statistics. 1. The sample characteristics were analysed using descriptive statistics. 2. We data-screened the items by examining the num- ber of missing items and each item’s mean and standard deviation. 2. We data-screened the items by examining the num- ber of missing items and each item’s mean and standard deviation. Background context of the ‘Prevention and Reactivation Care Pro- gramme’, which was designed to prevent loss of function in older patients due to hospitalisation and targeted older hospital patients (≥65 years of age) who were vul- nerable to loss of function after hospital admission. Three months after hospital admission is known to be a good moment to assess effects of a programme [20,21]. Therefore, patients were interviewed three months after hospital admission. Our research is based on the pilot study of 456 patients (≥65 years old) prior to imple- mentation of the ‘Prevention and Reactivation Care Pro- gramme’. The results of the pilot study have been used to identify possible practical implementation problems in preparation for the main evaluation study and serve as a base for power calculations for the main study. We interviewed 296 patients in their homes (response rate 65%). This work was supported by Netherlands organi- sation for health research and development (ZonMw) grant number: 60-61900-98-130. g Ethical approval The study protocol was approved by the medical ethics committee of the Erasmus Medical Centre, Rotterdam, the Netherlands, under protocol number MEC2011-041. Measures The 30-item SMAS consists of six five-item subscales. The scale’s overall internal consistency is 0.90 [19]. Within the subscales of taking initiative, investing, self- efficacy, variety, and multifunctionality, abilities are related to the physical and social dimensions of well- being in the SPF theory [13,14]. The ability to have a positive frame of mind is considered a more general cognitive frame; its subscale is thus not directly related to specific dimensions of well-being. Average overall SMAS scores range from 5 to 30, with higher scores indicating higher SMA. The 30-item Self-Management Ability Scale (SMAS) was developed to measure SMA [19]. Losses in function- ing - something that is especially associated with hospita- lisation - lead to a decreased reserve capacity for coping with losses. Self-management abilities become particu- larly important. Our first objective was to validate the SMAS among older people shortly after hospitalisation. The six subscales of the SMAS reflect the six SMA core abilities. Schuurmans and colleagues [19] concluded that future research could focus on shorter forms of the scale because (i) high correlations were found between some subscales and (ii) some items seemed to be less indicative of SMA (lower loadings). Our second objective was thus to reduce the number of items in the SMAS while main- taining validity and reliability. Overall subjective well-being was measured with the SPF-IL(s) (15-item Social Production Function Instru- ment for the Level of well-being) [17]. The scale inte- grates both affective and cognitive components of well- being, and measures levels of physical and social well- being. Cronbach’s alpha of the SPF-IL in our study was 0.72, indicating a reliable instrument. Cantrill’s Ladder was used to assess satisfaction with life and reflects a general, cognitive evaluation of a per- son’s overall well-being [22]. Methods Study 1 Our study was conducted in 2010 among older people who had recently been admitted to a hospital in the 3. To verify the factor structure of the questionnaire and to test whether the relationship between observed 3. To verify the factor structure of the questionnaire and to test whether the relationship between observed Page 3 of 7 Cramm et al. Health and Quality of Life Outcomes 2012, 10:9 http://www.hqlo.com/content/10/1/9 Page 3 of 7 Cramm et al. Health and Quality of Life Outcomes 2012, 10:9 http://www.hqlo.com/content/10/1/9 version resulted in N = 221. We re-ran the final short version on this sample. variables and their underlying latent constructs existed, confirmatory factor analysis was executed using the LIS- REL program version 8.80 [23]. By using structural equation modelling the overlap between items and dimensions can be traced via modification indices that were used to further refine the measurement model and eliminate potential overlap between items. No correla- tion errors either within or across sets of items were allowed in the model. variables and their underlying latent constructs existed, confirmatory factor analysis was executed using the LIS- REL program version 8.80 [23]. By using structural equation modelling the overlap between items and dimensions can be traced via modification indices that were used to further refine the measurement model and eliminate potential overlap between items. No correla- tion errors either within or across sets of items were allowed in the model. 6. Internal consistency of the subscales was assessed by calculating Cronbach’s alphas, inter-item correlations within each subscale, and correlations between subscales. 7. Validity is the degree to which a scale measures what it is intended to measure; here we focused on the construct validity of the questionnaire. Construct valid- ity is supported if instruments purported to assess the same concept correlate substantially with one another. Since the SPF-IL and SMAS are both based on the SPF theory we evaluated construct validity by comparing the SMAS scale scores with well-being measured by the SPF-IL scale. In addition, we will compare the SMAS scale scores with well-being measured by Cantril’s ladder. 4. Item reduction analysis was performed to develop a short version of the questionnaire. Methods Study 1 Item removal follow- ing several criteria: (i) items were excluded following modification indices provided by LISREL and the strength of the factor loadings; (ii) item elimination stopped when the reliability of each subscale dropped below 0.65; (iii) subscales were left with as few items as possible (but a minimum of three) without loss of con- tent and psychometric quality; and (iv) at least one phy- sical well-being item (comfort or stimulation) and one social well-being item (affection, behavioural confirma- tion or status) was kept in each subscale while maintain- ing validity and reliability. Listwise deletion of cases with missing data on the 30 items resulted in N = 204. Impu- tation was done by replacing missing values with the mean of the data, restoring the original sample of N = 296. Study 2 We additionally tested the SMAS (original and short version) in another longitudinal study sample, namely patients at risk for cardiovascular diseases (low and high-risk). These patients were selected by GPs of pri- mary healthcare practices. At both T0 and T1 Question- naires were mailed to patients’ homes. T1 was about 12 months after T0. A few weeks later, a reminder notice and another copy of the questionnaire were sent to non-respondents. Response rates were 72% (307 out of 426; T0) and 47% (200 out of 425; T1). A detailed description of the study can be found in our study pro- tocol [27]. We used four indices of model fit to test the measure- ment models, with cut-off criteria proposed by Hu and Bentler [24]. First, the overall test of goodness-of-fit assesses the discrepancy between the implied model and the sample covariance matrix by means of a normal-the- ory weighted least squares test. A plausible model has low, preferably non-significant c2 values. Chi-square is, however, overly sensitive when the sample size is large (over 200) [25], leading to difficulty in obtaining a desired non-significant level [26]. Second, the Root Means Square Error of Approximation (RMSEA) reflects the estimation error divided by the degrees of freedom as a penalty function. Values on RMSEA below 0.06 indicate small differences between the estimated and observed model. Values of up to 0.08 suggest a reason- able fit of the model in the population. Third, we used the Standardized Root Means square Residual (SRMR), which is a scale invariant index for global fit that ranges between 0 and 1. Values on SRMR lower than 0.08 indi- cate a good fit. Fourth, we calculated the Incremental Fit Index (IFI), which compares the independent model (i.e., observed variables are unrelated) to the estimated model. Values on IFI are preferably larger than 0.95. y) Analyses Internal consistency of the three subscales (SMAS and SMAS-S) at T0 was assessed by calculating Cronbach’s alphas. At T1 we calculated Cronbach’s alphas of all six SMAS-S subscales. In addition, we assessed correlations between three subscales of the SMAS and SMAS-S at T0 and between three subscales of the SMAS-S at T0 and T1. Study 1 5. After item reduction analyses the first full version and final short version of the instrument were tested on the non-imputed dataset (N = 204). Listwise deletion of missing data on the basis of the 18 items in the short Ethical approval The study was approved by the ethics committee of the Erasmus University Medical Centre of Rotterdam and informed consent was obtained from all participants. Measures At T0 we measured three subscales of the SMAS and SMAS-S (taking initiative, investment behavior and self- efficacy). At T1 we measured the full SMAS-S. Analyses Results 5. After item reduction analyses the first full version and final short version of the instrument were tested on the non-imputed dataset (N = 204). Listwise deletion of missing data on the basis of the 18 items in the short Item reduction analysis Following the factor loadings, modification indices, and an internal consistency check of each subscale, the step- wise procedure resulted in the elimination of 12 items. With respect to the ‘investment behavior’ subscale, modification indices and factor loadings showed that item 7 (getting enough exercise) could be eliminated. The results on the other items of the subscale showed some contradictory results. Eliminating item 6 (having a hobby) resulted in a better fit of the model; however, the physical component was no longer represented in Validity To estimate construct validity of the instrument, we looked at correlations between SMAS subscale scores and overall well-being scores. All SMAS subscales of the original and short versions significantly correlated with SPF-IL scores (all at p ≤0.001) and Cantril’s ladder (for cognitive well-being p ≤0.01; all other subscales p ≤ 0.001), indicating convergent validity. The relative strength of association with SPF-IL scores are the same for the original SMAS (range = 0.311-0.593) and the short version (0.311-0.580), which also applies to the association between Cantril’s ladder and SMAS (0.155- 0.430) and SMAS-S (0.150-0.420). Sample characteristics Respondents’ median age was 75.8 (sd 6.8; range = 65- 94); slightly more were female (54.2%). Just over half Page 4 of 7 Page 4 of 7 Cramm et al. Health and Quality of Life Outcomes 2012, 10:9 http://www.hqlo.com/content/10/1/9 the remaining items (8, 9 and 10) and led to a Cron- bach’s alpha below 0.70. Therefore, based on a lower factor loading of item 8 and construct validity, item 6 remained in the selection and item 8 (actively maintain contact with acquaintances) was eliminated. were married/living together (56.6%); the others were single, widowed or divorced (43.4%). Most lived inde- pendently with others (55.9%); about a third lived inde- pendently alone (37.3%); the remaining lived in elderly or nursing homes (6.8%). The final short version consisted of 18 items with three items for each subscale (table 1). Item reduction was possible without loss of model fit; in fact, its overall fit was better than the full version. Due to a decrease in the number of estimated parameters, the Normal The- ory Weighted Least Square c2 significantly decreased to 530.427. RMSEA still indicated reasonable fit. The value of IFI improved to 0.967, indicating that the specified relations between variables were well supported by the data. The SRMR index decreased to 0.0669, still consid- erably below the cut-off point of 0.08, indicating good global fit. The final short model on imputed data resulted in comparable factor loadings. A re-run of the full model and item reduction analysis on the non- imputed dataset (N = 217) resulted in somewhat less favourable but still acceptable fit indices and comparable factor loadings. Data screening All items were screened for univariate and bivariate nor- mality, and to detect outliers. Data screening informa- tion was taken into account in the stepwise procedure of the item reduction analysis. In general, the percen- tages of missing items were below 10%, except for item 15 (being good at certain things) of the variety subscale (table 1). This was taken into account when interpreting the results of confirmatory factor analysis. C fi t F t A l i All items (table 1) had factor loadings above 0.40 on the intended factor except item 12 (having different ways to relax) and item 18 (doing things for pleasure that benefit others), which were 0.34 and 0.31 respectively. Each SMA measure (except positive frame of mind) was designed with regard to the five dimensions of well- being. We tested the matrix model where each SMA is linked to the dimensions of well-being. The indices in table 2 clearly showed a good fit: a relatively small c2; SRMR had small residuals, indicating good global fit; a small RMSEA within its 90% confidence interval; and a large IFI indicating a good model. Although significant, the Normal Theory Weighted Least Square c2 statistic is not surprising given its sensitivity to sample size. Together the analyses showed that the underlying factors of the items were indeed the dimensions of abilities and well-being. A one-factor model without distinguishing the six subscales resulted in a worse fit (c2 = 2394.115 (p ≤0.0); RMSEA 0.0978; IFI 0.909; SRMR 0.0939). Internal consistency and inter-correlations Internal consistency and inter-correlations Internal consistency as represented by Cronbach’s alpha ranged from sufficient for the ‘variety’ and ‘multifunc- tionality’ subscales to very good for the ‘taking initiative’ subscale (table 3). The correlations between the full ori- ginal subscales and short subscales were also good (0.90-0.95) indicating acceptable coverage of the original sub-dimensions. The six subscales were significantly and positively correlated, indicating conceptually related sub- scales. A one-factor model without distinguishing the six subscales resulted in a worse fit (c2 = 977.270 (p ≤0.0); RMSEA 0.109; IFI 0.929; SRMR 0.0900). In addition, factor loadings were high on the six dimensions, which indicates that although the SMAS-S subscales are related they do represent separate concepts. Validity If we a priori do not link each measure of SMA to the five dimensions of well-being the indices of model fit also showed that the model fit was sufficient (table 2). The RMSEA was just above cut-off value, indicating rea- sonable fit. IFI value was 0.955, near the cut-off value of .95, and SRMR was well below the cut-off value of 0.08. All indices indicated that the model not systematically linked to the five dimensions of well-being was accepta- ble, but left room for improvement. Study 2 We additionally tested the SMAS and SMAS-S in another study sample, namely patients at risk for Page 5 of 7 Cramm et al. Health and Quality of Life Outcomes 2012, 10:9 http://www.hqlo.com/content/10/1/9 Table 1 Item characteristics and factor loadings of the first full model Table 1 Item characteristics and factor loadings of the first full model Item valid N missing mean sd l Taking Initiatives 1. How often do you take the initiative to keep yourself busy? 292 4 1.99 1.09 .65 2. How often are you engaged in making your home or room as comfortable as possible? 291 5 1.53 1.14 .46 3. How often do you take the initiative to get in touch with people who are dear to you? 291 5 1.93 1.05 .82 4. Do you sometimes try to be good at something? 283 13 1.67 1.22 .48 5. How often do you make an effort to have friendly contacts with other people? 291 5 1.67 1.06 .80 Investment Behavior 6. Do you ensure that you have enough interests on a regular basis (such as a hobby) to keep you active? 291 5 2.16 1.16 .65 7. Do you make sure that you get enough physical exercise in order to stay fit longer? 292 4 1.72 1.18 .49 8. Do you occasionally do something so that your contact with your acquaintances remains good? 288 8 1.47 .98 .64 9. Do you devote some time and attention to those who are dear to you in order to maintain good contact? 288 8 2.02 1.00 .77 10. Do you keep busy with the things you are good at so that you stay good at them? 285 11 1.81 1.21 .66 Variety 11. How many hobbies or activities do you have on a regular basis? 289 7 2.03 1.18 .67 12. Do you have different ways to relax when necessary? 289 7 .62 .82 .34 13. Do you have different occasions on which you have friendly contacts with others? 286 10 2.26 1.27 .73 14. With how many people do you have a confidential relationship? 285 11 2.54 1.32 .43 15. Are there certain things that you are good at? 266 30 1.41 1.26 .61 Multifunctionality 16. The activities I enjoy, I do together with others. 290 6 1.60 1.15 .59 17. I sometimes help the people I care about. 285 11 1.82 1.08 .75 18. Study 2 Others benefit from the things I do for my pleasure. 277 19 1.66 1.04 .74 19. I generally spend my holidays with others. 289 7 2.68 1.39 .32 20. I practice my hobbies together with others. 288 8 1.29 1.17 .45 Self-efficacy 21. Are you able to find agreeable activities? 288 8 2.18 1.03 .77 22. Are you capable of taking good care of yourself? 287 9 2.74 1.07 .56 23. Are you able to have friendly contacts with others? 290 6 2.23 1.10 .86 24. Are you able to let others know that you care about them? 286 10 2.24 1.01 .67 25. Are you good at something? 282 14 1.63 1.13 .55 Positive Frame of Mind 26. How often are you able to see the positive side of the situation when something disagreeable happens? 278 18 1.89 1.25 .63 27. When things go against you, how often do you think that it could always be worse? 280 16 2.22 1.34 .79 28. When you are not doing well, how often do you think that there are others who are worse off? 276 20 2.21 1.30 .76 29. When you have a bad day, how often do you think that things will be better tomorrow? 275 21 2.37 1.27 .71 30. When things are not going so well, how often do you succeed in thinking positively? 284 12 2.37 1.13 .72 Items in bold are included in the short version Items in bold are included in the short version behavior’ (0.83 SMAS vs 0.78 SMAS-S), and ‘self-effi- cacy’ (0.84 SMAS vs 0.80 SMAS-S). At T1 we tested all six subscales of the SMAS-S. These results showed that the SMAS-S is a reliable instrument (range from 0.73 for ‘positive frame of mind’ to 0.85 for ‘self-efficacy’). The correlations between the three original SMAS sub- scales and short subscales (SMAS-S) at T0 were also good (0.93-0.95) indicating acceptable coverage of the original sub-dimensions. The three SMAS-S subscales measured at T0 and T1 were also significantly related (0.57-0.70) indicating reliability. behavior’ (0.83 SMAS vs 0.78 SMAS-S), and ‘self-effi- cacy’ (0.84 SMAS vs 0.80 SMAS-S). At T1 we tested all six subscales of the SMAS-S. These results showed that the SMAS-S is a reliable instrument (range from 0.73 for ‘positive frame of mind’ to 0.85 for ‘self-efficacy’). Cramm et al. Health and Quality of Life Outcomes 2012, 10:9 http://www.hqlo.com/content/10/1/9 Discussion Due to high risk of functional losses among older people after hospitalisation, SMA becomes particularly impor- tant. Our objectives were to (1) validate the SMAS among older people who had recently been admitted to a hospital and (2) reduce the number of items in the SMAS while maintaining validity and reliability. After performing an item reduction analysis, the resulting 18- item short version (SMAS-S) was shown to be reliable and valid. The results of the confirmatory factor ana- lyses revealed good indices of fit with the SMAS and SMAS-S. The SMAS-S is thus a good alternative to the lengthier SMAS. We also found high correlations between some subscales in the SMAS and several items may have been less indicative of SMA (lower loadings). Our study showed that the subscales of the SMAS-S represented separate concepts. Therefore, SMA may even be better measured using the SMAS-S. Each mea- sure of 30-item SMAS (except positive frame of mind) is however, specifically related to the five dimensions of We found support for convergent validity of the origi- nal SMAS and SMAS-S through high correlations between the SMA dimensions and subjective well-being as measured by SPF-IL and Cantril’s ladder. We could not evaluate several psychometric properties in this study: the relationship of the SMAS-S with other self-management instruments, assessment of the SMAS- S responsiveness, its predictive value (e.g., clinical out- comes), and different modes of administration. They thus remain undefined. The instrument’s sensitivity to change requires further investigation. We recommend testing the English version of the SMAS-S in other countries to ensure international validity. Last, our sam- ple size was relatively small and our sample population was older people who had recently been discharged from the hospital. Future research is necessary to test Table 3 Scale characteristics and inter-correlations of the shortened subscales (n = 296) items short version Cronbach’s alpha original full scale scale mean (sd) inter-item correlations range 1 2 3 4 5 1. Taking Initiatives 1, 3, 5 .77 .91* 1.86 (.88) .43-.70 2. Investment Behavior 6, 9, 10 .71 .93* 1.98 (.90) .42-.50 .62* 3. Variety 11, 13, 15 .69 .93* 1.90 (.97) .38-.50 .47* .53* 4. Multifunctionality 16, 17, 18 .69 .90* 1.70 (.86) .34-.62 .43* .61* .53* 5. Self-efficacy 21, 23, 24 .77 .94* 2.22 (.87) .47-.61 .63* .71* .49* .57* 6. On imputed data (n = 296) Final short version 18 items Listwise deletion 18 items (n = 221) Final short version 18 items Final short version 18 items well-being specified in the SPF theory and thus provides insight into all five well-being dimensions; the SMAS-S items are related to the two higher-level dimensions (physical and social well-being). On imputed data (n = 296) Listwise deletion 30 items (n = 204) Full and short models with 6 abilities not systematically linked to dimensions of well-being as latent variables Study 2 The correlations between the three original SMAS sub- scales and short subscales (SMAS-S) at T0 were also good (0.93-0.95) indicating acceptable coverage of the original sub-dimensions. The three SMAS-S subscales measured at T0 and T1 were also significantly related (0.57-0.70) indicating reliability. cardiovascular diseases (low and high-risk). At T0 respondents’ median age was 59.8 (sd 9.6; range = 31- 88); slightly more were female (56.4%). The majority were married/living together (76.6%). At T1 respon- dents’ median age was 60.2 (sd 9.1; range = 34-86); 58.2% female and 79.1% were married/living together. At T0 we tested the three subscales ‘taking initiative’, ‘investment behavior’ and ‘self-efficacy’ of both the SMAS and SMAS-S for internal consistency. Cronbach’s alpha of the SMAS and SMAS-S were both reliable: ‘tak- ing initiative’ (0.79 SMAS vs 0.78 SMAS-S), ‘investment Page 6 of 7 Cramm et al. Health and Quality of Life Outcomes 2012, 10:9 http://www.hqlo.com/content/10/1/9 Table 2 Model fit indices full model with 6 abilities systematically linked to dimensions of well-being as latent variables On imputed data (n = 296) Χ2 (p) RMSEA IFI SRMR 30 items 837.874 (0.0) .0438 .985 .0568 Listwise deletion 30 items (n = 204) 30 items 740.991 (0.0) .0472 .984 .0603 Full and short models with 6 abilities not systematically linked to dimensions of well-being as latent variables On imputed data (n = 296) 30 items 1507.845 (0.0) .0689 .957 .0718 Final short version 18 items 523.786 (0.0) .0740 .971 .0644 Listwise deletion 30 items (n = 204) 30 items 1274.298 (0.0) .0734 .955 .0804 Final short version 18 items 454.335 (0.0) .0807 .967 .0755 Listwise deletion 18 items (n = 221) Final short version 18 items 501.856 (0.0) .0845 .964 .0742 full model with 6 abilities systematically linked to dimensions of well-being as latent variables Discussion Positive Frame of Mind 27, 29, 30 .74 .95 2.31 (1.01) .48-.50 .33* .40* .19* .22* .51* * p < 0.01 Table 3 Scale characteristics and inter-correlations of the shortened subscales (n = 296) Page 7 of 7 Cramm et al. Health and Quality of Life Outcomes 2012, 10:9 http://www.hqlo.com/content/10/1/9 Cramm et al. Health and Quality of Life Outcomes 2012, 10:9 http://www.hqlo.com/content/10/1/9 7. von Faber M, Bootsma-van-der-Wiel A, van Exel E, et al: Successful aging in the oldest old: who can be characterized as successfully aged? Arch Intern Med 2001, 161:2694-2700. the SMAS-S on other as well as larger populations. While the SMAS is validated and designed to assess self-management abilities among older people, this study additionally tested the SMAS-S among patients at risk for cardiovascular diseases (aged 30+). Our study find- ings show promising results to assess self-management abilities with the SMAS-S among other populations. 8. Brandtstädter J, Rothermund K: The life-course dynamics of goal pursuit and goal adjustment: a two-process framework. Dev Rev 2002, 22:117-150. 9. Carstensen LL, Isaacowitz DM, Charles ST: Taking time seriously: a theory of socioemotional selectivity. Am Psychol 1999, 54:165-181. 10. Freund AM, Baltes PB: Selection, optimization, and compensation as strategies of life management: correlations with subjective indicators of successful aging. Psychol Aging 1998, 13:531-543. Conclusion 11. Schulz R, Heckhausen J: A life span model of successful aging. Am Psychol 1996, 51:702-714. We conclude that the psychometric properties of both the SMAS and SMAS-S are good and the subscales of SMAS-S clearly represent separate concepts. The SMAS-S is a promising alternate instrument to evaluate self-management abilities. Having a shorter instrument makes it more feasible to assess SMA in a broader num- ber of people, especially among frail older populations. 12. Nieboer AP, Koolman AHE, Stolk EA: Preferences for long-term care services: Willingness to pay estimates derived from a discrete choice experiment. Soc Sci Med 2010, 70(9):1317-1325. 13. Steverink N, Lindenberg S, Slaets JPJ: How to understand and improve older people’s self-management of wellbeing. Eur J Ageing 2005, 2:235-244. 14. Lindenberg S: Continuities in the theory of social production functions. In Verklarende Sociologie [Explanatory Sociology]. Edited by: Ganzeboom H, Lindenberg S. Amsterdam: Thesis Publishers; 1996:169-184. 15. Steverink N, Lindenberg S, Ormel J: Towards understanding successful ageing: patterned change in resources and goals. Ageing Soc 1998, 18:441-467. Acknowledgements S d f d d Study 1 is funded with a grant from the Netherlands organisation for health research and development (ZonMw) grant number: 606190098130. Study 2 was also supported by the ZonMw grant number 300030201. The views are those of the authors. 16. Steverink N, Lindenberg S: Do good self-managers have less physical and social resource deficits and more well-being in later life? Eur J Ageing 2008, 5:181-190. 17. Nieboer A, Lindenberg S, Boomsma A, et al: Dimensions of well-being and their measurement: the SPF-IL scale. Soc Indic Res 2005, 73:313-353. Author details 1 18. Nieboer A, Lindenberg S: Substitution, buffers and subjective well-being: a hierarchical approach. In The Universality of Subjective Well-Being Indicators. Edited by: Gullone E, Cummins RA. Dordrecht: Kluwer Academic Publishers; 2002:175-189. 1Institute of Health Policy & Management (iBMG), Erasmus University, Rotterdam, The Netherlands. 2Erasmus MC, Department of Public Health, Rotterdam, the Netherlands. 3Section Health Psychology, Department of Health Sciences, University Medical Center Groningen, University of Groningen. 19. Schuurmans H, Steverink N, Frieswijk N, et al: How to measure self- management abilities in older people by self-report? The development of the SMAS-30. Qual Life Res 2005, 14:2215-2228. Competing interests Th h d l h The authors declare that they have no competing interests The authors declare that they have no competing interests. The authors declare that they have no competing interests. The authors declare that they have no competing interests. 24. Hu L, Bentler PM: Cutoff criteria for fit indexes in covariance structure analysis: conventional criteria versus new alternatives. Structural Equation Modeling 1999, 6:1-55. Received: 29 September 2011 Accepted: 24 January 2012 Published: 24 January 2012 Received: 29 September 2011 Accepted: 24 January 2012 Published: 24 January 2012 25. Hayduk LA: Structural Equation Modeling with LISREL: Essentials and Advances. Baltimore: Johns Hopkins University Press; 1987. Advances. Baltimore: Johns Hopkins University Press; 1987. Authors’ contributions 20. Beswick AD, Rees K, Dieppe P, Ayis S, Gooberman-Hill R, Horwood J, Ebrahim S: Complex interventions to improve physical function and maintain independent living in elderly people: a systematic review and meta-analysis. Lancet 2008, 371:725-35. JC Preparation of the paper; analyses and interpretation of data; final approval of the version to be published. MS Preparation of the paper; analyses and interpretation of data; final approval of the version to be published. PV Preparation of the paper; acquisition of subjects and data; final approval of the version to be published. NS Preparation of the paper; analyses and interpretation of data; final approval of the version to be published. AN Acquisition of subjects and data; study concept and design; preparation of the paper; analyses and interpretation of data; final approval of the version to be published. 21. Cunliffe AL, Gladman JR, Husbands SL, Miller P, Dewey ME, Harwood RH: Sooner and healthier: a randomised controlled trial and interview study of an early discharge rehabilitation service for older people. Age Ageing 2004, 33(3):246-52. 22. Cantril H: The pattern of human concern. New Brunswick: Rutgers University Press; 1965. 23. Jöreskog K, Sörbom D: User’s Reference Guide. Chicago. Scientific Software International; 1996. doi:10.1186/1477-7525-10-9 Cite this article as: Cramm et al.: Validation of the self-management ability scale (SMAS) and development and validation of a shorter scale (SMAS-S) among older patients shortly after hospitalisation. Health and Quality of Life Outcomes 2012 10:9. References 26. Bagozzi RP, Yi Y, Phillips LW: Assessing Construct Validity in 26. Bagozzi RP, Yi Y, Phillips LW: Assessing Construct Validity in Organizational Research. Admin Sci Q 1991, 36:421-458. 1. Lewis J: Should We Worry About Family Change? Toronto: University of Toronto Press; 2003. 1. Lewis J: Should We Worry About Family Change? Toronto: University of Toronto Press; 2003. Organizational Research. Admin Sci Q 1991, 36:421-458. 27. Lemmens KM, Rutten-Van Molken MP, Cramm JM, Huijsman R, Bal RA, Nieboer AP: Evaluation of a large scale implementation of disease management programmes in various Dutch regions: a study protocol. BMC Health Serv Res 2011, 11(1):6. 2. Baltes PB, Baltes MM: Psychological perspectives on successful aging: the model of selective optimization with compensation.Edited by: Baltes PB, Baltes MM. Successful Aging: Perspectives from the Behavioral Sciences. Cambridge: Cambridge University Press; 1990:1-34. doi:10.1186/1477-7525-10-9 Cite this article as: Cramm et al.: Validation of the self-management ability scale (SMAS) and development and validation of a shorter scale (SMAS-S) among older patients shortly after hospitalisation. Health and Quality of Life Outcomes 2012 10:9. doi:10.1186/1477-7525-10-9 Cite this article as: Cramm et al.: Validation of the self-management ability scale (SMAS) and development and validation of a shorter scale (SMAS-S) among older patients shortly after hospitalisation. Health and Quality of Life Outcomes 2012 10:9. doi:10.1186/1477-7525-10-9 Cite this article as: Cramm et al.: Validation of the self-management ability scale (SMAS) and development and validation of a shorter scale (SMAS-S) among older patients shortly after hospitalisation. Health and Quality of Life Outcomes 2012 10:9. 3. Wolff JL, Starfield B, Anderson G: Prevalence, expenditures, and complications of multiple chronic conditions in the elderly. Arch Intern Med 2002, 126(20):2269-2276. 4. Clark NM, Rakowski W, Wheeler JR, et al: Impact of self-management education on the functional health status of older adults with heart disease. Gerontologist 1992, 32:438-443. 5. Hopman-Rock M, Westhoff MH: The effects of a health educational and exercise program for older adults with osteoarthritis for the hip or knee. J Rheumatol 2000, 27:1947-1954. 6. Holman HR, Lorig KR: Overcoming barriers to successful aging. Self- management of osteoarthritis. West J Med 1997, 167:265-268.
W4286382742.txt	https://zenodo.org/records/6874584/files/Revista%20Casos%20MPNL_N%C2%BA7-337-345.pdf	en		WITUKA: ART FOR EVERYONE	Zenodo (CERN European Organization for Nuclear Research)	2,022	cc-by	4,060	de Marketing Público y No Lucrativo Casos Vol. 7, pp: 337-345 casos-aimpn.org (ISSN: 2530-3422) DOI: https://www.doi.org/ 10.5281/zenodo.6874584 de Marketing Público e Não Lucrativo WITUKA: ART FOR EVERYONE Ana Lanero Carrizo César Sahelices Pinto Universidad de León (España) Noelia Sastre Esteban Marta Rodríguez Prieto Andrea Sánchez Resumen: Wituka, la marca de ropa y complementos sevillana capaz de ofrecer moda sostenible, innovadores diseños y productos de calidad a un precio accesible, a la vez que colabora con la ONG americana EdenReforestationProjects. La firma ha llevado a cabo el magnífico proyecto ecológico “Una camiseta=Un árbol”, con el cual ayuda a la organización a la repoblación de árboles en Haití por cada producto vendido, al mismo tiempo que ofrece la oportunidad de trabajo bien remunerado a los habitantes del país necesitado. Esta empresa es apta para todo tipo de públicos, los productos se fabrican a partir de algodón orgánico y sus originales diseños serigrafiados son elaborados por artistas locales e internacionales. La empresa mantiene así un compromiso medioambiental y social que ha marcado su progresivo éxito desde su fundación en 2016 y que ha conllevado a su crecimiento por España instalando nuevas tiendas físicas. Abstract: Wituka, a Sevilla’s brand of clothes and accessories which is able to offer sustainable fashion. A new style with awesome quality in reasonable prizes as well as cooperating with the American ONG Eden Reforestation Projects. The brand has made a ecological project that consist of planting a tree for each Tshirt sold: “One tee = One tree”. It helps Haiti’s reforestation with each sale, and at the same time, it offer this country’s habitants to get a good payed job. This company fit in all type of consumers. The products are made of organic cotton and their designs are drawn by local and international artists. With this method, the business keep an environmental and social agreement which shows their ongoing success in 2016 and it untils opening new physical stores. 338 Casos de Marketing Público y/e No/Não Lucrativo 1. Introducción Wituka es una pequeña empresa revolucionaria que se dio de alta como sociedad limitada en Junio de 2016 en la ciudad española de Sevilla, motivados por la confección y venta al por menor de productos textiles, accesorios y calzado comprometidos con el medio ambiente. A partir de ahí, la firma española decidió asociarse a la organización americana Eden Reforestation Projects, originada en 2004 en Etiopía y cuya labor se basa en la reforestación en países devastados y sumidos en la pobreza, dando de esta forma un paso más grande para la marca en su labor de contribuir favorablemente al planeta. Su colaboración con la ONG Eden Reforestation Projects se fundamenta en la ayuda a la repoblación forestal en un país como Haití, una región muy dañada por el ser humano y por continuos desastres naturales. Mediante la venta de cualquiera de sus productos hacen partícipes a sus clientes en la plantación de un nuevo árbol en este país ofreciéndole un trabajo digno, seguro y bien remunerado a sus habitantes. Así es como ya han conseguido plantar más de 200.000 árboles y han ofrecido 2.018 días de empleo, teniendo en cuenta que cada día que pasa estas cifras van en progresivo aumento. Algunas de las características más importantes de la marca son, en primer lugar, el uso 100% de algodón orgánico proveniente de Bangladesh en sus productos. Por otra parte, la originalidad de los mismos, ya que son diseñados por reconocidos artistas locales e internacionales, que presentan así la oportunidad de difundir y publicitar su trabajo artístico a la vez que también reciben ciertos beneficios en colaboración por cada artículo vendido; cada vez son más los artistas que participan y no se le cierran puertas a nuevos descubrimientos para poder ampliar su gama de diseños. Además, el proceso de serigrafía de éstos, es llevado a cabo en España apostando por el trabajo tradicional nacional. Las prendas en sí, se fabrican en España, Portugal o Bangladesh y siempre cumpliendo con los más altos estándares de calidad y de manera sostenible, respetando estrictamente los derechos y leyes laborales siendo fieles a sus principios y objetivos de ser una empresa responsable. Son de destacar asimismo dos aspectos, la fabricación de sus zapatillas desarrollada en nuestro país, mediante vulcanizado de caucho y un elaborado proceso artesanal, favoreciendo a la industria española; y el ánimo de la empresa a la sociedad a contribuir más al medioambiente con sencillos trucos para reducir el uso de plástico mediante la venta de bolsas de tela para cualquier ocasión, reflejando esta idea en la frase “plastic bag are for losers” que incluyen en algún diseño y en el embalaje de sus envíos, siempre de papel o cartón. Figura 1. Pantalla de inicio de la página web oficial de Wituka Fuente: https://www.wituka.com/], a fecha 09/10/2019. Con todo esto, sus productos se venden a una increíble calidad-precio que los hace muy atractivos para el cliente. A día de hoy, existen varias tiendas físicas en diferentes lugares de España e Italia, además de la Casos 7, 2020, pp. 337-345 Wituka: Art for everyone 339 propia web online de la marca, con envíos aptos tanto para dentro como fuera de nuestro país pudiendo realizar su seguimiento y con unos tiempos de espera relativamente pequeños, dependiendo de la zona. Se reconoce en cualquier caso que esta empresa concede al cliente un gran poder de satisfacción, ya siendo a través de la plantación un árbol con la compra de sus productos como de otras muchas maneras haciéndole sentir que es capaz de aportar su pequeño granito de arena, para llegar a un futuro mejor en un mundo mejor. La filosofía de Wituka se basa en “crear una comunidad donde todas las partes salgan ganando” y en el caso que se desarrollará a continuación se podrá observar más de cerca y concretando aspectos de su labor. Figura 2. Logotipo oficial de la marca Wituka Fuente: Google Imágenes, a fecha 09/10/2019. 2. Desarrollo del Caso La empresa sevillana creada en el año 2016 por Ricardo Madeiro Galardi, cuenta a día de hoy con un rango de 11 a 50 empleados. La cual se dedica a la fabricación y venta de ropa y complementos generando una moda sostenible a través del uso de prendas fabricadas con algodón orgánico, con un mínimo impacto ambiental y con un compromiso social y laboral. Las prendas de esta empresa están fabricadas por artistas locales y de todo el mundo, quienes reciben beneficios por cada artículo vendido. La filosofía que sigue Wituka es crear una comunidad en la que todas las partes salgan ganando. Cuantos más diseños lanzan más pueden seguir ajustando los precios, por eso la calidad del diseño, el respeto por el planeta y por sus trabajadores no tiene por qué estar reñido con el precio. Las prendas como se indica en la introducción se fabrican en España, Portugal o Bangladesh cumpliendo siempre los más altos estándares de calidad, respetando el medioambiente, los derechos y las leyes laborales bajo valores éticos definidos. Sus prendas cuentan con certificados internacionales que así los reconocen. Cada uno de estos certificados tiene un significado diferente, que son: - FairWear:es una organización independiente de múltiples partes interesadas que trabaja con marcas de prendas de vestir, recibir el sello de aprobación de FWF no garantiza ningún estándar de calidad de trabajo existente, sino que solo demuestra un interés declarado en trabajar para mejorar. - Fairtrade: los productos con este certificado han sido producidos en condiciones de trabajo dignas y comprados a un precio justo que apoya el desarrollo sostenible de la organización productora. - GOT: es la certificación ecológica por parte de un organismo de certificación como Ecocert, que permite colocar en los productos la afirmación «producto textil ecológico» junto con el logotipo de GOTS. - OCS:permite una evaluación y verificación independiente transparente, consistente e integral de las declaraciones de contenido de material orgánico en los productos.Cubre el procesamiento, fabricación, empaque, etiquetado, comercialización y distribución de un producto que contiene al menos 5 por ciento demateriales 'orgánicos' certificados. Casos 7, 2020, pp. 337-345 340 Casos de Marketing Público y/e No/Não Lucrativo - Peta- Approved Vegan: permite a las empresas que venden ropa, accesorios, muebles o decoración del hogar resaltar sus ofertas veganas, ayudando a los consumidores a encontrar productos libres de animales de un vistazo y realizar compras que se alineen con sus valores. Figura 3.Certificados internacionales de Wituka Fuente: Elaboración propia, a fecha 10/10/2019 Es una marca aún joven pero con el paso de los años seguirá trabajando y desarrollando sus ideas sostenibles en un futuro lo que le ayudará a conseguir nuevos certificados con mayor alcance, tal vez a nivel Europeo o incluso internacional. Además de tener en cuenta la fabricación de una forma sostenible, han llevado a cabo una iniciativa por cada venta de una de una camiseta, una zapatilla o una sudadera que no solo están fomentando el talento de diseñadores autónomos por el mundo, sino que también están plantando un árbol y generando trabajo digno en Haití, Nepal, Madagascar,Indonesia, Mozambique y Kenia. Son países muy dañados por el hombre y la naturaleza, en el que ya se han plantado más de 100.000 árboles y se han generados 1023 días de trabajo. Por tanto esta iniciativa ha traído consigo, un enriquecimiento ecológico, el crecimiento económico y humano. La forma de ayudar de Wituka es a través de la organización Eden Reforestation Projects como se ha comentado anteriormente. Es una organización sin fines de lucro y su misión es dar empleo justo a los habitantes empobrecidos. Este trabajo consiste en cultivar, plantar y proteger diversos bosques en distintos países. Su metodología consiste en ‘emplear para plantar’ y esta da como resultado un aumento de medidas socioeconómicas y ambientales positivas. Después de décadas de trabajo y millones de dólares invertidos por la comunidad internacional, Haití sigue siendo uno de los países más degradados del medio ambiente en la tierra. En cuanto a los bosques han desaparecido el 98%, y la Organización de Naciones Unidas estima que el 3% restante de los árboles del país será destruido cada año. El carbón es usado como combustible primario por la mayoría de los habitantes de Haití. Es usado para cocinar, por eso la deforestación es una grave situación en Haití. Además, los efectos de fenómenos meteorológicos como huracanes hace que la deforestación aumente, y esto contribuye a la degradación del suelo, lo que provoca que los habitantes estén obligados a minorar sus cultivos e incluso no poder cultivar. A continuación, son descritas las acciones llevadas a cabo en cada país que ayuda la organización. En Haití, el impacto destructivo masivo que tienen los sistemas ambientales se debe a la falta de protección que se proporciona a los bosques nativos y a los sistemas manglares. Para luchar contra esto, Eden Projects ha estado trabajando directamente con los líderes de la comunidad local a través de asociaciones, para plantar y proteger los árboles hasta la madurez. Con ello, se espera ayudar a restaurar el medio ambiente natural y proporcionar árboles para la seguridad alimentaria. Casos 7, 2020, pp. 337-345 Wituka: Art for everyone 341 Figura 4.Organización EdenReforestationProjects Fuente:https://www.wituka.com/pages/about, a fecha 15/10/2019 Figura 5 - 6.Organización EdenReforestationProjects Fuente:https://edenprojects.org/, a fecha 15/10/2019 Nepal es uno de los países más pobres y menos desarrollado del mundo. Los aldeanos rurales dependen de su entorno natural para obtener sus alimentos y por tanto, sus ingresos. Eden Reforestation Projects se encarga de ayudarles cuando este se daña o se destruye ya que los habitantes más necesitados son los más afectados cuando esto ocurre. La organización ha conseguido plantar en Nepal más de 3 millones de árboles dentro de las tres regiones distintas en las que han trabajado. Una de ellas incluye un parque nacional bastante valioso en el que habitan numerosas especies de animales. Madagascar es una nación con más de 200.000 especies de plantas y animales que no existen en ningún otro lugar del mundo. Pero con el paso del tiempo más del 90% de los bosques han sido destruidos, lo que ha provocado el desplazamiento de especies de animales enteras y se ha eliminando la capacidad de cultivar. La organización está restaurando numerosos manglares ecológicamente devastados en el país. Lo que empezó como una restauración y reforestación de manglares ha terminado en un crecimiento para incluir una variedad de especies nativas. Indonesia en los últimos 30 años ha perdido más del 40% de sus bosques de manglares. Esta situación afecta al medio ambiente y a las especies como ha ocurrido en Madagascar. Eden Reforestation Projects está trabajando con aldeanos locales en la isla de Biak para restaurar, plantar y proteger estos sistemas forestales únicos y vitales. Casos 7, 2020, pp. 337-345 342 Casos de Marketing Público y/e No/Não Lucrativo Mozambique, un país en el que el 68% de su población vive en zonas rurales del país, alberga 20 especies de aves amenazadas a nivel mundial y más de 200 especies endémicas de mamíferos. Además la población vive por debajo del umbral de pobreza, dependiendo en gran medida de sus recursos naturales y bosques para sobrevivir. Los manglares de Mozambique han sido en gran parte destruidos. La organización está trabajando para restaurar, replantar y proteger estos preciosos sistemas forestales. En Kenia y Mozambique tienen una increíble diversidad de tipos de bosques que han mantenido comunidades y vida silvestre desde hace mucho tiempo. Por tanto, el trabajo de la organización tiene como objetivo aliviar la pobreza y restaurar los sistemas socioecológicos. Actualmente se han plantado gracias a la organización y sus colaboradores como en el caso de Wituka la cantidad de 201.780 árboles y se han creado 7.666 empleos. En definitiva Wituka ha demostrado que se puede ofrecer moda sostenible, grandes diseños, creatividad y productos de calidad a un precio asequible para todos los clientes. En definitiva, la visión que tienen, es que para el año 2025 se haya plantado un mínimo de 500 millones de árboles cada año y ofrecer esperanza a través del empleo de decenas de miles de personas en países donde la pobreza extrema es rampante. Al mantener bajos sus costes generales, ahora son reconocidos como uno de los proyectos de reforestación más rentables del planeta. Siendo el objetivo principal sacar a las personas de la pobreza extrema, además de convertirse en un modelo para la restauración ambiental y la gestión de la tierra. En cuanto a las tiendas que tiene disponible, aunque haya nacido en Sevilla, tiene tiendas físicas en Málaga, Granada, Valencia, A Coruña, Barcelona y una tienda en Italia en la localidad de Padua, además de su tienda online . Donde se ofrecen productos como camisetas, zapatillas, sudaderas con diseños muy atractivos. Teniendo los productos tanto para niños como para adultos en varias tallas, y siendo la mayoría de la ropa unisex para que de esta manera no haya distinciones. Figura 7.Muestra de productos de la empresa Fuente:https://www.wituka.com/collections, a fecha 18/10/2019 Asimismo la forma de contactar con la empresa con el fin de dar respuesta a las dudas que pueden surgir es a través del correo electrónico, al cual contestan de manera rápida y eficiente. Otra vía mediante la cual se dan a conocer es a través de sus redes sociales, teniendo página en Facebook, Pinterest e Instragram. En las que la empresa es bastante activa compartiendo novedades, promociones y descuentos que tiene mensualmente. En la página web, además de consultar cuáles son sus redes sociales, su catálogo o el proceso de creación de la empresa, también se puede encontrar una guía de las tallas para orientar al cliente al realizar un pedido online, y unas claves para saber cómo cuidar las prendas a la hora de lavarlas. Una vez que se ha realizado el pedido, de manera inmediata la empresa manda un correo con información del estado del pedido, así como el día que va a llegar y con qué compañía de transporte se recibirá el pedido. Cuando este te llega puedes ver que el paquete en el que viene la ropa es ecológico. Casos 7, 2020, pp. 337-345 343 Wituka: Art for everyone 3. Preguntas para discusión Pregunta 1. ¿Cuál es la repercusión de la RSC en la cartera de clientes de la empresa? Wituka implementa la plantación de árboles con objetivos de reforestación en distintos países y esta campaña está incluida en su Responsabilidad Social Corporativa. Además, sus materias primas son totalmente sostenibles por lo que la empresa consigue tener un marketing verde o ecológico. Su eslogan enfoca a la perfección la estrategia de la empresa, siendo este, 1 Tee = 1 Tree. Este eslogan quiere recordar a los clientes que por la compra de una camiseta se está plantando un árbol para evitar la deforestación en numerosos países, además de un trabajo digno de los trabajadores de estos países menos desarrollados. Desde su creación, en el año 2016, han conseguido un crecimiento continuo tanto en la evolución de sus empleados como en la evolución de sus ventas. Esto se traduce en un éxito total de la empresa y por lo tanto, podemos afirmar que sus estrategias de marketing (ofertas o promociones) como su RSC son efectivas. Por tanto, si sus ventas se han visto incrementadas, la repercusión en la cartera de clientes habrá sido el incremento y ampliación de la misma. Figura 8.Gráfico sobre la evolución de la empresa Fuente:https://guiaempresas.universia.es/WITUKA-ART-FOR-EVERYONE.html, a fecha 14/10/2019 Pregunta 2. ¿Cuáles son las estrategias que usa la empresa para llegar al mercado? La empresa tiene seis tiendas físicas en España y una tienda física en Italia. Son tiendas pequeñas, pero con toda su variedad de productos. También tienen la tienda online (https://www.wituka.com/) en la que puedes adquirir las prendas y además explican sus procesos de reforestación y sus certificados internacionales con respecto a la sostenibilidad y el medio ambiente. Todo esto hace que sus clientes hayan conocido la marca Wituka. Pero, además, la empresa usa otras estrategias de marketing para darse a conocer y intentar crecer más. Una de sus técnicas es que sus diseños son creados por reconocidos artistas locales que se benefician de las ventas. Estos para su beneficio, harán publicidad y se darán a conocer a ellos mismos y así aumentarán las ventas. Es un beneficio para los artistas y para la empresa. Asimismo, usan sus redes sociales como vía de marketing, publicando promociones o fotos de modelos con sus prendas. El caso lo escogimos por ver publicidad en redes sociales. Fue una colaboración de la ‘influencer’ Mery Turiel, que promocionaba unas de las sudaderas de la marca Wituka. Además, contaba a través de un video en ‘Instagram’ como era el proceso de reforestación por cada pedido de la marca. Son frecuentes estas colaboraciones con personas que hoy en día consiguen influir a la sociedad mediante las redes sociales y consiguen que aumenten las ventas de la empresa notablemente. Por último, numerosos periódicos como ABC Sevilla han publicado algún artículo sobre la marca o incluso en alguna revista de producción sostenible. Casos 7, 2020, pp. 337-345 344 Casos de Marketing Público y/e No/Não Lucrativo Pregunta 3. ¿Es necesario este tipo de medidasde RSC en empresas que de por si son sostenibles? Hoy en día la sostenibilidad se ha convertido en un pilar fundamental para todas las empresas, que apuestan más por los sistemas de producción y materias primas más sostenibles. Este proceso no solo se hace por miedo a nuevas regulaciones enfocadas en este ámbito, sino que también se debe al compromiso con el planeta y la sociedad y con el objetivo de dar respuesta a los consumidores. Este tipo de negocios con estrategias sostenibles muy consolidadas crean una imagen de marca mejorada y ventaja competitiva. Si la estrategia es correcta, puede llevar a aumentar la productividad y una disminución de los costes en general. En España, las empresas más conocidas que tengan ámbito sostenible son empresas grandes como Banco Santander, Naturgy o Endesa. Pero estas empresas no tienen nada que ver con empresas pequeñas que realmente se dedican a la producción sostenible. Pregunta 4. ¿Realmente la gente está concienciada de este tipo de tiendas y del precio de estos productos? Hoy en día, pensar en verde ya no es cosa de unos pocos. La población cada vez está más concienciada de que el planeta está atravesando unos momentos críticos y que está de nuestra mano detener el desastre ecológico que se prevé sí diversas acciones mundiales no cambian. Los últimos datos que se han recogido sobre venta de este tipo de productos ecológicos, aseguran que el consumo de ellos ha incrementado en nuestro país estos últimos años. La producción ecológica en España se multiplica por 40 en los últimos 20 años. Pero realmente, es más demandado el sector de la alimentación sostenible que el sector de la moda del mismo tipo. Las grandes superficies y las marcas más conocidas predominan antes que las tiendas ecológicas de menor tamaño como es el caso de Wituka. Por lo tanto, estas tiendas tienen que hacer un esfuerzo mayor por darse a conocer mediante publicidad en redes sociales o llamativas ofertas y promociones. Algunas otras marcas sostenibles del estilo de Wituka son NaturBrush que vende cepillos de dientes ecológicos elaborados con bambú, o Tambalea que también se dedica a la venta de ropa ecológica. 4. Conclusiones Actualmente la responsabilidad social corporativa se ha convertido en algo tan importante que puede llegar a condicionar notablemente el éxito o fracaso de una empresa, por esto cada vez son más las empresas que se suman a ello y no solo para conectar positivamente con sus consumidores sino también por la preocupación de nuestro futuro en un mundo mejor y menos contaminado. Son sencillos y estratégicos los pasos que una empresa puede dar para conseguir estos objetivos, como la marca que se ha analizado y que tanto ha llamado la atención a medida que se va descubriendo cada uno de sus actos. Wituka se ha adentrado en una iniciativa original y sostenible en la que además de ser ellos mismos quienes trabajan de una manera responsable hacen partícipes a sus clientes por lo que deben estar orgullosos y en un futuro podrían verse como los pioneros de una actividad envidiable ante sus competidores. Se puede decir que además de no basarse en sus propios intereses lo que quieren buscar es el bienestar de los demás preocupándose por la gente que actualmente está de verdad necesitada, como son los pobres de esos países que carecen de recursos, de esta manera y con esta iniciativa les ofrecen unas mejores condiciones. Esta iniciativa llama la atención a los clientes ya que muchas otras empresas no se centran en esto y no le dan tanta importancia, lo que hace que a largo plazo tenga más beneficios debido a que es un tema que tiene bastante importancia en la actualidad. Otra observación sobre la empresa, es la preocupación por el medio, es decir el hecho de que se preocupen en la reforestación de los bosques con sus ventas les hace más llamativa a la hora de la venta, porque muchos consumidores deciden comprar para aportar su grano de arena en los beneficios hacia un mundo mejor, y estas son iniciativas que hasta hace poco tiempo no se estaban teniendo en cuenta. Casos 7, 2020, pp. 337-345 Wituka: Art for everyone 345 Además de la forma de producir realizan todas sus prendas con materias primas orgánicas, por tanto se puede concluir que todas aquellas medidas e iniciativas que toman son sostenibles y teniendo en cuenta el medioambiente para conseguir un beneficio que todos estamos buscando hoy en día. A pesar de dudar sobre los certificados de la empresa, después de la investigación y el desarrollo del caso, se puede concluir que es una empresa bastante fiable debido a que uno de los certificados es de la Unión Europea, aunque los demás son de entidades privadas. La empresa podría mejorar en este ámbito consiguiendo más certificados ajustándose a los requisitos que se necesitan aunque son bastante más estrictos. Por último, se afirma que su éxito está en su preocupación por un mundo mejor sin basarse en sus propios intereses de obtener beneficios. Bibliografía Guia Empresas: página web https://guiaempresas.universia.es/WITUKA-ART-FOR-EVERYONE.html Wituka (2016): página web oficial https://www.wituka.com/ Periódico Expansion: noticia sobre la empresa: https://www.expansion.com/directorio-empresas/witukaart-for-everyone-sociedad limitada_8745747_G56_41.html Página web Econoticias: https://www.ecoticias.com/moda-sostenible/184646/Wituka-la-marca-de-ropasostenible-andaluza Red social Facebook: https://www.facebook.com/WitukaShop/ Red social Instagram: https://www.instagram.com/witukaonline/ Red social Pinterest: https://www.pinterest.es/witukashop/ Página web Ecosectores: https://www.ecosectores.com/DetalleArticulo/tabid/64/ArticleId/2762/Witukamoda-sostenible-y-conciencia-social.aspx Página official de EdenReforestationProjects: https://edenprojects.org/ Periódico Andalucia Económica: http://andaluciaeconomica.com/2018/05/wituka-la-marca-de-ropasostenible-sevillana-que-planta-un-arbol-por-cada-producto-que-vende/ Presentación Prezi de Wituka: https://prezi.com/20extobmwg0p/tienda-wituka/ Casos 7, 2020, pp. 337-345
https://openalex.org/W2895954425	http://www.biopolymers.org.ua/pdf/en/29/1/079/biopolym.cell-2013-29-1-079-en.pdf	Latin	null	Модель in vitro для изучения взаимодействия стромальных клеток и клеток опухолевого происхождения	null	2,013	cc-by	3,440	K. A. Shkarina, O. V. Cherednyk, I. O. Tykhonkova, A. I. Khoruzhenko State Key Laboratory on Molecular and Cellular Biology Institute of Molecular Biology and Genetics, NAS of Ukaine 150, Akademika Zabolotnoho Str., Kyiv, Ukraine, 03680 State Key Laboratory on Molecular and Cellular Biology Institute of Molecular Biology and Genetics, NAS of Ukaine 150, Akademika Zabolotnoho Str., Kyiv, Ukraine, 03680 a.i.khoruzhenko@imbg.org.ua Aim. To develop a model to study the interaction between tumor and stromal cells in three-dimensional culture. Methods. Cultivation of HeLa cell lines and human dermal fibroblasts in monolayer and three-dimensional culture, immunofluorescent and immunohistochemical analysis. Results. In this work we present an approach based on a direct interaction between the cells of multicellular tumor spheroids and spheroids of fibroblasts. Subsequent immunofluorescence analysis allows to determine an origin of cells in the area of their contact. Conclusions. This model will be useful to study the basic mechanisms of carcinogenesis, and to find targets for anticancer therapy. Keywords: interaction of stromal and tumor cells, three-dimen- sional cultures of malignant cells, multicellular spheroids. fibroblasts, similar to that observed during wound healing and tissue fibrosis [3]. The process of fibroblast activation includes the change in a number of morpho- functional features as well as in the expression of some specific biochemical markers, such as smooth muscle actin, fibroblast-specific protein (FSP) and fibroblast- activating protein (FAP) etc. The activated fibroblasts stimulate further tumor progression through the secre- tion of a high number of growth factors, hormones and cytokines, including PDGF-/, TGF1, bFGF, IL-6, LPA and eicosanoids; they also affect the tumor development via establishment of direct intercellular contacts and remodeling the surrounding matrix [1, 4]. The tumor-associated fibroblasts do not only stimulate the proliferative activity of tumor cells, but also participate in the initiation of their epithelial-mesen- chymal transition. In addition, they are capable to ini- tiate the formation of cancer stem cells phenotype in the population of malignant cells [4] and play an important role in the formation of organ-specific metastatic niches, which are favorable for the extravasion of circu- lating cancer cells and the formation of secondary tumors [5]. Introduction. The growth and development of a malignant tumor in organism occur in the context of complex homo- and heterotypical intercellular inter- actions that have significant effect on the survival and proliferation of tumor cells as well as on the processes of invasion, angiogenesis and metastasis [1]. Fibro- blasts are the main cellular component of the malignant tumor stroma. UDC 576 UDC 576 Institute of Molecular Biology and Genetics, NAS of Ukraine, 2013 doi: 10.7124/bc.00080A doi: 10.7124/bc.00080A ISSN 1993–6842. Biopolymers and Cell. 2013. Vol. 29. N 1. P. 79–82 (Translated from Russian) In vitro model for study the interaction between tumor and stromal cells K. A. Shkarina, O. V. Cherednyk, I. O. Tykhonkova, A. I. Khoruzhenk K. A. Shkarina, O. V. Cherednyk, I. O. Tykhonkova, A. I. Khoruzhenko They have been shown to secrete the components of extracellular matrix such as fibrillar collagens, fibronectin and tenascin C along with the production of a considerable amount of auto- and para- crine growth factors [2, 3]. Besides, fibroblasts have also been shown to participate in extracellular matrix remodeling via the secretion of several kinds of proteo- lytic enzymes, such as matrix metalloproteinases MMP2, MMP3 and MMP9 [1]. Cancer cells are known to modulate the properties of surrounding connective-tissue stroma and change its features, thus creating favorable microenvironment for further tumor progression. The process of malignant tumor development involves the activation of stromal 79 SHKARINA K. A. ET AL In vivo and in vitro experiments have demonstrated that stromal elements affect the sensitivity of malignant cells to anti-cancer drugs and participate in the forma- tion of tumor chemoresistance. At the same time there is convincing evidence of the fact that in some cases the normal cellular microenvironment is capable to affect the development of a malignant tumor, decreasing its proliferative activity and reducing the risk of invasion [6]. Therefore, heterotypical in vitro models employing stromal cells (e.g. fibroblasts) may become a conve- nient tool for investigation of molecular basis of reciprocal effect of stromal and transformed cells and their role in the initiation and inhibition of oncogenesis processes. Moreover, they will supply the information, required for the development of new approaches to prevention and therapy of oncological diseases [2, 4, 6]. formation of multicellular spheroids [7]. Recently, the increased application of 3D culture allowed the deve- lopment of a great number of methods to obtain these aggregates, beginning from traditional techniques, like a hanging drop method, the roller flasks method and the sandwich method, and up to the creation of various bio- engineering systems and microreactors for rapid isola- tion of multiple spheroids [8, 9]. The study of cancer and stromal cells behavior in three-dimensional conditions gives a better understan- ding of the heterotypical intercellular interactions in the structure of a malignant tumor and of the role of stroma components in the regulation of different stages of oncogenesis [9, 10]. K. A. Shkarina, O. V. Cherednyk, I. O. Tykhonkova, A. I. Khoruzhenko The following methods are most frequently used to obtain heterotypical aggregates: the formation of spheroids of a mixed type from the cell suspension, the cultivation of spheroids of one type of cells with a monolayer of cells of another type, the addition of the cell suspension to formed spheroids of another type of cells [10, 11]. There are a lot of methods which have been applied to study intercellular interactions in vitro, including indirect methods, based on the determination of the effect of a conditioned medium, obtained from fibro- blast or other stroma cells, on the behavior of cancer cells in cell culture, different technologies for co-cultu- ring cells of two and more types, etc. Both two- and three-dimensional systems, consisting of two and more cell types are used as well as other additional modi- fications [7, 8]. Here we present the method based on the inter- action of two types of multicellular spheroids origi- nated from human fibroblasts and HeLa cells. Materials and Methods. Cell culture. HeLa cells originated from human cervical carcinoma were cul- tured in the DMEM medium (Sigma, USA) supple- mented with 10% fetal calf serum (FBS, Hy Clone, USA), 50 units/ml penicillin, 50 µg/ml streptomycin, 4 mM glutamine (PAA, USA) until reaching 80–90 % monolayer confluence. Then the cells were detached using 0.25 % trypsin-0.02 % EDTA (PAA) solution and calculated in Goryaev's chamber. 1 % agarose gel was prepared ahead of time (#11400, Serva, USA) and autoclaved. 1 ml of hot agarose was layered in the wells of 24-well plate. After cooling it was washed with the serum-containing medium. Approximately 5·10 3 cells were added to each well and cultivated for 5 days to observe the formation of multicellular aggregates. The main advantage of three-dimensional cultiva- tion is the maintenance of the polarity and morphology as well as of the similarity of gene expression profile and activation of intracellular signaling cascades, specific for the initial tissue. It has been demonstrated that cultured cells in 2D and 3D conditions significantly differ in a variety of morphological and biochemical parameters, including the expression of receptors to the growth factors and a number of paracrine regulators, as well as in the level of proteolytic enzyme activity, migration ability and sensitivity to some therapeutic agents [8]. The primary culture of human dermal fibroblasts was kindly provided by K. A. Nizheradze (A. V. K. A. Shkarina, O. V. Cherednyk, I. O. Tykhonkova, A. I. Khoruzhenko After the completion of cultivation term the combined three-dimensional cell cultures obtained were fixed with 10 % formalin and the paraffin sections were pre- pared according to the histological protocol. Immuno- fluorescent analysis was applied to determine fibroblast and epithelial cells in the formed aggregates. The nuclei of cells were stained with the propidium iodide. A strong immunofluorescent reaction with antibodies to the surface of human fibroblasts on monolayer culture was observed, but there was a considerably weak reac- tion on histological sections. Therefore, the antibodies to vimentin were used to distinguish epithelial cells and fibroblasts. Strong immunohistochemical reaction with anti-vimentin antibodies was observed in fibro- blasts from the sections of interacting aggregates, but a weak reaction was present in HeLa cells as well, which corresponds to the literature data (Fig. 4, 5, see insert). Thus, the character of reaction in the cells was ob- viously different. The antibodies to epithelial antigens have demonstrated a positive reaction in HeLa cells, but there was no reaction in fibroblasts (Fig. 2, 3, see insert). The immunohistochemical and immunofluorescent analysis. The histological sections were boiled in 10 mM citrate buffer, pH 6, in the microwave oven (700 W) twice for 5 min with the 5 min break, for antigen retrieval procedure [12]. The autofluorescence of specimens was inhibited by the addition of 10 mM copper sulfate, 50 mM ammonium acetate buffer, pH 5, for 20 min at room temperature. The epithelial antigens were determined using mouse monoclonal antibodies to cytokeratins (anti-Pan cytokeratin, Clone 11, Sigma) in 1:100 ratio, and the fibroblasts – using antibodies to vimentin (Dako, Denmark) in 1:100 ratio. The visu- alization of antigens was conducted using secondary anti-mouse FITC-stained antibodies (Jacksonimmuno- research, USA), and the immunoperoxidase reaction – using Ultra Vision LPValue Large Volume Detection system (Thermo Scientific, USA). The specimens were analyzed using Leica DM 1000 microscope (Germany). Results and Discussion. The problem of inte- ractions between a malignant tumor and its stroma is very complicated. Earlier it was demonstrated that during the initial stages of breast cancer formation the stromal compartment can significantly inhibit the tumor growth and spreading, but at the later stages it may stimulate the tumor progression [5]. The problem of bilateral interactions of tumor and stromal cells has been a subject for a great number of studies, thus many models have been elaborated to investigate these inte- ractions. K. A. Shkarina, O. V. Cherednyk, I. O. Tykhonkova, A. I. Khoruzhenko Dumansky Institute of Colloid and Water Chemistry, NAS of Ukraine). The fibroblasts were cultivated in F-12 medium with 10 % fetal bovine serum, 50 units/ml penicillin, 50 µg/ml streptomycin, 4 mM glutamine. As described above for HeLa cells, the fibroblasts were trypsinized by 0.25 % trypsin-0.02 % EDTA solution. There are a number of approaches to organotypic three-dimensional cultures, the most often used is the formation of multicellular spheroids of one or more cell types. To obtain three-dimensional cellular aggregates the cell suspension is applied to Matrigel or non-adhesive substrate like agarose or their analogues, which pro- motes the initiation of intercellular contacts and the 80 ÌÎÄÅËÜ ÂÇÀÈÌÎÄÅÉÑÒÂÈß ÎÏÓÕÎËÅÂÛÕ È ÑÒÎÌÀËÜÍÛÕ ÊËÅÒÎÊ 5·10 3 cells per well were transferred into the 24-well plate, covered with agarose to allow the formation of multicellular aggregates. After five days of cultivation the aggregates, formed by HeLa cells and fibroblasts, were placed in one well and their fusion was observed. The aggregates were fixed in 10 % formalin, paraffin sections were obtained by the standard histological method. 5·10 3 cells per well were transferred into the 24-well plate, covered with agarose to allow the formation of multicellular aggregates. After five days of cultivation the aggregates, formed by HeLa cells and fibroblasts, were placed in one well and their fusion was observed. The aggregates were fixed in 10 % formalin, paraffin sections were obtained by the standard histological method. The initial culture of dermal fibroblasts was ana- lyzed by the immunohistochemical method using anti- bodies to the surface antigen of human fibroblasts; no cells with negative reaction were observed (the data are not presented). The mentioned fibroblasts served as a source to obtain multicellular aggregates, spheroids. HeLa cells formed spheroids on the second day of cultivation on the agar coated surface. The fusion of ag- gregates was observed during the co-cultivation of fib- roblast spheroids and HeLa spheroids on the following day (Fig. 1, see the insert). In many cases the number of fibroblast spheroids was lower than the number of He- La aggregates. In course of the interaction of aggrega- tes, formed by HeLa cells with the fibroblast aggrega- tes, the epithelial cells tended to locolize on top of the fibroblast spheroids, that corresponds to the general histogenesis mechanisms (Fig. 2, 3, see the insert). K. A. Shkarina, O. V. Cherednyk, I. O. Tykhonkova, A. I. Khoruzhenko The current work presents a modified ap- proach to the modeling ofstroma-tumor interaction. Classic models suggest the use of mixed cell sus- pension, placed in non-adhesive conditions. Our work presents the method, based on the interaction of already formed multicellular aggregates of fibroblasts (as main cellular stromal elements) and tumor cells. This ap- proach enables studying the direct interactions between tumor cells and fibroblasts using morphological and immunohistochemical methods. So, we managed to fix the area of the contact of tumor cells and fibroblasts at the histological sections of 3D cultures. This co-culture of stromal and mali- gnant cells will be used in our further research on the intracellular signaling in tumor cells and fibroblasts. The presented model to study the interaction between tumor and stroma can be useful for the investigation of both fundamental mechanisms of 81 ISSN 0233-7657. Biopolymers and Cell. 2013. Vol. 29. N 1 Figure to article by K. A. Shkarina et al. a b HeLa HeLa Fibroblast Fibroblast Fig.1. Co-culture of spheroids of human dermal fibroblasts and HeLa cells. a - 1 day of cultivation. Oc. 10x, Ob. 20x; b - 4 days of cultivation. Oc. 10x, Ob. 10x HeLa Fibroblast Fig.2. Multicellular aggregate obtained after fusion of spheroids of HeLa and fibroblast spheroid. Immunoperoxidase reaction (brown staining) revealed epithelial, cytokeratin positive cells surrounding fibroblast spheroid. Oc. 10x, Ob. 10x Fig.3. Immunohistochemical revealing of epithelial antigens (cytokeratins, intermediate filaments of epithelium) in HeLa cells forming spheroid. Oc. 10x, Ob. 40x Figure to article by K. A. Shkarina et al. a b HeLa HeLa Fibroblast Fibroblast Fig.1. Co-culture of spheroids of human dermal fibroblasts and HeLa cells. a - 1 day of cultivation. Oc. 10x, Ob. 20x; b - 4 days of cultivation. Oc. 10x, Ob. 10x HeLa Fibroblast Fig.2. Multicellular aggregate obtained after fusion of spheroids of HeLa and fibroblast spheroid. Immunoperoxidase reaction (brown staining) revealed epithelial, cytokeratin positive cells surrounding fibroblast spheroid. Oc. 10x, Ob. 10x Fig.4. Immunofluorescent revealing of vimentin (fibroblast inter- mediate filaments) in spheroid formed fibroblasts. There is a green staining of cytoplasm. Oc. 10x, Ob. 40x Fig.3. Immunohistochemical revealing of epithelial antigens (cytokeratins, intermediate filaments of epithelium) in HeLa cells forming spheroid. Oc. 10x, Ob. 40x HeLa Ôèáðîáëàñòû Fig. 5. Immunofluorescent detection of fibroblast by vimentin content (green) in aggregate in co-culture of HeLa and fibroblast spheroids. Cell nuclei are stained with PrI. Oc. 10x, Ob. K. A. Shkarina, O. V. Cherednyk, I. O. Tykhonkova, A. I. Khoruzhenko Õîðóæåíêî Ìîäåëü in vitro äëÿ èçó÷åíèÿ âçàèìîäåéñòâèÿ ñòðîìàëüíûõ êëåòîê è êëåòîê îïóõîëåâîãî ïðîèñõîæäåíèÿ Ìîäåëü in vitro äëÿ èçó÷åíèÿ âçàèìîäåéñòâèÿ ñòðîìàëüíûõ êëåòîê è êëåòîê îïóõîëåâîãî ïðîèñõîæäåíèÿ Summary Öåëü. Ðàçðàáîòàòü ìîäåëü äëÿ èçó÷åíèÿ âçàèìîäåéñòâèÿ îïóõî- ëåâûõ è ñòðîìàëüíûõ êëåòîê â óñëîâèÿõ òðåõìåðíîé êóëüòóðû. Ìåòîäû. Êóëüòèâèðîâàíèå êëåòîê ëèíèè HeLa è äåðìàëüíûõ ôèáðîáëàñòîâ ÷åëîâåêà â ìîíîñëîéíîé è òðåõìåðíîé êóëüòóðå, èììóíîôëóîðåñöåíòíûé è èììóíîãèñòîõèìè÷åñêèé àíàëèç. Ðå- çóëüòàòû. Ïðåäëîæåí ïîäõîä, áàçèðóþùèéñÿ íà èññëåäîâàíèè íåïîñðåäñòâåííûõ âçàèìîäåéñò- âèé êëåòîê ìíîãîêëåòî÷íûõ ñôåðîèäîâ îïóõîëåâûõ êëåòîê è ñôåðîèäîâ ôèáðîáëàñòîâ. Ïîñ- ëåäóþùèé èììóíîôëóîðåñöåíòíûé àíàëèç äàåò âîçìîæíîñòü îïðåäåëèòü ïðîèñõîæäåíèå êëåòîê â çîíå èõ êîíòàêòà. Âûâîäû. Îïèñàííàÿ ìîäåëü áóäåò ïîëåçíà êàê äëÿ èçó÷åíèÿ áàçîâûõ ìåõà- íèçìîâ êàíöåðîãåíåçà, òàê è ïðè ïîèñêå ìèøåíåé äëÿ ïðîòèâîî- ïóõîëåâîé òåðàïèè. 4. Cirri P., Chiarugi P. Cancer associated fibroblasts: the dark side of the coin // Am. J. Cancer Res.–2011.–1, N. 4.–P. 482–497. 5. Joyce J. A., Pollard J. W. Microenvironmental regulation of me- tastasis // Nat. Rev. Cancer.–2009.–9, N 4.–P. 239–252. 6. Wadlow R. C., Wittner B. S., Finley S. A., Bergquist H., Upad- hyay R., Finn S., Loda M., Mahmood U., Ramaswamy S. Sys- tems-level modeling of cancer-fibroblast interaction // PLoS One.–2009.–4, N 9.–e6888. 7. Nyga A., Cheema U., Loizidou M. 3D tumor models: novel in vit- ro approaches to cancer studies // J. Cell Commun. Signal.– 2011.–5, N 3.–P. 239–248. 8. Yamada K. M., Cukierman E. Modeling tissue morphogenesis and cancer in 3D // Cell.–2007.–130, N 4.–P. 601–610. Êëþ÷åâûå ñëîâà: âçàèìîäåéñòâèå ñòðîìàëüíûõ è îïóõîëåâûõ êëåòîê, òðåõìåðíûå êóëüòóðû çëîêà÷åñòâåííûõ êëåòîê, ìíîãîê- ëåòî÷íûå ñôåðîèäû. 9. Lin R. Z., Chang H. Y. Recent advances in three-dimensional multicellular spheroid culture for biomedical research // Biotech- nol. J.–2008.–3, N 9–10.–P. 1172–1184. Ê. À. Øêàð³íà, Î. Â. ×åðåäíèê, ². Î. Òèõîíêîâà, À. ². Õîðóæåíêî Ê. À. Øêàð³íà, Î. Â. ×åðåäíèê, ². Î. Òèõîíêîâà, À. ². Õîðóæåíêî 10. Miki Y., Ono K., Hata S., Suzuki T., Kumamoto H., Sasano H. The advantages of co-culture over mono cell culture in simulating in vivo environment // J. Steroid Biochem. Mol. Biol.–2012.– 131, N 3–5.–P. 68–75. Ìîäåëü in vitro äëÿ äîñë³äæåííÿ âçàºìîä³¿ ñòðîìàëüíèõ êë³òèí ³ êë³òèí ïóõëèííîãî ïîõîäæåííÿ 11. Hirschhaeuser F., Menne H., Dittfeld C., West J., Mueller-Klie- ser W., Kunz-Schughart L. A. Multicellular tumor spheroids: an underestimated tool is catching up again // J. Biotechnol.– 2010.–148, N 1.–P. 3–15. REFERENCES. Ãîñóäàðñòâåííàÿ êëþ÷åâàÿ ëàáîðàòîðèÿ ìîëåêóëÿðíîé è êëåòî÷- íîé áèîëîãèè Ãîñóäàðñòâåííàÿ êëþ÷åâàÿ ëàáîðàòîðèÿ ìîëåêóëÿðíîé è êëåòî÷- íîé áèîëîãèè Ãîñóäàðñòâåííàÿ êëþ÷åâàÿ ëàáîðàòîðèÿ ìîëåêóëÿðíîé è êëåòî÷- íîé áèîëîãèè 1. Kalluri R., Zeisberg M. Fibroblasts in cancer // Nat. Rev. Cancer.– 2006.–6, N 5.–P. 392–401. Èíñòèòóò ìîëåêóëÿðíîé áèîëîãèè è ãåíåòèêè ÍÀÍ Óêðàèíû Èíñòèòóò ìîëåêóëÿðíîé áèîëîãèè è ãåíåòèêè ÍÀÍ Óêðàèíû 2. Pietras K., Ostman A. Hallmarks of cancer: interactions with the tumor stroma // Exp. Cell Res.–2010.–316, N 8.–P. 1324–1331. Óë. Àêàäåìèêà Çàáîëîòíîãî, 150, Êèåâ, Óêðàèíà 03680 Óë. Àêàäåìèêà Çàáîëîòíîãî, 150, Êèåâ, Óêðàèíà 03680 3. Xing F., Saidou J., Watabe K. Cancer associated fibroblasts (CAFs) in tumor microenvironment // Front Biosci.–2010.– 15.–P. 166–179. K. A. Shkarina, O. V. Cherednyk, I. O. Tykhonkova, A. I. Khoruzhenko 40x Figure to article by K. A. Shkarina et al. a b HeLa HeLa Fibroblast Fibroblast Fig.1. Co-culture of spheroids of human dermal fibroblasts and HeLa cells. a - 1 day of cultivation. Oc. 10x, Ob. 20x; b - 4 days of cultivation. Oc. 10x, Ob. 10x Figure to article by K. A. Shkarina et al. Fig.1. Co-culture of spheroids of human dermal fibroblasts and HeLa cells. a - 1 day of cultivation. Oc. 10x, Ob. 20x; b - 4 days of cultivation. Oc. 10x, Ob. 10x Fig.1. Co-culture of spheroids of human dermal fibroblasts and HeLa cells. a - 1 day of cultivation. Oc. 10x, Ob. 20x; b - 4 days of cultivation. Oc. 10x, Ob. 10x HeLa Fibroblast Fig.2. Multicellular aggregate obtained after fusion of spheroids of HeLa and fibroblast spheroid. Immunoperoxidase reaction (brown staining) revealed epithelial, cytokeratin positive cells surrounding fibroblast spheroid. Oc. 10x, Ob. 10x Fig.2. Multicellular aggregate obtained after fusion of spheroids of HeLa and fibroblast spheroid. Immunoperoxidase reaction (brown staining) revealed epithelial, cytokeratin positive cells surrounding fibroblast spheroid. Oc. 10x, Ob. 10x Fig.3. Immunohistochemical revealing of epithelial antigens (cytokeratins, intermediate filaments of epithelium) in HeLa cells forming spheroid. Oc. 10x, Ob. 40x HeLa Ôèáðîáëàñòû Fig. 5. Immunofluorescent detection of fibroblast by vimentin content (green) in aggregate in co-culture of HeLa and fibroblast spheroids. Cell nuclei are stained with PrI. Oc. 10x, Ob. 40x Fig.4. Immunofluorescent revealing of vimentin (fibroblast inter- mediate filaments) in spheroid formed fibroblasts. There is a green staining of cytoplasm. Oc. 10x, Ob. 40x Fig. 5. Immunofluorescent detection of fibroblast by vimentin content (green) in aggregate in co-culture of HeLa and fibroblast spheroids. Cell nuclei are stained with PrI. Oc. 10x, Ob. 40x Fig.4. Immunofluorescent revealing of vimentin (fibroblast inter- mediate filaments) in spheroid formed fibroblasts. There is a green staining of cytoplasm. Oc. 10x, Ob. 40x SHKARINA K. A. ET AL carcinogenesis and the search for new anticancer the- rapy targets. carcinogenesis and the search for new anticancer the- rapy targets. êè. Îïèñàíà ìîäåëü áóäå êîðèñíîþ ÿê äëÿ âèâ÷åííÿ áàçîâèõ ìåõà- í³çì³â êàíöåðîãåíåçó, òàê ³ çà ïîøóêó ì³øåíåé äëÿ ïðîòèïóõëèí- íî¿ òåðàï³¿. Êëþ÷îâ³ ñëîâà: âçàºìîä³ÿ ñòðîìàëüíèõ ³ ïóõëèííèõ êë³òèí, òðèâèì³ðí³ êóëüòóðè çëîÿê³ñíèõ êë³òèí, áàãàòîêë³òèíí³ ñôåðî¿äè. Å. À. Øêàðèíà, Î. Â. ×åðåäíèê, È. À. Òèõîíêîâà, À. È. Õîðóæåíêî Ìîäåëü in vitro äëÿ èçó÷åíèÿ âçàèìîäåéñòâèÿ ñòðîìàëüíûõ êëåòîê è êëåòîê îïóõîëåâîãî ïðîèñõîæäåíèÿ Å. À. Øêàðèíà, Î. Â. ×åðåäíèê, È. À. Òèõîíêîâà, À. È. Ðåçþìå Ìåòà. Ðîçðîáèòè ìîäåëü äëÿ âèâ÷åííÿ âçàºìîä³¿ ïóõëèííèõ ³ ñòðî- ìàëüíèõ êë³òèí çà óìîâ òðèâèì³ðíî¿ êóëüòóðè. Ìåòîäè. Êóëüòè- âóâàííÿ êë³òèí ë³í³¿ HeLa ³ äåðìàëüíèõ ô³áðîáëàñò³â ëþäèíè â ìî- íîøàðîâèé ³ òðèâèì³ðí³é êóëüòóð³, ³ìóíîôëóîðåñöåíòíèé òà ³ìóíîã³ñòîõ³ì³÷íèé àíàë³ç. Ðåçóëüòàòè. Çàïðîïîíîâàíî ï³äõ³ä, ùî áàçóºòüñÿ íà äîñë³äæåíí³ áåçïîñåðåäíüî¿ âçàºìîä³¿ êë³òèí áàãà- òîêë³òèííèõ ñôåðî¿ä³â ïóõëèííèõ êë³òèí ³ ñôåðî¿ä³â ô³áðîáëà- ñò³â. Ïîäàëüøèé ³ìóíîôëóîðåñöåíòíèé àíàë³ç äàº ìîæëèâ³ñòü âèçíà÷èòè ïîõîäæåííÿ êë³òèí ó çîí³ ¿õíüîãî êîíòàêòó. Âèñíîâ- 12. Khoruzhenko A. I. Optimization of tumor cell culture conditions in soft agar for subsequent immunohistochemical analysis // Bio- polym. Cell.–2012.–28, N 4.–P. 302–305. Received 01.12.12 82 82
https://openalex.org/W4390449440	https://link.springer.com/content/pdf/10.1007/978-981-99-5873-3_7.pdf	English	null	Generic Competencies	null	2,023	cc-by	10,075	© The Author(s) 2024 F. Habbal et al., Reshaping Engineering Education, https://doi.org/10.1007/978-981-99-5873-3_7 Chapter 7 Generic Competencies We are using the term generic competen- cies to indicate the character of the competencies, as general human competencies for work collaboration and life orientation. For many years, teamwork, communication, and project management have been added to the list of core generic competencies for engineers. Furthermore, we have chosen the competency concept to indicate the potential together with the qualiﬁcation of application in practice (Fortuin et al.; Le Deist & Winterton, 2005). Policies and regulation frameworks do exist to address both generic competen- cies in the engineering standards, SDGs, and accreditation. The transformation of engineering education from a teacher-centered approach toward student-centered learning, as described in Chap. 6, is also facilitated by the development of interna- tional standards and accreditation, and the mutual recognition of engineering quali- ﬁcations and professional competencies (International Engineering Alliance, 2012). Accreditation can be seen as a transformative driver of change but, at the end of the day, it is the educational culture and the learning methodologies applied that will have a signiﬁcant impact on students’ learning. Accreditation can form the outside framework, but the inside life depends on academia and the culture. Engineering institutions have responded to this challenge with very different strategies, from adding on workshops to integrating competency development in the curriculum. However, very often generic competencies have suffered from a ‘lack of respect’ in the engineering curricula and have not been highly valued among engineering academics but have been tolerated in the curriculum as something ‘soft’ and poten- tially relevant. On the other hand, the corporate world has emphasized this as a very important component in graduate competency proﬁles, along with technical knowledge and competencies. It might be hard to solve complex problems in education, but the corporate world problems should be part of students’ learning processes to prepare them for their later careers and their understanding of the diversity of problems. In the future, we will also see more and more digital learning and blended formats for engineering students, and this will create even more possibilities for active learning methodologies, to apply blended learning modes that may allow for the facilitation of university and corporate collaboration. Teamwork, project management, communication, problem solving, etc. can all be covered under the concept of generic competencies, which are to be understood as cross-cutting competencies. Chapter 7 Generic Competencies A partner from one of the big international companies asked: how can we optimize human and social interaction? In our innovation projects, it is as though each time we start a new project, we start from scratch. Are there any tools or methods that I can use to optimize these processes? I think this is a new competition parameter. This question is becoming more and more relevant. We have learned that technologies are becoming more complex both in terms of technical combinations and in terms of the societal problems that must be solved. But what is often forgotten is that behind the technologies, there are humans. As technology becomes more complex, more humans will have to collaborate and there will be many more boundaries to cross. These include academics with different disciplinary backgrounds and users with different levels of education and social position in society. Engineering has traditionally been more product- than process-focused, which means that human interaction has received less attention and been valued less. The corporate world seeks optimization in every process to achieve more efﬁciency and better economy—and if the optimization for speciﬁc technologies might have reached its limit on the product side, the next step will be collaboration among human beings. Companies realize the need to cross boundaries between sales and innovation depart- ments, between contractors and subcontractors. In any innovation process today, the number of actors involved is increasing. Therefore, it is not enough for engineering students to learn to develop the content of technology and how technologies are combined and form systems. The students also need to learn how they, as humans and collaborators interact together in the design of systems, and here both cognitive and emotional aspects are at stake which are described in Chap. 5. As described in Chap. 5, reframing engineering education will involve rethinking the interaction between knowledge and competencies. We have several concepts for these types of competencies ranging from key skills, 21st- century skills, transversal skills, transferable skills, and generic competencies (Boelt et al., 2022; EU Commision, 2008; Kallioinen, 2010; Kearns, 2001). 139 © The Author(s) 2024 F. Habbal et al., Reshaping Engineering Education, https://doi.org/10.1007/978-981-99-5873-3_7 140 7 Generic Competencies 7 Generic Competencies Generic competencies work across different contexts, in contrast to speciﬁc ones related to a subject or a disciplinary context. 7.1 Generic and Meta-Competencies 141 Therefore, the learning of generic competencies has to be an integrated part of the curriculum and the students’ learning outcomes (Sánchez & Ruiz, 2008). But can we enhance students’ awareness of their own learning practices, and can we bring that awareness into their education in such a way that we can formalize learning and assessment processes? Of course, there are tools that can structure—there are some quick ﬁxes—but in the long run, this is not enough. It goes deeper, and the students need to learn to adapt, to participate in a complex collaboration. This cannot just be an external phenomenon; it must be integrated with the knowledge that one possesses, and it must be part of an entire culture and curriculum. Chapter 7 Generic Competencies Transferable skills, 21st-century skills, professional competencies, and transversal skills are other concepts for trying to conceptualize these competencies, which are related to codes for human interaction and behavior. The European Tuning project, which has been one of the research-based approaches for the process to a more student-centered learning model in higher education, deﬁned generic competencies as a long list of competencies ranging from critical thinking, ethics, and language to problem solving and interpersonal and organizational compe- tencies (González & Wagenaar, 2003). Scientiﬁc and technical competencies are not sufﬁcient but have to be seen in light of society, context, human relations, and ethical purpose. 7.1 Generic and Meta-Competencies 7.1 Generic and Meta-Competencies Increased complexity at both the technical and societal levels will require collabo- ration, communication, and management. But it will also require competencies to continuously develop and learn how individual and collaborative competencies can be contextualized. In one situation, it will be communication by digital means; the next situation might be 24 h of face-to-face workshop. The collaborative contexts in which an engineer will have to work will differ enormously—and it is no longer enough to have experienced teamwork or to have participated in projects. We must go above this level as the requirement today is to be able to participate in a variety of situations and to optimize the learning processes at work. It is no longer enough just to learn teamwork—the requirement is to develop teamwork skills to be efﬁcient in various situations and to be able to choose the right collaboration strategy for the speciﬁc situation—in other words, we need to move the competencies to a meta-level. Meta-competencies are deﬁned, with reference to Brown (1993, p. 32) as ‘the higher-order abilities, which have to do with being able to learn, adapt, anticipate, and create, rather than with being able to demonstrate that one has the ability to do so. Moving the competencies level to a meta-level does not only concern teamwork skills, project management and communication. A much broader concept is needed in order to understand what kind of problem we are aiming to solve and what kind of methods we have available to analyze and identify the core problem (Brown, 1993). A related concept is meta-cognition which is beyond factual, conceptual, and proce- dural knowledge and a question of strategy in combining knowledge, knowledge about cognitive tasks, and types of self-knowledge (Krathwohl, 2002). Whereas meta-competencies are about developing generic competencies, meta-cognition is the process of acquiring knowledge on how to acquire knowledge. There is a distinction between the concepts of meta-cognition, competency, and meta-competency and yet they are closely related. For example, meta-cognition strategies to acquire knowledge within thermodynamics can be developed. Compe- tencies go one step further, providing strategies to handle knowledge in action as thereby the ﬁeld of thermodynamics is appropriated to a given context, e.g., the design of heating and cooling systems in different types of buildings, climates, or 142 7 Generic Competencies 7 Generic Competencies cultures. 7.1 Generic and Meta-Competencies 7.1 Generic and Meta-Competencies 143 We cannot increase the technological complexity without addressing the human interaction. Complex challenges, such as the SDGs and the integration and develop- ment of AI and IoT systems, all involve the competencies of being able to advance learning and set future goals. Thus, strategic leadership and anticipation are two other important future competencies in complex problem approaches, as system approaches will include both strategic leadership and anticipation embedded in a system perspective (Chap. 2). There is nothing new in emphasizing generic competencies. As described earlier, this has been part of the accreditation for engineering education for a long time in terms of teamwork, collaboration, and project management. What is new is that more types of generic competencies are in play, such as entrepreneurial and digital competencies, and also that generic competencies are combined with a meta- level moving across generic competencies. Thereby, methodologies for developing lifelong competencies through the learning of meta-competencies are needed. As complexity and the need for system thinking increase, engineers not only need to learn to solve problems in the right way, but indeed they need to analyze the problems and to devise new solutions. If there are recurring ﬂooding situations and the decision is to build a new river crossing, a bridge may not be the best solution. In the ﬁrst design phase, engineers will have to analyze the weather conditions, the trafﬁc situation, the ground, possible ways of crossing the river, and many more elements, before deciding on the solution. In a complex problem analysis, engineers step backward to ﬁnd new solutions and combine knowledge and expertise in a new way. The ability to step back, at the right time and use the right lenses to get an overview of the whole system, including the underlying rules and values, becomes an important competency. The same is true for the collaboration among the involved actors and the engi- neers. If they have been used to collaborating on bridge projects, now there might be new expertise domains involved—and they not only have to analyze the knowl- edge domains and contextualize their expertise to a new innovation but also the way they collaborate has to be considered. This is a totally new element in engineering education. 7.1 Generic and Meta-Competencies Meta-competencies on the other hand include strategies to handle compe- tencies in action as competencies within a speciﬁc ﬁeld is transformed in interaction with other competencies, e.g., competencies within the ﬁeld of heating and cooling systems are combined with competency to design smart buildings or ensure business models for sustainable products. There is a dialectic relationship between competency and meta-competency and it can have blurred boundaries. Meta-competencies can be deﬁned by the development, adjustment and application of competencies and, therefore, at a higher reﬂection level (Brown, 1993). This links the meta-competencies to a notion of lifelong learning as students learn methods for how to develop their competencies (Cunningham et al., 2015). The relation between generic competencies and meta-competencies is similar blurred. The generic competencies consist of three components: understanding of the societal problem including ethics, interpersonal collaboration, and organization of the process. The meta-competencies in this domain are to combine, adjust, apply, and develop the interaction among the speciﬁc generic competencies (see Fig. 7.1). Generic competencies are types of practice competencies, and learning takes place through both reﬂection on practice and conceptualization and analysis. We have already described the importance of viewing problems in a societal context, and as the problem complexity increases, the complexity for interpersonal relations and organization will follow. The real-life problem, interpersonal collabora- tion, and organization act as a trinity that is fused together. As technology is a tightly woven system, the human side will likewise be tightly woven into the functions in the system. As the complexity increases, the interactions between knowledgeable people will increase. Fig. 7.1 Dimensions of generic and meta-competencies Fig. 7.1 Dimensions of generic and meta-competencies 7.1.1 Tacit Knowledge—Potentials and Risks Development of meta-competencies is difﬁcult, as the human interaction gets mixed in many ways with the inner world of the individual. One’s way of interacting might depend on one’s upbringing, personal identity, and personal life. But regardless of this, engineers need to learn to master collaboration, diverse contexts, and compli- cated communications and develop these capabilities as competencies to be applied in working life and life in general. Awareness and articulation of communication strate- gies and collaboration strategies can be learned. Most often, past social interactions form a body of tacit knowledge for the individual. 144 7 Generic Competencies 7 Generic Competencies Dreyfus and Dreyfus emphasize that the expert has such a rich pool of knowledge and experiences, that the intuitive processes of knowledge creation will not neces- sarily be conscious, but tacit (Dreyfus, 2004). For the expert, this might be true, while for the novice, it is important to be much more aware of the rules and the process itself—hence the need for explicit reﬂection. Tacit knowledge is the opposite of explicit knowledge and can be explained in two different ways, either as knowledge we cannot articulate or explain or as embodied knowledge that can only be expressed during practice. These two perspectives are not contradictory. Knowledge can become embodied; sometimes, it can reach a stage where it can be articulated and sometimes not. The concept derives from Polanyi, who argued that scientists should recognize that not all knowledge is propositional and in order, but that a lot of our ideas and learning comes from this messy and unordered embodied knowledge, which can hardly be communicated in words but rather in action (Polanyi & Sen, 2009). Generic and meta-competencies will remain tacit, either as unarticulated or as practice competencies, if the learning is not facilitated by reﬂection. Nonaka and Konno worked with tacit culture and knowledge in organizations and the interaction between the explicit and the tacit level (Nonaka & Konno, 1998, Engeström, 2001). Baumard took this approach to another level as he distinguished between individual and collective knowledge (Baumard, 1999). Baumard points out that there is a tacit element both for the individual and the team when there is a continuous interaction and complicated relations. Tacit and non-verbal communication might create both potentials and problems for learning from practice. 7.1.1 Tacit Knowledge—Potentials and Risks The potential with tacit knowledge is that it is a source of intuition for the individual and for the team. In group creativity research, it is a well-known phenomenon that ideas can be developed in a process of smaller interactions and iterations. One member of the team can present an idea, which will create responses from the other members, building on the idea in a continuous brainstorm of associ- ations. The process of interaction will form the team members’ culture and might very often remain tacit and form tacit patterns of interactions. The disadvantage of tacit knowledge for both the individual and the team is that it is difﬁcult to transfer or transform knowledge and competencies from one area to the next without articulation. There is a need for both the individual and the team to be able to articulate, communicate and conceptualize. Furthermore, this creates difﬁ- culties for the development of the individual and for collective competency develop- ment, especially for the individual student in articulating and conceptualizing their own competencies when ending a project. Therefore, a process of creating atten- tion and awareness by reﬂecting on these practice experiences will be an important element in learning generic competencies. In a single team consisting of four to six students, the tacit element can lead to team creativity. Schön’s reﬂective practitioners and their collaboration can be compared to a blues band doing a jam session. Each participant continues add-on to the contri- butions of the other participants and invites them to continue the development of the communication, which can be described as open, creative, uncritical, and reﬂective in relation to the theme and creative. Collaborative creativity and development might 7.1 Generic and Meta-Competencies 145 7.1 occur. In online teams, tacit knowledge also exists, although it is more difﬁcult to create a common tacit culture as there are limitations when communicating through a screen (Sawyer, 2005). occur. In online teams, tacit knowledge also exists, although it is more difﬁcult to create a common tacit culture as there are limitations when communicating through a screen (Sawyer, 2005). However, the potential of tacit knowledge related to the team interactions often vanishes when there is a change in team members, team size, the length or credits of the projects, physical versus digital space, problem types, or diversity pattern. 7.1.1 Tacit Knowledge—Potentials and Risks Then it becomes difﬁcult to apply the knowledge and competencies the learner has obtained in a new situation. As an example, students might articulate their reﬂections on teamwork expe- riences continually and state action points for change in a log-book format. To work collectively, teamwork experiences from the individuals must be articulated to align and negotiate understandings and perspectives for change. Furthermore, students must realize that if considerable changes in the team constellation happen, teamwork competencies have to be transformed. For example, a predescribed team culture might be beneﬁcial in one team, while counteractive in another team setting. Likewise, project management systems, approaches to problem solving, etc. will change with the type of problem, the intended learning outcomes, and the actual team constellation. In other words, there is a need for meta-teamwork competen- cies to provide strategies for the interaction of diverse teamwork competencies, such as inclusiveness, collaboration, communication, project management, problem analysis, etc. 7.1.2 Reﬂection and Meta-Reﬂection If practice experiences are not reﬂected, these will become trial and error, which can be beneﬁcial but remain tacit knowledge. Reﬂection on practice learning from the education is not only crucial for the competency development of young engineers; it also inﬂuences how ready they feel for employment and their lifelong qualiﬁcation and career strategies. Reﬂection is essential for progressing learning of generic competencies. Regard- less of the domain, the learning is framed by the learning methods and learning environments, which will create opportunities for students to experience collabo- ration among peers, knowledge management, creativity, and innovation. It can be argued that there is nothing new in reﬂecting in and on practice to attract attention to tacit knowledge. But it is new to think of the relation between generic and meta- competencies and that the learning of meta-competencies is based on a combination of reﬂection on practice and theory. Back in the 90s, the learning of the Alverno College culture was known far beyond the US borders (Gibbs, 1999) for their integration of reﬂection as a cultural factor. Later came the Olin College culture, where ongoing reﬂections in teams and for individuals after class were also an impressive contribution to the learning culture and the education of the individual. Students need to learn to reﬂect and articulate 146 7 Generic Competencies 7 Generic Competencies their experiences from collaboration, managing projects, managing cultures, problem analysis methods, collaboration with external stakeholders, presentations, etc. Kolb (1984) does not identify reﬂection as a method but as an element in a learning process consisting of concrete experience, reﬂective observation, abstract conceptu- alization, and active experimentation. This underlines reﬂection as a key to combine practice (active experimentation) and theory (abstract conceptualization) and can be the on-reﬂection which is looking backward as in Schön’s conceptions on reﬂections (Kolb, 1984). Without abstract conceptualization, reﬂection-in-action can quickly become a tacit process mostly characterized by trial and error. What works and what does not work are concluded based on immediate responses in the process (see Fig. 7.2). When reﬂection is brought into engineering education, engineering students may not be attracted by reﬂection, but rather much more by experimentation—they want to experiment; they want to design and build things. 7.1.2 Reﬂection and Meta-Reﬂection Therefore, it might be an advan- tage to ask students to set up experiments in their learning process and collabora- tive processes, and then ask them to reﬂect on these experiments (Kofoed et al., 2003). Especially in a team, discussions on how to set up new plans for knowl- edge management might also create awareness of the variety of possibilities that the students will face. For example, an experiment could be to ensure knowledge manage- ment by setting up team seminars, organized with presentations of the knowledge acquired from subgroups, opponents to discuss the application of knowledge, and inputs for further knowledge acquisition. If the action and future orientation drive the process, they increase the motivation to learn from the experiences gained from their experiments. Reﬂection on practice might remain at a lower taxonomy level for learning if it is not combined and integrated with more general theories. For example, students need to be able to create attention to, and awareness of, their collaboration patterns and EExperiences Reflective Observation Practice dimension Trial and error Metacognitive Dimension Theory and Reflection on variation Abstract Conceptualisation Active Experimentation Fig. 7.2 Reﬂection relating practice and theory based on Kolb (1984) EExperiences Practice dimension Trial and error Active Experimentation Metacognitive Dimension Fig. 7.2 Reﬂection relating practice and theory based on Kolb (1984) 7.1 Generic and Meta-Competencies 147 7.1 their use of project management methods. That means that the student will become much better at reﬂecting on how well they are interacting in speciﬁc collaboration. Reﬂection is an inner inductive process, which can be facilitated by oneself, by peers in a team, by comparing to previous similar experiences or facilitated by academic staff. Attention, awareness, and articulation enable understanding of existing practice but do not necessarily provide ideas for new innovative ways of collaborating. If reﬂection is not combined with theories and concepts, reﬂection might not contribute to learning. For example, if students state a team culture without considering theories of effective teamwork or fail to align their understandings of the team culture concept, the statement might be narrow-sided and open for diffuse interpretations. p We have to be careful that we are not teaching our students just to reﬂect on practice, but that they have to move around in the Kolbian learning circle and integrate the reﬂection of practice experiences into the learning and understanding of theories (see Fig. 7.2). 7.1.2 Reﬂection and Meta-Reﬂection Reﬂection on comparing several practice experiences will form the basis for understanding the variation and possible methods of project management or collaboration. Engineering students should not only be able to apply and reﬂect on theory or methods in order to make constant improvements. They should also be able to ques- tion whether these are the right theories and methods or if other solutions are needed in relation to the given challenges. In integrating complex problems in engineering education, students will need to integrate values, analyze contexts, and question established norms and institutions. As an example, energy systems are highly depen- dent on the political climate, social movements and strong institutional dependencies. Complex problem solving is part of the new core in engineering to be able to deal with the societal as well as the human challenges and contribute to strategic leadership. Thereby, there is a need to engage in meta-reﬂection as an integrated part of meta-cognition. Basically, meta-reﬂection happens when we reﬂect on how we reﬂect. Meta-reﬂection involves cross-cutting reﬂection on the appropriateness of the different interactions between theory and practice. Integration of meta-reﬂection in education is important, as it is not expected that students will develop these competen- cies by themselves. Rather, through guided reﬂections on varieties of learning experi- ences, students can gain a deeper insight into their problem, project, and collaborative and learning skills. In a learning context, we characterize meta-reﬂection as a comprehensive reﬂec- tion including different levels of reﬂection, see Chap. 6. In alignment with Argyris and Schön (1997) and their concepts of single- and double-loop learning, as well as the concept of triple-loop learning introduced by Tosey et al. (2012), we characterize three levels of reﬂection: 1. Single-loop reﬂection—that is reﬂection on activities (are we doing things in the right way). 1. Single-loop reﬂection—that is reﬂection on activities (are we doing things in the right way). 2. Double-loop reﬂection—that is reﬂection on the governing variables (are we doing the right things). 2. Double-loop reﬂection—that is reﬂection on the governing variables (are we doing the right things). 7 Generic Competencies 148 Table 7.1 Variation in reﬂection Type of reﬂection Example Single-loop reﬂection Do I collaborate in the right way? Does the collaboration proceed as planned? Do I use the right methods to follow expected collaboration patterns? Double-loop reﬂection Do I choose the right way to collaborate? 7.1.2 Reﬂection and Meta-Reﬂection What possible collaboration strategies should I apply in this situation? Is there a need for developing new collaboration patterns—and what new ideas can I contribute with? Triple-loop reﬂection Which values are reﬂected in the way we collaborate? Are these values what we want? Which politics should we develop to aligned collaboration strategies with our values? Table 7.1 Variation in reﬂection 3. Triple-loop reﬂection—that is reﬂection on the underlying assumptions leading to the governing variables (how do we consider what is right). 3. Triple-loop reﬂection—that is reﬂection on the underlying assumptions leading to the governing variables (how do we consider what is right). Meta-reﬂection thereby includes reﬂection considering whether we are doing things right, whether we are doing the right things and more fundamentally, what we consider as being ‘right’ (see Table 7.1). For example, in designing human collab- oration, we often forget to ask ourselves what possible collaboration strategies we can apply, and we forget to ask why we chose a speciﬁc one. We often forget to step backward and analyze the tasks ahead of us and form adequate organizations, and instead, we jump onto known pathways. Furthermore, the complexity of doing things right, doing the right things and considering what is ‘right’ increases considerably when several interests are involved. Negotiations between actors are a well-known part of complex problem solving. A simple problem is solvable within the disciplines; a complicated problem connects to known collaboration among disciplines and subdisciplines, whereas a complex problem does not have a known solution and learners will have to step backward to understand the problem and to design solutions across traditional disciplinary bound- aries. Thereby, meta-reﬂection also becomes a matter of reﬂecting on the boundary crossing between different disciplines in interdisciplinary projects. It becomes a meta-competency to handle interdisciplinary competencies in action, e.g., to foresee the limits of one’s own discipline, knowing who to consult to interact with other disciplines, and ensure aligned interaction with mutual beneﬁts. 7.2 Interdisciplinarity and Boundary Work Interdisciplinarity is not a subject matter, it is a process that ends up building a format of thinking. Each student has their own contribution to understanding and solving a speciﬁc problem. When students from diverse disciplines work together, they build a level of trust, and they may start consulting each other or work together outside Interdisciplinarity and Boundary Work 149 7.2 the classroom on different subjects. Interdisciplinary collaboration among students in the same course, brings together new ways of thinking, combined elements of solutions and more global ﬁt of an outcome (Mausoom & Vengadeshwaran, 2021). What the dimension of tacit knowledge reminds us of is that it is important to reﬂect not only at the individual level, but also among peers at the team level, and not least be aware of the diversity issues. Teams consist of individuals, and it can be very hard to look through what is happening in the team among the team members, but it is essential that the team members understand both their own and each other’s perspectives. What can be beneﬁcial to identify are many of the disciplinary and diversity factors that create boundaries. When boundaries are identiﬁed, it is much easier to establish common ground. It is important to create a language and set of concepts in order to set common goals and be able to reﬂect on the process and the outcomes, and as mentioned in Chap. 5, language and linguistics acts include a multitude of symbols which is open for interpretation. That might be easier to say than do. But a language is part of the organization of the process and the application of structural competencies. What is much more complicated is overcoming the boundaries of disciplines and cultures. The degree of interdisciplinarity is linked to the type of problem that the students are working on. As the problem becomes more complex, it will involve more actors and disciplines in both the identiﬁcation and solving phases. Most of the research on interdisciplinarity is primarily focused on research, and the literature on how to deal with interdisciplinarity in education or in collaboration is limited (Everett, 2016). For research, the literature mentions three variations of interdisciplinarity: multi, inter, and transdisciplinarity (Keestra & Menken, 2016; Repko et al., 2019). As illustrated in Fig. 7.3, a multidisciplinary approach ensures that the problems are looked at from different disciplinary angles and different discipline solutions are provided. 7.2 Interdisciplinarity and Boundary Work There is an exchange of information and knowledge, but there is no real integration in the product. The interdisciplinary approach is an integrated approach and there will be a common solution in the end. The transdisciplinary approach is deﬁned a bit differently in diverse literature, but there is general acceptance that boundaries of academia and the non-academic sector are crossed, and new knowledge will emerge. It is also a process involving new perspectives from the outside that will question own disciplinary origins and perspectives. There are other conceptualizations of the variation in the interdisciplinary approach, e.g., Klein deﬁnes a narrow and a broad interdisciplinary approach, where the narrow is characterized by a shared knowledge paradigm, while the broad is characterized by different knowledge paradigms, e.g., engineering versus humani- ties (Klein, 2006, 2010). However, even if it might be possible to distinguish between Integrated output TRANS Triple-helix integration Discipline A&B Public partner Private partner Fig. 7.3 Transdisciplinarity (based on Keestra & Menken, 2016) Fig. 7.3 Transdisciplinarity (based on Keestra & Menken, 2016) 150 7 Generic Competencies 7 Generic Competencies multi, inter, and transdisciplinarity, in practice the concepts are used in abundance, and therefore, it is an advantage to regard interdisciplinarity as an overall concept embracing a scale from multi, narrow inter and broad inter to transdisciplinary approaches. We have to be careful with the narrow disciplinary approach. Although engineers from, for example, electronics and mechanical engineering, might ﬁnd it hard to work together in systems, they do share scientiﬁc and engineering practices and cross-cutting concepts. Scientiﬁc and engineering practices include, for example, deﬁning problems, developing, and applying models, investigations, mathematical and computational thinking, etc., and there are cross-cutting examples like cause and effect, scales, systems, and system models, please see more in Chap. 2. There are also disciplinary-speciﬁc areas within the physical sciences, life sciences, earth and space sciences and engineering technologies (Council, 2012). The narrower interdisciplinary collaboration can be related to working on an innovation system, but we have to be careful that the technological systems are based on human and societal needs, which will involve a much broader interdisciplinary collaboration. Interdisciplinary educational models will apply more attention to cross-cutting generic and meta-competencies to bridge the different disciplines. The generic competencies can be used across domains and disciplines, but these have to be combined with meta-competencies to capture the variation in the disciplinary approaches. 7.2 Interdisciplinarity and Boundary Work There is a request, in particular, for interdisciplinary teamwork compe- tencies in various types of projects, and as an example, learning generic competencies in an interdisciplinary team of law and engineering could also be applied in groups of social sciences, humanities, and engineering (Male, 2010; Male et al., 2011). In education, there is a need for more attention to authentic problems to ensure that students learn methods for how to deal with complex, real-world problems, such as sustainability problems. Repko, Szostak, and Buchberger focus on interdisciplinary studies and emphasize that there are a series of characteristics or skills that we need to apply in interdisciplinary collaboration, such as an entrepreneurial mind (taking risks), a love of learning (excited to learn something new), self-reﬂection (self-awareness of strengths and weaknesses), intellectual courage (acceptance of, and respect for, other viewpoints), and patience and empathy (active listening) (Repko et al., 2019). All these characteristics are an extension of teamwork competencies in projects, but even more advanced, and they involve deep reﬂections and project skills as an extended part of the generic competencies. The presence of these characteristics differs in different contexts. One moment, it is listening, and the next moment, it is having the courage to move across boundaries and take risks. The really difﬁcult element is learning when we are doing the right thing, taking the context into account. Maybe it is not so difﬁcult to learn to listen and to learn to act, but the hard part is decoding the situation and applying appropriate skills in a given moment. This is the art that experts can carry out, but which novices will have to learn in a more structured way (Dreyfus, 2004). Similar abilities can be applied to intercultural collaboration as the individual will have to step outside their comfort zone to be able to understand another perspective like understanding variation of perceptions. Or it can be graduates going into work Transfer, Transformation, and Boundary Work 151 7.3 where they have boundaries to cross as they will meet new work cultures. Interdis- ciplinary collaboration is linked to academia, whereas the concept of boundary is a much broader concept and can be seen in relation to complexity, which is a philo- sophical concept, and to systems engineering, which is much more of an engineering and production approach. 7.2 Interdisciplinarity and Boundary Work Regardless of the approach, there are knowledge domains and communities, organizations, and cultures, which are to work on common goals. Many scholars use the concept of boundaries to describe an increasingly heterogenic society, which has increased its focus on developing expertise (Akkerman & Bakker, 2011). ) Boundaries of domains constitute what is regarded as expertise and what is not, as Lave and Wenger clearly describe in their concept of legitimate peripheral partic- ipation in communities (Lave & Wenger, 1991; Wenger, 1999). The technological development creates the need for more specializations, and thus, there will be an increased number of smaller expert communities but still with a need to reach out to other expert communities. For all aspects, being able to work on boundaries seems to be a common compe- tency—no matter whether we are talking about disciplines or cultures. Although it can be argued that boundaries will always represent analytical discourses for humans to be able to talk about, to negotiate and to create identities, boundaries will also be a connecting point. Boundaries do not mean that there is a strict black/white border, but that there is a sliding transition from one site to another or as a shared space (Leigh Star, 2010). Ecotones as a concept from biology could supplement the under- standing of boundaries as an area with a mixed and merged zone in between two different domains. It could be the zone between a wood and a farmer’s ﬁeld, where the natural law for trees and woods is to spread the seeds to grow, while the grass ﬁeld creates a counterpart by wanting to enlarge (Ryberg et al., 2021). Boundaries do not necessarily cause ﬁghts, but there might be tensions between different ways of understanding and contextualizing the same concept or action. 7.3 Transfer, Transformation, and Boundary Work Boundary crossing, generic competencies, and interdisciplinary learning relate to the concept of transfer and/or transformation. Transfer is a complicated concept. The concept has multiple meanings, such as transfer from education to work or as a concept for learning. In many learning theories, learning transfer is a concept or metaphor meeting a lot of criticism from different angles. The concept signals that once things are learned, they can be transferred to other situations as replications. But if we only replicate, there will be no progress, so an understanding of transfer as replication is totally out of the question. As introduced in Chap. 5, the social-constructivist theories emphasize that artifacts are created through the social interactions in a team. Each situation or context will be different depending on the individuals and the interactions in the group. Students can bring earlier experiences with them into a new situation, but they will never 152 7 Generic Competencies 7 Generic Competencies be able to replicate their learning in a new group; such a replication would have no meaning. But they can apply elements of their past experiences and knowledge and combine the learning elements to expand their learning and interaction with other group members. They can adjust and situate their knowledge and experiences together with the other members and learn how to apply their combined efforts in these new situations. Compared to the understanding of replication, this is a signiﬁcantly different approach as it is not enough to be aware of one’s own competencies; it is also necessary to analyze and understand new situations. What are the purposes? What could be a beneﬁcial organization? Who are the other group members? What are their expectations? Which of their experiences can the group beneﬁt from? Hager and Hodkinson (2009) argue against using the concept of ‘“learning trans- fer” and think instead of learning as becoming within a transitional process of boundary crossing’ (p. 635) (Hager & Hodkinson, 2009). They argue that the concept of transfer itself signals a narrow and instrumental way of approaching learning, as has already been argued. But they also argue that the concept could misleadingly emphasize academic and educational knowledge in the transition from education to work without any considerations of the culture, interactions, organization, tasks, or visions applicable for work. 7.3 Transfer, Transformation, and Boundary Work Meta-reﬂection and meta-competencies are necessary for the progression of learning and so that learners can apply knowledge from one area to the next. Learners can transfer some generic skills, e.g., how to handle phases in a project management process. This is a type of declarative knowledge. But each time learners are in a new situation; they will have to create a transformation process by appropriately adjusting experiences and knowledge to be recontextualized. They must learn to read the new project according to the new type of problem, the length of the project, and the composition of the team to go into a transformation process (see Fig. 7.4). The new team might be interdisciplinary or disciplinary, the collaboration with external part- ners might be new. Therefore, the way the students have learned to collaborate in project A will have to be reconstructed and transformed in the new project B. For students to be able to transform their experiences into a new context, they need to learn to analyze the problem and the new situation. Reﬂection on previous experiences might not be enough as this very much concerns questions like: Did I collaborate right, or did I choose the right way to collaborate? What is needed to come from A to B and what have I learned? What possible collaboration strategies do I have? What possible methods do I know? To get to this level, there is a need not only to compare experiences from practice but also to compare and analyze experiences in relation to the theories. Engeström points out that transfer and transformation take place from one activity system to another where transformation of meanings and activities takes place. The degree of variation and difference of these contexts or problems will inﬂuence the boundary crossing process and which competencies will be needed. Dohn et al. (2020) emphasize that from an activity system perspective, the goal of education (and of learning) is to facilitate students’ capacities for transfer and transformation to support ill-structured and complex problems (Dohn et al., 2020). 7.3 Transfer, Transformation, and Boundary Work 7.3 Transfer, Transformation, and Boundary Work 153 Project A • One team of 4-6 students • Working on narrow discipline problem • For ½ semester • Only internal stakeholders Project B • Two teams working together • Working on complex SDG problem • For a whole semester • External stakeholders transfer transformation Project A Project B transfer transformation • For ½ semester • Only internal stakeholders Fig. 7.4 Transfer and transformation Fig. 7.4 Transfer and transformation When learners reﬂect on their experiences and try to articulate these, they will never know when and where they will need these experiences again. Many different conditions will inﬂuence the need for prior experiences in new contexts: the need, the task, and the transition from idea to practice. Carlile’s work on boundaries in product development argues for three different ways of crossing boundaries: syntactic, semantic, and pragmatic, see Fig. 7.5 (Carlile, 2004). The transfer level concerns the transfer of known and factual knowledge. When the problem and contexts are known, it will be relevant to apply transfer as a concept to understand the learning. It is types of declarative knowledge that can be memorized. For generic skills, it can be phases in speciﬁc management systems. The translation level concerns translation between relatively new situations but still with recognizable elements for the problem and the context. Here, there is a focus on the language and understanding of the different team members. It makes a lot of sense to bring the translation level to generic competencies as students have to learn how to create dialogues of understanding instead of cheating oneself and each other by pretending they know. The transformation level is a kind of pragmatic boundary crossing and concerns unknown problems and contexts in which knowledge is going to be developed. The concept of transformation indicates that it is not just to replicate existing knowledge but to adjust and understand how knowledge and experiences can be applied. Hager and Hodkinson (2009) are pointing out that transformation is a comprehensive process that includes not only awareness of what competencies the learning is bringing to a new situation but indeed the ability to understand new situations. This understanding of transformation aligns with the understanding of meta-competencies. 154 7 Generic Competencies TRANSFER of existing knowledge and procedures TRANSLATION of current understandings and practices TRANSFORMATION to new understandings, practices, procedures and knowledge structures ACTOR/DISCIPLINE A ACTOR/DISCIPLINE B Fig. 7.3 Transfer, Transformation, and Boundary Work 7.5 Integrated/3-T framework for managing knowledge across boundaries, based on Carlile (2004) TRANSFORMATION to new understandings, practices, procedures and knowledge structures TRANSLATION of current understandings and practices ACTOR/DISCIPLINE A Fig. 7.5 Integrated/3-T framework for managing knowledge across boundaries, based on Carlil (2004) Carlile also reminds us that boundaries are diverse and the competencies to work on boundaries will vary accordingly. With increased novelty and innovation, there will also be a need for increased interdisciplinary collaboration and a need not only learning to transfer, but to situate, construct, and innovate new technologies as well as human collaboration. 7.3.1 Boundary Objects and Brokers Wenger deﬁned boundary objects and boundary brokers (Wenger, 1999). A boundary object is the reiﬁcation, the physical expression of common goals bridging diverse communities, and Leigh Star (2010) emphasizes that boundary objects can be charac- terized as being a material/organizational structure with scalability as a function that allows people from different communities to work together without it being neces- sary to have a consensus (Leigh Star, 2010). However, the work across communities must have a purpose or a problem as starting point; otherwise, it would make no sense to work together. Star also describes boundary objects as work arrangements that are at once material and processual, e.g., project management systems. Brokers are the humans involved from various communities who are working on a common goal and making use of the boundary objects to communicate, see Fig. 7.6. But the function of brokers is to build relationships, facilitate knowledge sharing and progress in collaboration, and to bridge and link the communities (Long et al., 2013; Neal et al., 2021). 7.3 Transfer, Transformation, and Boundary Work 155 7.3 Transfer, Transformation, and Boundary Work Discipline A Characterised by patterns of Participation (living, acting, interacting) and related Reifications (focus points, documents, artefacts) Discipline B Characterised by patterns of Participation (living, acting, interacting) and related Reifications (focus points, documents, artefacts) BROKERING Negotiation of participation and boundary objects Fig. 7.6 Brokers and boundary objects in a community of practice (based on Wenger, 1999) BROKERING Negotiation of participation and boundary objects Fig. 7.6 Brokers and boundary objects in a community of practice (based on Wenger, 1999) For example, for interdisciplinary teams, it will mean that the problem and project serve as a boundary object and the team members might need a period at the beginning where they create a common understanding of each other’s perspectives. However, it is also the boundary object that will require negotiation among the team members. The negotiation concerns both the scientiﬁc approach and the structure of the process and the interpersonal aspects. For the brokers or the humans working on the boundaries, it is essential to be able to understand diverse perspectives. No matter whether the boundary is primary cultural or disciplinary, the openness and willingness to try to understand an opposite point of view will be essential. 7.3.1 Boundary Objects and Brokers In this respect, meta-reﬂection is essential as it is not only a question of translation; it is a question of understanding other contexts to be able to grasp the meaning in a conversation. This might not be an easy process, but students need to be exposed to the issues that they will most likely face later in their professional life and learn how to overcome diversity issues in the teams. Maybe the conﬂicts among team members are to be understood as disagreement in the problem-solving approach, but the individual learner might understand this as a personal conﬂict. If the latter is the situation, this learner will bring along a self-understanding of personal conﬂict strategies that might not be beneﬁcial for scientiﬁc dispute. Therefore, reﬂection on, and articulation of, the experiences of both individuals and peers is required for progressing the development of generic competencies. Learning various strategies for negotiating and coping with disagreements is essential. Maybe we have to rethink the competencies the students need to learn along these lines to work as a negotiator and broker. For many years, we have talked about competencies for teamwork and collaboration, but when focusing on the process of becoming a negotiator, it becomes clearer that educators have to facilitate boundary work. Thereby, teachers can strengthen students’ abilities to use and create boundary objects and work as brokers to connect to core stakeholders. 7 Generic Competencies 156 7.4 Creating Learning Trajectories as a Lifelong Learning Strategy Lifelong learning has been on the agenda for the last 30–40 years. In Europe, ﬁrst it was a question of getting the formal education system to offer professional master educations part-time for employees. During the last ten years, this approach has to some degree changed to offering micro-credentials, which employees or learners can apply in different ways for their own competency development. Therefore, the focus has changed from getting institutions to offer educations to adults, to letting the responsibility for lifelong learning be an individual matter. Educational institu- tions then support individual learning pathways by making minor educational course credentials available. Being able to create and handle individual learning trajectories can be seen as a new concept of lifelong learning. Learning trajectories are a much broader concept that is based on the concept of personal learning and have individual ﬂexibilities in the creation of one’s own competency development. Learning takes place in many situations and the individual learner must be able to advance learning based on work in different networks and groups in both formal and informal settings. The individual must be able to develop their own professional and organizational competencies, both to assimilate knowledge to existing frameworks and to accumulate, transfer or trans- form learning from one context to another, and from one conceptual understanding to a new one. Accumulating, transferring, and transforming knowledge and practices are also about being able to choose strategies, methods, and techniques for speciﬁc situations. Therefore, the awareness of learning and of the progression and combination of generic competencies—both individually and in teams—will be a core in future engineering competencies, and it should be addressed in education. Besides, this is also what companies are asking for. Both the variation theory and Vygotsky’s zone of proximal development theory point to the fact that the learner remains within prior developed schema’s for learning, if there is nothing new in the learning situations (Vygotsky, 1978). There needs to be a balance between what is known and the challenge of the unknown. If the learner only meets totally new unknown challenges and has no experiences, it will be too difﬁcult to solve the task and then frustration arises, and motivation might decrease proportionally. It would be like asking an English language student to develop a piece of software or write an essay in French. 7.3.2 From Management to Leadership In the same vein, learning project management in engineering education might not be enough, and it might be that we should move the bar to leadership. The principle of reformulating generic competencies to include meta-competencies also counts for leadership. During the last 20 years, there has been a trend of including both project management and leadership in the list of competency requirements (Boelt et al., 2022). Especially in light of the requirement of new competencies along with the increasing technological and societal complexity, engineers will face the need for more future-oriented and strategic thinking (see Chap. 2). In the UNESCO report on SDGs in education, seven more general compe- tencies are identiﬁed: systems thinking, anticipation, normative competencies, strategic competencies, collaboration, critical thinking, self-awareness, and inte- grated problem solving (UNESCO, 2017). The last four of these competencies have been highlighted for a long time and are all part of the formal accreditation systems in many countries. Also, systems thinking is mentioned in some of the accredita- tion criteria; however, anticipation as well as normative and strategic competencies are relatively new. For example, competencies of forecast and scenario building are often applied, e.g., in the environmental and sustainability ﬁelds. Scenarios for climate change are based on projection patterns relying on chosen variables and their relations. Such general and generic competencies are also part of leadership and impor- tant for creating visions and strategic goals. Compared to management, leadership competencies are the competencies to align the organization, to set direction and to motivate employees, whereas management is the competency to plan, set up budgets, formulate subgoals, keep deadlines, organize the process and staff, and control the progression. The management part can be carried out primarily by competencies at the generic level, whereas leadership will primarily require meta-competencies in pointing out possibilities. Leadership and project management are far more comprehensive than described here, but the point is that engineering students do need to experience these types of competencies in education in order to be prepared for work. The scaling of projects from discipline projects to interdisciplinary projects in the curriculum (see Chap. 6) will allow students to experience the variation in project processes and project management. A single project in a course will not develop leadership competencies as this really requires a complex situation involving several disciplines and possible stakeholders. 7.4 Creating Learning Trajectories as a Lifelong Learning Strategy 157 7.4 Creating Learning Trajectories as a Lifelong Learning Strategy Both tasks would present totally unknown languages to the student, although there might be more transferable knowledge in the French language case than in the software case. The same applies to the transfer and transformation of generic competencies. If an engineering student has never applied or gained knowledge of methods for user involvement, there is nothing they can transform to a new situation. But as soon as the learner has had their ﬁrst experience with how to interview actors and methods of user involvement, there is a potential for developing these competencies 158 7 Generic Competencies 7 Generic Competencies by transformation from one project to another. Similarly, there is potential for the development of, for example, team skills and project management skills. by transformation from one project to another. Similarly, there is potential for the development of, for example, team skills and project management skills. However, it is not only a matter of ensuring progression of practice experiences. The new trend of offering digital micro-credentials is in line with the digitalization agenda and the notion that university degree programs should become more ﬂexible. Perhaps in this policy shift, the learner has been forgotten. Perhaps there has been too much focus on developing knowledge resources without considering their integration in various learning paths. It is however the individual learner, who participates in different communities or projects, who are to select and combine available micro- credentials to create his or her own learning trajectory. Also, in cases in companies where engineers participate in project after project, there is a requirement to create progress in capacity building. But if the learner or the team has had a tacit or non-articulated collaboration, how can the individual then develop his or her understanding and competencies based on this collaboration? It is a core aspect of learning to create individual transfer or transformation of individual competencies achieved in a team by understanding the task, understanding the individuals and their competencies, having the knowledge of how to design work processes aligned to the task, and being able to reﬂect and negotiate during the process. References Akkerman, S. F., & Bakker, A. (2011). Crossing boundaries between school and work during apprenticeships. Vocations and Learning, 5, 153–173. https://doi.org/10.1007/s12186-011- 9073-6 Argyris, C., & Schön, D. A. (1997). Organizational learning: A theory of action perspective. Reis, (77/78), 345–348. Baumard, P. (1999). Tacit knowledge in organizations. Sage. Boelt, A. M., Kolmos, A., & Holgaard, J. E. (2022). Literature review of students’ perceptions of generic competence development in problem-based learning in engineering education. European Journal of Engineering Education, 1–22. Brown, R. B. (1993). Meta-competence: A recipe for reframing the competence debate. Personne Review. Carlile, P. R. (2004). Transferring, translating, and transforming: An integrative framework for managing knowledge across boundaries. Organization Science, 15(5), 555–568. Council, N. R. (2012). A framework for K–12 science education: Practices, crosscutting concepts, and core ideas. National Academies Press. Cunningham, P., Matusovich, H. M., Hunter, D. A. N., & McCord, R. E. (2015). Teaching metacogni- tion: Helping engineering students take ownership of their own learning. Frontiers in Education Conference (Fie), 367–371. <Go to ISI>://WOS:000371705200065 f Dohn, N. B., Markauskaite, L., & Hachmann, R. (2020). Enhancing knowledge transfer. In Handbook of research in educational communications and technology (pp. 73–96). Springer. Dreyfus, S. E. (2004). The ﬁve-stage model of adult skill acquisition. Bulletin of Science Technology & Society, 24(3), 177–181. Engeström, Y. (2001). Making expansive decisions: An activity-theoretical study of practitioners building collaborative medical care for children. In Decision making: Social and creative dimensions (pp. 281–301). Springer. 159 References EU Commision. (2008). Improving competences for the 21st century: An agenda for European cooperation on schools. Brussels: Commision of the European Communities Retrieved from http://eur-lex.europa.eu/legal-content/EN/TXT/?uri=celex:52008DC0425 Everett, M. C. (2016). Interdisciplinary studies: A site for bridging the skills divide. Journal o Effective Teaching, 16(2), 20–31. Fortuin, K., Uiterweer, N. P., Gulikers, J., Oonk, C., & Tho, C. Training students to cross bound- aries between disciplines, cultures, and between university and society: Developing a boundary crossing learning trajectory. Gibbs, G. (1999). Using assessment strategically to change the way students learn. In S. A. Brown & A. Glasner (Eds.), Assessment matters in higher education: Choosing and using diverse approaches. McGraw-Hill International. González, J., & Wagenaar, R. (2003). Tuning Educational structures in Europe—A pilot project by and for higher education institutions supported by the European Commission in the framework of the Socrates programme. f p g Hager, P., & Hodkinson, P. (2009). Moving beyond the metaphor of transfer of learning. References British Educational Research Journal, 35(4), 619–638. International Engineering Alliance. (2012). The Sydney Accord. http://www.washingtonaccord.org/ sydney/ Kallioinen, O. (2010). Deﬁning and comparing generic competences in higher education. European Educational Research Journal, 9(1), 56–68. Kearns, P. (2001). Generic skills for the new economy: A review of research relating to generi skills. Adelaide: National Center for Vocational Education Research. Keestra, M., & Menken, S. (2016). An introduction to interdisciplinary research: Theory and practice. Amsterdam University Press. Klein, J. T. (2006). A platform for a shared discourse of interdisciplinary education. JSSE-Journa of Social Science Education, 5(4). Klein, J. T. (2010). A taxonomy of interdisciplinarity. The Oxford Handbook of Interdisciplinarity, 15, 15–30. Kofoed, L. B., Kolmos, A., Kofoed, L. B., & Kolmos, A. (2003). Development og Process Competen- cies by Reﬂection, Experiment and Creativity. In Teaching and Learning in Higher Education: New Trends and Innovations (p. 12), University of Aveiro, Portugal. Kolb, D. (1984). Experiential learning: Experience as the source of learning and development ( edn.). Prentice Hall. Krathwohl, D. R. (2002). A revision of Bloom’s taxonomy: An overview. Theory into Practice 41(4), 212–218. Lave, J., & Wenger, E. (1991). Situated learning: Legitimate peripheral participation. Cambridge University Press. Le Deist, F. D., & Winterton, J. (2005). What is competence? Human Resource Developmen International, 8(1), 27–46. Leigh Star, S. (2010). This is not a boundary object: Reﬂections on the origin of a concept. Science Technology, & Human Values, 35(5), 601–617. Long, J. C., Cunningham, F. C., & Braithwaite, J. (2013). Bridges, brokers and boundary spanner in collaborative networks: A systematic review. BMC Health Services Research, 13(1), 1–13. Male, S. A. (2010). Generic engineering competencies: A review and modelling approach Education Research and Perspectives, 37(1), 25. Male, S. A., Bush, M. B., & Chapman, E. S. (2011). Understanding generic engineering competencies. Australasian Journal of Engineering Education, 17(3), 147–156. Mausoom, M. M., & Vengadeshwaran, V. (2021). A novel and holistic approach for the estimation of the attainment of learning outcomes for an effective OBE implementation. International Journal of Social Research & Innovation, 5(2), 17–35. f , ( ), Neal, J. W., Neal, Z. P., & Brutzman, B. (2021). Deﬁning brokers, intermediaries, and boundary spanners: A systematic review. Evidence & Policy: A Journal of Research, Debate and Practice. 160 7 Generic Competencies 7 Generic Competencies Nonaka, I., & Konno, N. (1998). The concept of “Ba”: Building a foundation for knowledge creation California Management Review, 40(3), 40–54. References Polanyi, M., & Sen, A. (2009). The tacit dimension. University of Chicago press. Repko, A. F., Szostak, R., & Buchberger, M. P. (2019). Introduction to interdisciplinary studies. Sage Publications. Ryberg, T., Davidsen, J., Bernhard, J., & Larsen, M. C. (2021). Ecotones: A conceptual contribution to post-digital thinking. Postdigital Science and Education, 3(2), 407–424. Sánchez, A. V., & Ruiz, M. P. (2008). Competence-based learning. A proposal for the assessment of generic competences, University of Deusto. g p y Sawyer, R. K. (2005). Social emergence: Societies as complex systems. Cambridge University Press Sawyer, R. K. (2005). Social emergence: Societies as complex systems. Cambridge University Press. Tosey, P., Visser, M., & Saunders, M. N. (2012). The origins and conceptualizations of ‘triple-loop’ learning: A critical review. Management Learning, 43(3), 291–307. UNESCO. (2017). Sustainable Development Goals. http://en.unesco.org/sdgs Vygotsky, L. (1978). Interaction between learning and development (M. Cole, Trans.). In Mind in Society (pp. 79–91). Harvard University Press. Wenger, E. (1999). Communities of practice: Learning, meaning, and identity. Cambridge University Press. Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made. g The images or other third party material in this chapter are included in the chapter’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the chapter’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
https://openalex.org/W2163713991	https://earth-planets-space.springeropen.com/counter/pdf/10.1186/BF03352366	English	null	One-dimensional dynamic simulations of slip complexity of earthquake faults	Earth, planets and space	2,014	cc-by	8,162	2. Theory The N-degree-of-freedom dynamic spring-slider model (see Fig. 1) consists of N sliders of equal mass, m, and springs with one slider being linked by a coil spring of strength, Kc, with the other. A slider is also pulled by a leaf spring of strength, Kl, on a moving plate with a constant velocity, Vp. At time t = 0, all the sliders rest in the individual equilibrium states. Thei-th slider (i = 1, . . . , N) is located at position xi, measured from its initial equilibrium position, along the x- axis. A slider is subjected to a velocity- and state-dependent frictional force, Fi(θi, vi), where θi and vi (= dxi/dt) are the state variable and the sliding velocity, respectively, of the i-th slider. The equation of motion of the i-th slider is q Numerous authors also theoretically studied this prob- lem. Carlson and her co-authors (Carlson, 1991; Carlson and Langer, 1989; Carlson et al., 1991) stated that self- organization of repeated ruptures of uniform faults gives rise to slip complexity when inertial effects and nonlinear velocity-weakening friction are included. They stressed the importance of non-linearity of friction on slip complexity. On the other hand, Rice (1993) strongly argued that all of simulation results from the self-organizing models have been sensitive to the spatial discretization used. He stressed the importance of material heterogeneities and deﬁned a co- herent slip patch size, h∗. So-called self-organizing model faults, having a numerical cell size h > h∗, are inherently dis- crete, thus resulting in Gutenberg-Richter-type complexity for small events. Based on a model through the discretization from a continuum one, Shaw (1994) stated that slip complex- ity is mainly caused by non-linearity of velocity- and state- dependent friction rather than by the matter that the grid size h is larger than a critical length h∗of the model. Knopoff (1996) considered both heterogeneous breaking strengths and non-linearity of friction to be two important factors in m(d2xi/dt2) = Kc(xi+1 −2xi + xi−1) −Kl(xi −Vpt) −Fo(θi, vi). (1) (1) In Eq. (1), Fi(θi, vi) equals to μiσn A, where μi is the friction strength, σn is the normal pressure and A is the contact area between a slider and the moving plate. Experimental and theoretical studies show that two effects, i.e., a direct effect and an evolution one, affect the dynamic friction strength (cf. Dieterich, 1979; Ruina, 1983; Marone, 1998). Copy right c⃝The Society of Geomagnetism and Earth, Planetary and Space Sciences (SGEPSS); The Seismological Society of Japan; The Volcanological Society of Japan; The Geodetic Society of Japan; The Japanese Society for Planetary Sciences. One-dimensional dynamic simulations of slip complexity of earthquake faults Jeen-Hwa Wang and Ruey-Der Hwang Institute of Earth Sciences, Academia Sinica, P.O. Box 1-55, Nankang, Taipei, Taiwan Institute of Earth Sciences, Academia Sinica, P.O. Box 1-55, Nankang, Taipei, Taiwan (Received October 18, 1999; Revised November 13, 2000; Accepted November 13, 2000) (Received October 18, 1999; Revised November 13, 2000; Accepted November 13, 2000) Slip complexity of earthquake faults is studied based on an N-degree-of-freedom dynamical spring-slider system in the presence of slip-law-type, velocity- and state-dependent friction. Simulation results based on such a friction law show that slip complexity depends on the inhomogeneous distribution of the breaking strengths (including its pattern and degree) along the fault and nonlinear velocity- and state-dependent friction. However, for the given model parameters the former is more important than the latter in controlling slip complexity. Frictional effects obviously appear only when the distribution of the breaking strengths is inhomogeneous. In addition, the stiffness ratio, deﬁned as the ratio of the coil spring strength, Kc, to the leaf spring strength, Kl, is also a factor in controlling slip complexity. inﬂuencing slip complexity. 1. Introduction Slip complexity has long been observed from both the ﬁeld survey and from the inversion of earthquake source ruptures (cf. Scholz, 1990; Kanamori, 1994). A complex slip distribu- tion has been interpreted in terms of two phenomenological models. One of them is speciﬁed with barriers (areas where no slip occurs during a mainshock) proposed by Das and Aki (1977) and the other with asperities (areas where large slip takes place during a mainshock) suggested by Kanamori and Stewart (1978). However, the physical nature of asperities and barriers is not well known (cf. Kanamori, 1994). Hence, it is signiﬁcant to explore the causes to yield slip complexity of earthquake faults. In this work, we ﬁrst study the effects due to two factors: the inhomogeneous distribution of the breaking strengths and non-linearity of velocity- and state-dependent friction. Sec- ondly, from simulation results we explore which is the major factor in controlling slip complexity of earthquake faults. In this study, we use an N-degree-of-freedom spring-slider model, proposed by Burridge and Knopoff (1967), in the presence of velocity- and state-dependent friction, to approx- imate earthquake dynamics. Earth Planets Space, 53, 91–100, 2001 Earth Planets Space, 53, 91–100, 2001 2. Theory One of them is the slip law: dθ/dt = −(θv/Dc) ln(θv/Dc). When v = 0, dθ/dt approximates to zero. This implicates that no evo- lution occurs when v = 0. The other is the slowness law: dθ/dt = 1 −θv/Dc. When dθ/dt = 0, state is propor- tional to slowness, i.e., θ = Dc/v. For the two laws, the steady-state friction strength, μss, is μo + (a −b) ln(v/vo) when dθ/dt = 0. This leads to μss = μo for both laws when v = vo. From the simulation results based on a one- degree-of-freedom spring-slider model, Wang (1999) stated that the slowness law strongly resists the slider to move when the driving velocity is less than 10−3 m/s. Thus, it is not appropriate to apply the slowness law to simulate dynamic earthquake ruptures. Hence, in this study only the slip law is taken into account. o Based on the above-mentioned quantities, Eq. (2) is nor- malized to the following form: d2Xi/dτ 2 = s(Xi+1 −2Xi + Xi−1) −(Xi −νpτ) −[γoi + α ln(νi/νo) + β ln(φiνo/ )]. (3) (3) −[γoi + α ln(νi/νo) + β ln(φiνo/ )]. (3) The normalized form of the slip law is dφi/dτ = −(φiνi/ ) ln(φiνi/ ). It is obvious that from the friction law, the value of μ cannot be deﬁned at νi = 0. However, for dynamic simulations, we cannot discard the case at vi = 0. Hence, the values of φi are all set to be /νo when νi = 0 This can lead to μ = μo and avoid the appearance of inﬁnity when νi = 0. The main parameters controlling the motion of a slider are s (the stiffness ratio), α (the parameter for the direct effect), β (the parameter for the evolution effect), νp (the dimensionless velocity of the moving plate), and (the dimensionless characteristic slip displacement). According to the deﬁnition of friction law, the reference velocity νo is not usually regarded as a parameter affecting the friction strength. However, Wang (1999) stressed that νo must be one of the signiﬁcant parameters controlling the dynamic friction strength. This can be seen in the followings. In addition, the boundary condition at each end slider should be a factor in inﬂuencing simulation results. In the following numerical simulations, the two end sliders are ﬁxed and do not move at all. 2. Theory stable slip in rock can result only when the evolution effect is larger than the direct one. The one-state-variable, velocity- and state-dependent friction strength, μ, which employs a state variable, θ, deﬁned by Ruina (1983), has a form of μ = μo +a ln(v/vo)+b ln(θvo/Dc), where μo is the break- ing strength (or the static friction strength), v is the sliding velocity, vo is a constant reference velocity, and Dc is the characteristic slip distance. This leads to μ = μo when v = vo and θ = Dc/vo. Contributions to the total friction strength is scaled by a for the direct effect and by b for the evolution one. Two evolution laws are proposed to control the state variable (cf. Marone, 1998). One of them is the slip law: dθ/dt = −(θv/Dc) ln(θv/Dc). When v = 0, dθ/dt approximates to zero. This implicates that no evo- lution occurs when v = 0. The other is the slowness law: dθ/dt = 1 −θv/Dc. When dθ/dt = 0, state is propor- tional to slowness, i.e., θ = Dc/v. For the two laws, the steady-state friction strength, μss, is μo + (a −b) ln(v/vo) when dθ/dt = 0. This leads to μss = μo for both laws when v = vo. From the simulation results based on a one- degree-of-freedom spring-slider model, Wang (1999) stated that the slowness law strongly resists the slider to move when the driving velocity is less than 10−3 m/s. Thus, it is not appropriate to apply the slowness law to simulate dynamic earthquake ruptures. Hence, in this study only the slip law is taken into account. stable slip in rock can result only when the evolution effect is larger than the direct one. The one-state-variable, velocity- and state-dependent friction strength, μ, which employs a state variable, θ, deﬁned by Ruina (1983), has a form of μ = μo +a ln(v/vo)+b ln(θvo/Dc), where μo is the break- ing strength (or the static friction strength), v is the sliding velocity, vo is a constant reference velocity, and Dc is the characteristic slip distance. This leads to μ = μo when v = vo and θ = Dc/vo. Contributions to the total friction strength is scaled by a for the direct effect and by b for the evolution one. Two evolution laws are proposed to control the state variable (cf. Marone, 1998). 2. Theory The direct effect shows an instantaneous change of the friction strength with a change in velocity, while the evo- lution effect evolves with slip following a change in velocity and is proportional to the negative log of sliding velocity. Un- 91 J.-H. WANG AND R.-D. HWANG: SLIP-COMPLEXITY SIMULATIONS OF EARTHQUAKE FAULTS 92 Fig. 1. An N-degree-of-freedom dynamical spring-slider system. Fig. 1. An N-degree-of-freedom dynamical spring-slider system. s = Kc/Kl, ωo = (Kl/m)1/2, and Do = Fo/Kl. Wang (1995) called the quantity s the stiffness ratio. The quantity ωo/2π is the frequency of oscillation of a single slider at- tached to a leaf spring in the absence of friction. Do is the characteristic displacement of a slider exerted by a force Fo through a spring with strength of Kl. Longer Fo yields longer Do when Kl isﬁxed. Obviously, Do andωo canbeconsidered to be two signiﬁcant units to scale the spatial coordinates, xi, and time, t, respectively. Let Xi = xi/Do, τ = ωot, and = Dc/Do. This leads to dxi/dt = [Fo/(mKl)1/2]∂Xi/∂τ, d2xi/dt2 = (Fo/m)d2Xi/dτ 2, and Vi = d Xi/dτ. Hence, thenormalized(dimensionless)quantitiesareνi = Vi/Doωo, φi = ωoθi, νp = Vp/Doωo, and νo = Vo/Doωo. Do/Vp is the loading time for a leaf spring to stretch enough for over- coming the breaking strength, and νp, is equivalent to the ratio of the slipping time ω−1 o to the loading time. s = Kc/Kl, ωo = (Kl/m)1/2, and Do = Fo/Kl. Wang (1995) called the quantity s the stiffness ratio. The quantity ωo/2π is the frequency of oscillation of a single slider at- tached to a leaf spring in the absence of friction. Do is the characteristic displacement of a slider exerted by a force Fo through a spring with strength of Kl. Longer Fo yields longer Do when Kl isﬁxed. Obviously, Do andωo canbeconsidered to be two signiﬁcant units to scale the spatial coordinates, xi, and time, t, respectively. Let Xi = xi/Do, τ = ωot, and = Dc/Do. This leads to dxi/dt = [Fo/(mKl)1/2]∂Xi/∂τ, d2xi/dt2 = (Fo/m)d2Xi/dτ 2, and Vi = d Xi/dτ. Hence, thenormalized(dimensionless)quantitiesareνi = Vi/Doωo, φi = ωoθi, νp = Vp/Doωo, and νo = Vo/Doωo. Do/Vp is the loading time for a leaf spring to stretch enough for over- coming the breaking strength, and νp, is equivalent to the ratio of the slipping time ω−1 o to the loading time. J.-H. WANG AND R.-D. HWANG: SLIP-COMPLEXITY SIMULATIONS OF EARTHQUAKE FAULTS 93 Fig. 2. The variations of the friction strength with normalized sliding velocity, v/vmax (with vmax = 10−3 m/s), under the controlling of the slip-type friction law for six values of vo: (a) for 10−9 m/s, (b) for 10−6 m/s, (c) for 10−5 m/s, (d) for 10−4 m/s, (e) for 10−3 m/s, and (f) for 1 m/s, when a = 6 × 10−3 and b = 9 × 10−3. For each vo, four values of Dc: 10−5 m denoted by a solid line, 10−4 m by a dashed line, 10−3 m by a dashed-dotted-dashed line, and 10−2 m by a dotted line, are considered. Fig. 2. The variations of the friction strength with normalized sliding velocity, v/vmax (with vmax = 10−3 m/s), under the controlling of the slip-type friction law for six values of vo: (a) for 10−9 m/s, (b) for 10−6 m/s, (c) for 10−5 m/s, (d) for 10−4 m/s, (e) for 10−3 m/s, and (f) for 1 m/s, when a = 6 × 10−3 and b = 9 × 10−3. For each vo, four values of Dc: 10−5 m denoted by a solid line, 10−4 m by a dashed line, 10−3 m by a dashed-dotted-dashed line, and 10−2 m by a dotted line, are considered. are set to be 0.6, 10−3 and 1.5 × 10−3, respectively. For all cases, the initial value of θ for numerical computations is Dc/vo. The variations of μ with v/vmax for six values of vo, i.e., 10−9 m/s, 10−6 m/s, 10−5 m/s, 10−4 m/s, 10−3 m/s, and 1 m/s, are shown in Fig. 2. For each vo, four values of Dc, i.e., 10−4 m, 10−3 m, 10−2 m, and 10−1 m, are considered. In each diagram, the solid line, dashed line, dashed-dotted-dashed line, and dotted line represent the results for Dc = 10−5 m, 10−4 m, 10−3 m, and 10−2 m, respectively. strength with velocity decreases with increasing Dc. A rapid increase in the dynamic friction strength can make a stronger resistance to an object for sliding than a slow one. Hence, large vo and large Dc are beneﬁcial to an object for sliding when vo ≧10−5 m/s. Earthquake fault zones are usually quite complex, and, thus, the distribution of the breaking strengths is not homo- geneous. Inhomogeneous friction strengths might be caused either by a non-uniform distribution of crustal materials or by variable pore pressures (cf. Rice, 1992). J.-H. WANG AND R.-D. HWANG: SLIP-COMPLEXITY SIMULATIONS OF EARTHQUAKE FAULTS Wang (1995) con- sidered a fractal function speciﬁed with a fractal dimension, D, to describe the inhomogeneous distribution of the break- ing strengths. A fractal distribution of the breaking strengths is also considered in this study. The Fourier ﬁltering method developed by Saupe (1988) is applied to yield a fractal distri- bution with 2n (n is an integer) discrete points. Let γo max and γo min be, respectively, the maximum and minimum values of thebreakingstrengths, anddeﬁne R = (γo max−γo min)/γo max to represent the degree of the inhomogeneous distribution of the breaking strengths. There are two patterns for variations of μ with v/vmax, depending upon vo. The value of vo = 10−5 m/s is a critical one to separate the two patterns. When vo < 10−5 m/s, the value of μ increases rapidly ﬁrst from the initial value to a peak value, and then decreases. The velocity range, within which the friction strength increases from the initial value to the peak one, increases with Dc and slightly with vo. This in- dicates that the friction force makes resistance to prohibit an object for sliding. The amount of resistance increases with Dc and slightly with vo when vo < 10−5 m/s. On the other hand, when vo ≧10−5 m/s, the value of μ ﬁrst drops to a value less than 0.6, then increases from that to a peak value, and ﬁnally decreases again. The decrease in the dynamic friction strength at low velocities behaves like an impulse, which supplies additional energy to an object. After the fric- tion strength drops, the dynamic friction strength increases with velocity. The increasing rate of the dynamic friction 2. Theory The velocity- and state-dependent dynamic friction strength at the i-th slider is represented by μi = μoi + ai ln(vi/vio) + bi ln(θivio/Dic). The breaking strength, μoi, is a function of position, with a maximum value of μo max. In order to simplify numerical computations, the values of the parameters ai, bi, voi, and Doi are individually consid- ered to be constant at all sliders, i.e., letting ai = a, bi = b, vio = vo and Dic = Dc. This leads to Fi(θi, vi) = σn A[μo+ a ln(vi/vo) + b ln(θivo/Dc)] = Fo[γoi + α ln(vi/vo) + β ln(θivo/Dc)], where Fo = μo maxσn A, γoi = μoi/μo max, α = a/μo max, and β = b/μo max. The slip law becomes dθi/dt = −(θivi/Dc) ln(θivi/Dc). Replacing Fi(θi, vi) into Eq. (1), we have m(d2xi/dt2) = Kc(xi+1 −2xi + xi−1) −Kl(xi −Vpt) −Fo[γoi + α ln(vi/vo) + β ln(θivo/Dc)]. (2 To understand the variation of friction strength, μ, with sliding velocity, we calculate the variation of μ with v, which increases, in a sine-function form, from 0 to vmax (= 10−3 m/s), throughthefourth-orderRunge-Kuttamethod(cf.Press et al., 1986). The values of model parameters μo, a, and b (2) To simply deal with the problem, Eq. (2) is normalized in advance. We deﬁne three characteristic parameters, i.e., 3. Simulation Results In order to explore slip complexity of earthquake faults, we here study the spatial distribution of fault slip through numerical simulations based on Eq. (3) with slip-law-type velocity- and state-dependent friction. The friction strength J.-H. WANG AND R.-D. HWANG: SLIP-COMPLEXITY SIMULATIONS OF EARTHQUAKE FAULTS 94 Fig. 3. The spatial distributions of ﬁnal slip of sixty simulation events for three distributions of the breaking strengths: (a) for a fractal distribution with D = 1.5, (b) for a random distribution, and (c) for a uniform distribution, when s = 4, γo min = 0.98, α = 10−2, β = 1.5×10−2, νo = 1, and = 10−3. The site with γo max (= 1) is denoted by a cross and that with γo min shown by an arrow sign. The dipping line segments at the two end sliders demonstrate the ﬁxed boundary condition. The dotted line in the lower diagrams represents the site with γo min. Fig. 3. The spatial distributions of ﬁnal slip of sixty simulation events for three distributions of the breaking strengths: (a) for a fractal distribution with D = 1.5, (b) for a random distribution, and (c) for a uniform distribution, when s = 4, γo min = 0.98, α = 10−2, β = 1.5×10−2, νo = 1, and = 10−3. The site with γo max (= 1) is denoted by a cross and that with γo min shown by an arrow sign. The dipping line segments at the two end sliders demonstrate the ﬁxed boundary condition. The dotted line in the lower diagrams represents the site with γo min. the two end sliders. is deﬁned only for the positive velocities. This means that no backward motions in the fault are allowed. The simulation procedure is brieﬂy described here. First, we integrate Eq. (3) with friction forward in time. Secondly, we scan all sliders again to ﬁnd the sliders for which the forces are closest to the individual breaking strengths. The velocity νi(τ + δτ) and the displacement Xi(τ + δτ) of the i-th slider at time τ + δτ is calculated from the acceleration ai(τ) (= d2Xi/dτ 2) at time τ, i.e., νi(τ + δτ) = ai(τ)δτ and = ai(τ)(δτ)2/2. The value of φi(τ) is computed from the slip-type evolution law through the fourth-order Runge-Kutta method. The compu- tation processes are conducted until all sliders have come to rest. 3. Simulation Results Thirdly, after all sliders come to rest, we add again a loading force from the moving plate to each slider, and then repeat the ﬁrst and second processes. For simpliﬁcation, the inertial effect is considered to be constant for all cases, thus letting m = 1 kg. The strength of the leaf spring, Kl, is set to be 1 nt/m, thus leading to ωo = 1 Hz. Four values of s (= Kc/Kl), i.e., 2, 3, 4, and 5, are used. It is noted that in this study, since only the normalized equations are used, and only ωo and s are taken into account, the absolute units of m, Kc, and Kl are less important. For simpliﬁcation, Do is set to be 1 m for all cases. Generally, the value of Vp is generally about 10−9 m/s, and, thus, the value of νp is 10−9 when Doωo = 1 m/sec. Since Eq. (3) is a normalized equation, γo max is 1 for all cases in this study. The difference a −b or α −β depends upon the gouge layer thickness and the shear displacement (cf. Marone, 1998). For a bare fault plane, the value of a −b ranges from −0.004 to 0. The value changes when a gouge layer exists. It changes from a negative value to a positive one when the gouge layer thickness is increased. The value of a −b also decreases with increasing shear displacement. Since the parameter α (or a) only changes the amount of the direct effect, its value is set to be a constant (= 10−2) in this study. Two values of the parameter β, i.e., 0.5 × 10−2 and 2 × 10−2, are considered. The dimensionless reference velocity, νo, and The boundary condition must be a factor in affecting sim- ulation results. Since in this study we are only focusing the slip distribution of the model events, the two ends of the model fault do not break in the seismic cycles. Hence, the breaking strengths at two end sliders are set to be inﬁnity for avoiding numerical computations. In other words, we consider the ﬁxed boundary condition. To show the ﬁxed boundary condition, two dipping line segments are plotted at J.-H. WANG AND R.-D. HWANG: SLIP-COMPLEXITY SIMULATIONS OF EARTHQUAKE FAULTS 95 Fig. 4. 3. Simulation Results The spatial distributions of ﬁnal slip of sixty simulation events for three values of R: (a) for 10% (γo min = 0.90), (b) for 6% (γo min = 0.94), and (c) for 2% (γo min = 0.98), when s = 4, νo = 1, α = 10−2, β = 1.5 × 10−2, and = 10−3. The distribution of the breaking strengths is a fractal function with D = 1.5. The symbols used are the same as those shown in Fig. 3. Fig. 4. The spatial distributions of ﬁnal slip of sixty simulation events for three values of R: (a) for 10% (γo min = 0.90), (b) for 6% (γo min = 0.94), and (c) for 2% (γo min = 0.98), when s = 4, νo = 1, α = 10−2, β = 1.5 × 10−2, and = 10−3. The distribution of the breaking strengths is a fractal function with D = 1.5. The symbols used are the same as those shown in Fig. 3. For exploring the effects due to the distribution of the breaking strengths, we consider three kinds of spatial distri- butions of the breaking strengths: a fractal distribution with D = 1.5, a random distribution, and a uniform distribution. For the ﬁrst two distribution, γo min = 0.98. For the last one, the breaking strength is set to be 0.98 at slider 9 and to be 1 at the rest sliders. The values of s, νo, and are 1, 1, and 10−3, respectively. The spatial distributions of ﬁnal slip of sixty events for the three distributions of the breaking strengths are shown in Fig. 3. For the uniform distribution, three large events, rupturing along the whole fault (expect for the two end sliders), appear in the ﬁrst three time intervals. After that, a large event, rupturing almost along the whole fault, followed by two small ones, occurring at the two second-end sliders, appears repeatedly. When the fractal and random distributions are applied, the ruptures are non-uniform along the fault. For the fractal distribution (see Fig. 3(a)), the fault can be divided mainly into two segments: one, in the right- handed side, with lower breaking strengths and the other, in the left-handed side, with higher breaking strengths. In addi- tion, there is a transition zone, where the breaking strengths increase from right to left, between the two above-mentioned segments. 3. Simulation Results A rupture cycle along the fault essentially con- the ratio (= Dc/Do) are two frictional factors in affecting the motions of sliders. As mentioned above, for most of cases νo is set to be 1. The parameter Dc is affected by several factors (cf. Marone, 1998). Its values measured from laboratory experiments are generally in the range of 10−6 − 10−5 m, which are many orders of magnitude smaller than ﬁeld-based estimates and those values required by theoretical simulations of ruptures (cf. Beeler and Tullis, 1996). In this study we consider four values of Dc, i.e., 10−4 m, 10−3 m, 10−2 m, and10−1 m, whicharealmostbetweenthelaboratory values and the ﬁeld-based ones. Hence, the values of related to them are 10−4, 10−3, 10−2, and 10−1, respectively, because Do = 1 m. The number of sliders in use is 27 = 128. The time step δτ must be a factor in affecting numerical results. A big value of δτ can cause several starting points for one event and is able to produce numerical instability. However, the simulation results for various δτ’s are similar when δτ < 0.05. In other words, numerical stability exists when δτ < 0.05. For resulting in numerical stability, the time step size is δτ = 0.01 (or δt = 0.01 sec) in the following computations. Displayed also in the following ﬁgures are the dimensionless maximum and minimum displacements (denoted as Smax and Smin, respectively). J.-H. WANG AND R.-D. HWANG: SLIP-COMPLEXITY SIMULATIONS OF EARTHQUAKE FAULTS 96 Fig. 5. The spatial distributions of ﬁnal slip of sixty simulation events for two values of β and νo: (a) for β = 10−2 and νo = 1, (b) for β = 2 × 10−2 νo = 1, (c) for β = 10−2 and νo = 10−9, and (d) for β = 2 × 10−3 and νo = 10−9, when s = 4, α = 10−2, and = 10−3. The distribution o breaking strengths is a fractal function with D = 1.5 and R = 2%. The symbols used are the same as those shown in Fig. 3. Fig. 5. 3. Simulation Results The spatial distributions of ﬁnal slip of sixty simulation events for two values of β and νo: (a) for β = 10−2 and νo = 1, (b) for β = 2 × 10−2 and νo = 1, (c) for β = 10−2 and νo = 10−9, and (d) for β = 2 × 10−3 and νo = 10−9, when s = 4, α = 10−2, and = 10−3. The distribution of the breaking strengths is a fractal function with D = 1.5 and R = 2%. The symbols used are the same as those shown in Fig. 3. Fig. 5. The spatial distributions of ﬁnal slip of sixty simulation events for two values of β and νo: (a) for β = 10−2 and νo = 1, (b) for β = 2 × 10−2 and νo = 1, (c) for β = 10−2 and νo = 10−9, and (d) for β = 2 × 10−3 and νo = 10−9, when s = 4, α = 10−2, and = 10−3. The distribution of the breaking strengths is a fractal function with D = 1.5 and R = 2%. The symbols used are the same as those shown in Fig. 3. because different fractal dimensions and different values of α and β could lead to different results. Different fractal dimensions produce different fractal distributions, thus be- ing able to yield different spatial distributions of ﬁnal slip. However, tests show that when the coefﬁcients of friction law are ﬁxed, the maximum ﬁnal slips for different fractal distributions with fractal dimensions varying from 1.1 to 1.9 are essentially similar, and the spatial distributions of slip for different fractal dimensions are also similar to that displayed in Fig. 3. (The spatial distributions of the displacements for different fractal distributions of the breaking strengths are not shown in this study.) Hence, the above-mentioned statement can be correct when the values of α and β are ﬁxed. Results obviously show the importance of inhomogeneous breaking strengths (the highest one for the random distribution, the intermediate one for the fractal distribution and the lowest one for the uniform distribution) on the spatial distribution and the maximum value of ﬁnal slip. sists of three steps: First, a larger-sized event occurs along the right segment with lower breaking strengths. Secondly, few small events appear in the transition zone and at the two second-end sliders. 3. Simulation Results The ﬁrst 52 events belong to a rupture cycle. The energy in the fault system exerted by the moving plate completely releases after the occurrences of the 52 events. One could consider that the 52 events are sub-events of a large one. However, the total rupture time of the 52 events is too long to consider them to be sub-events of a large one. The simulation results for R = 6% (as shown in Fig. 4(b)) show that there are two rupture cycles, each occurring in a somewhat long time interval, and in each cycle there are small and intermediate-sized events. The seismicity pattern for R = 6% seems to be a middle state between those for R = 2% and 10%. However, the values of the maximum ﬁnal slip for the three values of R are almost equal and in- dependent upon R. For each case, except for the two end sliders, almost all sliders move in each rupture cycle. In addition, the total rupture times for individual cycles in dif- ferent diagrams are approximately equal. Of course, the total times for the sixty events are different for the three cases: the largest one for R = 2%, the second largest one for R = 6%, and the smallest one for R = 10%. In addition to the effects due to the distribution of the breaking strengths, we need to study the frictional effects. As mentioned above, the major factors of the friction law include α, β, and , and νo is set to be 1. For comparison, we also consider two cases with νo = 10−9. Figure 5 shows the plots for two values of β and two values of νo: (a) for β = 1 × 10−2 and νo = 1, (b) for β = 2 × 10−2 and νo = 1, (c) for β = 1 × 10−2 and νo = 10−9, and (d) for β = 2 × 10−3 and νo = 10−9, when s = 4, α = 10−2, and = 10−3. The distribution of the breaking strengths is a fractal function with D = 1.5 and R = 2%. The case for β = 1.5 × 10−2 and νo = 1 can be seen in Fig. 6(b). 3. Simulation Results Finally, the other larger-sized event takes place along the left segment with higher breaking strengths. The energy provided by the moving plate releases completely during a rupture cycle. Such a rupture cycle almost appears repeatedly. Of course, the number of small events, occurring in the transition zone, changes with time. However, these small events play a minor role for energy release during a rupture cycle. The number of rupture cycles for the uniform distribution is larger than that for each of the two inhomo- geneous distributions. For each case, almost all the sliders, except for the two end ones, slide in a rupture cycle. The total rupture time for a rupture cycle is approximately a con- stant, yet about ten times smaller the time interval between two rupture cycles. This phenomenon can also be seen in the following cases with different model parameters. Figure 3 also shows that the maximum ﬁnal slip somewhat depends upon the type of distribution: the largest one for the uniform distribution, the intermediate one for the fractal dis- tribution, and the smallest one for the random distribution. Of course, this statement is not necessary to be a general one, For an inhomogeneous distribution of the breaking strengths, in addition to the pattern of inhomogeneities, the degree of inhomogeneities would be also a signiﬁcant factor in affecting slip complexity. In this study, the parameter R is considered as an indication to show the degree of inho- .-H. WANG AND R.-D. HWANG: SLIP-COMPLEXITY SIMULATIONS OF EARTHQUAKE FAULTS 97 Fig. 6. The spatial distributions of ﬁnal slip of sixty simulation events for four values of : (a) for 10−4, (b) for 10−3, (c) for 10−2, and (d) for 10−1, when s = 4, α = 10−2, β = 1.5 × 10−2, and νo = 1. The distribution of the breaking strengths is a fractal function with D = 1.5 and R = 2%. The symbols used are the same as those shown in Fig. 3. Fig. 6. The spatial distributions of ﬁnal slip of sixty simulation events for four values of : (a) for 10−4, (b) for 10−3, (c) for 10−2, and (d) for 10−1, when s = 4, α = 10−2, β = 1.5 × 10−2, and νo = 1. The distribution of the breaking strengths is a fractal function with D = 1.5 and R = 2%. 3. Simulation Results The symbols used are the same as those shown in Fig. 3. mogeneities. The spatial distributions in ﬁnal slip of sixty simulation events for three value of R, i.e., 10%, 6%, and 2%, which are related to γo min = 0.90, 0.94, and 0.98, re- spectively, for a fractal distributions (with D = 1.5) of the breaking strengths are shown in Fig. 4. The values of s, νo, and used are 1, 1, and 10−3, respectively. When R = 2%, the simulation results (see Fig. 4(c)) are the same as those shown in Fig. 3(a), and the results were described previously. The simulation results for R = 10% (as shown in Fig. 4(a)) show that all events are small, and each of them is associ- ated with either a small number of slid sliders or only one slid slider. The ﬁrst 52 events belong to a rupture cycle. The energy in the fault system exerted by the moving plate completely releases after the occurrences of the 52 events. One could consider that the 52 events are sub-events of a large one. However, the total rupture time of the 52 events is too long to consider them to be sub-events of a large one. The simulation results for R = 6% (as shown in Fig. 4(b)) show that there are two rupture cycles, each occurring in a somewhat long time interval, and in each cycle there are small and intermediate-sized events. The seismicity pattern for R = 6% seems to be a middle state between those for R = 2% and 10%. However, the values of the maximum mogeneities. The spatial distributions in ﬁnal slip of sixty simulation events for three value of R, i.e., 10%, 6%, and 2%, which are related to γo min = 0.90, 0.94, and 0.98, re- spectively, for a fractal distributions (with D = 1.5) of the breaking strengths are shown in Fig. 4. The values of s, νo, and used are 1, 1, and 10−3, respectively. When R = 2%, the simulation results (see Fig. 4(c)) are the same as those shown in Fig. 3(a), and the results were described previously. The simulation results for R = 10% (as shown in Fig. 4(a)) show that all events are small, and each of them is associ- ated with either a small number of slid sliders or only one slid slider. 4. Discussion maximum ﬁnal slip decrease with increasing β. Figures 5(c) and 5(d) show that when νo = 10−9, the seismicity patterns and the values of the maximum ﬁnal slip for the two values of are the same. In the following simulations, the values of α and β are 10−2 and 1.5 × 10−2, respectively. The inhomogeneous distribution of the breaking strengths andnon-linearityofvelocity-andstate-dependentfrictionare usually considered as two important factors in affecting slip complexity. The debate about which is the predominant fac- tor has been lasted for a long time. Carlson and her co-author (cf. Carlson and Langer, 1989; Carlson et al., 1991) stressed that nonlinear velocity-dependent friction is the major fac- tor. Cochard and Madariaga (1994, 1996) also emphasized the importance of friction on slip complexity. On the other hand, Rice (1993) strongly argued that the complexity comes from the self-organized models and the BK models has been sensitive to the (inherent) spatial discretizations used. He proposed that the dynamic effect suggested by Carlson and her co-authors is not the major factor in inﬂuencing slip com- plexity in the earthquake faults. He also stated that models thatincorporateapproximatelygeometricfault-zonedisorder can produce slip histories with features that are comparable to observations. Based on a model through the discretiza- tion from a continuum one, Shaw (1994) stressed that slip complexity is mainly caused by non-linearity of velocity- and state-dependent friction rather than by the matter that the grid size h is larger than a critical length h∗of the model. Figure 6 shows the spatial distributions of ﬁnal slip for sixty simulation events for the four values of , i.e., 10−4, 10−3, 10−2, and 10−1, when s = 4 and νo = 10−3. The distribution of the breaking strengths is a fractal function with D = 1.5 and R = 2%. It is evident that large is more capable of generating larger-sized events than small , and the maximum ﬁnal slip increases with . There are several rupture cycles for > 10−4, but only two cycles for = 10−4. The rupture times for individual cycles for the four cases are almost equal. Finally, the effects caused by the stiffness ratio are taken into account. Figure 7 shows the spatial distributions of ﬁnal slip for sixty simulation events for four values of s, i.e., 2, 3, 4, and 5, when νo = 1 and = 10−3. 3. Simulation Results These plots for νo = 1 show that large events, rupturing almost along the whole fault, are generated when β = 1 × 10−2 or α −β = 0, and both the size of the largest event and the J.-H. WANG AND R.-D. HWANG: SLIP-COMPLEXITY SIMULATIONS OF EARTHQUAKE FAULTS 98 Fig. 7. The spatial distributions of ﬁnal slip of sixty simulation events for four values of s: (a) for 2, (b) for 3, (c) for 4, and (d) for 5, when α = 10−2, β = 1.5 × 10−2, νo = 1, and = 10−3. The distribution of the breaking strengths is a fractal function with D = 1.5 and R = 2%. The symbols used are the same as those shown in Fig. 3. Fig. 7. The spatial distributions of ﬁnal slip of sixty simulation events for four values of s: (a) for 2, (b) for 3, (c) for 4, and (d) for 5, when α = 10−2, β = 1.5 × 10−2, νo = 1, and = 10−3. The distribution of the breaking strengths is a fractal function with D = 1.5 and R = 2%. The symbols used are the same as those shown in Fig. 3. 4. Discussion The distribution of the breaking strengths is the same as that in Fig. 6. It is obvious that the range of ruptures increases slightly with s, but the maximum ﬁnal slip does not. J.-H. WANG AND R.-D. HWANG: SLIP-COMPLEXITY SIMULATIONS OF EARTHQUAKE FAULTS 99 than smaller s. Wang (1995) also stated that large s is more capable of generating larger-sized events, thus leading to a smaller b-value, than small s. The reasons are described as below: The coil spring (with a stiffness of Kc) between two sliders only transfers the mechanical energy from one slider to the nearest ones. Thus, it does not take the elastic energy in the system away. On the other hand, the leaf spring (with a stiffness Kl) between a slider and the moving plate plays two roles: One is to transfer the mechanical energy, exerted from the moving plate, into the system through the KlVpt term, and the other to remove the energy through the −Klxi term. Hence, the stiffness ratio is also a parameter representing the level of conservation of energy in the system. Larger s indicates a higher level of conservation or a lower level of dissipation of energy in the system. Hence, an increase in the capability of yielding larger-sized events with s must be reasonable. Of course, the difference in the conclusions obtained by the rival groups might also be due to the use of different friction laws used by different authors. Figure 3 obviously show that when the distribution of the breaking strengths is uniform, the spatial distribution of sim- ulated ﬁnal slip is also uniform whatever the values of model parameters are. Figure 4 shows the effects caused by the de- gree of inhomogeneity, represented by the parameter R, on slip complexity. A decrease in R results in an increase in the capability to generate large events. This is due to an increase in the degree of homogeneity with decreasing R. The dis- tribution of the breaking strengths is uniform when R = 0. Results shown in Figs. 3 and 4 indicate that the degree and the pattern of inhomogeneous breaking strengths over the fault plane are the major factors in affecting slip complexity, while non-linearity of velocity- and state-dependent friction is less signiﬁcant. However, the present results cannot directly con- ﬁrm Rice’s proposition. In addition, Fig. 5. Conclusions The characteristic slip, Dc, and the reference velocity, vo, of the slip law are two important parameters affecting dy- namic friction strength. The velocity-weakening process is produced when the reference velocity, vo, is smaller than a critical value, voc (= 10−5 m/s in this study), while the velocity-hardening process is yielded when vo > voc. The changing rate of μ with sliding velocity depends on Dc. Since the friction law itself could affect simulation results, the following conclusions mainly hold for the velocity- and state-dependentfrictionlawusedinthisstudy. Simulationre- sults show that slip complexity depends on the heterogeneity of the breaking strengths and nonlinear velocity- and state- dependent friction. The former is more important than the latter. When the distribution of the breaking strengths is homogeneous, only large events, which rupture almost uni- formly along the fault, are generated. When the distribution of the breaking strengths is inhomogeneous, the friction law and the stiffness ratio are also signiﬁcant factors in affecting slip complexity. Figure 6 shows the spatial distributions of ﬁnal slip for sixty simulation events for the four values of , i.e., 10−4, 10−3, 10−2, and 10−1, when s = 4 and νo = 10−3. It is evi- dent that large is more capable of generating larger-sized events than small , and the maximum ﬁnal slip increases with . From Fig. 2(e) with vo = 10−3 m/s (or νo = 10−3), we can see that the friction strength and the increasing rate of friction strength at low velocities generally increases with decreasing Dc. This indicates that smaller Dc (or ) is less appropriate for generating larger-sized model events. This leads to the phenomena observed in Fig. 6. Acknowledgments. The authors would like to express thanks to two anonymous reviewers for helpful comments and suggestions. This study was ﬁnancially supported from Academia Sinica and the National Sciences Council, ROC under Grant NSC87-2116-M- 001-020. 4. Discussion 4 also shows that the capability to generate larger-sized earthquakes decreases with increasing R, and only small events are produced when R = 10%. Simulation results in Figs. 3–7 also show that the maxi- mum ﬁnal slip depends upon mainly the characteristic slip distance and the reference velocity, but only slightly affected by stiffness ratio, the pattern of the inhomogeneous distri- bution of the breaking strengths, and the degree of inho- mogeneities. The former two parameters directly affect the dynamical rupture processes (see Fig. 2), but the latter three parameters do not. Hence, the former two parameters are more important on the size of the ﬁnal slip than the latter three. When the distribution of the breaking strengths is inhomo- geneous, the effects on slip complexity caused by nonlinear velocity- and state-dependent friction obviously appear. The main parameters in the velocity- and state-dependent fric- tion law include α (the parameter for the direct effect), β (the parameter for the evolution effect), (the dimension- less characteristic slip distance), and νo (the dimensionless reference velocity). Figure 5(a) (for β = 10−2), Fig. 5(b) (for β = 2×10−2), and Fig. 6(b) (for β = 1.5×10−2) show the effect of β as well as α −β on slip complexity when νo = 1. When β = 10−2 or α −β = 0, only large events, rupturing almost along the whole fault, are generated. The size of the largest event and the maximum ﬁnal slip decrease with increasing β. This indicates that when νo = 1, large β is not appropriate for the generation of larger-sized events. An increase in β (or b) results in an increase in the increasing rate of the dynamic friction strength with velocity. The effect is the same as that for decreasing (or decreasing Dc) as shown in Fig. 2. In addition, Figs. 5(c) and 5(d) also show that no larger-sized events can be yielded when νo = 10−9. As shown in Fig. 2(a) for νo = 10−9, the dynamic friction strength obviously increases with velocity. Hence, after a slider moves, the increasing dynamic friction force resists the slider to slide further. Hence, small νo cannot result in large events. References Beeler, N. M. and T. E. Tullis, Self-healing slip pulses in dynamic rupture models due to velocity-dependent strength, Bull. Seism. Soc. Am., 86, 1130–1148, 1996. The stiffness ratio, s, would be a factor in inﬂuencing slip complexity. The results shown in Fig. 7 display that larger s is more appropriate for producing larger-sized events Burridge, R. and L. Knopoff, Model and theoretical seismicity, Bull. Seism. Soc. Am., 57, 341–371, 1967. Carlson, J. M., Time intervals between characteristic earthquakes and cor- 100 J.-H. WANG AND R.-D. HWANG: SLIP-COMPLEXITY SIMULATIONS OF EARTHQUAKE FAULTS relation with smaller events: An analysis based on a mechanical model of fault, J. Geophys. Res., 96, 4255–4267, 1991. Press, W.H., B.P.Flannery, S.A.Teukolsky, andW.T.Vetterling, Numerical Recipes, 818 pp., Cambridge Univ. Press, Cambridge, 1986. relation with smaller events: An analysis based on a mechanical model of fault, J. Geophys. Res., 96, 4255–4267, 1991. p y Carlson, J. and J. S. Langer, Mechanical model of an earthquake fault, Phys. Rev. A, 40, 6470–6484, 1989. Rice, J. R., Fault stress states, pore pressure distributions, and the w Rice, J. R., Fault stress states, pore pressure distributions, and the weakness of the San Andreas Fault, in Fault Mechanics and Transport Properties of Rocks, pp. 475–503, Academic Press, 1992. Carlson, J. M., J. S. Langer, B. E. Shaw, and C. Tang, Intrinsic properties of a Burridge-Knopoff model of an earthquake fault, Phys. Rev. A, 44, 884–897, 1991. Rice, J. R., Spatio-temporal complexity of slip on a fault, J. Geophys. Res., 98, 9885–9907, 1993. Cochard, A. and R. Madariaga, Dynamic faulting under rate-dependent friction, Pure Appl. Geophys., 142, 419–445, 1994. Ruina, A. L., Slip instability and state variable friction laws, J. Geophys. Res., 88, 10,359–10,370, 1983. Saupe, D., Algorithmsforrandomfractals, Chapter2, inTheScienceofFrac- tal Images, edited by H. O. Peitgen and D. Saupe, pp. 71–136, Springer Verlag, New York, 1988. Cochard, A. and R. Madariaga, Complexity of seismicity due to highly rate-dependent friction, J. Geophys. Res., 101, 25,321–25,336, 1996. tal Images, edited by H. O. Peitgen and D. Saupe, pp. 71–136, Springer Verlag, New York, 1988. Das, S. and K. Aki, Fault planes with barriers: a versatile earthquake model, J. Geophys. Res., 82, 5658–5670, 1977. Scholz, C. H., The Mechanics of Earthquakes and Faulting, 439 pp., Cam- bridge Univ. Press, Cambridge, 1990. Dieterich, J. H., Modeling of rock friction 1. Experimental results and constitutive equations, J. Geophys. Res., 84, 2161–2168, 1979. Shaw, B. J.-H. Wang (e-mail: jhwang@earth.sinica.edu.tw) and R.-D. Hwang References E., Complexity in a spatially uniform continuum fault model, Geophys. Res. Lett., 21, 1983–1986, 1994. Kanamori, H., Mechanics of Earthquake, Annu. Rev. Earth Planet. Sci., 22, 207–237, 1994. Wang, J.H., Effectofseismiccouplingonthescalingofseismicity, Geophys. J. Int., 121, 475–488, 1995. Kanamori, H. and G. Stewart, Seismological aspects of the Guatemala earth- quake of February 4, 1976, J. Geophys. Res., 83, 3427–3434, 1978. Wang, J. H., A dynamical study of two one-state-variable, rate- and state- dependent friction laws, Bull. Seism. Soc. Am., 1999 (in revision). Knopoff, L., The organization of seismicity on fault networks, Proc. Natl. Acad. Sci., USA, 93, 3830–3837, 1996. Marone, C., Laboratory-derivedfrictionlawsandtheirapplicationtoseismic faulting, Annu. Rev. Earth Planet. Sci., 26, 643–696, 1998. J.-H. Wang (e-mail: jhwang@earth.sinica.edu.tw) and R.-D. Hwang
https://openalex.org/W4383684274	http://vital.lib.tsu.ru/vital/access/services/Download/koha:001000277/SOURCE1	English	null	Interest rate liberalization and debt risk in China	Vestnik Tomskogo gosudarstvennogo universiteta. Èkonomika	2,023	cc-by	7,951	Song Yuxuan1, Gao Yan2 Song Yuxuan1, Gao Yan2 1 Saint Petersburg State University, Saint Petersburg, Russian Federation, 1056022128@qq.com 2 Jining Medical University, Jining, China, 1056022128@qq.com 1 Saint Petersburg State University, Saint Petersburg, Russian Federation, 1056022128@qq.com 2 Jining Medical University, Jining, China, 1056022128@qq.com Abstract. Interest rate liberalization, which will have a multi-impact on the economy in the process of marketization, especially on the macro-debt market, is a requirement to develop the market economy. It is a current important economic topic that the risk control of the macro-debt market is related to the healthy degree of the economic de- velopment after the debt financing leverage turns to assets investment, and the study around these contents should be a cause for concern. As an important component of deepening financial reforms, the effect of interest rate liberalization on owners’ opera- tions and management behavior of a micro-enterprise is prominent; on the other hand, the continuous climbing debt risk of local government will affect the financial deci- sions and implementation effect of enterprises by spillover effect. We made an empir- ical research on the extent of interest rate liberalization in China integrating relevant modes with the statistical data of PBC and our nation. Research on the reform in the Chinese financial market showed that prompting the reform of interest rate liberaliza- tion can obviously increase enterprises’ financing capabilities which will be restrained by the aggravation of the local government’s debt risk, and the significant enhance- ment of it enhances the impetus to enterprises’ financing capabilities from interest rate liberalization. Keywords: interest rate, China’s debt, risk management, China’s financial system, liberalization, marketization, China’s financial reform, financial risk For citation: Song Yuxuan & Gao Yan. (2023) Interest rate liberalization and debt risk in China. Vestnik Tomskogo gosudarstvennogo universiteta. Ekonomika – Tomsk State University Journal of Economics. 61. pp. 136–151. doi: 10.17223/19988648/61/9 For citation: Song Yuxuan & Gao Yan. (2023) Interest rate liberalization and debt risk in China. Vestnik Tomskogo gosudarstvennogo universiteta. Ekonomika – Tomsk State University Journal of Economics. 61. pp. 136–151. doi: 10.17223/19988648/61/9 For citation: Song Yuxuan & Gao Yan. (2023) Interest rate liberalization and debt risk in China. Vestnik Tomskogo gosudarstvennogo universiteta. Ekonomika – Tomsk State University Journal of Economics. 61. pp. 136–151. doi: 10.17223/19988648/61/9 Вестник Томского государственного университета. Экономика. 2023. № 61. С. 136–151. Tomsk State University Journal of Economics. 2023. 61. рр. 136–151. Вестник Томского государственного университета. Экономика. 2023. № 61. С. 136–151. Tomsk State University Journal of Economics. 2023. 61. рр. 136–151. Вестник Томского государственного университета. Экономика. 2023. № 61. С. 136–151. Tomsk State University Journal of Economics. 2023. 61. рр. 136–151. Original article UDC 08.00.10 doi: 10.17223/19988648/61/9 Original article UDC 08.00.10 doi: 10.17223/19988648/61/9 Preface China’s national economy grew rapidly in the process of economic globaliza- tion after we joined the WTO, and our economic potential was developed after participating in international trade and technological transformation. The growth of disposable income stimulates the need for market-based investment assets to a great extent. But, in a short time, China’s financial market is not mature enough: its relatively few kinds of financial products cannot satisfy customers’ rapidly growing needs, there is a relative lack of products that are available for investors, the asset price inflated under the situation of the increasing supply relationship which cannot reflect the real value of it, and then formed the asset bubble [1, 2]. China is also facilitating the opening of the financial market, re- © Song Yuxuan, Gao Yan, 2023 Song Yuxuan, Gao Yan. Interest rate liberalization and debt risk in China forming the interest rate liberalization gradually; opening currency and bond markets step by step; allowing deposit and loan interest rates to have a limited fluctuation; permitting foreign capital to have shares in bond companies, banks, and other financial institutes while enlarging the opening of domestic commodi- ty market, and then raise the opening degree of the financial market to foreign investment and totally deregulate the loan interest rate [3, 4]. During the previous period, Chinese enterprises obtained much capital from domestic financial institutions, purchased massive foreign assets, and formed a huge debt system of enterprises with a loose domestic financing environment. Meanwhile, the investment vehicle taking real estate as the representation also attracts residents and other private sectors’ investment and increases the resident debt efficiently. The government expands its investment scale year by year, borrows money from banks, and releases long-term bonds with cities’ investment companies and other platforms to stimulate economic growth, which made the rapid growth of the government’s debt scale and the rapid development of local governments’ financing platform. Local governments are more positive to borrow money, which therefore leads to the rapidly growing leverage ratio of Chinese local governments below the provincial level [5, 6]. Chinese financial institutions re- lease financial bonds for major capital projects to serve the real economy and coordinate with our country’s construction of the main fundamental facilities, which push the long-term growth of China’s economy effectively, and financial departments’ debt ratio is rising quickly. Preface In the Chinese financial market, the debt of residents, governments, enter- prises, and financial institutions together make up the Chinese debt system, the debt risk of the four departments also transmits mutually, and the rapidly rising debt ratio raises possibilities of systematic debt risk. The current situation of Chinese economic development indicates that the debt risk in the real economy will ultimately impact the capital safety of financial institutions, the happening of economy’s systematic debt risk will also erupt as a form of the financial mar- ket’s debt risk [7, 8]. The reform process of China’s interest rate liberalization With the quick increase in the Chinese economy, the regulatory measures of interest rates limit capital market development and restrain the economic growth rate. With the impacts of the low speed of economic globalization’s develop- ment, the escalating of Sino–US trade frictions, the misappropriation of regional trade protectionism measures, and other factors, China’s economic growth and its capital market withstand immense domestic and foreign pressure. It is crucial for Chinese financial market to speed up the promotion of interest rate liberali- zation and reduce the risk of debt market efficiently [9, 10]. While interest rate liberalization reform, which needs the protection of a sta- ble macro-economic environment, involves interests of various aspects, the overheated economy will raise the interest rate, cause the market entities’ com- 137 Финансы / Finance petition for capital, seduce excessive needs for capital, and inspire inflation [9]. On the contrary, economic depression may cause too high financing costs for real businesses, the debt covers the operational profit, the credit risks of banks are high; the property prices declined, the bankruptcy risk of enterprises and banks increased and induced systematic financial risk [11]. China adopted a gradual path, which first opens the money market and inter- est rate controls for the bonds market, then realizes the marketization of deposit and loan interest rates progressively according to the experience of interest-rate reforms in developed countries. China set up a unified interbank borrowing market among banks in 1996, which allows both sides to decide on the rate of the interbank borrowing market among banks independently based on the supply and demand situation of the funds. This is also a symbol of starting the interest rate liberalization reform. At the same time, the Ministry of Finance released market-oriented bonds success- fully depending on the Stock Exchange and adopting numerous issuance meth- ods. PBC activated the repurchasing business of bonds among banks and liberal- ized the control of existing bonds’ transaction prices and repo rates on the na- tional unified interbank borrowing market in 1997 [12]. The China Development Bank sold financial bonds by public bidding firstly relying on the bond issue system of PBC in 1998, the Export-Import Bank of China sold financial bonds publicly in a similar way, policy banks and other financial institutes selling bonds publicly by market-oriented interest rate became a new financing way [13]. The reform process of China’s interest rate liberalization MOF began to sell national debt publicly by interest rate bidding in 1999. In 2020 the Chinese government started the marketization reform of deposit and loan interest rates, and in 2004 it canceled the controlling measures of deposit and loan interest rates of all foreign currencies [14]. PBC normally regarded the interbank interest rate as the benchmark interest rate of the market among banks in Shanghai in 2007 with 10 years of explora- tion and development, which propelled the further interest rate liberalization reform whose most basic symbol is the deposit and loan interest rate liberaliza- tion [15]. The control of local currency loan interest rates was fully liberalized in 2013, which basically realized the reform target of loan interest rates liberaliza- tion [10]. PBC controlled the upper limit of the fluctuation in 2015, which marks our country realized the reform target of interest rate liberalization [16]. PBC planned to totally open the limit of broker-dealer shareholding within 3 years in 2018. China took measures to deepen liberalizing finance and the mod- ern service industry, which advanced removing limitations of foreign investment proportion of securities, futures, and life insurance below 51% in 2020. Then China fully opened the financial market in 2020; there is no limitation to the shareholding ratio of foreign investment to security companies, funds, and fu- tures [2, 17]. China allowed investors of qualified foreign institutions, including foreign central banks, sovereign funds, commercial banks, and pension funds, to invest in securities in China’s foreign exchange market in June 2022. Overseas financial institutions can trade securities, investing derivatives, and other finan- cial instruments permitted by PBC and the China Securities Regulatory Com- 138 Song Yuxuan, Gao Yan. Interest rate liberalization and debt risk in China mission from 30 June 2022, the target of interest rate liberalization of the finan- cial market and opening the securities market completely in China was reached. Evaluation variables and data selection of China’s interest rate liberal- ization China’s interest rate marketization reform specifically includes interest rate liberalization process of the money market, bond market, and deposits and loans of financial institutions. Chinese scholars Tao Xionghua and Chen Mingjue chose statistical indicators such as the floating range and scope of interest rates, the deci- sion form of interest rate, the size of the real interest rate, and so on to measure the interest rate marketization, but the scholar Peng Yingqi believes that there is some overlap between the first two indicators, which essentially both reflect the control degree of official interest rates, so she constructed the index of the interest rate floating range based on other scholars’ research results [10, 11]. The benchmark interest rate and the floating range of interest rates determine the interest rate marketization degree according to relevant scholars’ research and the practical experience of China’s financial market, so the liberalization of the real interest rate and the interest rate floating range are selected as the re- search variable in the research process of measuring the interest rate marketiza- tion level in China. If the formation of the benchmark interest rate depends en- tirely on the supply and demand sides of funds in the market, the real interest rate will be consistent with the benchmark interest rate in businesses of China’s financial market. The larger the range of interest rate fluctuations the financial regulatory authorities allow and the fewer control measures they take, the higher the marketization degree of interest rates will be. The interest rate classification in financial markets of PBC was used in this article in order to effectively measure the research variables, and the control or determination level to different interest rate fluctuations of PBC was analyzed according to its control of interest rate fluctuations. Based on factors such as changes in interest rates and real interest rates in China’s money market, bond market, and interbank market, this article selected the data of China’s interest rate index changes from 2000 to 2020 for analysis, and the research data is from the China Financial Yearbook, the database of the National Data, and the database of PBC [17]. Research methods and models In this article, the index of real interest rate level and the fluctuation of inter- est rate are selected to calculate the selected research variables. The method of the fuzzy comprehensive evaluation is selected to easily quantify the market- oriented level of interest rate represented by the annual real interest rate, and the difference between the interest rate of a one-year deposit and the inflation rate is used as the real interest rate level [18, 19]. So we can get specific mathematical expressions: 139 Финансы / Finance IRit = INt −Pt (1) IRLt = IRt−IRmin IRmax−IRmin (2) (1) (2) (1) (2) (1) (2) IRit is the real interest rate in t year; INt is the nominal interest rate in t year, see the interest rate expression by the one-year deposit; Pt is the inflation rate in t year; IRLt is the level index of the real interest rate in t year; IRmax indicates the maximum level of real interest rates during the 1-t year; IRmin indicates the minimum level of real interest rates during the 1-t year; IRmax −IRmin indi- cates the level range of real interest rates during the 1-t period. Since China’s reform process of interest rate liberalization has been divided into interest rate reform of the money market, bond market, and financial institu- tions, the impact of bank deposits and loans on the supply and demand of funds in China’s financial system exceeds that of other self-financial markets. There- fore, in the research process, the CNY deposit and loan market is assigned 3, the upper and lower limits of the flexibility in the market of deposit and loan interest rates are assigned 1.5, and other financial submarkets are assigned 1, and the weights of each financial market segment are assigned as shown in Table 1. Table 1. Variable and variable assignment list Name Money market Bond market Financial institution Interbank Lending Bill Discounting Bond Buybacks Short-Term Treas- ury Bills Enterprise Bond Financial Bond Long-Term Treas- ury Bills Foreign Currency Deposit Foreign Currency Deposit Foreign Currency Loan RMB Deposit Segment financial markets (Lr) L1 L2 L3 L4 L5 L6 L7 L8 L9 L10 L11 Initial assignment 1 1 1 1 1 1 1 1 1 3 3 Weighting (Wr) 1/15 1/15 1/15 1/15 1/15 1/15 1/15 1/15 1/15 1/15 1/15 Source: databases of PBC and national data. Table 1. Research methods and models Variable and variable assignment list Name Money market Bond market Financial institution Interbank Lending Bill Discounting Bond Buybacks Short-Term Treas- ury Bills Enterprise Bond Financial Bond Long-Term Treas- ury Bills Foreign Currency Deposit Foreign Currency Deposit Foreign Currency Loan RMB Deposit Segment financial markets (Lr) L1 L2 L3 L4 L5 L6 L7 L8 L9 L10 L11 Initial assignment 1 1 1 1 1 1 1 1 1 3 3 Weighting (Wr) 1/15 1/15 1/15 1/15 1/15 1/15 1/15 1/15 1/15 1/15 1/15 Source: databases of PBC and national data. Table 1. Variable and variable assignment list 140 Song Yuxuan, Gao Yan. Interest rate liberalization and debt risk in China Based on Table 1, the natural exponential function is constructed as a mathe- matical expression to measure the floating level index of the interest rate of the submarket. Lrt = 1 2 (3) IRLft = ∑ LrtWr 11 r=1 (4) (3) (3) (4) 2 IRLft = ∑ LrtWr 11 r=1 (4) (4) Lr is the r financial market segment; Lrt – Lr is the free-floating index of year t; uprt – limrt– Lr is the upper limit of floating of interest rate; lowrt − limrt– Lr is the lower limit of floating of interest rate; IRLft is the floating in- dex of interest rate in year t; Wi – Lr’s weighting. From equation (3), it can be seen that when the upper and lower fluctuation ranges of interest rates in the market segment are both 0, the floating range in- dex of interest rates in the market segment is 0. When the range of interest rate fluctuations in the market segment is infinite, the interest rate fluctuation index of the segment is 1. From equation (4), it can be seen that the range of interest rate floating range index in year t is [0.1]. The real interest rate level index and the interest rate fluctuation range index together determine the interest rate market-oriented composite index. The real interest rate level index reflects the degree to which the benchmark interest rate level matches the marketization, and has a positive correlation with the degree of interest rate marketization. Research methods and models At the same time, the interest rate floating index reflects the degree to which the Chinese People’s Bank of China restricts interest rate changes, the higher the interest rate floating index, the higher the degree of interest rate marketization, and the interest rate floating index is also positively correlated with the level of interest rate marketization. In order to maximize the information contained in the index, we use the sim- ple averaging method to synthesize the two indices as an interest rate market- oriented composite index that measures the market-oriented interest rate, so that we can obtain an interest rate market-oriented composite index with an annual value between 0 and 1. Thus, a calculated model of the Synthetic Interest Rate Marketization Composite Index (IRLt) can be obtained (5): IRLst = 1 2 (IRLt + IRLft) (5) (5) IRLst is the composite index of interest rate marketization in year t; IRLt is the level index of the real interest rate in year t; IRLft is the floating index of interest rate in year t. IRLst is the composite index of interest rate marketization in year t; IRLt is the level index of the real interest rate in year t; IRLft is the floating index of interest rate in year t. At the same time, it is assumed that the degree of complete marketization of interest rates is 1, the marketization degree of fully controlling interest rates is 0, and the range of values for changes in index values is between 0 and 1. Taking into account the influence of monetary policy changes, the benchmark interest rate will be adjusted within one year, and, in the research process, the bench- mark interest rate after the last adjustment of the current year is used as the one- year deposit rate to evaluate the real interest rate level index. 141 Финансы / Finance Empirical analysis on results of the level of interest rate marketization in China According to the data released by PBC and financial departments, we calcu- lated the inflation rate of the one-year interest rate of fixed deposits in China from 2002 to 2021, and calculated the real interest rate and the level index of real interest rate on this basis, as shown in Table 2. Obviously, even though there are fluctuations in the level of China’s interest rate over the past 20 years, on the whole, it has kept transitioning stably, and China’s interest rates and real interest rates have maintained a slow decline since 2011. Table 2. The fluctuations of the level index and influencing factors of China’s real interest rate from 2002 to 2021 Year One-Year Interest Rate of Fixed Deposit (%) Inflation Rate (%) Real Interest Rate (%) Level Index of Real Interest Rate 2002 1.980 -0.8 2.78 0.7813 2003 1.980 1.2 0.78 0.5383 2004 2.250 3.9 -1.65 0.2430 2005 2.250 1.8 0.45 0.4982 2006 2.520 1.5 1.02 0.5674 2007 4.140 4.8 -0.66 0.3633 2008 2.250 5.9 -3.65 0.0000 2009 2.250 -0.7 2.95 0.8019 2010 2.750 3.3 -0.55 0.3737 2011 3.500 5.4 -1.90 0.2126 2012 3.000 2.6 0.40 0.4921 2013 3.000 2.6 0.40 0.4921 2014 2.750 2.6 0.75 0.5346 2015 1.500 1.4 0.10 0.4557 2016 1.500 2.0 -0.50 0.0677 2017 1.500 1.6 -0.10 0.4121 2018 1.500 2.1 -0.60 0.4743 2019 1.300 2.9 -1.60 0.3188 2020 1.500 2.5 -1.00 0.9565 2021 1.500 0.9 0.60 0.6610 Source: databases of PBC and national data Table 2. The fluctuations of the level index and influencing factors of China’s real interest rate from 2002 to 2021 Table 3 shows the fluctuations of the floating index of interest rates influ- enced together by the money market, bond market, and banking and financial institutions, we can see that there are only slight fluctuations in Interest rates of China’s money and bond markets from 2002 to 2021, and the floating index of interest rate was mainly affected by fluctuations in interest rates at banking insti- tutions. This also indicates that the main factors influencing the marketization of interest rates in China are banks and other financial institutions, and China’s floating index of interest rate has been significantly improved, which indicates that regulators and monetary authorities allow financial markets to price funds to 142 Song Yuxuan, Gao Yan. Interest rate liberalization and debt risk in China a higher degree through market competition. Empirical analysis on results of the level of interest rate marketization in China The study found that interest rate control is a key factor for non-state-owned enterprises to face problems that financing was inaccessible and unaffordable which led to a large number of high-quality credit funds flowing into state- owned listed enterprises, mainly because banks and other financial institutions lack independent pricing power under interest rate control which is easy to form 143 Финансы / Finance systemic risks, so there will be more funds injected into state-owned enterprises or large enterprises with high protection capacity which will aggravate the de- gree of resource misallocation in the financial market, and it is detrimental to the healthy development of the market [3]. Table 4. Composite index of China’s interest rate marketization from 2002 to 2021 Year Level Index of Real Interest Rate Floating Range Index of Interest Rate Composite Index of Interest Rate Mar- ketization 2002 0.781 0.506 0.644 2003 0.538 0.522 0.530 2004 0.243 0.732 0.488 2005 0.498 0.732 0.615 2006 0.567 0.735 0.651 2007 0.363 0.735 0.549 2008 0.000 0.736 0.368 2009 0.802 0.737 0.769 2010 0.374 0.737 0.556 2011 0.213 0.742 0.477 2012 0.492 0.742 0.617 2013 0.492 0.872 0.682 2014 0.535 0.873 0.704 2015 0.456 0.970 0.713 2016 0.068 0.970 0.519 2017 0.412 0.971 0.691 2018 0.474 0.975 0.725 2019 0.319 0.976 0.648 2020 0.956 0.979 0.968 2021 0.661 0.979 0.820 Source: Compiled by the author. Table 4. Composite index of China’s interest rate marketization from 2002 to 2021 The reform of the interest rate marketization is essential to leave the capital allocation right and pricing power in the financial market to be determined by the supply and demand of funds in the market, that is to say, the enterprise with better profitability capacity and better quality disclosed by accounting infor- mation will be better supported by financial institutions after putting forward financing demands, which is specifically manifested in lower financing costs and a more reasonable structure of financing period. Thus, interest rate market- ization reform can have a better guiding effect on the financial market and inject more high-quality funds into high-quality enterprises with more growth to better allocate fund [15, 20]. Empirical analysis on results of the level of interest rate marketization in China Since 2015, the floating index of interest rate in China’s financial market has remained above 0.97, and the gov- ernment, which affects the financial market mainly by relying on market- oriented measures, has continuously weakened its administrative intervention in the financial market. Table 3. Changes in China’s floating index and influencing factors of interest rate from 2002 to 2021 Year Money Market Bond Market Banking Institu- tions Floating Index of Interest Rate 2002 0.231 0.171 0.104 0.506 2003 0.231 0.171 0.120 0.522 2004 0.231 0.171 0.330 0.732 2005 0.231 0.171 0.330 0.732 2006 0.232 0.171 0.331 0.735 2007 0.232 0.171 0.331 0.735 2008 0.232 0.171 0.332 0.736 2009 0.233 0.171 0.332 0.737 2010 0.233 0.172 0.332 0.737 2011 0.233 0.172 0.337 0.742 2012 0.233 0.172 0.337 0.742 2013 0.267 0.172 0.433 0.872 2014 0.267 0.172 0.434 0.873 2015 0.267 0.172 0.531 0.970 2016 0.267 0.172 0.531 0.970 2017 0.267 0.173 0.531 0.971 2018 0.268 0.174 0.533 0.975 2019 0.268 0.175 0.533 0.976 2020 0.269 0.176 0.534 0.979 2021 0.269 0.176 0.534 0.979 Source of data: Compiled by the author. Table 3. Changes in China’s floating index and influencing factors of interest rate from 2002 to 2021 Although China deregulated the floating of deposit and loan interest rates in 2015, the degree of interest rate liberalization has not increased significantly from the measurement results of it, mainly because China’s current interest rate implements a dual-track system, and the benchmark interest rate of deposits and loans is still determined by PBC. Table 4 shows the changes in the composite index of China’s interest rate marketization from 2002 to 2021, under the influence of real interest rates and interest rate floating changes, the changing trend of the composite index of the interest rate marketization is not fixed, but fluctuating, and the fluctuation range is increasing persistently. Empirical analysis on results of the level of interest rate marketization in China In the reform of the interest rate marketization, banks and other financial in- stitutions will use more capital for high-risk and high-yield enterprises under the pressure of fierce competition, in the meantime, after the lower limit of loan in- terest rates is liberalized under interest rate liberalization, the attraction of state- owned enterprises or large enterprises will be significantly reduced to banks and other financial institutions, which will prompt financial institutions to inject 144 Song Yuxuan, Gao Yan. Interest rate liberalization and debt risk in China more capital into non-state-owned enterprises or small and medium-sized enter- prises, and provide enterprises with more flexible and independent financing products through product innovation and other means, and the financing ability of enterprises will be significantly improved under the joint action of many as- pects [21]. The continuous advancement of the reform of interest rate marketization puts forward higher standards and requirements for enterprises’ accounting infor- mation disclosure, capital allocation efficiency, and many other aspects, and effectively alleviates the information asymmetry in the market, and financial institutions can formulate more reasonable financing strategies based on high- quality accounting information to avoid uncertainty and high risks effectively, and then inject more high-quality funds into entity enterprises [16, 22]. As an important part of deepening financial reform, interest rate marketiza- tion can effectively improve the unbalanced credit and other problems formed under interest rate control to make various enterprises obtain a relatively fair financing environment. At the same time, it can also integrate more high-quality private funds into the entire financial market, provide more adequate financing support for enterprises, and introduce a more effective mechanism for debt gov- ernance. We can see from Table 1 that the main influencing factor of the interest rate marketization index fluctuation is the fluctuation of the real interest rate level index, which shows that, since the benchmark interest rate of deposits and loans is still controlled by PBC, the level of the real interest rate cannot really play the financial market regulation effect in economic development, that is also the rea- son why China fully liberalizes the financial market and gives foreign-funded financial enterprises a larger operational scope, and Chinese government intensi- fies the competition in its own financial market and reduces capital costs by opening up more fields to introduce foreign financial enterprises [4, 23]. Figure 1. Changes in China’s real interest rate and interest rate marketization At present, a large number of personal funds in China’s capital market want capital preservation and appreciation, seeking reliable targets, which will deviate asset prices. The motivations of risk avoidance or speculation aggravate the de- gree of asset bubbles if external risks have a greater impact on them. In the case of the high debt ratio, asset prices fluctuate sharply and decline, it is easy to pro- duce vicious liquidity crisis and credit risk. What is worse, it will evolve into systemic risk and resulting in a sustainable economic recession and form the financial crisis [24]. In the reform process of the financial market, Chinese society has also com- pleted the transformation of the economy from household savings to debt, and both household debts and the debts of market entities such as governments, en- terprises, and so on have increased significantly. As economic growth expecta- tions and the speed of GDP growth are declining, the possibility of shifting from inflation to deflation increases, systemic debt increases, and systematic decline of asset prices will inevitably accelerate the Fei Xue style “debt-deflation” vi- cious circle of substantial debt burden. Therefore, the marketization of interest rates cannot be deregulated alone, and it is necessary to establish a market-oriented interest rate system supporting the financial system to effectively affect the economic expectations of market entities, which directly determines the investment willingness, financing ability, and debt repayment ability of market entities in the economic cycle [15, 17]. From a macro point of view, China’s debt leverage is higher, asset bubbles are also turning bigger, and prospects of economic profitability are bleak in the expectation of a long-term decline in the growth speed. Enterprises and residents will inevitably reduce their debt burden as much as possible under the expecta- tion that the asset bubble burst will make a rapid increase in debt, take various measures to reduce leverage. The more direct and effective measure is to sell assets, but large-scale asset sales will cause market panic, resulting in rapid de- preciation of assets and forming a debt crisis. Empirical analysis on results of the level of interest rate marketization in China Composite index 0 0,2 0,4 0,6 0,8 1 1,220022003200420052006200720082009201020112012201320142015201620172018201920202021 Interest rate market index Actual interest rate Figure 1. Composite index 145 Финансы / Finance Changes in China’s real interest rate and interest rate marketization Referring to the successful experience of foreign countries, in order to strengthen controlling the debt market’s risks and reduce the risks arising from the marketization of interest rates, we must first choose the correct reform path, gradually relax the control degree over interbank lending rates, bill interest rates, and bond interest rates mostly under the control of the monetary management authorities to take the lead in realizing the marketization of money market inter- est rates [7, 19]. We must guide and regulate the market behavior of commercial banks and other financial institutions, first liberalize loans and then open depos- its, financial institutions can independently set the prices of the cost of funds under the premise that costs can be controlled on a priority basis to enhance pricing capabilities and risk identification capabilities, to avoid chaos in the fi- nancial market and improve bank profitability [21]. Relating to deposits in fi- nancial institutions, the interest rate control of long-term deposits should be opened first, and then the interest rate fluctuations of short-term deposits should 146 Song Yuxuan, Gao Yan. Interest rate liberalization and debt risk in China be relaxed to strengthen the monitoring and guidance of deposit liquidity. Con- sidering capital quota management, liquidity control is gradually relaxed to real- ize the free regulation and control of the market, smoothly realize the market- oriented transition period, reasonably control the mismatch risk of debt alloca- tion of market entities, and optimize the capital allocation. In addition, the progress of interest rate marketization should also coordinate the relationship with economic structural adjustment, especially the interrela- tionship between capital account marketization, exchange rate marketization, and the degree of financial business liberalization, and balance international cap- ital liquidity to avoid the impact of short-term arbitrage transactions of capital to the domestic financial system [4]. Discussion The volume of nonperforming loans in China has grown rapidly since 2015, and the debt default and liquidity risks of small and medium-sized commercial banks are at a high level, which has increased the default risk of banks and other financial institutions in the debt market [17]. Changes in the Sino-US trade war and other external economic environments have led to a rapid decline in profits or even deficits in some enterprises and industries, resulting in difficulties in their capital flow and a reduction in their debt repayment ability and willingness to repay debts. These companies have obtained large amounts of loans from banks in the early stage, or issued corporate bonds in the bond market, so the risk of debt default of the real economy will also be transmitted to the financial market. China’s financial market has successively appeared with the bankruptcy of small and medium-sized banks, the inability to pay corporate bonds on schedule, and other problems since 2019, due to the imperfect governing structure of cor- porations, subjecting to shareholders and other constraints. Independent opera- tion ability needs to be strengthened, no matter the insurance, fund trust enter- prises, or commercial banks have both management and liquidity risks [21]. At the same time, the increase in financial debt and lacking supervision of off- balance-sheet assets to banks by regulatory authorities have a negative impact on the risk control of China’s debt market, implementing the interest rate mar- ketization in the case of imperfect deposit insurance systems; immediately aban- don the control to capital liquidity, driven by the short-term profit, financial in- stitutions will carry out higher risk operational activities, so when the capital account is opened, it is necessary to strengthen supervision capabilities to pre- vent excessive international capital speculation from causing the exchange rate to get out of control. When Chinese commercial banks are resolving non-performing assets, they use measures of the debt-to-equity swap, which has helped many enterprises avoid the risk of debt default, but this method cannot substantially change the enterprises’ benefit and does not meet the shareholders’ long-term goals. Discussion The government’s administrative intervention is large in the implementation process, 147 Финансы / Finance which does not conform to the law of enterprise market operation making it dif- ficult to withdraw from enterprise equity in the later stage, it may increase the systemic risk of the debt market after implementing interest rate marketization [15, 25]. Moreover, this debt conversion method is generally carried out through government-controlled companies of asset management whose loss is ultimately undertaken by the government’s finances, and it remains to be observed whether state-owned enterprises can fundamentally solve the debt crisis by reorganizing enterprises with debt crises. In fact, the interest rate marketization process is also the deleveraging pro- cess of Chinese enterprises, and a successful deleveraging method requires the implementation of reasonable monetary easing, a steady rise in asset prices, a gradual decline in leverage ratios, a steady growth of the economy, and the na- tional economic growth rate higher than the growth of the interest rate. In the increasingly fierce market competition, financing ability has become the key for enterprises to implement a series of operation and management strat- egies such as innovation strategy, diversification strategy, and so on. But im- proving financing ability has become the key to enterprises, which generally face different degrees of financing constraints, in the critical period of economic system transformation. Enterprises are faced with a “dual-track system”, a high marketization in money market transmissions and a relatively low marketization of deposit and loan benchmark interest rates such as reverse repo interest rates and other situations under the environment of credit rationing and interest rate re- strictions in China, which makes a double distortion of regulating prices and quan- tities. In this context, the reform of interest rate marketization, which can effec- tively improve the dual-track system existing in the current financial market, pro- mote the optimal allocation of funds in the entire financial market, and improve the transmission channels’ efficiency of monetary policy interest rate, and have an impact on the financing ability of micro-level enterprises, is particularly crucial under this segmentation of the financial market [17]. Conclusion Direct debt, indirect debt, and economic system debt of China’s local gov- ernments rise continuously, meanwhile, China’s rapid economic development depends on the government’s stimulus policies excessively, which shows that the debt risk of local governments has a strong spatial spillover effect, and such a high scale of local governments’ debt will have a direct impact on the financ- ing of listed enterprises in them. In order to obtain better GDP, local governments will attract more enterprises to settle in through large-scale borrowing and provide more financing opportuni- ties for their financial market. However, when the scale of debt is too large or the debt ratio is too high, it will have a strong shock and impact on the stability of their financial market, which made banks and other financial institutions rela- tively skeptical about the repayment ability of their enterprises and will set high- er constraints to restrict the financing of enterprises. That is to say, in areas with 148 Song Yuxuan, Gao Yan. Interest rate liberalization and debt risk in China relatively high debt risks, because of high debt ratios, the attractiveness to the financial market and even the entire capital market will be significantly reduced, and governments cannot stimulate banks and other financial institutions to ac- tively participate in the development of real enterprises through policy incen- tives. Since the allocation efficiency of considerable high-quality resources is too low, which will aggravate the information asymmetry in the entire market, en- terprises lack funds to invest in technological innovation and other high-risk items, while, due to the profit-seeking characteristics of banks and other finan- cial institutions, more funds will be invested in areas with relatively low debt risk, the financing ability of enterprises is greatly suppressed under the spillover effect of local governments’ debt risk and lacking guidance ability. g g g y Implementing any policy and its economic consequences will be affected by the environment, and so does the interest rate liberalization, which is an important component of financial market reform. The debt risk of local gov- ernments is raising, and the degree of systemic risk banks and other financial institutions faced is increasing significantly, they will protect their own rights and interests by shrinking the monetary base, reducing the scale of loans, and shortening the loan term. Conclusion Enterprises face inaccessible and unaffordable fi- nancing, and other difficulties; in this case, promoting interest rate marketiza- tion makes interest rates determined by the supply and demand of funds in the financial market, that is to say, commercial banks and other financial institu- tions will invest more high-quality funds in high-risk and high-return enter- prises, which have effectively improved the financing constraints of enterpris- es, and this kind of improvement effect is more significant in small and medi- um-sized enterprises [7]. On one hand, interest rate marketization can effectively curb excessive in- vestment and other inefficient investment phenomena by improving the corpora- tions’ governing effect to retain more free cash flow for enterprises, which means that there is a significant positive correlation between interest rate mar- ketization and cash holdings of corporations [12]. When the debt risk of local governments is relatively high, the high cash holdings formed by the interest rate marketization make enterprises face a lower degree of financing constraints and have better debt repayment ability, and financial institutions such as com- mercial banks will also cooperate with enterprises by reducing interest rates, and then the financing ability of enterprises will be improved. On the other hand, enterprises with higher financing ability in the interest rate market-oriented environment often reflect accounting information with higher quality, which means that interest rate marketization will put forward higher standards and requirements for the disclosure quality of enterprises’ ac- counting information, and that will be transmitted to the entire market through the transmission effect of the interest rate [2]. It will be favored by commercial banks and other financial institutions, especially in areas where local govern- ments have relatively high debt risk. As a result, its financing capacity will be enhanced. 149 Финансы / Finance The debt risk of local governments has a spatial spillover effect. While im- proving the degree of freedom in accounting and finance of local governments, it is also necessary to formulate a more normative management and assessment mechanism, take multiple measures to effectively reduce debt risks of local gov- ernments, and fully develop governments’ macro-adjustment in the entire mar- ket deepening the market-oriented reform of interest rates to provide the most solid guarantee for improving the financial market and expanding enterprises’ financing channels. References 1. Wang, T. (2020) The Current Policy Choice to Reduce Macro Leverage. 1st ed. eijing: CITIC Press Group. 2. He, D. & Wang, X. (2020) Has The Risk of Default on Local Government Debt Been Reduced? Journal of Public Finance. 2. pp. 9–26. 3. Wang Zhengyang. (2017) Research on Market Risk and Mitigation Countermeasures of China’s Macro Debt. Nanjing: Nanjing University. 4. Li, Zh. & Du, S. (2019) Interest Rate Liberalization and Corporate Cash Holding. Finance and Economics. 11. pp. 4–14. 5. Ren, Z. (2016) Status, Response, Risks and Investment Opportunities of Deleveraging. Shanghai: Guotai Junan Securities Research Report. 6. Xiang, H., Li, L. & Chen, X. (2018) Loose Stimulus Policy, Interest Rate Liberalizatio nd Bank Risk-Taking Channels. Journal of Financial Economics. 3. pp. 3–14. 7. Liang Si. (2018) Research on the Transmission Mechanism of Monetary Policy Interest 7. Liang Si. (2018) Research on the Transmission Mechanism of Monetary Policy Interest Rate Under the Background of Interest Rate Liberalization. Journal of Financial Regulatory Research. 2. pp. 13–19. Rate Under the Background of Interest Rate Liberalization. Journal of Financial Regulator Research. 2. pp. 13–19. 8. Xiao, Zh. & Chen, W. (2017) Monetary Policy, Interest Rate Transmission, and Small and Medium-Sized Enterprises Financing Costs. Economic Review. 5. pp. 79–90. 9. Huang, W., Zhao, B. & Liu, Ch. (2019) The Impact of Financing Transactions on the Actual External Financing Ability of Enterprises – An Empirical Study Based on Shanghai and Shenzhen A-Shares. Financial Forum. 2. pp. 56–68. 10. Zhong, Y. & Zhang, L. (2019) Has the State Audit Reduced the Risk of Loca Government Debt? Journal of Nanjing Audit University. 3. pp. 1–10. 11. Li, Q. (2016) Monetary Policy, Interest Rate Marketization and Credit Resource Misallocation: A Balance Model of Credit Market Capital Supply Based on Unequal Market Status. Journal of Financial Economics. 2. pp. 3–13. 12. Zhan, M., Wang, X. & Ying, C. (2013) Interest Rate Control, Bank Credit Rationing Behavior Variation and Financing Constraints of Listed Companies. Economics. 12. pp. 1255–1276. 13. Qu, Sh. (2014) The Role of Government in Interest Rate Liberalization. China Finance. 8. pp. 83–85. 14. Hu G. & Hong, L. (2020) Research on the Relationship Between Corporate Financial Cash Flexibility and Financing Ability. Finance and Accounting Communications. 6. pp. 75–79. 15. Yi, G. (2009) The Process of Interest Rate Liberalization in the Past 30 Years of China’s Reform and Opening up. Journal of Financial Research. 1. pp. Conclusion The reform of interest rate liberalization will create the illu- sion of rising financing costs in a short period, so relevant departments of gov- ernment should actively guide and relax the regulation of local finance to let market mechanisms make a difference to the fullest extent. Information about the authors: Song Yuxuan, Saint Petersburg State University (Saint Petersburg, Russian Federation E-mail: songyuxuan_0225@163.com Song Yuxuan, Saint Petersburg State University (Saint Petersburg, Russian Federation). E-mail: songyuxuan_0225@163.com Gao Yan, associate professor, Jining Medical University (Jining, China). E-mail: 1056022128@qq.com Gao Yan, associate professor, Jining Medical University (Jining, China). E-mail: 1056022128@qq.com References 1–14. 16. Ma, X. (2017) The Impact of Green Credit on The Profitability of Commercial Banks in China: Based on the Panel Data Analysis of 14 Listed Banks. West China Finance. 3. pp. 28–33. 150 Song Yuxuan, Gao Yan. Interest rate liberalization and debt risk in China 17. Wang, Ch. & Han, F. (2021) Construction of Intelligent Financial System of Commercial Banks. Business Accounting. 20. pp. 23–26. 18. Liu, L., Yu, J. & Yang, J. (2017) Is Competition a Double-Edged Sword for Bank Credit Structure Adjustment? – Micro Evidence of China’s Interest Rate Liberalization Process. Economic Research Journal. 5. pp. 131–145. 19. Yang, P., Zhou, X. & Yang, W. (2019) Collaborative Promotion of Green Innovation Between Green Finance and Taxation and Green Finance: A Case Study of Hangzhou. Finance and Finance. 3. pp. 17–22. 20. Xu, L. (2018) Comprehensive Risk Study of Market-Oriented Green Credit. Wuhan: Wuhan University of Technology. 21. Zhang, W. & Li, L. (2019) Interest Rate Marketization, Price Competition, an C i l B k ’ Ri k T ki D i GMM T B d P l D f 16 Li Commercial Banks’ Risk Taking – Dynamic GMM Test Based on Panel Data of 16 Listed Commercial Banks. Journal of Shandong University of Finance and Economics. 4. pp. 50–60. 22. Wang, J. (2012) Empirical Analysis of the Profitability of Commercial Banks in the Process of Interest Rate Marketization in China. Journal of Finance and Economics. 8. pp. 74–77. 23. Zhang, H. (2015) Will the Marketization of Loan Interest Rates Increase the Risk of Commercial Banks? – Based on the Perspective of Non-Interest Income Business Expansion and Excessive Credit Growth. Finance and Economics. 11. pp. 56–61. 23. Zhang, H. (2015) Will the Marketization of Loan Interest Rates Increase the Risk of Commercial Banks? – Based on the Perspective of Non-Interest Income Business Expansion and Excessive Credit Growth. Finance and Economics. 11. pp. 56–61. 24. Xiang, H. (2019) Internal Control, Inefficient Investment and Internet Financing. Finance and Accounting Newsletter. 7. pp. 120–123. 24. Xiang, H. (2019) Internal Control, Inefficient Investment and Internet Financing. Finance and Accounting Newsletter. 7. pp. 120–123. g pp 25. Zhang, L. & Lian, Y. (2019) Green Credit, Bank Heterogeneity and Bank Financial Performance. Journal of Financial Regulatory Research. 2. pp. 43–61. g pp 25. Zhang, L. & Lian, Y. (2019) Green Credit, Bank Heterogeneity and Bank Financial Performance. Journal of Financial Regulatory Research. 2. Song Yuxuan, Gao Yan. Interest rate liberalization and debt risk in China 17. Wang, Ch. & Han, F. (2021) Construction of Intelligent Financial System of Commercial Banks. Business Accounting. 20. pp. 23–26. g pp 18. Liu, L., Yu, J. & Yang, J. (2017) Is Competition a Double-Edged Sword for Bank Credit Structure Adjustment? – Micro Evidence of China’s Interest Rate Liberalization Process. Economic Research Journal. 5. pp. 131–145. References pp. 43–61. The article was submitted 02.01.2023; approved after reviewing 13.02.2023; accepted for publication 21.02.2023. 17. Wang, Ch. & Han, F. (2021) Construction of Intelligent Financial System of Commercial Banks. Business Accounting. 20. pp. 23–26. 18. Liu, L., Yu, J. & Yang, J. (2017) Is Competition a Double-Edged Sword for Bank Credit Structure Adjustment? – Micro Evidence of China’s Interest Rate Liberalization Process. Economic Research Journal. 5. pp. 131–145. 19. Yang, P., Zhou, X. & Yang, W. (2019) Collaborative Promotion of Green Innovation Between Green Finance and Taxation and Green Finance: A Case Study of Hangzhou. Finance and Finance. 3. pp. 17–22. 20. Xu, L. (2018) Comprehensive Risk Study of Market-Oriented Green Credit. Wuhan: Wuhan University of Technology. 21. Zhang, W. & Li, L. (2019) Interest Rate Marketization, Price Competition, and Commercial Banks’ Risk Taking – Dynamic GMM Test Based on Panel Data of 16 Listed Commercial Banks. Journal of Shandong University of Finance and Economics. 4. pp. 50–60. 22. Wang, J. (2012) Empirical Analysis of the Profitability of Commercial Banks in the Process of Interest Rate Marketization in China. Journal of Finance and Economics. 8. pp. 74–77. 23. Zhang, H. (2015) Will the Marketization of Loan Interest Rates Increase the Risk of Commercial Banks? – Based on the Perspective of Non-Interest Income Business Expansion and Excessive Credit Growth. Finance and Economics. 11. pp. 56–61. 24. Xiang, H. (2019) Internal Control, Inefficient Investment and Internet Financing. Finance and Accounting Newsletter. 7. pp. 120–123. 25. Zhang, L. & Lian, Y. (2019) Green Credit, Bank Heterogeneity and Bank Financial Performance. Journal of Financial Regulatory Research. 2. pp. 43–61. The authors declare no conflicts of interests. 151
https://openalex.org/W3127302365	https://www.e3s-conferences.org/articles/e3sconf/pdf/2021/12/e3sconf_icersd2020_05099.pdf	English	null	Tracing back to the origin--Research on the application Value of Visual Media Design of National Traditional Patterns	E3S web of conferences	2,021	cc-by	3,183	! " # $ % & ' () ! !+, - , # $, ., ! " # $ % & ' () ! !+, - , # $, . , & )/ )/ 0 , , ( , ( ) % 1 2 , , ) , ) +, ! , / % 1 2 ( , 0 , ) ! ! $ , 3 ! 4 ) 0) ! + ( 0 + © The Authors, published by EDP Sciences. This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0 (http://creativecommons.org/licenses/by/4.0/). https://doi.org/10.1051/e3sconf/202123605099 https://doi.org/10.1051/e3sconf/202123605099 E3S Web of Conferences 236, 05099 (2021) ICERSD 2020 ! " # $ # ) + ) ! ! ( ) & + ( ! ! ! ( ) ! & ! ! ( ) , + ! ( 5 ! ! ( $ ( 6 ) ! ( ! % 1 2 ( ! ) - ) ! ), 0 ( ! % 1 2 , + ! 0 ! % 1 2 , & + ! , % 1 2 3 , , ! ) 00 ! , , 7 % 1 2 , 0 ! + + ) 0 ! ! DP Sciences. This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0 ses/by/4 0/) &%& ") (6 + ) & ! + ! , + 0 ! 0! 0) 0 ! ! ) ! 5 ! , + ( ! (6 , ) ( ! ! ( , ! + ! +3 & ! ( 0 ( ! (6 & ), , (, / ( " ( ! ), ( + * ( 0 !! ( ! ) ) + + + ! , 0 + % 1 2 , + !% " # # 8 0 ()+ ) !, ()+, , ! ( , ! ! ! $ , ! ! ) ! ( , !! ) , !! , !! + !! ) (, ! 0 / ! ! ! ! ( ), )6 ! + , ( ! (0 ! ! ( ( ) !! ) , , , +( , ( , , & 0! ( ) ! ) & ! ) 0 ! , ! , + 0 ! , + ! ! !! , !! 0! !! ) ( !3 , ) , ( ), https://doi.org/10.1051/e3sconf/202123605099 E3S Web of Conferences 236, 05099 (2021) ICERSD 2020 E3S Web of Conferences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eb of Conferences 236, 05099 (2021) ICERSD 2020 https://doi.org/10.1051/e3sconf/202123605099 ' , 0 ! ), 00 ! ) 3 0, 0) (( ) ) ! ), + 0, !! () !! 3 0 , ) 5 ( ! @9 A7 4) B, !) + 4) 0 , 0 ) & + !) + , 00 ! + + , ! , , 3 , $) ), + ! , $ )/ 4) B, ! )/ , (, ' %! , & )0 3 ! $ , ! $, ! ) 0 3 ), + $ () 00 , ! )0 ! 0 3, ! 0 6 , 0 ! + ? 3, 0 ), ( !! , + 0+ & ! " C, - 5 ! & ), # $, - .. > C " D#) @,9.E, , ( ? $ , F , ) F , 1 , ? , G 0 > .( & ! # $ 8( > 8 87 H 87 0 G ) ( ! # $ !3 : 2) )HCB@9@ 5 ! ' , 0 0 ! % 1 2 & ! ' 00 ! < 3 < , $ ! +3 %( ) ! +3 , 0 , + 0+ M "D@E & 8 2( & ! L 1 G! J#K F0 L /'//#/0 ",IDE8) !J#K !! L @ F,1GDE-+ !J#K0 ? ,ADE%+ ! L 1 G!J#K1 * ! >0 ? ,ADE%+ ! L 1 G!J#K1 * ! >0 . " ,2F" ,B? D@E& 5 ! ) L 1 G!J#K . " ,2F" ,B? D@E& 5 ! ) L 1 G!J#K 3 E3S Web of Conferences 236, 05099 (2021) ICERSD 2020 https://doi.org/10.1051/e3sconf/202123605099 3 + ) M "D@E & 8 2( & ! L 1 G! J#K F0 L 4
https://openalex.org/W4319979411	https://cyberleninka.ru/article/n/algoritmy-generatsii-obuchayuschih-mnozhestv-v-sisteme-s-pretsedentnym-vyvodom-na-osnove-situatsiy-primerov/pdf	Russian	null	Алгоритмы генерации обучающих множеств в системе с прецедентным выводом на основе ситуаций-примеров	Programmnye produkty i sistemy	2,022	cc-by	6,862	1 Институт математики и компьютерных наук, Тюменский государственный университет, г. Тюмень, 625003, Россия 1 Институт математики и компьютерных наук, Тюменский государственный университет, г. Тюмень, 625003, Россия В статье рассматривается проблема создания обучающих множеств и их масштабирования в задачах машинного обучения. Предметом исследования является процесс генерации обучающих множеств на основе примеров в целях их аугментации. Для реализации идеи расширения предлагается использовать преобразование имеющихся примеров ситуаций. Преобразование примеров осуществляется на основе известного метода оптимизации – ме- тода покоординатного спуска. Описывается постановка задачи преобразований ситуаций-примеров в терминах введенной модели представлений. Предлагаются алгоритмы, позволяющие из исходного множества ситуаций-примеров, заданных с помощью формальных представлений, получать расширенное множество, которое будет включать в себя ситуации, отвечающие критериям сходства с данными примерами. В статье представлена апробация предложенных алгоритмов при исследовании нейросетей для от- бора ситуаций в системах вывода по прецедентам. Полученные результаты имеют практическую зна- чимость для обучения искусственных нейросетей, используемых в интеллектуальных системах под- держки принятия решений. Предложенные алгоритмы позволяют автоматизировать формирование наборов данных дата-сетов, используя имеющиеся подготовленные и одобренные примеры характер- ных ситуаций и решая задачу преобразований как задачу поиска оптимума целевой функции схожести. Ключевые слова: обучение нейросетей, обучающие данные, case-based reasoning, искусственный интеллект, координатный спуск. Метод вывода решений по прецедентам (Case-Based Reasoning, CBR) – один из извест- ных методов искусственного интеллекта, кото- рый находит применение при разработке си- стем поддержки принятия решений (СППР) в различных областях [1–3]. В СППР на основе CBR при возникновении проблемной ситуации решение для нее ищется в базе знаний, в кото- рой хранятся прецеденты – пары <Ситуация, Решение>, где Ситуация представляет извест- ную из прошлого опыта проблемную ситуа- цию, а Решение – то решение, которое счита- ется рациональным или необходимым в этой ситуации. В качестве решения могут выступать набор управляющих воздействий, план меро- приятий, программа действий персонала для разрешения проблемной ситуации и др. Если такая база знаний, которая еще называется ба- зой (библиотекой) прецедентов, имеется в CBR-системе, то остается найти в ней ситуа- цию, которая наиболее похожа на вновь воз- никшую проблемную ситуацию, чтобы потом выдать пользователю решение из найденной пары. Сравнение и оценка похожести ситуаций является одной из ключевых задач CBR-мето- да. Для ее решения используются разные спо- собы, среди которых наиболее распространен- ным, по-видимому, является применение разнообразных метрик в пространстве пара- метров, которыми описываются ситуации [4]. Вычисление сходства между ситуациями и возможность применения тех или иных метрик определяются их формальным представлени- ем [5, 6]. 4 (35) 2022 4 (35) 2022 Программные продукты и системы / Software & Systems УДК 004.85 Дата подачи статьи: 11.05.22, после доработки: 22.07.22 DOI: 10.15827/0236-235X.140.660-669 2022. Т. 35. № 4. С. 660–669 Алгоритмы генерации обучающих множеств в системе с прецедентным выводом на основе ситуаций-примеров И.Н. Глухих 1, д.т.н., профессор Д.И. Глухих 1, аспирант, d.i.glukhikh@utmn.ru Дата подачи статьи: 11.05.22, после доработки: 22.07.22 2022. Т. 35. № 4. С. 660–669 Задача преобразования ситуаций-примеров Метод преобразований для создания обуча- ющего набора данных на основе примеров со- стоит в следующем. Пусть задано множество примеров ситуаций SIT = {Sitr\|r = 1, …, R} таких, что любые две различные ситуации из этого множества не являются схожими в смысле некоторой функции похожести ситуа- ций Sim (.), то есть Sim (Sitr, Sitp) < Th, где Th – порог, после превышения которого можно го- ворить о схожести ситуаций. Однако этот подход требует значительного объема обучающих примеров с известными ре- акциями, которые должны быть достигнуты обучаемой моделью. В рассматриваемой за- даче такие обучающие наборы данных образу- ются парами сравниваемых ситуаций с мет- ками, в качестве которых выступают количе- ственные оценки их похожести или, в упрощенном случае, качественные оценки вида похож/непохож [7, 9]. Применительно к базе знаний в системе с выводом на прецедентах это означает, что есть R пар <Ситуация (Sitr), Решение (Solr)>, где все решения отличаются друг от друга. Иначе го- воря, признак решения выступает в качестве классификационного признака, разделяющего все множество ситуаций на классы. Поскольку на практике в реальных систе- мах далеко не всегда есть достаточные по объ- ему наборы обучающих данных, формирова- ние таких наборов – обучающих и валидирую- щих дата-сетов становится самостоятельной и актуальной научно-практической задачей. Ее решение в условиях недостатка обучающих данных и при отсутствии возможностей сфор- мировать их по наблюдениям за системой свя- зывают с идеей расширения (аугментации) того небольшого числа обучающих примеров, которые уже имеются [10]. Задача – путем преобразований расширить имеющееся множество SIT новыми элементами так, чтобы для каждого r получить множество SITr = {Sitr, Sit1, Sit2, …}, где все элементы – си- туации, похожие с Sitr в соответствии с задан- ным критерием схожести. Задача – путем преобразований расширить имеющееся множество SIT новыми элементами так, чтобы для каждого r получить множество SITr = {Sitr, Sit1, Sit2, …}, где все элементы – си- туации, похожие с Sitr в соответствии с задан- ным критерием схожести. В данной работе для реализации идеи рас- ширения предлагается использовать преобра- зование имеющихся примеров ситуаций, за- данных в рамках разработанной ранее модели формальных представлений ситуаций [11]. В этой модели ситуация предметной области трактуется как совокупность состояний, в ко- торых находятся элементы сложного объекта и связи между ними. Формальным представле- нием ситуации выступает вектор в простран- стве состояний или, в случае сложного объекта, набор таких векторов (мультивектор ситуа- ции), каждый из которых соответствует своему элементу сложного объекта или связи между ними. 1 Институт математики и компьютерных наук, Тюменский государственный университет, г. Тюмень, 625003, Россия Однако применение метрик и соответству- ющих математических вычислений не всегда может обеспечить точную и детальную оценку сходства с тем, чтобы в базе знаний уверенно выбрать ситуацию с «правильным» решением и не выбрать ситуацию с решением «непра- вильным». Проблема особенно актуальна для комплексных ситуаций, которыми характери- зуются сложные объекты, состоящие из многих неоднородных элементов и отношений между ними. В подобных случаях для описания ситу- аций приходится использовать большое число параметров разного типа с использованием разных шкал измерений (количественных и ка- чественных), разрабатывая локальные метрики сходства [7], агрегирование которых, в свою 660 4 (35) 2022 Программные продукты и системы / Software & Systems ний соответствует свое состояние из множе- ства возможных. На этапе идентификации текущей ситуации распознаются состояния элементов сложного объекта и формируются векторы состояний, компоненты которых при- нимают значение 0 или 1 (при точном распо- знавании состояний) или в интервале от 0 до 1 (при нечетком распознавании состояний) [12]. Сформированный мультивектор ситуации вме- сте с другим мультивектором – ситуации из базы знаний подается на вход нейронной сети, которая вычисляет значение функции схоже- сти Sim, то есть определяет степень сходства между двумя ситуациями [13,14]. очередь, связано с рисками потери информа- ции и неоднозначностью получаемых резуль- татов. Такие ситуации возникают, например, в си- стемах эксплуатации сложных технологических объектов (на крупных производствах, в нефтега- зодобывающей и перерабатывающей промыш- ленности, на предприятиях городской инфра- структуры), когда при принятии решений необ- ходимо учитывать состояния различных подсистем, связи между ними, а также их опера- ционное окружение, внешнюю среду и др. [8]. Перспективным подходом к определению схожести ситуаций является подход на основе машинного обучения и, в частности, искус- ственных нейронных сетей, которые уже пока- зали свою работоспособность в ряде научных исследований [7–9]. Задача преобразования ситуаций-примеров Отдельному компоненту вектора состоя- Введем новый индекс для упрощения за- писи: SITr = {Sitk\|k = 1, …, Rr}, где Sitk  Sitr при k = 1. В общем виде задача преобразований фор- мулируется как следующая оптимизационная задача. Найти такую Sit, что h(Sit) = Sim(Sit, Sit) → max (1) при Sit  Sit. Найти такую Sit, что (1) Ограничение вводится для того, чтобы в процессе решения данной задачи не получить в ответ Sit, полностью совпадающую с Sit, то есть новая ситуация должна отличаться от ис- ходной. 661 4 (35) 2022 Программные продукты и системы / Software & Systems Если Sit – вектор в пространстве состояний, то решение этой задачи состоит в переборе компонентов данного вектора, которые могут принимать значение 0 или 1 (в случае точной классификации) или в интервале от 0 до 1 (в случае нечеткой классификации) [8]. Далее рассматриваем только случай точной класси- фикации. Для нечеткой классификации вместо перебора нулей и единиц будут изменяться зна- чения в соответствующей позиции вектора от 0 до 1 с некоторым шагом . 3. Для j от 1 до M делать // M – число компонентов вектора ситуации в про- странстве состояний { 4. Преобразуем значение j-го компонента век- тора ситуации на противоположное Sitout[j] = \|Sit [j] –1 \| 5. Вычисляем h(Sitout) 6. Если h(Sitout) > Th То 6. Если h(Sitout) > Th То // Включаем Sitout в искомое множество SIT, если новая ситуация удовлетворяет порогу сходства и значение h(Sitout) в множество меток SIM 7. Sitout in SIT и h(Sitout) in SIM 6. Если h(Sitout) > Th То // Включаем Sitout в искомое множество SIT, если новая ситуация удовлетворяет порогу сходства и значение h(Sitout) в множество меток SIM 7. Sitout in SIT и h(Sitout) in SIM 8. Упорядочиваем элементы множества SIT по значению h(Sitout) 8. Упорядочиваем элементы множества SIT по значению h(Sitout) 9. Отбираем из SIT первые V элементов (где V – заданное число искомых элементов) и, добавляя в него начальное преобразуемое Sit, формируем выходное множество SIT – расширенное множе- ство для ситуации Sit и, соответственно, множе- ство меток SIM. 10 Конец Алгоритм 1. Модификацией этого алгоритма становится проверка на шаге 6 не только условия h(Sitout) > Th, но и дополнительных условий- ограничений, например, равенства двух ситуа- ций по заданному подмножеству позиций (компонентов) в сравниваемых векторах состо- яний. 1. Начало 1. Начало 2. Устанавливаем MAXH = 0, Sitout = Sit 2. Устанавливаем MAXH = 0, Sitout = Sit 3. Для j от 1 до M делать // M – число компонентов вектора ситуации в про- странстве состояний // M – число компонентов вектора ситуации в про- странстве состояний { 4. Преобразуем значение j-го компонента век- тора ситуации на противоположное { 4. Преобразуем значение j-го компонента век- тора ситуации на противоположное Повторение этого алгоритма по всем r поз- воляет из SIT = {Sitr\|r = 1, …, R} получить рас- ширенные множества исходных ситуаций-при- меров {SITr\| r = 1, …, R} и их меток: р у ц р Sitout[j] = \|Sit [j] –1 \| 5. Вычисляем h(Sitout) 6. Если h(Sitout) > MAXH То { 7. MAXH = h(Sit) 8. Sit = Sitout } } 9. Вывод Sit 10. Конец р у ц р Sitout[j] = \|Sit [j] –1 \| 5. Вычисляем h(Sitout) 6. Если h(Sitout) > MAXH То { 7. MAXH = h(Sit) 8. Sit = Sitout } } 9. Вывод Sit 10. Конец Sitout[j] = \|Sit [j] –1 \| Sitout[j] = \|Sit [j] –1 \| Вычисляем h(Sit ) 5. Вычисляем h(Sitout) 6. Если h(Sitout) > MAXH То { р { \| } SITr = {Sitk \|k = 1, …, Rr}, SIMr = {Simk \|k = 1, …, Rr }, где для всех k имеем Sim(Sit1, Sitk) > Th, то есть все элементы множества ситуаций удовлетво- ряют требованию сходства с исходной ситуа- цией (исходная ситуация включена в множе- ство в виде Sit1). 9. Вывод Sit 10. Конец По окончании работы данного алгоритма получаем такую ситуацию Sit, для которой вы- полняется h(Sit) → mаx на всем множестве си- туаций, полученных преобразованием (пере- счетом) отдельных компонентов вектора ситу- ации. Алгоритмы решения задачи преобразований В основе предложенных алгоритмов лежит известный метод оптимизации – метод покоор- динатного спуска, где в качестве исходной точки поиска выступает преобразуемый при- мер ситуации, точнее, ее вектор в пространстве состояний [14]. Первый алгоритм является ба- зовым, он позволяет получить на выходе один дополнительный вектор Sit, который отвечает критерию наибольшей похожести с исходным примером Sit. Вычисление сходства h(Sitout) Для оценки степени сходства h(Sitout) = = Sim(Sit, Sit) используются экспертный и вы- числительный подходы. Развитием базового алгоритма является ал- горитм 2, на выходе которого выдается не единственный вариант Sit, а некоторое множе- ство ситуаций, удовлетворяющих требованиям сходства с исходным примером Sit. При экспертном подходе пары ситуаций предъявляются экспертам и решается задача экспертного оценивания, которая может быть поставлена в различных формулировках. В частности, это прямое оценивание похоже- 662 4 (35) 2022 Программные продукты и системы / Software & Systems сти в терминах «Похоже (Sim(.) = 1)/Непохоже (Sim(.) = 0)» или на более детальной шкале с промежуточными значениями; оценка возмож- ности применить одно и то же решение для обеих ситуаций «Возможно (Sim(.) =1)/Невоз- можно (Sim(.) = 0)»; оценка принадлежности к одному классу ситуаций и т.п.; H(Sit) = (Sim1, …, Simi , …, SimN) → max (2) при ограничениях Simk  Thk при k K, где K – множество индексов тех элементов сложного объекта, по которым обязательно должно быть достигнуто сходство не ниже некоторого по- рога Thk. Здесь Simi – функция сходства между ситу- ациями по i-му элементу, Simi = Sim(Siti, Siti). При вычислительном подходе вводится до- полнительная функция или набор правил, поз- воляющие оценить значение Sim(Sit, Sit) путем сравнения векторов двух ситуаций. В частно- сти, это могут быть правила, построенные на основе экспертных знаний и выполняющие классификацию двух ситуаций (оценка схоже- сти в значениях 0 или 1 по принадлежности к одному классу). Более детальной будет оценка схожести векторов состояний, если она вычис- ляется как отношение одинаковых компонен- тов двух векторов к их общему числу (с воз- можностью дальнейшей модификации путем введения дополнительных весов и ограниче- ний). В частности, это ограничение может отра- жать тот факт, что при оценке сходства ситуа- ций и выборе решений может потребоваться учитывать не только состояние собственно управляемого объекта, но и его контекст, окру- жение, состояние которого может влиять на ре- шение, но управлять которым невозможно. Чтобы сравнивать ситуации и выбирать реше- ния с учетом данного требования, для элемен- тов такого контекста задается значение Thk = 1. Для решения задачи (2) очевидным образом может быть использован алгоритм 2, на вход которого подается конкатенация векторов со- стояний. Тогда в цикле на шаге 3 число M за- меняется на MN (количество из N векторов по M компонентов), а на шагах 5 и 6 вместо ло- кального сходства h(.) определяется глобаль- ное сходство H(.). Вычисление сходства h(Sitout) Стоит отметить, что качественная оценка схожести путем выбора из двух возможных значений 0 или 1 не способна отделить порож- даемые ситуации друг от друга по уровню схо- жести. Такая оценка не позволяет использовать алгоритм 1 и его модификации. Однако, как бу- дет показано далее, в сложных ситуациях, ха- рактеризующихся не одним вектором состоя- ний, а набором таких векторов – мультивекто- ром, такая качественная оценка тоже дает работоспособный способ расширения обучаю- щих данных. Следующие алгоритмы позволяют допол- нить набор инструментов расширения исход- ных множеств за счет дополнительных воз- можностей для их комбинирования. Они опе- рируют результатами применения алгоритма 2 к каждому из i-х векторов в пространстве со- стояний. Таким образом получаются множе- ства SITi с соответствующими метками – ло- кальными оценками h(.). Далее перебираются комбинации из элементов этих множеств и вы- полняется отбор этих комбинаций согласно критерию и ограничениям задачи (2). 5. Для i от 1 до N делать 4. Sitout = (Siti1 \| i = 1, …, N) Алгоритм 4. Алгоритм 4. у Описанные алгоритмы 2 и 4 служат для формирования множеств ситуаций, являю- щихся похожими на исходные ситуации отно- сительно порога сходства Th. Для формирова- ния обучающего дата-сета, который также бу- дет включать в себя непохожие ситуации, используются эти же алгоритмы, в которых условия h(Sitout) > Th и H(Sitout) > Th заменяются в соответствующих шагах алгоритма на проти- воположные h(Sitout) ≤ \|Th – d\| и H(Sitout) ≤ ≤ \|Th – d\|, где d – дополнительный коэффициент для более надежного разделения похожих и не- похожих ситуаций. Итоговый обучающий дата-сет составляется из полученных расши- ренных множеств вместе с сохраняемыми оценками сходства Sim их элементов, которые выступают в качестве меток обучающих дан- ных. Алгоритм 4. 1. Начало … 9. Если H > Th то // Если схожесть ситуации с исходной Sit выше принятого порога, то она включается в целевое множество вместе с включением H(Sitout) в соот- ветствующее множество меток 10. Sitout in SIT, H(Sitout) in SIM 11. Упорядочение элементов SIT по H(Sitout) и формирование на выходе SIT, SIM 12. Конец 1. Начало // Если схожесть ситуации с исходной Sit выше принятого порога, то она включается в целевое множество вместе с включением H(Sitout) в соот- ветствующее множество меток 10. Sitout in SIT, H(Sitout) in SIM // Если схожесть ситуации с исходной Sit выше принятого порога, то она включается в целевое множество вместе с включением H(Sitout) в соот- ветствующее множество меток 10. Sitout in SIT, H(Sitout) in SIM 11. Упорядочение элементов SIT по H(Sitout) и формирование на выходе SIT, SIM 12. Конец В результате работы алгоритма формиру- ется искомое множество ситуаций в их фор- мальном представлении мультивекторами, каждая из которых удовлетворяет требованиям пороговой схожести с исходной ситуацией. Комплексные ситуации на сложном объекте Рассмотрим случай с комплексной (слож- ной) ситуацией. Такая ситуация возникает, например, на сложных объектах, которые со- стоят из множества разнородных элементов [8]. Если каждому из N элементов сложного объекта сопоставить свой вектор в пространстве состоя- ний, то вся комплексная ситуация будет харак- теризоваться набором из N таких векторов. Да- лее он будет называться мультивектором. Если до сих пор ситуации Sit соответствовал один вектор в пространстве состояний, то сложной ситуации Sit соответствует более одного такого вектора: Sit  (Sit1, Sit2, …, SitN). Таким же об- разом новой порождаемой ситуации соответ- ствует Sit  (Sit1, Sit2, …, SitN). Алгоритм 3. 1. Начало Алгоритм 3. 1. Начало 2. Для i от 1 до N делать { 3. Выполнить Алгоритм 2 // На выходе Алгоритма 2 формируется множе- ство множеств SIT1, SIT2, …, SITi,…, SITN, где SITi = {Sitik \| k = 1, …, Ri}, Ri – число векторов ситуаций, сгенерированных путем преобразова- ний исходной ситуации и удовлетворяющих кри- терию схожести (1) } // Устанавливаем начальный набор мультивек- тора выходной ситуации, в который включаются первые элементы каждого из множеств 4. Sitout = (Siti1 \| i = 1, …, N) 5. Для i от 1 до N делать р 1. Начало 1. Начало 2. Для i от 1 до N делать { 3. Выполнить Алгоритм 2 // На выходе Алгоритма 2 формируется множе- ство множеств SIT1, SIT2, …, SITi,…, SITN, где SITi = {Sitik \| k = 1, …, Ri}, Ri – число векторов ситуаций, сгенерированных путем преобразова- ний исходной ситуации и удовлетворяющих кри- терию схожести (1) // Устанавливаем начальный набор мультивек- тора выходной ситуации, в который включаются первые элементы каждого из множеств Теперь задача (1) может быть записана как многокритериальная оптимизационная задача: 4. Sitout = (Siti1 \| i = 1, …, N) Теперь задача (1) может быть записана как многокритериальная оптимизационная задача: 5. Для i от 1 до N делать 663 4 (35) 2022 Программные продукты и системы / Software & Systems 6. Для k от 1 до Ri делать { 7. Sitout = (Sitik) 8. H = H(Sitout) 9. Если H > MAXH То { 10. MAXH = H 11. Sit = Sitout } } 12. Конец 6. Для k от 1 до Ri делать { 7. Sitout = (Sitik) 8. H = H(Sitout) 9. Если H > MAXH То { 10. MAXH = H 11. Sit = Sitout } } 12. Комплексные ситуации на сложном объекте Конец H(Sit) = 1 N i i i Sim =   (3) при i  [0, 1] и i = 1. Трудность состоит в том, что в сложном объекте весовые коэффициенты формулы (3) не являются постоянными, так как важность одних элементов может зависеть от состояний других. Это учитывается при формировании исходного множества примеров SIT = {Sitr\|r= 1, …, R}, где полагается, что для каждой ситуа- ции экспертным путем сопоставлен свой век- тор весовых коэффициентов, который будет использован в формуле (3). Модификация спо- соба агрегирования частных оценок сходства с помощью (3) использует дополнительную функцию активации: H(Sit) =f( ) 1 N i i i Sim =   , где f(.) – функция активации, учитывающая допол- нительные условия. } 12. Конец На выходе алгоритма формируется один мультивектор ситуации Sit, компоненты кото- рого – векторы из множеств SIT1, SIT2, …, SITi, …, SITN и оценка схожести H(Sit) удо- влетворяет критерию (2). Алгоритм 4 аналогичен алгоритму 2, на его выходе формируется не один мультивектор си- туации, а упорядоченное по критерию (2) мно- жество, каждый из элементов которого явля- ется представлением комплексной ситуации, полученной из исходной ситуации Sit и отвеча- ющей требованиям схожести. Чтобы не приво- дить полное содержание этого алгоритма, по- кажем только те шаги, которые заменяют шаги 9–11 в алгоритме 3. В частности, f( ) 1 N i i i Sim =   = 1 N i i i Sim =   при В частности, f( ) 1 N i i i Sim =   = 1 N i i i Sim =   при условии выполнения всех ограничений и f(.) = 0 при невыполнении хотя бы одного тре- бования относительно сходства векторов со- стояний для элементов, входящих в контекст ситуации. Вычисление сходства H(Sit) Общая блок-схема процесса подготовки дата-сетов и обучения моделей оценки схоже- сти ситуаций, использующая разработанные алгоритмы, приведена на рисунке. Здесь репо- зиторий предполагает наличие заранее реали- зованных архитектур (многослойная нейро- сеть, регрессионная модель, ансамбль моделей и т.п.), которые в процессе обучения настраи- ваются на вычисление сходства ситуаций и, та- ким образом, подготавливаются для участия в Векторное представление целевой функции схожести в (2) требует выбора способа вычис- ления H(Sit) путем агрегирования локальных оценок сходства Simi. Для этих целей использу- ется аддитивная свертка, где весовые коэффи- циенты  отражают относительную важность для общей оценки сходства двух ситуаций их похожести по i-му элементу сложного объекта: 664 4 (35) 2022 Программные продукты и системы / Software & Systems Блок-схема процесса обучения с расширением множества примеров A flowchart of the learning process with the expansion of many examples Формирование множества примеров из R ситуаций Расширение множества примеров похожими ситуациями с формированием множества меток Sim(.) Расширение множества примеров непохожими ситуациями с формированием множества меток Sim(.) Формирование исходного дата-сета из R наборов {Ситуация-пример, <Похожая ситуация, Метка>, <Непохожая ситуация, Метка>} Разделение исходного дата-сета на обучающий дата-сет, объем 70 % от R, и валидирующий дата-сет, 30 % от R Выбор и обучение модели на обучающем дата-сете Проверка качества модели на валидирующем дата-сете Репозиторий моделей для оценки схожести Метрики качества удовлетворительны? Передача обученной модели заказчику Да Нет Формирование множества примеров из R ситуаций Расширение множества примеров похожими ситуациями с формированием множества меток Sim(.) Расширение множества примеров непохожими ситуациями с формированием множества меток Sim(.) Формирование исходного дата-сета из R наборов {Ситуация-пример, <Похожая ситуация, Метка>, <Непохожая ситуация, Метка>} Разделение исходного дата-сета на обучающий дата-сет, объем 70 % от R, и валидирующий дата-сет, 30 % от R Репозиторий моделей для оценки схожести Проверка качества модели на валидирующем дата-сете Метрики качества удовлетворительны? Да Блок-схема процесса обучения с расширением множества примеров A flowchart of the learning process with the expansion of many examples A flowchart of the learning process with the expansion of many examples процессах отбора ситуаций в предметно-ори- ентированной CBR-системе. Для оценки каче- ства обученных моделей используются извест- ные метрики из арсенала машинного обучения: MAPE (Mean Absolute Percentage Error) – абсо- лютная средняя процентная ошибка значения Sim [15]; nDCG@k (Normalized Discounted Commulative Gain для лучших k) – метрика для оценки верности ранжирования k ситуаций луч- ших по величине Sim [16]; Accuracy – метрика для оценки верности определения класса [17], в рассматриваемом случае класс пар ситуаций похож/непохож при заданном пороге Sim. Вычисление сходства H(Sit) моделей оценки схожести ситуаций в рабо- те [8]. Первоначально с помощью экспертов был сформирован набор примеров ситуаций, который далее расширен с помощью предло- женных алгоритмов. Программная реализация алгоритмов выполнена посредством языка VBA в среде MS Excel, реализация нейросете- вых моделей и их обучение для оценки схоже- сти ситуаций – средствами Python с библиоте- ками Keras, Scikit-learn. Для проведения экспериментов был рас- смотрен сложный технологический объект – тепловой пункт здания. Технологическая схема представляет собой независимую двухконтур- ную систему отопления, где внешний теплоно- ситель через теплообменник передает тепло- вую энергию теплоносителю системы отопле- ния дома. В результате работы алгоритма 1 получаем дополнительную ситуацию Sit2 для сравнения с исходной: (1,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,1,0,0, 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0, 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,1,0,0, 0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,1,0,0,0, 0,0,0,0,1,0,0,0,0,0,0,0,0,1). Запишем представление данной ситуации в виде вектора Sit, длина которого составляет 112 позиций: Новая полученная ситуация имеет с изна- чальной ситуацией Sit такую же схожесть 0.952, однако отличается состоянием другого элемента, который обладает таким же весовым коэффициентом (Природные явления). Для получения большего числа ситуаций исполь- зован алгоритм 2, где был принят порог h(Sitout) > Th = 0,85. В результате применения этого алгоритма получено расширенное мно- (1,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0, 1,0,0,0,0,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0, 1,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,1, 0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,1, 0,0,0,0,0,0,0,1). Экспериментальная часть Описанный подход апробирован при подго- товке дата-сетов и проведении исследований 665 4 (35) 2022 Программные продукты и системы / Software & Systems Элементы сложного объекта сформированы в группы: тором каждая из ситуаций является похожей на исходную Sit. С помощью экспертов подобран пример Sit1, отвечающий критерию сходства с Sit. Далее целевое множество получается путем преобразований этого примера с помощью вве- денных алгоритмов. Элементы сложного объекта сформированы в группы: − технологическое оборудование (насос, теплообменник, внутренняя труба); − обеспечение (ПО, электричество, прочее оборудование); Изначально имеется пара векторов Sit и Sit1, второй из которых выглядит следующим обра- зом: (1,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0,0, 0,0,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0, 0,0,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,1,0,0,0,0,0, 0,0,1,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,1,0,0,0,0,0,0, 0,0,1,0,0,0,0,0,0,1,0). − персонал (электрик, сантехник, аварий- ная служба); − окружение (помещение теплового пунк- та, соседние здания, природные объекты и яв- ления). − окружение (помещение теплового пунк- та, соседние здания, природные объекты и яв- ления). Элементы групп «персонал» и «окружение» не связаны напрямую с объектом, но считаются его частью, поскольку могут влиять на него. Например, снегопад может усложнить доступ персонала к объекту и повлиять на программу действий (то есть решение) при устранении проблемной ситуации. Ситуации отличаются состоянием послед- него элемента (Помещение). Схожесть ситуа- ций Sim = 0.952, оценка схожести определялась по методике, описанной в исследовании [8] с учетом нормированного фактора важности каждого элемента. Представим некоторый пример ситуации на объекте «тепловой пункт здания» через мат- рицу состояний, отражающую ситуацию на сложном техническом объекте, где единица в столбце соответствует данному состоянию эле- мента (см. таблицу). В результате работы алгоритма 1 получаем дополнительную ситуацию Sit2 для сравнения с исходной: (1,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,1,0,0, 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0, 0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,1,0,0, 0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,1,0,0,0, 0,0,0,0,1,0,0,0,0,0,0,0,0,1). Заключение В исследовании предложены алгоритмы, позволяющие из исходного множества ситуа- Исследование выполнено при финансовой поддержке РФФИ и Тюменской области в рамках научного проекта № 20-47-720004. жество SIT, куда вошли, в частности, такие до- полнительные ситуации, схожие с Sit: жество SIT, куда вошли, в частности, такие до- полнительные ситуации, схожие с Sit: ций-примеров, заданных с помощью формаль- ных представлений, получить расширенное множество, которое будет включать в себя си- туации, отвечающие критериям сходства с дан- ными примерами. Применение этих алгорит- мов позволяет расширять исходные данные не- достаточного объема и создавать дата-сеты, которые будут использоваться для подготовки моделей машинного обучения в задачах оценки схожести ситуаций и выбора решений в СППР с выводом на прецедентах. Sit3: (1,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0, 0,0,0,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0, 0,0,0,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,1,0,0,0,0, 0,0,0,1,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,1,0,0,0,0, 0,0,0,1,0,0,0,0,0,0,0,1); оценка схожести 0.903; Sit4: (1,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0, 0,0,0,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0, 0,0,0,0,0,1,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,1,0,0,0,0, 0,0,0,1,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,1,0,0,0,0,0, 0,0,0,1,0,0,0,0,0,0,0,1), оценка схожести 0.879; Подход к определению схожих ситуаций на основе машинного обучения актуален для сложных, комплексных ситуаций, когда оценка сходства между ними сталкивается с необхо- димостью определения локальных метрик сходства и трудностями их агрегирования. Ре- зультативность и применимость на практике моделей машинного обучения во многом опре- деляются исходными обучающими данными. Предлагаемые результаты позволяют автома- тизировать создание дата-сетов, используя для этого подготовленные и одобренные примеры характерных ситуаций и решая задачу преобра- зований как задачу поиска оптимума целевой функции схожести. Sit5: (1,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0, 0,0,0,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0, 0,0,0,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,1,0,0,0,0, 0,0,1,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,1,0,0,0,0,0, 0,0,0,1,0,0,0,0,0,0,0,1), оценка схожести 0.952. Таким образом, выполнено расширение набора обучающих векторов для ситуации Sit. Далее, задавая с помощью экспертов новые примеры ситуаций и применяя их преобразова- ния, были сформированы обучающий и вали- дирующий дата-сеты общей размерностью 150 пар похожих и непохожих ситуаций. Резуль- таты работы подтвердили возможность автома- тического формирования больших объемов обучающих данных из относительно неболь- шого числа примеров, для подготовки которых могут привлекаться эксперты. Разработанные алгоритмы не оптимизиро- ваны с точки зрения объема вычислений. Их совершенствование для сокращения числа пе- реборов, а также создание новых алгоритмов на основе других методов оптимизации может стать предметом дополнительного исследова- ния предложенных задач. Исследование выполнено при финансовой поддержке РФФИ и Тюменской области в рамках научного проекта № 20-47-720004. (1,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0, 1,0,0,0,0,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0, 1,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,1, 0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,1, 0,0,0,0,0,0,0,1). Для обучения нейросетевой модели, кото- рая будет отбирать в базе похожие ситуации, требуется сформировать множество SIT, в ко- The state matrix Sitact Исправно Сломано В работе Оста- новлен Недо- ступно До- ступно Влияет Не влияет Ввод 1 0 0 0 0 0 0 0 Внутренняя труба 1 0 0 0 0 0 0 0 Теплообменник 0 1 0 0 0 0 0 0 Насос 0 0 1 0 0 0 0 0 Прочее оборудование 0 0 0 1 0 0 0 0 IT 0 1 0 0 0 0 0 0 Электричество 0 0 0 0 1 0 0 0 Аварийная служба 0 0 0 0 0 1 0 0 Сантехник 0 0 0 0 0 1 0 0 Электрик 0 0 0 0 0 1 0 0 Здания 0 0 0 0 0 0 0 1 Природные объекты 0 0 0 0 0 0 1 0 Природные явления 0 0 0 0 0 0 0 1 Помещение 0 0 0 0 0 0 0 1 666 4 (35) 2022 Программные продукты и системы / Software & Systems жество SIT, куда вошли, в частности, такие до- полнительные ситуации, схожие с Sit: Sit3: (1,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0, 0,0,0,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0, 0,0,0,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,1,0,0,0,0, 0,0,0,1,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,1,0,0,0,0, 0,0,0,1,0,0,0,0,0,0,0,1); оценка схожести 0.903; Sit4: (1,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0, 0,0,0,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0, 0,0,0,0,0,1,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,1,0,0,0,0, 0,0,0,1,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,1,0,0,0,0,0, 0,0,0,1,0,0,0,0,0,0,0,1), оценка схожести 0.879; Sit5: (1,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0, 0,0,0,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0, 0,0,0,0,1,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,1,0,0,0,0, 0,0,1,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,1,0,0,0,0,0, 0,0,0,1,0,0,0,0,0,0,0,1), оценка схожести 0.952. Таким образом, выполнено расширение набора обучающих векторов для ситуации Sit. Далее, задавая с помощью экспертов новые примеры ситуаций и применяя их преобразова- ния, были сформированы обучающий и вали- дирующий дата-сеты общей размерностью 150 пар похожих и непохожих ситуаций. Резуль- таты работы подтвердили возможность автома- тического формирования больших объемов обучающих данных из относительно неболь- шого числа примеров, для подготовки которых могут привлекаться эксперты. 8. Glukhikh I., Glukhikh D. Case-based reasoning with an artificial neural network for decision support in situations at complex technological objects of urban infrastructure. Applied System Innovation, 2021, vol. 4, no. 4, art. 73. DOI: 10.3390/asi4040073. Литература Aamodt A., Plaza E. Case-based reasoning: foundational issues, methodological variations, and system approaches. AI Communications, 1994, vol. 7, no. 1, pp. 39–59. DOI: 10.3233/AIC-1994-7104. 10. Chen H., Birkelund Y., Zhang Q. Data-augmented sequential deep learning for wind power forecasting. Energy Conversion and Management, 2021, vol. 248, art. 114790. DOI: 10.1016/j.enconman.2021.114790. 11. Glukhikh I., Glukhikh D. Case based reasoning for managing urban infrastructure complex technological objects. CEUR Workshop Proceedings, 2021, vol. 2843, no. 038. URL: http://ceur-ws.org/Vol- 2843/paper038.pdf (дата обращения: 20.04.2022). 12. Глухих И.Н., Глухих Д.И., Карякин Ю.Е. Представление и отбор ситуаций на сложном техноло- гическом объекте в условиях неопределенности // Вестн. РосНОУ. Сер. Сложные системы модели, ана- лиз и управление. 2021. № 2. С. 65–73. 13. Глухих И.Н., Глухих Д.И., Карякин Ю.Е. Нейросетевая архитектура вывода решений в опасных ситуациях на сложном технологическом объекте // Прикладная информатика. 2021. Т. 16. № 5. DOI: 10.37791/2687-0649-2021-16-5-99-107. 14. Wright S.J. Coordinate descent algorithms. Mathematical Programming, 2015, vol. 151, no. 1, pp. 3–34. DOI: 10.1007/s10107-015-0892-3. pp 15. de Myttenaere A., Golden В., Le Grand D., Rossi F. Mean absolute percentage error for regression models. Neurocomputing, 2016, vol. 192, pp. 38–48. DOI: 10.1016/j.neucom.2015.12.114. 16. Wang Y., Wang L., Li Y., Di H., Tie-Yan L., Wei Ch. A Theoretical analysis of NDCG type ranking measures. Proc. PMLR, 2013, no. 30, pp. 25–54. 16. Wang Y., Wang L., Li Y., Di H., Tie-Yan L., Wei Ch. A Theoretical analysis of NDCG type ranking measures. Proc. PMLR, 2013, no. 30, pp. 25–54. pp 17. Taylor J.R. An Introduction to Error Analysis: The Study of Uncertainties in Physical Measurements. University Science Books Publ., Sausalito, CA, 1997, 327 p. 17. Taylor J.R. An Introduction to Error Analysis: The Study of Uncertainties in Physical Measurements. University Science Books Publ., Sausalito, CA, 1997, 327 p. Software & Systems DOI: 10.15827/0236-235X.140.660-669 Received 11.05.22, Revised 22.07.22 2022, vol. 35, no. 4, pp. 660–669 Software & Systems DOI: 10.15827/0236-235X.140.660-669 Software & Systems DOI: 10.15827/0236-235X.140.660-669 Software & Systems 9. Aamodt A., Plaza E. Case-based reasoning: foundational issues, methodological variations, and system approaches. AI Communications, 1994, vol. 7, no. 1, pp. 39–59. DOI: 10.3233/AIC-1994-7104. Литература 1. Башлыков А.А. Применение методов теории прецедентов в системах поддержки принятия реше- ний при управлении трубопроводными системами // Автоматизация, телемеханизация и связь в нефтя- ной промышленности. 2016. № 1. С. 23–33. 1. Башлыков А.А. Применение методов теории прецедентов в системах поддержки принятия реше- ний при управлении трубопроводными системами // Автоматизация, телемеханизация и связь в нефтя- ной промышленности. 2016. № 1. С. 23–33. р Kuzyakov O.N., Andreeva M.A. Applying case-based reasoning method for decision making in IIoT . Proc. FarEastCon, 2020, pp. 1–5. DOI: 10.1109/FarEastCon50210.2020.9271301. 3. Eremeev A., Varshavskiy P., Alekhin R. Case-based reasoning module for intelligent decision support systems. Proc. I Int. Sci. Conf. IITI, 2016, vol. 1, pp. 207–216. DOI: 10.1007/978-3-319-33609-1_18. 4. Feuillâtre H., Auffret V., Castro M., Lalys F., Le Breton H., Garreau M. Similarity measures and attribute selection for case-based reasoning in transcatheter aortic valve implantation. PLoS ONE, 2020, vol. 15, no. 9, art. e0238463. DOI: 10.1371/journal.pone.0238463. j p 5. Еремеев А.П., Кожухов А.А., Голенков В.В., Гулякина Н.А. О реализации средств машинного обучения в интеллектуальных системах реального времени // Программные продукты и системы. 2018. Т. 31. № 2. С. 239–245. DOI: 10.15827/0236-235X.122.239-245. 5. Еремеев А.П., Кожухов А.А., Голенков В.В., Гулякина Н.А. О реализации средств машинного обучения в интеллектуальных системах реального времени // Программные продукты и системы. 2018. Т. 31. № 2. С. 239–245. DOI: 10.15827/0236-235X.122.239-245. 6. Еремеев А.П., Кожухов А.А. Реализация методов обучения с подкреплением на основе темпо- ральных различий и мультиагентного подхода для интеллектуальных систем реального времени // Про- граммные продукты и системы. 2017. Т. 30. № 1. С. 28–33. DOI: 10.15827/0236-235X.117.028-033. 6. Еремеев А.П., Кожухов А.А. Реализация методов обучения с подкреплением на основе темпо- ральных различий и мультиагентного подхода для интеллектуальных систем реального времени // Про- граммные продукты и системы. 2017. Т. 30. № 1. С. 28–33. DOI: 10.15827/0236-235X.117.028-033. р р у 7. Gabel T., Godehardt E. Top-down induction of similarity measures using similarity clouds. In: Case- Based Reasoning Research and Development, 2015, pp. 149–164. DOI: 10.1007/978-3-319-24586-7_11. р р у 7. Gabel T., Godehardt E. Top-down induction of similarity measures using similarity clouds. In: Case- Based Reasoning Research and Development, 2015, pp. 149–164. DOI: 10.1007/978-3-319-24586-7_11. 667 4 (35) 2022 Программные продукты и системы / Software & Systems 8. Glukhikh I., Glukhikh D. Case-based reasoning with an artificial neural network for decision support in situations at complex technological objects of urban infrastructure. Applied System Innovation, 2021, vol. 4, no. 4, art. 73. DOI: 10.3390/asi4040073. 9. Acknowledgements. The research was funded by RFBR and Tyumen Region, Russia, project no. 20-47-720004. 10. Chen H., Birkelund Y., Zhang Q. Data-augmented sequential deep learning for wind power forecasting. Energy Conversion and Management, 2021, vol. 248, art. 114790. DOI: 10.1016/j.enconman.2021.114790. Algorithms for generating training sets in a system with case-based inference based on example situations I.N. Glukhikh 1, Dr.Sc. (Engineering), Professor D.I. Glukhikh 1, Postgraduate Student, d.i.glukhikh@utmn.ru 1 Department of Information Systems, University of Tyumen, Tyumen, 625004, Russian Federation 1 Department of Information Systems, University of Tyumen, Tyumen, 625004, Russian Federatio Для цитирования Глухих И.Н., Глухих Д.И. Алгоритмы генерации обучающих множеств в системе с прецедентным выводом на основе ситуаций-примеров // Программные продукты и системы. 2022. Т. 35. № 4. С. 660–669. DOI: 10.15827/0236-235X.140.660-669. Глухих И.Н., Глухих Д.И. Алгоритмы генерации обучающих множеств в системе с прецедентным выводом на основе ситуаций-примеров // Программные продукты и системы. 2022. Т. 35. № 4. С. 660–669. DOI: 10.15827/0236-235X.140.660-669. References In: Case- Based Reasoning Research and Development, 2015, pp. 149–164. DOI: 10.1007/978-3-319-24586-7_11. 8. Glukhikh I., Glukhikh D. Case-based reasoning with an artificial neural network for decision support in situations at complex technological objects of urban infrastructure. Applied System Innovation, 2021, vol. 4, no. 4, art. 73. DOI: 10.3390/asi4040073. 8. Glukhikh I., Glukhikh D. Case-based reasoning with an artificial neural network for decision support in situations at complex technological objects of urban infrastructure. Applied System Innovation, 2021, vol. 4, no. 4, art. 73. DOI: 10.3390/asi4040073. , 9. Aamodt A., Plaza E. Case-based reasoning: foundational issues, methodological variations, and system approaches. AI Communications, 1994, vol. 7, no. 1, pp. 39–59. DOI: 10.3233/AIC-1994-7104. , 9. Aamodt A., Plaza E. Case-based reasoning: foundational issues, methodological variations, and system approaches. AI Communications, 1994, vol. 7, no. 1, pp. 39–59. DOI: 10.3233/AIC-1994-7104. 10. Chen H., Birkelund Y., Zhang Q. Data-augmented sequential deep learning for wind power forecasting. Energy Conversion and Management, 2021, vol. 248, art. 114790. DOI: 10.1016/j.enconman.2021.114790. 10. Chen H., Birkelund Y., Zhang Q. Data-augmented sequential deep learning for wind power forecasting. Energy Conversion and Management, 2021, vol. 248, art. 114790. DOI: 10.1016/j.enconman.2021.114790. 11. Glukhikh I., Glukhikh D. Case based reasoning for managing urban infrastructure complex technolog- ical objects. CEUR Workshop Proceedings, 2021, vol. 2843, no. 038. Available at: http://ceur-ws.org/Vol- 2843/paper038.pdf (accessed April 20, 2022). 12. Glukhikh I.N., Glukhikh D.I., Karyakin Yu.E. Representation and retrieve of the situation on a complex technological object in the uncertainty conditions. Bull. of RosNOU. Ser. Complex Systems: Models, Analysis, Management, 2021, no. 2, pp. 65–73 (in Russ.). 13. Glukhikh I.N., Glukhikh D.I., Karyakin Yu.E. Neural network architecture for outputting solutions in dangerous situations at a complex technological facility. J. of Applied Informatics, 2021, vol. 16, no. 5, pp. 99–107. DOI: 10.37791/2687-0649-2021-16-5-99-107 (in Russ.). Wright S.J. Coordinate descent algorithms. Mathematical Programming, 2015, vol. 151, no. 1, 34. DOI: 10.1007/s10107-015-0892-3. 15. de Myttenaere A., Golden В., Le Grand D., Rossi F. Mean absolute percentage error for regression models. Neurocomputing, 2016, vol. 192, pp. 38–48. DOI: 10.1016/j.neucom.2015.12.114. 16. Wang Y., Wang L., Li Y., Di H., Tie-Yan L., Wei Ch. A Theoretical analysis of NDCG type ranking measures. Proc. PMLR, 2013, no. 30, pp. 25–54. 17. Taylor J.R. An Introduction to Error Analysis: The Study of Uncertainties in Physical Measurements. University Science Books Publ., Sausalito, CA, 1997, 327 p. References References 1. Bashlykov A.A. Precedent theory methods applyed in the systems of decision-making when managing pipeline systems. Automation, Telemechanization and Communication in Oil Industry, 2016, no. 1, pp. 23–33 (in Russ.). 1. Bashlykov A.A. Precedent theory methods applyed in the systems of decision-making when managing pipeline systems. Automation, Telemechanization and Communication in Oil Industry, 2016, no. 1, pp. 23–33 (in Russ.). ) Kuzyakov O.N., Andreeva M.A. Applying case-based reasoning method for decision making in IIoT . Proc. FarEastCon, 2020, pp. 1–5. DOI: 10.1109/FarEastCon50210.2020.9271301. 3. Eremeev A., Varshavskiy P., Alekhin R. Case-based reasoning module for intelligent decision support systems. Proc. I Int. Sci. Conf. IITI, 2016, vol. 1, pp. 207–216. DOI: 10.1007/978-3-319-33609-1_18. 3. Eremeev A., Varshavskiy P., Alekhin R. Case-based reasoning module for intelligent decision support systems. Proc. I Int. Sci. Conf. IITI, 2016, vol. 1, pp. 207–216. DOI: 10.1007/978-3-319-33609-1_18. 4. Feuillâtre H., Auffret V., Castro M., Lalys F., Le Breton H., Garreau M. Similarity measures and attrib- ute selection for case-based reasoning in transcatheter aortic valve implantation. PLoS ONE, 2020, vol. 15, no. 9, art. e0238463. DOI: 10.1371/journal.pone.0238463. 4. Feuillâtre H., Auffret V., Castro M., Lalys F., Le Breton H., Garreau M. Similarity measures and attrib- ute selection for case-based reasoning in transcatheter aortic valve implantation. PLoS ONE, 2020, vol. 15, no. 9, art. e0238463. DOI: 10.1371/journal.pone.0238463. 5. Eremeev A.P., Kozhukhov A.A., Golenkov V.V., Gulyakina N.A. On the implementation of machine learning tools in real-time intelligent systems. Software & Systems, 2018, vol. 31, no. 2, pp. 239–245. DOI: 10.15827/0236-235X.122.239-245 (in Russ.). 5. Eremeev A.P., Kozhukhov A.A., Golenkov V.V., Gulyakina N.A. On the implementation of machine learning tools in real-time intelligent systems. Software & Systems, 2018, vol. 31, no. 2, pp. 239–245. DOI: 10.15827/0236-235X.122.239-245 (in Russ.). 6. Eremeev A.P., Kozhukhov A.A. Implementation of reinforcement learning methods based on temporal differences and a multi-agent approach for real-time intelligent systems. Software & Systems, 2017, vol. 30, no. 1, pp. 28–33. DOI: 10.15827/0236-235X.117.028-033 (in Russ.). 6. Eremeev A.P., Kozhukhov A.A. Implementation of reinforcement learning methods based on temporal differences and a multi-agent approach for real-time intelligent systems. Software & Systems, 2017, vol. 30, no. 1, pp. 28–33. DOI: 10.15827/0236-235X.117.028-033 (in Russ.). pp 7. Gabel T., Godehardt E. Top-down induction of similarity measures using similarity clouds. In: Case- Based Reasoning Research and Development, 2015, pp. 149–164. DOI: 10.1007/978-3-319-24586-7_11. pp 7. Gabel T., Godehardt E. Top-down induction of similarity measures using similarity clouds. 1 Department of Information Systems, University of Tyumen, Tyumen, 625004, Russian Federation Abstract. The paper considers the issue of creating training sets and their scaling in machine learning problems. The subject of the study is the process of generating training sets based on examples in order to augment them. g To implement the idea of expansion, it is proposed to use the transformation of existing examples of situa- tions. The transformation of examples is based on a well-known optimization method - the method of coordi- nate descent. The paper describes the statement of the problem of transformations of example situations in terms of the introduced representation model. There are proposed algorithms that make it possible to obtain an extended set from the initial set of example situations specified using formal representations, which will include situations that meet the similarity criteria with these examples. The paper presents the testing of the proposed algorithms for expanding a set of example situations, carried out in order to form a data set for the studying artificial neural networks. The obtained results are of practical importance for training artificial neural networks used in intelligent decision support systems. The proposed algorithms make it possible to automate the formation of datasets using the available prepared and approved examples of typical situations and solving the transformation problem as the problem of finding the optimum of the similarity objective function. y j Keywords: neural network training, training data, case-based reasoning, artificial intelligence, coordinate descent. Acknowledgements. The research was funded by RFBR and Tyumen Region, Russia, project no. 20-47-720004. 668 4 (35) 2022 Программные продукты и системы / Software & Systems For citation Glukhikh I.N., Glukhikh D.I. Algorithms for generating training sets in a system with case-based inference based on example situations. Software & Systems, 2022, vol. 35, no. 4, pp. 660–669 (in Russ.). DOI: 10.15827/0236-235X.140.660-669. Glukhikh I.N., Glukhikh D.I. Algorithms for generating training sets in a system with case-based inference based on example situations. Software & Systems, 2022, vol. 35, no. 4, pp. 660–669 (in Russ.). DOI: 10.15827/0236-235X.140.660-669. 669
https://openalex.org/W2136588715	https://portal.research.lu.se/files/1349681/1745134.pdf	English	null	Therapeutic efficacy of alpha-1 antitrypsin augmentation therapy on the loss of lung tissue: an integrated analysis of 2 randomised clinical trials using computed tomography densitometry	Respiratory research	2,010	cc-by	7,667	Link to publication Citation for published version (APA): Stockley, R. A., Parr, D. G., Piitulainen, E., Stolk, J., Stoel, B. C., & Dirksen, A. (2010). Therapeutic efficacy of alpha-1 antitrypsin augmentation therapy on the loss of lung tissue: an integrated analysis of 2 randomised clinical trials using computed tomography densitometry. Respiratory Research, 11. https://doi.org/10.1186/1465- 9921-11-136 Citation for published version (APA): Stockley, R. A., Parr, D. G., Piitulainen, E., Stolk, J., Stoel, B. C., & Dirksen, A. (2010). Therapeutic efficacy of alpha-1 antitrypsin augmentation therapy on the loss of lung tissue: an integrated analysis of 2 randomised clinical trials using computed tomography densitometry. Respiratory Research, 11. https://doi.org/10.1186/1465- 9921-11-136 Total number of authors: 6 General rights U l th legal requirements associated with these rights. • Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal Read more about Creative commons licenses: https://creativecommons.org/licenses/ Therapeutic efficacy of alpha-1 antitrypsin augmentation therapy on the loss of lung tissue: an integrated analysis of 2 randomised clinical trials using computed tomography densitometry Therapeutic efficacy of alpha-1 antitrypsin augmentation therapy on the loss of lung tissue: an integrated analysis of 2 randomised clinical trials using computed tomography densitometry Stockley, Robert A.; Parr, David G.; Piitulainen, Eeva; Stolk, Jan; Stoel, Berend C.; Dirksen, Asger Published in: Respiratory Research Stockley, Robert A.; Parr, David G.; Piitulainen, Eeva; Stolk, Jan; Stoel, Berend C.; Dirksen, Asger Published in: Respiratory Research DOI: 10.1186/1465-9921-11-136 Abstract Background: Two randomised, double-blind, placebo-controlled trials have investigated the efficacy of IV alpha-1 antitrypsin (AAT) augmentation therapy on emphysema progression using CT densitometry. Methods: Data from these similar trials, a 2-center Danish-Dutch study (n = 54) and the 3-center EXAcerbations and CT scan as Lung Endpoints (EXACTLE) study (n = 65), were pooled to increase the statistical power. The change in 15th percentile of lung density (PD15) measured by CT scan was obtained from both trials. All subjects had 1 CT scan at baseline and at least 1 CT scan after treatment. Densitometric data from 119 patients (AAT [Alfalastin® or Prolastin®], n = 60; placebo, n = 59) were analysed by a statistical/endpoint analysis method. To adjust for lung volume, volume correction was made by including the change in log-transformed total lung volume as a covariate in the statistical model. Results: Mean follow-up was approximately 2.5 years. The mean change in lung density from baseline to last CT scan was -4.082 g/L for AAT and -6.379 g/L for placebo with a treatment difference of 2.297 (95% CI, 0.669 to 3.926; p = 0.006). The corresponding annual declines were -1.73 and -2.74 g/L/yr, respectively. Results: Mean follow-up was approximately 2.5 years. The mean change in lung density from baseline to last CT scan was -4.082 g/L for AAT and -6.379 g/L for placebo with a treatment difference of 2.297 (95% CI, 0.669 to 3.926; p = 0.006). The corresponding annual declines were -1.73 and -2.74 g/L/yr, respectively. Conclusions: The overall results of the combined analysis of 2 separate trials of comparable design, and the only 2 controlled clinical trials completed to date, has confirmed that IV AAT augmentation therapy significantly reduces the decline in lung density and may therefore reduce the future risk of mortality in patients with AAT deficiency- related emphysema. Conclusions: The overall results of the combined analysis of 2 separate trials of comparable design, and the only 2 controlled clinical trials completed to date, has confirmed that IV AAT augmentation therapy significantly reduces the decline in lung density and may therefore reduce the future risk of mortality in patients with AAT deficiency- related emphysema. Trial registration: The EXACTLE study was registered in ClinicalTrials.gov as ‘Antitrypsin (AAT) to Treat Emphysema in AAT-Deficient Patients’; ClinicalTrials.gov Identifier: NCT00263887. © 2010 Stockley et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Therapeutic efficacy of alpha-1 antitrypsin augmentation therapy on the loss of lung tissue: an integrated analysis of 2 randomised clinical trials using computed tomography densitometry obert A Stockley1, David G Parr2, Eeva Piitulainen3, Jan Stolk4, Berend C Stoel4, Asger Dirksen5 Correspondence: r.a.stockley@bham.ac.uk 1Lung Investigation Unit, University Hospitals of Birmingham, Edgbaston, Birmingham B15 2TH, UK Full list of author information is available at the end of the article Read more about Creative commons licenses: https://creativecommons.org/licenses/ Take down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim. Take down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim. LUND UNIVERSITY PO Box 117 221 00 Lund +46 46-222 00 00 Stockley et al. Respiratory Research 2010, 11:136 http://respiratory-research.com/content/11/1/136 Data analysis and CT densitometry The rate of emphysema progression was determined by change in lung density measured by whole lung CT scan, and reported as the annual change in the 15th per- centile lung density (PD15) (determined from the end- point in the original trials). The PD15 value is extracted from the frequency histogram of lung voxels and is the density value (g•L-1) at which 15% of the voxels have lower densities [9,10] (Figure 2). This analysis combines the raw data from both trials, thereby increasing the numbers of patients and the robustness of the analysis. In addition, because of the similar study design and method of CT densitometry, we combined the raw data from both studies to increase the statistical power as suggested in the previous Danish-Dutch study [14]. Introduction restoring the inhibitory capacity of AAT in the lungs. The net result is argued to be retardation of the destructive process and, therefore, the progressive decline in lung physiology [2]. A strategy of weekly aug- mentation with AAT was thus introduced in the 1980s, confirming that the attainment of a putative protective level was possible with weekly infusions of AAT at a dose of 60 mg•kg-1 body weight [3]. In subjects with a hereditary deficiency of alpha-1 anti- trypsin (AAT), the pathophysiology of emphysema is believed to be a direct consequence of tissue damage caused by a reduced ability of AAT to inactivate neutro- phil elastase, which is released by migrating neutrophils in response to inflammatory stimuli [1]. It is logical that augmentation of the circulating levels (and hence lung levels) of AAT would confer normal protection by Because the numbers required to perform a controlled clinical trial using forced expiratory volume in 1 second (FEV1) are thought to be prohibitive (requiring inclusion of a large number of individuals with a rare disease over many years [4,5]), no such study has been undertaken. * Correspondence: r.a.stockley@bham.ac.uk 1Lung Investigation Unit, University Hospitals of Birmingham, Edgbaston, Birmingham B15 2TH, UK Full list of author information is available at the end of the article Stockley et al. Respiratory Research 2010, 11:136 http://respiratory-research.com/content/11/1/136 Page 2 of 8 Despite this, augmentation therapy is widely prescribed using varying treatment intervals and doses of plasma- derived AAT [6]. further details of the patient inclusion and exclusion cri- teria for the 2 studies, can be found in the original pub- lications [14,15]. In the past, the mainstay of clinical assessment of emphysema was lung function and especially gas transfer measurements, although recent data have indicated that there is differential progression depending on disease severity [7]. Computed tomography (CT) densitometry is a validated and more direct measure of pathological emphy- sema [8-10] that relates well to physiological and clinical features of disease [11,12], progresses uniformly across dis- ease severity [10] and has specifically been shown to be the best independent predictor of mortality [13]. Patients Pooled patient data from the 2 previously described trials, the 2-centre Danish-Dutch study (Copenhagen, Denmark; Leiden, The Netherlands) [14] and the 3-centre EXACTLE study (Copenhagen, Denmark; Birmingham, United Kingdom; Malmö, Sweden) [15], are summarised in Table 2. All patients had been recruited from AAT deficiency registries. The Danish- Dutch study randomised 56 patients and there were 77 from EXACTLE; in total, 125 patients were valid for CT data analysis (Figure 1). However, 6 patients originally enrolled in the Danish-Dutch trial also participated in the EXACTLE study. The data for these 6 subjects from EXACTLE were therefore excluded from the integrated analysis. The original studies had been approved by local ethics committees and were conducted in accor- dance with the Declaration of Helsinki and Good Clini- cal Practice Guidelines. In 1999, Dirksen, et al reported a 3-year Danish- Dutch controlled study of intravenous (IV) AAT aug- mentation therapy, with loss of lung tissue measured by CT densitometry as a secondary outcome parameter in 56 patients [14]. The study suggested a reduction in emphysema progression with AAT augmentation ther- apy measured by CT, although the p value for the treat- ment difference obtained (p = 0.07) failed to achieve the conventional level of significance, which may reflect the number of subjects in the trial. More recently, the EXAcerbations and CT scan as Lung Endpoints (EXACTLE) study (77 patients studied over 24-30 months), using a similar placebo-controlled trial design of IV AAT, explored CT densitometry as the primary outcome [15]. Lung density was analysed using 4 different methods of adjustment that corrected for varia- tion in inspiratory levels between scans, and all showed a trend towards efficacy. However, endpoint analysis using a statistical correction for lung volume not only proved to be the most sensitive method of analysis (based on monitoring progression in the placebo group), but also achieved a conventional level of statistical significance with regard to lung tissue loss between both treatment groups. Interestingly, in both the Danish-Dutch and EXACTLE studies, there was little difference in density loss between the AAT and placebo groups within the first year while, subsequently, the difference between the groups increased with time. Furthermore, the effect of therapy in clinical trials is usually determined by end- point analysis. For these reasons, we chose to re-analyse the Danish-Dutch study using an endpoint analysis, utilising only the first and last available measurement. Study designs h d Both studies were randomised, placebo-controlled, double- blind, parallel-group trials [14,15]. Patients in the Danish- Dutch study were randomised to receive infusions of either AAT (Alfalastin®; Laboratoire Français du Fraction- nement et des Biotechnologies, 3 avenue des Tropiques, BP 305, Les Ulis, 91958 Courtaboeuf Cedex, France; 250 mg•kg-1 body weight) or placebo (human albumin solution; 625 mg•kg-1 body weight) every 4 weeks for ≥3 years [14]. Patients in the EXACTLE study were rando- mised to weekly infusions of AAT (Prolastin®; Talecris Biotherapeutics, Inc., Research Triangle Park, NC, USA; 60 mg•kg-1 body weight) or placebo (2% albumin) for 24 months, with an optional extension to 30 months in subjects who agreed to continue in the study [15]. Materials and methods For the Danish-Dutch study, scans were acquired during a breath hold (Dutch patients) or during quiet tidal breathing (Danish patients). The EXACTLE trial acquired scans during a breath hold at maximum inspiration as summarised in the online supplement for Dirksen et al [15]. In both trials, CT scanners were carefully calibrated and all scan data were centrally analysed by BioImaging Technolo- gies, Inc. Materials and methods CT scans were performed at baseline and annually thereafter. In the EXACTLE study, there was an option for additional scans at 30 months in those subjects who had their participation prolonged from 24 months [15]. Characteristics of the study subjects and designs of the Danish-Dutch and EXACTLE studies are presented in Table 1. Full methodological details, together with Page 3 of 8 Stockley et al. Respiratory Research 2010, 11:136 http://respiratory-research.com/content/11/1/136 Stockley et al. Respiratory Research 2010, 11:136 http://respiratory-research.com/content/11/1/136 Table 1 Comparison of study characteristics Table 1 Comparison of study characteristics Danish-Dutch trial EXACTLE trial Genotype/phenotype PiZZ on IEF PiZZ or severe deficiency with AAT concentrations <11 μM Lung function, FEV1 30-80% 25-80% and FEV1/VC ≤70% or Kco NA ≤80% if spirometry normal Exacerbations NA ≥1 exacerbation in the past 2 years Smoking history Never or ex-smokers for >6 months Cotinine checked every 4 weeks Never or ex-smokers for >6 months Cotinine checked at 1, 6, 24 and 30 months Previous augmentation therapy NA Never or ≤1 month in past 2 years Study design Randomised, double-blind, placebo-controlled Randomised, double-blind, placebo-controlled AAT dosing 250 mg•kg-1 body weight AAT 60 mg•kg-1 body weight AAT Treatment interval Every 4 weeks Every week Placebo 625 mg•kg-1 body weight albumin 2% albumin Centres 2 (Copenhagen, Leiden) 3 (Copenhagen, Birmingham, Malmö) Duration of study Minimum 3 years 24 months (optional 6-months extension) Study period January 1991 to August 1997 November 2003 to December 2006 Primary endpoints FEV1 measured by home spirometry twice daily Change in PD15 measured by CT Other endpoints Change in PD15 measured by CT Exacerbations Lung function (FEV1, KCO) Quality of life (SGRQ) AAT: alpha-1 antitrypsin; EXACTLE: Exacerbations and Computed Tomography scan as Lung Endpoints; IEF: isoelectric focusing; Kco: carbon monoxide transfer coefficient; NA: not applicable; PD15: 15th percentile lung density; SGRQ: St George’s Respiratory Questionnaire; VC: vital capacity. CT: computed tomography. AAT: alpha-1 antitrypsin; EXACTLE: Exacerbations and Computed Tomography scan as Lung Endpoints; IEF: isoelectric focusing; Kco: carbon monoxide transfer coefficient; NA: not applicable; PD15: 15th percentile lung density; SGRQ: St George’s Respiratory Questionnaire; VC: vital capacity. CT: computed tomography. CTLE: Exacerbations and Computed Tomography scan as Lung Endpoints; IEF: isoelectric focusing; Kco: carbon monoxide transfer ; PD15: 15th percentile lung density; SGRQ: St George’s Respiratory Questionnaire; VC: vital capacity. CT: computed tomography. CT scans were obtained during both trials using differ- ent scanner protocols. y AAT: alpha-1 antitrypsin; DLco: diffusion capacity of the lung for carbon monoxide; EXACTLE: Exacerbations and Computed Tomography scan as Lung Endpoints; Kco: carbon monoxide transfer co-efficient; PD15: 15th percentile lung density; TLC: total lung capacity; VC: vital capacity. The combined analysis was based on the modified ITT population and did not include the data for 6 subjects who participated in EXACTLE, but who had also participated in the earlier Danish-Dutch study. y p p The combined analysis was based on the modified ITT population and did not include the data for 6 subjects who participated in EXACTLE, but who had also alues are mean ± SD unless otherwise indicated. C dj d PD15 CT l d i l i li d b CT d l l l d di id d b h i di id l i ’ di d TLC †TLC-adjusted PD15: CT lung density multiplied by CT-measured total lung volume and divided by the individual patient’s predicted For the CT densitometric analyses, the modified ITT population was used. s are mean ± SD unless otherwise indicated. justed PD15: CT lung density multiplied by CT-measured total lung volume and divided by the individual patient’s predicted TLC. n ± SD unless otherwise indicated. D15: CT lung density multiplied by CT-measured total lung volume and divided by the individual patient’s predicted TLC. t t i l th difi d ITT l ti d Data analysis and FEV1 We also took the opportunity to review the FEV1 decline from both studies using all available data and a slope ana- lysis for the patients included in the integrated analysis. From the original Danish-Dutch study we were, however, unable to retrieve spirometry from 4 of the subjects. Materials and methods (Leiden, The Netherlands) using PulmoCMS® Table 2 Patient baseline demographic characteristics* Danish-Dutch trial EXACTLE trial Combined data AAT (n = 27) Placebo (n = 27) AAT (n = 38) Placebo (n = 39) AAT (n = 60) Placebo (n = 59) p value Age (y) 48.0 ± 7.99 47.5 ± 7.29 54.7 ± 8.4 55.3 ± 9.8 51.6 ± 9.03 51.8 ± 9.73 0.808 Sex (n) male/female 18/9 16/11 25/13 16/23 38/22 29/30 0.093 Smoking status (n, ex/never) 27/0 27/0 34/4 35/4 56/4 56/3 0.748 Body mass index (kg•m2) 23.3 ± 3.15 24.4 ± 2.70 24.3 ± 3.3 24.3 ± 3.5 24.0 ± 3.3 24.5 ± 3.2 0.355 FEV1 (L), median 1.63 ± 0.49 1.63 1.72 ± 0.53 1.61 1.44 ± 0.60 1.14 1.35 ± 0.62 1.14 1.55 ± 0.56 1.47 1.48 ± 0.63 1.38 0.553 FEV1% predicted, median 47.3 ± 11.4 48.6 51.2 ± 14.5 49.0 46.3 ± 19.6 41.1 46.6 ± 21.0 39.5 48.0 ± 16.4 47.2 47.9 ± 18.6 43.1 0.949 Danish-Dutch trial EXACTLE trial Combined data AAT (n = 27) Placebo (n = 27) AAT (n = 38) Placebo (n = 39) AAT (n = 60) Placebo (n = 59) p value VC % predicted 114 ± 14.7 117 ± 16.4 94 ± 21.8 98 ± 23.2 103.1 ± 21.8 104.7 ± 23.9 0.789 DLCO% predicted Median 59.7 ± 16.0 57.0 60.1 ± 16.3 65.0 50.7 ± 19.5 47.6 52.2 ± 15.2 50.1 56.3 ± 17.3 56.1 55.7 ± 15.9 56.0 0.797 KCO % predicted 62.2 ± 17.62 59.9 ± 16.9 55.3 ± 21.0 56.5 ± 14.8 60.0 ± 18.9 58.6 ± 15.5 0.619 Unadjusted PD15 (g•L-1) 71.41 ± 20.87 75.56 ± 25.53 47.98 ± 19.07 45.48 ± 16.95 58.88 ± 23.03 59.79 ± 25.83 0.844 TLC-adjusted PD15† (g•L-1) 59.9 ± 11.03 62.98 ± 13.49 54.6 ± 17.4 53.9 ± 16.0 57.1 ± 15.2 58.2 ± 15.7 0.691 Lung volume (L) 5.71 ± 1.27 5.52 ± 1.34 7.46 ± 1.60 7.27 ± 1.78 6.61 ± 1.67 6.35 ± 1.69 0.300 * Values are mean ± SD unless otherwise indicated Table 2 Patient baseline demographic characteristics* * Values are mean ± SD unless otherwise indicated. † y AAT: alpha-1 antitrypsin; DLco: diffusion capacity of the lung for carbon monoxide; EXACTLE: Exacerbations and Computed Tomography scan as Lung Endpoints; Kco: carbon monoxide transfer co-efficient; PD15: 15th percentile lung density; TLC: total lung capacity; VC: vital capacity. Statistical analysis The raw data from the Danish-Dutch and EXACTLE studies were retrieved and combined. A study ID vari- able was included in the integrated analysis database to identify the records in the Danish-Dutch or EXACTLE studies. Materials and methods Stockley et al. Respiratory Research 2010, 11:136 http://respiratory-research.com/content/11/1/136 Page 4 of 8 Figure 1 Patient disposition by treatment (patients providing CT data). AAT: alpha-1 antitrypsin; EXACTLE: Exacerbations and Computed Tomography scan as Lung Endpoints. Figure 1 Patient disposition by treatment (patients providing CT data). AAT: alpha-1 antitrypsin; EXACTLE: Exacerbations and Computed Tomography scan as Lung Endpoints. (Medis Specials, Leiden, The Netherlands) for the EXACTLE study, and by Leiden University Medical Centre for the Danish-Dutch study. reproducibility of CT. In the chosen method (statistical/ endpoint analysis method), volume correction was made by including the change in log-transformed total lung volume (TLV) as a covariate in the statistical model as described [14]. This method corrects for intra-patient differences in inspiration between scans as well as inter- patient differences in technique between centres. Discussion Until now, a suitably powered double-blind randomised trial of the clinical effectiveness of AAT augmentation therapy has been lacking. The individual and combined analysis of the Danish-Dutch and EXACTLE trials con- firms that AAT augmentation therapy has a beneficial effect on the decline in lung density, which is a measure of the progression of emphysema. In the Danish-Dutch study, the mean (range) length of exposure was 2.52 (0.9-4.2) years to AAT, and 2.55 (0.9- 3.9) years to placebo. The corresponding values in the EXACTLE study were 2.23 (1.1-2.6) and 2.18 (0.8-2.6) years, respectively. For the combined data from both studies, the mean (range) length of exposure to AAT was 2.36 (0.9-4.2) years and to placebo, 2.33 (0.9-3.9) years. AAT augmentation therapy is an accepted therapeutic regimen [6], and an earlier observational study showed better overall survival and reduced FEV1 decline (albeit in a subset with moderate airflow obstruction) for patients receiving therapy with varying regimens [16]. Whereas the recommended regimen is 60 mg•kg-1 body weight per week, other adopted approaches are likely to be as effective if the nadir AAT level is mostly above the putative protective threshold of 11 μM. The characteristics for patients at baseline are sum- marised in Table 2. Baseline demographics for patients enrolled into the Danish-Dutch and EXACTLE studies were comparable, although patients in the EXACTLE study were slightly older and had a lower FEV1% pre- dicted. For the combined data, there were no statistically significant differences between the group receiving AAT or placebo with respect to age or body mass index. There were some gender differences between the treat- ment groups, with more male subjects in the active treatment group, although this was not statistically sig- nificant (p = 0.093). Preservation of normal lung structure has been the long-term aim of preventive therapy in chronic obstruc- tive pulmonary disease (COPD). However, studies of this concept have used FEV1 as the endpoint, since it is not only a defining feature of COPD but also reflects patients with a variety of phenotypes, including those with small airways disease and emphysema. Moreover, FEV1 is a reasonable marker of a patient’s health status and exercise capacity [17], and has previously been con- sidered to be the best predictor of respiratory and all- cause mortality [18]. This has led to the tenet that the maintenance of FEV1 reflects disease stability or a con- sequent reduction in mortality. Discussion Nevertheless, FEV1 is a poor surrogate measure for the presence and severity of emphysema and its progression. For instance, it has been demonstrated that patients with apical emphysema may have a preserved FEV1 in both AAT deficiency [8,19] and usual COPD [20]. All patients fulfilled the physiological inclusion criteria shown in Table 1. There were no statistically significant differences at baseline between the treatment groups with regard to these parameters. There was also no sig- nificant difference in total lung capacity-adjusted PD15 between the 2 groups at baseline (p = 0.691). Volume correction of CT Scans The level of inspiration during scan acquisition is re- cognised to influence lung density and reduce the All CT scan analyses were based on the modified intent-to-treat (ITT) population, which included (in common with the ITT) all randomised subjects who received the study therapy. However, those subjects in the modified ITT population also had to have one valid CT scan measurement at baseline and at least one valid CT scan assessment at the Month 12 visit or after. Figure 2 Measurement of progression of emphysema. For the Danish-Dutch and EXACTLE studies, PD15 was analysed using an analysis of covariance (ANCOVA) model with change from baseline to the last CT scan measurement in PD15 as the dependent variable, treat- ment and centre as fixed factors, and change in loga- rithm of CT-measured TLV and baseline measurement as covariates (statistical/endpoint analysis method). For the combined data of the integrated analysis, the study ID was added to the model as a fixed effect. The ANCOVA model included the change from baseline to Figure 2 Measurement of progression of emphysema. Stockley et al. Respiratory Research 2010, 11:136 http://respiratory-research.com/content/11/1/136 Page 5 of 8 the 2 trials demonstrated a significant reduction in the loss of lung tissue for subjects receiving treatment with IV AAT in comparison with those receiving placebo. the last CT scan as the dependent variable; study (EXACTLE versus Danish-Dutch), treatment, centre and change in logarithm of lung volume as fixed factors, and baseline measurement as covariate. Patient disposition and baseline characteristics CT densitometric measurements from a total of 119 patients were analysed (AAT, n = 60; placebo, n = 59). In the Danish-Dutch study, CT data were obtained from 54 patients, comprising 26 patients from Denmark and 28 patients from The Netherlands. In the EXACTLE study, 65 patients provided data, 27 from Denmark, 23 from the United Kingdom and 15 from Sweden. The patient disposition by treatment is shown in Figure 1. Results The FEV1 declined significantly in both the combined treated and placebo groups. The average annualised differ- ence in FEV1 loss was 13 mL•yr-1 greater in the treated group although this is within the error of measurement (95% CI, -38 to 13; p = 0.321). CT densitometric progression This indicates that the reduction in densitometry quantified here (ྜ1 HU/year) is equivalent to about a 38 ml difference in FEV1 decline in patients in GOLD stage 2. With AAT augmentation therapy becoming widely accepted throughout the United States and Europe, the ability to deliver appropriately powered placebo-controlled clinical trials, particularly those requiring a physiological measurement outcome, has become difficult to justify ethi- cally and even more difficult to deliver. The current analy- sis, however, provides evidence of augmentation therapy reducing the rate of progression of lung tissue loss. The data, therefore, permit future studies to be powered for comparison of different therapeutic regimens using CT scans rather than physiology (either FEV1 or gas transfer). However, it should also be noted that even CT scans, as well as accepted physiological measurements, are only sur- rogate measures of emphysema. Importantly, the change in physiological endpoints varies throughout the course of the disease, with FEV1 decline being greatest in subjects with moderate airflow obstruction (35-79% of predicted) [16] and gas transfer decline being greatest in those with most severe disease [7]. On the other hand, loss of lung density as assessed by PD15 shows a more constant change across all stages of disease severity [10], suggesting that it is a better marker of the continuing disease process. It is not always feasible to conduct powered clinical However as indicated above the decline in FEV1 is not linear throughout the disease process. Therefore, for this and other reasons, stabilisation of emphysema progres- sion, as indicated by CT densitometry, would be as important an aim, if not more so, than preserving FEV1. The current combined analysis of the only 2 controlled Figure 3 Progression of emphysema in AAT-treated versus placebo-treated subjects (modified ITT). Estimated treatment difference between mean changes in unadjusted 15th percentile lung density from baseline. AAT: alpha-1 antitrypsin; LS: least squares; PD15: 15th percentile lung density. It is not always feasible to conduct powered clinical studies [21], and sometimes a combination of compar- able studies is necessary. For example, meta-analysis of several studies has been used to support the use of anti- biotics in acute exacerbations of COPD [22]. In clinical medicine, meta-analyses are accepted and useful tools that combine results from several studies to draw conclusions about clinical effectiveness. CT densitometric progression Table 3 Changes in unadjusted 15th percentile lung density (g•L-1) using endpoint analysis Danish-Dutch trial EXACTLE trial Combined data Statistic AAT (n = 27) Placebo (n = 27) AAT (n = 36) Placebo (n = 35) AAT (n = 60) Placebo (n = 59) Change from baseline to last CT scan, LS mean -6.409 -9.564 -2.645 -4.117 -4.082 -6.379 Estimated treatment difference between changes from baseline, 95% CI† 3.155 (0.008-6.301) 1.472 (0.009-2.935) 2.297 (0.669-3.926) p value for treatment difference 0.049 0.049 0.006 For the CT densitometric analyses, the modified ITT population was used. The combined analysis was based on the modified ITT population and did not include the data for 6 subjects who participated in EXACTLE, but who had their data included in the earlier Danish-Dutch study. †AAT treatment minus placebo. AAT: alpha-1 antitrypsin; CT: computed tomography; EXACTLE: Exacerbations and Computed Tomography scan as Lung Endpoints; LS: least squares. clinical trials completed to date has confirmed that AAT augmentation therapy significantly reduces the decline in lung density, and may thus reduce the future risk of mortality as well as the deterioration in health status. FEV1 does generally relate to loss of lung density, but only if sufficient data are analysed [10]. Extensive obser- vational studies of lung density in AAT deficiency using CT scanning have demonstrated that this parameter not only relates to progressive reduction in FEV1 [10], health status and exercise capacity [11], but is indeed a better predictor of all-cause mortality than FEV1 [13]. It is pos- sible to extrapolate the findings of this combined analy- sis to conventional measures such as the FEV1 using previously published data [10]. This indicates that the reduction in densitometry quantified here (ྜ1 HU/year) is equivalent to about a 38 ml difference in FEV1 decline in patients in GOLD stage 2. FEV1 does generally relate to loss of lung density, but only if sufficient data are analysed [10]. Extensive obser- vational studies of lung density in AAT deficiency using CT scanning have demonstrated that this parameter not only relates to progressive reduction in FEV1 [10], health status and exercise capacity [11], but is indeed a better predictor of all-cause mortality than FEV1 [13]. It is pos- sible to extrapolate the findings of this combined analy- sis to conventional measures such as the FEV1 using previously published data [10]. CT densitometric progression From the Danish-Dutch study, the least squares mean change in PD15 from baseline to endpoint was greater in the placebo group than in the active group (3.155; p = 0.049; Table 3). Combined data from the Danish-Dutch and EXACTLE studies confirmed the reduction in pro- gression in patients receiving augmentation therapy (-6.379 g•L-1 [placebo] versus -4.082 g•L-1 [AAT]; p = 0.006; Figure 3), which is approximately equivalent to -2.74 and -1.73 g•L-1•yr-1, respectively. Therefore, using the most sensitive statistical/endpoint analysis method of volume correction, the separate and integrated analysis of The FEV1 data from this combined study confirm that even doubling the number of subjects is inadequate to verify whether augmentation therapy affects this non- specific and relatively insensitive outcome of emphy- sema. Much larger numbers of subjects studied over a longer period of time are required [4] in order to deter- mine the response of therapy on FEV1, even though longitudinal CT data have confirmed that decline in Stockley et al. Respiratory Research 2010, 11:136 http://respiratory-research.com/content/11/1/136 Page 6 of 8 Stockley et al. Respiratory Research 2010, 11:136 http://respiratory-research.com/content/11/1/136 Table 3 Changes in unadjusted 15th percentile lung density (g•L-1) using endpoint analysis Danish-Dutch trial EXACTLE trial Combined data Statistic AAT (n = 27) Placebo (n = 27) AAT (n = 36) Placebo (n = 35) AAT (n = 60) Placebo (n = 59) Change from baseline to last CT scan, LS mean -6.409 -9.564 -2.645 -4.117 -4.082 -6.379 Estimated treatment difference between changes from baseline, 95% CI† 3.155 (0.008-6.301) 1.472 (0.009-2.935) 2.297 (0.669-3.926) p value for treatment difference 0.049 0.049 0.006 For the CT densitometric analyses, the modified ITT population was used. The combined analysis was based on the modified ITT population and did not include the data for 6 subjects who participated in EXACTLE, but who had their data included in the earlier Danish-Dutch study. †AAT treatment minus placebo. AAT: alpha-1 antitrypsin; CT: computed tomography; EXACTLE: Exacerbations and Computed Tomography scan as Lung Endpoints; LS: least squares. Authors’ contributions The Jadad scale is widely used to assess the methodo- logical quality of clinical trials [27,28]. When evaluated on this scale, the design of the 2 studies met the stan- dards required for their results to be included in a meta- or integrated analysis. Although the principle of meta- or integrated analyses is based on the inclusion of several studies, p values are reported without statistical adjustment of the alpha level. RAS was an investigator in the EXACTLE study and proposed the combined analysis. He wrote the first draft of the manuscript and has fine-tuned the final version, following input from all co-authors and with subsequent support from a medical writer. DGP has been involved in the methodology for CT analysis of the EXACTLE study and the integrated data. He has revised the submitted article for important intellectual content, and has approved the final version. EP was responsible for the Swedish arm of the EXACTLE study. She has reviewed and approved the manuscript. JS was an investigator in the Dutch part of the Danish-Dutch study and was involved in the design of the EXACTLE study. He has revised the submitted article critically for important intellectual content, and has provided final approval of the version to be published. BCS has been involved in the methodology for CT analysis used in both studies. He has revised the submitted article critically for important intellectual content, and has provided final approval of the version to be published. AD was the principal investigator of the 2 multicentre, randomised clinical trials of augmentation therapy with AAT. He has revised the submitted article critically for important intellectual content, and has provided final approval of the version to be published. All authors have read and approved the final manuscript. Integrating the data from the 2 studies increased the numbers and hence the power of the observations. By using the most sensitive method for assessing emphy- sema progression (as measured by tissue loss) with end- point analysis of PD15, the mean data demonstrate a deceleration of lung tissue loss with AAT augmentation therapy with a high degree of statistical significance. It is, however, recognised that progression even in CT densitometry varies between individuals. Thus adequate historical data will remain a prerequisite to therapeutic decision making. Furthermore, it should be noted that the treatment effect may not be demonstrable for the first 12 months of therapy [14,15]. Disclosure of prior abstract publications Disclosure of prior abstract publications Abstracts of this study have been published by the American Thoracic Society (Am J Respir Crit Care Med, Apr 2008;177), and by the European Respiratory Society (Eur Respir J, Oct 2008;32(Supplement 52):738s). reliable than aggregate meta-analyses and may thus lead to different conclusions [23,24]. This approach has been used more frequently in recent years [24] and also allows, as aggregate analyses similarly do, for the inclusion of dif- ferent drug substances belonging to the same drug class, and different predefined clinical endpoints in the source studies [25,26], provided that the studies have common characteristics to enable the pooling of data. Author details 1 1Lung Investigation Unit, University Hospitals of Birmingham, Edgbaston, Birmingham B15 2TH, UK. 2Department of Respiratory Medicine, University Hospitals of Coventry and Warwickshire, Clifford Bridge Road, Coventry CV2 2DX, UK. 3Department of Respiratory Medicine, Malmö University Hospital, Lund University, Malmö, 205 02, Sweden. 4Leiden University Medical Center, Albinusdreef 2, 2333 ZA Leiden, The Netherlands. 5Gentofte Hospital, Copenhagen University, DK-2900 Hellerup, Denmark. Acknowledgements Support Statement Although there were some differences in study charac- teristics, the EXACTLE and Danish-Dutch trials both had a randomised, placebo-controlled, blinded, parallel design and had a similar CT scan protocol. The 2 stu- dies were comparable with regard to treatment drug, treatment duration and patient characteristics. There is a general belief that maintaining AAT above a protective level of 11 μM is the key to a successful therapeutic out- come, and both studies had treatment regimens that are able to maintain protective levels of AAT, either consis- tently, or for at least 3 out of the 4 weeks in the monthly regimen used in the Danish-Dutch trial [14]. This study was sponsored by Talecris Biotherapeutics, Inc. (Research Triangle Park, NC 27709, USA). This study was sponsored by Talecris Biotherapeutics, Inc. (Research Triangle Park, NC 27709, USA). Technical editorial assistance was provided under the direction of the authors by M Kenig at PAREXEL (Worthing, UK) and was supported by Talecris Biotherapeutics, Inc. Authors’ contributions The exact reasons remain unknown but it is possible that a period of time is required to reverse the established, destructive inflam- matory process. This observation clearly has potential impact on the design of future phase 2 and 3 studies in AAT deficiency and support an end point analysis as the best primary outcome. CT densitometric progression These can be either based on the analysis of published data (so-called ‘aggregated analysis’) or by pooling individual patient data (also termed ‘integrated analysis’) [23]. Trials with different protocols, but with common characteris- tics, can be pooled for these analyses. An integrated ana- lysis based on individual patient data offers numerous advantages over the use of aggregated data; it is more Figure 3 Progression of emphysema in AAT-treated versus placebo-treated subjects (modified ITT). Estimated treatment difference between mean changes in unadjusted 15th percentile lung density from baseline. AAT: alpha-1 antitrypsin; LS: least squares; PD15: 15th percentile lung density. Stockley et al. Respiratory Research 2010, 11:136 http://respiratory-research.com/content/11/1/136 Page 7 of 8 References Burrows B: A clinical trial of efficacy of antiproteolytic therapy: can it be done? Am Rev Respir Dis 1983, 127:S42-S43. 5. Burrows B: A clinical trial of efficacy of antiproteolytic therapy: can it be done? Am Rev Respir Dis 1983, 127:S42-S43. 28. Olivo SA, Macedo LG, Gadotti IC, Fuentes J, Stanton T, Magee DJ: Scales to assess the quality of randomized controlled trials: a systematic review. Phys Ther 2008, 88:156-175, Epub 2007 Dec 11. 6. American Thoracic Society/European Respiratory Society Statement: Standards for the diagnosis and management of individuals with alpha- 1 antitrypsin deficiency. Am J Respir Crit Care Med 2003, 168:818-900. doi:10.1186/1465-9921-11-136 Cite this article as: Stockley et al.: Therapeutic efficacy of alpha-1 antitrypsin augmentation therapy on the loss of lung tissue: an integrated analysis of 2 randomised clinical trials using computed tomography densitometry. Respiratory Research 2010 11:136. doi:10.1186/1465-9921-11-136 Cite this article as: Stockley et al.: Therapeutic efficacy of alpha-1 antitrypsin augmentation therapy on the loss of lung tissue: an integrated analysis of 2 randomised clinical trials using computed tomography densitometry. Respiratory Research 2010 11:136. y y 7. Dawkins PA, Dawkins CL, Wood AM, Nightingale PG, Stockley JA, Stockley RA: Rate of progression of lung function impairment in alpha-1- antitrypsin deficiency. Eur Respir J 2009, 33:1338-1344. 7. Dawkins PA, Dawkins CL, Wood AM, Nightingale PG, Stockley JA, Stockley RA: Rate of progression of lung function impairment in alpha-1- antitrypsin deficiency. Eur Respir J 2009, 33:1338-1344. 8. Parr DG, Stoel BC, Stolk J, Stockley RA: Pattern of emphysema distribution in alpha1-antitrypsin deficiency influences lung function impairment. Am J Respir Crit Care Med 2004, 170:1172-1178. 9. Dirksen A, Friis M, Olesen KP, Skovgaard LT, Sørensen K: Progress of emphysema in severe alpha 1-antitrypsin deficiency as assessed by annual CT. Acta Radiol 1997, 38:826-832. 10. Parr DG, Stoel BC, Stolk J, Stockley RA: Validation of computed tomographic lung densitometry for monitoring emphysema in alpha1- antitrypsin deficiency. Thorax 2006, 61:485-490. 11. Dowson LJ, Newall C, Guest PJ, Hill SL, Stockley RA: Exercise capacity predicts health status in alpha(1)-antitrypsin deficiency. Am J Respir Crit Care Med 2001, 163:936-941. 12. Dowson LJ, Guest PJ, Hill SL, Holder RL, Stockley RA: High-resolution computed tomography scanning in α1-antitrypsin deficiency: relationship to lung function and health status. Eur Respir J 2001, 17:1097-1104. 13. Dawkins PA, Dowson LJ, Guest PJ, Stockley RA: Predictors of mortality in alpha 1-antitrypsin deficiency. Thorax 2003, 58:1020-1026. 14. Received: 4 June 2010 Accepted: 5 October 2010 Published: 5 October 2010 21. Drummond MF, Wilson DA, Kanavos P, Ubel P, Rovira J: Assessing the economic challenges posed by orphan drugs. Int J Technol Assess Health Care 2007, 23:36-42. Stockley et al. Respiratory Research 2010, 11:136 http://respiratory-research.com/content/11/1/136 Stockley et al. Respiratory Research 2010, 11:136 http://respiratory-research.com/content/11/1/136 from the Renfrew and Paisley prospective population study. BMJ 1996, 313:711-715. from the Renfrew and Paisley prospective population study. BMJ 1996, 313:711-715. randomised clinical trials of augmentation therapy with alpha-1 antitrypsin that are integrated in the manuscript, has received grant monies from Bayer and Talecris Biotherapeutics, and has participated in travel and meetings sponsored by Bayer and Talecris. Furthermore, he has received grant funding from the Danish Lung Association for a PhD, who shall analyse data from the Danish Lung Cancer Screening Trial that has no relation to the manuscript. randomised clinical trials of augmentation therapy with alpha-1 antitrypsin that are integrated in the manuscript, has received grant monies from Bayer and Talecris Biotherapeutics, and has participated in travel and meetings sponsored by Bayer and Talecris. Furthermore, he has received grant funding from the Danish Lung Association for a PhD, who shall analyse data from the Danish Lung Cancer Screening Trial that has no relation to the manuscript. 19. Holmes J, Stockley RA: Radiologic and clinical features of COPD patients with discordant pulmonary physiology: lessons from a-1-antitrypsin deficiency. Chest 2007, 132:909-915. 20. Klein JS, Gamsu G, Webb WR, Golden JA, Müller NL: High-resolution CT diagnosis of emphysema in symptomatic patients with normal chest radiographs and isolated low diffusing capacity. Radiology 1992, 182:817-821. Received: 4 June 2010 Accepted: 5 October 2010 Published: 5 October 2010 References 22. Saint SK, Bent S, Vittinghoff E, Grady D: Antibiotics in chronic obstructive pulmonary disease exacerbations: a meta-analysis. JAMA 1995, 273:957-960. 1. Snider GL, Lucey EC, Christensen TG, Stone PJ, Calore JD, Catanese A, Franzblau C: Emphysema and bronchial secretory cell metaplasia induced in hamsters by human neutrophil products. Am Rev Respir Dis 1984, 129:155-160. 23. Stewart LA, Parmar MK: Meta-analysis of the literature or of individual patient data: is there a difference? Lancet 1993, 341:418-422. 2. Seersholm N, Wencker M, Banik N, Viskum K, Dirksen A, Kok-Jensen A, Konietzko N: Does alpha1-antitrypsin augmentation therapy slow the annual decline in FEV1 in patients with severe hereditary alpha1- antitrypsin deficiency? Eur Respir J 1997, 10:2260-2263. 2. Seersholm N, Wencker M, Banik N, Viskum K, Dirksen A, Kok-Jensen A, Konietzko N: Does alpha1-antitrypsin augmentation therapy slow the annual decline in FEV1 in patients with severe hereditary alpha1- antitrypsin deficiency? Eur Respir J 1997, 10:2260-2263. 24. Riley RD, Lambert PC, bo-Zaid G: Meta-analysis of individual participant data: rationale, conduct, and reporting. BMJ 2010, 340:c221. 25. Liu PY, Swerdloff RS, Christenson PD, Handelsman DJ, Wang C: Rate, extent, and modifiers of spermatogenic recovery after hormonal male contraception: an integrated analysis. Lancet 2006, 367:1412-1420. y y 3. Wewers MD, Casolaro MA, Sellers SE, Swayze SC, McPhaul KM, Wittes JT, Crystal RG: Replacement therapy for alpha 1-antitrypsin deficiency associated with emphysema. N Engl J Med 1987, 316:1055-1062. 3. Wewers MD, Casolaro MA, Sellers SE, Swayze SC, McPhaul KM, Wittes JT, Crystal RG: Replacement therapy for alpha 1-antitrypsin deficiency associated with emphysema. N Engl J Med 1987, 316:1055-1062. 26. Sin DD, Wu L, Anderson JA, Anthonisen NR, Buist AS, Burge PS, Calverley PM, Connett JE, Lindmark B, Pauwels RA, Postma DS, Soriano JB, Szafranski W, Vestbo J: Inhaled corticosteroids and mortality in chronic obstructive pulmonary disease. Thorax 2005, 60:992-997. 4. Schluchter MD, Stoller JK, Barker AF, Buist AS, Crystal RG, Donohue JF, Fallat RJ, Turino GM, Vreim CE, Wu MC: Feasibility of a clinical trial of augmentation therapy for alpha(1)-antitrypsin deficiency. The Alpha 1- Antitrypsin Deficiency Registry Study Group. Am J Respir Crit Care Med 2000, 161(3 pt 1):796-801. obstructive pulmonary disease. Thorax 2005, 60:992-997. 27. Jadad AR, Moore RA, Carroll D, Jenkinson C, Reynolds DJ, Gavaghan DJ, McQuay HJ: Assessing the quality of reports of randomized clinical trials: Is blinding necessary? Control Clin Trials 1996, 17:1-12. 5. Competing interests Robert A Stockley has received an unrestricted grant from Talecris Biotherapeutics for the Alpha-1 Detection and Programme for Treatment (ADAPT UK registry). He has advised Baxter and Kamada on their augmentation programmes and received international lecture fees from Talecris. He has lectured widely as part of pharmaceutical sponsored symposia, sat on numerous advisory boards for drug design and trial implementation and received non-commercial grant funding from some companies. David G Parr has served on company advisory board meetings for Talecris Biotherapeutics and acts as a consultant on the technical steering committees of Talecris Biotherapeutics and F Hoffmann-La Roche. He has received honoraria and payment of expenses from Talecris Biotherapeutics for presentations at international meetings. Eeva Piitulainen has no conflicts of interest to disclose. Jan Stolk has served on company advisory board meetings of various companies and served as consultant to some of them. Fees were directly donated to the bank account of the Alpha-1 International Registry Foundation. Berend C Stoel has received honoraria for presentations from Talecris Biotherapeutics. He is a consultant for Roche Pharmaceuticals, Talecris Biotherapeutics, Bioclinica and CSL Behring. His institution has received grant monies from Bio-Imaging (now Bioclinica), Roche, Talecris and Medis Medical Imaging Systems for a research project. Asger Dirksen, as the principal investigator of the 2 multicenter, In conclusion, the overall results are supportive of the efficacy of AAT augmentation therapy and, importantly, provide confirmatory data to power and analyse future alternative strategies for which long-term IV placebo arms cannot be justified ethically. Page 8 of 8 doi:10.1186/1465-9921-11-136 Cite this article as: Stockley et al.: Therapeutic efficacy of alpha-1 antitrypsin augmentation therapy on the loss of lung tissue: an integrated analysis of 2 randomised clinical trials using computed tomography densitometry. Respiratory Research 2010 11:136. References Dirksen A, Dijkman JH, Madsen F, Stoel B, Hutchison DC, Ulrik CS, Skovgaard LT, Kok-Jensen A, Rudolphus A, Seersholm N, Vrooman HA, Reiber JH, Hansen NC, Heckscher T, Viskum K, Stolk J: A randomized clinical trial of alpha 1-antitrypsin augmentation therapy. Am J Respir Crit Care Med 1999, 160:1468-1472. Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color ﬁgure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit Submit your next manuscript to BioMed Central and take full advantage of: 15. Dirksen A, Piitulainen E, Parr DG, Deng C, Wencker M, Shaker SB, Stockley RA: Exploring the role of CT densitometry: a randomised study of augmentation therapy in alpha-1 antitrypsin deficiency. Eur Respir J 2009, 33:1345-1353. 16. The Alpha-1-Antitrypsin Deficiency Registry Study Group: Survival and FEV1 decline in individuals with severe deficiency of alpha 1-antitrypsin. Am J Respir Crit Care Med 1998, 158:49-59. 17. Ketelaars CA, Schlosser MA, Mostert R, Huyer Abu-Saad H, Halfens RJ, Wouters EF: Determinants of health-related quality of life in patients with chronic obstructive pulmonary disease. Thorax 1996, 51:39-43. 17. Ketelaars CA, Schlosser MA, Mostert R, Huyer Abu-Saad H, Halfens RJ, Wouters EF: Determinants of health-related quality of life in patients with chronic obstructive pulmonary disease. Thorax 1996, 51:39-43. 18. Hole DJ, Watt GC, Davey-Smith G, Hart CL, Gillis CR, Hawthorne VM: Impaired lung function and mortality risk in men and women: findings 18. Hole DJ, Watt GC, Davey-Smith G, Hart CL, Gillis CR, Hawthorne VM: Impaired lung function and mortality risk in men and women: findings
https://openalex.org/W4240104939	http://vital.lib.tsu.ru/vital/access/services/Download/koha:000720359/SOURCE1	Russian	null	МЕХАНОХИМИЧЕСКОЕ ЛЕГИРОВАНИЕ ПОРОШКОВОЙ СМЕСИ ТИТАН - НИКЕЛИДА ТИТАНА И ИЗМЕНЕНИЕ ФАЗОВОГО СОСТАВА	Международная конференция "Физическая мезомеханика. Материалы с многоуровневой иерархически организованной структурой и интеллектуальные производственные технологии"	2,021	cc-by	905	DOI: 10.17223/978-5-907442-03-0-2021-059 МЕХАНОХИМИЧЕСКОЕ ЛЕГИРОВАНИЕ ПОРОШКОВОЙ СМЕСИ ТИТАН - НИКЕЛИДА ТИТАНА И ИЗМЕНЕНИЕ ФАЗОВОГО СОСТАВА 1,2Абдульменова Е.В., 2Поповичев К.Е., 1,2Кульков С.Н. 1Национальный исследовательский Томский политехнический университет, Томск 2Институт физики прочности и материаловедения СО РАН, Томск у у 1Национальный исследовательский Томский политехнический университет, Томс 2Институт физики прочности и материаловедения СО РАН, Томск Согласно диаграмме состояния [1] системы Ti-Ni, существует двухфазная область в диапазоне содержания никеля ~ 38-56 % масс., которая состоит из Ti2Ni и TiNi, поэтому методом механохимического легирования порошка никелида титана вблизи эквиатомного состава титаном с последующим отжигом возможно варьировать содержание этих фаз. Цель работы – исследовать структуру и фазовый состав порошка никелида титана после механохимического легирования титаном. Механохимическое легирование промышленного порошка никелида титана вблизи эквиатомного состава (ПН55Т45) промышленным порошков титана (ПТО-1) проводили в планетарной шаровой мельнице АГО-2 в течение 300 секунд (при 60 g). Отжиг полученных смесей проводили в вакуумной печи СНВЭ 1.3.1/16 при 1000 °С, выдержка составляла 30 минут. Распределения частиц по размерам было построено на основе обработки изображений, полученных на растровом микроскопе, результат представлен на рис. 1. Средние размеры частиц исходных порошков ПН55Т45 и ПТО-1 составляли 11,1 мкм (дисперсия по размеру 7,5 мкм) и 6,5 мкм (дисперсия по размеру 6,7 мкм) соответственно. Из рис. 1а видно, что при добавлении 7 % масс. титана в исходную смесь, средний размер составляет 9,6 мкм (дисперсия по размеру 8,2 мкм), по-видимому, вследствие разрушения агломератов в процессе механической обработки и маленького размера частиц титана. С увеличением содержания титана в исходной смеси (рис. 1б), после отжига средний размер увеличивается до 19,2 мкм, вероятно, это связано как с агрегацией более мелких частиц титана в процессе механической обработки, так и с увеличение размера частиц после последующего отжига, так порошок ПН55Т45 после отжига имеет средний размер частиц 15,7 мкм (дисперсия по размеру 11,0 мкм), что больше по сравнению с исходным состоянием. а б Рис. 1. Распределения частиц по размерам, полученных порошков после отжига: 7 (а) и 25 % масс (б) титана в исходной смеси б а а а б Рис. 1. Распределения частиц по размерам, полученных порошков после отжига: 7 (а) и 25 % масс (б) титана в исходной смеси Анализ фазового состава был проведён на дифрактометре с CuKα излучением, типичная рентгенограмма, полученного порошка после отжига, представлена на рис. 2. Видно, что присутствуют дифракционные отражения фазы Ti2Ni и фазы TiNi, которая при комнатной температуре может существовать в трёх формах B2, B19`, R согласно [1,2]. При этом, параметр решётки фазы Ti2Ni с увеличением содержания введенного титана в исходную смесь уменьшается с 1,1294 ± 5·10-4 нм до 1,1251 ± 5·10-4 нм, что несколько меньше чем в [3]. МЕЖДУНАРОДНАЯ КОНФЕРЕНЦИЯ «Физическая мезомеханика. Материалы с многоуровневой иерархически организованной структурой и интеллектуальные производственные технологии» 6–10 сентября 2021 г. Томск, Россия 1. Massalski T.B., Murray J.L., Bennett L.H., Baker H. Binary alloy phase diagrams // ASM International. 1990. V. 3. P. 2874 – 2876. 2. Otsuka K., Sawamura T., Shimizu K. Crystal structure and internal defects of equatomic TiNi martensite // Physica Status Solidi. 1971. V. 5. P. 457-470. 3. Abdulmenova E.V. Kulkov S.N. Mechanical high-energy treatment of TiNi powder and phase changes after electrochemical hydrogenation // International journal of hydrogen energy. 2021. V. 46. P. 823-836. 4. Yurko G.A., Barton J.W., Parr J.G. The crystal structure of Ti2Ni // Acta Crystallographica. 1959. V. 12. P. 909-911. y g j y g gy A., Barton J.W., Parr J.G. The crystal structure of Ti2Ni // Acta Crystallographica. 1959. V. 12. P. 909-911 DOI: 10.17223/978-5-907442-03-0-2021-059 МЕХАНОХИМИЧЕСКОЕ ЛЕГИРОВАНИЕ ПОРОШКОВОЙ СМЕСИ ТИТАН - НИКЕЛИДА ТИТАНА И ИЗМЕНЕНИЕ ФАЗОВОГО СОСТАВА 1,2Абдульменова Е.В., 2Поповичев К.Е., 1,2Кульков С.Н. 1Национальный исследовательский Томский политехнический университет, Томск 2Институт физики прочности и материаловедения СО РАН, Томск Для того, чтобы оценить относительное содержание этих фаз, определялись интегральные интенсивности всех фаз в диапазоне углов от 15 до 100 ° и строилась зависимость отношения интегральных интенсивностей фаз Ti2Ni к TiNi (B2+B19`+R) от содержания введенного в 104 исходную смесь титана. Из зависимости видно, что пересечение аппроксимирующих прямых при содержании титана 15,3 % масс., соответствует составу с максимальным содержанием фазы Ti2Ni. Согласно диаграмме состояния системы Ti-Ni [4] при оценке содержания фаз в полученных порошках по правилу «отрезков» обнаружено, что при добавлении 25 % масс. титана к порошку никелида титана вблизи эквиатомного состава при отжиге формируется фаза Ti2Ni с относительным содержанием 71 % и фаза TiNi (B2+B19`+R) с относительным содержанием 29 %, в то время как при добавлении 7 % масс. титана к порошку никелида титана формируется фаза Ti2Ni с относительным содержанием 23 % и фаза TiNi (B2+B19`+R) с относительным содержанием 79 %. По-видимому, такое расхождение может быт связано с градиентной структурой частиц порошков после механического легирования, когда на поверхности частиц присутствует фаза Ti2Ni, а внутри частиц фаза TiNi (B2, B19`, R). Размер ОКР фазы Ti2Ni c увеличением титана не изменяется и составляет 53 ± 10 нм. Рис. 2. Зависимость отношения интегральных интенсивностей фаз TiNi (B2+B19`+R) к Ti2Ni от содержания титана в исходных смесях после отжига и типичная рентгенограмма, полученного порошка после отжига состава TiNi-25 % масс Рис. 2. Зависимость отношения интегральных интенсивностей фаз TiNi (B2+B19`+R) к Ti2Ni от содержания титана в исходных смесях после отжига и типичная рентгенограмма, полученного порошка после отжига состава TiNi-25 % масс Таким образом, показано, что методом механохимического легирования с последующим отжигом получены порошки с разным содержанием фазы Ti2Ni. Максимальное содержание фазы Ti2Ni формируется при добавлении 15 % масс. титана к порошку никелида титана. Работа выполнена при финансовой поддержке гранта РФФИ № 19-38-90196 Аспиранты в части проведения механического легирования и в рамках государственного задания ИФПМ СО РАН, тема номер FWRW-2021-0005 в части проведения исследований размера частиц и рентгенофазового/рентгеноструктурного анализа. 105
https://openalex.org/W4389606907	https://journals.iaepan.pl/sa/article/download/3605/3282	English	null	Yampil barrows from the fourth and IIIrd millenium BC in the light of Polish-Ukrainian investigations 2010-2014	Sprawozdania Archeologiczne	2,023	cc-by	11,972	Sprawozdania Archeologiczne 75/1, 2023 pl issn 0081-3834 DOI: 10.23858/SA75.2023.1.3605 Sprawozdania Archeologiczne 75/1, 2023 pl issn 0081-3834 DOI: 10.23858/SA75.2023.1.3605 Aleksander Kośko1, Viktor Klochko2, Mikhailo Potupchyk3, Piotr Włodarczak4, Danuta Żurkiewicz5 1. Introduction For more than a century, research on the prehistory of Central Europe has delved into the migration patterns of mobile (nomadic or semi-nomadic) peoples originating from the North Pontic region. This research explores their role in driving cultural and technological innovations, ultimately leading to the emergence of a new form of Eneolithic societies in the 4th millennium BC. This investigation also encompasses the origins of Early Bronze Age European civilization. Over the past 15 years, considerable attention has been dedi- cated to examining the expansion of populations exhibiting new genetic characteristics. This attention is rooted in the striking genetic similarities between the people of the Yamna culture (YC) and the Corded Ware culture (CWC; Allentoft et al. 2015; Haak et al. 2015). A significant part of this research has focused on reconstructing prehistoric events that occurred in the border regions between the East and the West, where human groups first made contact and new societies began to take shape. One of these critical crossroads con- necting two worlds is the region of Podillia, which is currently part of Ukraine, and to a lesser extent, Transnistria. The environmental boundary separating Central and Eastern Europe runs through Podillia (e.g., Makohonienko 2009). Historically, this region served as a frontier for the settlement of steppe communities in the North Pontic zone during the 4th and 3rd millennia BC. It also acted as an ecological barrier, restricting the expansion of Central European settlements during the Neolithic and early Eneolithic periods. Nevertheless, as far back as the beginning of Neolithization in the 6th millennium BC, this barrier was traversed by migrating communities seeking favourable settlement areas and engaging in expeditions for trade and raw material procurement. Throughout the mil- lennia, convenient communication routes traversed Podillia (Kośko and Klochko 2009; Makohonienko 2009). In the 4th millennium BC, the population of the Funnel Beaker cul- ture expanded eastward in this region (Rybicka 2017, 152, fig. 86). The existence of Podil- lian contact routes is further substantiated by the presence of Trypillia culture features in artefacts found in Central European areas (Kośko 1981; Kośko and Szmyt 2009). In the latter half of the 4th millennium, there was an observable westward expansion of settle- ments from phase C/II of the Trypillia culture (Rybicka 2017). This period also marked the emergence of the first barrow cemeteries in the Podillia region. Abstract Abstract Kośko A., Klochko V., Potupchyk M., Włodarczak P. and Żurkiewicz D. 2023. Yampil barrows from the fourth and third millennia BC in the light of Polish-Ukrainian investigations 2010-2014, Sprawozdania Archeologiczne 75/1, 247-281. In the vicinity of Yampil (Vinnytsia oblast, Ukraine), there exists a cluster of barrows dating back to the Eneo- lithic and Early Bronze Age. Nestled upon the Podillia Upland, this concentration lies at the crossroads of two cultural spheres: the Eastern European steppe and Central European region. The exploration of the Yampil bar- rows began during the 1980s by archaeologists from Vinnytsia. This endeavour was enriched by a Polish-Ukrai- nian expedition that conducted fieldwork from 2010 to 2014. Seven barrows were then examined. Today, an abundance of radiocarbon data allows the construction of a precise chronological framework for the Yampil bar- row graves. We can now discern four principal stages in this sequence: (1) Late Eneolithic, (2) early Yamna, (3) late Yamna era, and (4) Catacombna. During the first two periods (3350-2800 cal BC), these barrows were me- ticulously constructed, sometimes evolving in multiple phases. In the latter two stages (2800-2400 cal BC), cemeteries took shape, marked by graves deliberately dug into the fully formed mounds. Keywords: late Eneolithic, Early Bronze Age, Podolia, barrows, Yamna culture, Yampil 1 Adam Mickiewicz University, Faculty of Archaeology, Uniwersytetu Poznańskiego st. 7, 61-614 Poznań, Poland; antokol@amu.edu.pl, ORCID: 0000-0002-0529-3004 2 Department of Archaeology National University of Kyiv-Mohyla Academy, Skovorody 2, 04665 Kyiv, Ukraine; vklochko@ukr.net, ORCID: 0000-0002-7557-3829 3 Department of Culture and Arts; Khmelnytske St. 7, 21036 Vinnytsia, Ukraine; m_potupchik@ukr.net; ORCID: 0009-0005-6839-3845 4 Institute of Archaeology and Ethnology, Polish Academy of Sciences, Sławkowska st. 17, 31-016 Kraków, Poland; p.wlodarczak@iaepan.edu.pl; ORCID: 0000-0003-0359-7386 5 Adam Mickiewicz Univeristy, Faculty of Archaeology, Uniwersytetu Poznańskiego st. 7, 61-614 Poznań, Poland; danuta@amu.edu.pl, ORCID: 0000-0002-5219-0622 4 Institute of Archaeology and Ethnology, Polish Academy of Sciences, Sławkowska st. 17, 31-016 Kraków, Poland; p.wlodarczak@iaepan.edu.pl; ORCID: 0000-0003-0359-7386 5 Adam Mickiewicz Univeristy, Faculty of Archaeology, Uniwersytetu Poznańskiego st. 7, 61-614 Poznań, Poland; danuta@amu.edu.pl, ORCID: 0000-0002-5219-0622 5 Adam Mickiewicz Univeristy, Faculty of Archaeology, Uniwersytetu Poznańskiego st. 7, 61-614 Poznań, Poland; danuta@amu.edu.pl, ORCID: 0000-0002-5219-0622 This article is published in an open-access under the CC BY 4.0 license (https: //creativecommons.org/licenses/by/4.0/). This article is published in an open-access under the CC BY 4.0 license (https: //creativecommons.org/licenses/by/4.0/). 248 A. Kośko, V. Klochko, M. Potupchyk, P. Włodarczak, D. Żurkiewicz Fig. 1. Porohy, Yampil rayon, Barrow 2, Grave 6. Fig. 1. Porohy, Yampil rayon, Barrow 2, Grave 6. Burial of YC (1) equipped with a Corded Ware amphora (2, 3). After Harat et al. 2014 1. Introduction In the early 3rd millen- nium BC, two expansive movements converged in this area: the Globular Amphora culture in the east (Szmyt 1999) and the YC in the west (Włodarczak 2017). Moreover, around 2900-2800 BC, CWC communities appeared in the western part of Podillia, along with the introduction of funerary practices involving the construction of barrows. The easternmost barrows associated with the CWC are situated along the Zbruch River, a tributary of the Dnister. The origins of barrow communities in this region have become an intriguing issue in the context of research on the genesis of the CWC and the advent of the Bronze Age in Central Europe (this work follows the primary archaeological period divisions commonly used in studies of the North Pontic region). 249 Yampil barrows from the fourth and third millennia BC in the light of Polish-Ukrainian... 249 249 p g Fig. 1. Porohy, Yampil rayon, Barrow 2, Grave 6. Burial of YC (1) equipped with a Corded Ware amphora (2, 3). After Harat et al. 2014 1 2 3 1 1 1 2 3 2 3 3 2 250 A. Kośko, V. Klochko, M. Potupchyk, P. Włodarczak, D. Żurkiewicz One vital region for investigating the aforementioned issues of intercultural relations during the late 4th and the first half of the 3rd millennium BC is the territory surrounding the town of Yampil, located in the middle Dnister region (Vinnytsia Oblast). This area en- compasses the westernmost cluster of YC burial mounds within the Podillia Upland. Grave inventories from this region include amphorae with stylistic and technological connec- tions to Central European cultures (Kośko 2011, 184, 185, 188, fig. 6). Notably, a vessel from Grave 6 in Barrow 2 in Porohy (Fig. 1) exhibits an undeniable resemblance to the amphorae from the A-horizon of the CWC (Ivanova et al. 2014). This evidence strongly suggests that the Podillia region along the middle Dnister basin served as a hub of exten- sive interactions between Eastern and Western communities. It might also have been a place where communities with new cultural features, characteristic of the Central European Fi- nal Eneolithic (CWC), began to take shape. Recognizing the significance of this issue in prehistoric studies, the Yampil barrows became the focus of research for a Polish-Ukrainian field expedition conducted from 2010 to 2014. 1. Introduction This article aims to provide a summary of chronometric data and a review of taxo- nomic findings concerning grave materials excavated during that period. 2. Yampil Barrow Complex The Yampil rayon area lies within the forest-steppe zone, situated in the middle Dnis- ter basin on the fringes of the Podillia Upland. The landscape is profoundly influenced by its loess subsoil, currently covered by chernozem, resulting in a terrain characterized by deep, branching gorges and plateau-like watershed areas (Makohonienko and Hildeb- randt-Radke 2014). This topography offers expansive vistas, with sweeping views extend- ing for tens of kilometres across gentle hills. In contrast, the deep valleys of rivers and streams in the Podillia Upland remain hidden from view due to the elevated plateaus. These visual attributes are shaped by the prevailing vegetation, predominantly meadows of a steppe nature, although many have now transformed into arable fields and pastures. Historically, these areas were favoured by nomadic herding communities. Thanks to the efforts of the archaeological heritage services in Vinnytsia, 156 burial mounds have been documented within the administrative boundaries of the Yampil re- gion, covering approximately 790 square kilometres (Jachimowicz and Żurkiewicz 2017, 11, 12). It is important to note that this picture represents only the mounds identified dur- ing field inspections, as they stand out in relief against the landscape (Fig. 2). Further analysis of satellite images suggests the existence of hundreds of more eroded mounds, now only discernible as dark patches in cultivated fields. Within the Yampil cluster, one finds individual mounds as well as compact groups and chains of barrows. A spatial analysis indicates that these mounds are strategically situated on elevated terrain (Fig. 3), usually facing south and southwest, highlighting the impor- Yampil barrows from the fourth and third millennia BC in the light of Polish-Ukrainian... 251 Yampil barrows from the fourth and third millennia BC in the light of Polish-Ukrainian... 251 251 Yampil barrows from the fourth and third millennia BC in the light of Polish Ukrainian... Fig. 2. Yampil barrow complex. A – barrows, B – barrows excavated by Polish-Ukrainian expedition in 2010-2014 (1 – Pidlisivka, 2 –Porohy, 3 – Klembivka, 4 – Prydnistryanske), C – borders of the Yampil rayon. Illustrated by R. Skrzyniecki Fig. 2. Yampil barrow complex. A – barrows, B – barrows excavated by Polish-Ukrainian expedition in 2010-2014 (1 – Pidlisivka, 2 –Porohy, 3 – Klembivka, 4 – Prydnistryanske), C – borders of the Yampil rayon. Illustrated by R. Skrzyniecki tance of visibility for the builders in establishing a network of relationships throughout the region (Jachimowicz and Żurkiewicz 2017). 2. Yampil Barrow Complex Sixteen burial mounds in the Yampil rayon were excavated between 1984 and 1993 during various rescue operations. Among them, three barrows in Porohy were explored, where three graves containing amphorae with stylistic connections to Central European pottery were discovered (Iwanowa et al. 2014). Unfortunately, a fire in the archaeological A. Kośko, V. Klochko, M. Potupchyk, P. Włodarczak, D. Żurkiewicz A. Kośko, V. Klochko, M. Potupchyk, P. Włodarczak, D. Żurkiewicz 252 Fig. 3. Barrow IV from Prydnistryanske, Yampil rayon (excavated in 2014 by Polish-Ukrainian expedition). Photo by D. Żurkiewicz Fig. 3. Barrow IV from Prydnistryanske, Yampil rayon (excavated in 2014 by Polish-Ukrainian expedition). Photo by D. Żurkiewicz storage facilities has resulted in the loss of bone materials gathered during that time, which could have otherwise been valuable subjects for a variety of analyses. An important accom- plishment of the Polish-Ukrainian project was the collection, preparation, and publication of the research results from the years 1984-1993 (Harat et al. 2014). The primary aim was to gather as much evidence as possible about the barrow complexes, providing a clear un- derstanding of their chronology and attributes. Nevertheless, in order to obtain appropri- ate material for specialized analyzes (including chronometric and bioarchaeological ones), new excavations of the barrows from the Yampil cluster were carried out. Yampil barrows from the fourth and third millennia BC in the light of Polish-Ukrainian... 253 Yampil barrows from the fourth and third millennia BC in the light of Polish-Ukrainian... Yampil barrows from the fourth and third millennia BC in the light of Polish-Ukrainian.. Ukrainian team during the years 2010-2012 was Sergey Razumov, and in 2014, Viktor Klochko took the helm. On the Polish side, Aleksander Kośko managed the expedition, with contributions from Piotr Włodarczak and Danuta Żurkiewicz. Mykhailo Potupchyk from the heritage office in Vinnytsia, a researcher of the Yampil barrows in the 1980s and 1990s, also participated in the fieldwork. The excavations involved the use of mechanical equipment and the application of a meth- odology developed by Ukrainian researchers for the exploration of burial mounds (Kośko and Razumow 2014). The barrow layers were systematically examined within trenches oriented along the west-east axis, with widths ranging from 4.5 to 7 metres. Efforts were made to document the vertical sections of the barrow, both in the central area and on the outskirts of the mound. A comprehensive description of the investigated features, materi- als, and analytical work carried out can be found in two volumes of the “Baltic-Pontic Studies” journal (issues 20 and 22) and in volume 6 of the “Archaeologia Bimaris” series (Kośko ed. 2015; 2017; Kośko et al. eds 2014). Genetic research results on prehistoric Yampil populations within the broader context of Central Europe are discussed in a sepa- rate paper (Juras et al. 2018). A significant objective of the project was to obtain chronometric data for determining the ages of individual cultural phenomena in Yampil, particularly those from the Late En- eolithic and Early Bronze Age periods. A total of over 60 radiocarbon (14C) dates were ob- tained for 40 archaeological features, initially at a laboratory in Kyiv and later at the Poznań Radiocarbon Laboratory. The dated materials primarily included human bones, alongside wood samples, animal bones, and charcoal. Previous publications have on mul- tiple occasions presented the dating results and discussions regarding the chronology of the barrow cemeteries in the Yampil region (e.g., Goslar et al. 2014; Goslar et al. 2015; Wło- darczak 2017). Expanding the 14C dating database and conducting new taxonomic studies offer opportunities to validate earlier findings. The summary of research results presented below for individual burial mounds takes into account several revisions in taxonomic and chronological definitions compared to the previous presentations. These observations pertain solely to burials from the 4th and 3rd millennia BC. Yampil barrows from the fourth and third millennia BC in the light of Polish-Ukrainian... More comprehensive excavation results are published in volume 20 of “Bal- tic-Pontic Studies” from 2015 (Kośko ed. 2015). 3. Excavations of Yampil expedition (2010-2014) Between 2010 and 2014, the Yampil expedition conducted excavations on seven bar- rows at four different sites: Pidlisivka (Klochko et al. 2015a), Porohy (Klochko et al. 2015b), Klembivka (Klochko et al. 2015c), and Prydnistryanske (four burial mounds; Klochko et al. 2015d). This extensive project was made possible through agreements established between the Institute of Archaeology of the National Academy of Science of Ukraine in Kyiv and the Adam Mickiewicz University in Poznań, Poland, with the collaboration of the Institute of Archaeology and Ethnology of the Polish Academy of Sciences in Kraków. Leading the Pidlisivka, Barrow 1 (2010) Barrow 1 was situated on the plateau of the right bank of the Yalanka River, near its confluence with the Markivka, approximately 7 km north of the Dnister valley. At the out- set of the 2010 research, the barrow’s diameter was approximately 30 metres, and its height did not exceed 1 metre due to extensive ploughing. It had also suffered partial de- struction during World War II as it was dug into to create artillery positions. A 4.5-7 metre wide ditch, resulting from soil extraction, delineated the original boundary of the barrow (Kośko et al. 2014; Klochko et al. 2015a). A. Kośko, V. Klochko, M. Potupchyk, P. Włodarczak, D. Żurkiewicz 254 Fig. 4. Pidlisivka, Yampil rayon. Plan of Barrow 1. YC – Yamna culture, CC – Catacombna culture, BC – Babyno culture. Illustrated by D. Żurkiewicz Fig. 4. Pidlisivka, Yampil rayon. Plan of Barrow 1. g p y YC – Yamna culture, CC – Catacombna culture, BC – Babyno culture. Illustrated by D. Żurkiewicz YC – Yamna culture, CC – Catacombna culture, BC – Babyno culture. Illustrated by D. Żurkiewicz The extensive damage to the barrow makes it challenging to reconstruct its chrono- logical phases. The arrangement of the graves suggests that the mound was constructed in two stages, associated with Graves 1A and 1B located in the central portion (Fig. 4). The primary burial for the earlier stage was Grave 1B (Fig. 5), where an adult male was in- terred. This was classified as Late Eneolithic based on the arrangement and orientation of the chamber, as well as the absence of ochre usage in the burial ritual (Klochko et al. 2015a). However, radiocarbon dating of the bones from this burial points to a date in the first half of the 3rd millennium BC (2846-2577 cal BC, 68.3% probability). A similar age range was established for Grave 1A, which was probably dug into a small mound constructed over Grave 1B (Goslar et al. 2014, 308, fig. 4.1:1). An adult man was buried in a rectangular pit, with his head facing west. The skeleton of a child aged 7-8 was also found approximately 20 centimetres above this burial in the same pit. The man’s burial 255 Yampil barrows from the fourth and third millennia BC in the light of Polish-Ukrainian... Fig. 5. Pidlisivka, Yampil rayon, Barrow 1. Grave 1B – the primary burial of the first barrow phase. Photo by D. Porohy, Barrow 3A (2011) Barrow 3A was situated on a prominent promontory at the confluence of the Dnister and Rusava rivers. On the southern part of this headland, there were four mounds in a linear north-south arrangement, including Mound 2, which had been examined in 1984 (Potup- czyk and Razumow 2014, 37, fig. 1.2: 2). There, two royally equipped burials from the Sarmatian period had been discovered (Simonenko and Lobay 1991), along with a YC grave containing an A-amphora of the CWC (Fig. 1; Harat et al. 2014, 84-87). Barrow 3A, the northernmost in this group, had a diameter of about 40 metres. When the research com- menced in 2011, its height was 1.2 metres, with a 1933 topographic map noting a height of 3.6 metres. Therefore, the barrow had undergone significant damage, mainly due to ploughing. The mound’s boundary was marked by a ditch up to 10 metres wide and an average depth of 0.5 metres (Klochko et al. 2015b). The central grave (No. 14) had been destroyed by a substantial modern trench dug by treasure hunters (Fig. 7). This intrusion also obliterated most of Grave 2 (YC), potentially the central burial for the second phase of barrow construction. Taking into account the presence of a stone cromlech, a stele (subsequently used to build the circle – Fig. 7: B), and two large post-holes, the oldest phase of the barrow in Porohy was dated to the Late Eneo- lithic period (Klochko et al. 2015b). Regrettably, an attempt to date the bone remains found in a secondary position in the destroyed Feature 14 was unsuccessful (the obtained age corresponds to the Babyno culture; Poz-74396: 3675±35 BP). Nonetheless, the Eneo- lithic age of Barrow 3A appears likely. Grave 2 (YC) was situated in the central part of the mound, and its discovery depth (0.5 metres) implies that its floor was originally approximately 3 metres below the mound’s summit. Hence, it was most likely that the central grave was linked to the expansion of the barrow’s size. Nevertheless, due to the destruction of the central part, it is impossible to ascertain whether it was related to the second phase or any subsequent phases (given the depth of discovery, the latter is more plausible). Pidlisivka, Barrow 1 (2010) Żurkiewicz Fig. 5. Pidlisivka, Yampil rayon, Barrow 1. Grave 1B – the primary burial of the first barrow phase. Photo by D. Żurkiewicz Fig. 6. Pidlisivka, Yampil rayon, Barrow 1. Grave 1A – the primary burial of the second (?) barrow phase. Photo by D. Żurkiewicz Fig. 6. Pidlisivka, Yampil rayon, Barrow 1. Grave 1A – the primary burial of the second (?) barrow phase. Photo by D. Żurkiewicz A. Kośko, V. Klochko, M. Potupchyk, P. Włodarczak, D. Żurkiewicz 256 exhibited early YC characteristics due to the arrangement, head orientation, and use of ochre (Fig. 6). Presumably, this grave was connected with the second phase of barrow construction, although the precise stratigraphic relationship between Features 1A and 1B remains uncertain. Burials from the Early Bronze Age (Nos 4, 7, 9, 10, and 11), the Babyno culture (Nos 5, 7, 13, and possibly 8), and the early Iron Age (No. 12) were placed within the body of the mound . Beneath the mound, a cluster of cattle bones deposited in a basin-shaped depres- sion was discovered (Feature 6, possibly a sacrificial pit). It was probably associated with the barrow’s construction phase. The Early Bronze Age burials varied in terms of burial construction features, the positioning of the deceased, and the use of ochre. Porohy, Barrow 3A (2011) In addition to Feature 2, eleven other YC graves were discovered (Nos 1, 3, 7, 10-12, 15, 17-20), most of which formed an curved zone encircling the centre of the mound at a dis- tance of roughly 5-7 metres. Outside this area, only graves 18 and 20 were found. 257 Yampil barrows from the fourth and third millennia BC in the light of Polish-Ukrainian... 257 Yampil barrows from the fourth and third millennia BC in the light of Polish-Ukrainian... Fig. 7. Porohy, Yampil rayon. A – plan of Barrow 1, B – stone stele,YC – Yamna culture, BC – Babyno culture. Illustrated by D. Żurkiewicz A B BC B A Fig. 7. Porohy, Yampil rayon. A – plan of Barrow 1, B – stone stele,YC – Yamna culture, BC – Babyno culture. Illustrated by D. Żurkiewicz In the Late Bronze Age, three Babyno culture graves were excavated in the southern part of the barrow (Nos 5, 8, and 22). In the Sarmatian period, grave 21 was dug into its southern edge. Porohy’s burials stand out due to the distinctive nature of the YC funeral rituals (Kloch- ko et al. 2015b; Włodarczak 2017). The deceased were placed in a semi-supine position (Graves 1, 10, and 20) or on their sides (Graves 3, 7, 12, 15, and 17), with only occasional instances of a supine position (Grave 11). No burials featuring early YC characteristics were found. Typologically, the Porohy burials represent the late YC phase (though not the 258 A. Kośko, V. Klochko, M. Potupchyk, P. Włodarczak, D. Żurkiewicz Fig. 8. Porohy, Yampil rayon, Barrow 3A, Grave 10. Barrow of adult woman (late Yamna culture). Photo by P. Włodarczak Fig. 8. Porohy, Yampil rayon, Barrow 3A, Grave 10. Barrow of adult woman (late Yamna culture). Photo by P. Włodarczak latest, as seen in the mounds on the Budzhak steppe), a conclusion supported by the radio- carbon dating discussed in subsequent sections. A unique discovery in Barrow 3A was Grave 10, belonging to a woman aged 25-30 years (Fig. 8). Traces of pigment on the forearm bones of both limbs were documented and interpreted as either paintings (Lorkiewicz-Muszyńska et al. 2017) or being from tattoos preserved due to the corpse’s mummification (Włodarczak 2021). Klembivka, Barrow 1 (2012) Barrow 1 was located on the northern edge of the Yampil cluster, on a high watershed running along the north-south axis between the Rusava and Korytna rivers, about 15 km from the Dnister Valley. It had a diameter of 40 metres and a height of 1.2 metres. The barrow was encircled by a ditch up to 6 metres wide and 0.5 metres deep. During the exca- vations, three graves dating back to the 3rd millennium BC were discovered, along with two sacrificial deposits of animal bones from the same period (Fig. 9). In addition, ten graves from the Late Bronze Age, associated with the Babyno and Noua cultures, were found (Klochko et al. 2015c). 259 Yampil barrows from the fourth and third millennia BC in the light of Polish-Ukrainian... Fig. 9. Klembivka, Yampil rayon. A – plan of Barrow 1, B – stone stele. YC – Yamna culture, BC – Babyno culture, NC – Noua culture. Illustrated by D. Żurkiewicz A B B A Fig. 9. Klembivka, Yampil rayon. A – plan of Barrow 1, B – stone stele. Ż Fig. 9. Klembivka, Yampil rayon. A plan of Barrow 1, B stone stele. – Yamna culture, BC – Babyno culture, NC – Noua culture. Illustrated by D. Żurkiewicz YC – Yamna culture, BC – Babyno culture, NC – Noua culture. Illustrated by D. Żurk Due to significant erosion, it is challenging to accurately reconstruct the stratigraphic sequence of Klembivka. There appear to have been two stages of barrow construction. Grave 15 was the primary burial for the earlier phase. In a non-regular pit, the poorly pre- served remains of a person aged 15-20 years were discovered (Fig. 10). The deceased was placed in a supine position with the lower limbs bent, and the head was oriented to the northwest. The position of the preserved proximal parts of the forearm bones suggests that the upper limbs were extended along the body. A lump of ochre was found near the left A. Kośko, V. Klochko, M. Potupchyk, P. Włodarczak, D. Żurkiewicz 260 Fig. 10. Klembivka, Yampil rayon, Barrow 1. Grave 15 – primary burial of the first barrow phase. Photo by P. Włodarczak Fig. 10. Klembivka, Yampil rayon, Barrow 1. Grave 15 – primary burial of the first barrow phase. Photo by P. Włodarczak shoulder. The remains of the mat on which the deceased was placed were also documented. Klembivka, Barrow 1 (2012) However, no wooden roof construction was found. In general, most aspects of the funeral ritual for Grave 15 point to early YC trends. The only exceptions are the irregular shape of the burial pit and the absence of traces of a wooden roof. Grave 5 was excavated in the central part of the mound constructed above Grave 15. It was probably associated with the second phase of the mound’s construction. The third grave from the Early Bronze Age was Feature 14, also dug into the oldest burial mound (Fig. 11). It is possible that Grave 14 was linked to a second layer of barrow construction, which went unrecorded due to erosion. In a deep square pit, a man (25-30 years) was bur- ied in a tightly contracted position on his left side. The burial was considered Eneolithic (Klochko et al. 2015c, 175), corresponding to type IIIC (Rassamakin 2004, 55-59), typical of the Zhivotilovka-Volchansk group and similar cultures. However, its absolute dating to the beginning of the 3rd millennium BC contradicts this interpretation. On the southeast edge of the older barrow, at an approximate depth of 0.6 metres (i.e., at the original ground level), an anthropomorphic stele made of limestone measuring 0.7 x 0.7 x 0.1 metres was discovered (Fig. 9: B). Moreover, two small concentrations of animal bones were found in two places, potentially related to sacrificial deposits made during the barrow’s construction. A domestic horse bone was uncovered on the eastern edge of the 26 Yampil barrows from the fourth and third millennia BC in the light of Polish-Ukrainian... Yampil barrows from the fourth and third millennia BC in the light of Polish-Ukrainian... Fig. 11. Klembivka, Yampil rayon, Barrow 1. Grave 14. Photo by P. Włodarczak Fig. 11. Klembivka, Yampil rayon, Barrow 1. Grave 14. Photo by P. Włodarczak younger layer of the barrow, and the bone of a small ruminant was found on the eastern edge of the older mound. Additionally, two sacrificial pits with animal bones were associ- ated with the older mound’s construction phase (Features 4 and 9). Subsequently, seven graves of the Babyno culture were dug into the fully formed bar- row (Nos 1-3, 6, 8, 10, and 13), with 14C dates indicating a range from 1880 to 1771 BC. Klembivka, Barrow 1 (2012) Three Noua culture graves (Nos 7, 11, and 12) are dated to a slightly later period (1443-1311 BC), forming a cluster in the southern zone, in the barrow ditch. Prydnistryanske, Barrows I-IV (2014) On a lengthy promontory situated between the Dnister and Markivka rivers, a series of barrows is aligned along a north-south axis. This collection includes four mounds that underwent examination in 2014, as reported by Klochko et al. in 2015. Among them, we A. Kośko, V. Klochko, M. Potupchyk, P. Włodarczak, D. Żurkiewicz 262 Fig. 12. Prydnistryanske, Yampil rayon, Barrow I. Grave 4 – burial of Catacombna culture. Photo by M. Podsiadło Fig. 12. Prydnistryanske, Yampil rayon, Barrow I. Grave 4 – burial of Catacombna culture. Photo by M. Podsiadło find three smaller, significantly eroded mounds (designated as Nos I-III) closely aligned in a row. In addition, there is a larger, better-preserved barrow (No. IV) located approxi- mately 60 metres to the north (Fig. 3). find three smaller, significantly eroded mounds (designated as Nos I-III) closely aligned in a row. In addition, there is a larger, better-preserved barrow (No. IV) located approxi- mately 60 metres to the north (Fig. 3). Barrow I: This barrow, with a diameter of about 20 metres and a remaining height of just 0.3 metres (Klochko et al. 2015d, 188, fig. 5), is encircled by a ditch (a result of soil extraction). Notably, no traces of burial were found in the primary grave (I/1). The fill contained only two fragments of pottery from the late phase of the Trypillia culture, as well as a fragment of a wooden object. In the eastern part of the barrow, a double burial of the Catacombna culture (I/4) was discovered (Fig. 12), and later, two features, including a grave, from the Sarmatian period (I/2 and I/3). Barrow II: This mound, with an oval shape and dimensions of approximately 23 by 20 metres, maintains a height of no more than 0.2 metres (Klochko et al. 2015d, 196, fig. 11). Like Barrow I, no burial traces were identified in the primary grave (II/2), but a few wood fragments were discovered in its fill. Radiocarbon dating of one of these fragments places it in the Late Eneolithic period. In the vicinity of Grave II/2, a hearth (Feature II/1) was found, marked by a cluster of charcoal and burnt earth lumps. In the southern section of the mound, a pit with an unknown purpose (Features II/3) contained limestone rocks, as 263 Yampil barrows from the fourth and third millennia BC in the light of Polish-Ukrainian... Prydnistryanske, Barrows I-IV (2014) 263 Yampil barrows from the fourth and third millennia BC in the light of Polish-Ukrainian... Fig. 13. Prydnistryanske, Yampil rayon. Plan of Barrow III. Illustrated by D. Żurkiewicz Fig. 13. Prydnistryanske, Yampil rayon. Plan of Barrow III. Illustrated by D. Żurkiewicz well as nine human bone fragments in its lower layers. Radiocarbon dating of Feature II/3 indicates a modern age. well as nine human bone fragments in its lower layers. Radiocarbon dating of Feature II/3 indicates a modern age. Barrow III: This circular mound, approximately 24 metres in diameter and with a re- maining height of about 0.15 metres, was enclosed by a basin-shaped ditch (Fig. 13). Lime- stone pieces were discovered on the border of this ditch and the mound, with a greater concentration on the northern edge. Under the mound’s central part, two pits were uncov- ered. Feature III/1 held disturbed remains of an adult individual and fragments of a Gor- dineşti-type vessel (Klochko et al. 2015d, 205, fig. 18). At the bottom of Feature III/3, an amphora, a beaker (of the Gordineşti group pottery type – Fig. 14: 2, 3), and a stone battle- axe (Fig. 14: 1) were found. In the upper part of the barrow, a concentration of human 264 A. Kośko, V. Klochko, M. Potupchyk, P. Włodarczak, D. Żurkiewicz Fig. 14. Prydnistryanske, Yampil rayon, Barrow III. Grave 3 and its inventory. Illustrated by D. Żurkiewicz and M. Podsiadło Fig. 14. Prydnistryanske, Yampil rayon, Barrow III. Grave 3 and its inventory. Illustrated by D. Żurkiewicz and M. Podsiadło remains (an adult and a child in Feature III/2) was observed, possibly originating from the disturbed Grave III/1. In the southeast part of the mound, a human burial dating back to the early Middle Ages was located (Feature III/4). Barrow IV: This mound, boasting a diameter of 35 metres and a height of 2.5 metres, was encircled by a circular ditch up to 15 metres wide (Fig. 15). Regrettably, the eastern edge of the barrow had to remain unexcavated due to the presence of a powerline pole. Stratigraphic analysis revealed three construction phases of the barrow. The oldest phase, dating to the second half of the 4th millennium BC, involved the creation of a small mound with a diameter of 17-19 metres above Grave IV/10, featuring a catacomb construction. Radiocarbon dating of this feature places it in the Late Eneolithic period. Prydnistryanske, Barrows I-IV (2014) The oldest bar- row suffered disturbance through an extensive, irregular trench (Feature IV/11), which obliterated a significant portion of its central area, possibly due to animal burrows. Subse- quent construction phases of the barrow are connected to the YC. A complex grave, associ- ated with the early YC, known as Grave IV/4, was excavated in the eastern part of the Eneolithic barrow (Fig. 16). This elaborate construction, including stone block covering and additional wooden elements, signifies the high status of the interred individual. This burial led to the expansion of the mound, primarily on the eastern side, increasing its di- ameter to approximately 25 metres. A second grave from the YC (Feature IV/6) was exca- 265 Yampil barrows from the fourth and third millennia BC in the light of Polish-Ukrainian... 265 Yampil barrows from the fourth and third millennia BC in the light of Polish-Ukrainian... Fig. 15. Prydnistryanske, Yampil rayon. Plan of Barrow IV. YC –Yamna culture. Illustrated by D. Żurkiewicz A B B A Fig. 15. Prydnistryanske, Yampil rayon. Plan of Barrow IV. YC –Yamna culture. Illustrated by D. Żurkiewicz vated in the mound, around two hundred years later. The final stage of the barrow’s con- struction, marking its ultimate dimensions, is associated with this burial, also from the early YC. Subsequently, three late YC graves were inserted into the southern side (desig- nated as IV/3, IV/8 - Fig. 17, and IV/9). Much later, a Sarmatian period grave (IV/1) was discovered on the southern slope of the barrow. vated in the mound, around two hundred years later. The final stage of the barrow’s con- struction, marking its ultimate dimensions, is associated with this burial, also from the early YC. Subsequently, three late YC graves were inserted into the southern side (desig- nated as IV/3, IV/8 - Fig. 17, and IV/9). Much later, a Sarmatian period grave (IV/1) was discovered on the southern slope of the barrow. A. Kośko, V. Klochko, M. Potupchyk, P. Włodarczak, D. Żurkiewicz 266 Fig. 16. Prydnistryanske, Yampil rayon, Barrow IV. Grave 4 – burial of the early Yamna culture. Photo by M. Podsiadło Fig. 16. Prydnistryanske, Yampil rayon, Barrow IV. Grave 4 – burial of the early Yamna culture. Photo by M. Podsiadło Fig. 17. Prydnistryanske, Yampil rayon, Barrow IV. Grave 8 – burial of the late Yamna culture. Photo by P. Włodarczak Fig. 17. Prydnistryanske, Yampil rayon, Barrow IV. Prydnistryanske, Barrows I-IV (2014) Grave 8 – burial of the late Yamna culture. Photo by P. Włodarczak 267 Yampil barrows from the fourth and third millennia BC in the light of Polish-Ukrainian... 4. Chronometry of the Yampil barrows The list of radiocarbon dates for the Yampil barrows, as published by Goslar et al. in 2015, has been expanded with additional results obtained for Barrow 3A in Porohy and one for Pidlisivka. These findings provide an opportunity to propose a new chronology for the Podillian barrow communities. Previous research had primarily relied on results from four burial mounds in Prydnistryanske to establish the periodization of the Yampil bar- rows, marking the Late Eneolithic stage, late Trypillia Gordineşti, or Zhivotilovka-Vol- chansk groups, followed by the Early and Late/Classic YC phases. Now, a more extensive series of dates from Porohy allows us to refine the chronology of the late YC. Additionally, the taxonomic and chronological assessments of the burials in Klembivka need to be par- tially revised. The following dating lists of the Yampil barrows take into account the results obtained at the Poznań Radiocarbon Laboratory. The dates obtained at the Kyiv labora- tory were previously compiled in a publication by Goslar et al. in 2015. While some of these dates confirm the results presented below, in other cases, they deviate from our ex- pectations. * OxCal calibration program v4.4.4 (Bronk Ramsey 2021) Stage I – Late Eneolithic For the central graves in all four barrows at Prydnistryanske, we obtained similar ab- solute ages, suggesting years around 3350-3150 cal BC (Table 1; Goslar et al. 2015). These dates represent the oldest horizon of barrow graves in Podillia to date (Fig. 18). Determin- ing the taxonomic relationships of the Eneolithic burials from Prydnistryanske proved challenging. The presence of characteristic pottery in Features III/1 and III/3 suggested a connection with the Gordineşti group (Klochko et al. 2015d). Some researchers also as- sociated the barrow ritual observed at various Moldovan sites with this Late Trypillia cul- tural phenomenon (Larina 2003). However, burials from the Zhivotilovka-Volchansk group exhibited similar features, frequently featuring pottery from the Gordineşti group Table 1. Radiocarbon dates of late Eneolithic graves from barrows nos I-IV in Prydnistryanske, Yampil rayon Barrow Grave no. Dated material Lab code Age BP Calendar age BC (68.2%)* Prydnistryanske I 1 Wood Poz-66214 4640 ± 40 3504-3365 Prydnistryanske II 2 Wood Poz-66222 4655 ± 35 3507-3370 Prydnistryanske II 1 Charcoal Poz-66221 4485 ± 30 3331-3100 Prydnistryanske III 1 Human bone Poz-66224 4540 ± 35 3363-3109 Prydnistryanske III 2 Human bone Poz-66225 4530 ± 35 3359-3107 Prydnistryanske III 3 Wood Poz-71367 4510 ± 40 3346-3104 Prydnistryanske IV 10 Human bone Poz-66234 4520 ± 40 3355-3105 * OxCal calibration program v4 4 4 (Bronk Ramsey 2021) A. Kośko, V. Klochko, M. Potupchyk, P. Włodarczak, D. Żurkiewicz 268 Fig. 18. Calibration results of 14C ages of samples from Barrows I-IV at Prydnistryanske (late Eneolithic phase). For the calibration, the Oxcal v4.4.4 software (Bronk Ramsey 2021) was used Fig. 18. Calibration results of 14C ages of samples from Barrows I-IV at Prydnistryanske (late Eneolithic phase). For the calibration, the Oxcal v4.4.4 software (Bronk Ramsey 2021) was used (Rassamakin 2004, 126). No inhumation burials with preserved body positioning were discovered in the graves at Prydnistryanske. The pits were oriented along the NW-SE axis, and in the case of Features III/1 and III/3, it can be suggested that the deceased were placed with their heads facing SE. This orientation aligns with the Zhivotilovka-Volchansk group’s burial ritual (Dergachev 2022, 161). On the other hand, secondary interventions in burial chambers, leading to the decomposition or removal of human remains, are charac- teristic of the Gordineşti group’s ritual and have been documented at sites in northern Moldova near Prydnistryanske (Larina 2003; 66). Stage I – Late Eneolithic The 14C dating of Eneolithic features from Prydnistryanske aligns with the results obtained for the Zhivotilovka-Volchansk group graves in Bursuceni (Petrenko and Kovaliukh 2003) and Cimişlia (Popovici and Ciobanu 2021), as well as for settlements of the Gordineşti group (Rybicka et al. 2020). (Rassamakin 2004, 126). No inhumation burials with preserved body positioning were discovered in the graves at Prydnistryanske. The pits were oriented along the NW-SE axis, and in the case of Features III/1 and III/3, it can be suggested that the deceased were placed with their heads facing SE. This orientation aligns with the Zhivotilovka-Volchansk group’s burial ritual (Dergachev 2022, 161). On the other hand, secondary interventions in burial chambers, leading to the decomposition or removal of human remains, are charac- teristic of the Gordineşti group’s ritual and have been documented at sites in northern Moldova near Prydnistryanske (Larina 2003; 66). The 14C dating of Eneolithic features from Prydnistryanske aligns with the results obtained for the Zhivotilovka-Volchansk group graves in Bursuceni (Petrenko and Kovaliukh 2003) and Cimişlia (Popovici and Ciobanu 2021), as well as for settlements of the Gordineşti group (Rybicka et al. 2020). Stage II – Early Yamna The marker for the early YC in the Podillia region is Grave 4 from Barrow IV in Pryd- nistryanske. It was excavated in the eastern part of the Eneolithic mound and was associ- ated with a significant extension of the barrow. This grave featured a monumental con- struction, including a wooden chamber casing and stone slabs, probably reused stelae, as well as mats and wooden poles forming a roof (Fig. 19). The chamber had a rectangular shape with vertical, straight, and carefully finished walls. The deceased, a man aged 35-50 and approximately 187 cm tall, was placed on a mat in a supine position with the lower limbs curled up and the upper limbs stretched along the body. The body was oriented Yampil barrows from the fourth and third millennia BC in the light of Polish-Ukrainian... 269 Yampil barrows from the fourth and third millennia BC in the light of Polish-Ukrainian... Fig. 19. Reconstruction of Grave IV/4 from Prydnistryanske, Yampil rayon. Illustrated by M. Podsiadło Fig. 19. Reconstruction of Grave IV/4 from Prydnistryanske, Yampil rayon. Illustrated by M. Podsiadło Fig. 19. Reconstruction of Grave IV/4 from Prydnistryanske, Yampil rayon. Illustrated by M. Podsiadło along the W-E axis, with the head facing W. The bones were adorned with ochre in several places, and the upper part of the skull exhibited an intense coloration. A separate lump of bright red ochre was found near the head. This ritual is characteristic only of male burials from the older phase of YC and has strong parallels in graves from the burial mounds of this culture in the steppe zone, especially along the lower Dnister, Inhul, Inhulets, and the along the W-E axis, with the head facing W. The bones were adorned with ochre in several places, and the upper part of the skull exhibited an intense coloration. A separate lump of bright red ochre was found near the head. This ritual is characteristic only of male burials from the older phase of YC and has strong parallels in graves from the burial mounds of this culture in the steppe zone, especially along the lower Dnister, Inhul, Inhulets, and the A. Kośko, V. Klochko, M. Potupchyk, P. Włodarczak, D. Żurkiewicz 270 Table 2. Radiocarbon dates of graves from the early-Yamna period Barrow Grave no. Stage II – Early Yamna Dated material Lab code Age BP Calendar age BC (68.2%)* Klembivka 1 15 Human bone Poz-77470 4290 ± 35 2920-2884 Klembivka 1 14 Human bone Poz-52422 4260 ± 40 2916-2782 Klembivka 1 14 Wood Poz-52605 4135 ± 35 2862-2630 Klembivka 1 5 Human bone Poz-70670 4225 ± 35 2898-2707 Pidlisivka 1 1Aa Wood Poz-52423 4190 ± 35 2883-2700 Pidlisivka 1 1A Human bone Poz-38529 4195 ± 35 2885-2701 Pidlisivka 1 1A Human bone Poz-39214 4080 ± 40 2843-2500 Pidlisivka 1 1A Wood Poz-52424 4085 ± 35 2842-2504 Pidlisivka 1 1B Human bone Poz-102979 4100 ±35 2846-2577 Prydnistryanske IV 4 Human bone Poz- 66230 4455 ± 40 3327-3026 Prydnistryanske IV 4 Wood Poz-66229 4380 ± 35 3024-2921 Prydnistryanske IV 6 Human bone Poz-70673 4090 ± 40 2846-2505 Prydnistryanske IV 6 Wood Poz-66231 4185 ± 35 2882-2698 * OxCal calibration program v4.4.4 (Bronk Ramsey 2021) Table 2. Radiocarbon dates of graves from the early-Yamna period * OxCal calibration program v4.4.4 (Bronk Ramsey 2021) lower Southern Buh (see, e.g., Shaposhnikova et al. 1986, 15; Rychkov 2001, 45). The two radiocarbon dates for this grave differ significantly (Table 2; Fig. 20). A typical result for early YC was obtained for a sample of wood (ash) from the grave construction, indicating the transition from the 4th to the 3rd millennium BC, similar to most western YC results (e.g., Preda-Bălănică et al. 2020, 87; Alexandrov and Włodarczak 2022). The date ob- tained for the bone sample from Grave IV/4 is earlier (3327-3026 cal BC), although the younger part of the range allows reconciliation, suggesting an age around the last century of the 4th millennium BC. This is currently the only such early dating for a YC grave in the Yampil region. Another early YC burial from Prydnistryanske, Grave IV/6, shares simi- larities with Grave IV/4 in terms of grave construction and body arrangement. However, its 14C dates indicate a later period, approximately 2850-2600 cal BC. The early YC rite also extends to Grave 1A from Barrow 1 in Pidlisivka, which was con- sidered primary for the second phase of the barrow’s construction. In terms of arrange- ment, orientation, and the use of ochre, this burial exhibits characteristics typical of YC. Grave 1B, probably older than 1A, differs in its orientation, with the burial at a slightly lower level. Stage II – Early Yamna Despite the suggested chronology, the arrangement of the deceased aligns gen- erally with the tradition of the early YC. The skull also shows signs of ochre staining. Based on these features and the radiocarbon dating results, it is necessary to revise previous proposals and classify Graves 1A and 1B within the early YC period. Another early YC burial, Grave 15 in Barrow 1 in Klembivka, presents an exception in terms of the irregular shape of the grave pit. This may partly result from the poor preservation condition. Nevertheless, 2 Yampil barrows from the fourth and third millennia BC in the light of Polish-Ukrainian... 2 Yampil barrows from the fourth and third millennia BC in the light of Polish-Ukrainian... 271 p g Fig. 20. Calibration results of 14C ages of samples from the early Yamna graves. For the calibration, the Oxcal v4.4.4 software (Bronk Ramsey 2021) was used Fig. 20. Calibration results of 14C ages of samples from the early Yamna graves. For the calibration, the Oxcal v4.4.4 software (Bronk Ramsey 2021) was used the position, orientation of the deceased, and the presence of an ochre lump near the head indicate a typical early YC burial. This assessment is confirmed by its absolute dating, placing it in the period of 2920-2884 cal BC. However, two subsequent burials from this mound, Nos 5 and 14, were previously considered Eneolithic. Radiocarbon dating, however, suggests an early YC period. The position of the deceased in Grave 14 is particularly characteristic of the Zhivotilovka-Volchansk group and the rite in phase C/II of the Trypillia culture. the position, orientation of the deceased, and the presence of an ochre lump near the head indicate a typical early YC burial. This assessment is confirmed by its absolute dating, placing it in the period of 2920-2884 cal BC. However, two subsequent burials from this mound, Nos 5 and 14, were previously considered Eneolithic. Radiocarbon dating, however, suggests an early YC period. The position of the deceased in Grave 14 is particularly characteristic of the Zhivotilovka-Volchansk group and the rite in phase C/II of the Trypillia culture. A. Kośko, V. Klochko, M. Potupchyk, P. Włodarczak, D. Żurkiewicz 272 Accepting the obtained absolute age, the burials from Klembivka should be considered an example of the preservation of older traditions (Eneolithic funeral rites) into the 3rd mil- lennium BC. Stage II – Early Yamna Similar exceptions to the dominant YC ritual are known from other Podillia barrows, such as Grave 7 in Barrow 3A in Porohy (Klochko et al. 2015b, 99, fig. 18), or Grave 2 from Barrow 2 in Kuzmin (Bubulich and Khakheu 2002, 130, fig. 10: A). * OxCal calibration program v4.4.4 (Bronk Ramsey 2021) Stage III – Late Yamna Burials from the later YC phase were discovered in Pidlisivka, Porohy, and Prydnistry- anske (Barrow IV). In Prydnistryanske, these features (Nos 3, 8, and 9) were stratigraphi- cally younger than graves from an earlier phase (Nos 4 and 6). Stratigraphic observations confirm the characteristics of the grave construction and burial arrangements. It is possi- ble that Grave 3, located near the top of the barrow, represents a slightly older chrono- logical position, with the same body arrangement and W-E orientation characteristic of the older phase. Features 8 and 9, on the other hand, were dug into the outskirts of the Table 3. Radiocarbon dates obtained for graves dated to late YC period Barrow Grave no. Dated material Lab code Age BP Calendar age BC (68.2%)* Pidlisivka 1 11 Human bone Poz-81793 4085 ± 30 2839-2573 Porohy 3A 1 Human bone Poz-70668 3760 ± 35 2276-2065 Porohy 3A 1 Wood Poz-103120 4095 ± 35 2846-2576 Porohy 3A 2 Human bone Poz-74392 4140 ± 35 2866-2632 Porohy 3A 3 Wood Poz-103052 4105 ± 30 2846-2580 Porohy 3A 7 Human bone Poz-70667 4115 ± 35 2852-2584 Porohy 3A 7 Human bone Poz-102978 3925 ± 35 2470-2346 Porohy 3A 10 Human bone Poz-74393 4105 ± 35 2847-2580 Porohy 3A 10 Human bone Poz-81824 4040 ± 35 2622-2488 Porohy 3A 11 Human bone Poz-47741 4075 ± 35 2837-2500 Porohy 3A 11 Bone of goat Poz-101625 4070 ± 35 2835-2497 Porohy 3A 12 Human bone Poz-47742 3985 ± 35 2568--2467 Porohy 3A 17 Human bone Poz-47743 4050 ± 35 2627-2492 Porohy 3A 17 Human bone Poz-74394 3930 ± 35 2472-2347 Porohy 3A 18 Bone of goat Poz-101626 4115 ± 35 2852-2584 Porohy 3A 19 Human bone Poz-70665 4185 ± 35 2882-2698 Porohy 3A 20 Human bone Poz-47744 4190 ± 35 2883-2700 Porohy 3A 20 Human bone Poz-74397 4175 ± 35 2879-2680 Prydnistryanske IV 3 Human bone Poz-66228 4090 ± 35 2845-2573 Prydnistryanske IV 8 Human bone Poz-66232 4090 ± 35 2845-2573 Prydnistryanske IV 9 Human bone Poz-66233 4120 ± 35 2856-2585 Table 3. Radiocarbon dates obtained for graves dated to late YC period 2 Yampil barrows from the fourth and third millennia BC in the light of Polish-Ukrainian... Yampil barrows from the fourth and third millennia BC in the light of Polish-Ukrainian... 273 Fig. 21. Calibration results of 14C ages of samples from the late Yamna graves. Stage III – Late Yamna For the calibration, the Oxcal v4.4.4 software (Bronk Ramsey 2021) was used Fig. 21. Calibration results of 14C ages of samples from the late Yamna graves. For the calibration, the Oxcal v4.4.4 software (Bronk Ramsey 2021) was used A. Kośko, V. Klochko, M. Potupchyk, P. Włodarczak, D. Żurkiewicz 274 mound, marking a younger phase. Barrow IV in Prydnistryanske provides an example of clear stratigraphy, demonstrating patterns in the northwestern Pontic region. In Barrow 3A in Porohy, a cemetery from the late YC phase was discovered, consisting of 11 graves. Such a significant accumulation of burials is rarely found in the forest-steppe zone of the Podillia Upland and was not observed in the other four burial mounds examined in Porohy. In the Yampil cluster, only Barrow 2 in Severynivka exhibits a similar arrangement of eleven YC graves to that in Porohy 3A (Harat et al. 2014, 173, fig. 2). In the neighbouring Kamenka cluster, numerous late YC graves were discovered in Barrows 3 and 6 in Ocniţa, situated around the centre of the mound (Manzura et al. 1992, 109, 121, figs 7 and 19). This method of organizing the cemetery reflects a model from the Budzhak steppe (see Yarovoy 1985, 52-61; Dergachev 1986, 36-42). The Podillian sites lack pottery equipment typical of the latest (= Budzhak) YC period. An interesting and challenging issue in this context is the chronological relationship between the dated Yampil graves and features with Budzhak- style pottery. The latter may belong to a later period than the cemeteries of the Porohy 3A type (cf. Dergachev 1986, 82-87). The primary grave in Mound 3A in Porohy (No. 14) was unfortunately destroyed by modern treasure hunters. Due to the barrow’s construction, the age of its older phase was determined to be Late Eneolithic (Klochko et al. 2015b). The modern intrusion also destroyed a significant part of Feature 2 – the YC burial, probably Fig. 22. Porohy, Yampil rayon, Barrow 3A. Grave 7 – late Yamna stage. Photo by P. Włodarczak Fig. 22. Porohy, Yampil rayon, Barrow 3A. Grave 7 – late Yamna stage. Photo by P. Włodarczak 275 Yampil barrows from the fourth and third millennia BC in the light of Polish-Ukrainian... associated with the barrow’s superstructure phase. It is unknown whether this burial was linked to the early or late YC. Stage III – Late Yamna The radiocarbon date obtained for it is similar to the results from other YC graves (Table 3; Fig. 21). It can only be assumed that Grave 2 is slightly older than the surrounding features. Radiocarbon dating suggests that the absolute age of the Porohy graves falls within the range of approximately 2800-2600 cal BC. Graves 7 and 12 probably have a slightly younger age (around 2600-2400 cal BC), with two distinct 14C results for Feature 7. The correctness of the younger one is supported by the specific burial arrangement and the presence of a fragment of a vessel, secondarily used as a “zharovnia” (a stand for storing embers), an element characteristic of the Catacombna culture stage (Fig. 22). Three late YC graves (Nos 4, 9, and 11) were discovered in the Pidlisivka barrow. Radiocarbon dating was conducted only for Feature 11, yielding a date of 2839-2573 cal BC (Table 3). This date aligns with the analyses performed for the cemetery in Porohy 3A. * OxCal calibration program v4.4.4 (Bronk Ramsey 2021) Stage IV – Catacombna culture The documentation of the Catacombna culture phase in the Yampil region is limited. A grave unequivocally connected with the Inhul Catacombna group (Feature I/4) was found only in Barrow I in Prydnistryanske. The absolute age for this feature (Table 4) was estimated to be around 2550-2450 cal BC (Fig. 23). Although this date is surprisingly early, considering the typological features of the burial, it is slightly younger than the estimated age for the late YC in the Yampil region. Burials displaying features of the Catacombna culture were also uncovered in the barrows in Pidlisivka and Porohy. In the former, the poorly preserved Grave 7 probably had a semi-catacomb construction (Klochko et al. 2015a, 57, fig. 13). In horizontal projection, its chamber had an oval shape, unlike the rectangular structures of the YC. However, a relatively early 14C date was obtained for this grave, cor- responding to the late YC results (Table 4). Attempting to reconcile the typological and absolute dating, one would need to assume the presence of early Catacombna culture fea- tures in the Yampil region, dating to around 2850-2600 cal BC. The coexistence of early Catacombna culture and the YC in the Podillia Upland is supported by the discovery of a typical, early-dated grave with a catacomb construction in Barrow 3 in Ocniţa (Klochko Table 4. Radiocarbon dates for graves of the Catacombna type from Yampil region Barrow Grave no. Dated material Lab code Age BP Calendar age BC (68.2%)* Pidlisivka 1 7 Human bone Poz-38531 4120 ± 35 2856-2585 Prynistryanske I 4 Wood Poz-66218 4105 ± 40 2848-2579 Prynistryanske I 4 Human bone Poz-66219 4070 ± 35 2835-2497 Prynistryanske I 4 Human bone Poz-66220 3940 ± 40 2556-2347 Prynistryanske I 4 Human bone Poz-66732 3940 ± 35 2553-2347 * OxCal calibration program v4.4.4 (Bronk Ramsey 2021) Table 4. Radiocarbon dates for graves of the Catacombna type from Yampil region A. Kośko, V. Klochko, M. Potupchyk, P. Włodarczak, D. Żurkiewicz 276 Fig. 23. Calibration results of 14C ages of samples from the Catacombna graves. For the calibration, the Oxcal v4.4.4 software (Bronk Ramsey 2021) was used Fig. 23. Calibration results of 14C ages of samples from the Catacombna graves. For the calibration, the Oxcal v4.4.4 software (Bronk Ramsey 2021) was used 1990; Manzura et al. 1992, 20, 21). Stage IV – Catacombna culture The burial from Grave 7 in Porohy may be an example of a late-dated YC feature with elements of the Catacombna culture’s funeral rite. Two distinct 14C dates were obtained for this grave. The younger date – 2470-2346 cal BC (68.3% probability) – appears to align better with the typological assessment of this burial. 5. Conclusions During the field research conducted between 2010 and 2014, intriguing sequences of graves were unearthed in four barrow cemeteries located in the Yampil region. The list of 48 radiocarbon dates presented above serves as a crucial reference point for the entire eastern Podillia area, offering valuable insights into the historical context of this region. It is essential to note that the materials collected by the Polish-Ukrainian expedition do not comprehensively represent the taxonomic and chronological diversity of the Yampil bar- rows, as evident from comparisons with earlier studies in the region, including those in the neighbouring Kamenka region. Chronometric investigations have revealed that the majority of Yampil barrows were constructed during two distinct periods: the late Eneolithic and the early YC. Notably, research on barrows I-IV in Prydnistryanske led to the discovery of the oldest barrow graves in Podillia, marked by pottery from the late Trypillia Gordineşti group and dating back to 3350-3150 cal BC. A slightly different Eneolithic tradition may have been represented by Barrow 3A in Porohy, though its precise age remains undeter- mined. The area also features extended inhumations, including the Post-Mariupol/Kvityan- 277 Yampil barrows from the fourth and third millennia BC in the light of Polish-Ukrainian.. natype, and burials reflecting the Cernavodă/Nizhna Mikhailivka tradition, though their absolute chronology is presently uncertain. A second phase of barrow construction, associated with the early YC, is documented in Klembivka and Pidlisivka. This period also witnessed the utilization of Eneolithic mounds, such as Prydnistryanske IV and probably Porohy 3A. The construction activities com- menced around 3100-3000 cal BC (Prydnistryanske IV, Grave 4) and persisted until around 2800 cal BC (Klembivka, Grave 15; Pidlisivka, Graves 1A and 1B). It is probably during this time that distinctive chains of barrows were established, similar to those in the well-identified Pysarivka cemetery from previous research. Between 2800-2600 cal BC, YC graves were predominantly dug into pre-existing mounds, often featuring superstruc- tures associated with early YC burials. Late YC cemeteries typically contain multiple graves, although, in some cases, like Porohy 3A, they exhibit characteristic curved zones with a higher number of graves, akin to the barrows found in the steppe zone. Singular graves confirm the continued use of the Yampil barrow cemeteries approximately between 2600- 2400 cal BC. Only one grave from this period exhibits classic characteristics of the Cata- comb culture (Prydnistryanske, Grave I/4). 5. Conclusions Subsequently, there was a hiatus lasting sev- eral hundred years until the early 2nd millennium BC when graves of the Babyno culture were incorporated into the mounds. Based on typological analysis, Yampil graves featuring pottery with Central European characteristics are linked to the late YC stage (approximately 2800-2600 cal BC). These burials were interred within barrows that date back to the Eneolithic and early YC periods. The chronometric findings presented here provide a solid foundation for further research aimed at organizing barrow sequences in the Podillia zone and studying the cultural trans- formations between the North Pontic steppe and Central Europe. References Alexandrov S. and Włodarczak P. 2021. Chronological sequence of the Early Bronze Age graves in the Pamukli Bair Barrow at Malomirovo and the Pit-Grave Culture expansion in the Middle Tundzha Valley. Studia Praehistorica 16, 207-240. Allentoft M. E., Sikora M., Sjögren K.-G., Rasmussen S., Rasmussen M. , Stenderup J., Damgaard P. B., Schroeder H., Ahlström T., Vinner L., Malaspinas A.-S., Margaryan A., Higham T., Chivall D., Lynnerup N., Harvig L., Baron J., Della Casa P., Dąbrowski P., Duffy P. R., Ebel A. V., Epimakhov A., Frei K., Furmanek M., Gralak T.,Gromov A., Gronkiewicz S., Grupe G., Hajdu T., Jarysz R., Khar- tanovich V., Khokhlov A., Kiss V., Kolář J., Kriiska A.,Lasak I., Longhi C., McGlynn G., Merkevi- cius A., MerkyteI., Metspalu M., Mkrtchyan R., Moiseyev V., Paja L., PálfiG., Pokutta D., Pos- pieszny Ł., Price T. D., Saag L., Sablin M., Shishlina N., Smrčka V., Soenov V. I., Szeverényi V., Tóth G., Trifanova S. V., Varul L., Vicze M., Yepiskoposyan L., Zhitenev V., Orlando L., Sicheritz- Pontén T., Brunak S., Nielsen R., Kristiansen K. and Willerslev E. 2015. Population genomics of Bronze Age Eurasia. Nature 522, 167-173. Allentoft M. E., Sikora M., Sjögren K.-G., Rasmussen S., Rasmussen M. , Stenderup J., Damgaard P. B., Schroeder H., Ahlström T., Vinner L., Malaspinas A.-S., Margaryan A., Higham T., Chivall D., Lynnerup N., Harvig L., Baron J., Della Casa P., Dąbrowski P., Duffy P. R., Ebel A. V., Epimakhov A., Frei K., Furmanek M., Gralak T.,Gromov A., Gronkiewicz S., Grupe G., Hajdu T., Jarysz R., Khar- tanovich V., Khokhlov A., Kiss V., Kolář J., Kriiska A.,Lasak I., Longhi C., McGlynn G., Merkevi- cius A., MerkyteI., Metspalu M., Mkrtchyan R., Moiseyev V., Paja L., PálfiG., Pokutta D., Pos- pieszny Ł., Price T. D., Saag L., Sablin M., Shishlina N., Smrčka V., Soenov V. I., Szeverényi V., Tóth G., Trifanova S. V., Varul L., Vicze M., Yepiskoposyan L., Zhitenev V., Orlando L., Sicheritz- Pontén T., Brunak S., Nielsen R., Kristiansen K. and Willerslev E. 2015. Population genomics of Bronze Age Eurasia. Nature 522, 167-173. Bronze Age Eurasia. Nature 522, 167-173. 278 A. Kośko, V. Klochko, M. Potupchyk, P. Włodarczak, D. Żurkiewicz Bronk Ramsey C. 2021. OxCal 4.4 Manual. https://c14.arch.ox.ac.uk/oxcal/OxCal.html, accessed September 2023. Bronk Ramsey C. 2021. OxCal 4.4 Manual. https://c14.arch.ox.ac.uk/oxcal/OxCal.html, accessed September 2023. Bubulich V. and Khakheu V. 2002. Issledovaniya kurganov v Kamenskom rayone na levoberezhe Srednego Dnestra. In N. A. References Ketraru (ed.), Severnoe Prichernomore ot eneolita k antichnosti, Tiraspol: Pridnestrovskiy Gosudarstvennyi Universitet im. T. G. Shevchenko, 112-148. Dergachev V. A. 1986. Moldaviya i sosednie territorii v epokhu bronzy. Kishinev: Shtiintsa. Dergachev V. A. 2022. Pozdnee Tripolye – Maykop (s katalogom pogrebalnykh kompleksov pozdnego Tripolya). Kishineu: CEP USM. Goslar T., Klochko V. I., Kośko A., Włodarczak P. and Żurkiewicz D. 2015. Chronometry of late Eneo- lithic and ‘Early Bronze’ cultures in the middle Dniester area: investigations of the Yampil barrow complex. Baltic-Pontic Studies 20, 256-291. Goslar T., Kośko A. and Razumov S. 2014. Identyfikacja kulturowo-chronologiczna cmentarzysk „wczesnobrązowych” na stanowisku Pidlisivka 1. In A. Kośko, M. Potupczyk and S. Razumow (eds), Naddniestrzańskie kompleksy cmentarzysk kurhanowych z III tysiąclecia przed Chr. w okoli- cach Jampola, obwód winnicki (= Archaeologia Bimaris. Monografie 6). Poznań: Wydawnictwo Nauka i Innowacje, 305-311. Haak W., Lazaridis I., Patterson N., Rohland N., Mallick S., Llamas B., Brandt G., Nordenfelt S., Har- ney E., Stewardson K., Fu Q., Mittnik A., Bánffy E., Economou C., Francken M., Friederich S., Garrido Pena R., Hallgren F., Khartanovich V., Khokhlov A., Kunst M., Kuznetsov P., Meller H., Mochalov O., Moiseyev V., Nicklisch N., Pichler S. L., Risch R., Rojo Guerra M. A., Roth C., Széc- sényi-Nagy A., Wahl J., Meyer M., Krause J., Brown D., Anthony D., Cooper A., Alt K. W. and Reich D. 2015. Massive migration from the steppe was a source for Indo-European languages in Europe. Nature 522, 207-211. Harat K., Potupczyk M. and Razumow S. 2014. Charakterystyka źródeł archeologicznych pozyskanych w latach 1984-1993. In A. Kośko, M. Potupczyk and S. Razumow (eds), Naddniestrzańskie kompleksy cmentarzysk kurhanowych z III tysiąclecia przed Chr. w okolicach Jampola, ob- wód winnicki (= Archaeologia Bimaris. Monografie 6). Poznań: Wydawnictwo Nauka i Inno- wacje, 47-204. Ivanova S., Kośko A. and Włodarczak P. 2014. Komponent tradycji kultur ceramiki sznurowej. Amfory w północno-zachodnio-nadczarnomorskich grobach kultury jamowej. In In A. Kośko, M. Potup- czyk and S. Razumow (eds), Naddniestrzańskie kompleksy cmentarzysk kurhanowych z III tysiąc- lecia przed Chr. w okolicach Jampola, obwód winnicki (=Archaeologia Bimaris. Monografie 6). Poznań: Wydawnictwo Nauka i Innowacje, 351-386. Jachimowicz K. and Żurkiewicz D. 2017. Spatial analysis of Yampil barrow complex. Baltic-Pontic Studies 22, 7-29. Jachimowicz K. and Żurkiewicz D. 2017. Spatial analysis of Yampil barrow complex. Baltic-Pontic Studies 22, 7-29. References Juras A., Chyleński M., Ehler E., Malmström H., Żurkiewicz D., Włodarczak P., Wilk S., Peška J., Fojtík P., Králík M., Libera J., Bagińska J., Tunia K., Klochko V. I., Dabert M., Jakobsson M. and Kośko A. 2018. Mitochondrial genomes reveal an east to west cline of steppe ancestry in Corded Ware populations. Scientific Reports 8, 11603. https://doi.org/10.1038/s41598-018-29914-5. Yampil barrows from the fourth and third millennia BC in the light of Polish-Ukrainian... Yampil barrows from the fourth and third millennia BC in the light of Polish-Ukrainian.. 279 Klochko E. O. 1990. Redkiy pogrebalnyi kompleks katakombnoy kultury na Srednom Dnestre. In O. G. Shaposhnikova, V. V. Otroschenko, S. Z. Pustovalov, G. N. Toschev and G.I. Shapovalov (eds), Problemy izucheniya katakombnoy kulturno-istoricheskoy obschnosti. Zaporozhe: Zaporoz hskiy gosudarstvennyi universitet, 28-30. Klochko V. I., Kośko A., Razumov S. M., Włodarczak P., Żurkiewicz D. 2015a. Eneolithic, Yamnaya, Catacomb and Babyno culture cemeteries, Pidlisivka, barrow 1, Yampil region, Vinnitsa oblast: archaeometry, chronometry and taxonomy. Baltic-Pontic Studies 20, 40-77. Klochko V. I., Kośko A., Razumov S. M., Włodarczak P. and Żurkiewicz D. 2015b. Eneolithic, Yamnaya and Noua culture cemeteries from the first half of the 3rd and the middle of the 2nd millenium BC, Porohy, site 3A, Yampil region, Vinnitsa oblast: archaeometric and chronometric description, ritual and taxonomic-topogenetic identification. Baltic-Pontic Studies 20, 78-141. Klochko V. I., Kośko A., Razumov S. M., Włodarczak P. and Żurkiewicz D. 2015c. Eneolithic, Babyno and Noua culture cemeteries, Klembivka, site 1, Yampil region, Vinnitsa oblast: archaeometry, taxonomy and topogenetics. Baltic-Pontic Studies 20, 142-182. Klochko V. I., Kośko A., Potupchyk M. V., Włodarczak P., Żurkiewicz D. and Ivanova S.V. 2015d. Tripolye (Gordineşti group), Yamnaya and Catacomb culture cemeteries, Prydnistryanske, site 1, Yampil region, Vinnitsa oblast: an archaeometric and chronometric description and a taxonomic and topogenetic discussion. Baltic-Pontic Studies 20, 183-255. Kośko A. 1981. Udział południowo-wschodnioeuropejskich wzorców kulturowych w rozwoju niżo- wych społeczeństw kultury pucharów lejkowatych. Grupa mątewska (= Uniwersytet im. Ada- ma Mickiewicza w Poznaniu. Seria Archeologia 19). Inowrocław: Wydawnictwo Naukowe Uni- wersytetu im. Adama Mickiewicza w Poznaniu. Kośko A. 2011. Z badań nad kontekstem kultur wczesnobrązowych strefy pontyjskiej w rozwoju spo- łeczności środkowoeuropejskich obszaru dorzecza Wisły w III tys. BC. In U. Stankiewicz and A. Wawrusiewicz (eds), Na rubieży kultur. Badania nad okresem neolitu i wczesną epoką brązu. Białystok: Muzeum Podlaskie w Białymstoku, 183-193. Kośko A. ed. 2015. Podolia as a cultural contact area in the 4th/3rd-2nd millennium BC. Baltic-Pontic Studies 20. Kośko A. ed. 2017. Podolia “Barrow culture” communities: 4th/3rd-2nd mill. BC. The Yampil barrow complex: interdisciplinary studies. Baltic-Pontic Studies 22. Kośko A. ed. 2017. Podolia “Barrow culture” communities: 4th/3rd-2nd mill. BC. The Yampil barrow complex: interdisciplinary studies. Baltic-Pontic Studies 22. Kośko A. and Klochko V. I. 2009. Yampil barrows from the fourth and third millennia BC in the light of Polish-Ukrainian... Transit routes between the Baltic and Black seas: early development stages – from the 3rd to the middle of the 1st millennium BC an outline of research project. Baltic- Pontic Studies 14, 9-18. Kośko A. and Klochko V. I. 2009. Transit routes between the Baltic and Black seas: early development stages – from the 3rd to the middle of the 1st millennium BC an outline of research project. Baltic- Pontic Studies 14, 9-18. Kośko A., Potupczyk M. and Razumow S. eds 2014. Naddniestrzańskie kompleksy cmentarzysk kur- hanowych z III tysiąclecia przed Chr. w okolicach Jampola, obwód winnicki (= Archaeologia Bimaris. Monografie 6). Poznań: Wydawnictwo Nauka i Innowacje. Kośko A. and Razumow S. 2014. Cele i metody badań terenowych. In A. Kośko, M. Potupczyk and S. Razumow (eds), Naddniestrzańskie kompleksy cmentarzysk kurhanowych z III tysiąclecia przed Chr. w okolicach Jampola, obwód winnicki (= Archaeologia Bimaris. Monografie 6). Po- znań: Wydawnictwo Nauka i Innowacje, 41-44. 280 A. Kośko, V. Klochko, M. Potupchyk, P. Włodarczak, D. Żurkiewicz Kośko A., Razumow S. and Żurkiewicz D. 2014. Cechy formalno-typologiczne obiektów oraz ich wy- posażenia. In A. Kośko, M. Potupczyk and S. Razumow (eds), Naddniestrzańskie kompleksy cmentarzysk kurhanowych społeczności z III i z pierwszej polowy II tysiąclecia BC w okolicach Jampola, obwód winnicki. Z badań nad północno–zachodnią rubieżą osadnictwa społeczności kręgu kultur „wczesnobrązowych” strefy pontyjskiej. Badania z lat 1984-2014 (= Archaeologia Bimaris – Monografie 6). Poznań: Wydawnictwo Nauka i Innowacje, 213-235. Kośko A. and Szmyt M. 2009. Udział społeczności Niżu Środkowoeuropejskiego w poznawaniu środo- wisk biokulturowych Płyty Nadczarnomorskiej: IV-IV/III tys. BC. In M. Ignaczak, A. Kośko and M. Szmyt (eds), Między Bałtykiem a Morzem Czarnym. Szlaki międzymorza IV-I tys. przed Chr. (= Archaeologia Bimaris. Dyskusje 4). Poznań: Wydawnictwo Poznańskie, 205-221. Larina O. 2003. Pozdnetripolskoe pogrebenie gordineshtskogo tipa na Dnestre. In E. Sava (ed.), In- terferenţe cultural-chronologice in spaţiul nord-pontic, Chişinău: Akademiya Nauk Respubliki Moldova, Institut Arkheologii i Etnografii, 57-80. Makohonienko M. 2009. Natural scientific aspects of prehistoric and early historic transit routes in the Baltic-Pontic cultural area. Baltic-Pontic Studies 14, 19-71. Makohonienko M. 2009. Natural scientific aspects of prehistoric and early historic transit routes in the Baltic-Pontic cultural area. Baltic-Pontic Studies 14, 19-71. Makohonienko M. and Hildebrandt-Radke I. 2014. Kontekst przyrodniczy i perspektywy studiów pa- leośrodowiskowych stanowiska kultury jamowej w Pidlisiwce nad Dniestrem, zachodnia Ukraina. In A. Kośko, M. Potupczyk and S. Razumow (eds), Naddniestrzańskie kompleksy cmentarzysk kurhanowych z III tysiąclecia przed Chr. Yampil barrows from the fourth and third millennia BC in the light of Polish-Ukrainian... w okolicach Jampola, obwód winnicki (= Archaeolo- gia Bimaris. Monografie 6). Poznań: Wydawnictwo Nauka i Innowacje, 247-274. Popovici S. and Ciobanu I. 2021. Cercetări arheologice la Cimişlia. Chişinău: Agenţia Naţională Ar- heologică, Rezervaţia Cultural-Naturală „Orheiul Vechi”. Popovici S. and Ciobanu I. 2021. Cercetări arheologice la Cimişlia. Chişinău: Agenţia Naţională Ar- heologică, Rezervaţia Cultural-Naturală „Orheiul Vechi”. Potupczyk M. and Razumow S. 2014. Zgrupowanie cmentarzysk kurhanowych w rejonie jampolskim. In A. Kośko, M. Potupczyk and S. Razumow (eds), Naddniestrzańskie kompleksy cmentarzysk kurhanowych z III tysiąclecia przed Chr. w okolicach Jampola, obwód winnicki (= Archaeolo- gia Bimaris. Monografie 6). Poznań: Wydawnictwo Nauka i Innowacje, 35-39. Potupczyk M. and Razumow S. 2014. Zgrupowanie cmentarzysk kurhanowych w rejonie jampolskim. In A. Kośko, M. Potupczyk and S. Razumow (eds), Naddniestrzańskie kompleksy cmentarzysk kurhanowych z III tysiąclecia przed Chr. w okolicach Jampola, obwód winnicki (= Archaeolo- gia Bimaris. Monografie 6). Poznań: Wydawnictwo Nauka i Innowacje, 35-39. Preda-Bălănică B., Frînculeasa A. and Heyd V. 2020. The Yamnaya Impact North of the Lower Danube: A Tale of Newcomers and Locals. Bulletin de la Société préhistorique française, 117/1, 85-101. Rassamakin Yu. Y. 2004. Die nordpontische Steppe in der Kupferzeit Gräber aus der Mitte des 5. Jts. bis Ende des 4. Jts. v. Chr. Teil I – Text (= Archäologie in Eurasien 17), Mainz: Philipp von Zabern. 004 e o dpo sc e S eppe de pfe e G äbe a s de e des 5 J s bis Ende des 4. Jts. v. Chr. Teil I – Text (= Archäologie in Eurasien 17), Mainz: Philipp von Zabern. Rybicka M. 2017. Kultura trypolska – kultura pucharów lejkowatych. Natężenie kontaktów i ich chronologia (= Collectio Archaeologica Ressoviensis 37). Rzeszów: Wydawnictwo Uniwersytetu Rzeszowskiego i Wydawnictwo Zimowit. Rybicka M., Sîrbu G., Król D. and Bicbaev V. 2020. New radiocarbon dates for stage CII Tripolye culture, northern Moldova. Baltic-Pontic Studies 24, 87-103. Rychkov N.A. 2001. Ob uslovnykh tsentrakh nositeley Yamnoy kultury. In N. A. Ketraru (ed.), Sever- noe Prichernomore ot eneolita k antichnosti. Tiraspol: Pridnestrovskiy Gosudarstvennyi Univer- sitet im. T. G. Shevchenko, 42-66. Shaposhnikova O. G., Fomenko V. N. and Dovzhenko N. D. 1986. Yamnaya kulturno-istoricheskaya Oblast (yuzhnobugskiy variant). Kiev: Naukova Dumka. Yampil barrows from the fourth and third millennia BC in the light of Polish-Ukrainian... Yampil barrows from the fourth and third millennia BC in the light of Polish-Ukrainian... 281 Simonenko A. V. and Lobay B. I. 1991. Sarmaty Severo-Zapadnogo Prichernomoria v I v. n. e. Szmyt M. 1999. Between West and East. People of the Globular Amphora Culture in Eastern Europe. Baltic-Pontic Studies 8. Yarovoy E. V. 1985. Drevneyshie skotovodcheskie plemena yugo-zapada SSSR (klassifikatsiya pogrebalnogo obriada). Kishinev: Shtiintsa. Yarovoy E. V. 1985. Drevneyshie skotovodcheskie plemena yugo-zapada SSSR (klassifikatsiya pogrebalnogo obriada). Kishinev: Shtiintsa. Włodarczak P. 2017. Kurgan rites in the Eneolithic and Early Bronze Age Podolia in light of materials from the funerary-ceremonial centre at Yampil. Baltic-Pontic Studies 22, 246-283. Włodarczak P. 2017. Kurgan rites in the Eneolithic and Early Bronze Age Podolia in light of materials from the funerary-ceremonial centre at Yampil. Baltic-Pontic Studies 22, 246-283. Włodarczak P. 2021. Estetyczny wymiar pochówku u społeczności kurhanowych w III tysiącleciu BC. In O. Diachenko, T. K. Harper, Yu. Rassamakin and I. Sobkowiak-Tabaka (eds), Data systemati- zation in the Neo-Eneolithic of Southeastern and Central Europe. Essays in honor of Sergej Ryzhov. Kyiv: National Academy of Sciences of Ukraine, Institute of Archaeology, 182-202. Włodarczak P. 2021. Estetyczny wymiar pochówku u społeczności kurhanowych w III tysiącleciu BC. In O. Diachenko, T. K. Harper, Yu. Rassamakin and I. Sobkowiak-Tabaka (eds), Data systemati- zation in the Neo-Eneolithic of Southeastern and Central Europe. Essays in honor of Sergej Ryzhov. Kyiv: National Academy of Sciences of Ukraine, Institute of Archaeology, 182-202.
https://openalex.org/W2952993480	https://research.utwente.nl/files/120240998/Ihwagi_et_al_2019_The_Journal_of_Wildlife_Management.pdf	English	null	Poaching lowers elephant path tortuosity: implications for conservation	The Journal of wildlife management/The journal of wildlife management	2,019	cc-by	9,734	Poaching Lowers Elephant Path Tortuosity: Implications for Conservation FESTUS W. IHWAGI,1,2 Department of Natural Resources, Faculty of Geo‐Information Science and Earth Observation (ITC), University of Twente, P.O. Box 217, 7500 AE, Enschede, The Netherlands FESTUS W. IHWAGI,1,2 Department of Natural Resources, Faculty of Geo‐Information Science and Earth Observation (ITC), University of Twente, P.O. Box 217, 7500 AE, Enschede, The Netherlands ANDREW K. SKIDMORE,3 Department of Natural Resources, Faculty of Geo‐Information Science and Earth Observati University of Twente, P.O. Box 217, 7500 AE, Enschede, The Netherlands TIEJUN WANG, Department of Natural Resources, Faculty of Geo‐Information Science and Earth Observation (ITC), University of Twente, P.O. Box 217, 7500 AE, Enschede, The Netherlands GUILLAUME BASTILLE‐ROUSSEAU,2 Department of Fish, Wildlife, and Conservation Biology, Colorado State University, Fort Collins, CO 80523, USA GUILLAUME BASTILLE‐ROUSSEAU,2 Department of Fish, Wildlife, and Conservation Biology, Colorado State University, Fort Collins, CO 80523, USA ALBERTUS G. TOXOPEUS, Department of Natural Resources, Faculty of Geo‐Information Science and Earth Observat University of Twente, P.O. Box 217, 7500 AE, Enschede, The Netherlands IN DOUGLAS‐HAMILTON,4 Save the Elephants, P.O. Box 54667 ‐ 00200, Nairobi, Kenya, Department of Zoology, Univ OX1 3PS, United Kingdom ABSTRACT Poaching is the most immediate threat to African elephants (Loxodonta africana). Several con- tinental‐wide surges in poaching have occurred since the latter half of the twentieth century, and the latest surge occurred from 2007 to 2012. The behavioral responses of elephants to poaching risk has not been studied widely because of a lack of high‐resolution movement data collected simultaneously with veriﬁed causes of mortality. We managed to collate 2 such datasets from 2004 to 2013. We studied the spatial‐temporal changes in movement behavior of 11 elephants in their core areas. Past studies have focused on elephant movement along corridors. We tested for the eﬀect of poaching risk on their path straightness (i.e., tortuosity) while controlling for other environmental and human activities in the landscape using a set of generalized linear mixed models. To test for temporal variation of tortuosity, we used a time‐series linear model. Elephants turned less frequently while they were in poaching locations and at times with a high level of poaching activity, even though their speed did not change. The variation of tortuosity is a good indicator of diﬀerences in poaching risk as perceived by the elephants, which could complement patrol‐based anti‐poaching eﬀorts by wildlife managers, especially in remote, inaccessible landscapes. © 2019 The Authors. The Journal of Wildlife Management published by Wiley Periodicals, Inc. The Journal of Wildlife Management 1–10; 2019; DOI: 10.1002/jwmg.21688 Research Article 1E‐mail: ihwagi@gmail.com 2Other aﬃliation: Save the Elephants, P.O. Box 54667‐00200, Nairobi, Kenya. 3Current aﬃliation: Department of Environmental Science, Macquarie University, NSW, 2106, Australia. 4Other aﬃliation: Department of Zoology, University of Oxford, Oxford, OX1 3PS, United Kingdom. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. r r t a w s - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - This is an open access article under the terms of the Creative Commons Attribution Licens original work is properly cited. 1E‐mail: ihwagi@gmail.com 2Other aﬃliation: Save the Elephants, P.O. Box 54667‐00200, Nairobi, Kenya. 3Current aﬃliation: Department of Environmental Science, Macquarie University, NSW, 2106, Australia. 4Other aﬃliation: Department of Zoology, University of Oxford, Oxford, OX1 3PS, United Kingdom. Poaching Lowers Elephant Path Tortuosity: Implications for Conservation There are, therefore, links between an animal’s path tortuosity, its foraging tenacity, and the current and historical exposures to risk. There is a widespread need for scientists, conservationists and wildlife managers to better understand the variation in animals’ behavioral responses to habitats that are undergoing rapid changes due to human activities like habitat loss and illegal harvesting (Sih 2013). p Between 2007 and 2012, the levels of poaching increased gradually, aﬀecting all African elephant populations (Wit- temyer et al. 2014, Chase et al. 2016). The population of the Laikipia‐Samburu ecosystem, which lies under diﬀerent conservation statuses, is the most intensely monitored in Africa regarding the causes of elephant mortality. A participatory network veriﬁes each report of a dead elephant (Kahindi et al. 2010). During the 2007 to 2012 poaching surge, the levels of poaching in the ecosystem increased in tandem with the continental trend, and even some of the previously safe land units, like private ranches, were aﬀected by the surge (Douglas‐Hamilton et al. 2010, Ihwagi et al. 2015). Our study goal was to evaluate elephants’ behavioral adaptation to the increasing levels of illegal killing in their core foraging areas using path tortuosity as the main study parameter. We predicted that the tortuosity of elephants would be lower in places and times when there were high levels of illegal killing. Using movement and mortality data collected before and during the period when poaching levels rose so markedly, we investigated how the tortuosity of elephants varied with poaching levels spatially and temporally. g Besides the loss of habitat, the most immediate threat to African elephants (Loxodonta africana) is poaching and other deadly conﬂicts with humans (Nellemann et al. 2013, Wittemyer et al. 2014). Hunting by humans is a form of predation risk (Frid and Dill 2002, Shannon et al. 2014). The poacher‐elephant system has already oﬀered numerous insights into how prey can respond to the risk of encountering humans. For instance, by increasing speed when traversing unsafe areas, switching from the day to night‐time travel, and avoiding settlements during the daytime (Douglas‐Hamilton et al. 2005, Graham et al. 2009, Wittemyer et al. 2017, Ihwagi et al. 2018). These studies focused on the alteration of speed when elephants venture out of their (presumably safer) core areas. Elephants increase their speed in migratory corridors to traverse unsafe areas quickly, a behavior called streaking (Douglas‐Hamilton et al. 2005). Poaching Lowers Elephant Path Tortuosity: Implications for Conservation 2013, Wittemyer et al. 2014, Chase et al. 2016), it is still unclear how elephants adjust their movement in response to poaching risk on short time‐scales. Several studies have reported on elephant movement behavior along their migration corridors (Douglas‐Hamilton et al. 2005, Galanti et al. 2006, Mpanduji et al. 2009, Jachowski et al. 2013, Roever et al. 2013). The home ranges of migratory elephants comprise core areas linked with narrow corridors through which they migrate at faster speeds than usual (Douglas‐Hamilton et al. 2005). Elephants walking through safe or risky areas, however, can have the same average speed between path segments but show a varied tortuosity in each area because the risks inﬂuence the animals’ rates of turning along the way (Angilletta et al. 2008). Speed, therefore, has limited applicability to understanding the eﬀect of the risk landscape on animal movement. larger ones moving in straighter paths (i.e., with lower tortuosity; Hodges et al. 2014). How risk inﬂuences the tortuosity of animals is poorly understood because most studies have been conducted in laboratories or using mathematical models, rather than in complex natural habitats (Domenici et al. 2011b). It has been hypothesized that if large animals can perceive danger and make eﬀorts to avoid it, the rate of tortuosity should decrease in risky habitats to minimize their exposure to risk (Domenici et al. 2011a, Hodges et al. 2014). An animal’s ability to assess and inﬂuence the risk of encountering predators by changing its behavior aﬀects its decision‐making (Sih 1987). Behavioral plasticity is essen- tial for an animal to be able to respond to fast changes in the environment brought about by rapid human activities (Sih 2013). The landscape of fear is an individual animal's perception of fear, a concept that is not geographically dependent (Laundré et al. 2010). The mean rate of change of an animal’s foraging tenacity is the best measure of its landscape of fear; the higher the diﬀerence in risk, the higher the mean diﬀerence in foraging tenacity should be (Laundré 2010, Bleicher 2017). The path tortuosity of a herbivore when foraging is commensurate with the hetero- geneity and spatial distribution of its preferred forage items (Etzenhouser et al. 1998). The animal’s past experiences in the landscape, including encounters with predators, inﬂu- ence its current landscape of fear (Bleicher 2017) and, in turn, how it forages. Poaching Lowers Elephant Path Tortuosity: Implications for Conservation on behalf of The Wildlife Society EY WORDS elephants, GPS tracking, Laikipia ‐ Samburu, movement behavior, path tortuosity, poaching. inversely proportional to the eﬃciency of the orientation mechanism involved, unless major obstacles abound (Benhamou 2004). The nature of survival strategies inﬂuences the path tortuosity of animals that live in heterogeneous landscapes, but it is aﬀected by intrinsic and extrinsic factors (Schooley and Wiens 2004, Prevedello et al. 2010, Sih 2013). An animal should exhibit its maximum eﬀort towards avoiding predators during high‐ risk times (Lima and Bednekoﬀ1999). If animals can reliably detect and make eﬀorts to avoid exposure to the risk, movement tortuosity should be inversely proportional to the magnitude of risk in space and time to reduce the animals’ exposure (Hodges et al. 2014). Maneuverability while escaping risk is also inﬂuenced by body size, with smaller, faster animals being able to turn more sharply, and Animals that live in landscapes that have been recently altered by humans exhibit sub‐optimal movement behavior as they seek to evade predators (Fahrig 2007). The path tortuosity of an animal is a measure of how direct a path is from a certain point to the intended destination, and in open landscapes with no physical obstacles or barriers, it is Received: 16 March 2018; Accepted: 11 March 2019 1E‐mail: ihwagi@gmail.com 2Other aﬃliation: Save the Elephants, P.O. Box 54667‐00200, Nairobi, Kenya. 3Current aﬃliation: Department of Environmental Science, Macquarie University, NSW, 2106, Australia. 4Other aﬃliation: Department of Zoology, University of Oxford, Oxford, OX1 3PS, United Kingdom. Ihwagi et al. • Tortuosity and Poaching Risk 1 Poaching has long‐term eﬀects on the population of elephants aﬀected, and this is evident from the sustained high levels of stress hormones detected for as long as 6 years after a poaching surge (Gobush et al. 2008). Elephants switch to moving more at night than in the day in times and places with high levels of illegal killing (Ihwagi et al. 2018); however, little is known about what factors inﬂuence the ﬁne‐scale variation in behavior in relation to poaching risk. The hourly adjustment of movement behavior in relation to risk is poorly understood because few studies have access to a multi‐year high‐resolution movement dataset and a concurrent dataset of causes of elephant mortality veriﬁed in the ﬁeld. Our study has compiled and made use of such datasets. Although poaching causes a decline in many populations of African elephants (Nellemann et al. Poaching Lowers Elephant Path Tortuosity: Implications for Conservation In a landscape dominated by humans, in which diﬀerent core areas for elephants lie in distinct land management units that have diﬀerent levels of poaching, we would expect elephants to change their short‐ term movement behavior as they forage in full vigilance. Change in behavior is an aspect that movement speed alone would probably not capture. STUDY AREA We conducted this study from 2002–2013 in the Laikipia‐ Samburu ecosystem, Kenya, which covers an area bounded by 0.2°S to 1.5°N, and 36.2°E to 38°E (Fig. 1). The The Journal of Wildlife Management • 2 Figure 1. Map of Laikipia‐Samburu ecosystem showing the coverage of global positioning system (GPS) ﬁxes for 11 African elephants tracked between 2004 and 2013, and the ecosystem’s complexity with regard to land use types. Figure 1. Map of Laikipia‐Samburu ecosystem showing the coverage of global positioning system (GPS) ﬁxes for 11 African elephants tracked between 2004 and 2013, and the ecosystem’s complexity with regard to land use types. ecosystem is delineated by the extent of the Ewaso Nyiro River and its tributaries, encompassing roughly 33,000 km2 (Thouless 1995). It has a north‐south rainfall gradient, with a marked decline in rainfall towards the north (Georgiadis 2011). The rainfall pattern in the ecosystem is highly variable and bimodal, with peaks in April and November and a yearly range of <400 mm in the north to a maximum of 600 mm in the south (Barkham and Rainy 1976, Ihwagi et al. 2012). The onset of the rains in April and November marks the beginning of short and long rainy seasons, respectively. The terrain in the northern part of the landscape comprises expansive plains interrupted by rugged terrain and isolated hills. In the northern part of the landscape, the surface geology is characterized by complex rocks that consist of horn‐blend gneisses and schists, and banded biotite gneisses (Krhoda et al. 2015). The southern part of the landscape is marked by tertiary and more recent ﬂows of olivine basalt giving rise to a plateau of poorly structured and excessively drained volcanic soils. ecosystem is delineated by the extent of the Ewaso Nyiro River and its tributaries, encompassing roughly 33,000 km2 (Thouless 1995). It has a north‐south rainfall gradient, with a marked decline in rainfall towards the north (Georgiadis 2011). The rainfall pattern in the ecosystem is highly variable and bimodal, with peaks in April and November and a yearly range of <400 mm in the north to a maximum of 600 mm in the south (Barkham and Rainy 1976, Ihwagi et al. 2012). The onset of the rains in April and November marks the beginning of short and long rainy seasons, respectively. STUDY AREA The terrain in the northern part of the landscape comprises expansive plains interrupted by rugged terrain and isolated hills. In the northern part of the landscape, the surface geology is characterized by complex rocks that consist of horn‐blend gneisses and schists, and banded biotite gneisses (Krhoda et al. 2015). The southern part of the landscape is marked by tertiary and more recent ﬂows of olivine basalt giving rise to a plateau of poorly structured and excessively drained volcanic soils. elatior) and fever tree (Vachellia xanthophloea), which attracts large numbers of animals in the dry months (Ihwagi et al. 2010). The diverse plant community supports an equally diverse fauna. In addition to elephants, the ecosystem is home to other large mammals that include reticulated giraﬀe (Girrafa camelopardalis), plains zebra (Equus quagga), Grevy’s zebra (Equus grevyi), Cape buﬀalo (Syncerus caﬀer), impala (Aepyceros melampus), defassa waterbuck (Kobus ellipsiprymnus defassa), and common waterbuck (Kobus ellipsiprymnus; Ihwagi et al. 2010). Other animal species found in the landscape but in relatively fewer numbers include beisa oryx (Oryx beisa), eland (Taurotragus oryx), and gerenuk (Litocranius walleri). The landscape is a complex mosaic of land use types owned by individuals, government, or communities, and elephants walk freely through most of the land units. There is a signiﬁcant variation in levels of poaching within diﬀerent land units, with higher poaching levels being associated with communal grazing areas (Ihwagi et al. 2015). The vegetation of the Laikipia‐Samburu ecosystem is dominated by wooded grassland with an underlying large cover of perennial and annual grasses and thorny acacia {Acacia spp.) scrubland (Pratt et al. 1966, Barkham and Rainy 1976). The major river in the landscape, the Ewaso, supports a riverine forest dominated by river acacia (Acacia Ihwagi et al. • Tortuosity and Poaching Risk METHODS Monitoring Illegal Killing of Elephants (MIKE) Under the auspices of the MIKE program of the Convention on International Trade in Endangered Species (CITES), 3 separate family of ≥10 elephants. With the help of government veterinarians, we ﬁtted the elephants with GPS collars sourced from African Wildlife Tracking (Pretoria, South Africa) and FOLLOWIT (Lindesberg, Sweden); each recorded 1 ﬁx/hour. We abbreviated sex as F for a female, M for a male and assigned a number for ease of reference (e.g., F1 for female 1). The full identities and sex of elephants tracked were Loldaiga (F1), Wangari (F2), Ngelesha (M1), Ol ari Nyiro (M2), Sera (F3), Genghis (M3), Mpala (M4), Mutara (M5), Olpejeta (M6), Tia Maria (F4), and Drachmae (F5), but hereafter we refer to each of them using the number in parentheses. The elephants exhibited more directed movements on a ﬁne temporal scale (1 hr) compared to larger temporal scales (Street et al. 2018), which made use of hourly GPS ﬁxes useful in inferring behavioral change. Five of the elephants; F1, F2, F3, M1, and M2 were wide‐ ranging (migratory) and used multiple, distant land units, making them ideal candidates to test spatial trends, whereas the other 6 were residents of distinct land units in diﬀerent parts of the ecosystem and thus were ideal candidates for testing temporal trends within each elephant’s home. We ﬁltered out GPS points that corresponded to unlikely speeds of >7 km/hour (Hutchinson et al. 2003, Hutchinson et al. 2006). causes of elephant mortality were veriﬁed in the ﬁeld by a participatory community network of landowners, herders, conservancy managers, and government wildlife rangers (Kahindi et al. 2010). We analyzed data collected from 2002 to 2013. The proportion of illegally killed elephants (PIKE) has been adopted as an unbiased estimator of the levels of poaching (Douglas‐Hamilton et al. 2010, Jachmann 2013). The formula for calculating PIKE is: 100 × The number of illegally killed elephants comprises those that died from poaching and those that died from human‐ elephant conﬂicts. The PIKE is a reliable metric for comparing levels of illegal killing even between sites with diﬀerent sampling eﬀorts per unit area (Douglas‐Hamilton et al. 2010, Jachmann 2013). Despite PIKE comprising both human‐elephant conﬂict incidents and poaching ones, in the Laikipia‐Samburu MIKE site, poaching accounts for the highest numbers of illegally killed elephants, with over 91% of illegally killed elephants in the MIKE site in any of the years studied (Ihwagi et al. 2015). Because the number of poached carcasses represents the majority of all illegally killed elephants, the estimates of PIKE in the ecosystem are almost synonymous with the actual poaching levels, and we can thus make direct reference to poaching levels in many instances. We used a ﬁxed‐kernel home range estimator to identify core areas, which we delineated as those areas within 50% contour lines (Worton 1987, 1989). We tracked the elephants on diﬀerent dates between September 2004 and December 2013 (Table 1). For each hourly GPS ﬁx, we calculated the tortuosity as log(L/R2), where L is the hourly segment length and R is the net displacement (Whittington et al. 2004). We used the current and previous 2 GPS points to calculate tortuosity at any point as follows: assuming hourly measurements, the net displacement at 1100 hours would be the straight‐line distance between the elephant’s location at 0900 hours and 1100 hours. Tortuosity measures have an exponential distribution. A straight line (i.e., a line with the lowest tortuosity) has a value of zero or negative. The magnitude of the negative value depends on the actual ratios of the 2 successive, unidirectional line segments whose simple addition equals the displacement. The longer the immediate step length compared to the previous one on We calculated PIKE values for each core area by averaging the PIKE of the land units used by the elephants. To determine the PIKE value associated with an elephant’s home range, we included all mortality records for the time before and during its global positioning system (GPS) tracking period because elephants have a long‐term spatial memory (Gobush et al. 2008, Polansky et al. 2015) and we assumed they were still responding to past experiences encountered before we put collars on them. The Journal of Wildlife Management • PIKE % Number of illegally killed elephants Total number of dead elephants recorded ( ) = PIKE % Number of illegally killed elephants Total number of dead elephants recorded ( ) = 100 × Collating other Environmental Variables Collating other Environmental Variables To disentangle the role of poaching risk from other environmental factors aﬀecting the tortuosity of elephants, we modeled their tortuosity as a function of the poaching levels, land cover, type of terrain, distance from water points, livestock density, and the presence of permanent and temporary human dwellings in the landscape. We adopted a grid size of 500 m × 500 m for environmental variables, which aligns conveniently to the Universal Transverse Mercator grid and is also the same size as the grid adopted for measures of tortuosity. We performed a sensitivity analyses of tortuosity to the changes in grid sizes by calculating the variability of mean tortuosity values within progressively increasing size of grids (50 m, 100 m, 150 m, 200 m, 300 m, 400 m, 500 m, 750 m, and 1,000 m) using GPS data for the wide‐ranging elephants (i.e., those with multiple core areas; Fig. S1, available online in Supporting Information). We found that tortuosity is resistant to the changes in grid sizes. Land cover inﬂuences the permeability of habitat to animals (Saunders et al. 1993, Goad et al. 2014, Cooney et al. 2015). We used land cover data provided by the Food and Agriculture Organization to assign the dominant cover type in each grid square (Food and Nations 2015). The 14 land cover types were rain‐fed herbaceous crop, scattered (in natural vegetation or other) rain‐fed herbaceous crop (ﬁeld density 20–40% of polygon area), isolated (in natural vegetation or other) rain‐fed herbaceous crop (ﬁeld density 10–20% polygon area), closed trees, open trees (65–40% crown cover), very open trees (40–15% crown cover), closed to open woody vegetation (thicket), open shrubs (45–40% crown cover), open low shrubs (65–40% crown cover), open to closed herbaceous vegetation, shrub savannah, tree‐ and shrub savannah, open to closed herbaceous vegetation on temporarily ﬂooded areas, and natural waterbodies. Elephants and livestock in the ecosystem overlap in space and time and share critical resources like water and forage (Raizman et al. 2013). From the results of 2 total‐count aerial censuses conducted in 2008 and 2012, we combined the numbers of all livestock species (i.e., cattle, sheep, goats, donkeys, and camels) for each census year, and calculated their average density per square kilometer in each grid. The aerial census dataset also included locations of occupied bomas. The pastoral communities shift their bomas over time and build new ones. GPS Tracking and Calculation of Path Tortuosity We collected hourly GPS positions of 11 mature elephants (7 females aged 30–40 yr, and 4 males aged >30 yr). The males were solitary, but each of the females represented a Table 1. The dates of tracking for each of the 6 male (M) and 5 female (F) elephants in Laikipia‐Samburu ecosystem, Kenya, and the number of hours they spent in their respective core areas. Elephant identity (sex and number) Dates tracked Hours in core area 1 Hours in core area 2 F1 Aug 2006–Jan 2009 3,383 3,597 F2 Sep 2004–Jul 2006 10,649 1,511 M1 Aug 2008–Feb 2012 2,874 1,524 M2 Aug 2007–Feb 2012 7,241 14,327 F3 Aug 2006–Apr 2011 11,078 12,773 M3 May 2004–Jun 2012 41,943 NAa M4 Feb 2007–Dec 2011 39,134 NA M5 Feb 2009–Jun 2013 13,095 NA M6 Sep 2006–Feb 2009 17,022 NA F4 Feb 2008–Aug 2012 31,715 NA F5 Feb 2008–Dec 2013 37,353 NA a Indicates the elephant had only 1 core area (was not migratory). Table 1. The dates of tracking for each of the 6 male (M) and 5 female (F) elephants in Laikipia‐Samburu ecosystem, K spent in their respective core areas. 4 Human settlements negatively aﬀect the distribution of elephants and, in most instances, tend to occupy prime resource areas for the elephants, thereby inadvertently setting the stage for competition (Verlinden 1997). The landscape is dominated by temporary shelters for nomadic pastoralists (bomas, which were vacated seasonally), and by permanent settlements around the few, scattered towns. We calculated Euclidean distance to each category of settlement. We hypothesized that the bomas and permanent settlements would aﬀect elephants diﬀerently because of the contrasting occupations of residents, and thus categorized them as separate features. We classiﬁed the 2 settlement types (2 features) and calculated for each the distances from each grid’s centroid to the nearest feature. a straight line, the more negative the tortuosity value. A high tortuosity is achieved when an elephant makes an acute angle resulting in very short displacement, and the tortuosity approximates a value of 2. Should the elephant turn and walk back through the previous location, the net displacement would be less than that step length line segment, and the tortuosity would thus be negative, as expected for such a highly tortuous path. This explains why the use of actual displacement alone is not as good a parameter for elephant movement as calculating tortuosity. Collating other Environmental Variables The shift, however, is often only over a short distance and within the same land management unit that we adopted for calculating PIKE; it does not therefore aﬀect our analyses. Wildlife fences create edge eﬀects and modify elephant movement behavior directly by coercing them onto a unidirectional path (Newmark 2008, Vanak et al. 2010, Nams 2014). We excluded all GPS tracking data points that fell within a distance of 2 km of any wildlife fence from our analyses. Ihwagi et al. • Tortuosity and Poaching Risk GPS Tracking and Calculation of Path Tortuosity Theoretically, should the elephant walk back to its precise previous location, this would yield a displacement of zero, and the calculation for tortuosity would yield an error; however, in our analyses, such an occurrence is extremely rare, and could be ﬂagged, manually checked, and corrected. We adhered to the guidelines for the care of elephants by employing the services of a government veterinarian from the Kenya Wildlife Service to immobilize the elephants for ﬁtting their collars. The elephants had multiple core areas, although the migration from one core to another was not in any regular cycle that could be linked to the seasons. The availability of surface‐water inﬂuences elephant distribution because they need to drink at least once a day (Buss 1961, Western 1975). Elephants turn less when they are farther away from a source of water (Duﬀy et al. 2011). We calculated the distance from each grid’s centroid to the nearest known water point. Elephants avoid climbing hills, with a gradient of 30 degrees being prohibitive (Wall et al. 2006), so we assumed that even at a distance away from the base of a continuous ridge or escarpment, elephants would align their movement to circumvent it. We deﬁned the base of the hills as the level with a gradient of >30 degrees on a 30‐m digital elevation model. We calculated the distance from each grid’s center to the base of the nearest hill or escarpment. Statistical Analysis We analyzed data using R 3.3.0 (R Development Core Team 2012). The calculation of tortuosity included a log transformation that corrected for non‐normality. To test for spatial variation of the tortuosity of the 5 migratory elephants (F1, F2, F3, M1, and M2) in relation to the level of illegal killing and other covariates, we used 28 generalized linear mixed models grouped into 3 categories: physical environmental factors, human activity related factors, and plausible combinations of these in a model 5 selection framework (Table 2). We performed a quasi‐ experiment that controlled for individual elephant, sex, and spatial autocorrelation of points. We scaled all covariates to enable direct comparison of the coeﬃcients in the model summary. To account for spatial autocorrelation, we included an exponential correlation function based on the coordinates of the centers of each grid cell. The auto- correlation term also included control for repeated measures from individual elephants. We included the elephant’s identity in the model as a random eﬀect. Some of the models included interactions between some variables, where plausible. We selected the best model as the one with the lowest second‐order Akaike’s Information Criterion (AICc) value within the model selection framework (Burnham and Anderson 2002). We implemented the mixed eﬀects regression models using the function lme within the package nlme. Table 2. Candidate models of factors aﬀecting tortuosity of 5 migratory elephants (F1, F2, F3, M1, and M2) that inhabited diﬀerent land man- agement units in the Laikipia‐Samburu ecosystem, Kenya, 2004–2013. All models included sex and elephant identiﬁcation as a random factor in addition to the listed covariates. Scenario 3 PIKE + livestock density + land cover type PIKE l k d d f The Journal of Wildlife Management • Statistical Analysis Model Variables Scenario 1 Only physical environment aﬀects path tortuosity Water Hills Land cover Scenario 2 Only human activity aﬀects path tortuosity PIKEa + livestock density × distance from boma + distance from town PIKE × distance from town + livestock density PIKE + livestock density + distance from boma PIKE + distance from town PIKE × distance from town + distance from boma PIKE + livestock density × distance from boma PIKE + livestock density Livestock density Livestock density + distance from hill Livestock density + distance from boma Livestock density + distance from town + distance from boma Livestock density × distance from town Livestock density × distance from boma Distance from town + distance from boma PIKE + sex Scenario 3 Both human and physical environmental factors aﬀect tortuosity PIKE + livestock density × distance from town + distance from water + distance from hills + distance from boma PIKE + livestock density + land cover type PIKE + livestock density + distance from town + distance from water PIKE × livestock density + distance from town + distance from water + land cover type PIKE + land cover type Distance from town + distance from water + distance from hills PIKE + livestock density + distance from hills PIKE + distance from hills Livestock density + land cover type Distance from town + distance from water + distance from hills + land cover type p g To examine temporal changes in tortuosity and poaching, we used data from the 2 elephants (M2 and M4) that inhabited 2 regions: land units that experienced a gradual increase in levels of illegal killing (i.e., private ranches) and poaching‐free sanctuaries. The other 4 were residents of 2 poaching‐free rhino (white rhino [Ceratotherium simum] and black rhino [Diceros bicornis]) sanctuaries, Lewa and Ol Pejeta conservancies, at distant locations from each other. We modeled tortuosity as a function of time and land use type (i.e., poaching‐free sanctuary or private ranches) with elephant identity as a random eﬀect, using a linear mixed‐ eﬀect model. The modeling of change on tortuosity over time focused only on the 2 elephants where there was a temporal trend in tortuosity. We tested statistical signiﬁ- cance at α = 0.05. a Proportion of illegally killed elephants. Data Availability The site summaries of elephant mortality data are publicly available on the CITES website (www.cites.org). The GPS coordinates of elephant positions are not publicized because they are an endangered species. Exposing their a Proportion of illegally killed elephants. Table 3. The ranking for diﬀerent top‐ranked models predicting tortuosity of elephant movement paths in the Laikipia‐Samburu ecosystem, Kenya, 2004–2013. Modela Kb AICc c ΔAICc d Like wi f LLg Cum. wth (PIKE × town) + boma + sex 9 6,520.89 0.00 1.00 0.32 −3,251.43 0.32 (PIKE × livestock) + town + water + land cover + sex 25 6,521.10 0.21 0.90 0.29 −3,235.42 0.61 PIKE + (livestock × boma) + town + sex 10 6,522.19 1.30 0.52 0.17 −3,251.07 0.78 PIKE + (livestock × town) + water + hills + boma + sex 12 6,522.65 1.76 0.41 0.13 −3,249.29 0.92 nt top‐ranked models predicting tortuosity of elephant movement paths in the Laikipia‐Samburu ecosystem, Kenya, Table 3. The ranking for diﬀerent top‐ranked models predicting tortuosity of elephant movement paths in the La 2004 2013 a PIKE is the proportion of illegally killed elephants; boma is the distance to temporary settlements of nomadic pastoralists; water refers to the distance to nearest water point; and hills represents the distance to the base of the nearest hill, town refers to the distance to the nearest town, land cover refers to the physical material at the surface of the earth, livestock refers to the average density of livestock species per square kilometer counted form air in 2008 and 2018. p c Second‐order Akaike’s Information Criterion. d d The diﬀerence in AICc between the current and the most appropriate model. 6 -8 -7 -6 -5 -4 -3 -2 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 Log(tortuosity) M3 M4 F4 F5 M5 M6 0% 10% 20% 30% 40% 50% 60% 70% 80% -7.5 -7 -6.5 -6 -5.5 -5 -4.5 -4 2002 2004 2006 2008 2010 2012 PIKE Log(tortuosity) M3 M4 PIKE Poly. (PIKE) -3.5 -3 -2.5 -2 -1.5 -1 -0.5 0 0.5 (Intercept) Year Ranches Year × Ranches Estimate (C) (B) (A) Figure 2. A) The mean annual tortuosity values of 6 male and 5 female elephants tracked in the Laikipia‐Samburu ecosystem, Kenya, at diﬀerent dates between 2004 and 2013. Four elephants, i.e. F4, F5, M5, and M6 that inhabited poaching‐free sanctuaries showed a constant mean annual tortuosity throughout the period. RESULTS All models that included PIKE ranked above all other models. We chose the top‐ranked model, which comprised of PIKE, distance from town, distance from boma, and sex, because of its parsimony (Table 3). The signiﬁcant factors in the best model included PIKE and distance from settle- ments (towns and bomas; Table 4). The interaction between level of illegal killing and town was not signiﬁcant even though it was included in the top model (P = 0.548), and the coeﬃcients for the distance from settlements were very small despite being statistically signiﬁcant. The tortuosity of elephants reduced by a factor of 0.4, i.e., exp(Estimate), for every 1% increase in the proportion of illegal killing. For every increase in 500 m of distance from towns and bomas, the median tortuosity values increased by a factor of 1.1 and decreased by 0.9, respectively. The second‐ranked model included PIKE, the density of livestock, distance from town, distance from water, land cover, and sex and had a similar AICc value, despite having more parameters. Figure 2. A) The mean annual tortuosity values of 6 male and 5 female Figure 2. A) The mean annual tortuosity values of 6 male and 5 female elephants tracked in the Laikipia‐Samburu ecosystem, Kenya, at diﬀerent dates between 2004 and 2013. Four elephants, i.e. F4, F5, M5, and M6 that inhabited poaching‐free sanctuaries showed a constant mean annual tortuosity throughout the period. M3 and M4 inhabited the Laikipia Ranches, where poaching levels increased gradually, and their tortuosity decreased. B) The yearly mean tortuosity of M3 and M4, decreased over time as the yearly proportion of illegally killed elephants (PIKE) increased. A second order polynomial (Poly) best represented the trendline of PIKE and for this dataset we plotted data collected from 2002 for the purpose of presenting the historical trend. C) The coeﬃcients of the linear mixed‐ eﬀects model of change in tortuosity of elephants as a function of (regional) poaching levels over time. Ranches represents a contiguous set of land units managed primarily for large‐scale cattle production but also hosts wildlife. town indicated that elephants walk with reduced tortuosity as they come near towns. On the contrary, we found that elephants walked with higher tortuosity when they were near the dwellings of nomadic people (bomas). Data Availability M3 and M4 inhabited the Laikipia Ranches, where poaching levels increased gradually, and their tortuosity decreased. B) The yearly mean tortuosity of M3 and M4, decreased over time as the yearly proportion of illegally killed elephants (PIKE) increased. A second order polynomial (Poly) best represented the trendline of PIKE and for this dataset we plotted data collected from 2002 for the purpose of presenting the historical trend. C) The coeﬃcients of the linear mixed‐ eﬀects model of change in tortuosity of elephants as a function of (regional) poaching levels over time. Ranches represents a contiguous set of land units managed primarily for large‐scale cattle production but also hosts wildlife. Table 4. The standardized coeﬃcients for the top‐ranked model of tor- tuosity of elephant movement paths in the Laikipia‐Samburu ecosystem, Kenya, 2004–2013, as a function of the proportion of illegally killed ele- phants (PIKE) and other human and environmental variables. -8 -7 -6 -5 -4 -3 -2 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 Log(tortuosity) M3 M4 F4 F5 M5 M6 (A) (A) Estimate Lower 95% CI Upper 95% CI SE (Intercept) −6.705 −7.451 −5.958 0.381 PIKE −1.049 −1.113 −0.984 0.033 Distance from town (km) 0.057 0.009 0.105 0.025 Distance from bomas (km) −0.136 −0.187 −0.085 0.026 Sex‐male 1.599 −0.319 3.516 0.602 PIKE × distance from town 0.012 −0.032 0.057 0.023 F5 M5 M6 0% 10% 20% 30% 40% 50% 60% 70% 80% -7.5 -7 -6.5 -6 -5.5 -5 -4.5 -4 2002 2004 2006 2008 2010 2012 PIKE Log(tortuosity) M3 M4 PIKE Poly. (PIKE) (B) -3.5 -3 -2.5 -2 -1.5 -1 -0.5 0 0.5 (Intercept) Year Ranches Year × Ranches Estimate (C) (C) (B) locations may be counterproductive to ongoing anti‐ poaching eﬀorts. locations may be counterproductive to ongoing anti‐ poaching eﬀorts. RESULTS The latter was not surprising because elephants have traditionally co‐existed with the nomadic pastoralists (Gadd 2005, Kideghesho 2008), and bomas are thus not perceived as a risk by elephants. These temporal dwellings of nomads in northern Kenya are often located close to key resource areas like water and forage (Berger 2003), resources that elephants also seek (Ngene et al. 2009, Bohrer et al. 2014). The tortuosity of the 2 elephants (M3 and M4) that inhabited private ranches, where PIKE increased gradually throughout the study period, were lower and decreased more over time compared to the tortuosity of those that inhabited poaching‐free rhino sanctuaries (Fig. 2A). The level of illegal killing over time was inversely related to the change in tortuosity for the 2 elephants in private ranches (Fig. 2B). The coeﬃcients of the model showed a strong inﬂuence of PIKE levels for the elephants in private ranches (Fig. 2C). Temporally, the variation in tortuosity corresponded to the annual variation of PIKE in the aﬀected land use types, and the inclusion of poaching‐free land units was a perfect control case. Our results support earlier observations that habitat quality alone is not suﬃcient to explain elephant movement behavior (Boettiger et al. 2011) because human activities also aﬀect it (Lima 1987). This study identiﬁes the level of illegal killing as the best predictor of tortuosity of elephant movements in their crucial foraging areas. Ihwagi et al. • Tortuosity and Poaching Risk DISCUSSION Environmental factors inﬂuenced the tortuosity of all the elephants we studied, and, in addition, human factors inﬂuenced those elephants that were outside the (highly secured, low human density) rhino sanctuaries. The models with the highest support from the data featured PIKE as the most important variable. The distance from town is a proxy for human population density, likely disturbances, and the positive relationship between tortuosity and distance from Many studies have emphasised the importance of speed in describing elephant movement behavior within diverse contexts (Douglas‐Hamilton et al. 2005, Hutchinson 7 et al. 2006, Graham et al. 2009, Chamaille‐Jammes et al. 2013, Jachowski et al. 2013). Elephants use a direct movement strategy in which they maintain straighter paths rather than increase speed to reach an intended foraging spot (Duﬀy et al. 2011). Because of the high energy requirements of large animals, a change of direction is more favorable than a change of speed when navigating through a complex landscape or risk area (Wall et al. 2006, Angilletta et al. 2008). Unlike smaller species, where the trade‐oﬀbetween speed and tortuosity is more important (Angilletta et al. 2008), elephants with their large mass and high energy requirements have minimal ﬂexibility with regards to increasing their speed (Wall et al. 2006). They thus choose to walk straighter paths than to increase speed. The constant risk model postulates that vigilance is minimal in a visually obstructive environment, where scanning takes time and is too costly to the animal (Lima 1987). Logically, for animals with low visual acuity such as elephants (Rensch and Altevogt 1955), active vigilance would entail frequent turns and as a result higher tortuosity. The elephants, however, exhibited lower tortuosity in risky areas, implying they exercised low vigilance, in line with the constant risk model. 2004). The foraging of elephants involves variation in their daily displacement in relation to resource availability (Polansky et al. 2013). A risk‐induced reduction in tortuosity in the prime foraging areas thus implies a change to optimal foraging strategy. The foraging success of elephants in core areas in risky environments may have been compromised by the need to keep moving, instead of actively searching and exploiting food resources thoroughly. A study with a higher temporal resolution of data is warranted to better understand the eﬀect of reduced tortuosity on evolutionarily adapted foraging strategy. DISCUSSION In the locations and times with higher poaching risk, the elephants probably maximized their foraging at speciﬁc safe spots, which may not necessarily be the richest in forage quality. The major limitation to establishing a relationship between elephant movement tortuosity and poaching levels has been the lack of suﬃcient records of individually veriﬁed causes of elephant mortality recorded concurrently with movement data. Our study was able to collate such data and show that a change in elephant tortuosity values in circumstances other than migration or physical obstacles can be used as an indicator of the stress induced by poaching. We have shown that poaching aﬀects their natural pace and movement behavior. Future studies characterizing the movement behavior of elephants in risky habitats should therefore incorporate tortuosity because it is more informative than speed alone. We propose that relative change of tortuosity is a useful variable for assessing elephants’ landscape of fear in their crucial foraging areas; with regard to poaching it has a steep gradient that speed alone does not illustrate. The absolute values of tortuosity, however, depend on the temporal scale of the data and for a comparative study, the resolutions should be the same (Street et al. 2018). p The human population density, poverty levels, and livestock densities around MIKE sites are strongly corre- lated with poaching levels (CITES 2012). As the human population increases, the encroachment on elephant home ranges is expected to increase, and elephant movement behavior will be more aﬀected. In our study, however, despite featuring in the top model, the variables measuring distance from human dwellings had very low coeﬃcients. Also, there was no signiﬁcant interaction between PIKE and human settlements. This lack of interaction suggests that the inﬂuence of settlements is not necessarily through poaching, but possibly elephants simply avoiding encounters with humans. Another possible explanation is that ele- phants, intelligent as they are, do not associate risk with the mere presence of human dwellings. Instead, the elephants respond to speciﬁc human activities, and one response is to avoid real‐time encounters (Graham et al. 2009). Elephants are intelligent enough to distinguish levels of threat and they can even distinguish pastoralists from other persons by odor and clothing color (Bates et al. 2008). An animal’s behavioral response to a known risk depends on their assessment of the risk magnitude (Sih 2013). DISCUSSION A recent study explored the variation of tortuosity of elephant movement within corridors and core areas (Jachowski et al. 2013), but neither the variation of tortuosity between core areas nor its relationship with poaching risk was explored. We found that elephants turn more when in their respective low‐risk areas. This ﬁnding is consistent with past observations that when elephants are foraging, they turn less often (instead of increasing speed) to reach their intended point faster; this strategy saves energy (Duﬀy et al. 2011). Our results indicate that a high level of illegal killing and other risk factors make a habitat less favorable and lead elephants to walk straighter paths through it. The presence of both males and females in our group of elephants residing in poaching‐free rhino sanctuaries was a good control dataset to illustrate that tortuosity of both sexes changed primarily according to the risk from poaching and not because of other environmental factors or sex‐related diﬀerences. We none- theless found that the change in tortuosity was inﬂuenced by the level of risk and that although sex as a key variable did not have a signiﬁcant eﬀect in the top model, male elephants had a marginally higher tortuosity than females. The Journal of Wildlife Management • LITERATURE CITED Angilletta, M. J., T. C. Roth II, R. S. Wilson, A. C. Niehaus, and P. L. Ribeiro. 2008. The fast and the fractalous: speed and tortuosity trade oﬀ in running ants. Functional Ecology 22:78–83. Georgiadis, N. J., editor. 2011. Introduction: conserving wildlife in Kenya's Ewaso landscape. Smithsonian Contributions to Zoology 632:1–10. g gy Barkham, J. P., and M. E. Rainy. 1976. The vegetation of the Samburu‐ Isiolo Game Reserve. African Journal of Ecology 14:297–329. Goad, E. H., L. Pejchar, S. E. Reed, and R. L. Knight. 2014. Habitat use by mammals varies along an exurban development gradient in northern Colorado. Biological Conservation 176:172–182. Bates, L. A., K. N. Sayialel, N. W. Njiraini, C. J. Moss, J. H. Poole, and R. W. Byrne. 2008. Elephants classify human ethnic groups by odor and garment color. Current Biology 17:1938–1942. Gobush, K. S., B. M. Mutayoba, and S. K. Wasser. 2008. Long‐term impacts of poaching on relatedness, stress physiology, and reproductive output of adult female African elephants. Conservation Biology 22:1590–1599. g gy Benhamou, S. 2004. How to reliably estimate the tortuosity of an animal's path: straightness, sinuosity, or fractal dimension? Journal of Theoretical Biology 229:209–220. Graham, M. D., I. Douglas‐Hamilton, W. M. Adams, and P. C. Lee. 2009. The movement of African elephants in a human‐dominated land‐ use mosaic. Animal Conservation 12:445–455. Berger, R. 2003. Conﬂict over natural resources among pastoralists in northern Kenya: a look at recent initiatives in conﬂict resolution. Journal of International Development 15:245–257. Hodges, K. E., J. A. F. Cunningham, and L. S. Mills. 2014. Avoiding and escaping predators: movement tortuosity of snowshoe hares in risky habitats. Ecoscience 21:97–103. Bleicher, S. S. 2017. The landscape of fear conceptual framework: deﬁ- nition and review of current applications and misuses. PeerJ 5:e3772. Hutchinson, J., D. Famini, R. Lair, and R. Kram. 2003. Are fast‐moving elephants really running? Nature 422:493–494. Boettiger, A. N., G. Wittemyer, R. Starﬁeld, F. Volrath, I. Douglas‐ Hamilton, and W. M. Getz. 2011. Inferring ecological and behavioral drivers of African elephant movement using a linear ﬁltering approach. Ecology 92:1648–1657. Hutchinson, J. R., D. Schwerda, D. J. Famini, R. H. I. Dale, M. S. Fischer, and R. Kram. 2006. The locomotor kinematics of Asian and African elephants: changes with speed and size. Journal of Experimental Biology 209:3812–3827. Bohrer, G., P. Beck, S. Ngene, A. Skidmore, and I. Douglas‐Hamilton. 2014. LITERATURE CITED Elephant movement closely tracks precipitation‐driven vegetation dynamics in a Kenyan forest‐savanna landscape. Movement Ecology 2:2. gy Ihwagi, F. W., R. M. Chira, G. Kironchi, F. Vollrath, and I. Douglas‐ Hamilton. 2012. Rainfall pattern and nutrient content inﬂuences on African elephants' debarking behaviour in Samburu and Buﬀalo Springs National Reserves, Kenya. African Journal of Ecology 50:152–159. y y p gy Burnham, K. P., and D. R. Anderson. 2002. Model selection and multi- model inference: a practical information theoretic approach. Second edition. Springer‐Verlag, New York, New York, USA. p g g Buss, I. O. 1961. Some observations on food habits and behavior of the African elephant. Journal of Wildlife Management 25:131–148. Ihwagi, F. W., C. Thouless, T. Wang, A. K. Skidmore, P. Omondi, and I. Douglas‐Hamilton. 2018. Night‐day speed ratio of elephants as indicator of poaching levels. Ecological Indicators 84:38–44. Chamaille‐Jammes, S., G. Mtare, E. Makuwe, and H. Fritz. 2013. African elephants adjust speed in response to surface‐water constraint on foraging during the dry‐season. PLoS One 8:e59164. Ihwagi, F. W., F. Vollrath, R. M. Chira, I. Douglas‐Hamilton, and G. Kironchi. 2010. The impact of elephants, Loxodonta africana, on woody vegetation through selective debarking in Samburu and Buﬀalo Springs National Reserves, Kenya. African Journal of Ecology 48:87–95. Chase, M. J., S. Schlossberg, C. R. Griﬃn, P. J. C. Bouché, S. W. Djene, P. W. Elkan, S. Ferreira, F. Grossman, E. M. Kohi, K. Landen, P. Omondi, A. Peltier, S. A. Jeanetta Selier, and R. Sutcliﬀe. 2016. Continent‐wide survey reveals massive decline in African savannah elephants. PeerJ 4:e2354. Ihwagi, F. W., T. Wang, G. Wittemyer, A. K. Skidmore, A. G. Toxopeus, S. Ngene, J. King, J. Worden, P. Omondi, and I. Douglas‐Hamilton. 2015. Using poaching levels and elephant distribution to assess the conservation eﬃcacy of private, communal and government land in northern Kenya. PLoS One 10:e0139079. Convention on International Trade in Endangered Species [CITES]. 2012. Interpretation and implementation of the convention, species trade and conservation, elephants. Monitoring the illegal killing of elephants. Including Cop16 Doc. 53.1 Addendum. Pages 3–14 in Sixteenth meeting of the Conference of the Parties, Bangkok, Thailand. y Jachmann, H. 2013. Pilot study to validate PIKE‐based inferences at the site level. Pachyderm 52:72–87. Cooney, S. A., E. M. Schauber, and E. C. Hellgren. 2015. Comparing permeability of matrix cover types for the marsh rice rat (Oryzomys pa- lustris). Landscape Ecology 30:1307–1320. Jachowski, D. S., R. Slotow, and J. J. Millspaugh. 2013. ACKNOWLEDGMENTS Douglas‐Hamilton, I., T. Krink, and F. Vollrath. 2005. Movements and corridors of African elephants in relation to protected areas. Nat- urwissenschaften 92:158–163. We thank the Government of Kenya for making the mortality data available for this study and publication through the Kenya Wildlife Service (KWS). This research was conducted with direct support to the ﬁrst author from the Netherlands Fellowship Program (NFP‐PhD.14/ 44) and the Disney Conservation Fund through Wildlife Conservation Network. We thank Jackie Senior, English and Science editor, for editing the ﬁnal version of the manuscript. The long‐term elephant tracking program under the auspices of Save the Elephants (www.savetheelephants.org) has been supported in part by diﬀerent organizations at various times, including the Safaricom Foundation, JRS Biodiversity Foundation, and The Nature Conservancy. Douglas‐Hamilton, I., G. Wittemyer, and F. Ihwagi. 2010. Levels of il- legal killing of elephants in the Laikipia‐Samburu MIKE site. Pages 1–14 in Fifteenth meeting of the Conference of the Parties, Doha, Qatar. Duﬀy, K. J., X. Dai, G. Shannon, R. Slotow, and B. Page. 2011. Move- ment patterns of African elephants (Loxodonta africana) in diﬀerent habitat types. South African Journal of Wildlife Research 41:21–28. Etzenhouser, M. J., M. K. Owens, D. E. Spalinger, and S. B. Murden. 1998. Foraging behavior of browsing ruminants in a heterogeneous landscape. Landscape Ecology 13:55–64. Fahrig, L. 2007. Non‐optimal animal movement in human‐altered land- scapes. Functional Ecology 21:1003–1015. p gy Food and Nations. 2015. FAO GEONETWORK. FAO, Rome, Italy. Frid, A., and L. M. Dill. 2002. Human‐caused disturbance stimuli as a form of predation risk. Ecology and Society 6:11. Nature Conservancy. Gadd, M. E. 2005. Conservation outside of parks: attitudes of local people in Laikipia, Kenya. Environmental Conservation 32:50–63. Galanti, V., D. Preatoni, A. Martinoli, L. A. Wauters, and G. Tosi. 2006. Space and habitat use of the African elephant in the Tarangire‐Manyara ecosystem, Tanzania: implications for conservation. Mammalian Biology 71:99–114. MANAGEMENT IMPLICATIONS Elephants in landscapes with high levels of illegal killing move with reduced path tortuosity. Changes in path tortuosity can serve as a useful proxy for changes in levels of illegal killing at the site level and the success of diﬀerent models of conservation in contigous land units with diﬀerent conservation statuses. A reduction in path tortuosity implies reduced searching intensity per unit area, which might have negative implications in the foraging success of elephants in risky landscapes where food resources are limited. Lower path tortuosity of herbivores in heterogeneous habitats implies lower foraging eﬃciency because it corresponds to lower searching eﬀort (Doerr and Doerr The Journal of Wildlife Management • 8 LITERATURE CITED G., M. East, and H. Hofer. 2009. Analysis of habitat use by and preference of elephants in the selous‐niassa wildlife corridor, southern Tanzania. African Journal of Ecology 47:257–260. Street, G. M., T. Avgar, and L. Börger. 2018. Net displacement and temporal scaling: model ﬁtting, interpretation, and implementation. Methods in Ecology and Evolution 9:1503–1517. Nams, V. O. 2014. Tortuosity of habitat edges aﬀects animal movement. Landscape Ecology 29:655–663. Thouless, C. R. 1995. Long distance movements of elephants in northern Kenya. African Journal of Ecology 33:321–334. p gy Nellemann, C., R. K. Formo, J. Blanc, D. Skinner, T. Milliken, and T. De Meulenaer. 2013. Elephants in the dust – the African elephant crisis. A rapid response assessment. United Nations Environment Programme, Nairobi, Kenya. Vanak, A. T., M. Thaker, and R. Slotow. 2010. Do fences create an edge‐ eﬀect on the movement patterns of a highly mobile mega‐herbivore? Biological Conservation 143:2631–2637. g Verlinden, A. 1997. Human settlements and wildlife distribution in the southern Kalahari of Botswana. Biological Conservation 82:129–136. y Newmark, W. D. 2008. Isolation of African protected areas. Frontiers in Ecology and the Environment 6:321–328. Wall, J., I. Douglas‐Hamilton, and F. Vollrath. 2006. Elephants avoid costly mountaineering. Current Biology 16:527–529. Ngene, S. M., A. K. Skidmore, H. Van Gils, I. Douglas‐Hamilton, and P. Omondi. 2009. Elephant distribution around a volcanic shield dominated by a mosaic of forest and savanna (Marsabit, Kenya). African Journal of Ecology 47:234–245. Western, D. 1975. Water availability and its inﬂuence on the structure and dynamics of a savannah large mammal community. African Journal of Ecology 13:265–286. Polansky, L., I. Douglas‐Hamilton, and G. Wittemyer. 2013. Using diel movement behavior to infer foraging strategies related to ecological and social factors in elephants. Movement Ecology 1:13. gy Whittington, J., C. C. St. Clair, and G. Mercer. 2004. Path tortuosity and the permeability of roads and trails to wolf movement. Ecology and Society 9:4. Polansky, L., W. Kilian, and G. Wittemyer. 2015. Elucidating the sig- niﬁcance of spatial memory on movement decisions by African savannah elephants using state‐space models. Proceedings of the Royal Society B: Biological Sciences 282:20143042. Wittemyer, G., L. M. Keating, F. Vollrath, and I. Douglas‐Hamilton. 2017. Graph theory illustrates spatial and temporal features that structure elephant rest locations and reﬂect risk perception. Eco- graphy 40:598–605. Pratt, D. J., P. J. Greenway, and M. D. Gwynne. 1966. A classiﬁcation of East African rangeland, with an appendix on terminology. LITERATURE CITED Corridor use and streaking behavior by African elephants in relation to physiological state. Biological Conservation 167:276–282. p gy Doerr, V. A. J., and E. D. Doerr. 2004. Fractal analysis can explain in- dividual variation in dispersal search paths. Ecology 85:1428–1438. g Kahindi, O., G. Wittemyer, J. King, F. Ihwagi, P. Omondi, and I. Douglas‐Hamilton. 2010. Employing participatory surveys to monitor the illegal killing of elephants across diverse land uses in Laikipia‐ Samburu, Kenya. African Journal of Ecology 48:972–983. Domenici, P., J. M. Blagburn, and J. P. Bacon. 2011a. Animal escapology I: theoretical issues and emerging trends in escape trajectories. Journal of Experimental Biology 214:2463–2473. Kideghesho, J. R. 2008. Co‐existence between the traditional societies and wildlife in western Serengeti, Tanzania: its relevancy in contemporary wildlife conservation eﬀorts. Biodiversity and Conservation 17:1861–1881. p gy Domenici, P., J. M. Blagburn, and J. P. Bacon. 2011b. Animal escapology II: escape trajectory case studies. Journal of Experimental Biology 214: 2474–2494. Ihwagi et al. • Tortuosity and Poaching Risk 9 Krhoda, G. O., M. Amimo, and I. A. Nyandega. 2015. Geophysical in- vestigations of SuyienEarthdam in Maralal, Samburu County, Kenya. Africa Journal of Physical Sciences 2:33–49. Saunders, D. A., R. J. Hobbs, and P. R. Ehrlich. 1993. The reconstruction of fragmented ecosystems: global and regional perspectives. Surrey Beatty, Chipping Norton, Australia. y pp g Schooley, R. L., and J. A. Wiens. 2004. Movements of cactus bugs: patch transfers, matrix resistance, and edge permeability. Landscape Ecology 19:801–810. J y Laundré, J. W. 2010. Behavioral response races, predator‐prey shell games, ecology of fear, and patch use of pumas and their ungulate prey. Ecology 91:2995–3007. Shannon, G., L. S. Cordes, A. R. Hardy, L. M. Angeloni, and K. R. Crooks. 2014. Behavioral responses associated with a human‐mediated predator shelter. PLoS One 9:e94630. Laundré, J. W., L. Hernandez, and W. Ripple. 2010. The landscape of fear: ecological implications of being afraid. Open Ecology Journal 3:1–7. g p g p gy Lima, S. L. 1987. Vigilance while feeding and its relation to the risk of predation. Journal of Theoretical Biology 124:303–316. p Sih, A. 1987. Prey refuges and predator‐prey stability. Theoretical Popu- lation Biology 31:1–12. Lima, S. L., and P. A. Bednekoﬀ. 1999. Temporal variation in danger drives antipredator behavior: the predation risk allocation hypothesis. American Naturalist 153:649–659. Sih, A. 2013. Understanding variation in behavioural responses to human‐ induced rapid environmental change: a conceptual overview. Animal Behaviour 85:1077–1088. Mpanduji, D. LITERATURE CITED Journal of Applied Ecology 3:369–382. g p y Wittemyer, G., J. M. Northrup, J. Blanc, I. Douglas‐Hamilton, P. Omondi, and K. P. Burnham. 2014. Illegal killing for ivory drives global decline in African elephants. Proceedings of the National Academy of Sciences of the United States of America 111:13117–13121. pp gy Prevedello, J. A., G. Forero‐Medina, and M. V. Vieira. 2010. Movement behaviour within and beyond perceptual ranges in three small mammals: eﬀects of matrix type and body mass. Journal of Animal Ecology 79:1315–1323. Worton, B. J. 1987. A review of models of home range for animal movement. Ecological Modelling 38:277–298. R Development Core Team. 2012. R: a language and environment for sta- tistical computing. R Foundation for Statistical Computing, Vienna, Austria. Worton, B. J. 1989. Kernel methods for estimating the utilization dis- tribution in home‐range studies. Ecology 70:164–168. p g p g Raizman, E. A., H. B. Rasmussen, L. E. King, F. W. Ihwagi, and I. Douglas‐ Hamilton. 2013. Feasibility study on the spatial and temporal movement of Samburu's cattle and wildlife in Kenya using GPS radio‐tracking, remote sensing and GIS. Preventive Veterinary Medicine 111:76–80. Associate Editor: Jason P. Marshal. Associate Editor: Jason P. Marshal. Rensch, B., and R. Altevogt. 1955. Zahmung und Dressurleistungen in- discher Arbeitselefanten. Ethology 11:497–510. The Journal of Wildlife Management • SUPPORTING INFORMATION Roever, C. L., R. J. van Aarde, and K. Leggett. 2013. Functional con- nectivity within conservation networks: delineating corridors for African elephants. Biological Conservation 157:128–135. Additional supporting information may be found online in the Supporting Information section at the end of the article. 10
https://openalex.org/W2785644874	https://iptek.its.ac.id/index.php/jps/article/download/3348/2613	Malay	null	Peningkatan Akses Sanitasi Masyarakat melalui Promosi, Penguatan Wirausaha Sanitasi, dan Mikro Kredit Sanitasi	Aptech Proceeding International Seminar on Applied Technology, Science & Arts : Development of Green Agro-Industry to Support Human Life Sustainability	2,018	cc-by-sa	3,328	C I N I A The 2nd Conference on Innovation and Industrial Applications (CINIA 2016) The 2nd Conference on Innovation and Industrial Applications (CINIA 2016) Abstrak Abstrak PROPER merupakan Program Penilaian Peringkat Kinerja Perusahaan oleh Kementerian Negara Lingkungan Hidup dan Kehutanan yang merupakan salah satu upaya untuk mendorong dan mengevaluasi penaatan perusahaan dalam pengelolaan lingkungan hidup. Metode yang diterapkan pada PROPER dilakukan secara bertahap yaitu kategori ketaatan (peringkat hitam, merah dan biru) dan lebih dari ketaatan (peringkat hijau dan emas). Kategori lebih dari ketaatan memiliki salah satu kriteria penting yaitu kriteria inovasi, dimana industri yang telah memiliki inovasi dalam pengelolaan lingkungan akan berpotensi untuk mendapatkan peringkat hijau/emas. Efektivitas PROPER dalam memicu munculnya inovasi teknologi yang ramah lingkungan dapat dilihat dengan pendekatan jumlah peserta yang ikut, peringkat yang diperoleh dan jumlah inovasi yang dihasilkan dari PROPER. Jumlah peserta PROPER 2015 mencapai 2.076 peserta yang berarti mengalami peningkatan rata-rata pertahun sebesar 28% sejak sepuluh tahun terakhir. Persentase ketaatan peserta PROPER mengalami peningkatan hingga tahun 2015 mencapai 74%. Peringkat lebih dari ketaatan mencapai 120 peserta dengan jumlah total inovasi yang dihasilkan sebanyak 151 inovasi diberbagai bidang seperti efisiensi energi, penurunan beban pencemar udara, 3R limbah B3, 3R limbah padat Non B3, efisiensi air dan penurunan beban pencemar air, keanekaragaman hayati, dan pemberdayaan masyarakat. Hasil menunjukkan bahwa PROPER sudah sesuai sebagai alat evaluasi dan pemicu inovasi teknologi Industri. Kata Kunci : PROPER, Inovasi Teknologi, Industri Kata Kunci : PROPER, Inovasi Teknologi, Industri Penerapan PROPER sebagai Alat Pemicu Inovasi Teknologi Industri Berkelanjutan Febri Eko Wahyudianto dan Rachmat Boedisantoso 1Ilmu dan Teknologi Lingkungan, FST, Universitas Airlangga 2Teknik Lingkungan, FTSP, Institut Teknologi Sepuluh Nopember (ITS) Surabaya Jl. Arif Rahman Hakim, Sukolilo, Surabaya Indonesia febriekow@yahoo.com 1 PENDAHULUAN Industri adalah seluruh bentuk kegiatan ekonomi yang mengolah bahan baku dan/atau memanfaatkan sumberdaya industri sehingga menghasilkan barang yang mempunyai nilai tambah atau manfaat lebih tinggi termasuk jasa industri menurut UndangUndang Republik Indonesia Nomor 3 Tahun 2014 tentang Perindustrian. Suatu kegiatan industri memiliki dampak terhadap lingkungan baik dampak positif maupun dampak negatif. Dampak positif yang ditimbulkan oleh industri lebih cenderung pada peningkatan pendapatan masyarakat, kemudahan akses fasilitas kesehatan dan peningkatan kepemilikan fasilitas [1]. Terlepas dari dampah positif yang dihasilkan, suatu kegiatan industri juga dapat menimbulkan dampak negatif. Dampak negatif dari industri umumnya diakibatkan dari sisa proses produksi atau buangan dari proses produksi yang sering disebut sebagai limbah. Limbah yang dihasilkan dapat berupa limbah cair, gas, dan partikulat. Dampak negatif dari limbah yaitu dapat menyebabkan kekurangan oksigen pada badan air, kematian ikan, gangguan terhadap sistem ekologi, perubahan sifat-sifat air, hingga keracunan dan gangguan kesehatan pada manusia [2]. Dampak negatif yang dapat muncul akibat kegiatan industri telah dicegah atau diminimalisasi oleh pemerintah dengan adanya regulasi dalam pembuangan limbah yang dihasilkan oleh masing-masing industri. Regulasi yang ada mulai dari undangundang, peraturan pemerintah, peraturan mentri, peraturan gubernur, hingga peraturan bupati/walikota. Regulasi yang ada umumnya mengatur batasan-batasan yang harus dipenuhi oleh industri dalam pengelolaan lingkungan. Sehingga dalam pelaksanaan sebagian besar industri melakukan kegiatan pengelolaan lingkungan hanya untuk memenuhi regulasi yang ditetapkan, tidak untuk membentuk suatu budaya industri yang ramah lingkungan dan berkelanjutan. Oleh karena itu terdapat satu regulasi yang dikeluarkan oleh kementrian lingkungan hidup mengenai suatu alat untuk melakukan evaluasi terhadap kinerja industri dalam pengelolaan lingkungan yaitu PROPER. PROPER menurut Peraturan Menteri Lingkungan Hidup Nomor 3 Tahun 2014 adalah evaluasi ketaatan dan kinerja melebihi ketaatan pertanggung jawab usaha dan/atau kegiatan dibidang pengendalian pencemaran dan/atau kerusakan lingkungan hidup serta pengelolaan limbah bahan berbahaya dan beracun. PROPER sebagai alat evaluasi ketaatan kinerja suatu kegiatan industri dibedakan menjadi tiga peringkat yaitu: hitam, merah, dan biru. PROPER selain sebagai alat evaluasi kinerja ketaatan suatu kegiatan industri terhadap regulasi juga sebagai pemicu inovasi-inovasi teknologi industri yang ramah lingkungan dan berkelanjutan dimana PROPER sebagai alat evaluasi melebihi ketaatan kinerja suatu industri dibedakan 59 menjadi dua peringkat yaitu: hijau dan emas. Dalam upaya mengetahui efektifitas PROPER dalam fungsinya perlu dilakukan kajian terkait dengan jumlah peserta yang dinilai dalam PROPER, peringkat yang diperoleh, dan jumlah inovasi yang dihasilkan dalam PROPER menjadi dua peringkat yaitu: hijau dan emas. 1 PENDAHULUAN Dalam upaya mengetahui efektifitas PROPER dalam fungsinya perlu dilakukan kajian terkait dengan jumlah peserta yang dinilai dalam PROPER, peringkat yang diperoleh, dan jumlah inovasi yang dihasilkan dalam PROPER 2 TINJAUAN PUSTAKA PROPER merupakan suatu program kementrian lingkungan hidup yang telah dimullai pada tahun 2015. Awal penerapan proper berdasarkan Keputusan Mentri No. 35a Tahun 1995 tentang Program Penilaian Kinerja Perusahaan/ Kegiatan Usaha dalam Pengendalian Pencemaran dalam Lingkup Kegiatan PROKASIH (PROPER PROKASIH)[3]. PROPER dalam perjalanannya mengalami pasang surut dimana pada tahun 1998-2002 PROPER tidak dapat dilaksanakan karena krisis ekonomi dan perubahan struktur pada KLH sehingga baru mulai berjalan kembali pada tahun 2002 dengan integrasi 3 kriteria media (air, udara, dan B3)[4]. PROPER terus berkembang dengan beberapa kriteria yang ditetapkan dan terus diperbarui hingga pada tahun 2014 keluar peraturan terbaru terkait PROPER yaitu Peraturan Menteri Lingkungan Hidup Nomor 3 Tahun 2014 tentang Program Penilaian Peringkat Kinerja Perusahaan dalam Pengelolaan Lingkungan Hidup. Definisi PROPER berdasar Peraturan Menteri Lingkungan Hidup Nomor 3 Tahun 2014 adalah evaluasi ketaatan dan kinerja melebihi ketaatan pertanggung jawab usaha dan/atau kegiatan dibidang pengendalian pencemaran dan/atau kerusakan lingkungan hidup serta pengelolaan limbah bahan berbahaya dan beracun. Evaluasi ketaatan dan melebihi ketaatan terdiri dari beberapa aspek seperti tersaji pada tabel 1. Hasil dari evaluasi PROPER akan direpresentasikan dengan peringkat yang disimbolkan menjadi lima warna yang berbeda diantaranya yaitu hitam, merah, biru, hijau, dan emas. Masing-masing warna tersebut memiliki makna yang berbeda sesuai dengan tingkat ketaatan dan upaya melebihi ketaatan yang dilakukan oleh industri seperti tersaji pada tabel 2. 60 Tabel 1. Aspek Penilaian PROPER Berdasarkan Peraturan Menteri Lingkungan Hidup Nomor 3 Tahun 2014 Aspek Ketaatan Melebihi Ketaatan Pemenuhan ketentuan dalam izin lingkungan Penerapan sistem manajemen lingkungan Pengendalian pencemar air Pencapaian dibidang efisiensi energi Pengendalian pencemar udara Pengurangan dan pemanfaatan limbah bahan berbahaya dan beracun Pengelolaan limbah bahan berbahaya dan beracun Penerapan pronsip pengurangan, penggunaan kembali dan daur ulang Pengendalian kerusakan lingkungan hidup limbah padat non bahan berbahaya dan beracun Pengurangan pencemar udara dan emisi gas rumah kaca Pencapaian dibidang efisiensi air dan penurunan beban pencemar air Perlindungan keanekaragaman hayati Pemberdayaan Masyarakat Tabel 2. Peringkat PROPER Berdasarkan Peraturan Menteri Lingkungan Hidup Nomor 3 Tahun 2014 No. 4.1. Analisis Jumlah Peserta PROPER 4.1. Analisis Jumlah Peserta PROPER PROPER merupakan merupakan program pemerintah dibidang lingkungan untuk industri merupakan program sukarela (belum diwajibkan) untuk sebagian industri. Mengacu Peraturan Menteri Lingkungan Hidup Nomor 3 Tahun 2014 tentang Program Penilaian Peringkat Kinerja Perusahaan dalam Pengelolaan Lingkungan Hidup Pasal 3 menyebutkan bahwa Pelaksanaan PROPER dilakukan terhadap usaha dan/atau kegiatan wajib amdal atau UKL-UPL, yang: 1. Hasil Produksnya untuk tujuan ekspor; 2. Terdapat pada pasar bursa; 3. Menjadi perhatian masyarakat, baik dalam lingkup regional maupun nasional; dan/atau 4. Skala kegiatan signifikan untuk menimbulkan dampak terhadap lingkungan hidup j p y g p g p 4. Skala kegiatan signifikan untuk menimbulkan dampak terhadap lingkungan hidup Kriteria tersebut menunjukkan bahwa tidak semua kegiatan/usaha belum diwajibkan untuk mengikuti PROPER. Berdasarkan kondisi tersebut, salah satu parameter yang dapat digunakan sebagai indikator keberhasilan PROPER yaitu jumlah Peserta PROPER. Hasil data yang didapatkan dari laporan/publikasi tahunan kementrian lingkungan hidup mulai 2004-2015 dilakukan pengolahan menghasilkan grafik peningkatan peserta PROPER seperti pada gambar 1. Gambar 1. Grafik Peningkatan Peserta PROPER Tahun 2004-2015 Gambar 1. Grafik Peningkatan Peserta PROPER Tahun 2004-2015 Gambar 1. Grafik Peningkatan Peserta PROPER Tahun 2004-2015 Hasil tersebut menunjukkan bahwa setiap tahun pelaksanaan PROPER selama 10 tahun terakhir mengalami peningkatan yang cukup tinggi. Periode PROPER 2003-2004 menunjukkan peserta PROPER sebesar 251 perusahaan dari seluruh Indonesia. Angka tersebut terus mengalami peningkatan pada periode 2004-2005 menjadi 446 hingga pada periode 2014- 2015 menjadi 2.076 perusahaan yang mengikuti PROPER. Rata-rata peningkatan jumlah peserta selama 10 tahun terahir mencapai 28% pertahun. Peningkatan tersebut menunjukkan bahwa semakin bertambah banyak industri/kegiatan yang telah aware terhadap lingkungan. Hal tersebut menunjukkan bahwa pemerinta berhasil dalam menggunakan PROPER sebagai alat untuk meningkatkan kepedulian industri terhadap aspek lingkungan. Gambar 1. Grafik Peningkatan Peserta PROPER Tahun 2004-2015 Hasil tersebut menunjukkan bahwa setiap tahun pelaksanaan PROPER selama 10 tahun terakhir mengalami peningkatan yang cukup tinggi. Periode PROPER 2003-2004 menunjukkan peserta PROPER sebesar 251 perusahaan dari seluruh Indonesia. Angka tersebut terus mengalami peningkatan pada periode 2004-2005 menjadi 446 hingga pada periode 2014- 2015 menjadi 2.076 perusahaan yang mengikuti PROPER. Rata-rata peningkatan jumlah peserta selama 10 tahun terahir mencapai 28% pertahun. Peningkatan tersebut menunjukkan bahwa semakin bertambah banyak industri/kegiatan yang telah aware terhadap lingkungan. Hal tersebut menunjukkan bahwa pemerinta berhasil dalam menggunakan PROPER sebagai alat untuk meningkatkan kepedulian industri terhadap aspek lingkungan. 3 METODE PENELITIAN Metode yang digunakan yaitu metode studi literatur. Literatur didapatkan dari regulasi yang dikeluarkan pemerintah, publikasi pemerintah terkait PROPER, dan artikel ilmiah lainnya terkait dengan PROPER serta perkembangannya. Data yang diperoleh merupakan data sekunder dari literatur. Pengolahan data dilakukan dengan mengolah data menjadi grafik kecenderungan, persebaran peringkat sesuai dengan kriteria PROPER sepuluh tahun terakhir, dan perbandingan ketaatan dari seluruh peserta PROPER. 2 TINJAUAN PUSTAKA Aspek Peringkat Deskripsi 1 Ketaatan Hitam Penanggung jawab usaha dan/atau kegiatan yang sengaja melakukan perbuatan atau melakukan kelalaian yang mengakibatkan pencemaran dan/atau kerusakan lingkungan hidup serta pelanggaran terhadap Peraturan Perundang-Undangan atau tidak melaksanakan sanksi administrasi 2 Ketaatan Merah Penanggung jawab usaha dan/atau kegiatan yang upaya pengelolaan lingkungan hidupnya dilakukan tidak sesuai dengan persyaratan sebagaimana diatur dalam Peraturan Perundang-Undangan 3 Ketaatan Biru Penanggung jawab usaha dan/atau kegiatan yang upaya pengelolaan lingkungan hidupnya dilakukan sesuai dengan persyaratan sebagaimana diatur dalam Peraturan Perundang-Undangan 4 Melebihi Ketaatan Hijau Penanggung jawab usaha dan/atau kegiatan yang telah melakukan pengelolaan Tabel 1. Aspek Penilaian PROPER Berdasarkan Peraturan Menteri Lingkungan Hidup Nomor 3 Tahun 2014 Aspek Ketaatan Melebihi Ketaatan Pemenuhan ketentuan dalam izin lingkungan Penerapan sistem manajemen lingkungan Pengendalian pencemar air Pencapaian dibidang efisiensi energi Pengendalian pencemar udara Pengurangan dan pemanfaatan limbah bahan berbahaya dan beracun Pengelolaan limbah bahan berbahaya dan beracun Penerapan pronsip pengurangan, penggunaan kembali dan daur ulang Pengendalian kerusakan lingkungan hidup limbah padat non bahan berbahaya dan beracun Pengurangan pencemar udara dan emisi gas rumah kaca Pencapaian dibidang efisiensi air dan penurunan beban pencemar air Perlindungan keanekaragaman hayati Pemberdayaan Masyarakat Tabel 1. Aspek Penilaian PROPER Berdasarkan Peraturan Menteri Lingkungan Hidup Nomor 3 Tahun 2014 Tabel 2. Peringkat PROPER Berdasarkan Peraturan Menteri Lingkungan Hidup Nomor 3 Tahun 2014 No. Aspek Peringkat Deskripsi 1 Ketaatan Hitam Penanggung jawab usaha dan/atau kegiatan yang sengaja melakukan perbuatan atau melakukan kelalaian yang mengakibatkan pencemaran dan/atau kerusakan lingkungan hidup serta pelanggaran terhadap Peraturan Perundang-Undangan atau tidak melaksanakan sanksi administrasi 2 Ketaatan Merah Penanggung jawab usaha dan/atau kegiatan yang upaya pengelolaan lingkungan hidupnya dilakukan tidak sesuai dengan persyaratan sebagaimana diatur dalam Peraturan Perundang-Undangan 3 Ketaatan Biru Penanggung jawab usaha dan/atau kegiatan yang upaya pengelolaan lingkungan hidupnya dilakukan sesuai dengan persyaratan sebagaimana diatur dalam Peraturan Perundang-Undangan 4 Melebihi Ketaatan Hijau Penanggung jawab usaha dan/atau kegiatan yang telah melakukan pengelolaan 60 No. Aspek Peringkat Deskripsi lingkungan hidup melebihi ketaatan melalui pelaksanaan sistem manajemen lingkungan, pemanfaatan sumberdaya secara efisien dan melakukan upaya pemberdayaan masyarakat dengan baik. 5 Melebihi Ketaatan Emas Penanggung jawab usaha dan/atau kegiatan yang telah secara konsisten menunjukkan keunggulan lingkungan hidup dalam proses produksi dan/atau jasa, melaksanakan bisnis yang beretika dan bertanggung jawab terhadap masyarakat No. Aspek 4.2. Analisis Peringkat PROPER PROPER Memiliki tingkatan ketaatan atau melebihi ketaatan yang direpresentasikan menjadi lima warna peringkat yaitu hitam, merah, biru, hijau dan emas. Peringkat hitam, merah, dan biru menunjukkan tingkat ketaatan (compliance) sedangkan 61 peringkat hijau dan emas menunjukkan tingkat melebihi ketaatan (beyond compliance). Pringkat PROPER tersebut juga dapat digunakan sebagai salah satu keberhasilan pemerintah dalam melakukan pembinaan kepada industri terkait dengan kepedulian terhadap aspek lingkungan. Dengan mengetahui distribusi peringkat maka dapat diketahui proporsi industri/kegiatan yang tidak taat, taat, dan melebihi ketaatan. Data dan distribusi peringkat PROPER dari tahun 2004-2015 tersaji pada tabel 3 dan gambar 2. Tabel 3. Data Peringkat PROPER Tahun 2004-2015 [4]–[13] Tabel 3. Data Peringkat PROPER Tahun 2004-2015 [4]–[13] Tabel 3. Data Peringkat PROPER Tahun 2004-2015 [4]–[13] No. Tahun Peringkat Hitam Merah Biru Hijau Emas Jumlah Peserta 1 2003-2004 42 86 114 9 0 251 2 2004-2005 72 150 221 23 0 466 3 2006-2007 42 83 344 46 1 516 4 2008-2009 56 130 399 41 1 627 5 2009-2010 47 152 435 54 2 690 6 2010-2011 49 283 552 106 5 995 7 2011-2012 79 330 771 119 12 1.311 8 2012-2013 17 611 1.039 113 12 1.792 9 2013-2014 21 516 1.224 121 9 1.891 10 2014-2015 21 529 1.406 108 12 2.076 Gambar 2. Grafik Distribusi Peringkat PROPER Tahun 2004-2015 Gambar 2. Grafik Distribusi Peringkat PROPER Tahun 2004-2015 Data dan distribusi peringkat PROPER berdasarkan tabel 1 dan grafik distribusi pada gambar 2 menunjukkan bahwa sebagian besar peserta PROPER mendapatkan peringkat biru yang berarti industri tersebut taat dengan regulasi yang ditetapkan dimana industri telah melakukan pengelolaan lingkungan dan memenuhi standar yang berlaku. Peringkat merah menempati peringkat kedua terbanyak dimana industri pada peringkat merah telah melakukan pengelolaan lingkungan tetapi belum secara total (100%) memenuhi standar yang berlaku. Kondisi industri yang mendapatkan peringkat merah inilah yang harus dilakukan pembinaan untuk mendorong industri meningkatkan kualitas pengelolaan lingkungan sehingga dapat mencapai pengelolaan yang baik (100% memenuhi standar). Industri yang mendapatkan peringkat hitam atau industri yang belum melakukan pengelolaan lingkungan dari tahun ke tahun telah dapat ditekan jumlahnya sehingga memiliki trend menurun pada tahun 2015. Peringkat hijau dan emas (beyond compliance) merupakan peringkat yang menunjukkan industri tersebut tidak hanya sekedar memenuhi peraturan, tetapi industri pada peringkat ini memiliki komitmen lebih terhadap lingkungan. 4.2. Analisis Peringkat PROPER PROPER Jumlah industri yang mendapat peringkat hijau memiliki trend yang cukup baik dengan mengalami peningkatan dari tahun ketahun dengan persentase sebesar 5% (108 peserta) pada tahun 2015 sedangkan peringkat emas sebesar 1% (12 peserta) seperti pada gambar 3. Industri pada peringkat ini selalu berupaya untuk peduli terhadap lingkungan dengan menerapkan sistem manajemen lingkungan, efisiensi sumber daya, dan melakukan kegiatan pemberdayaan masyarakat (community development). Industri yang mendapat peringkat hijau dan emas juga telah menerapkan inovasi-inovasi dalam proses produksi maupun penunjang untuk mengefisiensikan kegiatan yang ada sehingga dapat mengurangi beban pencemar yang ada dan mendapatkan nilai lebih dari penurunan biaya produksi. 62 Gambar 3. Proporsi Peringkat PROPER Tahun 2015 Gambar 3. Proporsi Peringkat PROPER Tahun 2015 Gambar 3. Proporsi Peringkat PROPER Tahun 2015 Selain dari distribusi dan proporsi peringkat PROPER, analisis dapat dilakukan berdasarkan persentase ketaatan (peringkat biru, hijau, dan emas). Persentase dapat dihitung dengan menggunakan rumus (1) sebagai berikut: Selain dari distribusi dan proporsi peringkat PROPER, analisis dapat dilakukan berdasarkan persentase ketaatan (peringkat biru, hijau, dan emas). Persentase dapat dihitung dengan menggunakan rumus (1) sebagai berikut: Dengan menggunakan rumus tersebut dapat didapatkan data persentase ketaatan tiap tahun. Data yang didapatkan diplot dalam grafik mulai tahun 2004 hingga 2015 untuk mengetahui trend %ketaatan selama 10 tahun terakhir. Hasil analisis yang dilakukan tersaji pada gambar 4. Hasil menunjukkan bahwa %ketaatan 10 tahun terakhir berkisar antara 49%-76%. Semakin tinggi %ketaatan yang diraih mengindikasikan bahwa industri yang ada di Indonesia semakin taat dengan regulasi yang berlaku. Sebagai contoh pada tahun 2015 tingkat ketaatan mencapai 74% yang berarti bahwa 74% industri yang mengikuti PROPER pada tahun 2015 mendapatkan peringkat biru, hijau, dan emas sedangkan 26% peserta mendapat peringkat merah dan hitam atau tidak taat. Data tersebut dapat dijadikan sebagai acuan untuk mendorong 26% industri yang tidak taat untuk dilakukan pembinaan sehingga diharapkan pada tahun berikutnya menjadi industri yang taat terhadap regulasi yang berlaku. Gambar 4. Persentase Ketaatan (Peringkat Biru, Hijau, dan Emas) PROPER Tahun 2004-2015 Gambar 4. Persentase Ketaatan (Peringkat Biru, Hijau, dan Emas) PROPER Tahun 2004-2015 5. KESIMPULAN Penerapan PROPER di Indonesia menunjukkan efektifitas dalam mendorong industri dalam pengelolaan lingkungan dengan ditandai banyaknya peserta PROPER pada tahun 2015 sebesar 2.076 perusahaan dengan rata-rata peningkatan keikutsertaan sebesar 28% pertahun. Peringkat PROPER terdistribusi menjadi 1% peringkat hitam, 25% peringkat merah, 58% peringkat biru, 5% peringkat hijau, dan 1% peringkat emas. Persentase ketaatan peserta PROPER tahun 205 mencapai 74% dengan jumlah inovasi sebesar 151 inovasi dari 120 industri berperingkat hijau dan emas. Hasil tersebut menunjukkan bahwa PROPER telah sukses menjadi suatu alat evaluasi pengelolaan lingkungan dan pemicu inovasi teknologi industri berkelanjutan. 4.3. Inovasi Industri PROPER merupakan suatu alat (tools) evaluasi yang kompleks dimana dapat mengakomodir seluruh regulasi pemerintah dan aspek lingkungan lain yang tidak diatur dalam regulasi pemerintah. Hal tersebut yang membuat PROPER dapat meningkatkan kualitas pengelolaan lingkungan industri yang ada di Indonesia. Salah satu aspek penting dalam PROPER yakni inovasi. Inovasi merupakan aspek yang penting bagi industri untuk terus melakukan perubahan yang lebih baik disegala aspek baik produksi dan penunjang sehingga meningkatkan daya saing industri tersebut. Mengacu Peraturan Menteri Lingkungan Hidup Nomor 3 Tahun 2014 tentang Program Penilaian Peringkat Kinerja Perusahaan dalam Pengelolaan Lingkungan Hidup lampiran 5 kriteria Dokumen Ringkasan Kinerja Pengelolaan Lingkungan (DRKPL), PROPER mengapresiasi industri yang melakukan inovasi baik di proses produksi maupun penunjang dimana industri yang memiliki inovasi mendapatkan nilai lebih dalam penilaian PROPER. Inovasi yang dihasilkan oleh industri juga didorong untuk meningkatkan inovasi tersebut menjadi lebih tinggi berupa hak paten. PROPER juga mengapresiasi hak paten yang didapatkan oleh industri dengan memberikan nilai lebih tinggi berupa nilai adisionalitas dan nilai paten. PT Badak NGL merupakan salah satu industri yang menjadikan inovasi yang dikembangkan menjadi paten. Inovasi yang menjadi paten tersebut yaitu “T-Plug instaler” dalam aspek efisiensi energi dan penurunan beban emisi, “Draying oven” dalam aspek 3R limbah B3 dan 3R limbah padat non B3, dan “Biskuit dari tepung cangkang kepiting” dalam aspek pemberdayaan masyarakat [4]. Adanya inovasi yang mendapatkan paten tersebut diharapkan dapat memberikan kontribusi nyata baik bagi industri maupun stakeholder lain yang terkait.inovasi yang dihasilkan pada periode PROPER tahun 2015 tercatat sebanyak 151 inovasi yang telah dikembangkan oleh 120 industri di Indonesia yang mendapat peringkat melebihi ketaatan [13]. Inovasi-inovasi yang telah dihasilkan oleh berbagai industri seperti pada tabel 4. 63 Tabel 4. Inovasi yang telah Dilakukan oleh Industri [4], [13] No. Program Industri Keterangan 1. T-Plug instaler PT Badak NGL Inovasi - Paten 2. Draying oven PT Badak NGL Inovasi - Paten 3. Biskuit dari tepung cangkang kepiting PT Badak NGL Inovasi - Paten 4. Pertanian hidroponik Masyarakat Pangalengan Star Energy Geothermal Wayang Windu Ltd Inovasi 5. Teknologi listrik hybrid di area kantor PT Pertamina (Persero) TBBM Rewulu Inovasi 6. Eco-driving dengan software android PT Pertamina (Persero) TBBM Rewulu Inovasi 7. Alat sterilisasi jamur menggunakan uap geothermal PT Pertamina Geothermal Energy Inovasi 8. Smart Exhaust Air Conditioning (AC) PT Pertamina (Persero) TBBM Rewulu Inovasi 9. Portable turbine dan bladeless turbine PT Pertamina Geothermal Energy Inovasi 5. KESIMPULAN Tabel 4. Inovasi yang telah Dilakukan oleh Industri [4], [13] 6. DAFTAR PUSTAKA [1] S. Nawawi, I., Ruyadi, Y., Komariah, “Pengaruh Keberadaan Industri Terhadap Kondisi Sosial Ekonomi dan Budaya Masyarakat Desa Lagadar Kecamatan Marga Asih Kabupaten Bandung,” J. Sosietas, vol. 5, 2015. [1] S. Nawawi, I., Ruyadi, Y., Komariah, “Pengaruh Keberadaan Industri Terhadap Kondisi Sosial Ekonomi dan Budaya Masyarakat Desa Lagadar Kecamatan Marga Asih Kabupaten Bandung,” J. Sosietas, vol. 5, 2015. [2] Supraptini, “Pengaruh Limbah Industri Terhadap Lingkungan di Indonesia,” Media Litbang Kesehat., vol. 5, p. 1019, 2002. [2] Supraptini, “Pengaruh Limbah Industri Terhadap Lingkungan di Indonesia,” Media Litbang Kesehat., vol. 5, p. 1019, 2002. [3] A. Reliantoro, S., Arsyad, I., Prihantono, D. S., Karim, M. A., Kusrihadi, Gunawan, Y., Susanti, N., Parwitosari, A., Khoirunisa, I., Nuryandain, D., Busori, Dwiastoro, The Gold For Green: Bagaimana Penghargaan PROPER Emas Mendorong Lima Perusahaan Mencapai Inovasi, Pencipataan Nilai dan Keunggulan Lingkungan. Jakarta: Kementerian Lingkungan Hidup, 2012. [4] Kementerian Lingkungan Hidup, “Keputusan Menteri Lingkungan Hidup dan Kehutanan Republik Indonesia Nomor 180 Tahun 2014 Tentang Hasil Penilaian Peringkat Kinerja Perusahaan dalam Pengelolaan Lingkungan Hidup 2013-2014,” 2014. [5] Kementerian Lingkungan Hidup, “Press Briefing PROPER 2014,” 2004. [6] Kementerian Lingkungan Hidup, “Press Release Hasil Penilaian PROPER 2004-2005,” 2005. [7] Kementerian Lingkungan Hidup, “Press Release Hasil Penilaian PROPER 2006-2007,” 2007. [8] Kementerian Lingkungan Hidup, “Laporan Hasil Penilaian Peringkat Kinerja Perusahaan Dalam Pengelolaan Lingkungan Hidup PROPER 2008-2009,” 2009. [9] Kementerian Lingkungan Hidup, “Laporan Hasil Penilaian Peringkat Kinerja Perusahaan Dalam Pengelolaan Lingkungan Hidup 2010,” 2010. [10] Kementerian Lingkungan Hidup, “Laporan Hasil Penilaian Peringkat Kinerja Perusahaan Dalam Pengelolaan Lingkungan Hidup,” 2011. [11] Kementerian Lingkungan Hidup, Keputusan Menteri Lingkungan Hidup Republik Indonesia Nomor 273 Tahun 2012. 2012. [12] Kementerian Lingkungan Hidup, “PROPER 2013 Program Penilaian Peringkat Kinerja Perusahaan Menyemai Kebijakan, Melindungi Lingkungan,” 2013. [13] Kementerian Lingkungan Hidup, “PROPER Program Penilaian peringkat Kinerja Perusahaan Dalam Pengelolaan Lingkungan hidup 2015,” 2015. 64
https://openalex.org/W2138125130	https://europepmc.org/articles/pmc3871563?pdf=render	English	null	Effect of Statin Therapy in the Outcome of Bloodstream Infections Due to Staphylococcus aureus: A Prospective Cohort Study	PloS one	2,013	cc-by	6,739	Abstract This is an open-access article distributed under the terms of the Creative Commons Att unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. nding: The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: JRB has served as consultant and speaker for Pfizer, Roche, Astellas, Novartis and Merck. A.P. has been a consultant for Pfizer; has served as speaker for Wyeth, and Pfizer; and has received research support from Pfizer, Wyeth, and Novartis. All other authors had no conflicts of interest. This does not alter the authors’ adherence to all the PLOS ONE policies on sharing data and materials. * E-mail: jesusrb@us.es Effect of Statin Therapy in the Outcome of Bloodstream Infections Due to Staphylococcus aureus: A Prospective Cohort Study Luis E. Lo´ pez-Corte´s1, Juan Ga´lvez-Acebal1,2,4, Marı´a D. del Toro1,2,4, Carmen Velasco2,3, Marina de Cueto1,2, Francisco J. Caballero2, Miguel A. Muniain1,2,4, A´ lvaro Pascual1,2,3, Jesu´ s Rodrı´guez-Ban˜ o1,2,4* 1 Unidad Clı´nica de Enfermedades Infecciosas y Microbiologı´a, Hospital Universitario Virgen Macarena, Seville, Spain, 2 Spanish Network for Research in Infectious Diseases, Instituto de Salud Carlos III, Madrid, Spain, 3 Departamento de Microbiologı´a, Universidad de Sevilla, Seville, Spain, 4 Departamento de Medicina, Universidad de Sevilla, Seville, Spain December 2013 \| Volume 8 \| Issue 12 \| e82958 Abstract Introduction: Statins have pleiotropic effects that could influence the prevention and outcome of some infectious diseases. There is no information about their specific effect on Staphylococcus aureus bacteremia (SAB). Methods: A prospective cohort study including all SAB diagnosed in patients aged $18 years admitted to a 950-bed tertiary hospital from March 2008 to January 2011 was performed. The main outcome variable was 14-day mortality, and the secondary outcome variables were 30-day mortality, persistent bacteremia (PB) and presence of severe sepsis or septic shock at diagnosis of SAB. The effect of statin therapy at the onset of SAB was studied by multivariate logistic regression and Cox regression analysis, including a propensity score for statin therapy. Results: We included 160 episodes. Thirty-three patients (21.3%) were receiving statins at the onset of SAB. 14-day mortality was 21.3%. After adjustment for age, Charlson index, Pitt score, adequate management, and high risk source, statin therapy had a protective effect on 14-day mortality (adjusted OR = 0.08; 95% CI: 0.01–0.66; p = 0.02), and PB (OR = 0.89; 95% CI: 0.27– 1.00; p = 0.05) although the effect was not significant on 30-day mortality (OR = 0.35; 95% CI: 0.10–1.23; p = 0.10) or presentation with severe sepsis or septic shock (adjusted OR = 0.89; CI 95%: 0.27–2.94; p = 0.8). An effect on 30-day mortality could neither be demonstrated on Cox analysis (adjusted HR = 0.5; 95% CI: 0.19–1.29; p = 0.15). Conclusions: Statin treatment in patients with SAB was associated with lower early mortality and PB. Randomized studies are necessary to identify the role of statins in the treatment of patients with SAB. Citation: Lo´pez-Corte´s LE, Ga´lvez-Acebal J, del Toro MD, Velasco C, de Cueto M, et al. (2013) Effect of Statin Therapy in the Outcome of Bloodstream Infections Due to Staphylococcus aureus: A Prospective Cohort Study. PLoS ONE 8(12): e82958. doi:10.1371/journal.pone.0082958 Editor: Georg Ha¨cker, University Freiburg, Germany Received April 16, 2013; Accepted October 30, 2013; Published December 23, 2013 Copyright: 2013 Lo´pez-Corte´s et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Copyright: 2013 Lo´pez-Corte´s et al. Variables and definitions y Crude comparisons were performed by using the Chi squared or Fisher tests as appropriate for percentages, and the Mann- Whitney U test for continuous variables. Multivariate analyses were performed by using logistic regression. Variables with a p value,0.2 in the univariate analysis were included. Selection of variables was performed using a stepwise backward process, variables with a p value,0.2 were kept in the models. Effect modifications between the exposure of interest and other variables were investigated. A propensity score for receiving statins was added to the models. The propensity score (the probability of receiving statins) was calculated using a non-parsimonious multivariate logistic regression model, in which the outcome variable was use of statins; all variables considered as potentially influencing the prescription of statins, including gender, service of admission, type of acquisition, all underlying diseases, recent surgery, and indwelling implants, were included. The validity of the models was assessed by estimating goodness-of-fit to the data with the Hosmer-Lemeshow test, and its discrimination ability with the area under the receiver operating characteristic (ROC) curve. A survival analysis, using time until death (limited to 30 days) or censorship as outcome variable was also performed. Bivariate comparison of Kaplan Meier curves was performed by log rank test, and multivariate analysis was carried out by Cox regression modelling. The software used for the analysis was SPSS v17.0. The main outcome variable was all-cause 14-day mortality, and the secondary outcome variables were all-cause 30-day mortality, persistent bacteremia (PB; see definition below) and presence of severe sepsis or septic shock at diagnosis of SAB. The outcome variables and their definitions were decided and defined previ- ously. The reasons for choosing 14-day mortality as main outcome variable are explained in the Discussion. The main exposure variable was therapy with statins. For the purpose of the study, patients were considered to be receiving statins if they had been treated with any drug belonging to this family (including pravastatin, atorvastatin, simvastatin, fluvastatin, lovastatin and pitavastatin) for at least 30 days and were still taking them when SAB was diagnosed. Exposure to statins was assessed by reviewing the charts and by directly interviewing the patients or their relatives. Other exposure variables considered included demographics, types and severity of underlying conditions, type of acquisition of SAB, source of infection, severity of systemic inflammatory response syndrome (SIRS) at presentation, antimi- crobial therapy, and support therapy. Introduction bloodstream infections (BSI), the use of statins was associated with lower mortality in a recent meta-analysis [10]. Statins act as selective and competitive inhibitors of HMG-CoA reductase. This enzyme converts 3-hydroxy-3-methyl glutaril- coenzime A into mevalonate, which is a precursor of sterols. The hypolipidemic action of statins is due to the inhibition of synthesis of cholesterol in the liver and the increase of the number of liver receptors for LDL cholesterol. In addition to this activity, other so- called pleiotropic effects have been described for statins, including anti-inflammatory, inmunomodulatory, antioxidant and anticoag- ulant activities [1,2]. Multiple studies have evaluated the effect of statins in the prevention and prognosis of diverse infectious diseases. Most of these studies ascribed a beneficial effect to exposition to statins [3–6]. However, other studies did not found any benefit or even showed harmful effects [7–9]. In patients with Staphylococcus aureus is one of the most important etiological agents of both nosocomial and community-onset BSIs [11,12]; S. aureus bacteremia (SAB) is associated with important morbidity and mortality [13]. To our knowledge, there are no studies specifically investigating the effect of statins in the outcome of patients with SAB, in spite of the fact that in vitro and animal model data suggest that they may play a role [14,15]. It has been hypothesized that this effect could be linked to the inhibition of host cell invasion. This inhibition would facilitate the extracellular activity of antibiotics, avoid intracellular persistence, and so reduce the hematogenous spread of S. aureus [14]. The aim of our study is, therefore, to assess the effect of the treatment with statins on the evolution and prognosis of SAB. December 2013 \| Volume 8 \| Issue 12 \| e82958 December 2013 \| Volume 8 \| Issue 12 \| e82958 1 PLOS ONE \| www.plosone.org Statins & S. aureus Bacteremia Setting, patients, and study design g p y g A prospective cohort study including all episodes of SAB diagnosed in patients aged $18 years admitted to Hospital Universitario Virgen Macarena (a 950-bed tertiary hospital located in Seville, Spain) from March 2008 to January 2011 was carried out. A specific ‘‘bacteremia program’’, carried out by clinical microbiologists and infectious diseases physicians is active in our hospital since 2003; the program includes early personalized reporting of blood culture results and unsolicited management advice for all patients with BSI [16]. Patients with SAB were detected through daily review of microbiology reports. SAB was defined as the isolation of S. aureus from at least one blood culture in patients with symptoms or signs of infection. Only one episode (the first) per patient was included in the analysis. Patients were followed for 90 days after the diagnosis of SAB. An experienced team of clinical microbiologists and infectious diseases doctors followed all included patients daily during their admission; the evolution of surviving patients discharged before day 30 was assessed by outpatient clinic visits and/or phone calls. The data were collected by one investigator (LELC) using a structured questionnaire and reviewed by a senior investigator (JRB). This analysis was reported following the STROBE recommendations [17]. The study was approved by the Ethic Committee of Hospital Universitario Virgen Macarena which waived the need to obtained written informed consent from patients because of the observational nature of the study. Microbiological Studies Two or three sets of two blood samples, separated by 20– 30 minutes and containing 15 mL of blood each, were drawn in patients who presented fever $38uC or when bacteremia was suspected because of clinical signs or symptoms. Blood samples were processed by API Staph system (BioMerieux, NC, USA). The identification of the isolates was assessed by API Staph (BioMer- ieux, NC, USA), and antimicrobial susceptibility was studied by WIDER (Soria Melguizo, Madrid, Spain) and microdilution according to the recommendations by the Clinical and Laboratory Standards Institute (CLSI) [29]. Methods according to standard definitions [23]. Clinical management was considered adequate when it fulfilled all the following criteria: appropriate antimicrobial therapy was administered (at least one active drug was administered during the first 24 hours, and active drugs were administered as definitive therapy); fluid resuscitation was administered in patients with severe sepsis or septic shock according to recommendations [24]; respiratory support was provided according to recommendations [24]; duration of treatment was according to complexity of infection [16,25,26]; and the source of infection was removed or drained whenever feasible [16,27,28]; otherwise, it was considered non-adequate. Persistent SAB was defined as the isolation of S. aureus in blood cultures obtained from peripheral veins for $3 days despite appropriate antimicrobial therapy according to susceptibility testing. Variables and definitions We used the Charlson comorbidity index to measure the severity of chronic underlying conditions [18]; this index has been validated as a mortality predictor in SAB [19]. The acute severity of the illness was retrospectively assessed on the day before the diagnosis of SAB using the Pitt bacteremia score, which has also been validated [20]. SAB were considered as hospital-acquired if occurring after 48 hours of hospital stay, and as community-onset in all other cases. The source of infection was established by the agreement between 2 investigators according to clinical and microbiological criteria. Sources of SAB associated with high mortality in previous studies were classified as high-risk sources; these included endocarditis, unknown source, endovascular infections other than catheter-related, central nervous system infections, and respiratory tract infections [21,22]. The severity of SIRS was classified as sepsis, severe sepsis or septic shock Results Because the source of bacteremia was an important predictor of outcome (Table 2) and the distributions of SAB sources were different among patient exposed and not exposed to statins, we performed different analysis to control for this confounder. To do so, we performed 4 multivariate models for which the variable source was defined in different ways (Table 3). Statin therapy was independently associated with statin therapy in all of them. All models showed a high prediction ability; as an mellitus and catheter-related bacteremia were more frequent among patients who were receiving statins. The treatment and outcomes are also shown in Table 1. Mortality at day 14 and 30 were 21.3% and 28.7%, respectively; crude comparison showed lower 14-day mortality in the statins group; the differences for PB and 30-day mortality were in the limit of significance. The univariate analysis of the association of exposure to categorical variables and 14-day mortality is shown in Table 2. Apart from statin use, other variables showing a crude association with 14-day mortality were the source of bacteremia, non-adequate clinical management, and presentation with severe sepsis and septic shock. As regards the continuous variables [measured as median value (interquartile range)], age and Pitt score were significantly higher among patients who died [age in years: 75 (63–83) vs. 66 (58– 75), p = 0.01; Pitt score: 3 (2–4) vs. 1 (0–2), p,0.001, respectively], but Charlson index was not [2 (2–4) vs. 2 (1–3), p = 0.09]. We calculated a propensity score for receiving statins. The model showed a p value of 0.42 for the Hosmer-Lemeshow Table 1. Features of patients with Staphylococcus aureus bacteremia. Results We included 160 episodes of SAB during the thirty five months of the study period. Thirty-three patients (21.3%) were receiving statins at the onset of SAB (all of them had been receiving these drugs for at least one month): 17 (5.15% of those receiving statins) received atorvastatin, 14 (42.4%) simvastatin, and 2 (6%) pravastatin. The epidemiological and clinical features of patients receiving and not receiving statins are shown in Table 1. Diabetes PLOS ONE \| www.plosone.org December 2013 \| Volume 8 \| Issue 12 \| e82958 December 2013 \| Volume 8 \| Issue 12 \| e82958 2 Statins & S. aureus Bacteremia goodness-of-fit test, and an area under the ROC curve of 0.877, showing good predictive ability for receiving statins. In order to control for confounding in the association between statins and 14- day mortality, we performed several multivariate analysis by logistic regression. The variables introduced were age, Charlson index, Pitt score, high risk source, adequate clinical management, and statin use, plus the propensity score. Neither presentation with severe sepsis/septic shock nor PB was introduced in the model because these variables are in the pathogenic pathway between development of SAB and death, and exposure to statins preceded them. Because the source of bacteremia was an important predictor of outcome (Table 2) and the distributions of SAB sources were different among patient exposed and not exposed to statins, we performed different analysis to control for this confounder. To do so, we performed 4 multivariate models for which the variable source was defined in different ways (Table 3). Statin therapy was independently associated with statin therapy in all of them. All models showed a high prediction ability; as an goodness-of-fit test, and an area under the ROC curve of 0.877, showing good predictive ability for receiving statins. In order to control for confounding in the association between statins and 14- day mortality, we performed several multivariate analysis by logistic regression. The variables introduced were age, Charlson index, Pitt score, high risk source, adequate clinical management, and statin use, plus the propensity score. Neither presentation with severe sepsis/septic shock nor PB was introduced in the model because these variables are in the pathogenic pathway between development of SAB and death, and exposure to statins preceded them. Patients who died within 96 hours are excluded. #High risk source: endocarditis, unknown source, endovascular infections other than catheter-related, central nervous system infections, and respiratory tract infections. Data are expressed as number of cases (percentage) except where specified. doi:10.1371/journal.pone.0082958.t001 Results Variable Subcategory All patients (n = 160) Statin use (n = 33) No statin use (n = 127) p-value Median age (interquartile range), years 68 (59–77) 67 (63–75) 68 (58–77) 0.4 Male sex 100 (63) 23 (69.7) 77 (60.6) 0.3 Comorbidities Chronic heart failure 28 (17.5) 3 (9.1) 16 (12.6) 0.5 Chronic pulmonary disease 34 (21.3) 4 (12.1) 24 (18.9) 0.3 Malignancy 63 (39.4) 6 (18.2) 28 (22) 0.6 Diabetes mellitus 19 (11.9) 20 (60.6) 43 (33.9) 0.005 Chronic heart failure 19 (11.9) 3 (9.1) 16 (12.6) 0.5 Hemodialysis 12 (7.5) 4 (12.1) 8 (6.3) 0.2 Intravenous drug abuse 5 (3.1) 0 (0) 5 (3.9) 0.2 Organ transplantation 1 (0.3) 0 (0) 1 (0.8) 0.6 Median Charlson index 2 2 3 0.1 Median Pitt score 2 2 2 0.8 Hospital-acquired infection 93 (58.1) 22 (66.7) 71 (55.9) 0.2 Source of BSI Vascular catheter 69 (43.1) 23 (69.7) 46 (36.2) 0.001 Unknown source 30 (18.8) 4 (12.1) 26 (20.5) 0.2 Respiratory tract 11 (6.9) 0 (0) 11 (8.7) 0.08 Skin and/or soft tissue 24 (15) 6 (18.2) 18 (14.2) 0.5 Endocarditis 4 (2.5) 0 (0) 4 (3.1) 0.9 Others 22 (13.7) 0 (0) 22 (100) 0.1 High risk source# 48 (42.8) 4 (12.1) 44 (34.6) 0.01 Methicillin-resistant S. aureus 26 (16.3) 4 (12.1) 22 (17.3) 0.4 ICU admission 42 (26.3) 10 (30.3) 32 (25.2) 0.5 Adequate management 145 (90.6) 30 (90.9) 115 (90.6) 0.9 Outcome variables 14-day mortality 34 (21.3) 2 (6) 32 (25.2) 0.01 30-day mortality 46 (28.7) 5 (15.2) 42 (32.3) 0.05 Persistent bacteremia* 34/150 (22.7) 3/31 (9.7) 31/119 (26.1) 0.05 Severe sepsis or septic shock 38 (23.7) 6 (18.2) 33 (25.2) 0.5 Patients who died within 96 hours are excluded. #High risk source: endocarditis, unknown source, endovascular infections other than catheter-related, central nervous system infections, and respiratory tract infections. Data are expressed as number of cases (percentage) except where specified. doi:10.1371/journal.pone.0082958.t001 ble 1. Features of patients with Staphylococcus aureus bacteremia. Table 1. Features of patients with Staphylococcus aureus bacteremia. Patients who died within 96 hours are excluded. #High risk source: endocarditis, unknown source, endovascular infections other than catheter-related, central nervous system infections, and respiratory tract infections. Data are expressed as number of cases (percentage) except where specified. doi:10.1371/journal.pone.0082958.t001 December 2013 \| Volume 8 \| Issue 12 \| e82958 PLOS ONE \| www.plosone.org 3 Statins & S. aureus Bacteremia Table 2. Results Univariate analysis of 14-day mortality among patients with S. aureus bacteremia according to exposure to different categorical variables. Variable Subcategory No. dead/No. exposed(percentage) RR (95% CI) P value Gender Male 19/100 (19) Ref. Female 15/60 (25) 0.76 (0.41–1.38) 0.3 Source Catheter 6/69 (8.7) Ref. Ref. Respiratory 8/11 (72.7) 8.36 (3.58–19.48) ,0.001 Unknown 9/30 (30) 3.45 (1.34–8.83) 0.009 Skin and/or soft tissue 6/24 (25) 2.87 (1.02–8.06) 0.04 Endocarditis 2/4 (50) 5.75 (1.65–19.92) 0.01 Others 2/12 (16.7) 1.91 (0.43–8.44) 0.86 Type of acquisition Community-onset 14/67 (20.9) Ref. Nosocomial 20/93 (21.5) 0.97 (0.53–1.78) 0.9 Persistent bacteremia" Yes 7/34 (20.6) 1.41 (0.53–3.72) 0.4 No 18/116 (15.5) Ref. Severe sepsis or septic shock Yes 21/35 (60) 12.92 (5.32–31.38) ,0.001 No 13/125 (10.4) Ref. Susceptibility MRSA 5/26 (19.1) Ref. MSSA 29/134 (21.6) 1.12 (0.48–2.63) 0.7 Empirical treatment Appropriate 24/122 (19.7) Ref. Non-appropriate 10/38 (26.3) 1.33 (0.70–2.54) 0.3 Clinical management Adequate 25/145 (17.2) 0.24 (0.06–0.94) ,0.001 Non-adequate 9/15 (60) Ref. Statins use Yes 2/33 (6.1) 0.24 (0.06–0.94) 0.01 No 32/127 (25.2) Ref. "Considered only among surviving patients at 96 hours. MRSA: Methicillin-resistant S. aureus. MSSA: Methicillin-susceptible S. aureus. doi:10.1371/journal.pone.0082958.t002 Table 2. Univariate analysis of 14-day mortality among patients with S. aureus bacteremia according to exposure to different categorical variables. Table 2. Univariate analysis of 14-day mortality among patients with S. aureus bacteremia according to exposure to different categorical variables. "Considered only among surviving patients at 96 hours. MRSA: Methicillin-resistant S. aureus. MSSA: Methicillin-susceptible S. aureus. doi:10.1371/journal.pone.0082958.t002 statin use with 30-day mortality (adjusted HR = 0.5; 95% CI: 0.19–1.29; p = 0.15). example, model 1, in which source was classified as high or low- risk, showed a p value of 0.63 for the Hosmer-Lemeshow goodness-of-fit test, and an area under the ROC curve of 0.90. Additionally, we performed another model (not shown) in which source was defined as a polichotomous variable (including catheter, which was taken as reference, respiratory tract, unknown, and others); in this model, statin therapy remained associated with lower 14-day mortality (adjusted OR = 0.08; CI 95%: 0.01–0.83; p = 0.03). Finally, because the models might be overfitted due to the high number of variables introduced in relation to the number of patients with the event, we also performed several models in which the variables were added to statins in a forward approach to a limit of 3. Statin therapy showed a protective effect for 14-day mortality in all of them (data not shown). Results We also performed a survival analysis until day 30. The Kaplan Meier curves are shown in figure 1; the log rank value for mortality among patients who received and not received statins was 0.06. The Cox regression model including the propensity score, did not show association of example, model 1, in which source was classified as high or low- risk, showed a p value of 0.63 for the Hosmer-Lemeshow goodness-of-fit test, and an area under the ROC curve of 0.90. Additionally, we performed another model (not shown) in which source was defined as a polichotomous variable (including catheter, which was taken as reference, respiratory tract, unknown, and others); in this model, statin therapy remained associated with lower 14-day mortality (adjusted OR = 0.08; CI 95%: 0.01–0.83; p = 0.03). Finally, because the models might be overfitted due to the high number of variables introduced in relation to the number of patients with the event, we also performed several models in which the variables were added to statins in a forward approach to a limit of 3. Statin therapy showed a protective effect for 14-day mortality in all of them (data not shown). We also performed a survival analysis until day 30. The Kaplan Meier curves are shown in figure 1; the log rank value for mortality among patients who received and not received statins was 0.06. The Cox regression model including the propensity score, did not show association of We also performed multivariate analyses for secondary outcome variables. An independent relationship between statin use and these variables could not be demonstrated using a similar strategy, neither for 30-day mortality (adjusted OR = 0.35; 95% CI: 0.10– 1.22; p = 0.10) nor for presentation with severe sepsis or septic shock (adjusted OR = 0.89; CI 95%: 0.27–2.94; p = 0.8). However, statins use showed a protective effect for PB (adjusted OR = 0.89; 95% CI: 0.27–1.00; p = 0.05). In the subgroup analyses, statin therapy showed the following association to 14-day mortality: among nosocomial SAB, OR = 0.08; CI 95%: 0.005–1.27; among community-onset SAB, OR = 0.81; CI 95%: 0.21–3.27; among patients aged #60 years, OR = 0.43; CI 95%: 0.02–6.95; among those older than 60 years, OR = 0.25; CI 95%: 0.006–0.98; among patients with a Charlson index #1, OR = 0.38; CI 95%: 0.02–6.23; finally, among those with a Charlson index $2, OR = 0.25; CI 95%: 0.06–0.98. Results December 2013 \| Volume 8 \| Issue 12 \| e82958 PLOS ONE \| www.plosone.org 4 Statins & S. aureus Bacteremia Table 3. Different multivariate models for 14-day mortality. Model1 Model 2 Model 3 Model 4 OR (95% CI) P value OR (95% CI) P value OR (95% CI) P value OR (95% CI) P value Statins use 0.08 (0.01–0.67) 0.02 0.09 (0.01–0.70) 0.02 0.08 (0.01–0.74) 0.03 0.07 (0.01–0.49) 0.008 Age (per unit) 1.05 (1.01–1.09) 0.007 1.05 (1.01–1.08) 0.02 1.06 (1.02–1.10) 0.006 1.05 (1.01–1.09) 0.01 Charlson index (per unit) 1.26 (0.98–1.62) 0.06 1.20 (0.93–1.55) 0.16 1.20 (0.93–1.53) 0.16 1.21 (0.96–1.55) 0.11 Pitt score (per unit) 1.79 (1.31–2.46) ,0.001 1.82 (1.34–2.46) ,0.001 1.81 (1.33–2.49) ,0.001 1.77 (1.34–2.41) ,0.001 Source according to below definition* 5.72 (2.07–15.79) 0.001 12.86 (2.54–65.21) 0.002 0.13 (0.04–0.46) 0.002 1.91 (0.63–5.79) 0.26 Adequate management 0.09 (0.02–0.42) 0.002 0.09 (0.02–0.38) 0.001 0.06 (0.01–0.29) 0.001 0.12 (0.03–0.46) 0.002 Source was defined as a dichotomous variable as follows. Model 1: high risk source (endocarditis, unknown source, endovascular infections other than catheter-related, central nervous system infections, and respiratory tract infections) vs. low risk source. Model 2: respiratory source vs. other sources. Model 3: catheter-related source vs. other sources. Model 4: Unknown source vs. known source. The propensity score for including statins was included in all of them. The definition for the variable source was different in each model (see footnote). doi:10.1371/journal.pone.0082958.t003 Source was defined as a dichotomous variable as follows. Model 1: high risk source (endocarditis, unknown source, endovascular infections other than catheter-related central nervous system infections, and respiratory tract infections) vs. low risk source. Model 2: respiratory source vs. other sources. Model 3: catheter-related source vs other sources. Model 4: Unknown source vs. known source. The propensity score for including statins was included in all of them. The definition for the variable source was different in each model (see footnote). doi:10.1371/journal.pone.0082958.t003 central nervous system infections, and respiratory tract infections) vs. low risk source. Model 2: respiratory source vs. other sources. Model 3: catheter-related source vs other sources. Model 4: Unknown source vs. known source. The propensity score for including statins was included in all of them. The definition for the variable source was different in each model (see footnote). doi:10.1371/journal.pone.0082958.t003 Figure 1. Kaplan-Meier survival curves for patients receiving and not receiving statins. doi:10.1371/journal.pone.0082958.g001 PLOS ONE \| www.plosone.org 5 December 2013 \| Volume 8 \| Issue 12 \| e82958 Figure 1. Discussion We found an association between statin therapy and reduced risk of early death among patients with SAB. Such association could not be demonstrated for 30-day mortality. We chose 14-day mortality as our main outcome variable because our hypothesis was that statin therapy may have an impact in early, acute infection-related mortality, and we considered this time frame would adequately represent the biological plausible window of effect, as has been recommended [30]. Since mortality in patients with SAB may be due either to a direct effect of the infection (because of septic shock, respiratory distress, disseminated intravascular coagulation) or to later complications related to endocarditis or metastatic lesions, and also to decompensation of underlying conditions or other non- infection-related caused, early mortality is more probably related to the direct impact of the infection rather than to late metastatic complications or unrelated issues. Patients receiving statins were similar to those not exposed to these drugs in demographic features and underlying conditions except diabetes mellitus. This was expected because diabetes mellitus is frequently associated with dyslipidemia, which would increase the probability of receiving statins due to a higher cardiovascular risk. Patients with statins had a higher frequency of catheter-related SAB and a trend towards lower frequency of respiratory tract infection-related SAB, which would clearly reduce their baseline risk of mortality, since catheter-related BSI is usually associated with lower mortality while the opposite is seen with respiratory tract infections [13]. These and other potential confounders were controlled by performing multivariate analyses in which a propensity score for receiving statins was also used. Our study must be interpreted considering its limitations. Since it is not a randomized study, confounding due to unmeasured variables may not have been controlled. However, we made a great effort to control the potential confounding factors, including a propensity analysis and performing several multivariate models. The healthy user effect (patients receiving statins may be subject to better medical follow-up and a healthier lifestyle) needs to be taken into account [44]. Although our propensity score analysis could not discard this effect, we do not think it is a relevant issue since the effect was also significant in patients with nosocomial infections. Our study does not allow determining whether the effect of statins is immediate or it requires a minimum treatment time to be established, as it is the case of the hypolipidemic effect. 3. Dobesh PP, Klepser DG, McGuire TR, Morgan CW, Olsen KM (2009) Reduction in mortality associated with statin therapy in patients with severe sepsis. Pharmacotherapy 29:621–30. Conclusion In conclusion, in this prospective observational study, treatment with statins showed a protective effect for 14-day mortality and PB among patients with BSA. Further studies are needed to determine the role of statins in the treatment of patients with BSA. The pathophysiological basis for the potential association between statin treatment and lower mortality in some infectious diseases is still unknown. In our study, patients under statin therapy developed severe sepsis and septic shock with less frequency, but the difference was not statistically significant, and multivariate analysis could not show a protective effect of statins for developing severe systemic inflammatory response syndrome Results Kaplan-Meier survival curves for patients receiving and not receiving statins. doi:10.1371/journal.pone.0082958.g001 December 2013 \| Volume 8 \| Issue 12 \| e82958 December 2013 \| Volume 8 \| Issue 12 \| e82958 PLOS ONE \| www.plosone.org 5 Statins & S. aureus Bacteremia (SIRS). This is in contrast to previous observations in patients with infections due to other microorganisms [36,37]. Beyond the potential effect on the development of severe SIRS, we found that patients receiving statins had a lower frequency of PB. PB has been associated with increased mortality in previous studies of SAB [30,38,39]. PB has been shown to be related to some hosts’ features as well as to clinical management [39,40]. However, the molecular mechanisms involved in PB are poorly known. It is known that secretion of platelet microbicidal proteins (PMPs) and intracellular killing by neutrophil-associated oxidative and non- oxidative mechanisms are involved in the initial phases of vascular invasion leading to bacteremia [27,41,42]. On the other hand, Xiong et al. [34] showed that MRSA causing PB tend to be more resistant to key innate cationic host defense molecules from both neutrophils (e.g., hNP-1) and platelets (e.g., tPMP-1), have more enhanced membrane fluidity, and substantially greater adhesion to fibronectin, fibrinogen, and endothelial cells. Also, isolates causing PB adhere better to host cells (i.e., endothelium) and matrix ligands relevant to endovascular pathogenesis (i.e., fibrinogen and fibronectin) [34]. Simvastatin has been shown to inhibit cell invasion by S. aureus in vitro [14] and to interfere with exotoxine- induced leukocyte-endothelial cell interactions [15]. Whether the lower rate of PB in patients treated with statins, if confirmed, might be related to any of these effects or others [43], would merit being investigated. Discussion Finally, since most of the patients were treated with atorvastatin or simvastatin (91%), we cannot assess differences in the impact of specific statins. The fact that we could not find an effect on 30-day mortality may be due to a lack of longer effect of statins, or to the limited statistical power of the study. Finally, we did not study the molecular features of the S. aureus isolates. The association of statin treatment with lower mortality in different types of infection has been previously described, but there are scarce data about S. aureus. A recent meta-analysis assessed the effect of statins on all-cause mortality among patients with severe infections [10]; the pooled analysis showed an overall survival benefit in patients receiving statins. Bacteremia-related mortality was reported in 4 of these studies; a meta-analysis of this subgroup of studies also showed a lower rate of bacteremia-related mortality in patients receiving statins (OR = 0.33; 95% CI: 0.09–0.75); specific data for SAB were not provided. In a cohort of 388 patients with BSI of diverse etiologies, statins use was associated with a 17% decrease in the attributable mortality [31]; however, in the sub-analysis performed in SAB patients the decrease in mortality was not statistically significant (no multivariate analysis in patients with SAB was performed). To our knowledge, there are no other studies specifically addressing the impact of statins on mortality among patients with SAB, although there are some features of SAB that worth a specific investigation: SAB is more frequently persistent and complicated than bacteremia due to other microorganisms [32–34], and clinical management has been consistently shown to influence the outcome [23–26,35]. Author Contributions Conceived and designed the experiments: JRB AP MAM LELC. Performed the experiments: LELC JGA MDT. Analyzed the data: LELC JGA MDT MAM JRB. Contributed reagents/materials/analysis tools: MAM AP LELC JRB. Wrote the manuscript: LELC JRB. Process and interpretation of microbiological isolates: CV MC FJC. Conceived and designed the experiments: JRB AP MAM LELC. Performed the experiments: LELC JGA MDT. Analyzed the data: LELC JGA MDT MAM JRB. Contributed reagents/materials/analysis tools: MAM AP LELC JRB. Wrote the manuscript: LELC JRB. Process and interpretation of microbiological isolates: CV MC FJC. 1. Koh KK (2000) Effects of statins on vascular wall: vasomotor function, inflammation, and plaque stability. Cardiovasc Res 47:648–57. 2. Bellosta S, Bernini F, Ferri N, Quarato P, Canavesi M, et al. (1998) Direct vascular effects of HMGCoA reductase inhibitors. Atherosclerosis 137:S101–9. 1. Koh KK (2000) Effects of statins on vascular wall: vasomotor function, inflammation, and plaque stability. Cardiovasc Res 47:648–57. 2. Bellosta S, Bernini F, Ferri N, Quarato P, Canavesi M, et al. (1998) Direct vascular effects of HMGCoA reductase inhibitors. Atherosclerosis 137:S101–9. References Janda S, Young A, Fitzgerald JM, Etminan M, Swiston J (2010) The effect of statins on mortality from severe infections and sepsis: a systematic review and meta-analysis. J Crit Care 25: 656.e7–22. 30. McGregor JC, Rich SE, Harris AD, Perencevich EN, Osih R, et al. (2007) A systematic review of the methods used to assess the association between appropriate antibiotic therapy and mortality in bacteremic patients. Clin Infect Dis 45:329–37. 11. Wisplinghoff H, Bischoff T, Tallent SM, Seifert H, Wenzel RP, et al. (2004) Nosocomial bloodstream infections in US hospitals: analysis of 24,179 cases from a prospective nationwide surveillance study. Clin Infect Dis 39:309–17. 31. Liappis AP, Kan VL, Rochester CG, Simon GL (2001) The effect of statins on mortality in patients with bacteremia. Clin Infect Dis 33:1352–7. 12. Rodrı´guez-Ban˜o J, Lo´pez-Prieto MD, Portillo MM, Retamar P, Natera C, et al. (2010) Epidemiology and clinical features of community-acquired, healthcare- associated and nosocomial bloodstream infections in tertiary-care and community hospitals. Clin Microbiol Infect 16:1408–1413. 32. Khatib R, Johnson LB, Sharma M, Fakih MG, Ganga R, et al. (2009) Persistent Staphylococcus aureus bacteremia: incidence and outcome trends over time. Scand J Infect Dis 41:4–9. community hospitals. Clin Microbiol Infect 16:1408–1413. y p 13. Retamar P, Portillo MM, Lo´pez-Prieto MD, Rodrı´guez-Lo´pez F, de Cueto M, et al. (2012) Impact of inadequate empirical therapy on the mortality of patients with bloodstream infections: a propensity score-based analysis. Antimicrob Agents Chemother 56:472–8. 33. Johnson JR (2008) Persistent Staphylococcus aureus bacteremia. Arch Intern Med 168:772–3. 34. Xiong YQ, Fowler VG, Yeaman MR, Perdreau-Remington F, Kreiswirth BN, et al. (2009) Phenotypic and genotypic characteristics of persistent methicillin- resistant Staphylococcus aureus bacteremia in vitro and in an experimental endocarditis model. J Infect Dis 199:201–8. g 14. Horn MP, Knecht SM, Rushing FL, Birdsong J, Siddall CP, et al. (2008) Simvastatin inhibits Staphylococcus aureus host cell invasion through modulation of isoprenoid intermediates. J Pharmacol Exp Ther 326: 135–143, 35. Jenkins TC, Price CS, Sabel AL, Mehler PS, Burman WJ (2008) Impact of routine infectious diseases service consultation on the evaluation, management, and outcomes of Staphylococcus aureus bacteremia. Clin Infect Dis 46:1000–8. p J p 15. Pruefer D, Makowski J, Schnell M, Buerke U, Dahm M, et al. (2002) Simvastatin inhibits inflammatory properties of Staphylococcus aureus alpha-toxin. Circulation 106:2104–2110. 36. Almog Y, Novack V, Eisinger M, Porath A, Novack L, et al. References (2007) The effect of statin therapy on infection-related mortality in patients with atherosclerotic diseases. Crit Care Med 35:372–8. 16. Rodrı´guez-Ban˜o J, de Cueto M, Retamar P, Ga´lvez-Acebal J (2010) Current management of bloodstream infections. Expert Rev Anti Infect Ther 8: 815–29. 17. von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, et al; STROBE- Initiative. (2007) The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting of observational studies. PLoS Med 4: e296. 37. Gupta R, Plantinga LC, Fink NE, Melamed ML, Coresh J, et al. (2007) Statin use and sepsis events in patients with chronic kidney disease. JAMA 297:1455– 64. 38. Khatib R, Johnson LB, Fakih MG, Riederer K, Khosrovaneh A, et al. (2006) Persistence in Staphylococcus aureus bacteremia: incidence, characteristics of patients and outcome. Scand J Infect Dis 38:7–14. 18. Charlson ME, Pompei P, Ales KL, MacKenzie CR (1987) A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chron Dis 40:373–383. 39. Yoon YK, Kim JY, Park DW, Sohn JW, Kim MJ (2010) Predictors of persistent methicillin-resistant Staphylococcus aureus bacteraemia in patients treated with vancomycin. J Antimicrob Chemother 65:1015–8. J 19. Lesens O, Methlin C, Hansmann Y, Remy V, Martinot M, et al. (2003) Role of comorbidity in mortality related to Staphylococcus aureus bacteremia: a prospective study using the Charlson weighted index of comorbidity. Infect Control Hosp Epidemiol 24:890–6. 40. Khatib R, Johnson LB, Sharma M, Fakih MG, Ganga R, et al. (2009) Persistent Staphylococcus aureus bacteremia: incidence and outcome trends over time. Scand J Infect Dis 41:4–9. p 20. Chang FY, MacDonald BB, Peacock JE Jr, Musher DM, Triplett P, et al. (2003) A prospective multicenter study of Staphylococcus aureus bacteremia: incidence of endocarditis, risk factors for mortality, and clinical impact of methicillin resistance. Medicine (Baltimore) 82:322–32. 41. Fowler VG Jr, Sakoulas G, McIntyre LM, Meka VG, Arbeit RD, et al. (2004) Persistent bacteremia due to methicillin-resistant Staphylococcus aureus infection is associated with agr dysfunction and low-level in vitro resistance to thrombin- induced platelet microbicidal protein. J Infect Dis 190:1140–1149. ( ) 21. Gasch O, Camoez M, Domı´nguez MA, Padilla B, Pintado V, et al. (2013) Predictive factors for early mortality among patients with methicillin-resistant Staphylococcus aureus bacteraemia. J Antimicrob Chemother 68:1423–30. 42. Yeaman M (1997) The role of platelets in antimicrobial host defense. Clin Infect Dis 25:951–68. 22. References December 2013 \| Volume 8 \| Issue 12 \| e82958 PLOS ONE \| www.plosone.org 6 Statins & S. aureus Bacteremia 4. su J, Andes DR, Knasinski V, Pirsch J, Safdar N (2009) Statins are associated with improved outcomes of bloodstream infection in solid-organ transplant recipients. Eur J Clin Microbiol Infect Dis 28:1343–51. 24. Dellinger RP, Levy MM, Rhodes A, Annane D, Gerlach H, et al. (2008) Surviving sepsis campaign guidelines for management of severe sepsis and septic shock. Crit Care Med 36:296–327. p recipients. Eur J Clin Microbiol Infect Dis 28:1343–51. p 5. Kruger P, Fitzsimmons K, Cook D, Jones M, Nimmo G (2006) Statin therapy is associated with fewer deaths in patients with bacteraemia. Intensive Care Med 32: 75–9. 25. Rieg S, Peyerl-Hoffmann G, de With K, Theilacker C, Wagner D, et al. (2009) Mortality of S. aureus bacteremia and infectious diseases specialist consultation: a study of 521 patients in Germany. J Infect 59: 232–239. 6. Martin CP, Talbert RL, Burgess DS, Peters JI (2007) Effectiveness of statins in reducing the rate of severe sepsis: a retrospective evaluation. Pharmacotherapy 27:20–6. 26. Raad II, Sabbagh MF (1992) Optimal duration of therapy for catheter-related Staphylococcus aureus bacteremia: a study of 55 cases and review. Clin Infect Dis 14:75–82. 7. Ferna´ndez R, De Pedro VJ, Artigas A (2006) Statin therapy prior to ICU admission: protection against infection or a severity marker? Intensive Care Med 32:160–4. 27. Fowler VG Jr, Sanders LL, Sexton DJ, Kong L, Marr KA, et al. (1998) Outcome of Staphylococcus aureus bacteremia according to compliance with recommenda- tions of infectious diseases specialists: experience with 244 patients. Clin Infect Dis 27:478–86. 8. Thomsen RW, Hundborg HH, Johnsen SP, Pedersen L, Sørensen HT, et al. (2006) Statin use and mortality within 180 days after bacteremia: a population- based cohort study. Crit Care Med 34:1080–6. 28. Lahey T, Shah R, Gittzus J, Schwartzman J, Kirkland K (2009) Infectious Diseases consultation lowers mortality from Staphylococcus aureus bacteremia. Medicine (Baltimore) 88, 263–267. 9. ang KC, Chien JY, Tseng WK, Hsueh PR, Yu CJ, et al. (2007) Statins do not improve short term survival in an oriental population with sepsis. Am J Emerg Med 25:494–501. 29. Tenover C, Moellering R (2007) The rationale for revising the clinical and laboratory standards institute vancomycin minimal inhibitory concentration interpretive criteria for Staphylococcus aureus. Clin Infect Dis 44:1208–1215. 10. References Kim SH, Park WB, Lee KD, Kang CI, Kim HB, et al. (2003) Outcome of Staphylococcus aureus bacteremia in patients with eradicable foci versus none- radicable foci. Clin Infect Dis 37:794–9. 43. Viasus D, Garcia-Vidal C, Gudiol F, Carratala´ J (2010) Statins for community- acquired pneumonia: current state of the science. Eur J Clin Microbiol Infect Dis 29:143–152. 23. Bone RC, Balk RA, Cerra FB, Dellinger RP, Fein AM, et al. (1992) American College of Chest Physicians/Society of Critical Care Medicine Consensus Conference Committee. Definitions for sepsis and organ failure and guidelines for use of innovative therapies in sepsis. Crit Care Med 20:864–74. 44. Majumdar SR, McAlister FA, Eurich DT, Padwal RS, Marrie TJ (2006) Statins and outcomes in patients admitted to hospital with community acquired pneumonia: population based prospective cohort study. BMJ 333:999. December 2013 \| Volume 8 \| Issue 12 \| e82958 PLOS ONE \| www.plosone.org 7
https://openalex.org/W2614235497	https://europepmc.org/articles/pmc5477335?pdf=render	English	null	Draft Genome Sequence of the Tacrolimus-Producing Bacterium <i>Streptomyces tsukubaensis</i> F601	Genome announcements	2,017	cc-by	1,482	PROKARYOTES crossm PROKARYOTES crossm Volume 5 Issue 20 e00385-17 genomea.asm.org 1 Draft Genome Sequence of the Tacrolimus-Producing Bacterium Streptomyces tsukubaensis F601 Gongli Zong,a Chuanqing Zhong,b Jiafang Fu,a Ronghuo Qin,a Guangxiang Caoa Shandong Academy of Medical Sciences, Shandong Medicinal Biotechnology Centre, Jinan, Chinaa; Shandong Jianzhu University, School of Municipal and Environmental Engineering, Jinan, Chinab ABSTRACT Streptomyces tsukubaensis strain F601 was found to be a producer of the immunosuppressive drug tacrolimus. The draft genome sequence of this strain was approximately 8.52 Mbp. Genes involved in the biosynthesis of tacrolimus were identiﬁed in the genome. This draft genome sequence will provide insights into the genetic basis of tacrolimus biosynthesis and regulation. Received 29 March 2017 Accepted 30 March 2017 Published 18 May 2017 Citation Zong G, Zhong C, Fu J, Qin R, Cao G. 2017. Draft genome sequence of the tacrolimus-producing bacterium Streptomyces tsukubaensis F601. Genome Announc 5: e00385-17. https://doi.org/10.1128/genomeA .00385-17. Copyright © 2017 Zong et al. This is an open- access article distributed under the terms of the Creative Commons Attribution 4.0 International license. Address correspondence to Guangxiang Cao, caozhong0402@163.com. G.Z. and C.Z. contributed equally to this article. Received 29 March 2017 Accepted 30 March 2017 Published 18 May 2017 Citation Zong G, Zhong C, Fu J, Qin R, Cao G. 2017. Draft genome sequence of the tacrolimus-producing bacterium Streptomyces tsukubaensis F601. Genome Announc 5: e00385-17. https://doi.org/10.1128/genomeA .00385-17. F K506 (tacrolimus) is a natural product with immunosuppressive and antifungal activities; it has been used clinically to prevent the rejection of transplanted organs and in the treatment of eczema (1, 2). Recently, clinical studies have shown numerous promising therapeutic properties of FK506, such as neuroprotection and regeneration (3, 4). FK506 was ﬁrst detected in the fermentation broth of Streptomyces tsukubaensis 9993, which was isolated from a soil sample from Tsukuba, Ibaraki Prefecture, Japan, in 1984 (5). In the decades following its discovery, FK506 and its structural and functional analogs (such as FK520 and rapamycin) have been isolated from a variety of soil- dwelling Streptomyces species (6). FK506 biosynthesis involves a multifunctional polyketide synthase and nonribosomal peptide synthetase systems. The FK506 gene cluster has been partially or entirely sequenced from several streptomycete strains, such as Streptomyces sp. strain MA6548 (ATCC 53770), Streptomyces sp. strain KCTC 11604BP, Streptomyces kanamyceticus KCTC 9225, and S. tsukubaensis 18488 (6–9). F K a Copyright © 2017 Zong et al. This is an open- access article distributed under the terms of the Creative Commons Attribution 4.0 International license. Address correspondence to Guangxiang Cao, caozhong0402@163.com. G.Z. and C.Z. Received 29 March 2017 Accepted 30 March 2017 Published 18 May 2017 Citation Zong G, Zhong C, Fu J, Qin R, Cao G. 2017. Draft genome sequence of the tacrolimus-producing bacterium Streptomyces tsukubaensis F601. Genome Announc 5: e00385-17. https://doi.org/10.1128/genomeA .00385-17. Copyright © 2017 Zong et al. This is an open- access article distributed under the terms of the Creative Commons Attribution 4.0 International license. Address correspondence to Guangxiang Cao, caozhong0402@163.com. G.Z. and C.Z. contributed equally to this article. genomea.asm.org 1 ACKNOWLEDGMENTS This work was supported, in part, by the Shandong Natural Science Foundation (grants ZR2016CP04 and ZR2015EM018) and the Innovation Project of Shandong Academy of Medical Sciences (grant 201604). REFERENCES macrolactone ring of the immunosuppressant FK506. Eur J Biochem 256:528–534. https://doi.org/10.1046/j.1432-1327.1998.2560528.x. macrolactone ring of the immunosuppressant FK506. Eur J Biochem 256:528–534. https://doi.org/10.1046/j.1432-1327.1998.2560528.x. 1. Meingassner JG, Stütz A. 1992. Immunosuppressive macrolides of the type FK 506: a novel class of topical agents for treatment of skin diseases? J Invest Dermatol 98:851–855. https://doi.org/10.1111/1523 -1747.ep12456939. 1. Meingassner JG, Stütz A. 1992. Immunosuppressive macrolides of the type FK 506: a novel class of topical agents for treatment of skin diseases? J Invest Dermatol 98:851–855. https://doi.org/10.1111/1523 -1747.ep12456939. 8. Goranovic D, Kosec G, Mrak P, Fujs S, Horvat J, Kuscer E, Kopitar G, Petkovic H. 2010. Origin of the allyl group in FK506 biosynthesis. J Biol Chem 285:14292–14300. https://doi.org/10.1074/jbc.M109.059600. 2. Wallemacq PE, Reding R. 1993. FK506 (tacrolimus), a novel immunosup- pressant in organ transplantation: clinical, biomedical, and analytical aspects. Clin Chem 39:2219–2228. 2. Wallemacq PE, Reding R. 1993. FK506 (tacrolimus), a novel immunosup- pressant in organ transplantation: clinical, biomedical, and analytical aspects. Clin Chem 39:2219–2228. 9. Goranovicˇ D, Blažicˇ M, Magdevska V, Horvat J, Kušcˇer E, Polak T, Santos- Aberturas J, Martínez-Castro M, Barreiro C, Mrak P, Kopitar G, Kosec G, Fujs S, Martín JF, Petkovic´ H. 2012. FK506 biosynthesis is regulated by two positive regulatory elements in Streptomyces tsukubaensis. BMC Microbiol 12:238. https://doi.org/10.1186/1471-2180-12-238. 3. Sierra-Paredes G, Sierra-Marcuño G. 2008. Ascomycin and FK506: phar- macology and therapeutic potential as anticonvulsants and neuropro- tectants. CNS Neurosci Ther 14:36–46. https://doi.org/10.1111/j.1527 -3458.2008.00036.x. 10. Bankevich A, Nurk S, Antipov D, Gurevich AA, Dvorkin M, Kulikov AS, Lesin VM, Nikolenko SI, Pham S, Prjibelski AD, Pyshkin AV, Sirotkin AV, Vyahhi N, Tesler G, Alekseyev MA, Pevzner PA. 2012. SPAdes: a new genome assembly algorithm and its applications to single-cell sequenc- ing. J Comput Biol 19:455–477. https://doi.org/10.1089/cmb.2012.0021. 4. Gold BG. 2000. Neuroimmunophilin ligands: evaluation of their thera- peutic potential for the treatment of neurological disorders. Expert Opin Invest Drugs 9:2331–2342. https://doi.org/10.1517/13543784.9.10.2331. 4. Gold BG. 2000. Neuroimmunophilin ligands: evaluation of their thera- peutic potential for the treatment of neurological disorders. Expert Opin Invest Drugs 9:2331–2342. https://doi.org/10.1517/13543784.9.10.2331. 5. Kino T, Hatanaka H, Hashimoto M, Nishiyama M, Goto T, Okuhara M, Kohsaka M, Aoki H, Imanaka H. 1987. FK-506, a novel immunosuppres- sant isolated from a Streptomyces. I. Fermentation, isolation, and physico-chemical and biological characteristics. J Antibiot (Tokyo) 40: 1249–1255. https://doi.org/10.7164/antibiotics.40.1249. 5. Kino T, Hatanaka H, Hashimoto M, Nishiyama M, Goto T, Okuhara M, Kohsaka M, Aoki H, Imanaka H. 1987. Draft Genome Sequence of the Tacrolimus-Producing Bacterium Streptomyces tsukubaensis F601 contributed equally to this article. Streptomyces tsukubaensis F601 is a potent producer of FK506, and the original strain was isolated from forest soil in Shandong, China. Here, we report the whole-genome sequence of S. tsukubaensis F601 and its features. The genome of strain F601 was sequenced using the Illumina HiSeq 4000 platform at the Beijing Genomics Institute (Shenzhen, China). The paired-end sequencing library was constructed with the NEBNext Ultra DNA library prep kit for Illumina (New England BioLabs, UK). Sequencing of the genome produced a raw data set of 10,766,446 paired-end reads. After removing low-quality reads, duplication reads, and adapter reads, about 1,500 Mb of clean data were obtained, which gave a 35.48-fold average coverage of the genome. The clean reads were assembled de novo with SPAdes (10) and then polished with SSPACE-Standard and GapFiller to obtain scaffold sequences (11, 12). Using SPAdes, a total of 8,506,753 bp of genomic sequence was assembled, which covers about 95.01% of the predicted genome; the GC content was calculated to be 70.45%. The genome was annotated using the Prokaryotic Genome Annotation Pipeline (PGAP) version 3.2 software on NCBI. Additional functional annotation was performed using the RASTtk server (13). The automated gene prediction identiﬁed 7,257 coding genes (CDSs) and 73 RNA-coding genes, including 64 tRNA, two 5S rRNA, one 16S rRNA, one 23S rRNA, and ﬁve noncoding RNA genes. The genome also harbors the FK506 biosynthetic gene cluster, which partially shares high sequence similarity with the FK506 gene clusters from S. kanamyceticus KCTC 9225 and S. tsukubaensis 18488. This Volume 5 Issue 20 e00385-17 genomea.asm.org 1 Zong et al. draft genome sequence will help us further understand the molecular mechanisms of FK506 biosynthesis and regulation. Accession number(s). The draft sequence of S. tsukubaensis strain F601 obtained in this whole-genome shotgun sequencing project has been deposited in DDBJ/ENA/ GenBank under accession no. MVFC00000000. The version described in this paper has accession no. MVFC01000000. Volume 5 Issue 20 e00385-17 REFERENCES FK-506, a novel immunosuppres- sant isolated from a Streptomyces. I. Fermentation, isolation, and physico-chemical and biological characteristics. J Antibiot (Tokyo) 40: 1249–1255. https://doi.org/10.7164/antibiotics.40.1249. 11. Boetzer M, Henkel CV, Jansen HJ, Butler D, Pirovano W. 2011. Scaffolding pre-assembled contigs using SSPACE. Bioinformatics 27:578–579. https://doi.org/10.1093/bioinformatics/btq683. 12. Boetzer M, Pirovano W. 2012. Toward almost closed genomes with GapFiller. Genome Biol 13:R56. https://doi.org/10.1186/gb-2012-13-6 -r56. 6. Mo S, Kim DH, Lee JH, Park JW, Basnet DB, Ban YH, Yoo YJ, Chen SW, Park SR, Choi EA, Kim E, Jin YY, Lee SK, Park JY, Liu Y, Lee MO, Lee KS, Kim SJ, Kim D, Park BC, Lee SG, Kwon HJ, Suh JW, Moore BS, Lim SK, Yoon YJ. 2011. Biosynthesis of the allylmalonyl-CoA extender unit for the FK506 polyketide synthase proceeds through a dedicated polyketide synthase and facilitates the mutasynthesis of analogues. J Am Chem Soc 133: 976–985. https://doi.org/10.1021/ja108399b. 13. Brettin T, Davis JJ, Disz T, Edwards RA, Gerdes S, Olsen GJ, Olson R, Overbeek R, Parrello B, Pusch GD, Shukla M, Thomason JA, III, Stevens R, Vonstein V, Wattam AR, Xia F. 2015. RASTtk: a modular and extensible implementation of the RAST algorithm for building custom annotation pipelines and annotating batches of genomes. Sci Rep 5:8365. https:// doi.org/10.1038/srep08365. g j 7. Motamedi H, Shaﬁee A. 1998. The biosynthetic gene cluster for the Volume 5 Issue 20 e00385-17 genomea.asm.org 2
https://openalex.org/W2811184474	https://curis.ku.dk/ws/files/215563486/jensen_2018_oi_180042.pdf	English	null	Endocrine Late Effects in Survivors of Cancer in Adolescence and Young Adulthood	JAMA network open	2,018	cc-by	9,233	Citation for published version (APA): Jensen, M. V., Rugbjerg, K., de Fine Licht, S., Johansen, C., Schmiegelow, K., Andersen, K. K., & Winther, J. F. (2018). Endocrine Late Effects in Survivors of Cancer in Adolescence and Young Adulthood: A Danish Population-Based Cohort Study. JAMA network open, 1(2), 1-12. [e180349]. https://doi.org/10.1001/jamanetworkopen.2018.0349 Download date: 24. Oct. 2024 Endocrine Late Effects in Survivors of Cancer in Adolescence and Young Adulthood A Danish Population-Based Cohort Study Jensen, Mette Vestergaard; Rugbjerg, Kathrine; de Fine Licht, Sofie; Johansen, Christoffer; Schmiegelow, Kjeld; Andersen, Klaus Kaae; Winther, Jeanette Falck Published in: JAMA network open DOI: 10.1001/jamanetworkopen.2018.0349 Publication date: 2018 Document version Publisher's PDF, also known as Version of record Document license: CC BY Citation for published version (APA): Jensen, M. V., Rugbjerg, K., de Fine Licht, S., Johansen, C., Schmiegelow, K., Andersen, K. K., & Winther, J. F. (2018). Endocrine Late Effects in Survivors of Cancer in Adolescence and Young Adulthood: A Danish Population-Based Cohort Study. JAMA network open, 1(2), 1-12. [e180349]. https://doi.org/10.1001/jamanetworkopen.2018.0349 Endocrine Late Effects in Survivors of Cancer in Adolescence and Young Adulthood A Danish Population-Based Cohort Study Jensen, Mette Vestergaard; Rugbjerg, Kathrine; de Fine Licht, Sofie; Johansen, Christoffer; Schmiegelow, Kjeld; Andersen, Klaus Kaae; Winther, Jeanette Falck Endocrine Late Effects in Survivors of Cancer in Adolescence and Young Adulthood A Danish Population-Based Cohort Study Document version Publisher's PDF, also known as Version of record university of copenhagen Endocrine Late Effects in Survivors of Cancer in Adolescence and Young Adulthood A Danish Population-Based Cohort Study Jensen, Mette Vestergaard; Rugbjerg, Kathrine; de Fine Licht, Sofie; Johansen, Christoffer; Schmiegelow, Kjeld; Andersen, Klaus Kaae; Winther, Jeanette Falck university of copenhagen Endocrine Late Effects in Survivors of Cancer in Adolescence and Young Adulthood A Danish Population-Based Cohort Study Jensen, Mette Vestergaard; Rugbjerg, Kathrine; de Fine Licht, Sofie; Johansen, Christoffer; Schmiegelow, Kjeld; Andersen, Klaus Kaae; Winther, Jeanette Falck university of copenhagen Original Investigation \| Oncology Endocrine Late Effects in Survivors of Cancer in Adolescence and Young Adulthood A Danish Population-Based Cohort Study A Danish Population-Based Cohort Study Mette Vestergaard Jensen, MD; Kathrine Rugbjerg, PhD; Sofie de Fine Licht, PhD; Christoffer Johansen, DMSc; Kjeld Schmiegelow, DMSc; Klaus Kaae Andersen, PhD; Jeanette Falck Winther, DMSc + Invited Commentary + Supplemental content Author affiliations and article information are listed at the end of this article. Document license: CC BY Citation for published version (APA): Jensen, M. V., Rugbjerg, K., de Fine Licht, S., Johansen, C., Schmiegelow, K., Andersen, K. K., & Winther, J. F. (2018). Endocrine Late Effects in Survivors of Cancer in Adolescence and Young Adulthood: A Danish Population-Based Cohort Study. JAMA network open, 1(2), 1-12. [e180349]. https://doi.org/10.1001/jamanetworkopen.2018.0349 Download date: 24. Oct. 2024 June 29, 2018 1/12 Key Points Question Are adolescent and young adult cancer survivors at increased risk for endocrine diseases? IMPORTANCE As survival rates from cancer have improved dramatically over the last decades, there is a need to explore the long-term consequences. Adolescents and young adults with cancer are at risk for several therapy-related late effects; however, these have not been studied extensively. Findings This Danish population-based cohort study that included 32 548 adolescent and young adult cancer survivors shows a 73% higher risk for endocrine diseases in these cancer survivors than a matched cancer-free population. This study shows the patterns of endocrine late effects associated with many cancer sites and how they were modified by patient factors. OBJECTIVE To investigate the lifetime risks of endocrine late effects of cancer and cancer treatment in adolescent and young adult cancer survivors. DESIGN, SETTING, AND PARTICIPANTS This Danish, nationwide, population-based cohort study was conducted from January 1, 1976, through December 31, 2009, and included follow-up from January 1, 1977, through December 31, 2010. A total of 32 548 one-year cancer survivors diagnosed at ages 15 to 39 years were identified using the Danish Cancer Registry and 188 728 cancer-free comparison participants matched by year of birth and sex were randomly chosen from the Danish Civil Registration system. Analyses were performed from July 3, 2015, to February 27, 2018. EXPOSURES Individuals in the survivor cohort were diagnosed with a first primary cancer at ages 15 to 39 years and received treatment according to recommendations and guidelines at time of diagnosis. Meaning This study represents the first step in identifying patients who are at risk for endocrine late effects and indicates the need for surveillance of these patients to prevent the most severe conditions. MAIN OUTCOMES AND MEASURES By linkage to the National Patient Register, all hospital contacts for endocrine diseases were identified, and standardized hospitalization rate ratios (RRs) and absolute excess risks (AERs) were calculated. Abstract (continued) CONCLUSIONS AND RELEVANCE The increased risk for endocrine diseases in adolescent and young adult cancer survivors indicates the need for counseling and follow-up, and could guide future preventive measures and surveillance strategies. Additional studies are required to determine exact associations between treatment regimens and endocrine diseases. JAMA Network Open. 2018;1(2):e180349. Corrected on July 27, 2018. doi:10.1001/jamanetworkopen.2018.0349 JAMA Network Open. 2018;1(2):e180349. Corrected on July 27, 2018. doi:10.1001/jamanetworkopen.2018.0349 JAMA Network Open. 2018;1(2):e180349. Corrected on July 27, 2018. doi:10.1001/jamanetworkopen.2018.0349 Introduction Adolescent and young adult oncology has recently become a subspecialty of cancer research.1,2 Adolescent and young adult cancer survivors, defined as those in whom cancer was diagnosed when they were aged 15 to 39 years,2-4 differ from younger and older patients with cancer in the biology, epidemiology, and clinical outcomes of cancer5 and are at risk for long-term morbidity associated with their cancer or cancer treatment. Cancer is 7-fold more frequent in adolescents and young adults than in children younger than 15 years,3,4 with the pattern of late effects depending on the age at diagnosis.6 Most data on the long-term sequelae in cancer survivors at a young age are derived from studies of childhood cancer survivors.3,7 Survivors of cancer in adolescence and young adulthood, their relatives, and the treating clinicians also require information on the long-term outcomes of treatment. Common late effects in this population of survivors include second primary cancers, cardiovascular and pulmonary complications, neurological complications, and endocrine and gonadal disorders.6 These are stages of life with many transitions in terms of education, employment, social relations, relocations, and family formation.3 Endocrine late effects, with hormonal disturbances and gonadal dysfunction, could have many physical and social consequences for cancer survivors. Several studies have assessed the late effects of treatment for site-specific cancers in the age range of adolescents and young adults,3 including testicular cancer,8 Hodgkin lymphoma,9-11 and breast cancer,12 reporting increased risks for hypogonadism, hypothyroidism, and premature menopause in survivors. However, there is little information about endocrine late effects in survivors of other cancers in this age group. We report the results of, to our knowledge, the first large-scale population- based study of all hospital contacts for endocrine diseases, which includes more than 32 000 adolescent and young adult 1-year cancer survivors and 5-fold as many population comparisons. June 29, 2018 2/12 MAIN OUTCOMES AND MEASURES By linkage to the National Patient Register, all hospital contacts for endocrine diseases were identified, and standardized hospitalization rate ratios (RRs) and absolute excess risks (AERs) were calculated. + Invited Commentary + Supplemental content Author affiliations and article information are listed at the end of this article. RESULTS A total of 32 548 adolescent and young adult 1-year cancer survivors (14 021 [43.1%] male) in the Danish Patient Registry were followed up for 379 157 person-years (median [range]: 10 [0-34] years) and 188 728 cancer-free participants (82 669 [43.8%] male) for comparison were followed up for 2 958 994 person-years (median [range]: 15 [0-34] years). A total of 2129 survivors (6.5%) had at least 1 hospital contact for an endocrine disease, while 1232.0 (3.8%) were expected, yielding a statistically significant increased RR of 1.73 (95% CI, 1.65-1.81). The RRs were highest for testicular hypofunction (75.12; 95% CI, 45.99-122.70), ovarian hypofunction (14.65; 95% CI, 8.29-25.86), and pituitary hypofunction (11.14; 95% CI, 8.09-15.34). The leading reasons for hospital contacts were thyroid disease (38.0% of total AER), testicular dysfunction (17.1% of total AER), and diabetes (14.4% of total AER). Leukemia survivors were at a high risk for any endocrine disease (RR, 3.97; 95% CI, 3.10-5.09), while Hodgkin lymphoma survivors (RR, 3.06; 95% CI, 2.62-3.57) had the highest disease-specific excess risk for hypothyroidism (AER, 362 per 100 000 person-years; 95% CI, 280-443 per 100 000 person-years). (continued) Open Access. This is an open access article distributed under the terms of the CC-BY License. June 29, 2018 1/12 JAMA Network Open. 2018;1(2):e180349. doi:10.1001/jamanetworkopen.2018.0349 Downloaded From: https://jamanetwork.com/ by a Copenhagen University Library User on 03/29/2019 Downloaded From: https://jamanetwork.com/ by a Copenhagen University Library User on 03/29/2019 Endocrine Late Effects in Survivors of Cancer in Adolescence and Young Adulthood JAMA Network Open \| Oncology Abstract (continued) Survivors and Comparison Participants The basic cohort of adolescent and young adult cancer survivors was identified from the Danish Cancer Registry,13 which contains records of all incident cases of cancer nationwide since 1943. Each cancer record includes the personal identification number, date of diagnosis, and type of cancer. Cancers were classified according to modified International Classification of Diseases, Seventh Revision codes between 1943 and 1977 and according to the International Classification of Diseases, Tenth Revision (ICD-10) for diagnosis and the International Classification of Diseases for Oncology, Third Revision for morphology and topography from 1978 onward.13 The unique personal identification number assigned to all Danish citizens since the start of the Civil Registration System on April 2, 1968, incorporates date of birth and sex and permits accurate linkage among Danish administrative and medical registers.14,15 From 1976 through 2009, 38 670 individuals aged 15 to 39 years were registered with a first primary cancer in the Danish Cancer Registry. These individuals were the basis for the 1-year survivor cohort to observe in the Danish National Patient Registry, which is a nationwide registry initiated on January 1, 1977 (eFigure 1 in the Supplement). We followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline. JAMA Network Open. 2018;1(2):e180349. doi:10.1001/jamanetworkopen.2018.0349 June 29, 2018 2/12 Downloaded From: https://jamanetwork.com/ by a Copenhagen University Library User on 03/29/2019 Downloaded From: https://jamanetwork.com/ by a Copenhagen University Library User on 03/29/2019 JAMA Network Open \| Oncology Endocrine Late Effects in Survivors of Cancer in Adolescence and Young Adulthood This study was approved by the Danish Data Protection Agency. Patient consent was waived because the data set did not include any identifiable sensitive personal data or identification number. The comparison cohort was identified through the Civil Registration System.14,15 For each survivor, we randomly chose 5 cancer-free comparison participants of the same sex and year of birth who were alive on the date of diagnosis of cancer in the survivor, resulting in 193 350 comparisons. In both cohorts, we excluded individuals who had died or emigrated during the first year after cancer diagnosis or a corresponding date for the comparison participants. This resulted in a cohort of 34 448 survivors and 192 254 comparison participants (eFigure 1 in the Supplement). Follow-up Follow-up for endocrine disease late effects started 1 year after cancer diagnosis and on a corresponding date for the equivalent comparison participants. Follow-up ended on the date of death, emigration, second primary cancer diagnosis in survivors or first primary cancer diagnosis in comparison participants, or the closing date (December 31, 2010), whichever occurred first. For participants with more than 1 hospital contact for a particular endocrine disease, only the first record was retained, which was presumed to be the date of diagnosis. Hospital Contacts for Endocrine Disease For both survivors and comparison participants, we identified all hospital admissions and outpatient visits with a primary or supplementary discharge diagnosis of endocrine disease (International Classification of Diseases, Eighth Revision [ICD-8] codes 240-258 and ICD-10 codes E01-E35 and E89) in the Danish Patient Registry.16 The registry includes information on all hospital admissions for somatic disease since 1977, including outpatient visits since 1995. Diagnoses made by general practitioners were not included. Diagnostic information was coded according to ICD-8 until 1993 and thereafter according to ICD-10.16 We identified hospital admissions for endocrine diseases defined in chapter 3 of ICD-8 and chapter 4 of ICD-10 and excluded nutritional and metabolic diseases, including metabolic syndrome, as we consider these diseases multifactorial and strongly associated with lifestyle or other factors rather than with cancer treatment. Further, metabolic syndrome was introduced in ICD-10 only recently. We also excluded congenital endocrine disorders, disorders of puberty, and diseases of the thymus, which occur primarily in childhood. We excluded patients in whom an endocrine disease had been diagnosed before the cancer or a corresponding date for comparison participants as well as survivors with chromosome abnormalities or congenital endocrine malformations (eFigure 1 in the Supplement) registered in the Danish Patient Registry. We also excluded patients with a pituitary tumor because of the direct risk for endocrine dysfunction. After these exclusions, 32 548 one-year survivors and 188 728 population comparison participants remained for analysis. June 29, 2018 3/12 Results We studied 32 548 adolescent and young adult 1-year cancer survivors (14 021 [43.1%] male) in the Danish Patient Registry for 379 157 person-years (median [range] time, 10 [0-34] years) and 188 728 population comparisons (82 669 [43.8%] male) for 2 958 994 person-years (median [range] time, 15 [0-34] years). A total of 2129 survivors (6.5%) had at least 1 hospital contact for an endocrine disease, while 1232.0 (3.8%) were expected, yielding a statistically significantly increased RR of 1.73 (95% CI, 1.65-1.81) (Table 1). The overall observed and expected rates of hospitalization for endocrine Table 1. Numbers of First Hospital Contacts for Endocrine Disease by Patient Factors Characteristic Hospital Contacts, No. Results Survivors at Risk, Person-Years All Endocrine Disease First Hospital Contacts, No.a Hospitalization Rate per 100 000 Person-Years Observed Expected RR (95% CI) Observed Expected AER (95% CI) Totalb 32 548 379 157 2129 1232.0 1.73 (1.65 to 1.81) 561.5 324.9 237 (212 to 261) Sex Male 14 021 167 616 826 342.5 2.41 (2.23 to 2.61) 492.8 204.3 288 (254 to 323) Female 18 527 211 541 1303 889.5 1.46 (1.38 to 1.55) 616.0 420.5 195 (160 to 230) Attained age, y 16-19 1369 2128 17 2.0 8.43 (4.14 to 17.16) 798.8 94.8 704 (321 to 1087) 20-29 9363 34 812 197 49.9 3.95 (3.31 to 4.70) 565.9 143.4 423 (342 to 503) 30-39 26 930 121 540 603 261.3 2.31 (2.10 to 2.53) 496.1 215.0 281 (240 to 322) 40-49 21 463 145 634 728 470.5 1.55 (1.43 to 1.68) 499.9 323.1 177 (139 to 215) 50-59 9997 60 499 410 331.3 1.24 (1.12 to 1.37) 677.7 547.6 130 (62 to 199) 60-69 3539 14 152 166 113.5 1.46 (1.24 to 1.72) 1172.9 802.1 371 (187 to 555) ≥70 285 391 8 3.5 2.29 (1.08 to 4.87) 2044.2 890.9 1153 (−287 to 2594) Age at cancer diagnosis, y 15-19 1838 21 014 139 34.4 4.04 (3.41 to 4.78) 661.5 163.8 498 (387 to 608) 20-24 3328 42 100 248 88.7 2.80 (2.46 to 3.17) 589.1 210.7 378 (305 to 452) 25-29 5920 73 297 362 200.0 1.81 (1.63 to 2.01) 493.9 272.9 221 (170 to 272) 30-34 8939 104 625 568 349.1 1.63 (1.50 to 1.77) 542.9 333.7 209 (164 to 254) 35-39 12 523 138 121 812 559.9 1.45 (1.35 to 1.56) 587.9 405.3 183 (142 to 224) Calender year for cancer diagnosis 1975-1989 12 411 217 562 1036 717.9 1.44 (1.35 to 1.54) 476.2 330.0 146 (116 to 176) 1990-2009 20 137 161 595 1093 514.2 2.13 (2.00 to 2.26) 676.4 318.2 358 (318 to 399) Years since cancer diagnosis 1-4 32 548 109 682 616 237.1 2.60 (2.39 to 2.82) 561.6 216.1 345 (301 to 390) 5-9 23 874 100 059 421 254.0 1.66 (1.50 to 1.83) 420.8 253.9 167 (126 to 208) 10-19 17 642 121 623 649 448.8 1.45 (1.34 to 1.57) 533.6 369.0 165 (123 to 206) >20 8325 47 793 443 292.2 1.52 (1.38 to 1.67) 926.9 611.3 315 (229 to 402) Abbreviations: AER, absolute excess risk; RR, hospitalization rate ratio. Statistical Analysis A Cox proportional hazards model was used to esti- mate hospitalization rate hazard ratios for any endocrine disease. All tests were 2-sided likelihood ratio tests with a P value less than .05 considered statistically significant. We estimated the cumulative inci- dence of overall and selected endocrine diseases, with death as a competing risk18 and age as the un- derlying timescale using left truncation. Statistical analyses were conducted using SAS software version 9.3 (SAS Institute Inc) and R statistical computing software version 3.2.3 (R Foundation), and the Sur- vival and Design packages were used. Statistical Analysis Risk analyses were performed for any endocrine disease and for 9 main diagnostic categories and 26 subcategories. We compared the rates of first hospital contact for survivors with expected rates de- rived from the comparison cohort. We estimated hospitalization rate ratios (RRs) with 95% confidence intervals as the ratio of observed to expected hospitalization rates for each defined disease entity sepa- rately and by censoring for competing events. We used Fieller theorem, assuming that the observed number of first hospital contacts followed a Poisson distribution.17 We calculated absolute excess risks (AERs)—the additional risk of adolescent and young adult cancer survivors for a hospital contact for an endocrine disease—as the difference between the observed and the expected first hospitalization rate for endocrine disease per 100 000 person-years of follow-up, with 95% confidence intervals. Risk analyses were performed for the total cohort of survivors and separately in subcohorts of survivors who had 1 of the 10 most frequent cancers occurring in adolescence and young adulthood. Analyses were stratified according to sex, age at cancer diagnosis, calendar period of cancer diagnosis, and attained age at diagnosis of the endocrine disease. The association of survivor demographic characteristics (sex, JAMA Network Open. 2018;1(2):e180349. doi:10.1001/jamanetworkopen.2018.0349 June 29, 2018 3/12 Downloaded From: https://jamanetwork.com/ by a Copenhagen University Library User on 03/29/2019 Downloaded From: https://jamanetwork.com/ by a Copenhagen University Library User on 03/29/2019 age, year of diagnosis, and type of cancer) with the risk for any endocrine disease was investigated in a multivariate analysis within the survivor cohort. A Cox proportional hazards model was used to esti- mate hospitalization rate hazard ratios for any endocrine disease. All tests were 2-sided likelihood ratio tests with a P value less than .05 considered statistically significant. We estimated the cumulative inci- dence of overall and selected endocrine diseases, with death as a competing risk18 and age as the un- derlying timescale using left truncation. Statistical analyses were conducted using SAS software version 9.3 (SAS Institute Inc) and R statistical computing software version 3.2.3 (R Foundation), and the Sur- vival and Design packages were used. JAMA Network Open \| Oncology Endocrine Late Effects in Survivors of Cancer in Adoles Endocrine Late Effects in Survivors of Cancer in Adolescence and Young Adulthood JAMA Network Open \| Oncology age, year of diagnosis, and type of cancer) with the risk for any endocrine disease was investigated in a multivariate analysis within the survivor cohort. June 29, 2018 4/12 Abbreviations: AER, absolute excess risk; RR, hospitalization rate ratio. Abbreviations: AER, absolute excess risk; RR, hospitalization rate ratio. b According to Statistics Denma Results a Observed and expected numbers of first hospital contacts for endocrine disease of any type among 32 548 adolescent and young adult 1-year cancer survivors in Denmark, b According to Statistics Denmark, 8% of the Danish population has immigrated from non-Western countries. Table 1. Numbers of First Hospital Contacts for Endocrine Disease by Patient Factors JAMA Network Open. 2018;1(2):e180349. doi:10.1001/jamanetworkopen.2018.0349 June 29, 2018 4/12 Downloaded From: https://jamanetwork.com/ by a Copenhagen University Library User on 03/29/2019 Endocrine Late Effects in Survivors of Cancer in Adolescence and Young Adulthood JAMA Network Open \| Oncology diseases were 561.5 and 324.9 per 100 000 person-years, respectively, resulting in an AER of 236.6 (95% CI, 212-261) per 100 000 person-years (ie, a new excess case of endocrine disease in 237 of 100 000 survivors for each additional year of follow-up) (Table 1). The risk for any endocrine disease compared with risk in the background population was higher in male survivors (RR, 2.41; 95% CI, 2.23-2.61) than in female survivors (RR, 1.46; 95% CI, 1.38-1.55), reflecting lower background rates of endocrine diseases in males. A younger age at cancer diagnosis was associated with a higher RR and a higher AER for any endocrine disease (Table 1). The risk for hospital contact for endocrine diseases decreased with time since cancer diagnosis, with an RR of 2.60 (95% CI, 2.39-2.82) for 1 to 4 years since diagnosis and an RR of 1.45 (95% CI, 1.34-1.57) for 10 to 19 years since diagnosis (Table 1). Survivors had significantly increased risks for 8 of 9 main diagnostic groups of endocrine disease and 18 of 26 subcategories (Table 2). The highest RRs were seen for testicular hypofunction (75.12; 95% CI, 45.99-122.70), ovarian hypofunction (14.65; 95% CI, 8.29-25.86), and pituitary hypofunction (11.14; 95% CI, 8.09-15.34). Diseases of the thyroid gland, testicular dysfunction, and diabetes were the leading reasons for hospital contacts (38.0%, 17.1%, and 14.4% of total AER, respectively). Analyses by sites of cancer revealed the highest RRs for any endocrine disease for survivors of leukemia (RR, 3.97; 95% CI, 3.10-5.09), Hodgkin lymphoma (RR, 3.06; 95% CI, 2.62-3.57), and brain cancer (RR, 3.03; 95% CI, 2.53-3.64) (Table 3; eFigure 2 in the Supplement). Survivors of Hodgkin lymphoma had a particularly high excess risk for hypothyroidism (AER, 362 per 100 000 person-years; 95% CI, 280-443 per 100 000 person-years). Results Brain cancer survivors were at increased risk for a broad spectrum of endocrine diseases, comprising diseases of the pituitary and thyroid glands as well as diabetes. For survivors of the most common cancers among adolescents and young adults, the risks for hospital contact for endocrine disease were significantly increased for testicular cancer (RR, 2.50; 95% CI, 2.25-2.78) and breast cancer (RR, 1.16; 95% CI, 1.02-1.32), whereas the risks among survivors of malignant melanoma (RR, 1.08; 95% CI, 0.95-1.24) and cervical cancer (RR, 0.94; 95% CI, 0.82-1.07) were similar to those of the comparison cohort (Table 3). When the survivors were stratified by period of treatment, the RR of hospital contact for endocrine disease for those diagnosed during the period of 1975 to 1989 was 1.44 (95% CI, 1.35-1.54), whereas the corresponding risk for those diagnosed during the period of 1990 to 2009 was 48% higher (RR, 2.13; 95% CI, 2.00-2.26) (Table 1). Overall, RRs and AERs diminished with age (Table 1 and Figure). The risk for diabetes increased markedly after age 50 years for both survivors and comparison participants, while the RR and AER for hypothyroidism were highest in the age group 20 to 39 years. The vast majority of cases of pituitary hypofunction were diagnosed before the age of 30 years. The cumulative risk of survivors for any endocrine disease by age 30 years was 6.6% (95% CI, 6.5%-6.6%) and by age 60 years was 12.5% (95% CI, 12.2%-12.7%), while the cumulative risk of comparison participants by age 30 years was 1.7% (95% CI, 1.7%-1.7%) and by age 60 years was 11.4% (95% CI, 11.2%-11.6%). Survivors of Hodgkin lymphoma were at considerably high risk for hypothyroidism, with a cumulative risk of 9.6% (95% CI, 7.5%-11.6%) by age 60 years; the corresponding cumulative risk for comparisons was 1.4% (95% CI, 1.3%-1.4%). In a multivariate analysis, significant variation in the risk of the survivor cohort for any endocrine disorder was found by sex, age at diagnosis, year of diagnosis, and type of cancer (Table 4). The adjusted hazard ratio of females for any endocrine disorder was 1.85 (95% CI, 1.65-2.09; P < .001) compared with males. June 29, 2018 5/12 Results In line with our findings in Table 1, the adjusted hazard ratios decreased significantly with increasing age at cancer diagnosis (ages 35-39 vs 15-19 years: adjusted HR, 0.55; 95% CI, 0.43-0.71; P < .001) and increased in the most recent calendar period of diagnosis (1990- 2009 vs 1975-1989: adjusted HR, 1.97; 95% CI, 1.78-2.18; P < .001). The adjusted hazard ratios according to cancer site matched the corresponding RR estimates compared with the general population (Table 3 and Table 4; eFigure 2 in the Supplement); ie, in both analyses, survivors of leukemia and Hodgkin lymphoma had higher risks for endocrine disease than survivors of brain cancer, whereas survivors of cancers at other sites had lower risks. JAMA Network Open. 2018;1(2):e180349. doi:10.1001/jamanetworkopen.2018.0349 June 29, 2018 5/12 Downloaded From: https://jamanetwork.com/ by a Copenhagen University Library User on 03/29/2019 Downloaded From: https://jamanetwork.com/ by a Copenhagen University Library User on 03/29/2019 Endocrine Late Effects in Survivors of Cancer in Adolescence and Young Adulthood JAMA Network Open \| Oncology Table 2. Results First Hospital Contacts for Endocrine Disease by Diagnostic Categories or Diagnosesa Category of Endocrine Disease and Diagnostic Entity (ICD-10 Code)b First Hospital Contacts, No.c RR (95% CI) AER (95% CI) % of Total AER Observed Expected Diseases of the thyroid gland (E01, E02, E03.2-E03.9, E04-E07, E35.0, and E89.0) 1039 654.9 1.59 (1.48 to 1.70) 100 (83 to 117) 38.0 Goiter, nontoxic (E01 and E04) 394 325.6 1.21 (1.09 to 1.35) 18 (7 to 28) Thyrotoxicosis (E05) 279 225.0 1.24 (1.09 to 1.41) 14 (5 to 23) Hypothyroidism (E02, E03.2-E03.9, and E89.0) 444 152.4 2.91 (2.61 to 3.25) 75 (64 to 85) Thyroiditis (E06) 55 46.8 1.17 (0.88 to 1.56) 2 (−2 to 6) Other diseases of the thyroid gland (E07 and E35.0) 14 15.5 0.91 (0.52 to 1.58) 0 (−2 to 2) Diabetes (E10-E14 and E89.1) 660 511.1 1.29 (1.19 to 1.40) 38 (25 to 52) 14.4 Type 1 (E10) 264 188.4 1.40 (1.23 to 1.60) 19 (11 to 28) Type 2 (E11) 520 408.4 1.27 (1.16 to 1.40) 29 (16 to 41) Other types (E12-E14 and E89.1) 146 89.4 1.63 (1.37 to 1.95) 14 (8 to 21) Disorders of pancreatic internal secretion other than diabetes (E15-E16) 49 41.1 1.19 (0.88 to 1.61) 2 (−2 to 6) 0.8 Diseases of parathyroid glands (E20-E21, and E89.2) 93 33.8 2.75 (2.17 to 3.49) 15 (10 to 20) 5.7 Hyperparathyroidism (E21.0-E21.3) 48 26.4 1.82 (1.33 to 2.49) 6 (2 to 9) Hypoparathyroidism (E20-E89.2) 49 7.9 6.19 (4.24 to 9.04) 10 (7 to 14) Other diseases of parathyroid glands (E21.4 and E21.5) 1 0.4 2.53 (0.26 to 24.42) 0 (0 to 1) Diseases of pituitary gland (E22-E23 and E89.3) 122 22.1 5.53 (4.38 to 6.99) 25 (20 to 31) 9.5 Pituitary hyperfunction (E22) 19 12.0 1.58 (0.96 to 2.59) 2 (0 to 4) Pituitary hypofunction (E23.0-E23.3 and E89.3) 93 8.4 11.14 (8.09 to 15.34) 22 (17 to 26) Other diseases of pituitary gland (E23.6 and E23.7) 23 2.8 8.28 (4.58 to 14.96) 5 (3 to 8) Diseases of adrenal glands (E25.8, E25.9, E26, E27, E35.1, and E89.6) 73 18.25 4.00 (3.01 to 5.31) 14 (10 to 18) 5.3 Adrenocortical hyperfunction (E25.8, E25.9, E26, and E27.0) 12 5.5 2.17 (1.14 to 4.13) 2 (0 to 3) Adrenomedullary hyperfunction (E27.5) 14 1.6 8.66 (4.02 to 18.68) 3 (1 to 5) Adrenal hypofunction (E27.1-E27.4 and E89.6) 43 7.7 5.58 (3.76 to 8.28) 9 (6 to 12) Other diseases of adrenal glands (E27.8, E27.9, and E35.1) 6 4.1 1.46 (0.61 to 3.51) 0 (−1 to 2) Ovarian dysfunction (E28 and E89.4)d 74 35.4 2.09 (1.61 to 2.70) 10 (5 to 14) 3.8 Ovarian hypofunction (E28.3 and E89.4) 35 2.4 14.65 (8.29 to 25.86) 8 (5 to 11) Other diseases of the ovaries (E28.2, E28.8, and E28.9) 39 31.5 1.24 (0.88 to 1.74) 2 (−1 to 5) Testicular dysfunction (E29 and E89.5)d 182 4.5 40.24 (27.91 to 58.01) 45 (38 to 52) 17.1 Testicular hypofunction (E29.1 and E89.5) 175 2.3 75.12 (45.99 to 122.70) 44 (37 to 51) Other diseases of testis (E29.8 and E29.9) 10 2.4 4.22 (1.95 to 9.16) 2 (0 to 4) Other diseases of endocrine glands (E24, E31, E34, E35.8, E89.8, and E89.9) 70 14.7 4.78 (3.54 to 6.45) 14 (10 to 18) 5.3 Cushing syndrome (E24) 10 4.4 2.29 (1.13 to 4.66) 1 (0 to 3) Other endocrine diseases (E34, E35.8, E89.8, and E89.9) 59 9.9 5.94 (4.22 to 8.36) 12 (9 to 16) Table 2. c Observed and expected numbers of first hospital contacts for endocrine disease among 32 548 adolescent and young adult 1-year cancer survivors in Denmark, by 9 main diagnostic categories and 26 subcategories or diagnoses. d Numbers and rates for ovarian and testicular diseases are calculated for the entire cancer survivor and comparison cohorts even though these disorders occur only in women and men, respectively. Tenth revision; RR, hospitalization rate ratio. a Excluded diagnoses are nutritional and metabolic diseases (E40-E88 and E90), metabolic syndrome (E888C), congenital endocrine disorders (E00, E03.0, E03.1, and E25), disorders of puberty (E30), and diseases of thymus (E32). b Endocrine diseases of which there were fewer than 5 cases are not included even if the RR was high and/or statistically significant. Abbreviations: AER, absolute excess risk; ICD-10, International Classification of Diseases, Tenth revision; RR, hospitalization rate ratio. Results First Hospital Contacts for Endocrine Disease by Diagnostic Categories or Diagnosesa JAMA Network Open. 2018;1(2):e180349. doi:10.1001/jamanetworkopen.2018.0349 June 29, 2018 6/12 c Observed and expected numbers of first hospital contacts for endocrine disease among 32 548 adolescent and young adult 1-year cancer survivors in Denmark, by 9 main diagnostic categories and 26 subcategories or diagnoses. JAMA Network Open. 2018;1(2):e180349. doi:10.1001/jamanetworkopen.2018.0349 Discussion a Only endocrine diseases for which the RR had a low 95% confidence limit of 1 or greater and AERs with lower 95% confidence limit of 0 or greater were included. Endocrine disease categories of which there were fewer than 5 cases, however, were not included, even if the RR was high and/or statisticall significant. b Sample size indicate number of survivors. Table 3. Endocrine Disease by 10 Most Frequent Cancer Sitesa Site of Cancer and Endocrine Diseaseb Hospital Contacts, No. Discussion This nationwide, population-based cohort study showed that adolescent and young adult cancer survivors were at 73% higher risk for endocrine disease than the background population. The dominating endocrine diseases were thyroid diseases, testicular dysfunction, and diabetes. A particularly high AER for hypothyroidism was observed in Hodgkin lymphoma survivors. Treatment for Hodgkin lymphoma includes irradiation to the thyroid region; previous reviews concluded that half of Hodgkin lymphoma survivors who were irradiated subsequently experienced thyroid disease,10,19 and the reported risk increased with radiation dose. Alkylating chemotherapy and irradiation increased the risk for testicular dysfunction in a dose-dependent manner.10,19 In this study, brain cancer survivors had a broad spectrum of endocrine diseases, some of which might be secondary to hypothalamic and pituitary dysfunction. We found particularly high RRs and AERs for Table 3. Endocrine Disease by 10 Most Frequent Cancer Sitesa Site of Cancer and Endocrine Diseaseb Hospital Contacts, No. Discussion RR (95% CI) AER (95% CI) per 100 000 Person-Years Brain cancer (n = 1895) 116 3.03 (2.53 to 3.64) 490 (357 to 623) Pituitary hypofunction 40 112.01 (75.55 to 166.07) 244 (168 to 320) Other pituitary diseases 15 104.23 (55.19 to 196.85) 90 (44 to 137) Pituitary hyperfunction 5 11.28 (4.57 to 27.84) 28 (1 to 54) Thyrotoxicosis 14 2.24 (1.32 to 3.79) 47 (2 to 92) Diabetes (all types) 32 2.05 (1.44 to 2.90) 100 (32 to 168) Leukemia (n = 944) 63 3.97 (3.10 to 5.09) 755 (506 to 1004) Ovarian hypofunction 13 236.74 (121.17 to 462.56) 200 (91 to 309) Testicular hypofunction 8 155.26 (69.59 to 346.42) 122 (37 to 207) Pituitary hypofunction 8 52.56 (25.28 to 109.27) 121 (35 to 206) Type 2 diabetes 17 3.46 (2.14 to 5.57) 187 (62 to 313) Hodgkin lymphoma (n = 1713) 165 3.06 (2.62 to 3.57) 509 (393 to 624) Testicular hypofunction 5 27.67 (10.53 to 72.72) 21 (2 to 40) Hypothyroidism 87 14.89 (11.93 to 18.59) 362 (280 to 443) Diabetes, other and unspecified 11 2.68 (1.48 to 4.87) 30 (2 to 58) Goiter 32 2.46 (1.73 to 3.49) 83 (35 to 132) Type 2 diabetes 37 2.00 (1.45 to 2.77) 81 (29 to 134) Non-Hodgkin lymphoma (n = 1201) 66 1.86 (1.46 to 2.37) 237 (113 to 361) Ovarian hypofunction 9 134.34 (58.76 to 307.11) 67 (23 to 112) Testicular hypofunction 6 57.88 (23.58 to 142.07) 44 (8 to 81) Pituitary hypofunction 5 17.89 (7.19 to 44.48) 36 (3 to 69) Hypothyroidism 14 3.78 (2.23 to 6.42) 78 (22 to 134) Testis (n = 5503) 393 2.50 (2.25 to 2.78) 305 (254 to 356) Testicular hypofunction 148 134.71 (82.75 to 219.29) 186 (156 to 216) Adrenal hypofunction 21 21.31 (11.14 to 40.77) 25 (14 to 37) Testicular dysfunction, other and unspecified 8 7.23 (3.14 to 16.60) 9 (2 to 16) Thyrotoxicosis 24 2.21 (1.45 to 3.39) 17 (4 to 29) Type 1 diabetes 73 1.58 (1.24 to 2.00) 34 (12 to 55) Type 2 diabetes 143 1.50 (1.27 to 1.78) 61 (30 to 91) Ovary (n = 775) 44 1.14 (0.85 to 1.54) 56 (−78 to 189) Diabetes, other and unspecified 7 4.18 (1.98 to 8.86) 53 (1 to 104) Breast (n = 4654) 237 1.16 (1.02 to 1.32) 71 (7 to 136) Diseases of the thyroid gland 172 1.28 (1.10 to 1.49) 80 (26 to 135) Malignant melanoma (n = 5133) 225 1.08 (0.95 to 1.24) 30 (−21 to 82) Goiter 82 1.44 (1.16 to 1.80) 43 (13 to 73) Colon (n = 596) 14 0.68 (0.40 to 1.14) −111 (−232 to 11) Cervix (n = 3987) 224 0.94 (0.82 to 1.07) −26 (−80 to 28) Type 2 diabetes 73 1.31 (1.03 to 1.65) 31 (0 to 61) Abbreviations: AER, absolute excess risk, RR, hospitalization rate ratio. Downloaded From: https://jamanetwork.com/ by a Copenhagen University Library User on 03/29/2019 Downloaded From: https://jamanetwork.com/ by a Copenhagen University Library User on 03/29/2019 JAMA Network Open \| Oncology JAMA Network Open \| Oncology Endocrine Late Effects in Survivors of Cancer in Adolescence and Young Adulthood Discussion RR (95% CI) AER (95% CI) per 100 000 Person-Years Brain cancer (n = 1895) 116 3.03 (2.53 to 3.64) 490 (357 to 623) Pituitary hypofunction 40 112.01 (75.55 to 166.07) 244 (168 to 320) Other pituitary diseases 15 104.23 (55.19 to 196.85) 90 (44 to 137) Pituitary hyperfunction 5 11.28 (4.57 to 27.84) 28 (1 to 54) Thyrotoxicosis 14 2.24 (1.32 to 3.79) 47 (2 to 92) Diabetes (all types) 32 2.05 (1.44 to 2.90) 100 (32 to 168) Leukemia (n = 944) 63 3.97 (3.10 to 5.09) 755 (506 to 1004) Ovarian hypofunction 13 236.74 (121.17 to 462.56) 200 (91 to 309) Testicular hypofunction 8 155.26 (69.59 to 346.42) 122 (37 to 207) Pituitary hypofunction 8 52.56 (25.28 to 109.27) 121 (35 to 206) Type 2 diabetes 17 3.46 (2.14 to 5.57) 187 (62 to 313) Hodgkin lymphoma (n = 1713) 165 3.06 (2.62 to 3.57) 509 (393 to 624) Testicular hypofunction 5 27.67 (10.53 to 72.72) 21 (2 to 40) Hypothyroidism 87 14.89 (11.93 to 18.59) 362 (280 to 443) Diabetes, other and unspecified 11 2.68 (1.48 to 4.87) 30 (2 to 58) Goiter 32 2.46 (1.73 to 3.49) 83 (35 to 132) Type 2 diabetes 37 2.00 (1.45 to 2.77) 81 (29 to 134) Non-Hodgkin lymphoma (n = 1201) 66 1.86 (1.46 to 2.37) 237 (113 to 361) Ovarian hypofunction 9 134.34 (58.76 to 307.11) 67 (23 to 112) Testicular hypofunction 6 57.88 (23.58 to 142.07) 44 (8 to 81) Pituitary hypofunction 5 17.89 (7.19 to 44.48) 36 (3 to 69) Hypothyroidism 14 3.78 (2.23 to 6.42) 78 (22 to 134) Testis (n = 5503) 393 2.50 (2.25 to 2.78) 305 (254 to 356) Testicular hypofunction 148 134.71 (82.75 to 219.29) 186 (156 to 216) Adrenal hypofunction 21 21.31 (11.14 to 40.77) 25 (14 to 37) Testicular dysfunction, other and unspecified 8 7.23 (3.14 to 16.60) 9 (2 to 16) Thyrotoxicosis 24 2.21 (1.45 to 3.39) 17 (4 to 29) Type 1 diabetes 73 1.58 (1.24 to 2.00) 34 (12 to 55) Type 2 diabetes 143 1.50 (1.27 to 1.78) 61 (30 to 91) Ovary (n = 775) 44 1.14 (0.85 to 1.54) 56 (−78 to 189) Diabetes, other and unspecified 7 4.18 (1.98 to 8.86) 53 (1 to 104) Breast (n = 4654) 237 1.16 (1.02 to 1.32) 71 (7 to 136) Diseases of the thyroid gland 172 1.28 (1.10 to 1.49) 80 (26 to 135) Malignant melanoma (n = 5133) 225 1.08 (0.95 to 1.24) 30 (−21 to 82) Goiter 82 1.44 (1.16 to 1.80) 43 (13 to 73) Colon (n = 596) 14 0.68 (0.40 to 1.14) −111 (−232 to 11) Cervix (n = 3987) 224 0.94 (0.82 to 1.07) −26 (−80 to 28) Type 2 diabetes 73 1.31 (1.03 to 1.65) 31 (0 to 61) b Sample size indicate number of survivors. JAMA Network Open. 2018;1(2):e180349. doi:10.1001/jamanetworkopen.2018.0349 Discussion June 29, 2018 7/12 Downloaded From: https://jamanetwork.com/ by a Copenhagen University Library User on 03/29/2019 JAMA Network Open \| Oncology Endocrine Late Effects in Survivors of Cancer in Adolescence and Young Adulthood JAMA Network Open \| Oncology pituitary diseases in brain cancer survivors based on 40 observations of hypopituitarism, 5 observations of hyperpituitarism, and 15 observations of other pituitary diseases in 1895 brain cancer survivors. These numbers were lower than those in a British follow-up study of 56 adult nonpituitary brain cancer survivors who received radiation therapy at ages 21 to 45 years.20 Hypopituitarism was reported in 41% of survivors, and the risk was associated with radiation dose but not with age at radiation, sex, or chemotherapy.20 In this study, the risks for endocrine diseases were highest in leukemia survivors. In a study by Tauchmanovà et al,21 in preparation for bone marrow transplantation, some patients with leukemia underwent total body irradiation, which resulted in exceptionally high risks for gonadal dysfunction (95% of women and 47% of men), thyroid dysfunction (46%), and adrenal abnormalities (10%). We found a 40-fold increased risk for testicular dysfunction. Testicular damage is common in men receiving chemotherapy, and men receiving cytotoxic chemotherapy had a significantly lower testosterone level than healthy controls in a study of men with cancer diagnosed at a mean age of 28.6 years.22 Older age at cancer diagnosis and treatment was negatively associated with testosterone level, and testosterone levels were negatively correlated with fasting glucose, insulin, and body fat mass. As an example of adverse health effects of untreated endocrine diseases, we found increased risk of diabetes in testicular cancer survivors, ovarian cancer survivors, and cervical cancer survivors, which might be related to sex hormone deficiencies. Associations between hypogonadism and men with type 2 diabetes are well described, and the relationship is Figure. Hospitalization Rates for Any Endocrine Disease and Selected Endocrine Diseases Figure. Hospitalization Rates for Any Endocrine Disease and Selected Endocrine Diseases Figure. Hospitalization Rates for Any Endocrine Disease and Selected Endocrine Diseases Figure. Discussion A questionnaire- based survey of late effects in cervical cancer survivors showed significantly increased risk for diabetes in women treated with radiotherapy but not in those treated by surgery or by chemotherapy and surgery.24 bidirectional.23 Type 1 diabetes may be caused by direct damage to the pancreas. A questionnaire- based survey of late effects in cervical cancer survivors showed significantly increased risk for diabetes in women treated with radiotherapy but not in those treated by surgery or by chemotherapy and surgery.24 Of testicular cancer survivors, breast cancer survivors, malignant melanoma survivors, and cervical cancer survivors, which are the most common cancers in the cohort, only testicular cancer survivors and breast cancer survivors had significantly increased risk for hospital contact for endocrine disease. To our knowledge, no previous studies have evaluated the risk of endocrine diseases specifically for survivors of these cancers diagnosed in adolescence and young adulthood. The within-cohort analysis showed that sex modified the risk for endocrine diseases, with female survivors having the highest risks. The difference in the RRs of male and female survivors might reflect the sex differences seen in the general population or biological differences between men and women (eg, female patients appeared to be more vulnerable to adverse effects of cancer treatment than males, perhaps as a result of differences in oxidative stress and body composition).25 The patients with the most recent diagnoses of cancer (1990-2009) had a significantly higher risk for endocrine diseases than those with diagnoses before 1990. Surveillance bias might be more likely in the latest period, but the higher risk is more likely to be because of changes in treatment protocols and survival. Survival rates have improved dramatically over many years, resulting in increased risks for late effects. The strengths of this study include high statistical power because of the very large cohort of cancer survivors. We used nationwide health registries of high-quality dating several decades back, allowing long follow-up and virtually no loss to follow-up. To our knowledge, this is the first large- scale study of the risks for endocrine diseases after cancer in adolescence and young adulthood, with comparisons of risks for a broad range of well-defined endocrine diseases with those expected for a large, well-defined, population-based comparison cohort. Furthermore, we were able to evaluate patterns in endocrine late effects across the spectrum of cancer sites and other patient-related Table 4. Discussion Hospitalization Rates for Any Endocrine Disease and Selected Endocrine Diseases 2500 2000 Hospitalization Rate per 100 000 Person-Years 1500 1000 500 0 16-19 30-39 40-49 50-59 60-69 ≥70 Attained Age, y 20-29 All endocrine diseases A 0 800 600 Hospitalization Rate per 100 000 Person-Years 400 200 16-19 30-39 40-49 50-59 60-69 ≥70 ≥70 Attained Age, y 20-29 Hypothyroidism B 0 1600 1400 Hospitalization Rate per 100 000 Person-Years 1200 1000 800 600 400 200 16-19 30-39 40-49 50-59 60-69 ≥70 Attained Age, y 20-29 Diabetes C 0 250 200 Hospitalization Rate per 100 000 Person-Years 150 100 50 16-19 30-39 40-49 50-59 60-69 Attained Age, y 20-29 Pituitary hypofunction D Comparison Survivors Age-specific observed and expected hospitalization rates for any endocrine disease and selected endocrine diseases among 32 548 adolescent and young adult 1-year cancer survivors. 2500 2000 Hospitalization Rate per 100 000 Person-Years 1500 1000 500 0 16-19 30-39 40-49 50-59 60-69 ≥70 Attained Age, y 20-29 All endocrine diseases A Comparison Survivors 0 800 600 Hospitalization Rate per 100 000 Person-Years 400 200 16-19 30-39 40-49 50-59 60-69 ≥70 Attained Age, y 20-29 Hypothyroidism B ≥70 0 250 200 Hospitalization Rate per 100 000 Person-Years 150 100 50 16-19 30-39 40-49 50-59 60-69 Attained Age, y 20-29 Pituitary hypofunction D 0 1600 1400 Hospitalization Rate per 100 000 Person-Years 1200 1000 800 600 400 200 16-19 30-39 40-49 50-59 60-69 ≥70 Attained Age, y 20-29 Diabetes C Pituitary hypofunction D Age-specific observed and expected hospitalization rates for any endocrine disease and selected endocrine diseases among 32 548 adolescent and young adult 1-year cancer survivors Age-specific observed and expected hospitalization rates for any endocrine disease and selected endocrine diseases among 32 548 adolesce cancer survivors. Age-specific observed and expected hospitalization rates for any endocrine disease and selected endocrine diseases among 32 548 adolescent and young adult 1-year cancer survivors. JAMA Network Open. 2018;1(2):e180349. doi:10.1001/jamanetworkopen.2018.0349 June 29, 2018 8/12 Downloaded From: https://jamanetwork.com/ by a Copenhagen University Library User on 03/29/2019 Endocrine Late Effects in Survivors of Cancer in Adolescence and Young Adulthood JAMA Network Open \| Oncology bidirectional.23 Type 1 diabetes may be caused by direct damage to the pancreas. Limitations Our study had limitations. The limitations include lack of information of less severe conditions diagnosed and treated by general practitioners. Outpatient visits were included in the patient register in 1995, meaning that only the more severe cases that required hospitalization were included before that date. This results in underestimation of the number of cases. Our results might be influenced by surveillance bias, as cancer survivors are observed more closely in the health care system than the general population. This might have caused an overestimation of the risk estimates reported in this study. The Danish Cancer Registry has limited information on cancer stage and treatment, so associations between treatment factors and endocrine late effects cannot be conclusively determined. Discussion Adjusted Hospitalization Rate Ratios for Endocrine Disease in a Within-Survivor Cohort Analysis Characteristic First Hospitalizations, No. Adjusted HR (95% CI) P Value Sex <.001 Male 826 1 [Reference] NA Female 1303 1.85 (1.65-2.09) <.001 Age at cancer diagnosis, y <.001 15-19 139 1 [Reference] NA 20-24 248 0.89 (0.70-1.12) .32 25-29 362 0.67 (0.53-0.85) .001 30-34 568 0.63 (0.50-0.81) <.001 35-39 812 0.55 (0.43-0.71) <.001 Calendar period of cancer diagnosis, y <.001 1975-1989 1036 1 [Reference] NA 1990-2009 1093 1.97 (1.78-2.18) <.001 Site of cancera <.001 Leukemia (n = 944) 63 1.36 (1.00-1.85) .05 Hodgkin lymphoma (n = 1713) 165 1.05 (0.83-1.33) .70 Brain cancer (n = 1895) 116 1 [Reference] NA Testicular cancer (n = 5503) 393 0.92 (0.74-1.15) .45 Non-Hodgkin lymphoma (n = 1201) 66 0.69 (0.51-0.93) .02 Breast cancer (n = 4654) 237 0.48 (0.38-0.61) <.001 Ovarian cancer (n = 775) 44 0.45 (0.32-0.64) <.001 Malignant melanoma (n = 5133) 225 0.43 (0.34-0.54) <.001 Cervical cancer (n = 3987) 224 0.39 (0.31-0.49) <.001 Colon cancer (n = 596) 14 0.28 (0.16-0.49) <.001 Abbreviatio a Sample s JAMA N t k O 2018 1(2) 180349 d i 101001/j t k 2018 0349 Table 4. Adjusted Hospitalization Rate Ratios for Endocrine Disease in a Within-Survivor Cohort Analysis June 29, 2018 9/12 Downloaded From: https://jamanetwork.com/ by a Copenhagen University Library User on 03/29/2019 factors. The inclusion of all 1-year adolescent and young adult cancer survivors diagnosed since 1976 minimizes possible selection bias. Only medically verified diagnoses of endocrine diseases were included, which ensures more correct estimates than those in studies based on self-reported data. JAMA Network Open \| Oncology Endocrine Late Effects in Surv Endocrine Late Effects in Survivors of Cancer in Adolescence and Young Adulthood factors. The inclusion of all 1-year adolescent and young adult cancer survivors diagnosed since 1976 minimizes possible selection bias. Only medically verified diagnoses of endocrine diseases were included, which ensures more correct estimates than those in studies based on self-reported data. Conclusions Although increased risks for a wide range of cardiovascular diseases,26,27 secondary malignancies,28,29 infectious diseases, and digestive diseases29,30 have been reported in cancer survivors, late effects in adolescent and young adult cancer survivors have received little attention. Our study provides new, accurate, and detailed information about these survivors and important clinical information on how the risks for such late effects are modified by patient factors. This is the first step in identifying patients who are at risk for endocrine late effects so that individual profiles can be drawn up to assess the probable risk for endocrine disease. This will require prospective capture and close phenotyping of these diseases. We hope that this study will inspire investigators in future studies to determine exact associations between treatment regimens and endocrine disease and ultimately incorporate them into individual, customized treatment plans. Each future adolescent and young adult cancer patient should be offered the least deleterious treatment to ensure high quality of life after cancer while maintaining the good cure rates. Cure has become an insufficient goal. June 29, 2018 10/12 ARTICLE INFORMATION Published: June 29, 2018. doi:10.1001/jamanetworkopen.2018.0349 Correction: This article was corrected on July 27, 2018, to fix an error in the text. Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2018 Jensen MV et al. JAMA Network Open. Corresponding Author: Mette Vestergaard Jensen, MD, Survivorship Unit, Danish Cancer Society Research Center, Strandboulevarden 49, DK-2900 Copenhagen, Denmark (mettev87@gmail.com). Corresponding Author: Mette Vestergaard Jensen, MD, Survivorship Unit, Danish Cancer Society Research Center, Strandboulevarden 49, DK-2900 Copenhagen, Denmark (mettev87@gmail.com). Author Affiliations: Survivorship Unit, Danish Cancer Society Research Center, Copenhagen, Denmark (Jensen, Rugbjerg, de Fine Licht, Johansen, Winther); Department of Clinical Medicine, Juliane Marie Centre, Rigshospitalet, Copenhagen, Denmark (Johansen, Schmiegelow); Statistics, Bioinformatics, and Registries, Danish Cancer Society Research Center, Copenhagen, Denmark (Andersen); Department of Clinical Medicine, Faculty of Health, Aarhus University, Aarhus, Denmark (Winther). Rugbjerg, de Fine Licht, Johansen, Winther); Department of Clinical Medicine, Juliane Marie Centre, Rigshospitalet, Copenhagen, Denmark (Johansen, Schmiegelow); Statistics, Bioinformatics, and Registries, Danish Cancer Society Research Center, Copenhagen, Denmark (Andersen); Department of Clinical Medicine, Faculty of Health, Aarhus University, Aarhus, Denmark (Winther). Author Contributions: Drs Jensen and Rugbjerg had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Concept and design: Jensen, Rugbjerg, Winther. Acquisition, analysis, or interpretation of data: All authors. Drafting of the manuscript: Jensen, Winther. Critical revision of the manuscript for important intellectual content: All authors. JAMA Network Open. 2018;1(2):e180349. doi:10.1001/jamanetworkopen.2018.0349 June 29, 2018 10/12 JAMA Network Open \| Oncology Statistical analysis: Jensen, Andersen. Obtained funding: Jensen, Johansen, Winther. Supervision: Rugbjerg, de Fine Licht, Johansen, Schmiegelow, Winther. Conflict of Interest Disclosures: None reported. Conflict of Interest Disclosures: None reported. Funding/Support: This study was supported by grant No. R117-A7242-14-S7 (Dr Jensen) from the Danish Cancer Society. Role of the Funder/Sponsor: The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. Role of the Funder/Sponsor: The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. Additional Contributions: We thank data managers Anja Krøyer, MSc, and Andrea Bautz, BSc, Danish Cancer Society Research Center, for programming and analyzing the data used in this article. They were not compensated aside from their salary. REFERENCES 1. Shaw PH, Reed DR, Yeager N, Zebrack B, Castellino SM, Bleyer A. Adolescent and young adult (AYA) oncology in the United States: a specialty in its late adolescence. J Pediatr Hematol Oncol. 2015;37(3):161-169. 2. National Institutes of Health. Closing the gap: research and care imperatives for adolescents and young adults with cancer: report of the Adolescent and Young Adult Oncology Progress Review Group. https://www.cancer. gov/types/aya/research/ayao-august-2006.pdf. Accessed March 30, 2018. 3. Coccia PF, Altman J, Bhatia S, et al. Adolescent and young adult oncology: clinical practice guidelines in oncology. J Natl Compr Canc Netw. 2012;10(9):1112-1150. 4. Coccia PF, Pappo AS, Altman J, et al. Adolescent and young adult oncology, version 2.2014. J Natl Compr Canc Netw. 2014;12(1):21-32. 5. Bleyer A, Barr R, Hayes-Lattin B, Thomas D, Ellis C, Anderson B; Biology and Clinical Trials Subgroups of the US National Cancer Institute Progress Review Group in Adolescent and Young Adult Oncology. The distinctive biology of cancer in adolescents and young adults. Nat Rev Cancer. 2008;8(4):288-298. 6. Woodward E, Jessop M, Glaser A, Stark D. Late effects in survivors of teenage and young adult cancer: does age matter? Ann Oncol. 2011;22(12):2561-2568. 7. Oeffinger KC, Mertens AC, Sklar CA, et al; Childhood Cancer Survivor Study. Chronic health conditions in adult survivors of childhood cancer. N Engl J Med. 2006;355(15):1572-1582. 8. Abouassaly R, Fossa SD, Giwercman A, et al. Sequelae of treatment in long-term survivors of testis cancer. Eur Urol. 2011;60(3):516-526. 9. De Bruin ML, Huisbrink J, Hauptmann M, et al. Treatment-related risk factors for premature menopause following Hodgkin lymphoma. Blood. 2008;111(1):101-108. 10. Ng AK, Mauch PM. Late effects of Hodgkin’s disease and its treatment. Cancer J. 2009;15(2):164-168. 11. Hancock SL, Cox RS, McDougall IR. Thyroid diseases after treatment of Hodgkin’s disease. N Engl J Med. 1991; 325(9):599-605. 12. Ewertz M, Jensen AB. Late effects of breast cancer treatment and potentials for rehabilitation. Acta Oncol. 2011;50(2):187-193. 13. Gjerstorff ML. The Danish Cancer Registry. Scand J Public Health. 2011;39(7, suppl):42-45. 14. Pedersen CB. The Danish Civil Registration System. Scand J Public Health. 2011;39(7, suppl)(suppl):22-25. 15. Schmidt M, Pedersen L, Sørensen HT. The Danish Civil Registration System as a tool in epidemiology. Eur J Epidemiol. 2014;29(8):541-549. 16. Lynge E, Sandegaard JL, Rebolj M. The Danish National Patient Register. Scand J Public Health. 2011;39(7, suppl):30-33. 17. Fieller EC. Some problems in interval estimation. J R Stat Soc Series B Stat Methodol. 1954 18. Downloaded From: https://jamanetwork.com/ by a Copenhagen University Library User on 03/29/2019 Endocrine Late Effects in Survivors of Cancer in Adolescence and Young Adulthood June 29, 2018 11/12 REFERENCES Rosthøj S, Andersen PK, Abildstrom SZ. SAS macros for estimation of the cumulative incidence functions based on a Cox regression model for competing risks survival data. Comput Methods Programs Biomed. 2004;74 (1):69-75. 19. Adams MJ, Constine LS, Lipshultz SE. Late effects of therapy for Hodgkin’s lymphoma. Curr Hematol Malig Rep. 2007;2(3):143-150. 20. Agha A, Sherlock M, Brennan S, et al. Hypothalamic-pituitary dysfunction after irradiation of nonpituitary brain tumors in adults. J Clin Endocrinol Metab. 2005;90(12):6355-6360. JAMA Network Open. 2018;1(2):e180349. doi:10.1001/jamanetworkopen.2018.0349 June 29, 2018 11/12 JAMA Network Open \| Oncology 21. Tauchmanovà L, Selleri C, Rosa GD, et al. High prevalence of endocrine dysfunction in long-term survivors after allogeneic bone marrow transplantation for hematologic diseases. Cancer. 2002;95(5):1076-1084. 22. Greenfield DM, Walters SJ, Coleman RE, et al. Prevalence and consequences of androgen deficiency in young male cancer survivors in a controlled cross-sectional study. J Clin Endocrinol Metab. 2007;92(9):3476-3482. 23. Ebrahimi F, Christ-Crain M. Metabolic syndrome and hypogonadism—two peas in a pod. Swiss Med Wkly. 2016;146:w14283. 24. Shin DW, Nam JH, Kwon YC, et al. Comorbidity in disease-free survivors of cervical cancer compared with the general female population. Oncology. 2008;74(3-4):207-215. 25. Lipshultz SE, Lipsitz SR, Mone SM, et al. Female sex and higher drug dose as risk factors for late cardiotoxic effects of doxorubicin therapy for childhood cancer. N Engl J Med. 1995;332(26):1738-1743. 26. Rugbjerg K, Mellemkjaer L, Boice JD, Køber L, Ewertz M, Olsen JH. Cardiovascular disease in survivors of adolescent and young adult cancer: a Danish cohort study, 1943-2009. J Natl Cancer Inst. 2014;106(6):dju110- dju110. 27. Kero AE, Järvelä LS, Arola M, et al. Cardiovascular morbidity in long-term survivors of early-onset cancer: a population-based study. Int J Cancer. 2014;134(3):664-673. 28. Lee JS, DuBois SG, Coccia PF, Bleyer A, Olin RL, Goldsby RE. Increased risk of second malignant neoplasms in adolescents and young adults with cancer. Cancer. 2016;122(1):116-123. 28. Lee JS, DuBois SG, Coccia PF, Bleyer A, Olin RL, Goldsby RE. Increased risk of second malignant neoplasms in adolescents and young adults with cancer. Cancer. 2016;122(1):116-123. 29. Youn P, Milano MT, Constine LS, Travis LB. Long-term cause-specific mortality in survivors of adolescent and young adult bone and soft tissue sarcoma: a population-based study of 28,844 patients. Cancer. 2014;120(15): 2334-2342. 29. Youn P, Milano MT, Constine LS, Travis LB. Long-term cause-specific mortality in survivors of adolescent and young adult bone and soft tissue sarcoma: a population-based study of 28,844 patients. Cancer. 2014;120(15): 2334-2342. 30. Rugbjerg K, Olsen JH. Long-term risk of hospitalization for somatic diseases in survivors of adolescent or young adult cancer. JAMA Oncol. 2016;2(2):193-200. 30. Rugbjerg K, Olsen JH. Long-term risk of hospitalization for somatic diseases in survivors of adolescent or young adult cancer. JAMA Oncol. 2016;2(2):193-200. June 29, 2018 12/12 Downloaded From: https://jamanetwork.com/ by a Copenhagen University Library User on 03/29/2019 Downloaded From: https://jamanetwork.com/ by a Copenhagen University Library User on 03/29/2019 Endocrine Late Effects in Survivors of Cancer in Adolescence and Young Adulthood SUPPLEMENT. eFigure 1. Flowchart of the Adolescent and Young Adult Cancer Survivor Cohort With a First Primary Cancer Diagnosed From 1976 to 2009 At Ages 15-39 Years and of Population Comparisons eFigure 2. Rate Ratios for Hospital Contacts for Any Endocrine Disorder in Survivors of Cancer At the 10 Most Frequent Cancer Sites in Adolescents and Young Adults eFigure 2. Rate Ratios for Hospital Contacts for Any Endocrine Disorder in Survivors of Cancer At the 10 Most Frequent Cancer Sites in Adolescents and Young Adults June 29, 2018 12/12 JAMA Network Open. 2018;1(2):e180349. doi:10.1001/jamanetworkopen.2018.0349 Downloaded From: https://jamanetwork.com/ by a Copenhagen University Library User on 03/29/2019
https://openalex.org/W3037092817	https://www.scielo.br/j/sa/a/4tm67qbtjPPT7P7jhBMGxQF/?lang=en&format=pdf	English	null	Defoliation intensity and leaf area index recovery in defoliated swards: implications for forage accumulation	Scientia Agrícola	2,021	cc-by	6,340	Defoliation intensity and leaf area index recovery in defoliated swards: implications for forage accumulation Medeiros Martins , Daniel Schmitt* , Paulo Gonçalves Duchini , Tiago Miqueloto , André Fischer Sbrissia Clóvis David Medeiros Martins , Daniel Schmitt* , Paulo Gonçalves Duchini , Tiago Miqueloto , André Fisc Universidade do Estado de Santa Catarina/CAV – Depto. de Produção Animal e Alimentos, Av. Luiz de Camões, 2090 – 88520-000 – Lages, SC – Brasil. Corresponding author <daniel.schmitt@veterinario.med.br> Edited by: Harley D. Naumann Received April 04, 2019 Accepted July 11, 2019 ABSTRACT: This study evaluated the leaf area index (LAI) recovery mechanisms and forage accumulation rates on the regrowth of different grass species subjected to different defoliation intensities. For that purpose, plots of Pennisetum clandestinum (kikuyugrass), Lolium multiflorum (annual ryegrass), and Avena strigosa (black oats) were defoliated from 20 to 80 % of their initial heights (25, 20, and 25 cm, respectively). At different increments in height, forage samples were collected to ground level and used to estimate tiller population density (TPD), leaf area per tiller (LA), and forage mass. From these data, we calculated the leaf area index (LAI), average leaf area index (aLAI), and average and instantaneous forage accumulation rate (FAR and IFAR, respectively). Data were plotted over time (days) to describe LAI recovery and forage accumulation rates. As the defoliation intensity increased, greater canopy heights were needed for pastures to achieve their maximum forage accumulation rates, which required longer regrowth intervals. The need for high tiller recruitment after defoliation, which delayed canopy LAI recovery, seemed to be one of the main cause. Thus, grazing management strategies that involve costly tiller recruitment could decrease both overall forage production and sward persistence over time. However, the plant ability to recover LAI after successive intense defoliations seems to be species-dependent and related to their phenotypic plasticity. Universidade do Estado de Santa Catarina/CAV – Depto. de Produção Animal e Alimentos, Av. Luiz de Camões, 2090 – 88520-000 – Lages, SC – Brasil. Corresponding author <daniel.schmitt@veterinario.med.br> Edited by: Harley D. Naumann Received April 04, 2019 Accepted July 11, 2019 Edited by: Harley D. Naumann Edited by: Harley D. Naumann Received April 04, 2019 Accepted July 11, 2019 Keywords: grazing heights, grazing management, intermittent stocking, sward targets Animal Science and Pastures DOI: http://dx.doi.org/10.1590/1678-992X-2019-0095 Research Article Introduction convergent findings. This is because defoliation intensities are frequently generated from different pre- defoliation targets, which tend to favor targets created with better criteria (e.g., critical LAI; Barbosa et al., 2007; Silva et al., 2009; Zanini et al., 2012). In addition, most studies did not assess possible interspecific variations in LAI recovery and in forage mass during pasture regrowth. Therefore, this study assessed LAI recovery mechanisms and forage accumulation rates throughout the regrowth of different grass species subjected to different defoliation intensities. The central hypothesis tested was that increasing the defoliation intensity reduces the productive potential of pastures and that it is a species-dependent response. The critical leaf area index ([LAI]; Brougham, 1958) has been suggested as the upper limit to interrupt regrowth of pastures under intermittent stocking method (Korte et al., 1982; Carnevalli et al., 2006; Sbrissia et al., 2018). However, the effects of residual LAI on pasture productive capacity and its relationship with defoliation management heights are still unclear. Theoretical models developed by Parsons et al. (1988) indicate that heavily defoliated pastures reach maximum growth rates later than leniently defoliated pastures do. According to Chapman (2016), intense defoliations would result in longer regrowth intervals, which could reduce total forage production. Materials and Methods The LAI recovery after defoliation occurs through leaf growth from remaining tillers and/or the emergence of new tillers (Lemaire and Chapman, 1996). High defoliation intensities can stimulate initial LAI increments through the recruitment of new tillers due to the improved light quality reaching the plant base (Gautier et al., 1999) and apex meristem decapitation (McSteen, 2009). However, frequent recruitment of new tillers can affect forage accumulation; thus, an appropriate management strategy should minimize tillering dependence (Nelson and Zarrough, 1981). Moreover, forage plants are incapable to maintain a high growth potential after defoliation, despite the possibility of mobilizing carbohydrate reserves and of changes in a specific leaf area (Irving, 2015; Chapman, 2016). Three experiments (I, II, and III) were conducted using different grass species in Lages, Santa Catarina State, Brazil (27°47’ S, 50°18’ W, altitude of 913 m), between Jan 2014 and Sept 2015. According to the Köppen classification system, the region has a Cfb climate (humid subtropical under oceanic influences), with cold winters, mild summers, and well-distributed rainfall throughout the year (Alvares et al., 2013). Weather data from the experimental period are presented in Table 1. The soil in the experimental area is classified as Clayey Inceptisol (Soil Survey Staff, 1999). Before the beginning of each experiment, soil samples were collected from 0-20 cm depth and submitted to the chemical analysis. Liming and fertilization were performed in order to reach a baseline pH of about 6.0 and to supply adequate levels Research data regarding the effects of defoliation intensity on forage production do not always present Sci. Agric. v.78, n.2, e20190095, 2021 Productive capacity of defoliated swards Martins et al. Table 1 – Average monthly climatic conditions recorded during the experimental months of 2014 and 2015 in Lages, SC, Brazil. 2014 Month Jan Feb Mar Apr May June July Aug Sept Oct Sunlight (h) 216.8 218.4 163.7 125.5 131.5 85.7 106.7 204.7 120.2 131.5 Precipitation (mm) 183.0 210.7 121.3 94.2 235.3 397.8 77.3 61.0 233.4 235.3 Maximum temp. (°C) 27.9 28.3 24.2 21.6 18.9 16.4 16.6 20.6 20.6 18.9 Average temp. (°C) 21.5 21 18.4 16.3 13.2 12.0 11.3 12.7 15.1 13.2 Minimum temp. Experiment I (kikuyugrass) p ( y g ) This experiment was conducted in a well- established (~ 20-year-old) 1.8-ha kikuyugrass (Pennisetum clandestinum Hochst ex. Chiov) pasture. The experimental treatments consisted of four defoliation intensities (lowering 40, 50, 60, or 70 % of the initial height) applied from a recommended pre-cutting height of 25 cm (Sbrissia et al., 2018) and replicated three times in a randomized block design (12 plots of 1500 m2). The experimental treatments were established as follows: i) in early Dec 2013, the entire area was mowed to a residual height of 10 cm (to standardize the pre-experimental conditions), fertilized (Table 2), and plots were delimitated; ii) when canopies reached 25 cm in height (from 14 Jan 2014 onward), grazing was performed by cattle to the intended residual heights (Table 3), defining the beginning of the data collection period. The canopy heights were measured with a sward stick (Barthram, 1985) at 50 points per plot after grazing and twice a week during the regrowth period. 2014 into 16 plots of 1 m2. The experimental treatments comprised four defoliation intensities (lowering 20, 40, 60, or 80 % of the initial height) applied from a recommended pre-cutting height of 20 cm (Santos et al., 2016) and replicated four times in a completely randomized design. Four weeks after sowing, canopies were approximately 20 cm tall. They were first lowered to 10-cm residual height to standardize the pre-experimental conditions and favor pasture establishment. On 09 Aug 2014, the canopies were 20 cm tall and were lowered to the intended residual heights by manual scissors trimming (Table 3), stablishing the beginning of the data collection period. The canopy heights were measured with a sward stick (Barthram, 1985) at 15 points per plot twice a week during plot establishment, and daily during the data collection period. 2014 into 16 plots of 1 m2. The experimental treatments comprised four defoliation intensities (lowering 20, 40, 60, or 80 % of the initial height) applied from a recommended pre-cutting height of 20 cm (Santos et al., 2016) and replicated four times in a completely randomized design. Four weeks after sowing, canopies were approximately 20 cm tall. They were first lowered to 10-cm residual height to standardize the pre-experimental conditions and favor pasture establishment. Materials and Methods (°C) 16.3 15.3 14.0 12.2 8.8 8.1 6.6 6.3 10.4 8.8 2015 Jan Feb Mar Apr May June July Aug Sept Oct Sunlight (h) 197.0 163.0 174.8 147.9 106.7 127.4 114.0 180.2 147.1 96.3 Precipitation (mm) 211.9 133.9 135.6 99.7 71.0 134.9 264.0 69.3 334.6 363.2 Maximum temp. (°C) 27.3 26.6 25.3 22.2 18.7 17.4 17.4 22.3 21.3 21.0 Average temp. (°C) 20.9 20.1 19.2 16.4 13.9 11.4 12.4 15.5 15.0 16.0 Minimum temp. (°C) 15.7 15.3 14.6 12.1 10.2 6.4 7.7 10.1 10.2 11.8 Source: National Institute of Meteorology (INMET, 2017). Table 2 – Application dates, fertilizer types, and amounts of phosphorus, potassium, calcium, and nitrogen applied during the three experiments. of N, P, and K (Table 2), according to the soil analysis and the Fertilizing and Liming Manual for the states of Rio Grande do Sul and Santa Catarina (CQFS, 2004). Experiment Date Element Fertilizer Amount kg ha–1 I 16 Dec 2013 P TSP 80 kg P2O5 I 16 Dec 2013 K KCl 80 kg K2O I 14 Jan 2014 N Urea 90 kg N I 10 Mar 2014 N Urea 90 kg N II 17 Apr 2014 Ca CaL 20 kg CaO II 25 May 2014 P TSP 80 kg P2O5 II 25 May 2014 K KCl 80 kg K2O II 10 June 2014 N Urea 200 kg N II 09 Aug 2014 N Urea 200 kg N III 15 Feb 2015 Ca CaL 30 kg CaO III 08 Apr 2015 P TSP 80 kg P2O5 III 08 Apr 2015 K KCl 80 kg K2O III 25 May 2015 N Urea 150 kg N TSP = triple superphosphate; KCl = potassium chloride; CaL = calcitic limestone (RTNP 98 %). Experiment I (kikuyugrass) On 09 Aug 2014, the canopies were 20 cm tall and were lowered to the intended residual heights by manual scissors trimming (Table 3), stablishing the beginning of the data collection period. The canopy heights were measured with a sward stick (Barthram, 1985) at 15 points per plot twice a week during plot establishment, and daily during the data collection period. Data were collected throughout two regrowth cycles (from 14 Jan 2014 to 20 Apr 2014). After each grazing event, three areas of 1 m2 per plot were marked with wooden stakes. From these areas, forage samples were collected at five points throughout the regrowth period as follows: i) after grazing; ii) during regrowth (increments of 25, 50, and 75 % of height); and iii) at pre-grazing (25 cm). The heights of each sampling are presented in Table 3. On all occasions, clipping was performed to ground level, using quadrats of 0.0625 m2 and scissors, and samples were taken to the laboratory for immediate processing. Productive capacity of defoliated swards Productive capacity of defoliated swards Martins et al. Table 3 – Cutting heights* (cm) from the residue in different defoliation intensity for kikuyugrass, annual ryegrass, and black oat pastures. Experiment I (kikuyugrass) II (annual ryegrass) III (black oat) Defoliation intensity 40 % 50 % 60 % 70 % 20 % 40 % 60 % 80 % 40 % 60 % 80 % Cutting heights - - - 7.5 - - - 4 - - 5 - - 10 11.9 - - 8 8 - 10 10 - 12.5 13.8 16.3 - 12 12 12 15 15 15 15 15.6 17.6 20.7 16 16 16 16 20 20 20 17.5 18.7 21.2 25 20 20 20 20 25 25 25 20.0 21.8 25 - 24 24 24 24 30 30 30 22.5 25 - - 28 28 28 28 35 35 35 25 - - - 32 32 32 32 40 40 40 - - - - 36 36 36 36 - - - - - - - 40 40 40 40 - - - Pre-cutting canopy height targets for each species are in bold. Pre-cutting canopy height targets for each species are in bold. Experiment III (black oat) Black oat ‘IAPAR 61’ (Avena strigosa Schreb.) was broadcast-sown (90 kg seeds ha–1) on 15 Apr 2015 into nine plots of 2.8 m2. The experimental treatments consisted of three defoliation intensities (lowering 40, 60, and 80 % of the initial height) applied from a recommended pre-cutting height of 25 cm (Guzatti et al., 2015; Duchini et al., 2014) and replicated three times in a complete randomized design. On 8 May 2015, canopies were trimmed to a residual height of 10 cm to standardize the pre-experimental conditions and favor pasture establishment. After reaching a height of 25 cm (25 May 2015), pastures were lowered for the first time to the intended residual heights (Table 3). From that point, as canopies reached 25 cm in height (from 06 June 2015 onward), another lowering to the intended residual heights was performed, starting a new data collection period. The canopy heights were measured with a sward stick (Barthram, 1985) at 15 points per plot, twice a week during the plots establishment, and daily during the data collection period. From these data, the LAI, the average LAI (aLAI), average forage accumulation rate (FAR; kg DM ha–1 d–1), and instantaneous forage accumulation rate (IFAR; kg DM ha–1 d–1) were calculated as follows: LAI = TPD × LA aLAI LAI LAI i = +     0 2 FAR FM FM t i −     0 IFAR FM FM t i i t = −     − where: LAIi represents the leaf area index at some specific point during regrowth; LAI0 is the leaf area index at the beginning of the regrowth period; FMi is the forage mass (kg DM ha–1) at some specific point during regrowth; FMi–1 is the forage mass (kg DM ha–1) recorded before the FMi; FM0 is the forage mass (kg DM ha–1) at the beginning of the regrowth period; and t is the time interval (days) between FM assessments. Thus, (FMi – FM0) represents the net forage mass variation between the beginning and some specific point during regrowth, and (FMi – FMi–1) equals the net forage mass variation between subsequent increments in height (Table 3). Laboratory assessments 40 cm in height. This procedure was performed in order to detect, at some point during regrowth, maximum forage accumulation rates (Parsons et al., 1988). Forage samplings were performed at every 4 cm increment in height throughout regrowth, from the residue until the canopy reached 40 cm (Table 3). Forage samples were collected to ground level using quadrats of 0.03 m2 and scissors, and were taken to laboratory for immediate processing. The laboratory assessments for experiments I, II, and III followed the same protocol. The number of tillers in each forage sample was counted to estimate the tiller population density (TPD; tillers m–2). Then, green leaf blades from 50 tillers chosen randomly per sample were detached and scanned in a leaf area meter. The results (cm2) were divided by 50 to estimate leaf area per tiller (LA; cm2 per tiller). After TPD and LA were estimated, each sample (i.e., 50 tillers + the remaining material) was separated into plant parts (green leaf blades, stems + sheath, and senescent material), and dried in a forced- air oven at 65 °C for 72 h. The dry weights of these fractions were used to calculate total forage mass (kg DM ha–1). Experiment II (annual ryegrass) Annual ryegrass ‘BRS Ponteio’ (Lolium multiflorum Lam.) was broadcast-sown (90 kg seeds ha–1) on 24 June Data were collected throughout a single regrowth cycle from 09 Aug 2014 until the canopies achieved 2 Sci. Agric. v.78, n.2, e20190095, 2021 Results In general, increments in defoliation intensity had negative effects on aLAICHT and FARCHT (Table 4); however, kikuyugrass maintained similar levels when lowered by 40 or 50 % of initial height. Similarly, canopies under greater defoliation intensities (≥ 40 % for annual ryegrass and > 50 % for black oats and kikuyugrass) had a lower FARmax and reached that condition above the CHT (CHFARmax; Table 4). The time required for canopies to reach IFARmax and FARmax rose as defoliation intensity increased (Figure 1A-F); however, kikuyugrass did not reach either FARmax or IFARmax when defoliated at 70 %, since the protocol used for this species comprised assessments of only up to 25 cm (Table 4). Means followed by the same upper case letter in rows are not different by the Tukey-Kramer test (p > 0.05). Abbreviations: aLAICHT = average LAI when pastures reached the canopy height target; CHFARmax = canopy height to reach maximum forage accumulation rate (cm); FARCHT = average forage accumulation rate when pastures reached the canopy height target (kg DM ha–1 d–1); FARmax = maximum average forage accumulation rate (kg DM ha–1 d–1); DaysFARmax = days needed to reach maximum average forage accumulation rate; SEM = standard error of the mean. Did not reach the CHFARmax. residual LAI, as residual LAI determines the amount of light intercepted by the canopy after defoliation and, consequently, the leaf area replacement and organic reserves usages during regrowth (Donaghy and Fulkerson, 1998). Another factor that might have contributed to these decreases was the initial LAI recovery pattern, since the initial TPD increase is associated to the emergence of new tillers, as they are light and have a low LA, they have little contribution to the LAI and forage accumulation (Tainton, 1974). Thus, the increase in defoliation intensity triggered greater tissue renewal from low residual LAI (low pasture photosynthetic capacity), reducing FARCHT. The LAI recovery in canopies under intense defoliation (> 50 %) depended on TPD and LA increments, mainly for kikuyugrass and annual ryegrass (Figures 2A- I). Black oat was less capable to recruit new tillers to recover LAI, even after intense defoliations (Figures 2G and 2I). Less-intense defoliations (≤ 50 %) resulted in LAI recovery through LA increments for all of the studied species (Figures 2B, 2E and 2H). Experiment III (black oat) where: LAIi represents the leaf area index at some specific point during regrowth; LAI0 is the leaf area index at the beginning of the regrowth period; FMi is the forage mass (kg DM ha–1) at some specific point during regrowth; FMi–1 is the forage mass (kg DM ha–1) recorded before the FMi; FM0 is the forage mass (kg DM ha–1) at the beginning of the regrowth period; and t is the time interval (days) between FM assessments. Thus, (FMi – FM0) represents the net forage mass variation between the beginning and some specific point during regrowth, and (FMi – FMi–1) equals the net forage mass variation between subsequent increments in height (Table 3). Similar to Experiment II, data were collected throughout a single regrowth cycle from 06 June 2015 (residual heights) until the canopies achieved 40 cm in height (as mentioned above). Forage samplings were performed at every 4 cm increment in height throughout regrowth, from the residue until the canopy reached 40 cm in height (Table 3). Forage samples were collected to ground level using quadrats of 0.06 m2 and scissors, and were taken to laboratory for immediate processing. 3 Sci. Agric. v.78, n.2, e20190095, 2021 Productive capacity of defoliated swards Martins et al. Table 4 – Variables related to canopy height and forage accumulation rate of kikuyugrass, annual ryegrass, and black oat pastures subjected to different defoliation intensities. Experiment III (black oat) Experiment I (Kikuyugrass) Variable Defoliation intensity SEM p-value 40 % 50 % 60 % 70 % CHFARmax 20 22 21 +25 - - aLAICHT 3.9A 3.9A 2.7B 2.6B 0.25 < 0.001 FARCHT 149A 147A 108B 97B 15.4 < 0.001 FARmax 165A 177A 140B 97C 10.7 < 0.001 DaysFARmax 9A 13A 19B 43C 2 < 0.001 Experiment II (Annual ryegrass) Variable Defoliation intensity SEM p-value 20 % 40 % 60 % 80 % CHFARmax 24 20 24 32 - - aLAICHT 4.0A 2.9B 2.0C 1.5D 3.3 < 0.001 FARCHT 126A 100B 68C 50D 10.5 < 0.001 FARmax 158A 100B 108B 109B 6.4 < 0.001 DaysFARmax 3A 3A 8B 26C 2 < 0.001 Experiment III (Black oat) Variable Defoliation intensity SEM p-value 40 % 60 % 80 % CHFARmax 25 30 35 - - aLAICHT 3.5A 2.5B 2.2C 0.3 < 0.001 FARCHT 149A 57B 45C 9.8 < 0.001 FARmax 149A 83B 80B 8.2 < 0.001 DaysFARmax 9A 21B 39C 3 < 0.001 Means followed by the same upper case letter in rows are not different by the Tukey-Kramer test (p > 0.05). Abbreviations: aLAICHT = average LAI when pastures reached the canopy height target; CHFARmax = canopy height to reach maximum forage accumulation rate (cm); FARCHT = average forage accumulation rate when pastures reached the canopy height target (kg DM ha–1 d–1); FARmax = maximum average forage accumulation rate (kg DM ha–1 d–1); DaysFARmax = days needed to reach maximum average forage accumulation rate; SEM = standard error of the mean. *Did not reach the CHFARmax. Table 4 – Variables related to canopy height and forage accumulation rate of kikuyugrass, annual ryegrass, and black oat pastures subjected to different defoliation intensities. Finally, the values for maximum average forage accumulation rate (FARmax, kg DM ha–1 d–1), maximum instantaneous forage accumulation rate (IFARmax, kg DM ha–1 d–1), canopy height to reach FARmax (CHFARmax, cm), average forage accumulation rate (FARCHT, kg DM ha–1 d–1), and average LAI (aLAICHT) when pastures reached the canopy height target, and the time (in days) to reach IFARmax and FARmax were determined for each species- defoliation intensity combinations by plotting the data against days at regrowth. Statistical analyses Data were analyzed separately for each experiment using the PROC GLM of SAS (SAS Institute, 2008). For Experiment I, data from regrowth cycles were averaged to compose a single mean per plot (n = 12), and analyzed using a model including the fixed effect of defoliation intensity and the random effect of blocks. Data from Experiment II (n = 12) and III (n = 9) were analyzed using a model including the fixed effect of defoliation intensity. Treatment means were estimated using LSMEANS and compared using the Tukey-Kramer test. The significance level was set at 5 % (p < 0.05). Results For kikuyugrass, lenient defoliations (40 and 50 %) caused tiller mortality and LA increases in the same proportions thus LAI was similar throughout the regrowth period (Figures 2A-C). (low pasture photosynthetic capacity), reducing FARCHT. Although the increase in defoliation intensity reduced the forage production capacity of all three species, some distinctions were observed (Table 4). The lack of difference between defoliations of 40 and 50 % in the kikuyugrass was probably due to compensatory mechanisms between tiller size and tissue flows, as previously reported for this species under moderate defoliation intensity (50 % of initial pre-grazing height; Sbrissia et al., 2018). On the other hand, the annual ryegrass and black oat growth habit (caespitose grasses) might have contributed, as small increments in defoliation Impact of defoliation intensity on the productive potential of pastures Overall, decreases in FARCHT and aLAICHT occurred with increasing defoliation intensity (Table 4). These decreases occurred due to a gradual reduction in Sci. Agric. v.78, n.2, e20190095, 2021 4 Productive capacity of defoliated swards Martins et al. Figure 1 – Instantaneous (IFAR; A, C, and E) and average (FAR; B, D, and F) forage accumulation rates evolution in kikuyugrass (A and B), annual ryegrass (C and D), and black oat (E and F) pastures submitted to different defoliation intensities. Diamonds correspond to the moment when pastures reached their pre-cutting canopy height targets (CHT). Figure 1 – Instantaneous (IFAR; A, C, and E) and average (FAR; B, D, and F) forage accumulation rates evolution in kikuyugrass (A and B), annual ryegrass (C and D), and black oat (E and F) pastures submitted to different defoliation intensities. Diamonds correspond to the moment when pastures reached their pre-cutting canopy height targets (CHT). Figure 2 – Tiller population density (TPD) and average leaf area per tiller (LA) contribution in LAI recovery for kikuyugrass (A, B, and C), annual ryegrass (D, E, and F), and black oat (G, H, and I) pastures submitted to different defoliation intensities. Diamonds correspond to the moment when pastures reached their pre-cutting canopy height targets (CHT). Figure 2 – Tiller population density (TPD) and average leaf area per tiller (LA) contribution in LAI recovery for kikuyugrass (A, B, and C), annual ryegrass (D, E, and F), and black oat (G, H, and I) pastures submitted to different defoliation intensities. Diamonds correspond to the moment when pastures reached their pre-cutting canopy height targets (CHT). intensity significantly decreased the residual LAIs and consequently their productive potentials (Figures 1B and 1C). Such responses suggest that these forage species intensity significantly decreased the residual LAIs and consequently their productive potentials (Figures 1B and 1C). Such responses suggest that these forage species intensity significantly decreased the residual LAIs and consequently their productive potentials (Figures 1B and 1C). Such responses suggest that these forage species exhibit mechanisms of adaptation to defoliation that are different from those observed in kikuyugrass, even under moderate defoliation intensities. exhibit mechanisms of adaptation to defoliation that are different from those observed in kikuyugrass, even under moderate defoliation intensities. Sci. Agric. v.78, n.2, e20190095, 2021 5 Productive capacity of defoliated swards Martins et al. According to Parsons et al. Defoliation intensity influence on LAI recovery Therefore, based on various studies suggesting a period for pastures to be defoliated to maximize forage production, for instance, the FARmax (Parsons et al., 1988) and IFARmax (Morley, 1968; Maeda and Yonetani, 1978), or when both of these are equivalently high (Parsons et al., 2011), our findings show that the productive response of management based on presumed ideal pre-defoliation heights are dependent on the defoliation intensity to which the pasture will be submitted. Generally, in canopies submitted to defoliations up to 60 %, the maximum growth rates were obtained at heights close to the pre-defined targets for our study (Table 4; Figures 1A-F). On the other hand, for defoliation intensities of 70 % and 80 %, the maximum forage accumulation rates were reached at greater heights as compared to the pre- determined ones (Table 4; Figures 1A-F). This lack of synchrony may be due to the average LAI reduction and to the delay in LAI recovery caused by greater defoliation intensity. Thus, for intensively defoliated pastures, the delay in reaching their maximum forage accumulation rates reflects not only the low initial growth rates (Chapman, 2016), but also their need to reach greater heights to minimize the negative effects of low residual LAIs. The use of heights greater than those defined in this study as the pre-defoliation target could The greater defoliation intensities resulted in lower residual LAI due to lower TPD and LA; thus, canopies used different means to recover the LAI after defoliation (Figures 2A-I). Canopies that were lowered by more than half of their canopy height target invested in increasing both LA and TPD at the initial regrowth, while those lowered up to 50 % recovered their LAI mainly through LA increments (Figures 2A-I). A Although every species exhibited TPD reduction after a period of regrowth, this occurred at different intensities and LAI for the three species, which could be a result of their different tillering and TPD maintenance abilities (Duchini et al., 2014; Sbrissia et al., 2018). Therefore, LAI recovery after defoliation depends not only on the defoliation intensity, but also on interspecific factors. In this sense, kikuyugrass submitted to moderate defoliation exhibited lower LAI variations, maintaining a high LA and adjusting TPD; however, when submitted to intense defoliation, this species resorted to LA and TPD adjustments (Figures 2A-C). Impact of defoliation intensity on the productive potential of pastures (1988), young leaves and/or those that grow under full light conditions (e.g., pastures managed at shorter heights) exhibit greater photosynthetic capacity as compared to old leaves and/ or those grown under shadowy conditions (e.g., pastures managed at taller heights), which could, in theory, explain some productive equivalencies of pastures managed under different conditions (e.g., different heights or residual LAI). However, in our study, this advantage does not seem to have been enough to compensate for the differences in residual LAI when the pastures were defoliated more than 50 %. This probably ocurred because once pasture regrowth, it partially depends on tiller recruitment, a process that involves a concomitant investment on tissues that do not contribute to LAI and aerial biomass increment (e.g., roots) (Tainton, 1974), which extend the period to reach the maximum growth rates. Therefore, Nelson and Zarrough (1981) showed that tiller growth is more important than TPD to determine forage accumulation rate in established swards. Additionally, greater LAIs (observed in the less- intense defoliations; Figures 2A, 2D and 2G) usually results in greater leaf elongation rates (Berone et al., 2007; Sbrissia et al., 2018), which, associated with the greater TPD (Figures 2C, 2F and 2I), allowed fast initial increments of average and instantaneous forage accumulation rates in the canopies that were defoliated less intensively (Figures 1A-F). Therefore, apparently, the intense tissue renewal per se cannot be considered a factor that justifies the use of defoliation intensities above 50 % of initial height, particularly when the objective is to maximize forage accumulation. defoliation intensity (Figures 2G-I). These results are in line with those reported in the literature, as annual ryegrass and kikuyugrass have higher tillering capacity while black oat has difficulty to maintain its TPD over time and depends on quick LA increments to recompose its LAI (Duchini et al., 2014; Sbrissia et al., 2018). It is worth noting that the deviations from these trends in specific moments of regrowth are due to the emergence of small and low LA tillers and/or tiller size/density compensation mechanisms (Matthew et al., 1995). Defoliation intensity effect on average and instantaneous forage accumulation rate Parsons et al. (1988) used mathematical models to show the LAI recovery pattern of perennial ryegrass (Lolium perenne L.) submitted to different defoliation intensities to dissert about the rates of gross photosynthesis (Pgross), net photosynthesis (Pnet; i.e., gross tissue production), senescence, and average and instantaneous growth rates. In their study, pastures that were intensively defoliated had lower initial Pnet and senescence rates, which led to a delay in reaching the maximum growth rates, while lenient defoliations had little effect on Pnet and senescence, allowing those pastures to quickly obtain the maximum growth rates. Similarly, our results showed that increasing defoliation intensity promoted delays to obtain the IFARmax and FARmax rates in kikuyugrass, annual ryegrass, and black oat pastures (Table 4; Figures 1A-F). An important difference compared to Parsons et al. (1988) is that, while they found different intervals to reach the FARmax, our results show that, over time, increasing height is needed to reach this condition in highly defoliated swards. Authors’ Contributions Conceptualization: Sbrissia, A.F.; Martins, C.D.M.; Schmitt, D.; Duchini, P.G.; Miqueloto, T. Data acquisition: Martins, C.D.M.; Miqueloto, T.; Schmitt, D.; Duchini, P.G. Data analysis: Martins, C.D.M.; Schmitt, D.; Duchini, P.G.; Miqueloto, T. Design of methodology: Sbrissia, A.F.; Martins, C.D.M.; Schmitt, D.; Duchini, P.G.; Miqueloto, T. Writing and editing: Martins, C.D.M.; Schmitt, D.; Duchini, P.G.; Sbrissia, A.F.; Miqueloto, T. Donaghy, D.J.; Fulkerson, W.J. 1998. Priority for allocation of water soluble carbohydrate reserves during regrowth of Lolium perenne. Grass and Forage Science 53: 211-218. Duchini, P.G.; Guzatti, G.C.; Ribeiro-Filho, H.M.N.; Sbrissia, A.F. 2014. Tiller size/density compensation in temperate climate grasses grown in monoculture or in intercropping systems under intermittent grazing. Grass and Forage Science 69: 655- 665. Gautier, H.; Varlet-Grancher, C.; Hazard, L. 1999. Tillering responses to the light environment and to defoliation in populations of perennial ryegrass (Lolium perenne L.) selected for contrasting leaf length. Annals of Botany 83: 423-429. Implicationsi p The findings reported here highlight the mechanisms of LAI recovery and forage accumulation rates in defoliated swards. As the defoliation intensity increases, greater canopy heights are needed for pastures to achieve their maximum forage accumulation rates, requiring longer regrowth intervals. The need for high tiller appearance after defoliation, which delays canopy LAI recovery, seems to be one of the main causes; thus, grazing management strategies that involve costly tiller recruitment could decrease both overall forage production and sward persistence over time. However, the plant ability to recover LAI after successive intense defoliations seems to be species-dependent and related to their phenotypic plasticity. Brougham, R.W. 1958. Interception of light by the foliage of pure and mixed stands of pasture plants. Australian Journal of Agricultural Research 9: 39-52. Carnevalli, R.A.; Silva, S.C.; Bueno, A.A.O.; Uebele, M.C.; Bueno, F.O.; Hodgson, J.; Silva, G.N.; Morais, J.P.G. 2006. Herbage production and grazing losses in Panicum maximum cv. Mombaça under four grazing managements. Tropical Grasslands 40: 165-176. Chapman, D.F. 2016. Using ecophysiology to improve farm efficiency: application in temperate dairy grazing systems. Agriculture 6: 17-36. Comissão de Química e Fertilidade do Solo [CQFS]. 2004. Fertilizing and Liming Manual for the States of Rio Grande do Sul and Santa Catarina = Manual de adubação e calagem para os Estados do Rio Grande do Sul e Santa Catarina. CQFS RS/ SC, Porto Alegre, RS, Brazil (in Portuguese). Defoliation intensity influence on LAI recovery Annual ryegrass showed adjustments in TPD and increments in LA to recover its LAI in all defoliation intensities (Figures 2D-F), whereas black oat regulated its LAI throughout the regrowth primarily through adjustments in LA, regardless of the 6 Sci. Agric. v.78, n.2, e20190095, 2021 Productive capacity of defoliated swards Martins et al. favor the accumulation of stem and dead material (Da Silva et al., 2015), structures that reduce the nutritional value of the material offered to the animals and impose physical restrictions on the foraging process (Stobbs, 1973; Benvenutti et al., 2006). Benvenutti, M.A.; Gordon, I.J.; Poppi, D.P. 2006. The effect of the density and physical properties of grass stems on the foraging behaviour and instantaneous intake rate by cattle grazing an artificial reproductive tropical sward. Grass and Forage Science 61: 272-281. Berone, G.D.; Lattanzi, F.A.; Colabelli, M.R.; Agnusdei, M.G. 2007. A comparative analysis of the temperature response of leaf elongation in Bromus stamineus and Lolium perenne plants in the field: intrinsic and size-mediated effects. Annals of Botany 100: 813-820. Acknowledgements The authors would like to thank FAPESC (Fundação de Apoio à Pesquisa Científica e Tecnológica do Estado de Santa Catarina; grant number 2019 TR- 584) for funding this research. The first author would like to thank FAPESC and CAPES (Coordenação de Aperfeiçoamento de Pessoal de Nível Superior) for the scholarship. Da Silva, S.C.; Bueno, A.A.O.; Carnevalli, R.A.; Uebele, M.C.; Bueno, F.O.; Hodgson, J.; Matthew, C.; Arnold, G.C.; Moraes, J.P.G. 2009. Sward structural characteristics and herbage accumulation of Panicum maximum cv. Mombaça subjected to rotational stocking managements. Scientia Agricola 66: 8-19. Da Silva, S.C.; Sbrissia, A.F.; Pereira, L.E.T. 2015. Ecophysiology of C4 forage grasses: understanding plant growth for optimising their use and management. Agriculture 5: 598-625. Authors’ Contributions References Alvares, C.A.; Stape, J.L.; Sentelhas P.C.; Gonçalves, J.L.M.; Sparovek, G. 2013. Köppen’s climate classification map for Brazil. Meteorologische Zeitschrift 22: 711-728. Guzatti, G.C.; Duchini, P.G.; Sbrissia, A.F.; Ribeiro-Filho, H.M.N. 2015. Qualitative aspects and biomass production in oats and ryegrass pastures cultivated pure or intercropping and subjected to lenient grazing. Arquivo Brasileiro de Medicina Veterinária e Zootecnia 67: 1399-1407 (in Portuguese, with abstract in English). Barbosa, R.A.; Nascimento Jr, D.; Euclides, V.P.B.; Silva, S.C.; Zimmer, A.H.; Torres, R.A.A.J. 2007. Tanzania grass subjected to combinations of intensity and frequency of grazing. Pesquisa Agropecuária Brasileira 42: 329-340 (in Portuguese, with abstract in English). Irving, J. 2015. Carbon assimilation, biomass partitioning and productivity in grasses. Agriculture 5: 1116-1134. Korte, C.J.; Watkin, B.R.; Harris, W. 1982. Use of residual leaf area index and light interception as criteria for spring-grazing management of a ryegrass-dominant pasture. New Zealand Journal of Agricultural Research 25: 309-319. Barthram, G.T. 1985. Experimental techniques: the HFRO sward stick. p. 29-30. In: Alcock, M.M., ed. Biennial report of the Hill Farming Research Organization. Hill Farming Research Organization, Midlothian, UK. 7 Sci. Agric. v.78, n.2, e20190095, 2021 Productive capacity of defoliated swards Martins et al. Parsons, A.J.; Johnson, I.R.; Harvey, A. 1988. Use of a model to optimize the interaction between frequency and severity of intermittent defoliation and to provide a fundamental comparison of the continuous and intermittent defoliation of grass. Grass and Forage Science 43: 49-59. Lemaire, G.; Chapman, D. 1996. Tissue flows in grazed plant communities. p. 3-36. In: Hodgson, J.; Illius, A.W., eds. The ecology and management of grazing systems. CAB International, Wallingford, UK. Maeda, S.; Yonetani, T. 1978. Optimum cutting stage of forage plants. II. Seasonal changes of CGR and average productivity in Italian ryegrass population. Journal of Japanese Society of Grassland Science 24: 10-16. Santos, G.T.; Zanini, G.D.; Padilha, D.A.; Sbrissia, A.F. 2016. A grazing height target to minimize tiller stem elongation rate in annual ryegrass swards. Ciência Rural 46: 169-175. Matthew, C.; Lemaire, G.; Hamilton, N.R.S.; Hernandez- Garay, A. 1995. A modified self-thinning equation to describe size/density relationships for defoliated swards. Annals of Botany 76: 579-587. Sbrissia, A.F.; Duchini, P.G.; Zanini, G.D.; Santos, G.T.; Padilha, D.A.; Schmitt, D. 2018. Defoliation strategies in pastures submitted to intermittent stocking method: underlying mechanisms buffering forage accumulation over a range of grazing heights. Crop Science 56: 1-10. McSteen, P. 2009. Hormonal regulation of branching in grasses. Plant Physiology 149: 46-55. Sci. Agric. v.78, n.2, e20190095, 2021 References Soil Survey Staff. 1999. Soil Taxonomy: A Basic System of Soil Classification for Making and Interpreting Soil Surveys. 2ed. USDA-NRCS, Washington, DC, USA. (Agricultural Handbook, 436). Morley, F.H.W. 1968. Pasture growth curves and grazing management. Australian Journal of Experimental Agriculture 8: 40-45. Nelson, C.J.; Zarrough, K.M. 1981. Tiller density and tiller weight as yield determination of vegetative swards. Plant Physiology and Herbage Production 13: 25-29. Stobbs, T.H. 1973. The effect of plant structure on the intake of tropical pastures. I. Variation in the bite size of grazing cattle. Australian Journal of Agricultural Research 24: 609-619. Parsons, A.J.; Rowarth, J.; Thornley, J.; Newton, P. 2011. Primary production of grasslands, herbage accumulation and use, and impacts of climate change. p. 3-13. In: Lemaire, G.; Hodgson, J.; Chabbi, A., eds. Grassland productivity and ecosystem services. CAB International, Wallingford, UK. Tainton, N. 1974. Effects of different grazing rotations on pasture production. Grass and Forage Science 29: 191-202. Zanini, G.D.; Santos, G.T.; Sbrissia, A.F. 2012. Frequencies and intensities of defoliation in Aruana guineagrass swards: morphogenetic and structural characteristics. Revista Brasileira de Zootecnia 41: 1848-1857. 8
https://openalex.org/W4249086282	https://www.researchsquare.com/article/rs-48960/v4.pdf?c=1631884549000	English	null	Effects of trees, gardens, and nature trails on heat index and child health: Design and methods of the Green Schoolyards Project	Research Square (Research Square)	2,021	cc-by	8,912	Effects of trees, gardens, and nature trails on heat index and child health: Design and methods of the Green Schoolyards Project Kevin Lanza (  Kevin.L.Lanza@uth.tmc.edu ) Research article Version of Record: A version of this preprint was published on January 7th, 2021. See the published version at https://doi.org/10.1186/s12889-020-10128-2. Page 1/19 Page 1/19 Page 1/19 Abstract Background: Latinx children in the United States are at high risk for nature-deficit disorder, heat-related illness, and physical inactivity. We developed the Green Schoolyards Project to investigate how green features—trees, gardens, and nature trails—in school parks impact heat index (i.e., air temperature and relative humidity) within parks, and physical activity levels and socioemotional well-being of these children. Herein, we present novel methods for a) observing children’s interaction with green features and b) measuring heat index and children’s behaviors in a natural setting, and a selection of baseline results. Methods: During two September weeks (high temperature) and one November week (moderate temperature) in 2019, we examined three joint-use elementary school parks in Central Texas, United States, serving predominantly low- income Latinx families. To develop thermal profiles for each park, we installed 10 air temperature/relative humidity sensors per park, selecting sites based on land cover, land use, and even spatial coverage. We measured green features within a geographic information system. In a cross-sectional study, we used an adapted version of System for Observing Play and Recreation in Communities (SOPARC) to assess children’s physical activity levels and interactions with green features. In a cohort study, we equipped 30 3rd and 30 4th grade students per school during recess with accelerometers and Global Positioning System devices, and surveyed these students regarding their connection to nature. Baseline analyses included inverse distance weighting for thermal profiles and summing observed counts of children interacting with trees. Results: In September 2019, average daily heat index ranged 2.0°F among park sites, and maximum daily heat index ranged from 103.4°F (air temperature = 33.8°C; relative humidity = 55.2%) under tree canopy to 114.1°F (air temperature = 37.9°C; relative humidity = 45.2%) on an unshaded playground. 10.8% more girls and 25.4% more boys interacted with trees in September than in November. Conclusions: We found extreme heat conditions at select sites within parks, and children positioning themselves under trees during periods of high heat index. These methods can be used by public health researchers and practitioners to inform the redesign of greenspaces in the face of climate change and health inequities. Background Children in modern times are experiencing “nature-deficit disorder,” described as the human costs of alienation of nature (1), and are consequently missing the benefits of engaging with nature such as a stronger sense of place, improvements in physical and mental health, greater environmental knowledge, and pro-environment attitudes as an adult (2). In the United States (US), children spend three times as many hours on the computer or watching television as they do playing outdoors (3). Furthermore, access to nature is an environmental justice issue: individuals who are Latinx, low-income, and/or with low levels of education have less access to vegetation (4, 5). These same populations lacking access to nature are also at risk for heat-related illnesses, such as heat exhaustion and heat stroke (6). Communities of color and low-income families disproportionately live in areas characterized by urban heat islands (7), the phenomenon in which cities experience higher temperatures than nearby areas due to high amounts of impervious materials, lack of vegetation, morphology, and waste heat from industrial processes (8). Along with their high risk for nature-deficit disorder and heat-related illness, Black and Latinx children from low-income families are less physically active than other groups (9, 10). Fewer than half of all US children are reaching recommended physical activity levels (11). This lack of physical activity poses a serious public health threat: children who are not sufficiently active are more likely to develop several chronic diseases, such as obesity and type 2 diabetes Page 2/19 (12). To compound the problem, elevated temperatures have been found to be negatively associated with children’s physical activity (13). With children spending a significant portion of their time at school (14), public health practitioners have promoted child health through green schoolyards—“natural spaces (that) are used as outdoor classrooms to enhance learning outcomes and create daily wellness for the children they serve” (15). Greenspaces have been found to lower air temperatures (through shading and evapotranspiration from vegetation) and improve human thermal comfort (16, 17). Schoolyard greening, in particular, has been shown to contribute to children’s physical, mental, and social-emotional well-being (18-21). However, public health researchers have not fully explored the relations between green schoolyards, temperatures, and child health. Background In response, we developed the Green Schoolyards Project to establish whether school parks can serve as a tool for urban heat island adaptation and health promotion in divested communities at risk of disconnect from nature (and its associated health consequences) and heat-related illness. The specific aims of the Green Schoolyards Project are to determine how green features—trees, gardens, and nature trails—in joint-use elementary school parks impact a) heat index within parks; b) physical activity levels of predominantly Latinx children from low-income families; and c) psychosocial and academic outcomes (i.e., social-emotional learning skills, misconduct at school, and standardized test scores) of these children. Herein, we detail the research design and methods of the Green Schoolyards Project, which are novel and innovative in two major ways: a) direct observation of children’s interaction with green features at multiple physical sites per park; and b) time-matching of objective measurements of heat index, children’s geographic location, and children’s physical activity levels in relation to location of green features. We demonstrate the importance of these methods by presenting a selection of baseline results: certain sites within parks were characterized by extreme heat conditions and children positioned themselves under trees during periods of high heat index. In response, we developed the Green Schoolyards Project to establish whether school parks can serve as a tool for urban heat island adaptation and health promotion in divested communities at risk of disconnect from nature (and its associated health consequences) and heat-related illness. The specific aims of the Green Schoolyards Project are to determine how green features—trees, gardens, and nature trails—in joint-use elementary school parks impact a) heat index within parks; b) physical activity levels of predominantly Latinx children from low-income families; and c) psychosocial and academic outcomes (i.e., social-emotional learning skills, misconduct at school, and standardized test scores) of these children. Herein, we detail the research design and methods of the Green Schoolyards Project, which are novel and innovative in two major ways: a) direct observation of children’s interaction with green features at multiple physical sites per park; and b) time-matching of objective measurements of heat index, children’s geographic location, and children’s physical activity levels in relation to location of green features. We demonstrate the importance of these methods by presenting a selection of baseline results: certain sites within parks were characterized by extreme heat conditions and children positioned themselves under trees during periods of high heat index. Research Design The Green Schoolyards Project consisted of two separate studies: a) a serial cross-sectional study focused on physical sites within parks, and b) a prospective cohort study focused on individual students. The cross-sectional study was designed to examine the associations between heat index, children’s physical activity, and interaction with green features for multiple sites (e.g., playground, track, and soccer field) per park on multiple times per day. The cohort study followed a selection of students from each school affiliated with these parks to assess the impact of green features—the amount of which differs per park—on student’s physical activity levels during recess for different heat index conditions, along with their connection to nature, social-emotional learning skills, misconduct, and standardized test scores. Methods Research Design Project Sites and Green Features Project Sites and Green Features Three elementary school parks within a school district in Central Texas, US, were used for this study. The project was a comparative analysis between similar parks with different levels of green features. Three schools met our initial selection criteria: serving populations greater than 85% economically disadvantaged Latinx; located in ZIP codes with low Nature Factor scores; joint-use agreements between the school district and the city Parks and Recreation Department permitting the surrounding community to use parks after school hours; and equivalent park features (e.g., playgrounds, soccer fields, running tracks, and basketball courts). The selection criterion of Nature Factor is defined as the sum of Nature Factor Ratings of all parks within a ZIP code (25). Nature Factor Rating is the sum of four park-level ratings: park acreage rating, Trust for the Public Land land use rating, National Recreation and Parks Association park status rating, and tree canopy rating (26). High values for Nature Factor Ratings (e.g., high park acreage, designed lands, open park status, and high levels of tree canopy) correspond to high Nature Factor scores (i.e., higher levels of nature present in that ZIP code). The three selected schools are in ZIP codes with Nature Factor scores of 121, 118, and 198, respectively, which are relatively low compared to those of other ZIP codes (n = 53; min. = 0; max. = 712; mean = 150; standard deviation = 143). The three school parks were characterized by different profiles of green features: the “intervention park” had added green features (i.e., trees, wildflower meadow, and nature trail); the “low-green park” had relatively low amounts of historical green features (i.e., trees); and the “high-green park” had relatively high amounts of historical green features (i.e., trees, wildlife habitat garden, and nature trail). Landscape vegetation—including trees and gardens—was planted in each park to provide visual enhancement; learning, play, and recreation opportunities; and a calming and welcoming environment. The school district is responsible for maintaining trees, and individual schools and their communities are responsible for maintaining gardens and nature trails. The wildflower meadow at the intervention park is a 1,383m2 pollinator-supporting grassland consisting of dense, tall wildflowers and grasses native to Texas (27). Conceptual Model Research design of the Green Schoolyards Project was based on the social-ecological model and attention restoration theory. The social-ecological model states multiple levels of influence—individual, social, environmental, and policy— impact health behaviors (e.g., physical activity), and these influences interact across levels to impact behavior (22). Effective health interventions focus on behavior-specific influences, and intervene at multiple levels of influence. In general, attention restoration theory proposes exposure to natural environments replenishes the ability to concentrate, a cognitive resource that can be depleted (23, 24). Natural settings are considered restorative because these environments provide a) respite from everyday stressors, b) space that feels extensive, c) objects that fascinate and require little concentration, and d) intrinsic compatibility with humans (24). Page 3/19 Page 3/19 From these theoretical underpinnings, we developed a conceptual model of the Green Schoolyards Project that illustrates the concepts of interest and relations between these concepts (Figure 1). The model provides our logic for testing green features in school parks as a strategy for improving health, psychosocial, and academic outcomes of children. We posited that trees and gardens will decrease outdoor heat and that nature trails, which surround gardens but are not directly vegetated, will not significantly impact outdoor heat. When subjected to outdoor heat, children will engage in less physical activity due to thermal discomfort. Trees, gardens, and nature trails will inherently provide more opportunities for children to physically interact with nature, yet outdoor heat will have a mixed impact on these physical interactions: children will seek natural settings providing protection from outdoor heat (e.g., park sites under tree shade) while not seeking and potentially avoiding natural settings without heat protection (e.g., unshaded park sites). The relation between children’s physical interactions with nature and their physical activity levels will depend on the interaction with nature (e.g., climbing trees versus resting under tree shade) and whether the setting is intended for physical activity (e.g., playground under tree canopy versus picnic tables under tree canopy). Lastly, children’s physical interactions with nature will offer opportunities for beneficial outcomes, such as mentally connecting with nature and restoring ability to concentrate. Improved concentration will increase children’s social-emotional learning skills and standardized tests scores, and decrease their misconduct at school. Figure 1. Conceptual model of Green Schoolyards Project. Project Sites and Green Features The meadow includes signage and a weaving, 20 meter-long, 1.5 meter-wide, nature trail—a secondary or tertiary path that branches from a primary track, trail, or pathway and composed of materials such as decomposed granite, compacted dirt, and stepping stones. To Page 4/19 Page 4/19 encourage use of the meadow, a nature themed-story in English- and Spanish-language was installed on temporary signage placed along the trail. The wildlife habitat garden at the high-green park is a 247m2 dense cover of native grasses and shrubs that serves as a haven for local and migratory species by providing food, cover, and places to raise young (28). This garden includes a nature trail that is nearly identical to the trail in the wildflower meadow at the intervention park. For the Green Schoolyards Project, both wildflower meadow and wildlife habitat garden are categorized as gardens because of similar characteristics. The low-green park had no gardens or nature trails. For each school park, we calculated tree canopy cover using i-Tree Canopy, a publicly available tool that uses random point sampling to estimate the percentage of tree canopy cover for a predefined area (29). Although trees at the intervention park were more abundant and evenly distributed than trees at parks at the other schools, tree canopy cover was only 8.5% (standard error = 1.97) because trees planted were saplings. The low-green park had 11.5% (standard error = 2.26) tree canopy cover, with most trees clustered in the far northwest corner. The high-green park had 22.5% (standard error = 2.95) tree canopy cover, with relatively large trees on the park’s periphery. Project Period We designed data collection to take place on 18 days over the fall semester in 2019, which will be duplicated in 2020 for a total of 36 study days. Each year, study days consist of two September weeks (i.e., five weekdays and one weekend day per week) and one November week (i.e., five weekdays and one weekend day). We selected September and November because these months have historically high and moderate temperature conditions, respectively: the weather station at the city’s major airport recorded monthly mean average air temperatures of 26.4°C in September and 15.3°C in November from 2009 to 2018 (30). Prior to undertaking any project activities, we received approval of project protocols by the institutional review board at The University of Texas Health Science Center at Houston (HSC-SPH-19- 0502) and the school district. We also received informed consent from participants’ parents and written assent from study participants. Measurement of Heat Index We measured heat index—the combination term for air temperature and relative humidity that captures what the temperature feels like (31)—by semi-permanently installing 10 HOBO MX2302A external air temperature/relative humidity sensor data loggers (Onset Computer Corporation, MA) at each park. Previous studies have used comparable networks of in situ sensors to monitor microclimatic conditions of a given area (32-34). Measurement of near-surface air temperatures is advantageous over the use of land surface temperatures as a proxy for air temperatures, as research has shown land surface temperatures are not directly comparable to air temperatures (35, 36). In situ measurement of air temperatures has been found to be more useful for estimating short-term, actual temperature exposures than using land surface temperatures or the percentage of impervious surface (37). Designed for outdoor use, the MX2302A model collects air temperature (±0.2°C from 0 to 70°C) and relative humidity data (±2.5% from 10% to 90%), and is configured to wirelessly link with the free HOBOmobile app on a cell phone or tablet (38), permitting efficient collection of air temperature and relative humidity data by project members. Before deploying HOBO sensors, we encased each sensor in an RS3-B solar radiation shield (Onset Computer Corporation, MA), which results in improved temperature measurement accuracy by protecting the sensor from absorption of incoming solar radiation and resultant heat gain. In another attempt to improve temperature measurement accuracy, we attached the radiation shield (encasing the external sensor) and data logger to a 2x2” piece of weather-treated lumber, which serves as a physical buffer minimizing heat transfer between the sensor and the installation surface in the park, such as a metal swing set pole (Figure 2). We programmed the sensors to record air temperature and relative humidity every five minutes, consistent with previous studies (34, 39, 40). Page 5/19 Page 5/19 Figure 2. HOBO MX2302A external air temperature/relative humidity sensor data logger on swing set at low-green park. Figure 2. HOBO MX2302A external air temperature/relative humidity sensor data logger on swing set at low-green park. For each park, we selected 10 sites for HOBO sensors based on land cover (e.g., grass, pavement, and mulch); land use (e.g., soccer field, basketball court, and playground); comparability across parks; and even spatial coverage (Figure 3). To include a highly impervious area for comparison within park sites, we installed one of the 10 sensors at each park’s parking lot, just outside park boundaries. Direct Observation of Parks The cross-sectional study utilized the System for Observing Play and Recreation in Communities (SOPARC), a validated direct observation tool for assessing the conditions and users of park sites (43). Following SOPARC protocol, we divided each park into target areas intended for physical activity, such as basketball courts and soccer fields (Figure 3). On study days during school (i.e., 7:00 and 12:00) and after school (i.e., 16:00 and 18:00), study staff—in pairs for interrater reliability—administered SOPARC by walking from target area to target area and recording what they observed. We adapted SOPARC to measure physical activity levels of children aged 1–12 years old and these children’s interactions with green features. Although previous research has used direct observation to examine the influence of nature on children’s play (44), no research employs SOPARC to quantify the number of children’s interactions with different green features at multiple park sites. On study days for each target area, trained staff used SOPARC to record the date and time of observation and target area conditions (i.e., whether area was accessible, usable, equipped, supervised, dark, empty, and organized); then, if applicable, to record the primary physical activity (e.g., playing basketball) of female and male children; scan for the physical activity levels of female and male children; and lastly scan for the number of female and male children interacting with green features (i.e., no interaction, under tree canopy or touching tree, interacting with garden, and on nature trail). To employ SOPARC, staff used the iSOPARC application on an electronic tablet for scan counts and input data on a data collection form (see Additional File 1). Figure 3. HOBO sensors and SOPARC target areas at (A1–A2) intervention, (B1–B2) low-green, and (C1–C2) high-green parks. Measurement of Green Features We identified the location, type, and quantity of green features using four-band, 60cm orthoimagery taken in November 2018 by the US Department of Agriculture’s National Agriculture Imagery Program (41). Within a geographic information system (ArcGIS 10.6.1, ESRI, Redlands, CA, USA), we digitized polygons of trees, gardens, and nature trails, an established technique deemed appropriate for the relatively small park areas (i.e., intervention = 21,448 m2; low- green = 27,923 m2; high-green = 16,187 m2) (42). Measurement of Heat Index To capture air temperature and relative humidity experienced by humans, we installed sensors at two meters above ground level, similar to previous studies (33, 34, 39). To promote community awareness of our project and deter vandalism, we attached a small laminated tag with a description of the sensor and our contact information to each sensor. Sensors were installed the day before a study week, and removed at the end of each study week. Measurement of Connection to Nature, Psychosocial and Academic Outcomes, and School Policies For baseline data collection in November 2019, we collected data on the cohort sample and school policies from three sources. First, we administered aloud a written survey—in both English and Spanish language—to the cohort sample, asking them about their connection to nature using two adapted instruments: Inclusion of Nature with Self and Connection to Nature Index (see Additional File 2) (49-51). Second, the school district provided individual data for each student on sociodemographic characteristics, social-emotional learning skills (from a student climate survey), number of disciplinary actions for misconduct, and standardized test scores. Lastly, we distributed an annual survey—adapted from a previous study (52)—to ask school principals about policies impacting park access, greening at school parks, and student physical activity. Statistical Analysis of Baseline Data For the selection of baseline findings shared within, we performed analysis over several steps. To develop thermal profiles for each park, we first calculated heat index—from air temperature and relative humidity data recorded by HOBO sensors—using validated equations utilized by US National Weather Service (53, 54). Heat index is measured in degrees Fahrenheit (55). Within GIS, we used inverse distance weighting to create spatially continuous thermal profiles for each park. A common interpolation method in urban heat island measurement (56), inverse distance weighting permits estimation of unsampled heat index values between HOBO sensors by averaging sampled heat index values from sensors surrounding each prediction location. We used SOPARC data to understand how children interact with trees during time periods with different temperature conditions, summing observed counts of children under tree canopy or touching trees by sex of child, park, and study period. Cohort Study Sample For the cohort study, we recruited 40 3rd and 40 4th grade students per school over two years, to achieve a final sample size of 30 students per grade, after accounting for attrition. From mid-August 2019 through early September 2019, we recruited participants by convenience sampling. Participant incentives were a total of $35 US dollars/year worth of supermarket gift cards (i.e., $10 for each September study week and $15 for each November study week). Page 6/19 Measurement of Geographic Location and Physical Activity On study days during recess (i.e., 30-minute period of unstructured play under teacher supervision), the cohort sample wore elastic belts around their waist equipped with a Qstarz BT-Q1000XT Global Positioning System (GPS) device (Qstarz Intl Co., Taipei, Taiwan) and an Actigraph wGT3X-BT accelerometer (ActiGraph LLC, FL) to measure geographic location and physical activity levels, over time (45, 46). We set sampling rates of 15 seconds for GPS devices and accelerometers (47, 48). For the separate recess periods per grade, belt distribution began five minutes before recess start, and belt collection occurred once teachers signaled recess end. On study days during recess (i.e., 30-minute period of unstructured play under teacher supervision), the cohort sample wore elastic belts around their waist equipped with a Qstarz BT-Q1000XT Global Positioning System (GPS) device (Qstarz Intl Co., Taipei, Taiwan) and an Actigraph wGT3X-BT accelerometer (ActiGraph LLC, FL) to measure geographic location and physical activity levels, over time (45, 46). We set sampling rates of 15 seconds for GPS devices and accelerometers (47, 48). For the separate recess periods per grade, belt distribution began five minutes before recess start, and belt collection occurred once teachers signaled recess end. Results Thermal Profiles of School Parks Time-Matching of Geographic Location, Physical Activity, and Heat Index Data from GPS devices, accelerometers, and HOBO sensors will be time-matched, allowing us to know a student’s location, student’s physical activity intensity level, and heat index at that location at 15-second intervals throughout recess. The location of green features will be joined to the time-matched device data, within GIS. Although previous studies have matched children’s geographic location and physical activity levels over time (45-48), this study enables assessment of a child’s experienced heat index and physical activity level at any particular location. Thermal Profiles of School Parks Average daily heat index was 87.4°F in September and 62.4°F in November across the three parks, with average daily heat index ranging from 86.8°F to 88.8°F from September 16–30th, in 2019 (Figure 4). The minimum and maximum values of this range originated from the intervention park at two sites less than 50m apart: a playground under heavy tree canopy (Target Area 18) and an unshaded playground (Target Area 9), respectively (Figure 3A2). Page 7/19 Page 7/19 In September 2019 at the intervention park (Figure 4A), the canopied playground reached a maximum heat index of 104.3°F (air temperature = 34.6°C; relative humidity = 51.0%), whereas the unshaded playground reached 114.1°F (air temperature = 37.9°C; relative humidity = 45.2%). At the low-green park (Figure 4B), the lowest maximum heat index was 103.4°F (air temperature = 33.8°C; relative humidity = 55.2%) under heavy tree canopy (Target Area 14), and the highest was 106.9°F (air temperature = 35.8°C; relative humidity = 47.4%) at an unshaded playground (Target Area 13). At the high-green park (Figure 4C), the lowest maximum heat index was 104.2°F (air temperature = 33.8°C; relative humidity = 55.8%) at a basketball court under an artificial shade structure (Target Area 9), and the highest was 109.4°F (air temperature = 35.9°C; relative humidity = 50.5%) at an unshaded playground (Target Area 14). Children’s Interaction with Green Features When performing SOPARC scans for the number of children interacting with green features, we observed a total of 1,229 children in target areas with trees, three children in target areas with gardens, and zero children in target areas with nature trails, during 12 days in September and six days in November 2019. In target areas with trees (Table 1), children observed at the low-green park constituted 67.1% of all children observed in September and 65.4% of all children observed in November. Target areas with trees were frequented by a slightly larger percentage of female children than male children in both September (52.4% female) and November (51.8% female). Across the three parks, these target areas had 10.8% more female children and 25.4% more male children under tree canopy or touching trees in September than in November 2019. However, a lower percentage of female and male children at the high-green park interacted with trees in September than in November. Across the three parks, we observed no children interacting with gardens or on nature trails during study days in September and November 2019. Table 1. Children’s interaction with trees at school parks on study days using SOPARC (September, Table 1. Children’s interaction with trees at school parks on study days using SOPARC (September, ction with trees at school parks on study days using SOPARC (September, School Park Study Period (2019) Children Under Tree Canopy or Touching Trees1 Total Children Observed1 Female (%) Male (%) Female (#) Male (#) Intervention September 59.7 65.7 72 67 November 50.0 53.7 30 41 Low-Green September 71.4 77.6 273 228 November 59.8 44.7 184 132 High-Green September 19.6 13.3 46 60 November 63.2 64.5 36 60 Total September 63.2 64.5 391 355 November 52.4 39.1 250 233 1Observations are for target areas with trees present within its boundaries. Discussion Methodology for Climate and Health Solutions 1Observations are for target areas with trees present within its boundaries. Methodology for Climate and Health Solutions Methodology for Climate and Health Solutions We provided baseline findings from the Green Schoolyards Project as evidence for the utility of the methods in understanding how green features can moderate place-based climate change and health inequities affecting children. Page 8/19 Page 8/19 Public health researchers and practitioners can use these methods as a model for exploring how joint-use parks with green features can serve as climate and health solutions for divested communities, in the wake of current climate change and health inequities impacting cities (7). Understanding how green features in school parks impact heat index and child health is essential due to projected increases in a) urban temperatures from population-driven development; b) global temperatures from greenhouse gas emissions; and c) the intensity, frequency, and duration of extreme heat events from climate change (57). From thermal profiles of each park (Figure 4), we found a two-degree range in average daily heat index (86.8–88.8°F) in September 2019 across the three parks, which is a change detectable by humans (58). Park sites with heavy tree canopy and artificial shade structures exhibited the lowest heat index values, and unshaded playgrounds exhibited the highest heat index values. These results corroborate those of a previous study that examined the temperatures of surface materials in parks in Phoenix, Arizona, in which researchers found tree canopy and artificial shade structures were associated with significant reductions in surface temperatures, and playground structures exhibited the highest surface temperatures (59). In September at the intervention park, we found a 9.8°F difference in maximum heat index between a canopied playground and an unshaded playground—the difference between “Extreme Caution” and “Danger” levels for likelihood of extreme heat disorders with prolonged exposure or strenuous activity, as defined by the US National Weather Service (60). From using SOPARC to understand children’s interaction with green features, we found more children were interacting with trees in September than in November across the three parks (Table 1). This may suggest children were actively seeking trees—a proven heat management strategy (16, 61)—during high heat index for thermal comfort. Yet this finding was reversed when examining the high-green park independently, which may be related to children’s preference for a playground characterized by sparse canopy cover (i.e., 53.8% of all children observed), and their unwillingness to travel to large trees on the park’s periphery during high heat index. Methodology for Climate and Health Solutions Our finding that no children interacted with gardens or used nature trails may be due to these features being located far from play elements (e.g., slides, ladders, and swings), which have been shown to be associated with more users and more moderate-to-vigorous physical activity (62). In addition, installation of green features in a park may not induce the use of these spaces: researchers have found that living near sidewalk improvements was not associated with accelerometer-derived physical activity (63). Lastly, the gardens at the parks (i.e., wildflower meadow and wildlife habitat garden) may not have been aesthetically pleasing enough or functional for children, unlike vegetable and fruit gardens. Children have been found to engage in more physical activity during outdoor, garden-based lessons than during indoor, classroom- based lessons (64). Seasonally, wildflower meadows may be more attractive to children when flowers and pollinators are present.Strengths and Limitations The methods presented herein are unique for adapting a direct observation tool to measure children’s interaction with green features, and for actively documenting heat index and children’s behaviors in a natural setting (i.e., recess period during a typical school day). Other strengths included the use of validated tools for the objective measurement of heat index (i.e., HOBO sensors), conditions and users of park sites (i.e., SOPARC), children’s physical activity levels (i.e., accelerometers), and children’s geographic location (i.e., GPS devices). One limitation of the Green Schoolyards Project was the uneven spatial distribution of HOBO sensors at each park, which could have reduced the accuracy of air temperature and relative humidity values in the thermal profile for each park. In inverse distance weighting interpolation, an uneven spatial distribution of observational data points results in less accurate predictions between observational data points (65). We found that achieving an even spatial coverage of sites for HOBO sensors at each park was difficult in practice. Installation at certain sites—in particular soccer fields— would obstruct use of those sites and/or sites lacked a surface for us to attach a sensor at two-meter height. Page 9/19 Page 9/19 Additional reasons for uneven spatial coverage included the potential for sensor damage, vandalism, and theft at each prospective site. Methodology for Climate and Health Solutions Future studies can develop more accurate thermal profiles by installing HOBO sensors with more even spatial coverage, supplementing the data recorded by installed HOBO sensors with periodic handheld measurements between installation sites, and/or modeling air temperatures from a combination of air temperature and land surface temperature inputs (66). We experienced two limitations with the use of SOPARC. For observations occurring on study days during school (i.e., 7:00 and 12:00), the number of children present at each park was directly linked to the schedule of the school and teachers, which resulted in significant differences in the number of children observed during school hours across school parks. If a recess period happened to coincide with an observation period, then more children were observed than if there was no overlap between observation and recess period. Each school had different schedules for recess across classes and grades (i.e., pre-kindergarten through 5th grade), and recess periods were sometimes shared among multiple classes and grades. Teachers were in control of the length of recess, and the start and end times of recess often fluctuated. As a solution, researchers can work with school staff to understand the recess schedule, and assign SOPARC observation periods accordingly. A second limitation of SOPARC was that we did not observe children’s physical activity and interaction with green features simultaneously. Because these two behaviors were observed on different scans, the female and male children observed during the scans for physical activity were not necessarily the same children observed during the scans for interaction with green features. In future work, we will test the use of a single scan to measure both physical activity and interaction with green features, to understand the physical activity levels of children while interacting with green features. Behavior mapping may be preferred over SOPARC because this direct observation method identifies individuals’ specific locations within a study area (67), which allows for the production of design-sensitive results. Lastly, the cohort sample may be prone to response bias—specifically acquiescence bias and social desirability bias— when responding to survey items about their connection to nature, potentially affecting data quality (68-71). Conclusions We designed the Green Schoolyards Project to determine the relations between green features and heat index at joint- use school parks and the health of children who are at high risk for nature-deficit disorder, heat-related illness, and physical inactivity. From baseline results, we found extreme heat index conditions at school parks, significant differences in heat index across park sites, and more children interacting with trees during periods of high heat index than periods of moderate heat index. The methods presented herein can be adopted by projects in other cities that are exploring how to redesign urban greenspaces to adjust to high heat conditions and eliminate health inequities facing communities. City officials can use findings from implementing these methods to inform the funding of green feature enhancements at parks in areas characterized by urban heat islands and poor health outcomes. In future analyses of the Green Schoolyards Project, we will further examine how trees, gardens, and nature trails at school parks impact heat index within parks and physical activity levels of predominantly Latinx children from low-income families; and how these children’s connection to nature relates to their social-emotional learning skills, misconduct, and standardized test scores. If we find these green features to decrease heat index within parks, increase children’s physical activity levels, and/or exhibit positive associations with children’s psychosocial and academic outcomes, then we can recommend future installments of green features in school parks for child health in a warming world. List Of Abbreviations List Of Abbreviations Page 10/19 Page 10/19 Page 10/19 HOBO: Not an abbreviation/accronym SOPARC: System for Observing Play and Recreation in Communities SOPARC: System for Observing Play and Recreation in Communities US: United States US: United States Acknowledgements: Not applicable Acknowledgements: Not applicable Declarations Ethics Approval and Consent to Participate: Prior to undertaking any project activities, we received approval of project protocols by the institutional review board at The University of Texas Health Science Center at Houston (HSC-SPH-19- 0502) and the school district. We received informed consent from participants’ parents and written assent from study participants. Consent for Publication: Not applicable Consent for Publication: Not applicable Availability of Data and Materials: The datasets used and analyzed during the current study are available from the corresponding author on reasonable request. Competing Interests: The authors declare that they have no competing interests. Competing Interests: The authors declare that they have no competing interests. Funding: This work was supported by the Robert Wood Johnson Foundation [grant number 76576]. The authors also received partial funding from the Michael & Susan Dell Foundation to the Michael & Susan Dell Center for Healthy Living, and City of Austin Parks & Recreation Department. Sponsors had no involvement in study design; in the collection, analysis and interpretation of data; in the writing of the report; and in the decision to submit the article for publication. Authors' Contributions: KL contributed to conceptualization, methodology, formal analysis, investigation, writing of the original manuscript draft, development of visual materials, project administration, and funding acquisition. MA contributed to conceptualization, investigation, reviewing and editing of the manuscript draft, project administration, and funding acquisition. DH and HK both contributed to conceptualization, reviewing and editing of the manuscript draft, and supervision. All authors read and approved the final manuscript. References 1. Louv R. Last child in the woods: Saving our children from nature-deficit disorder: Algonquin books; 2008. 1. Louv R. Last child in the woods: Saving our children from nature-deficit disorder: A 1. Louv R. Last child in the woods: Saving our children from nature-deficit disorder: Algonquin books; 2008. 2. Gill T. The benefits of children's engagement with nature: A systematic literature review. Children Youth and Environments. 2014;24(2):10-34. 2. Gill T. The benefits of children's engagement with nature: A systematic literature review. Children Youth and Environments. 2014;24(2):10-34. 3. Kellert SR, Case DJ, Escher D, Witter DJ, Mikels-Carrasco J, Seng PT. The nature of Americans national report: Disconnection and recommendations for reconnection. 2017. 3. Kellert SR, Case DJ, Escher D, Witter DJ, Mikels-Carrasco J, Seng PT. The nature of Americans national report: Disconnection and recommendations for reconnection. 2017. https://natureofamericans.org/sites/default/files/reports/Nature-of-Americans_National_Report_1.3_4-26-17.pdf. Accessed 25 Mar 2019. Page 11/19 Page 11/19 Page 11/19 4. Nesbitt L, Meitner MJ, Girling C, Sheppard SR, Lu Y. Who has access to urban vegetation? A spatial analysis of distributional green equity in 10 US cities. Landscape and Urban Planning. 2019;181:51-79. 5. Author et al. 2019 [details removed for peer review]. 6. Basu R, Pearson D, Malig B, Broadwin R, Green R. The effect of high ambient temperature on emergency room visits. Epidemiology. 2012:813-20. 7. Hoffman JS, Shandas V, Pendleton N. The Effects of Historical Housing Policies on Resident Exposure to Intra- Urban Heat: A Study of 108 US Urban Areas. Climate. 2020;8(1):12. 8. Stone Jr B. The city and the coming climate: Climate change in the places we live: Cambridge University Press; 2012. 9. Belcher BR, Berrigan D, Dodd KW, Emken BA, Chou C-P, Spuijt-Metz D. Physical activity in US youth: Impact of race/ethnicity, age, gender, & weight status. Medicine and science in sports and exercise. 2010;42(12):2211. 10. Chang SH, Kim K. A review of factors limiting physical activity among young children from low-income families. Journal of exercise rehabilitation. 2017;13(4):375. 11. Piercy KL, Troiano RP, Ballard RM, Carlson SA, Fulton JE, Galuska DA, et al. The physical activity guidelines for Americans. Jama. 2018;320(19):2020-8. 12. US Centers for Disease Control and Prevention. Physical activity facts. 2020. https://www.cdc.gov/healthyschools/physicalactivity/facts.htm. Accessed 5 Mar 2020. 12. US Centers for Disease Control and Prevention. Physical activity facts. 2020. https://www cdc gov/healthyschools/physicalactivity/facts htm Accessed 5 Mar 2020 12. US Centers for Disease Control and Prevention. Physical activity facts. 2020. https://www.cdc.gov/healthyschools/physicalactivity/facts.htm. Accessed 5 M 13. Edwards NM, Myer GD, Kalkwarf HJ, Woo JG, Khoury PR, Hewett TE, et al. Outdoor temperature, precipitation, and wind speed affect physical activity levels in children: a longitudinal cohort study. Journal of Physical Activity and Health. 2015;12(8):1074-81.13. 13. Edwards NM, Myer GD, Kalkwarf HJ, Woo JG, Khoury PR, Hewett TE, et al. Outdoor temperature, precipitation, and wind speed affect physical activity levels in children: a longitudinal cohort study. Journal of Physical Activity and Health. 2015;12(8):1074-81.13. 14. Chawla L. Benefits of nature contact for children. Journal of Planning Literature. 2015;30(4):433-52. 14. Chawla L. Benefits of nature contact for children. Journal of Planning Literature. 2015;30(4):433-52. 15. Children & Nature Network. Green schoolyard report. 2016. https://www.childrenandnature.org/wp- content/uploads/2015/03/CNN_GSY_Report2016_Final.pdf. Accessed 7 Mar 2020. 15. Children & Nature Network. Green schoolyard report. 2016. https://www.childrenandnature.org/wp- content/uploads/2015/03/CNN_GSY_Report2016_Final.pdf. Accessed 7 Mar 2020. 6. Author et al. 2019 [details removed for peer review]. 16. Author et al. Page 11/19 2019 [details removed for peer review]. 17. Aram F, Solgi E, Garcia EH, Mosavi A. Urban heat resilience at the time of global warming: evaluating the impact of the urban parks on outdoor thermal comfort. Environmental Sciences Europe. 2020;32(1):1-15. 17. Aram F, Solgi E, Garcia EH, Mosavi A. Urban heat resilience at the time of global warming: evaluating the impact of the urban parks on outdoor thermal comfort. Environmental Sciences Europe. 2020;32(1):1-15. 18. Van Dijk-Wesselius J, Maas J, Hovinga D, Van Vugt M, Van den Berg A. The impact of greening schoolyards on the appreciation, and physical, cognitive and social-emotional well-being of schoolchildren: A prospective intervention study. Landscape and urban planning. 2018;180:15-26. 18. Van Dijk-Wesselius J, Maas J, Hovinga D, Van Vugt M, Van den Berg A. The impact of greening schoolyards on the appreciation, and physical, cognitive and social-emotional well-being of schoolchildren: A prospective intervention study. Landscape and urban planning. 2018;180:15-26. 19. Chawla L, Keena K, Pevec I, Stanley E. Green schoolyards as havens from stress and resources for resilience in childhood and adolescence. Health & place. 2014;28:1-13. 19. Chawla L, Keena K, Pevec I, Stanley E. Green schoolyards as havens from stress and resources for resilience in childhood and adolescence. Health & place. 2014;28:1-13. 20. Raith A. Contact with nature in green schoolyards. Children, Youth and Environmen Contact with nature in green schoolyards. Children, Youth and Environments. 2018;28 21. Bell AC, Dyment JE. Grounds for health: the intersection of green school grounds and health‐promoting schools. Environmental Education Research. 2008;14(1):77-90. 21. Bell AC, Dyment JE. Grounds for health: the intersection of green school grounds and health‐promoting schools. Environmental Education Research. 2008;14(1):77-90. 22. Author et al. 2015 [details removed for peer review]. et al. 2015 [details removed for peer review]. 22. Author et al. 2015 [details removed for peer review]. 23. Kaplan R, Kaplan S. The experience of nature: A psychological perspective: CUP Archive; 1989. 23. Kaplan R, Kaplan S. The experience of nature: A psychological perspective: CUP Archive; 1989. 24. Kaplan S. The restorative benefits of nature: Toward an integrative framework. Journal of environmental psychology. 1995;15(3):169-82. 24. Kaplan S. The restorative benefits of nature: Toward an integrative framework. Journal of environmental psychology. 1995;15(3):169-82. 25. City of Austin. Cities connecting children to nature: GIS gap analysis. 2019. https://austin.maps.arcgis.com/apps/MapJournal/index.html?appid=f77ea201c8d04425a1942f5e5f234a6c. Accessed 15 Feb 2019. 25. City of Austin. Cities connecting children to nature: GIS gap analysis. 2019. https://austin.maps.arcgis.com/apps/MapJournal/index.html?appid=f77ea201c8d04425a1942f5e5f234a6c. Accessed 15 Feb 2019. Page 11/19 https://austin.maps.arcgis.com/apps/MapJournal/index.html?appid=f77ea201c8d04425a1942f5e5f234a6c. Accessed 15 Feb 2019. Page 12/19 Page 12/19 26. City of Austin. City_of_austin_parks_cccn. 2016. https://services.arcgis.com/0L95CJ0VTaxqcmED/ArcGIS/rest/services/CCCN_cityofaustinparks/FeatureServer/0. Accessed 7 Apr 2019. 27. City of Austin. Wildflower meadow. 2020. http://www.austintexas.gov/page/wildflower-meadow. Accessed 10 Oct 2020. 28. National Wildlife Federation. Certified wildlife habitat. 2020. https://www.nwf.org/certifiedwildlifehabitat. Accessed 10 Oct 2020. 29. Parmehr EG, Amati M, Taylor EJ, Livesley SJ. Estimation of urban tree canopy cover using random point sampling and remote sensing methods. Urban Forestry & Urban Greening. 2016;20:160-71. 30. US National Weather Service Forecast Office. NOW Data - NOAA online weather data. 2019. 30. US National Weather Service Forecast Office. NOW Data - NOAA online weather data. 2019. https://w2.weather.gov/climate/xmacis.php?wfo=ewx. Accessed 25 Feb 2019. 31. US National Weather Service. What is the heat index? 2020. https://www.weather.gov/ama/heatindex. Accessed 9 Apr 2020. 32. Matias M, Lopes A. Surface Radiation Balance of Urban Materials and Their Impact on Air Temperature of an Urban Canyon in Lisbon, Portugal. Applied Sciences. 2020;10(6):2193. 33. Carroll ML, Brown ME, Wooten MR, Donham JE, Hubbard AB, Ridenhour WB. In situ air temperature and humidity measurements over diverse land covers in Greenbelt, Maryland, November 2013–November 2015. Earth System Science Data. 2016;8(2):415. 34. Hawkins TW, Brazel AJ, Stefanov WL, Bigler W, Saffell EM. The role of rural variability in urban heat island determination for Phoenix, Arizona. Journal of Applied Meteorology. 2004;43(3):476-86. 35. Sheng L, Tang X, You H, Gu Q, Hu H. Comparison of the urban heat island intensity quantified by using air temperature and Landsat land surface temperature in Hangzhou, China. Ecological Indicators. 2017;72:738-46. 36. Mildrexler DJ, Zhao M, Running SW. A global comparison between station air temperatures and MODIS land surface temperatures reveals the cooling role of forests. Journal of Geophysical Research: Biogeosciences. 2011;116(G3). 37. White-Newsome JL, Brines SJ, Brown DG, Dvonch JT, Gronlund CJ, Zhang K, et al. Validating satellite-derived land surface temperature with in situ measurements: A public health perspective. Environmental health perspectives. 2013;121(8):925-31. 38. Onset Computer Corporation. HOBO external temperature/RH sensor data logger. 2019. https://www.onsetcomp.com/products/data-loggers/mx2302a. Accessed 19 Mar 2019. p p p gg https://www.onsetcomp.com/products/data-loggers/mx2302a. Accessed 19 Mar 2019. 39. Whiteman CD, Hoch SW. Pseudovertical temperature profiles in a broad valley from lines of temperature sensors on sidewalls. Journal of Applied Meteorology and Climatology. 2014;53(11):2430-7. 40. Meier F, Fenner D, Grassmann T, Otto M, Scherer D. Crowdsourcing air temperature from citizen weather stations for urban climate research. Urban Climate. 2017;19:170-91. 41. US Department of Agriculture Farm Service Agency. NAIP imagery. 2019. Page 11/19 https://www.fsa.usda.gov/programs- and-services/aerial-photography/imagery-programs/naip-imagery/. Accessed 19 Jul 2019. 42. Kremer P, Hamstead ZA, McPhearson T. A social–ecological assessment of vacant lots in New York City. Landscape and Urban Planning. 2013;120:218-33. 43. McKenzie TL, Cohen DA, Sehgal A, Williamson S, Golinelli D. System for Observing Play and Recreation in Communities (SOPARC): reliability and feasibility measures. Journal of Physical Activity and Health. 2006;3(s1):S208-S22. Page 13/19 Page 13/19 44. Dowdell K, Gray T, Malone K. Nature and its influence on children’s outdoor play. Journal of Outdoor and Environmental Education. 2011;15(2):24-35. 45. Remmers T, Thijs C, Ettema D, de Vries S, Slingerland M, Kremers S. Critical hours and important environments: Relationships between afterschool physical activity and the physical environment using GPS, GIS and accelerometers in 10–12-year-old children. International journal of environmental research and public health. 2019;16(17):3116. 46. Borghese MM, Janssen I. Duration and intensity of different types of physical activity among children aged 10–13 years. Canadian Journal of Public Health. 2019;110(2):178-86. 46. Borghese MM, Janssen I. Duration and intensity of different types of physical activity among children aged 10–13 years. Canadian Journal of Public Health. 2019;110(2):178-86. 47. Andersen HB, Christiansen LB, Pawlowski CS, Schipperijn J. What we build makes a difference–Mapping activating schoolyard features after renewal using GIS, GPS and accelerometers. Landscape and Urban Planning. 2019;191:103617. 47. Andersen HB, Christiansen LB, Pawlowski CS, Schipperijn J. What we build makes a difference–Mapping activating schoolyard features after renewal using GIS, GPS and accelerometers. Landscape and Urban Planning. 2019;191:103617. 48. Pawlowski CS, Andersen HB, Troelsen J, Schipperijn J. Children’s physical activity behavior during school recess: A pilot study using GPS, accelerometer, participant observation, and go-along interview. PloS one. 2016;11(2). 48. Pawlowski CS, Andersen HB, Troelsen J, Schipperijn J. Children’s physical activity behavior during school recess: A pilot study using GPS, accelerometer, participant observation, and go-along interview. PloS one. 2016;11(2). 49. Schultz PW. Inclusion with nature: The psychology of human-nature relations. Psychology of sustainable development: Springer; 2002. p. 61-78. 49. Schultz PW. Inclusion with nature: The psychology of human-nature relations. Psychology of sustainable development: Springer; 2002. p. 61-78. 50. Cheng JC-H, Monroe MC. Connection to nature: Children’s affective attitude toward nature. Environment and Behavior. 2012;44(1):31-49. 50. Cheng JC-H, Monroe MC. Connection to nature: Children’s affective attitude toward nature. Environment and Behavior. 2012;44(1):31-49. 51. Bragg R, Wood C, Barton J, Pretty J. Measuring connection to nature in children aged 8-12: A robust methodology for the RSPB. University of Essex. 2013. 51. Page 11/19 Kamoen N, Holleman B, van den Bergh H, Sanders T, editors. Positive, negative, and bipolar questions: The effect of question polarity on ratings of text readability. Survey Research Methods; 2013. Page 11/19 Bragg R, Wood C, Barton J, Pretty J. Measuring connection to nature in children aged 8-12: A robust methodology for the RSPB. University of Essex. 2013. 52. Author et al. 2016 [details removed for peer review]. 52. Author et al. 2016 [details removed for peer review]. 53. Rothfusz LP. The heat index equation (or, more than you ever wanted to know about heat index). Fort Worth, Texas: National Oceanic and Atmospheric Administration, National Weather Service, Office of Meteorology. 1990;9023. 53. Rothfusz LP. The heat index equation (or, more than you ever wanted to know about heat index). Fort Worth, Texas: National Oceanic and Atmospheric Administration, National Weather Service, Office of Meteorology. 1990;9023. 54. Steadman RG. A universal scale of apparent temperature. Journal of Climate and Applied Meteorology. 1984;23(12):1674-87. 54. Steadman RG. A universal scale of apparent temperature. Journal of Climate and Applied Meteorology. 1984;23(12):1674-87. 55. US National Weather Service Weather Prediction Center. The heat index equation. 2014. https://www.wpc.ncep.noaa.gov/html/heatindex_equation.shtml. Accessed 16 Mar 2020. q https://www.wpc.ncep.noaa.gov/html/heatindex_equation.shtml. Accessed 16 Mar 2020. 56. Honjo T, Yamato H, Mikami T, Grimmond C. Network optimization for enhanced resilience of urban heat island measurements. Sustainable Cities and Society. 2015;19:319-30. 56. Honjo T, Yamato H, Mikami T, Grimmond C. Network optimization for enhanced resilience of urban heat island measurements. Sustainable Cities and Society. 2015;19:319-30. 57. Melillo JM, Richmond T, Yohe GW. Climate change impacts in the United States: the third national climate assessment. 2017. 57. Melillo JM, Richmond T, Yohe GW. Climate change impacts in the United States: the third national climate assessment. 2017. 58. Stevens JC, Choo KK. Temperature sensitivity of the body surface over the life span. Somatosensory & Motor Research. 1998. 59. Vanos JK, Middel A, McKercher GR, Kuras ER, Ruddell BL. Hot playgrounds and children's health: a multiscale analysis of surface temperatures in Arizona, USA. Landscape and Urban Planning. 2016;146:29-42. 60. US National Oceanic and Atmospheric Administration National Weather Service. Heat index. 2020. https://www.weather.gov/safety/heat-index. Accessed 18 Mar 2020. 61. Author et al. 2016 [details removed for peer review]. 55(now 16). Author et al. 2019 [details removed for peer review]. 62. Cohen DA, Han B, Williamson S, Nagel C, McKenzie TL, Evenson KR, et al. Playground features and physical activity in US neighborhood parks. Preventive Medicine. 2020;131:105945. Page 14/19 Page 14/19 63. Author et al. 2018 [details removed for peer review]. 63. Author et al. 2018 [details removed for peer review]. 63. Author et al. Page 11/19 2018 [details removed for peer review]. 63. Author et al. 2018 [details removed for peer review]. 64. Wells NM, Myers BM, Henderson Jr CR. School gardens and physical activity: A randomized controlled trial of low- income elementary schools. Preventive medicine. 2014;69:S27-S33. 64. Wells NM, Myers BM, Henderson Jr CR. School gardens and physical activity: A randomized controlled trial of low- income elementary schools. Preventive medicine. 2014;69:S27-S33. 65. Luo W, Taylor M, Parker S. A comparison of spatial interpolation methods to estimate continuous wind speed surfaces using irregularly distributed data from England and Wales. International Journal of Climatology: A Journal of the Royal Meteorological Society. 2008;28(7):947-59. 65. Luo W, Taylor M, Parker S. A comparison of spatial interpolation methods to estimate continuous wind speed surfaces using irregularly distributed data from England and Wales. International Journal of Climatology: A Journal of the Royal Meteorological Society. 2008;28(7):947-59. 66. Venter ZS, Brousse O, Esau I, Meier F. Hyperlocal mapping of urban air temperature using remote sensing and crowdsourced weather data. Remote Sensing of Environment. 2020;242:111791. 66. Venter ZS, Brousse O, Esau I, Meier F. Hyperlocal mapping of urban air temperature using remote sensing and crowdsourced weather data. Remote Sensing of Environment. 2020;242:111791. 67. Cosco NG, Moore RC, Islam MZ. Behavior mapping: a method for linking preschool physical activity and outdoor design. Medicine & Science in Sports & Exercise. 2010;42(3):513-9. 67. Cosco NG, Moore RC, Islam MZ. Behavior mapping: a method for linking preschool physical activity and outdoor design. Medicine & Science in Sports & Exercise. 2010;42(3):513-9. 68. Jann B, Krumpal I, Wolter F. Social Desirability Bias in Surveys–Collecting and Analyzing Sensitive Data. Special Issue. 2019. 68. Jann B, Krumpal I, Wolter F. Social Desirability Bias in Surveys–Collecting and Analyzing Sensitive Data. Special Issue. 2019. 69. Allen M. The SAGE encyclopedia of communication research methods: Sage Publications; 2017. 69. Allen M. The SAGE encyclopedia of communication research methods: Sage Publications; 2017. 69. Allen M. The SAGE encyclopedia of communication research methods: Sage Pu 70. Salazar MS. The dilemma of combining positive and negative items in scales. Psicothema. 2015;27(2):192-9. 70. Salazar MS. The dilemma of combining positive and negative items in scales. Psicothema. 2015;27(2):192-9. 71. Kamoen N, Holleman B, van den Bergh H, Sanders T, editors. Positive, negative, and bipolar questions: The effect of question polarity on ratings of text readability. Survey Research Methods; 2013. 71. Figures Page 15/19 Figure 1 Conceptual model of Green Schoolyards Project. Figure 1 Conceptual model of Green Schoolyards Project. Conceptual model of Green Schoolyards Project. Conceptual model of Green Schoolyards Project. Page 16/19 Figure 2 HOBO MX2302A external air temperature/relative humidity sensor data logger on swing set at low-green park. al air temperature/relative humidity sensor data logger on swing set at low-green park Page 17/19 Page 17/19 Page 17/19 Figure 3 HOBO sensors and SOPARC target areas at (A1–A2) intervention, (B1–B2) low-green, and (C1–C2) Figure 3 HOBO sensors and SOPARC target areas at (A1–A2) intervention, (B1–B2) low-green, and (C1–C2) high-green parks. Page 18/19 Figure 4 Thermal profiles for (A) intervention, (B) low-green Supplementary Files Thi i li f l fil i d i Figure 4 Thermal profiles for (A) intervention, (B) low-green, and (C) high-green parks, September 16–30th 2019. Supplementary Files Figure 4 Thermal profiles for (A) intervention, (B) low-green, and (C) high-green parks, September 16–30th 2019. Thermal profiles for (A) intervention, (B) low-green, and (C) high-green parks, September 16–30th 2019. Supplementary Files This is a list of supplementary files associated with this preprint. Click to download. AdditionalFile1.pdf AdditionalFile2.pdf Page 19/19 Page 19/19
https://openalex.org/W3115293794	https://www.frontiersin.org/articles/10.3389/fphys.2020.607544/pdf	English	null	Monitoring Muscle-Tendon Adaptation Over Several Years of Athletic Training and Competition in Elite Track and Field Jumpers	Frontiers in physiology	2,020	cc-by	11,000	ORIGINAL RESEARCH published: 16 December 2020 doi: 10.3389/fphys.2020.607544 Monitoring Muscle-Tendon Adaptation Over Several Years of Athletic Training and Competition in Elite Track and Field Jumpers Kiros Karamanidis* and Gaspar Epro Sport and Exercise Science Research Centre, School of Applied Sciences, London South Bank University, London, United Kingdom Differences in muscle and tendon responsiveness to mechanical stimuli and time courses of adaptive changes may disrupt the interaction of the musculotendinous unit (MTU), increasing the risk for overuse injuries. We monitored training-induced alterations in muscle and tendon biomechanical properties in elite jumpers over 4 years of athletic training to detect potential non-synchronized adaptations within the triceps surae MTU. A combined cross-sectional and longitudinal investigation over 4 years was conducted by analyzing triceps surae MTU mechanical properties in both legs via dynamometry and ultrasonography in 67 elite track and ﬁeld jumpers and 24 age-matched controls. Fluctuations in muscle and tendon adaptive changes over time were quantiﬁed by calculating individual residuals. The cosine similarity of the relative changes of muscle strength and tendon stiffness between sessions served as a measure of uniformity of adaptive changes. Our cross-sectional study was unable to detect clear non-concurrent differences in muscle strength and tendon stiffness in elite jumpers. However, when considering the longitudinal data over several years of training most of the jumpers demonstrated greater ﬂuctuations in muscle strength and tendon stiffness and hence tendon strain compared to controls, irrespective of training period (preparation vs. competition). Moreover, two monitored athletes with chronic Achilles tendinopathy showed in their affected limb lower uniformity in MTU adaptation as well as higher ﬂuctuations in tendon strain over time. Habitual mechanical loading can affect the MTU uniformity in elite jumpers, leading to increased mechanical demand on the tendon over an athletic season and potentially increased risk for overuse injuries. Edited by: Hassane Zouhal, University of Rennes 2–Upper Brittany, France Reviewed by: Lauri Stenroth, University of Eastern Finland, Finland Kirsten Legerlotz, Humboldt University of Berlin, Germany Correspondence: Kiros Karamanidis k.karamanidis@lsbu.ac.uk Specialty section: This article was submitted to Exercise Physiology, a section of the journal Frontiers in Physiology Specialty section: This article was submitted to Exercise Physiology, a section of the journal Frontiers in Physiology Received: 17 September 2020 Accepted: 27 November 2020 Published: 16 December 2020 Received: 17 September 2020 Accepted: 27 November 2020 P bli h d 16 D b 2020 Accepted: 27 November 2020 Published: 16 December 2020 Keywords: Achilles tendon, triceps surae, time course, imbalance, athletic training, tendinopathy INTRODUCTION Published: 16 December 2020 Citation: This may be explained by discrepancies between muscle and tendon in their adaptive response to the experienced mechanical stimuli (Arampatzis et al., 2007a) and in the timescale of adaptation (Kubo et al., 2012) with distinctively slower response rates for tendons due to lower tissue turnover (Heinemeier et al., 2013). Hence, an imbalance between the generated muscle forces and the strength of the tendon may occur over the course of training process, placing the tendon under greater mechanical demand (i.e., higher strain). fact that athletic training for elite athletes consists of complex loading variations and includes diﬀerent training modalities (Haugen et al., 2019) could lead to considerable ﬂuctuations in muscle strength in the intermediate term, which may have implications not only for the performance but also for the integrity of the tendon. Regarding that a clear pathogenesis of the development of tendon overuse is still unknown, the above mentioned provides a more mechanistic explanation along other more physiological concepts behind pathological changes in tendon (e.g., immunoregulation/excessive immune response). g g p The Achilles tendon (AT) is susceptible to injury potentially due its noticeably low safety factor (the ratio of ultimate and operational stress) when compared with other tendons (Ker et al., 1988; Magnusson et al., 2001). Accordingly it does not seem so unexpected that there is a high prevalence of AT tendinopathies in adult athletes (Janssen et al., 2018). This supports the assumption that periods of asynchronous muscle-tendon adaptation in elite jumping athletes might be of clinical relevance and that further research is needed to explore whether athletic training is an inﬂuential factor in non-uniform adaptation between muscle and tendon. To our knowledge the time course of muscle and tendon adaptations in adult athletes over several years of athletic training and competition is unknown. The aim of the current investigation was to regularly monitor training-induced alterations in triceps surae (TS) muscle strength, tendon stiﬀness and tendon strain in the legs of elite jumpers over several years of training in order to detect potential non-uniformities within the TS MTU due to habitual mechanical loading. Potential alterations in TS MTU properties in elite athletes were compared with the variations in a control group of healthy recreationally active adults of similar age range. Citation: Karamanidis K and Epro G (2020) Monitoring Muscle-Tendon Adaptation Over Several Years of Athletic Training and Competition in Elite Track and Field Jumpers. Front. Physiol. 11:607544. doi: 10.3389/fphys.2020.607544 Karamanidis K and Epro G (2020) Monitoring Muscle-Tendon Adaptation Over Several Years of Athletic Training and Competition in Elite Track and Field Jumpers. Front. Physiol. 11:607544. doi: 10.3389/fphys.2020.607544 Tendons are an integral part of musculotendinous units (MTUs) and their primary function is to transmit forces from muscles to rigid bone levers to produce joint motion. As a part of their viscoelastic nature, elasticity of the tendons can enhance force and power generating capacity of MTUs by storing and releasing strain energy and by allowing muscles to operate under more favorable conditions related to the force-length-velocity relationship (Biewener and Roberts, 2000; Hof et al., 2002). Thus the mechanical properties of tendons need to be considered as important December 2020 \| Volume 11 \| Article 607544 Frontiers in Physiology \| www.frontiersin.org Monitoring MTU Adaptation in Jumpers Karamanidis and Epro determinants of musculoskeletal function, and it is not surprising that they have been shown to aﬀect athletic performance (Bojsen- Møller et al., 2005; Staﬁlidis and Arampatzis, 2007). Not only are tendons vitally important for movement and sports performance, movement and exercise are also essential for tendons. As living tissue, tendons respond to mechanical loading by changing their metabolism, which can lead to alterations in their material composition (e.g., changes in extra-cellular matrix or cross- linking) and their morphological properties (cross-sectional area) with the latter as a more long-term response (Bohm et al., 2015a; Wiesinger et al., 2015). Accordingly, when a MTU is constantly experiencing increased mechanical loading (e.g., by resistance training) it is generally observed that the associated adaptive enhancements in muscle strength lead also to a markedly greater tendon stiﬀness (Kubo et al., 2001, 2012; Arampatzis et al., 2007a,b; Bohm et al., 2014). The reported modiﬁcations in tendon stiﬀness may exhibit a protective homeostatic process to withstand greater functional demand due to muscle strength gains. Nevertheless, the muscle strength gains seem to be merely in a moderate association with modiﬁcations in tendon mechanical properties (Arampatzis et al., 2007a,b). Citation: On the basis that elite athletes are subjected to high variation in mechanical loading proﬁles over a season, we hypothesized that elite athletes experience greater ﬂuctuations in MTU mechanical properties over time and demonstrate temporary increases in mechanical demand on the tendon as a result of non-uniform MTU adaptation. For this reason we conducted a combined cross-sectional and longitudinal investigation by analyzing a total of 67 healthy young elite jump track and ﬁeld athletes competing at international level, from which a sub-group of 18 elite jumpers were monitored regularly over 1 year of training. As part of this long-term investigation six of the elite athletes were measured regularly over a 4-year period to investigate possible diﬀerences between the preparation and competition periods. Two additional elite athletes with unilateral AT tendinopathy and one athlete following unilateral AT reconstruction will be discussed separately. g Compared to tendons, muscles seem to possess superior adaptability in reaction to a broad variety of exercise modalities (Moss et al., 1997), the former responding most eﬀectively to mechanical loads creating high magnitudes of tendon strain over extended time durations (Arampatzis et al., 2007a; Bohm et al., 2014). These adaptive discrepancies within the MTU could become crucial for athletes constantly involved with high mechanical loading proﬁles in their training process, particularly in sport disciplines with frequent plyometric loading, such as track and ﬁeld sprinters and jumpers (Stefanyshyn and Nigg, 1997, 1998; Dorn et al., 2012; Willwacher et al., 2017). By its nature plyometric training is characterized by high magnitude mechanical loading, though the time spans over which those high forces are applied tend to be rather short. This might be a reason why various plyometric loading regimens do not always lead to clear adaptive increases in tendon stiﬀness along with muscle strength gains (Kubo et al., 2007; Fouré et al., 2010; Houghton et al., 2013; Bohm et al., 2014; Hirayama et al., 2017). From a biomechanical viewpoint, muscle strength improvements with insuﬃcient compensatory adaptations in tendon mechanical properties would potentially increase the mechanical demand for the tendon (i.e., higher strain). In fact, it has recently been demonstrated that habitual athletic training that consists mainly of plyometric loading aﬀects the uniformity of muscle and tendon adaptation in adolescents (Mersmann et al., 2016) with potential implications for injury and tendinopathy (Wren et al., 2003; Fredberg and Stengaard-Pedersen, 2008). The Frontiers in Physiology \| www.frontiersin.org Longitudinal Investigation Over 4 Years Longitudinal Investigation Over 4 Years Twelve jumpers from the initial pool of healthy elite athletes agreed to participate over an extended 4 year time period of TS MTU monitoring. However, only in six elite athletes it was possible to determine the bilateral MTU biomechanical properties about every three- to ﬁve-weeks over the 4 years period, therefore all other athletes were excluded from further analysis (dropped out due to missing time, injury or other reasons). Those regularly monitored six elite athletes served for the investigation of diﬀerences between the preparation and competition periods in bilateral MTU adaptive changes over 4 years. As a part of the day-to-day training routine at their corresponding Olympic training centers or training camps (Figure 1A), the TS MTU biomechanical properties (maximal ankle plantar ﬂexion moment, AT stiﬀness and maximal tendon strain) were assessed in both legs of all participating athletes, concurrently or precisely prior to the competition period (cross- sectional investigation). In order to distinguish between the limbs, the preferred jumping leg was considered as the take-oﬀ leg and the contralateral leg as the swing leg. The hop leg in triple jumpers was deﬁned as the take-oﬀleg. Similar to the athletes, measurements with the control group were performed on both limbs in the cross-sectional investigation (take-oﬀleg deﬁned via a questionnaire as the preferred jumping leg), whereas in the longitudinal investigation only the take-oﬀleg was analyzed. In the injured athletes, the comparison was made between the healthy and aﬀected (injured) limb. MATERIALS AND METHODS Participants and Experimental Design Participants and Experimental Design In this investigation on TS MTU adaptation we were able to voluntarily recruit 67 healthy young elite international level jumping event track and ﬁeld athletes (35 male and 32 female) from the German national team (age: 23 ± 4 years; body mass: 73 ± 11 kg; body height: 183 ± 9 cm; mean and December 2020 \| Volume 11 \| Article 607544 2 Monitoring MTU Adaptation in Jumpers Karamanidis and Epro standard deviation; for further description and personal records see Supplementary Material). Additionally, 24 young healthy recreationally active (12 male and 12 female; age: 24 ± 3 years; body mass: 72 ± 12 kg; body height: 177 ± 10 cm) acted as control group. Excluded were athletes with prior AT injuries (ruptures, tendinopathy etc.) during the preceding six months, as it may have possibly aﬀected the ﬁndings. Accordingly in total 91 subjects were included in our cross- sectional investigation. Next to the healthy participants, three additional athletes with current AT injuries (two athletes with chronic tendinopathy and one with a recent tendon rupture) voluntarily participated in the study and were separately included to the study. The study was ethically approved by the local committee of the German Sport University Cologne and all participants provided their written informed consent prior to the study in accordance to the recommendations in the Declaration of Helsinki. with AT rupture less frequent measurements were conducted). For the control group only the take-oﬀleg was considered, as previous studies have shown similar TS MTU mechanical properties between limbs in healthy adults (Bohm et al., 2015b) and thereby to reduce the amount of measurements taken. Analysis of Muscle and Tendon Mechanical Properties p The TS MTU biomechanical properties (maximal ankle plantar ﬂexion moment, AT stiﬀness and maximal AT strain) were assessed in all participants on a custom-made device (Figure 1B) using synchronous ultrasonography and dynamometry as in our previous study (Epro et al., 2019). Brieﬂy, the participants were positioned with their knee and ankle joints ﬁxed at 90◦ (with their foot and thigh perpendicular to the shank) and their foot set on a strain-gauge type dynamometer (1000Hz; TEMULAB R⃝, Protendon GmbH & Co. KG, Aachen, Germany; Figure 1B). The foot was set on the dynamometer (Figure 1B) with the midpoint of the malleolus lateralis aligned with the dynamometer’s axis of rotation (Ackermans et al., 2016). A laser-guided electrical potentiometer system measuring the linear displacement was used to position each participant in a standardized manner, which allowed recreating the exact position for each individual at any measurement time point. Prior to each measurement, all participants performed an individual warm-up, which was succeeded by a standardized visually guided warm- up in the software consisting of several submaximal and three maximal isometric contractions to “precondition” the tendon (Maganaris, 2003). Frontiers in Physiology \| www.frontiersin.org FIGURE 1 \| Measurement locations at the Olympic training centers in Germany incl. study design (A) and the used custom-made analysis device (TEMULAB®, Protendon GmbH & Co. KG, Aachen, Germany) to assess TS MTU mechanical properties incorporating dynamometry and ultrasonography (B). In order to determine the force-elongation relationship of the tendon the displacement of the myotendinous junction (MTJ) of the m. gastrocnemius medialis (1Lmeasured) was digitized at rest (0%) and at three sustained contractions (target joint moment constantly held for 3s) at ﬁxed target ankle joint moment levels (30, 50, and 80%; C). relation to the skin. The MTJ displacement was traced at the resting state (0% MVC) and in the following three sustained contractions (steadily held target joint moment for 3 s) at the target ankle joint moments (30, 50, 80% MVC; Figure 1C). The participants had to repeat the speciﬁc trial, when they were unsuccessful to hold a steady state for 3s within a range of ±5% of the target joint moment. A linear extrapolation of the elongation at 50 and 80% target joint moments was used to estimate the tendon elongation at maximal ankle joint moment (100%) as in Ackermans et al. (2016), because speciﬁc strict instructions about loading rate and maintaining the generated moments at certain given level may limit the ability to contract maximally. Moreover, this approach to calculate maximal tendon strain can be considered as a valid measure as tendon elongation between 50 and 100% of the MVC can be assumed to be rather linear and because a considerable amount of the net elongation occurs within the ﬁrst 25% of the MVC (McCrum et al., 2018). The eﬀect of the inevitable ankle joint angular rotation on the measured tendon elongation (Muramatsu et al., 2001) was taken into account during each contraction by subtracting the elongation caused by ankle joint changes. Ankle joint angle during each contraction was obtained through the inverse tangent of the ratio between the heel lift (measured via a heel-potentiometer) and the distance from the ankle joint axis to the ﬁfth metatarsal bone (Ackermans et al., 2016). Longitudinal Investigation Over 1 Year o g ud a es ga o O e ea Due to the a lack of appropriate longitudinal data of MTU adaptive changes in elite track and ﬁeld athletes a power analysis for a target sample size was not possible, therefore it was aimed to recruit as many participants as possible. Out of the 91 healthy participants from our cross-sectional investigation, we were able to recruit 40 elite athletes and 15 controls to participate over a minimum of 1 year observation period. However, only 18 athletes and 12 controls were considered as valid for our longitudinal investigation over 1 year by reaching our criteria for the statistical analysis (set criteria: regular measurement about every three- to ﬁve-weeks) while the rest dropped out due to missing time, various injuries (including lower limb MTU injuries) or due to other reasons. Next to the 18 monitored healthy athletes we were able to recruit and regularly measure two elite athletes with a chronic unilateral AT tendinopathy and one elite jumping athlete 10 months after AT rupture and reconstruction monitored over a follow up time period of 2 years. The data of these three athletes with AT injury were not included in the statistical analysis of the pool of healthy elite athletes and were considered separately. For the statistical analysis of the elite athletes (n = 21), measurements were conducted on both limbs at every three- to ﬁve-weeks over the 1 year observation window (only for the elite athlete Subsequently, the maximal ankle plantar ﬂexion moment as well as the force–elongation relationship of the tendon during the loading phase were assessed by performing isometric ankle plantar ﬂexion contractions at diﬀerent joint moment levels: three verbally encouraged maximal voluntary ankle plantar ﬂexion contractions (MVC), succeeded by three visually guided sustained contractions at 30, 50, and 80% of the maximal joint moment obtained in the prior MVCs. The resultant joint moments acting about the ankle joint were calculated using inverse dynamics and a preceding passive measurement (relaxed muscles in the ﬁxed position) was performed to account for the gravitational moments. The alignment of the axes of rotation of the ankle and the dynamometer (Figure 1B) allowed to consider the ankle joint moment to be quasi-equal with the acquired moment of the force plate (Ackermans et al., 2016). Frontiers in Physiology \| www.frontiersin.org Longitudinal Investigation Over 1 Year Nevertheless, December 2020 \| Volume 11 \| Article 607544 3 Monitoring MTU Adaptation in Jumpers Karamanidis and Epro FIGURE 1 \| Measurement locations at the Olympic training centers in Germany incl. study design (A) and the used custom-made analysis device (TEMULAB®, Protendon GmbH & Co. KG, Aachen, Germany) to assess TS MTU mechanical properties incorporating dynamometry and ultrasonography (B). In order to determine the force-elongation relationship of the tendon the displacement of the myotendinous junction (MTJ) of the m. gastrocnemius medialis (1Lmeasured) was digitized at rest (0%) and at three sustained contractions (target joint moment constantly held for 3s) at ﬁxed target ankle joint moment levels (30, 50, and 80%; C). Statistics The normality of distribution and homogeneity of variance of the data was conﬁrmed using the Shapiro-Wilk and Levene’s test. For the cross-sectional analysis, a two-way repeated measures analysis of variance (ANOVA) was performed with the group (athletes vs. control) and leg (take-oﬀvs. swing leg) as factors to investigate potential diﬀerences in TS muscle strength, AT stiﬀness and maximal AT strain. For the longitudinal investigations, ﬁrstly the uniformity of muscle and tendon adaptive changes over the 1 year time period in both elite athletes and control subjects were analyzed using individual absolute residuals (averaged over all measurement sessions) from a general linear model as a measure of ﬂuctuations in TS MTU properties and secondly by forming cosine similarity (CS) of the relative changes between the individual measurement sessions (Mersmann et al., 2016): For the cross-sectional investigation 67 healthy elite jumpers (35 male and 32 female) and 24 recreationally active young adults (12 male and 12 female; control group) were considered for statistical analysis with respect to potential leg and subject group eﬀects on MTU properties. Irrespective of the analyzed limb (take-oﬀvs. swing), elite jumpers demonstrated in comparison to the control group signiﬁcantly higher TS muscle strength (average values of both legs: athletes 303 ± 85 vs. controls 241 ± 65 Nm; subject group eﬀect: F(1, 89) = 10.415, P = 0.002, η 2 p = 0.105) as well as greater tendon stiﬀness (653 ± 187 vs. 570 ± 131 N/mm; F(1, 89) = 4.004, P = 0.048, η 2 p = 0.043) with similar relative diﬀerence between groups (∼20% in TS muscle strength vs. ∼15% in tendon stiﬀness). Hence, maximal tendon strain was not signiﬁcantly diﬀerent between groups independent of the analyzed leg (average values of both legs: athletes 4.8 ± 1.1 vs. controls 4.8 ± 1.0%). Concerning the between leg analysis, we found limb-speciﬁc diﬀerences in TS MTU mechanical properties in jumpers but not in controls. Elite jumpers showed signiﬁcantly higher muscle strength (310 ± 89 vs. 296 ± 84 Nm, P = 0.001, η 2 p = 0.127) and tendon stiﬀness (675 ± 195 vs. 630 ± 186 N/mm, P < 0.001, η 2 p = 0.226) in their take-oﬀin comparison to their swing leg. No signiﬁcant leg-speciﬁc diﬀerences were detected for maximal AT strain in the elite jumpers (take-oﬀleg 4.6 ± 1.0 vs. swing leg 4.8 ± 1.1%). Statistics A signiﬁcant group-diﬀerence between the elite jumpers and control group was found for the body height (183 ± 9 vs. 177 ± 10 cm for elite jumpers and control respectively, P = 0.008), but not for age (23 ± 4 vs. 24 ± 3 years) and body mass (73 ± 11 vs. 72 ± 12 kg). CS = cos (θ) = X · Y ∥X∥· ∥Y∥= Pn−1 i=1 XiYi qPn−1 i=1 X2 i qPn−1 i=1 Y2 i CS = cos (θ) = X · Y ∥X∥· ∥Y∥= Pn−1 i=1 XiYi qPn−1 i=1 X2 i qPn−1 i=1 Y2 i X and Y denote the vectors of the relative changes between sessions (i = 1,..., n) and n refers to the individual total number of measurements. CS values range between 1 (illustrating equal direction of the vectors, i.e., high adaptation similarity) to−1 (opposing direction, i.e., low adaptation similarity). For the 4 years investigation (eight consecutive seasons), the residuals and CS were averaged separately for the preparation (mid-March to mid-May and start of October to end of December) and competition periods (start of January to mid-March and mid of May to mid-September). A one-way ANOVA was implemented to investigate potential subject group-related diﬀerences (take- oﬀleg vs. swing leg of athletes vs. take-oﬀleg controls) in the ﬂuctuations (residuals) and CS of muscle and tendon biomechanical properties over the 1 year time period. Potential training period (preparation vs. competition period) inﬂuence in the ﬂuctuations and CS in MTU adaptation over 4 years of athletic training was examined using a two-way ANOVA with leg and training period as factors. In the case of a detected signiﬁcant main eﬀect or interaction a Bonferroni post hoc comparison was performed. In addition, in order to evaluate the strength of potential training-induced alterations in cross-sectional and longitudinal investigations the partial eta squared (η 2 p ) was calculated (Cohen, 2013). Potential subject group diﬀerences in age, body mass and body height between elite athletes and controls were compared using an independent samples t-test. All statistical calculations and analyses were performed using custom-made MATLAB scripts (version 2020b; The MathWorks, Natick, MA, USA) or SPSS statistics software (version 26.0; IBM, Armonk, NY, USA) with the results in the text and ﬁgures provided either as means and standard deviation (SD) or as Subsequently, the ratio between the change in the calculated tendon force and the resultant tendon elongation between 30 and 80% of maximum tendon force was it needs to be noted, that the resultant ankle plantar ﬂexion moment is an approximation of the net moment generated by the TS muscle, as the current setup did not account for the individual moment contributions of the synergistic agonist muscles and the antagonist dorsiﬂexors. Subsequently, the resultant ankle joint moment was further used to calculate the acting AT force by dividing it by the tendon moment arm acquired from a previous study (Maganaris et al., 1998). In order to exclude the eﬀect of any potential changes in body mass on the calculations of ﬂuctuations in muscle strength over time, we used the absolute maximal joint moments (Nm) instead of body mass normalized moments. After tendon preconditioning, the AT resting length was measured using the integrated laser-guided electrical potentiometer system as the distance from the myotendinous junction (MTJ) of the m. gastrocnemius medialis to the most proximal point of the tuber calcanei (both deﬁned and marked beforehand using ultrasonography). Similar to the positioning of each participant on the dynamometer, the laser-guided electrical potentiometer system allowed maintaining the probe position constant for each subject and leg during all measurements. During the following plantar ﬂexion contractions the movement of the MTJ of the m. gastrocnemius medialis was analyzed with a ﬁrmly ﬁxed 40 mm linear array ultrasound probe (27 Hz; MyLabTMOne, Esaote; Genoa, Italy) and manual digitization using a custom-made software in Python (ver. 2.7.0; Figure 1C). The ultrasound probe was secured above the MTJ on the shank using a casing with adjustable straps to avoid movement in December 2020 \| Volume 11 \| Article 607544 4 Monitoring MTU Adaptation in Jumpers Karamanidis and Epro boxplots (median and interquartile range between 25th and 75th percentile along with minimum and maximum values) and the level of signiﬁcance set at α = 0.05. used to estimate the AT stiﬀness. Before conducting the current study we performed several pilot studies to test the accuracy and validity of the implemented method to assess the TS MTU mechanical properties. The main ﬁndings of these pilot studies are provided in the Supplementary Material under the chapter “Methodological considerations and pilot data.” TS MTU Adaptive Changes Over 4 Years: Preparation vs. Competition Period For the longitudinal investigation over 4 years we were able to analyse six healthy male elite athletes on a regular basis (about every three- to ﬁve-weeks) over eight preparation and eight competition periods. Exemplarily Figure 3 represents the changes in TS muscle strength and tendon stiﬀness in both legs of one elite high jumper over a period of 4 years of athletic training separated into preparation and competition periods. The total number of measurements for those six elite athletes over the 4 years period ranged from 36 to 73 with 51% measurements in the preparation and 49% measurements TS MTU Adaptive Changes Over 1 Year: Elite Athletes vs. Controls Eighteen healthy athletes (8 males and 10 females) were considered as valid for our longitudinal investigation over 1 year as they reached our criteria for the longitudinal statistical analysis (set criteria: measurement about every three- to ﬁve- weeks). When considering all data points analyzed within the 12 months observation period we found greater ﬂuctuations of muscle strength over time in elite jumpers compared with controls irrespective of the analyzed leg (i.e., greater residuals; take-oﬀleg of athletes vs. take-oﬀleg control P = 0.010 and swing leg of athletes vs. take-oﬀleg control P = 0.049; Figure 2). Furthermore, although not reaching statistical signiﬁcance, there were tendencies for greater ﬂuctuation in tendon stiﬀness (P = 0.074; Figure 2) and in maximal tendon strain for elite athletes take-oﬀleg in comparison to control (P = 0.092; Figure 2). In addition, for the CS a main eﬀect was detected (P = 0.030), with the athletes take-oﬀleg in comparison to the December 2020 \| Volume 11 \| Article 607544 Frontiers in Physiology \| www.frontiersin.org 5 Monitoring MTU Adaptation in Jumpers Karamanidis and Epro FIGURE 2 \| Box plots of muscle strength (Max. PF moment), tendon stiffness (AT stiffness) and maximal tendon strain (Max. AT strain) as well as the cosine similarity (Cosine) of the relative changes of muscle strength and tendon stiffness between sessions over the 1 year observation period in 18 healthy elite jumpers (take-off and swing leg) and 12 healthy control subjects (take-off leg). Statistically signiﬁcant difference to control (P < 0.05). #Tendency for signiﬁcant difference to control (0.05 < p < 0.1). FIGURE 3 \| Adaptive changes in TS muscle strength (Max. PF moment) and tendon stiffness (AT stiffness) in the take-off and swing leg of one elite high jumper over a period of 4 years of athletic training separated into preparation (PREP) and competition (COMP) periods. FIGURE 2 \| Box plots of muscle strength (Max. PF moment), tendon stiffness (AT stiffness) and maximal tendon strain (Max. AT strain) as well as the cosine similarity (Cosine) of the relative changes of muscle strength and tendon stiffness between sessions over the 1 year observation period in 18 healthy elite jumpers (take-off and swing leg) and 12 healthy control subjects (take-off leg). Statistically signiﬁcant difference to control (P < 0.05). #Tendency for signiﬁcant difference to control (0.05 < p < 0.1). FIGURE 2 \| Box plots of muscle strength (Max. TS MTU Adaptive Changes Over 1 Year: Elite Athletes vs. Controls PF moment), tendon stiffness (AT stiffness) and maximal tendon strain (Max. AT strain) as well as the cosine similarity (Cosine) of the relative changes of muscle strength and tendon stiffness between sessions over the 1 year observation period in 18 healthy elite jumpers (take-off and swing leg) and 12 healthy control subjects (take-off leg). Statistically signiﬁcant difference to control (P < 0.05). #Tendency for signiﬁcant difference to control (0.05 < p < 0.1). FIGURE 3 \| Adaptive changes in TS muscle strength (Max. PF moment) and tendon stiffness (AT stiffness) in the take-off and swing leg of one elite high jumper over a period of 4 years of athletic training separated into preparation (PREP) and competition (COMP) periods. FIGURE 3 \| Adaptive changes in TS muscle strength (Max. PF moment) and tendon stiffness (AT stiffness) in the take-off and swing leg of one elite high jumper over a period of 4 years of athletic training separated into preparation (PREP) and competition (COMP) periods. FIGUR period ontro ower FIGURE 3 \| Adaptive changes in TS muscle strength (Max. PF moment) and tendon stiffness (AT stiffness) in the take-off and swing leg of one elite high jumper over a period of 4 years of athletic training separated into preparation (PREP) and competition (COMP) periods. control group demonstrated statistically signiﬁcant (P = 0.026) lower similarities (lower CS) in their adaptive changes of muscle strength and tendon stiﬀness (Figure 2). in the competition period. Independent of the analyzed leg, there were no signiﬁcant diﬀerences in ﬂuctuations of muscle strength, tendon stiﬀness and maximal tendon strain over time (no diﬀerences in the residuals) between preparation and competition period (Figure 4). Further, there were no signiﬁcant diﬀerences in CS in the adaptive changes of muscle strength and tendon stiﬀness for the preparation in comparison to the competition period (Figure 4). TS MTU Adaptive Changes in Elite Athletes With AT Injury AT strain) as well as the cosine similarity (Cosine) of the relative changes of muscle strength and tendon stiffness between sessions over the 1 year observation period for the affected and healthy leg in two elite jumpers (Ath 1 and Ath 2) with chronic AT tendinopathy and in the control group (CTR; take-off leg; n = 12). FIGURE 5 \| Box plots of residuals in muscle strength (Max. PF moment), tendon stiffness (AT stiffness) and maximal tendon strain (Max. AT strain) as well as the cosine similarity (Cosine) of the relative changes of muscle strength and tendon stiffness between sessions over the 1 year observation period for the affected and healthy leg in two elite jumpers (Ath 1 and Ath 2) with chronic AT tendinopathy and in the control group (CTR; take-off leg; n = 12). tendon stiﬀness in their aﬀected (injured) limb compared to their unaﬀected limb (average values for both athletes: aﬀected limb 564 ± 127 N/mm vs. healthy limb 690 ± 242 N/mm) with similar muscle strength values between limbs (303 ± 61 Nm vs. 312 ± 59 Nm). As a consequence, maximal tendon strain was for both athletes higher in their aﬀected limb (5.3 ± 0.1 vs. 4.6 ± 0.2%). We were able to monitor these two elite athletes over a period of 1 year of athletic training and the two athletes with AT tendinopathy showed in their aﬀected limb a lower uniformity (i.e., lower CS) in the adaptive changes of muscle strength and tendon stiﬀness in comparison to the healthy limb and healthy control group (Figure 5). Furthermore, the aﬀected limb showed greater ﬂuctuations (higher residuals) in muscle strength as well as in tendon stiﬀness and as a consequence in tendon strain over time in comparison to their healthy limb and healthy controls (Figure 5). able to monitor the jumper’s bilateral MTU adaptive changes over a follow up time period of 2 years. The data in Figure 6 indicates that the aﬀected leg shows on average lower muscle strength (∼57%) but about 15% higher Achilles tendon stiﬀness (with more than 60% lower maximal strain) in comparison to the healthy leg despite over 2.5 years of intense rehabilitation and athletics training. TS MTU Adaptive Changes in Elite Athletes With AT Injury In addition to the pool of 18 healthy athletes we were able to monitor two elite athletes with a unilateral AT tendinopathy and one elite athlete following a unilateral AT reconstruction over several years of athletic training. At baseline, the two elite athletes with a unilateral AT tendinopathy showed lower December 2020 \| Volume 11 \| Article 607544 Frontiers in Physiology \| www.frontiersin.org 6 Monitoring MTU Adaptation in Jumpers Karamanidis and Epro FIGURE 4 \| Box plots of residuals in muscle strength (Max. PF moment), tendon stiffness (AT stiffness) and maximal tendon strain (Max. AT strain) as well as in cosine similarity (Cosine) of the relative changes of muscle strength and tendon stiffness between sessions over 4 years during preparation (PREP) and competition (COMP) period in six elite jumpers (take-off and swing leg). FIGURE 5 \| Box plots of residuals in muscle strength (Max. PF moment), tendon stiffness (AT stiffness) and maximal tendon strain (Max. AT strain) as well as the cosine similarity (Cosine) of the relative changes of muscle strength and tendon stiffness between sessions over the 1 year observation period for the affected and healthy leg in two elite jumpers (Ath 1 and Ath 2) with chronic AT tendinopathy and in the control group (CTR; take-off leg; n = 12). tendon stiﬀness in their aﬀected (injured) limb compared to their unaﬀected limb (average values for both athletes: aﬀected able to monitor the jumper’s bilateral MTU adaptive changes over a follow up time period of 2 years The data in Figure 6 FIGURE 4 \| Box plots of residuals in muscle strength (Max. PF moment), tendon stiffness (AT stiffness) and maximal tendon strain (Max. AT strain) as well as in cosine similarity (Cosine) of the relative changes of muscle strength and tendon stiffness between sessions over 4 years during preparation (PREP) and competition (COMP) period in six elite jumpers (take-off and swing leg). FIGURE 4 \| Box plots of residuals in muscle strength (Max. PF moment), tendon stiffness (AT stiffness) and maximal tendon strain (Max. AT strain) as well as in cosine similarity (Cosine) of the relative changes of muscle strength and tendon stiffness between sessions over 4 years during preparation (PREP) and competition (COMP) period in six elite jumpers (take-off and swing leg). FIGURE 5 \| Box plots of residuals in muscle strength (Max. PF moment), tendon stiffness (AT stiffness) and maximal tendon strain (Max. Frontiers in Physiology \| www.frontiersin.org DISCUSSION Muscles and tendons both adapt to mechanical loading. However, diﬀerences in their responsiveness to mechanical stimuli and the rates of adaptive changes may disrupt the ﬁne- tuned interaction of the MTU, increasing the risk for overuse injuries. In the current work we monitored training-induced alterations in muscle strength and tendon stiﬀness in young elite jumpers over 4 years of athletic training in order to detect potential non-uniformities within the TS MTU due to habitual Muscles and tendons both adapt to mechanical loading. However, diﬀerences in their responsiveness to mechanical stimuli and the rates of adaptive changes may disrupt the ﬁne- tuned interaction of the MTU, increasing the risk for overuse injuries. In the current work we monitored training-induced alterations in muscle strength and tendon stiﬀness in young elite jumpers over 4 years of athletic training in order to detect potential non-uniformities within the TS MTU due to habitual The remaining injured elite jumper was measured 10 months after an Achilles tendon rupture and reconstruction and we were December 2020 \| Volume 11 \| Article 607544 Frontiers in Physiology \| www.frontiersin.org 7 Monitoring MTU Adaptation in Jumpers Karamanidis and Epro FIGURE 6 \| Adaptive changes in TS muscle strength (Max. PF moment) and tendon stiffness (AT stiffness) in the affected and healthy leg of an elite athlete with an Achilles tendon rupture and reconstruction during a follow up time period of 2 years. Maximal tendon strain ranged from 1.5 to 2.5% and from 3.9 to 4.9% for the affected and healthy leg over the entire observation window, respectively. mechanical loading. When considering the longitudinal data over several years of athletic training most of the jumpers demonstrated greater ﬂuctuations in TS muscle strength and tendon stiﬀness (and hence tendon strain) compared to controls, irrespective of training period (preparation vs. competition). Moreover, two elite athletes with chronic AT tendinopathy showed lower uniformity (lower CS) in their aﬀected leg in the adaptive changes of muscle strength and tendon stiﬀness and higher ﬂuctuations in tendon strain. These results support our hypothesis that habitual mechanical loading can aﬀect the uniformity within the MTU in elite jumpers, suggesting a temporarily increased mechanical demand on the tendon over a season and potentially higher injury risk. plyometric activities such as jumping or sprint running. Yet, our cross-sectional investigation comprising a total of 91 subjects did not detect clear non-concurrent increases in muscle strength and tendon stiﬀness in elite jumpers. Frontiers in Physiology \| www.frontiersin.org DISCUSSION #Statistically signiﬁcant difference between Low CS and High CS group (P < 0.05). FIGURE 7 \| Individual residuals and group means (dark line) in muscle strength (Max. PF moment), tendon stiffness (AT stiffness) and maximal tendon strain (Max. AT strain) as well as the cosine similarity (Cosine) of the relative changes of muscle strength and tendon stiffness between sessions over the 1 year observation in the take-off leg of the Low CS and High CS group of elite jumpers and in the control group. Statistically signiﬁcant difference between Low CS and control group (P < 0.05). #Statistically signiﬁcant difference between Low CS and High CS group (P < 0.05). training is unknown. As part of this work on muscle and tendon adaptation, we investigated for the ﬁrst time MTU biomechanical properties at about three- to ﬁve-week intervals over several years of athletic training and competition. When all data points analyzed within the 12-month observation period were taken into consideration (for the 18 healthy athletes monitored), we found greater ﬂuctuations of muscle strength over time in elite jumpers in comparison to controls irrespective of the analyzed leg (i.e., greater residuals). Furthermore, the take-oﬀleg showed statistically signiﬁcant lower uniformity (lower CS) in adaptive changes of muscle strength in comparison to the control group although statistical signiﬁcance was not achieved for the inter- group comparison of ﬂuctuations in tendon stiﬀness (P = 0.074) and maximal tendon strain (P = 0.092). There is therefore evidence that the notable ﬂuctuations of muscle strength in the athlete group were only partly matched by an adaptive response in the tendon. the tendon over a season and hence a greater threat to the tissue integrity. Moreover, the ﬁndings from our additional two elite athletes with chronic AT tendinopathy conﬁrmed previous reports of a decreased tendon stiﬀness and increased tendon strain in patients with tendinopathy (Wang et al., 2012; Kulig et al., 2016). The above ﬁndings, when combined with the relatively high prevalence (up to 40%) of AT overuse in elite sprinters and jumpers (Janssen et al., 2018) support the notion that unbalanced adaptation of muscle and tendon in elite jumpers may have a clinical relevance and increase the risk for overuse injuries. Given that ﬂuctuation of muscle strength was more predominant than in tendon stiﬀness, we suggest that the ﬂuctuation of strain was primarily due to the variations in muscle strength rather than tendon stiﬀness. DISCUSSION Irrespective of the analyzed leg, diﬀerences were detected for both TS muscle strength and tendon stiﬀness (0.043 ≤η 2 p ≤0.105), and hence similar levels of strain during maximum contractions between athletes and controls were observed (on average 4.8% in both the athlete and the control group). Given that tendon strain was not signiﬁcantly greater in athletes compared to controls, one might suggest that repetitive plyometric and habitual mechanical loading during training seems not to provoke any measurable non-uniform adaptation within the TS MTU in elite jumpers. Hence, it is reasonable to suggest that the reported alterations in tendon stiﬀness in elite athletes may present a protective mechanism to maintain the tendon integrity in response to the increased demand due to the changed muscular strength. Although most resistance exercise interventions implementing high tendon strain magnitudes are regularly reporting comparable changes in tendon and muscle biomechanical properties (Kubo et al., 2001; Arampatzis et al., 2007a), purely plyometric exercise interventions show tendon adaptations that are not concurrent with those of skeletal muscle (Kubo et al., 2007, 2017; Fouré et al., 2010; Sáez-Sáez de Villarreal et al., 2010; Houghton et al., 2013; Bohm et al., 2014; Hirayama et al., 2017). A feasible explanation for this may be that tendons adapt less eﬀectively to loading proﬁles characterized by short tendon strain durations (Bohm et al., 2014)—as in As for most previous investigations on MTU structure and function in adult athletes (Kongsgaard et al., 2005; Arampatzis et al., 2007b; Staﬁlidis and Arampatzis, 2007; Wang et al., 2012; Couppé et al., 2013; Epro et al., 2019), the above ﬁndings rely on a single time point (data were assessed directly prior to the competition period) and therefore potential ﬂuctuations in muscle strength and tendon stiﬀness over a period of athletic December 2020 \| Volume 11 \| Article 607544 Frontiers in Physiology \| www.frontiersin.org 8 Monitoring MTU Adaptation in Jumpers Karamanidis and Epro FIGURE 7 \| Individual residuals and group means (dark line) in muscle strength (Max. PF moment), tendon stiffness (AT stiffness) and maximal tendon strain (Max. AT strain) as well as the cosine similarity (Cosine) of the relative changes of muscle strength and tendon stiffness between sessions over the 1 year observation in the take-off leg of the Low CS and High CS group of elite jumpers and in the control group. Statistically signiﬁcant difference between Low CS and control group (P < 0.05). Frontiers in Physiology \| www.frontiersin.org DISCUSSION Given that tendon strain is accepted as the main mechanical parameter determining tendon damage accumulation (Wren et al., 2003; Andarawis-Puri et al., 2012), this implies to a temporarily increased mechanical demand on Adaptation of the MTU to various mechanical alternations has been the subject of numerous studies (Bohm et al., 2015a) and has provided evidence that the muscle and tendon homeostatic December 2020 \| Volume 11 \| Article 607544 Frontiers in Physiology \| www.frontiersin.org 9 Monitoring MTU Adaptation in Jumpers Karamanidis and Epro further supports the notion that an unbalanced adaptation within the MTU in athletes could heighten the risk for overuse injuries. In contrast to this, the case report data of one elite jumping athlete 10 months after AT rupture and reconstruction indicated opposing ﬁndings: there was a lower force generation capacity but a higher tendon stiﬀness in the aﬀected limb, leading to a lower level of strain. These bilateral diﬀerences and deﬁcient contractile strength in the aﬀected limb seemed persistent as the athlete did not show any adaptive improvements over a follow up period of 2.5 years post-surgery (Figure 6). Similar ﬁndings have been obtained in several cross-sectional and longitudinal investigations of non-active AT rupture patients even more than 10 years post rupture and reconstruction (Agres et al., 2015; Heikkinen et al., 2017). As the current longitudinal data were generated from an elite athlete encountering athletics training with high training loads and volumes over a period of 2.5 years post-surgery we may argue that AT rupture and reconstruction lead to irreversible MTU dysfunction. In contrast to an injury or pathology, an unbalanced adaptation of muscle and tendon due to athletic training in elite jumpers seems to be a temporary phenomenon. Further long-term investigations are needed to determine whether targeting balanced development of muscle strength and tendon stiﬀness through individualized MTU- focused exercise interventions can enhance the eﬀectiveness of tissue recovery caused by injury or prevent overuse tendon injuries in athletes. response to mechanical loading can vary considerably. For example, a discrepancy between muscle and tendon in the time course of adaptation in response to mechanical loading has been reported previously for both the patellar and the Achilles tendon (Kubo et al., 2012), revealing that muscle strength gains can precede signiﬁcant changes in tendon biomechanical properties. DISCUSSION While an increase in tendon stiﬀness relies on modulation of tissue metabolism and subsequent adaptive changes of tendon structure and morphology (Kjær et al., 2009), neuronal adaptation enables marked increases in muscle strength even before major morphological changes occur within a muscle (Moritani and DeVries, 1979; Narici et al., 1989). Moreover, there is strong evidence that muscles tend to have superior plasticity in response to a wide range of exercise modalities in comparison to tendons, as both high and low loading magnitudes can facilitate increases in muscle strength but only high magnitude loading promotes tendon adaptation (Arampatzis et al., 2007a). Therefore, in line with recent studies conducted on the patellar tendons of adolescent volleyball players (Mersmann et al., 2016, 2017), we propose that injury prevention strategies in elite track and ﬁeld athletes might need to integrate mechanical stimuli which primarily focuses on eﬀectively increasing the tolerance of the tendon to mechanical loading i.e., by incorporating exercise modalities involving optimal strain magnitudes over longer strain durations (Arampatzis et al., 2007a; Bohm et al., 2014). Interestingly, the individual tendon strain in the take-oﬀ leg showed greater ﬂuctuations over 1 year of athletic training and competition for most but not for all elite athletes in comparison to controls (Figure 7). Moreover, the diﬀerences to control subjects were in absolute terms obvious for the swing leg but did not reach statistical signiﬁcance (Figure 2). This substantial variation amongst elite athletes indicates that the relative time courses of adaptation in muscle and tendon might reveal diﬀerences at an individual level and that imbalances between muscle and tendon could often remain undetected when only group mean values are considered. Therefore we argue that individualized approaches (Arampatzis et al., 2020) and rapid subject-speciﬁc tendon strain estimates (Pizzolato et al., 2020) are needed to provide valuable information for coaches and athletes as well as their medical teams about the adaptive processes in muscles and tendons during the various phases of athletic training. This would enable detection of temporal disruption of the ﬁne-tuned interaction within the MTU during the course of training and could allow adjustment of individual athlete training loads through tailored intervention. In relation to our testing protocol one could argue that the used 90◦knee joint conﬁguration could place the m. DISCUSSION Taking into account that tendons have a lower tissue turnover than muscles (Heinemeier et al., 2013) we argue that muscular adaptation takes place at a higher rate in response to modiﬁed mechanical loading induced by variations in training volume and content over a season. Regarding this, it is, however, important to note that the six elite athletes that we monitored over 4 years showed no indications for diﬀerent temporal dynamics in muscle and tendon adaptation between the preparation and competition periods (Figure 4). These athletes demonstrated no signiﬁcant diﬀerences in ﬂuctuations of muscle strength and tendon stiﬀness and hence maximal tendon strain between preparation and competition periods, irrespective of the leg analyzed. Therefore, while we argue that the above observed non-uniform adaptation in elite athletes could be deﬁned by the diﬀerences in the MTU temporal dynamics, potential variations in the training stimulus between preparation and competition were insuﬃcient to cause measurable diﬀerences in the adaptive changes within the MTU in elite jumpers. A more detailed examination of our data, moreover, revealed that some elite athletes clearly showed higher ﬂuctuations over time in maximal tendon strain especially in their take-oﬀleg than other athletes and controls (see interquartile range and maximal value in Figure 2). Accordingly we decided in addition to use an approach based on median split for CS for the take- oﬀleg of the elite athletes and divided them into two groups: Low CS (n = 8) and High CS (n = 9) athletes (see Figure 7). Using this approach we found statistically signiﬁcant greater ﬂuctuations of muscle strength (P = 0.007) and tendon stiﬀness (P = 0.032) for Low CS elite jumpers in comparison to controls (Figure 7). As a consequence, tendon strain level during maximal contractions recorded over time showed in athletes signiﬁcantly greater ﬂuctuations as well (P < 0.001; Figure 7); there were episodes for which levels of strain were up to 1.5 times higher in comparison to controls. Frontiers in Physiology \| www.frontiersin.org December 2020 \| Volume 11 \| Article 607544 DATA AVAILABILITY STATEMENT The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fphys. 2020.607544/full#supplementary-material The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation. ETHICS STATEMENT various contractions in elite athletes. Further, the current work made use of generic AT moment arms from previous literature (Maganaris et al., 1998), which might aﬀect the calculated tendon stiﬀness in absolute terms. However, we believe this could not aﬀect the main outcomes from our longitudinal analyses and the non-uniformity of muscle and tendon adaptations in mature elite athletes. Finally, it is important to note that only six elite athletes were considered in our comparison between preparation and competition periods, irrespective of our monitoring both limbs over a period of 4 years of athletic training. Further studies are needed to investigate whether potential variations of training volume and content between preparation and competition can lead to ampliﬁcation of non-uniformity in adaptation of muscle and tendon. This may be of particular relevance as it has been reported that the frequency of musculoskeletal injuries occurring during preparation is threefold greater in comparison to competition in elite track and ﬁeld athletes (D’Souza, 1994). The studies involving human participants were reviewed and approved by German Sport University Cologne Ethics Committee. The patients/participants provided their written informed consent to participate in this study. DISCUSSION gastrocnemius medialis to an unfavorable position to generate force based on its force-length relationship, which might lead to a diﬀerence in tendon subcomponent deformations and aﬀect the determined tendon stiﬀness. In justiﬁcation, the primary reason for using the ﬂexed knee joint angle conﬁguration was to minimize the inevitable ankle joint rotation that occurs when using an extended knee joint, as this causes a substantial overestimation of the tendon elongation due to the generated force during the plantar ﬂexion contractions. In an earlier study (Ackermans et al., 2016) we investigated the current (90◦) and a more dorsiﬂexed ankle joint conﬁguration (85◦) with the same knee joint setup to enhance the gastrocnemii force potential and their contribution to the resultant joint moment. Even though the more dorsiﬂexed position demonstrated signiﬁcantly higher ankle joint moments and tendon elongation, no signiﬁcant diﬀerences were found for tendon stiﬀness between those two joint conﬁgurations (Ackermans et al., 2016; Supplementary Material of the current study). Hence, although a ﬂexed knee joint could potentially inﬂuence the interactions between the individual AT subtendons, it seems not to cause a clear measurable alteration in the m. gastrocnemius medialis tendon force-elongation relationship during loading, particularly in the “linear” region were the tendon stiﬀness was determined. However, we cannot exclude that the observed higher tendon strain ﬂuctuations in elite athletes may be inﬂuenced also by variations in the regional strain patterns and subtendon sliding (Handsﬁeld et al., 2020; Maas et al., 2020). Incorporating regional and subtendon strain diﬀerentiation could contribute additional understanding into the complex tendon loading pattern during With regard to our case study reporting of two elite athletes with unilateral chronic AT tendinopathy, we saw no evidence for diﬀerences in contractile strength between limbs but there was lower tendon stiﬀness in the aﬀected limb. As a consequence, increased tendon strain during maximal contraction was identiﬁed for the aﬀected limb, in accordance with previous results for both non-active adults and for athletes with tendinopathy (Wang et al., 2012; Kulig et al., 2016). Moreover, we were able to expand those ﬁndings, showing that AT tendinopathy is accompanied by lower uniformity (lower CS) in the adaptation changes of muscle strength and tendon stiﬀness, leading to higher ﬂuctuations in tendon strain over time. This December 2020 \| Volume 11 \| Article 607544 10 Monitoring MTU Adaptation in Jumpers Karamanidis and Epro ACKNOWLEDGMENTS We would like to thank Dr. Falk Schade as well as Hans-Jörg Thomaskamp and their teams at the Olympic Training Centre Rhineland and TSV Bayer 04 Leverkusen for their great support and their practical advice during the entire period of this 4 years nationwide research-service project. AUTHOR CONTRIBUTIONS KK and GE conception and design of research, performed experiments, analyzed data, interpreted results of experiments, prepared ﬁgures, drafted manuscript and approved ﬁnal version of manuscript. CONCLUSIONS In conclusion, the current ﬁndings demonstrate that repetitive plyometric and habitual mechanical loading during training and competition can aﬀect uniformity within the MTU in elite jumpers, leading to a temporarily increased demand on the tendon and potentially higher injury risk. We argue that the mechanism for unbalanced MTU adaptation during athletic training is related to diﬀerences in the time courses of muscle and tendon adaptation, which needs to be considered in exercise and rehabilitation. Further long-term investigations should address the potential of more individualized exercise interventions for prevention of disruptions to the ﬁne-tuned interactions within the MTU in athletes. FUNDING This work was supported by a research grant from the National Institute of Sport Science of Germany (Bundesinstitut für Sportwissenschaft, BISp: ZMVI4-072059/16-17), by the Forschungsservicestelle, German Sport University Cologne (Hochschulinterne Forschungsförderung) and by the Sport and Exercise Science Research Centre at the London South Bank University. All open access publication fees were ﬁnanced by the London South Bank University Open Access funds. REFERENCES doi: 10.1007/s00421-011-2248-x Bojsen-Møller, J., Magnusson, S. P., Rasmussen, L. R., Kjær, M., and Aagaard, P. (2005). Muscle perfomance during maximal isometric and dynamic contractions is inﬂuenced by the stiﬀness of the tendinous structures. J. Appl. Physiol. 99, 986–994. doi: 10.1152/japplphysiol.01305.2004 Kubo, K., Ishigaki, T., and Ikebukuro, T. (2017). Eﬀects of plyometric and isometric training on muscle and tendon stiﬀness in vivo. Physiol. Rep. 5, 1–13. doi: 10.14814/phy2.13374 Cohen, J. (2013). Statistical Power Analysis for the Behavioral Sciences. London: Routledge. Kubo, K., Kanehisa, H., Ito, M., and Fukunaga, T. (2001). Eﬀects of isometric training on the elasticity of human tendon structures in vivo. J. Appl. Physiol. 91, 26–32. doi: 10.1152/jappl.2001.91.1.26 Couppé, C., Kongsgaard, M., Aagaard, P., Vinther, A., Boesen, M., Kjær, M., et al. (2013). Diﬀerences in tendon properties in elite badminton players with or without patellar tendinopathy. Scand. J. Med. Sci. Sport. 23, 89–95. doi: 10.1111/sms.12023 Kubo, K., Morimoto, M., Komuro, T., Yata, H., Tsunoda, N., Kanehisa, H., et al. (2007). Eﬀects of plyometric and weight training on muscle-tendon complex and jump performance. Med. Sci. Sports Exerc. 39, 1801–1810. doi: 10.1249/mss.0b013e31813e630a Dorn, T. W., Schache, A. G., and Pandy, M. G. (2012). Muscular strategy shift in human running: dependence of running speed on hip and ankle muscle performance. J. Exp. Biol. 215, 2347–2347. doi: 10.1242/jeb.075051 Kulig, K., Chang, Y. J., Winiarski, S., and Bashford, G. R. (2016). Ultrasound-based tendon micromorphology predicts mechanical characteristics of degenerated tendons. Ultrasound Med. Biol. 42, 664–673. doi: 10.1016/j.ultrasmedbio.2015.11.013 D’Souza, D. (1994). Track and Field Athletics injuries - a one-year survey. Br. J. Sports Med. 28, 197–202. Epro, G., Hunter, S., König, M., Schade, F., and Karamanidis, K. (2019). Evidence of a uniform muscle-tendon unit adaptation in healthy elite track and ﬁeld jumpers: a cross sectional investigation. Front. Physiol. 10, 574. doi: 10.3389/fphys.2019.00574 Maas, H., Noort, W., Baan, G. C., and Finni, T. (2020). Non-uniformity of displacement and strain within the Achilles tendon is aﬀected by joint angle conﬁguration and diﬀerential muscle loading. J. Biomech. 101:109634. doi: 10.1016/j.jbiomech.2020.109634 Fouré, A., Nordez, A., and Cornu, C. (2010). Plyometric training eﬀects on Achilles tendon stiﬀness and dissipative properties. J. Appl. Physiol. 109, 849–854. doi: 10.1152/japplphysiol.01150.2009 Maganaris, C. N. (2003). Tendon conditioning: artefact or property? Proc. Biol. Sci. 270 Suppl., S39–S42. doi: 10.1098/rsbl.2003.0004 Maganaris, C. N., Baltzopoulos, V., and Sargeant, A. J. (1998). REFERENCES Changes in Achilles tendon moment arm from rest to maximum isometric plantarﬂexion: in vivo observations in man. J. Physiol. 510, 977–985. doi: 10.1111/j.1469-7793.1998.977bj.x Fredberg, U., and Stengaard-Pedersen, K. (2008). Chronic tendinopathy tissue pathology, pain mechanisms, and etiology with a special focus on inﬂammation: review. Scand. J. Med. Sci. Sport. 18, 3–15. doi: 10.1111/j.1600-0838.2007.00746.x Handsﬁeld, G. G., Greiner, J., Madl, J., Rog-Zielinska, E. A., Hollville, E., Vanwanseele, B., et al. (2020). Achilles Subtendon Structure and Behavior as Evidenced From Tendon Imaging and Computational Modeling. Front. Sport. Act. Living 2:70. doi: 10.3389/fspor.2020.00070 Magnusson, S. P., Aagaard, P., Rosager, S., Dyhre-Poulsen, P., and Kjær, M. (2001). Load-displacement properties of the human triceps surae aponeurosis in vivo. J. Physiol. 531, 277–288. doi: 10.1111/j.1469-7793.2001.0277j.x McCrum, C., Oberländer, K. D., Epro, G., Krauss, P., James, D. C., Reeves, N. D., et al. (2018). Loading rate and contraction duration eﬀects on in vivo human Achilles tendon mechanical properties. Clin. Physiol. Funct. Imaging 38, 517–523. doi: 10.1111/cpf.12472 Haugen, T., Seiler, S., Sandbakk, Ø., and Tønnessen, E. (2019). The training and development of elite sprint performance: an integration of scientiﬁc and best practice literature. Sport. Med. Open 5:221. doi: 10.1186/s40798-019- 0221-0 Mersmann, F., Bohm, S., Schroll, A., Marzilger, R., and Arampatzis, A. (2016). Athletic training aﬀects the uniformity of muscle and tendon adaptation during adolescence. J. Appl. Physiol. 121, 893–899. doi: 10.1152/japplphysiol.00493.2016 Heikkinen, J., Lantto, I., Piilonen, J., Flinkkil,ä, T., Ohtonen, P., Siira, P., et al. (2017). Tendon length, calf muscle atrophy, and strength deﬁcit after acute achilles tendon rupture: long-term follow-up of patients in a previous study. J. Bone Jt. Surg. Am. 99, 1509–1515. doi: 10.2106/JBJS.16.01491 Mersmann, F., Charcharis, G., Bohm, S., and Arampatzis, A. (2017). Muscle and tendon adaptation in adolescence: elite volleyball athletes compared to untrained boys and girls. Front. Physiol. 8:417. doi: 10.3389/fphys.2017. 00417 Heinemeier, K. M., Schjerling, P., Heinemeier, J., Magnusson, S. P., and Kjær, M. (2013). Lack of tissue renewal in human adult Achilles tendon is revealed by nuclear bomb14C. FASEB J. 27, 2074–2079. doi: 10.1096/fj.12-225599 Hirayama, K., Iwanuma, S., Ikeda, N., Yoshikawa, A., Ema, R., and Kawakami, Y. (2017). Plyometric training favors optimizing muscle–tendon behavior during depth jumping. Front. Physiol. 8:16. doi: 10.3389/fphys.2017.00016 Moritani, T., and DeVries, H. (1979). Neural factors vs. hypertrophy in the time course of muscle strength gain. Am. J. Phys. Med. 58, 115–130. Moss, B. M., Refsnes, P. E., Abildgaard, A., Nicolaysen, K., and Jensen, J. (1997). REFERENCES aponeurosis in relation to intensity of sport activity. J. Biomech. 40, 1946–1952. doi: 10.1016/j.jbiomech.2006.09.005 aponeurosis in relation to intensity of sport activity. J. Biomech. 40, 1946–1952. doi: 10.1016/j.jbiomech.2006.09.005 Ackermans, T. M. A., Epro, G., McCrum, C., Oberländer, K. D., Suhr, F., Drost, M. R., et al. (2016). Aging and the eﬀects of a half marathon on Achilles tendon force–elongation relationship. Eur. J. Appl. Physiol. 116, 2281–2292. doi: 10.1007/s00421-016-3482-z Arampatzis, A., Mersmann, F., and Bohm, S. (2020). Individualized muscle-tendon assessment and training. Front. Physiol. 11, 723. doi: 10.3389/fphys.2020. 00723 Biewener, A. A., and Roberts, T. J. (2000). Muscle and tendon contributions to force, work, and elastic energy savings: a comparitive perspective. Exerc. Sport Sci. Rev. 28, 99–107. Agres, A. N., Duda, G. N., Gehlen, T. J., Arampatzis, A., Taylor, W. R., and Manegold, S. (2015). Increased unilateral tendon stiﬀness and its eﬀect on gait 2-6 years after Achilles tendon rupture. Scand. J. Med. Sci. Sports 860–867. doi: 10.1111/sms.12456 Bohm, S., Mersmann, F., and Arampatzis, A. (2015a). Human tendon adaptation in response to mechanical loading: a systematic review and meta-analysis of exercise intervention studies on healthy adults. Sport. Med. Open 1:9. doi: 10.1186/s40798-015-0009-9 Andarawis-Puri, N., Sereysky, J. B., Jepsen, K. J., and Flatow, E. L. (2012). The relationships between cyclic fatigue loading, changes in initial mechanical properties, and the in vivo temporal mechanical response of the rat patellar tendon. J. Biomech. 45, 59–65. doi: 10.1016/j.jbiomech.2011.10.008 Bohm, S., Mersmann, F., Marzilger, R., Schroll, A., and Arampatzis, A. (2015b). Asymmetry of Achilles tendon mechanical and morphological properties between both legs. Scand. J. Med. Sci. Sport. 25, 124–132. doi: 10.1111/sms.12242 Arampatzis, A., Karamanidis, K., and Albracht, K. (2007a). Adaptational responses of the human Achilles tendon by modulation of the applied cyclic strain magnitude. J. Exp. Biol. 210, 2743–2753. doi: 10.1242/jeb.003814 Bohm, S., Mersmann, F., Tettke, M., Kraft, M., and Arampatzis, A. (2014). Human Achilles tendon plasticity in response to cyclic strain: eﬀect of rate and duration. J. Exp. Biol. 217, 4010–4017. doi: 10.1242/jeb.112268 Arampatzis, A., Karamanidis, K., Morey-Klapsing, G., De Monte, G., and Staﬁlidis, S. (2007b). Mechanical properties of the triceps surae tendon and December 2020 \| Volume 11 \| Article 607544 Frontiers in Physiology \| www.frontiersin.org 11 Monitoring MTU Adaptation in Jumpers Karamanidis and Epro during training and detraining in vivo. Eur. J. Appl. Physiol. 112, 2679–2691. doi: 10.1007/s00421-011-2248-x during training and detraining in vivo. Eur. J. Appl. Physiol. 112, 2679–2691. REFERENCES Eﬀects of maximal eﬀort strength training with diﬀerent loads on dynamic strength, cross-sectional area, load-power and load-velocity relationships. Eur. J. Appl. Physiol. Occup. Physiol. 75, 193–199. doi: 10.1007/s004210 050147 Hof, A. L., Van Zandwijk, J. P., and Bobbert, M. F. (2002). Mechanics of human triceps surae muscle in walking, running and jumping. Acta Physiol. Scand. 174, 17–30. doi: 10.1046/j.1365-201x.2002.00917.x Houghton, L. A., Dawson, B. T., and Rubenson, J. (2013). Eﬀects of plyometric training on Achilles tendon properties and shuttle running during a simulated cricket batting innings. J. Strength Cond. Res. 27, 1036–1046. doi: 10.1519/JSC.0b013e3182651e7a Muramatsu, T., Muraoka, T., Takeshita, D., Kawakami, Y., Hirano, Y., and Fukunaga, T. (2001). Mechanical properties of tendon and aponeurosis of human gastrocnemius muscle in vivo. J. Appl. Physiol. 90, 1671–1678. doi: 10.1152/jappl.2001.90.5.1671 Janssen, I., van der Worp, H., Hensing, S., and Zwerver, J. (2018). Investigating Achilles and patellar tendinopathy prevalence in elite athletics. Res. Sports Med. 26, 1–12. doi: 10.1080/15438627.2017.1393748 Narici, M. V., Roi, G. S., Landoni, L., Minetti, A. E., and Cerretelli, P. (1989). Changes in force, cross-sectional area and neural activation during strength training and detraining of the human quadriceps. Eur. J. Appl. Physiol. Occup. Physiol. 59, 310–319. doi: 10.1007/BF02388334 Ker, R. F., Alexander, R. M., and Bennett, M. B. (1988). Why are mammalian tendons so thick? J. Zool. 216, 309–324. doi: 10.1111/j.1469-7998.1988.tb02432.x Pizzolato, C., Shim, V. B., Lloyd, D. G., Devaprakash, D., Obst, S. J., Newsham- West, R., et al. (2020). Targeted achilles tendon training and rehabilitation using personalized and real-time multiscale models of the neuromusculoskeletal system. Front. Bioeng. Biotechnol. 8:878. doi: 10.3389/fbioe.2020. 00878 Kjær, M., Langberg, H., Heinemeier, K. M., Bayer, M. L., Hansen, M., Holm, L., et al. (2009). From mechanical loading to collagen synthesis, structural changes and function in human tendon. Scand. J. Med. Sci. Sport. 19, 500–510. doi: 10.1111/j.1600-0838.2009.00986.x Kongsgaard, M., Aagaard, P., Kjær, M., and Magnusson, S. P. (2005). Structural Achilles tendon properties in athletes subjected to diﬀerent exercise modes and in Achilles tendon rupture patients. J. Appl. Physiol. 99, 1965–1971. doi: 10.1152/japplphysiol.00384.2005 Sáez-Sáez de Villarreal, E., Requena, B., and Newton, R. U. (2010). Does plyometric training improve strength performance? A meta-analysis. J. Sci. Med. Sport 13, 513–522. doi: 10.1016/j.jsams.2009.08.005 Staﬁlidis, S., and Arampatzis, A. (2007). Muscle-tendon unit mechanical and morphological properties and sprint performance. J. Sports Sci. 25, 1035–1046. doi: 10.1080/02640410600951589 Kubo, K., Ikebukuro, T., Maki, A., Yata, H., and Tsunoda, N. (2012). Frontiers in Physiology \| www.frontiersin.org December 2020 \| Volume 11 \| Article 607544 REFERENCES Time course of changes in the human Achilles tendon properties and metabolism December 2020 \| Volume 11 \| Article 607544 Frontiers in Physiology \| www.frontiersin.org 12 Monitoring MTU Adaptation in Jumpers Karamanidis and Epro Wren, T. A. L., Lindsey, D. P., Beaupré, G. S., and Carter, D. R. (2003). Eﬀects of creep and cyclic loading on the mechanical properties and failure of human Achilles tendons. Ann. Biomed. Eng. 31, 710–717. doi: 10.1114/1.1569267 Stefanyshyn, D. J., and Nigg, B. M. (1997). Mechanical energy contribution of the metatarsophalangeal joint to running and sprinting. J. Biomech. 30, 1081–1085. doi: 10.1016/S0021-9290(97)00081-X Stefanyshyn, D. J., and Nigg, B. M. (1998). Contribution of the lower extremity joints to mechanical energy in running vertical jumps and running long jumps. J. Sports Sci. 16, 177–186. doi: 10.1080/026404198366885 Conﬂict of Interest: KK has equity in Protendon GmbH & Co. KG, whose measurement device and software was used for the data processing and analysis in this study. Wang, H. K., Lin, K. H., Su, S. C., Shih, T. T. F., and Huang, Y. C. (2012). Eﬀects of tendon viscoelasticity in Achilles tendinosis on explosive performance and clinical severity in athletes. Scand. J. Med. Sci. Sport. 22, 1–9. doi: 10.1111/j.1600-0838.2012.01511.x The remaining author declares that the research was conducted in the absence of any commercial or ﬁnancial relationships that could be construed as a potential conﬂict of interest. Wiesinger, H.-P., Kösters, A., Müller, E., and Seynnes, O. R. (2015). Eﬀects of increased loading on in vivo tendon properties: a systematic review. Med. Sci. Sports Exerc. 47, 1885–1895. doi: 10.1249/MSS.00000000000 00603 Copyright © 2020 Karamanidis and Epro. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Willwacher, S., Funken, J., Heinrich, K., Müller, R., Hobara, H., Grabowski, A. M., et al. (2017). Elite long jumpers with below the knee prostheses approach the board slower, but take-oﬀmore eﬀectively than non-amputee athletes. Sci. Rep. 7:16058. doi: 10.1038/s41598-017-16383-5 December 2020 \| Volume 11 \| Article 607544 Frontiers in Physiology \| www.frontiersin.org 13
https://openalex.org/W3089815360	http://real.mtak.hu/122335/1/Pasztor_Szep_Marko__PhysRevD.103.034511.pdf	English	null	Apparent convergence of Padé approximants for the crossover line in finite density QCD	Physical review. D/Physical review. D.	2,021	cc-by	11,392	PHYSICAL REVIEW D 103, 034511 (2021) PHYSICAL REVIEW D 103, 034511 (2021) Attila Pásztor ,1,* Zsolt Sz´ep ,2,† and Gergely Markó 3,‡ 1ELTE Eötvös Loránd University, Institute for Theoretical Physics, Pázmány P. s. 1/A, H-1117 Budapest, Hungary 2MTA-ELTE Theoretical Physics Research Group, Pázmány P. s. 1/A, H-1117 Budapest, Hungary 3Fakultät für Physik, Universität Bielefeld, D-33615 Bielefeld, Germany (Received 1 October 2020; accepted 19 January 2021; published 25 February 2021) Attila Pásztor ,1,* Zsolt Sz´ep ,2,† and Gergely Markó 3,‡ 1ELTE Eötvös Loránd University, Institute for Theoretical Physics, Pázmány P. s. 1/A, H-1117 Budapest, Hungary 2MTA-ELTE Theoretical Physics Research Group, Pázmány P. s. 1/A, H-1117 Budapest, Hungary 3Fakultät für Physik, Universität Bielefeld, D-33615 Bielefeld, Germany (Received 1 October 2020; accepted 19 January 2021; published 25 February 2021) We propose a novel Bayesian method to analytically continue observables to real baryochemical potential μB in finite density QCD. Taylor coefficients at μB ¼ 0 and data at imaginary chemical potential μI B are treated on equal footing. We consider two different constructions for the Pad´e approximants, the classical multipoint Pad´e approximation and a mixed approximation that is a slight generalization of a recent idea in Pad´e approximation theory. Approximants with spurious poles are excluded from the analysis. As an application, we perform a joint analysis of the available continuum extrapolated lattice data for both pseudocritical temperature Tc at μI B from the Wuppertal-Budapest Collaboration and Taylor coefficients κ2 and κ4 from the HotQCD Collaboration. An apparent convergence of ½p=p and ½p=p þ 1 sequences of rational functions is observed with increasing p. We present our extrapolation up to μB ≈600 MeV. DOI: 10.1103/PhysRevD.103.034511 Published by the American Physical Society Published by the American Physical Society under the terms of the Creative Commons Attribution 4.0 International license. Further distribution of this work must maintain attribution to the author(s) and the published article’s title, journal citation, and DOI. Funded by SCOAP3. Apparent convergence of Pad´e approximants for the crossover line in finite density QCD Attila Pásztor ,1,* Zsolt Sz´ep ,2,† and Gergely Markó 3,‡ 1ELTE Eötvös Loránd University, Institute for Theoretical Physics, Pázmány P. s. 1/A, H-1117 Budapest, Hungary 2MTA-ELTE Theoretical Physics Research Group, Pázmány P. s. 1/A, H-1117 Budapest, Hungary 3Fakultät für Physik, Universität Bielefeld, D-33615 Bielefeld, Germany (Received 1 October 2020; accepted 19 January 2021; published 25 February 2021) I. INTRODUCTION sampled inside some domain D is uniquely determined by the approximant used, the extension of a function differing on D by no matter how small an amount can lead to arbitrarily different values at points outside D. That is to say: analytic continuation is unique, but is not a continuous function of the data. For such ill-posed prob- lems, the only way to achieve convergence in the results is to use some kind of regularization. This makes sure that the noise in the data is not overemphasized by the analytic continuation. As the noise is reduced, the regularizing term is made weaker. This leads to a kind of double limit when the regularization and the noise are taken to zero together. The simplest kind of regularization for analytic continu- ation is the use of some ansatz, which is assumed to describe the physics both in the range where data is available, and in the range where one tries to extrapolate. The conservative view is to use for analytic continuation few-parameter approximants, which all fit the data well, and perform the continuation only in a range where they do not deviate much from each other, assessing the systematic error of the continuation from this deviation. Here we pursue a more adventurous approach, by considering a sequence of approximants of increasing functional com- plexity, and trying to observe whether they converge or not. Despite considerable effort invested so far, the phase diagram of QCD in the temperature(T)-baryon chemical potential(μB) plane still awaits determination from first principles. At the moment, the only solid information available is the curvature of the crossover temperature [1–3], together with some upper bound on the absolute value of the next Taylor coefficient of order Oðμ4 BÞ [2,3]. These results come either from the evaluation of Taylor coefficients with lattice simulations performed at μB ¼ 0 or via simulations performed at imaginary μB, where the sign problem is absent, with the Taylor coefficients obtained from a subsequent fit. Whether the input data are the Taylor coefficients or the values of a function at several values of the imaginary chemical potential, fact is that the numerical analytic continuation needed to extrapolate the crossover to real μB is a mathematically ill-posed problem [4,5]. I. INTRODUCTION This means that although the analytic continuation of a function The conservative view is to use for analytic continuation few-parameter approximants, which all fit the data well, and perform the continuation only in a range where they do not deviate much from each other, assessing the systematic error of the continuation from this deviation. Here we pursue a more adventurous approach, by considering a sequence of approximants of increasing functional com- plexity, and trying to observe whether they converge or not. *apasztor@bodri.elte.hu †szepzs@achilles.elte.hu ‡gmarko@physik.uni-bielefeld.de Inthe absence ofphysicallymotivatedansatz,a goodguess is to study the ½p=p (diagonal) and ½p=p þ 1 (subdiagonal) Pad´e sequences, as these are only slightly more complicated to work with than polynomials, but have far superior convergence properties. Ordinary Pad´e approximants 034511-1 2470-0010=2021=103(3)=034511(11) Published by the American Physical Society PHYS. REV. D 103, 034511 (2021) PÁSZTOR, SZ´EP, and MARKÓ (i.e., rational functions constructed using approximation- through-order conditions to match the Taylor expansion of a functionat agiven point)areknownto convergeuniformlyon the entire cut plane for functions of Stieltjes type [6] (which have a cutonthe negative real axis). Forthisclass offunctions the subdiagonal sequence of multipoint (or N-point) Pad´e approximants [7], also know as the Schlessinger point method in the context of scattering theory [8], is also convergent (see [9] and references therein). For a meromor- phic function, on the other hand, Pad´e approximants are known to converge in measure [10,11], i.e., almost every- where on the complex plane, in stark contrast to polynomial approximations, which stop converging at the first pole of such a function. construct the subdiagonal sequence as well. For the data fitting step we use a Bayesian analysis. The likelihood function ensures that approximants are close to both the data on the Taylor coefficients at μB ¼ 0 and the data at purely imaginary μB, while a Bayesian prior makes sure that spurious poles are excluded from the extrapolation. Considering two different types of Pad´e approximants is a nontrivial consistency check, mainly because the exact form of the prior distribution will be different for the two cases, as the number of interpolation point where the function values will be restricted is different. I. INTRODUCTION We note that while Bayesian methods—especially var- iations of the maximum entropy method with different entropy functionals—for the analytical continuation to real time are quite commonly used in lattice QCD [19–26], as far as we are aware, such methods have not been applied to the analytic continuation problem in μB so far. The only related example we are aware of is Ref. [27], where a Bayesian method is used to extract high order derivatives of the pressure around μB ¼ 0 from data at imaginary μB. One must note however, that the mathematical problem in that paper is that of numerical differentiation, which is distinct from the analytic continuation problem discussed here. This paper, therefore, is the first attempt of using this class of mathematical techniques to the analytic continuation problem in finite density QCD. While the convergence properties of Pad´e approximants in exact arithmetic are often very good, even in cases where the mathematical reason for the convergence is not fully understood yet, these approximations tend to be very fragile in the presence of noise. This often manifests itself in spurious poles, whose residue goes to zero as the noise level is decreased, as well as spurious zero-pole pairs (called Froissart doublets [12]). The distance between the zero and the pole goes to zero as the noise decreases, eventually leading to the annihilation disappearance of the doublet. There is a large body of mathematical literature devoted to the removal of these spurious poles. Procedures which do so typically involve some further regularization, like in Ref. [13], where this is based on singular value decomposition, or monitoring the existence of Froissart doublets for later removal, like in Ref. [14]. In cases where the noise level on the data cannot be arbitrarily decreased,1 the exclusion of spurious poles is mandatory if one wants to go to higher order approximants in the analysis. The above mentioned fragility of the Pad´e approxima- tion method when applied to noisy data is the main reason that most of the previous applications to finite density QCD employ low order approximants. Pad´e approximants were used in this context to analytically continue to real values of μB the pseudocritical temperature values obtained at imagi- nary chemical potential for various number of flavors and colors [28–31]. The convergence of a Pad´e sequence was seemingly not in the focus of these investigations, with the exception of [28]. 1E.g., if the noise is only coming from machine precision in floating point arithmetic, the Froissart doublets can often be removed by simply using multiple precision arithmetic for the “naive” algorithm. 2One can choose b0 ¼ 1 without loss of generality, having n þ m þ 1 independent coefficients. 3One equates the expansion of fðtÞ with RnmðtÞ, cross multiply and then equates the coefficients of t on both sides of the equation. I. INTRODUCTION For odd number of points N ¼ 2k þ 1, k ≥0, one obtains the approximant ½k=k, while for even number of points N ¼ 2k, k ≥1, one obtains the approximant ½k þ 1=k. More details can be found in Appendix A, where the construction of the multipoint approximant CN is summarized. novel Pad´e-type approximants in the absence of noise. Since the traditional multipoint Pad´e approximants are quite well known, they are relegated to Appendix A. Next, we discuss the Bayesian analysis in the presence of noise in a general manner that includes both the multipoint Pad´e and the mixed Pad´e approximant case. In Sec. III we turn to physical applications. We first demonstrate the effective- ness of Pad´e approximants in a chiral effective model. Finally, we perform a joint analysis of the continuum extrapolated lattice data on the Taylor coefficients at μB¼0 and the crossover line at imaginary μB. Appendix B summarizes the formulas relating our notational conven- tions on the Taylor coefficients to those found elsewhere in the literature. As mentioned in Ref. [38] (see p. 16), an obvious modification of the multipoint Pad´e approximation can be given if any number of successive derivatives exists at the points where the value of the function is known. Recently such a modification, called Pad´e-type rational approximant with n ¼ m ¼ k was constructed in [18], with k being the degree up to (and including) which the expansion of the approximant matches the Taylor expan- sion of the function. The denominator of this ½k=k approximant is fixed by function values at arbitrarily chosen interpolating points and the coefficients of the numerator are obtained by imposing the approximation- through-order conditions. around t ¼ 0 are known. A widely used method to tackle the problem is to fit a rational fraction2 A widely used method to tackle the problem is to fit a rational fraction2 In what follows we shall use Pad´e-type approximants of the form Rp p and Rp pþ1, with p ≥1, satisfying 2p ¼ k þ l and 2p þ 1 ¼ k þ l, respectively. To construct for example R3 4ðtÞ using c0, c1, and c2, one equates (1) with (2) and after cross multiplication one matches the coefficients of t0, t1, and t2. This gives a0 ¼ c0, a1 ¼ c1 þ b1c0, and a2 ¼ c2 þ b1c1 þ b2c0, which are common for all approximants with n ≥2. Using these expressions for a0, a1, and a2 in N3ðtÞ, one sees that the five conditions fiD4ðτiÞ ¼ N3ðτiÞ; i ¼ 1; …; 5 represents a system of linear equations for the five unknown a3, b1, b2, b3, b4, which can be easily solved numerically with some standard linear algebra algorithm. As for the R1 2ðtÞ approximant, this is constructed very similarly to the original Pad´e approximant, only the condition on the third derivative (unknown in our case) is replaced by R1 2ðτ1Þ ¼ f1, where τ1 is an interpolating point. ½n=m ≡RnmðtÞ ≡NnðtÞ DmðtÞ ≔ Pn i¼0 aiti Pm i¼0 biti; ð2Þ ð2Þ to the set of function values and/or available derivatives. When only derivatives at t ¼ 0 are used, one obtains the ordinary Pad´e approximant in which the coefficients of both the denominator and the numerator are fully determined by the Taylor coefficients by imposing the approximation-through-order conditions3 RnmðtÞ ¼ fðtÞþ Oðtmþnþ1Þ. This condition implies that the Taylor expan- sion of the Pad´e approximant around t ¼ 0 agrees with the Taylor expansion of the function up to and including the order of the highest Taylor coefficient known. to the set of function values and/or available derivatives. When only the set of function values at the interpolating points are used, one obtains the so-called multipoint Pad´e I. INTRODUCTION A related problem in finite density QCD, where Pad´e approximants have also been considered, is the calculation of the equation of state at finite chemical potential. An early work that uses a high order Taylor expansion in an effective model is Ref. [32]. Two recent examples in lattice QCD are Refs. [33,34]. The low order Pad´e approximants used in the above studies are not yet expected to take advantage of the superior convergence properties of the Pad´e series. This is in sharp contrast to the case in statistical physics, where in Ising-like models the Taylor coefficients are known exactly to high orders [35– 37]. However, even a low order Pad´e approximant repre- sents a resummation of the Taylor series, which is exploited when applied outside the radius of convergence of the Taylor series. The main advantage of the Bayesian approach presented here is the ability to go to considerably higher orders, at the cost of what we believe are physically reasonable extra assumptions. When dealing with numerical analytic continuation, we need to select the approximant from a class of possible functions (i.e., a model) and a method to take into account the data (i.e., a fitting method). For the former we use two types of rational approximants, the classical multipoint Pad´e approximants recently used for analytic continuation in Refs. [15–17] and a slight generalization of the Pad´e- type approximant introduced and studied recently in [18]. The parameters of the multipoint Pad´e approximant are determined solely in terms of the interpolating points and information on Taylor coefficients, if it exists, can be taken into account in the second, data fitting step. In contrast, the Pad´e-type approximant allows for a joint use of interpolat- ing points and Taylor coefficients in determining the parameters of the approximant. Although the focus in [18] was on the diagonal sequence ½p=p of Pad´e-type approximants, the method can be easily generalized to The paper is organized as follows. In Sec. II we introduce the mathematic tools used for our analysis. First we treat the 034511-2 PHYS. REV. D 103, 034511 (2021) APPARENT CONVERGENCE OF PADÉ APPROXIMANTS FOR … approximant, which is particularly useful in numerics in its continuous fraction formulation because the coefficients can be determined easily from recursion relations. A. Pad´e-type rational approximants in the absence of noise Using the notation of [18] the mathematical formulation of analytic continuation is as follows. Assuming the existence of a continuous real function f∶R →R, we would like to know its value for t > 0 given that: It is easy to generalize the construction used in [18] to obtain Pad´e-type approximants for which l ≠k. With k þ 1 coefficients of the Taylor expansion (including the value of the function at zero) one can construct many Pad´e-type approximants of this type, one just has to satisfy the relation n þ m ¼ k þ l. In this case k þ 1 coefficients of the numerator NnðtÞ are determined from the approximation- through-order conditions, meaning that strictly speaking Rnm satisfies by construction n ≥k, and the remaining n þ m −k ¼ l coefficients are fixed by function values at l number of interpolating points via fiDmðτiÞ ¼ NnðτiÞ; i ¼ 1; …; l. (i) at a number of interpolating points τi < 0, i ¼ 1; …; l, the values fi ≔fðτiÞ are known, fi fð iÞ (ii) a number of coefficients ci, i ¼ 0; …; k in the Taylor expansion ð Þ (ii) a number of coefficients ci, i ¼ 0; …; k in the Taylor expansion fðtÞ ¼ c0 þ c1t þ þ cktk ð1Þ ð1Þ around t ¼ 0 are known. B. Bayesian approach The index “i” goes over the interpolating (node) points, which are different from the data points j ¼ 1; …; L used in (4). In Method 1 the temperature values RnmðiμI B;i; ⃗a; ⃗bÞ ≡Tic at the node points are generated with a normal distribution whose standard deviation wi is chosen to be substantially larger than the error σfj of the lattice data, in which case the result of the analytic continuation is not sensitive to the actual value of wi. In Method 2 the temperature values at the node points are generated using importance sampling and the θ-function is only needed for a technical reason, as explained at the end of Sec. III B 1. The node points, as well as wi ≡wTic and ¯Tcðˆμ2 B;iÞ, are given in Fig. 3. ¯Tcðˆμ2 B;iÞ is obtained by interpolating the mean value of the lattice data points available at imaginary μB. ð4Þ with T being the number of derivatives known at μB ¼ 0 and L being the number of function values known for μ2 B < 0. Z is a normalization constant. The Taylor coef- ficients at μB ¼ 0 are clearly correlated, but their correla- tion matrix was not given in Ref. [2] so we ignore the correlations. If the correlations between the Taylor coef- ficients are known, including them in our method is completely straightforward. The data at different values of imaginary μB come from different Monte Carlo runs, and are thus uncorrelated. with T being the number of derivatives known at μB ¼ 0 and L being the number of function values known for μ2 B < 0. Z is a normalization constant. The Taylor coef- ficients at μB ¼ 0 are clearly correlated, but their correla- tion matrix was not given in Ref. [2] so we ignore the correlations. If the correlations between the Taylor coef- ficients are known, including them in our method is completely straightforward. The data at different values of imaginary μB come from different Monte Carlo runs, and are thus uncorrelated. The variables that the Bayesian analysis code uses for the construction of the Pad´e approximants are not the coef- ficients of the polynomial themselves. For the multipoint Pad´e approximants we use a number of interpolated values at fixed node points in ˆμ2 B ≔μ2 B=T2. B. Bayesian approach The Bayesian approach [39] that considers the data sample fixed and the model parameters as random variables 034511-3 PÁSZTOR, SZ´EP, and MARKÓ PHYS. REV. D 103, 034511 (2021) gives a perspective on the curve fitting problem which is particularly suited for a meta-analysis of data with noise. spurious poles both in the interpolated and the extrapolated range. Due to this factor of the prior, the method is only applicable when no physical poles are expected in the aforementioned ranges. We do not include Pad´e approximants of different order in one large meta-analysis, rather we perform a separate Bayesian analysis of the different order approximants, in order to study their convergence properties as the order of the approximation is increased. For an ½n=m Pad´e approx- imant, the model parameters are the coefficients ⃗a ¼ ða0; a1; …; anÞ and ⃗b ¼ ðb1; b2; …; bmÞ, with a total of n þ m þ 1 coefficients to be determined. The posterior prob- ability can be written as: The prior also contains a further factor—the exact form of which for the two different Pad´e approximants will be discussed in Sec. III B—which prevents extra oscillations of the interpolants in the μ2 B < 0 range, which are not warranted by the data. This is enforced by using a prior distribution of the interpolated values at the node points at fixed μ2 B=T2 range. We have checked that our results are not sensitive to the choice of the node points. This is also expected on mathematical grounds, since unlike polyno- mial interpolants, rational interpolants are not extremely sensitive to the choice of the node points used for the interpolation [40]. B. Bayesian approach Pð⃗a; ⃗bjdataÞ ¼ 1 Z Pðdataj⃗a; ⃗bÞPpriorð⃗a; ⃗bÞ; ð3Þ ð3Þ where assuming Gaussian errors around the correct model parameters, the likelihood is given by: Putting all the above information together, the prior can be given as an implicit condition on the model parameters ⃗a and ⃗b in the following form Pðdataj⃗a; ⃗bÞ ¼ exp −1 2 χ2 ; Ppriorð⃗a; ⃗bÞ ∝ Q i FðjRnmðiμI B;i; ⃗a; ⃗bÞ −¯Tcðˆμ2 B;iÞj;wiÞ; ∄pole ∈I; 0; ∃pole ∈I; ð5Þ χ2 ¼ χ2 Taylor þ χ2 ImμB; y μB χ2 Taylor ¼ X T i¼1 ci −∂iRnmðμB;⃗a;⃗bÞ ∂ðμ2 BÞi jμB¼0 2 σ2ci ; χ2 ImμB ¼ X L j¼1 fj −RnmðiμI B;j; ⃗a; ⃗bÞ 2 σ2 fj ; ð4Þ ∃pole ∈I; ð5Þ where Fðx; wÞ is either exp ð−x2=ð2w2ÞÞ or θðw −xÞ (Heaviside step function), corresponding, respectively, to Method 1 and Method 2 used in Sec. III B 1 and I represents a range of ˆμ2 B for which the absence of poles of the Pad´e approximants is required (we use I ¼ ½−π2; 60π2). The index “i” goes over the interpolating (node) points, which are different from the data points j ¼ 1; …; L used in (4). In Method 1 the temperature values RnmðiμI B;i; ⃗a; ⃗bÞ ≡Tic at the node points are generated with a normal distribution whose standard deviation wi is chosen to be substantially larger than the error σfj of the lattice data, in which case the result of the analytic continuation is not sensitive to the actual value of wi. In Method 2 the temperature values at the node points are generated using importance sampling and the θ-function is only needed for a technical reason, as explained at the end of Sec. III B 1. The node points, as well as wi ≡wTic and ¯Tcðˆμ2 B;iÞ, are given in Fig. 3. ¯Tcðˆμ2 B;iÞ is obtained by interpolating the mean value of the lattice data points available at imaginary μB. where Fðx; wÞ is either exp ð−x2=ð2w2ÞÞ or θðw −xÞ (Heaviside step function), corresponding, respectively, to Method 1 and Method 2 used in Sec. III B 1 and I represents a range of ˆμ2 B for which the absence of poles of the Pad´e approximants is required (we use I ¼ ½−π2; 60π2). 1. Convergence in the absence of noise 0 20 40 60 80 100 120 140 0 200 400 600 800 1000 Tc [MeV] B [MeV] N=3, [1/1] 4, [1/2] 5, [2/2] 6, [2/3] 7, [3/3] 8, [3/4] exact TCP -10 -5 0 5 10 0 50 100 150 200 250 300 350 400 (Tc approx/Tc exact-1)100 B 2/T2 m2 eff ¼ m2 þ λ 72 þ g2 12 Nc T2 þ g2Nc 36π2 μ2 B; ð6aÞ λeff ¼ λ 6 −g4Nc 8π2 Ψ 1 2 þ i ˆμB 6π þ Ψ 1 2 −i ˆμB 6π þ2 þ 2 ln 4πT M0 : ð6bÞ ð6aÞ ð6bÞ In the expressions above ΨðxÞ is the digamma function, ˆμB ¼ μB=T, Nc is the number of colors, g ¼ mq=Φ0 (with mq ¼ mN=3 and Φ0 ¼ fπ=2) is the Yukawa coupling between the pion and sigma mesons and the constituent quarks, and m2 and λ are the renormalized mass and the self- coupling in the OðNÞ symmetric mesonic sector of the CQM model, which at the value M0 ¼ 886 MeV of the re- normalization scale take the values m2 ¼ −326054 MeV2 and λ ¼ 400. FIG. 1. Apparent convergence of the multipoint Pad´e approx- imants CN (solid lines) determined from Tc values at μI B in the CQM model in comparison to the Taylor expansion of order N−1 around μB ¼ 0 (dashed lines). In the main plot the parametric curves for the Pad´e approximants are obtained as ð ﬃﬃﬃﬃﬃ ˆμ2 B p CNðˆμ2 BÞ; CNðˆμ2 BÞÞ. The bivaluedness of the subdiagonal approximants (and h f h dd d T l i ) fl h FIG. 1. Apparent convergence of the multipoint Pad´e approx- imants CN (solid lines) determined from Tc values at μI B in the CQM model in comparison to the Taylor expansion of order N−1 around μB ¼ 0 (dashed lines). In the main plot the parametric curves for the Pad´e approximants are obtained as ð ﬃﬃﬃﬃﬃ ˆμ2 B p CNðˆμ2 BÞ; CNðˆμ2 BÞÞ. The bivaluedness of the subdiagonal approximants (and the curves of the odd-order Taylor expansion) reflects that μB ≔ ﬃﬃﬃﬃﬃ ˆμ2 B p CNðˆμ2 BÞ has a maximum. The inset shows the percentage difference between the approximated and exact values of Tc, with the vertical dotted line indicating the radius of converges of the Taylor expansion. 4These expressions corresponds to Eqs. B. Bayesian approach For the case of Pad´e- type approximants we use a smaller number of interpolated values at node points and a number of derivatives at μB ¼ 0. These are of course in a one-to-one correspon- dence with the polynomial coefficients, once the restriction b0 ¼ 1 has been made in Eq. (2). Details of the imple- mentation will be discussed in Sec. III B. Our numerical results will be based on the posterior distribution. For a fixed value of μB=T, we study the posterior distribution of the crossover temperature Tc ¼ Rnmðˆμ2 BÞ and chemical potential μB ¼ ﬃﬃﬃﬃﬃ ˆμ2 B p Tc. The center point will in both cases be the median, while the asym- metric error bars represent the central 68% of the posterior distribution of both quantities. We will call these percentile based errors. We shall see that the asymmetry of the An important part of our procedure is that we do not work with the space of all Pad´e approximants of order ½n=m, rather, the allowed approximants are restricted by the prior, which always contains a factor that excludes 034511-4 APPARENT CONVERGENCE OF PADÉ APPROXIMANTS FOR … PHYS. REV. D 103, 034511 (2021) For μB ≥0 the model exhibits a second order chiral phase transition line in the μB −T plane, which is obtained from the condition m2 eff ¼ 0. This line of second order points ends in a tricritical point with coordinates deter- mined by m2 eff ¼ λeff ¼ 0. For μB > μTCP B the chiral phase transition is of first order and m2 eff ¼ 0 gives the location of the first spinodal down to T ¼ 0. The merit of the expressions in (6a) and (6b) is that the line of second order phase transitions, which is actually an ellipse in the μB −T plane, can be determined analytically together with the location of the TCP. This makes the analytic continu- ation very simple, as we just have to change ˆμ2 B →−ˆμ2 B in the expression posterior distribution increases as ˆμ2 B increases in the extrapolation range (real values of μB) and is the largest for the [2/2] Pad´e approximant. In practice, the integration over the prior distribution is carried out with simple Monte Carlo algorithms. 1. Convergence in the absence of noise In Ref. [41], using leading order large-N techniques resulting in an ideal gas approximation for the constituent quarks, the coefficients of the Landau-Ginzburg type effec- tive potential Veff ¼ m2 eff 2 Φ2 þ λeff 4 Φ4 þ for the chiral order parameter Φ were determined in the chiral limit to be4: 0 20 40 60 80 100 120 140 0 200 400 600 800 1000 Tc [MeV] B [MeV] N=3, [1/1] 4, [1/2] 5, [2/2] 6, [2/3] 7, [3/3] 8, [3/4] exact TCP -10 -5 0 5 10 0 50 100 150 200 250 300 350 400 (Tc approx/Tc exact-1)100 B 2/T2 FIG. 1. Apparent convergence of the multipoint Pad´e approx- imants CN (solid lines) determined from Tc values at μI B in the CQM model in comparison to the Taylor expansion of order N−1 around μB ¼ 0 (dashed lines). In the main plot the parametric curves for the Pad´e approximants are obtained as ð ﬃﬃﬃﬃﬃ ˆμ2 B p CNðˆμ2 BÞ; CNðˆμ2 BÞÞ. The bivaluedness of the subdiagonal approximants (and the curves of the odd-order Taylor expansion) reflects that μB ≔ ﬃﬃﬃﬃﬃ ˆμ2 B p CNðˆμ2 BÞ has a maximum. The inset shows the percentage difference between the approximated and exact values of Tc, with the vertical dotted line indicating the radius of converges of the Taylor expansion. A. Convergence of Pad´e approximants in a chiral effective model BeforeapplyingthemethoddescribedinSec.IItotheactual QCD data, we study the analytic continuation within the chiral limit of the two flavor (Nf ¼ 2) constituent quark-meson (CQM) model. We show that in this model both the diagonal and the subdiagonal sequences constructed from TcðμI BÞ exhibit apparent convergence to the exact TcðμBÞ curve. We also show that the Pad´e approximant knows nothing about the location of the tricritical point (TCP), as this information is not encoded in TcðμBÞ. Finally, we investigate the effect of the error on the analytical continuation. Tcðˆμ2 BÞ ¼ ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ −72m2 λ þ 6g2Ncð1 þ ˆμ2 B=ð3π2ÞÞ s ; ð7Þ ð7Þ ð7Þ obtained from (6), to go from real to imaginary chemical potentials. We can sample Tc at imaginary values of ˆμB and fit multipoint Pad´e approximants to the sampled data (see Appendix A). Then, we can evaluate the Pad´e approximant at real values of ˆμB and compare the value of analytic continued Tc with the exact values obtained from (7). This comparison is presented in Fig. 1, where the inset shows the percentage difference between CNðˆμ2 BÞ and Tcðˆμ2 BÞ using the same interpolating points as those shown in Fig. 3 in the case of the QCD data. The main figure shows that the diagonal sequence converges from above, while the B. Bayesian approach The statistics needed is such that the posterior distribution of the studied observables does not change anymore, which we explicitly checked to be the case in our analysis. 2. Effect of the error Next, we investigate what happens when analytic con- tinuation is performed in the presence of noise. As a reference point we start by generating Tic configurations with a normal distribution characterized by mean calculated from (7) and standard deviation corresponding to the relative error wTic=Tic ¼ 1% and investigate to what extent should we decrease the relative error in order to get close to the curves obtained in Fig. 1 in the absence of noise. Note that wTic is by a factor of two larger than the average error of the QCD data at imaginary chemical potential. the curves of Fig. 1 by more than the estimated statistical error. This reflects the ill-posedness of the analytic con- tinuation problem. However, even with the largest error used, the Pad´e sequence converges up to μB ≈600 MeV. For the mathematically curious, we also show the effect of increasing the range of the interpolation points. As expected, convergence is accelerated by the increase of the sampling range. This is of course not directly relevant for QCD, as the Roberge-Weiss transition puts a limit on the available range for the interpolation points. We determine the coefficients of the multipoint Pad´e approximant for each generated configuration, evaluate the approximant for positive ˆμ2 B and, using the Bayesian method presented in Sec. II B, calculate χ2 including or omitting information on the Taylor coefficients, and then study the posterior distribution of these values. The method is applied to the QCD data in the next subsection, where it is referred to as Method 1. The control points used to calculate χ2 have ˆμ2 B corresponding to the QCD data at imaginary chemical potential and Tc obtained from (7). We use a unique relative error of Tc at all interpolating and control points, whose value is indicated in the key of Fig. 2 (wTic used to generate Tic instances is twice the indicated value). When Taylor coefficients c1 and c2 are also used in the calculation of χ2, their values c1 ¼ −1.575 and c2 ¼ 0.0267 is determined from the Taylor expansion of (7), as for the reference value of their error, indicated in the key of Fig. 2, we use the error of the QCD data obtained from Ref. [2], namely σ0c1 ¼ 0.626 and σ0c2 ¼ 0.627. The sam- pling points in the range ˆμ2 B ∈ð−7.35; 0 are those used previously to obtain Fig. 1. Convergence in the absence of noise (13) and (14) of [41], just that we used the relation ∂ ∂n ðLinð−ezÞ þ Linð−e−zÞjn¼0 ¼ −γ −lnð2πÞ −½Ψðð1 þ iz=πÞ=2Þ þ Ψðð1 −iz=πÞ=2Þ=2, which can be proven by comparing the high temperature expansion used there with the one given in [42]. 034511-5 PÁSZTOR, SZ´EP, and MARKÓ PHYS. REV. D 103, 034511 (2021) 40 60 80 100 120 140 160 N=3 B [MeV] wTc i: 1.0%, ci=ci 0 wTc i: 0.1%, ci=ci 0 wTc i: 0.1%, ci=ci 0/10 wTc i: 0.01%, ci=ci 0/102 N=4 wTc i: 1.0%, ci=ci 0 wTc i: 0.1%, ci=ci 0/10 wTc i: 10-3%, ci=ci 0/103 wTc i: 10-4%, ci=ci 0/102 N=5 wTc i: 1.0%, ci=ci 0 wTc i: 0.1%, no ci wTc i: 0.1%, no ci, range: 2 wTc i: 10-3%, ci=ci 0/103 wTc i: 10-4%, ci=ci 0/104 wTc i: 10-5%, ci=ci 0/105 20 40 60 80 100 120 140 0 200 400 600 800 1000 N=6 wTc i: 1.0%, ci=ci 0 wTc i: 0.1%, ci=ci 0 wTc i: 10-3%, ci=ci 0/103 wTc i: 10-4%, ci=ci 0/104 wTc i: 10-5%, ci=ci 0/105, range: 3 200 400 600 800 1000 N=7 wTc i: 1.0%, ci=ci 0 wTc i: 0.1%, no ci wTc i: 10-5%, no ci wTc i: 10-5%, no ci, range: 2 wTc i: 10-7%, no ci wTc i: 10-7%, no ci, range: 2 200 400 600 800 1000 N=8 Tc [MeV] wTc i: 1.0%, ci=ci 0 wTc i: 0.1%, ci=ci 0/10 wTc i: 10-3%, ci=ci 0/103 wTc i: 10-5%, ci=ci 0/105 wTc i: 10-6%, ci=ci 0/106, range: 3 wTc i: 10-7%, no ci, range: 2 FIG. 2. Result of a mock analysis showing the effect of the error of the data and of the sampling range on the quality of the analytic continuation obtained via multipoint Pad´e approximants of various order and by including or omitting information (the latter is denoted by “no ci” in the key) on the error of the Taylor coefficients in the evaluation of χ2. For additional information see the main text. 1. Convergence in the absence of noise 2. Result of a mock analysis showing the effect of the error of the data and of the sampling range on the quality of the analytic continuation obtained via multipoint Pad´e approximants of various order and by including or omitting information (the latter is denoted by “no ci” in the key) on the error of the Taylor coefficients in the evaluation of χ2. For additional information see the main text. subdiagonal sequence converges from below to the line m2 eff ¼ 0, which for μB < μTCB B is the line of critical points and for μB > μTCB B is the first spinodal. Given that the sampling range is ˆμ2 B ∈ð−7.35; 0, the accuracy of the ½3=4 Pad´e approximant around the location of the TCP is remarkable, even though Tcðˆμ2 BÞ is a rather simple function, as it represents an ellipse. This is even more so when one compares to the radius of convergence of the Taylor expansion around μB ¼ 0 which is ˆμ2 B ≈44.2, as given by the pole in (7). We only refer to the TCP because μB (or ˆμ2 B) is rather large there; the Pad´e approximant does not know about the existence of the TCP, as this is encoded in the quartic part of the tree-level potential and the second derivative of h¯qqi=Φ (q is the constituent quark field) with respect to Φ, which jointly determine λeff. FIG. 2. Result of a mock analysis showing the effect of the error of the data and of the sampling range on the quality of the analytic continuation obtained via multipoint Pad´e approximants of various order and by including or omitting information (the latter is denoted by “no ci” in the key) on the error of the Taylor coefficients in the evaluation of χ2. For additional information see the main text. 2. Effect of the error 1. We also investigate the effect of changing the sampling range for fixed value of the error by increasing the lower bound of the interval by the factor indicated in the keys of Fig. 2. In the modified range the interpolating points are equidistant from each other. 1. Convergence in the absence of noise 40 60 80 100 120 140 160 N=3 B [MeV] wTc i: 1.0%, ci=ci 0 wTc i: 0.1%, ci=ci 0 wTc i: 0.1%, ci=ci 0/10 wTc i: 0.01%, ci=ci 0/102 N=4 wTc i: 1.0%, ci=ci 0 wTc i: 0.1%, ci=ci 0/10 wTc i: 10-3%, ci=ci 0/103 wTc i: 10-4%, ci=ci 0/102 N=5 wTc i: 1.0%, ci=ci 0 wTc i: 0.1%, no ci wTc i: 0.1%, no ci, range: 2 wTc i: 10-3%, ci=ci 0/103 wTc i: 10-4%, ci=ci 0/104 wTc i: 10-5%, ci=ci 0/105 20 40 60 80 100 120 140 0 200 400 600 800 1000 N=6 wTc i: 1.0%, ci=ci 0 wTc i: 0.1%, ci=ci 0 wTc i: 10-3%, ci=ci 0/103 wTc i: 10-4%, ci=ci 0/104 wTc i: 10-5%, ci=ci 0/105, range: 3 200 400 600 800 1000 N=7 wTc i: 1.0%, ci=ci 0 wTc i: 0.1%, no ci wTc i: 10-5%, no ci wTc i: 10-5%, no ci, range: 2 wTc i: 10-7%, no ci wTc i: 10-7%, no ci, range: 2 200 400 600 800 1000 N=8 Tc [MeV] wTc i: 1.0%, ci=ci 0 wTc i: 0.1%, ci=ci 0/10 wTc i: 10-3%, ci=ci 0/103 wTc i: 10-5%, ci=ci 0/105 wTc i: 10-6%, ci=ci 0/106, range: 3 wTc i: 10-7%, no ci, range: 2 40 60 80 100 120 140 160 N=3 B [MeV] wTc i: 1.0%, ci=ci 0 wTc i: 0.1%, ci=ci 0 wTc i: 0.1%, ci=ci 0/10 wTc i: 0.01%, ci=ci 0/102 N=4 wTc i: 1.0%, ci=ci 0 wTc i: 0.1%, ci=ci 0/10 wTc i: 10-3%, ci=ci 0/103 wTc i: 10-4%, ci=ci 0/102 N=5 wTc i: 1.0%, ci=ci 0 wTc i: 0.1%, no ci wTc i: 0.1%, no ci, range: 2 wTc i: 10-3%, ci=ci 0/103 wTc i: 10-4%, ci=ci 0/104 wTc i: 10-5%, ci=ci 0/105 20 40 60 80 100 120 140 0 200 400 600 800 1000 N=6 wTc i: 1.0%, ci=ci 0 wTc i: 0.1%, ci=ci 0 wTc i: 10-3%, ci=ci 0/103 wTc i: 10-4%, ci=ci 0/104 wTc i: 10-5%, ci=ci 0/105, range: 3 200 400 600 800 1000 N=7 wTc i: 1.0%, ci=ci 0 wTc i: 0.1%, no ci wTc i: 10-5%, no ci wTc i: 10-5%, no ci, range: 2 wTc i: 10-7%, no ci wTc i: 10-7%, no ci, range: 2 200 400 600 800 1000 N=8 Tc [MeV] wTc i: 1.0%, ci=ci 0 wTc i: 0.1%, ci=ci 0/10 wTc i: 10-3%, ci=ci 0/103 wTc i: 10-5%, ci=ci 0/105 wTc i: 10-6%, ci=ci 0/106, range: 3 wTc i: 10-7%, no ci, range: 2 FIG. 1. Numerical implementation of the Bayesian approach 1. Numerical implementation of the Bayesian approach In order to use the method presented in Sec. III A, we need to generate fTicg instances at chosen interpolating points (also values of c1 and c2 in the case of the Pad´e-type approximant) and then evaluate χ2, defined in (4), using the actual lattice data as control points. The interpolating points ˆμ2 B;i used for the two types of Pad´e approximants mentioned above are indicated in Fig. 3. The idea behind our choice was that each interpolating point of any of the used approximant fall in between two nearby lattice data points and be more or less equally distributed in the sampling range. The actual choice of the interpolating points is not important, however, in order to maximize the sampling range, one interpolating point is chosen close to the lattice data point with ˆμ2 Bðj ¼ 7Þ and, since we are interested in p g g The second method (Method 2) for generating input for the Pad´e approximants is the importance sampling using the Metropolisalgorithmwith “action” χ2=2. Theproposed value of Tic in the Markov chain is generated using a normal distribution for the noise with vanishing mean and SD of Oð1Þ MeV. In the case of the Pad´e-type approximant we use normal distribution with standard deviation σci; i ¼ 1, 2 to generate the noise for the Taylor coefficients ci. Configurations for which the corresponding Pad´e approx- imant has spurious poles in the range given above are excluded by assigning to them the value χ2 ¼ ∞. For the remaining configurations χ2 is calculated using all the available lattice data according to the formulas in (4). The averageandpercentilebasederrorofTc ¼ Rnmðˆμ2 BÞandμB for the Pad´e approximants were calculated in the standard way withtheconfigurationsprovidedbytheMetropolisalgorithm. 155 160 165 170 175 180 185 -8 -7 -6 -5 -4 -3 -2 -1 0 Tc [MeV] B 2/T2 W-Bp lattice data N=3, [1/1] 5, [2/2] 6, [2/3] 8, [3/4] mP: Pt: wTc i N N6 N4 N4 N=4,7,8 N5 N=8 N N N=8 N6 N=5,7,8 N=6 N7 N FIG. 3. Choice of the interpolating (node) points, whose position is indicated by the vertical dotted line, in comparison to the actual lattice data. The labels indicate the use of the interpolating point in the multipoint Pad´e approximant (bottom) and the Pad´e-type approximant (top) of a given order, charac- terized by the number of independent parameters N. 1. Numerical implementation of the Bayesian approach At the value of ˆμ2 B corresponding to the lattice data points we show the mean and the error of the Tc computed from the multipoint Pad´e approximants generated with importance sampling (for the sake of the presentation the abscissa is shifted). The band indicates the standard deviation of the normal distribution used in Method 1 to generate fTicg instances, i.e., it indicates the prior distribution, excluding the factor that removes the spurious poles. The values ¯Tcðˆμ2 B;iÞ at the node points ˆμ2 B;i used in the expression (5) of the prior are from the dotted curve in the band, which interpolates the mean values of the lattice data. 155 160 165 170 175 180 185 -8 -7 -6 -5 -4 -3 -2 -1 0 Tc [MeV] B 2/T2 W-Bp lattice data N=3, [1/1] 5, [2/2] 6, [2/3] 8, [3/4] mP: Pt: wTc i N N6 N4 N4 N=4,7,8 N5 N=8 N N N=8 N6 N=5,7,8 N=6 N7 N There are some peculiarities when doing importance sampling in this context. These are related to the spurious poles of the Pad´e approximants, which appear as “walls” of infinite action in the Metropolis update. Configurations with spurious poles are not guaranteed to be isolated points in the space of all configurations, rather, there can be regions in configurations space where all approximants have a pole. One can easily stumble on an accepted configuration that is surrounded in most directions by configurations with a pole, thereby trapping the algorithm. To avoid this problem it is a good idea to mark out a temperature range sampled by the algorithmduringtherandomwalkandassigninfinitevaluefor the action if a proposed Tic lies outside this range. I.e., even in thecaseofMethod2aprior,likethatinFig.3isused.Another reason to introduce this band is to exclude the Pad´e approx- imants from having features in the interpolated range not presentinthedata,evenifsuchanapproximanthasnopoleand fits the data points acceptably. This is also a possibility, since Pad´e approximants are rather flexible. In practice a two times wider band than the one shown in Fig. 3 proved sufficient. FIG. 3. Choice of the interpolating (node) points, whose position is indicated by the vertical dotted line, in comparison to the actual lattice data. The labels indicate the use of the interpolating point in the multipoint Pad´e approximant (bottom) and the Pad´e-type approximant (top) of a given order, charac- terized by the number of independent parameters N. APPARENT CONVERGENCE OF PADÉ APPROXIMANTS FOR … PPARENT CONVERGENCE OF PADÉ APPROXIMANTS FOR … analytic continuation through μB ¼ 0, we also choose ˆμ2 Bðj ¼ 0Þ ¼ 0 as an interpolating point. With the notation of Sec. II, the (assumed) function fðtÞ, which corresponds to Tcðˆμ2 BÞ, is known at seven points τj ¼ ˆμ2 BðjÞ < 0, corresponding to j ≠0 in the list given in (8), and we also know c0 ¼ TcðˆμBðj ¼ 0ÞÞ, as well as c1 and c2 in terms of κ2 and κ4. The values of κ2, κ4 and Tcð0Þ reported in [2] give through the explicit relations given in Appendix B c1 ¼ −1.878 and c2 ¼ 0.0451 with errors σc1 ¼ 0.626 and σc2 ¼ 0.627. We use two methods to generate input for the Pad´e approximants. In the first method (Method 1) we simply generate Tic from normal distribution with mean obtained by interpolating the mean of the lattice data and with the standard deviation (SD) indicated in Fig. 3. In this case c1 and c2, used in the Pad´e-type approximant, are generated from a normal distribution with mean and SD given by Eqs. (B2) and (B3), respectively. As a result, c1 and c2 are taken into account in the calculation of χ2 only when using the multipoint Pad´e approximant. According to our prior, we only accept those configurations for which the corre- sponding Pad´e approximant is free of spurious poles in the wide range ˆμ2 B ∈½−π2; 60π2. When using this method we calculate Tc at some value of ˆμ2 B as Tc ¼ Rnmðˆμ2 BÞ and the value of the real chemical potential as μB ¼ ﬃﬃﬃﬃﬃ ˆμ2 B p Tc and determine their percentile based error using theweight e−χ2=2. B. Analytic continuation of QCD data We apply he method presented in Sec. II to the continuation of the critical line of the QCD in the T − μB plane. Our main focus is the study of the convergence of Pad´e series of the form ½p=p and ½p=p þ 1 constructed: ½ ½ (1) based only on interpolating points (multipoint Pad´e approximants), or (1) based only on interpolating points (multipoint Pad´e approximants), or (2) using interpolating points and the expansion fðtÞ ≈ c0 þ c1t þ c2t2 around t ¼ 0, as explained in Sec. III A (Pad´e-type approximant). ( yp pp ) We use the continuum extrapolated values of Tc recently determined on the lattice at μB ¼ 0 and seven imaginary values of ˆμB ¼ μB=T, namely ˆμBðjÞ ¼ i jπ 8 ; j ¼ 0; 2; 3; 4; 5; 6; 6.5; 7; ð8Þ ð8Þ given in Table II of [3] and the Taylor coefficients κ2 and κ4, appearing in the parametrization TcðμBÞ¼Tcð0Þ½1−κ2ðμB=Tcð0ÞÞ2 −κ4ðμB=Tcð0ÞÞ4; ð9Þ also extrapolated to the continuum limit in [2]. TcðμBÞ¼Tcð0Þ½1−κ2ðμB=Tcð0ÞÞ2 −κ4ðμB=Tcð0ÞÞ4; ð9Þ also extrapolated to the continuum limit in [2]. TcðμBÞ¼Tcð0Þ½1−κ2ðμB=Tcð0ÞÞ2 −κ4ðμB=Tcð0ÞÞ4; ð9Þ also extrapolated to the continuum limit in [2]. ð9Þ Worth noticing in Fig. 2 is that in the presence of noise the bands for TcðμBÞ can deviate above some value of μB from 034511-6 PHYS. REV. D 103, 034511 (2021) IV. CONCLUSIONS AND OUTLOOK We presented a method for the numerical analytic continuation of data available at imaginary chemical poten- tial that uses also the Taylor coefficients of an expansion around μB ¼ 0. Using lattice data that became available recently, we have investigated the continuation to real μB of the crossover line with a sequence of Pad´e approximants, looking for apparent convergence as the number of inde- pendent coefficients increases. Such an analysis would have been less conclusive using the smaller data set available at imaginary μB in [31] and without taking into account the lattice data for the Taylor coefficients. Now we turn our attention to the extrapolation. Since both Method 1 and Method 2 used to generate input for both the multipoint Pad´e and Pad´e-type approximants resulted in very similar resultsfortheanalyticallycontinuedTcðμBÞ curve,we only present those obtained with Method 2 (importance sampling) in the case of the Pad´e-type approximant con- structed using the first and second derivative of TcðμBÞ at μB ¼ 0. The fact that with Method 1 the analytic continuation does not depend on the approximant used, means that it makes no difference whether we take into account the Taylor coefficients only in the approximant or only in the calculation of χ2. We remind that when importance sampling is used the Taylor coefficients are taken into account in the calculation of χ2 irrespective ofthetypeofapproximant, since otherwise the range in which c1 and c2 varies during the random walk would not be constrained. Our largest order Pad´e approximants is very close to the simplest quadratic curve obtained with just the κ2 coefficient. This means that if the observed apparent convergence is genuine, such a quadratic approximation might be applicable in a rather large range of μB. We would like to stress that, as discussed in the case of an effective model in Sec. III A, our results on the analytic continuation tell nothing on the possible existence and location of the critical end point (CEP). It is also not possible to clearly determine the value of μB up to which the analytic continuation could be trusted. Our main result on the analytic continuation is presented in Fig. 4 in comparison with a simple parametrization of the crossover line based on the Taylor coefficient κ2. One sees that with the exception of the [2/2] type, the Pad´e approximants tend to give smaller Tc with increasing μB. 2. Results for the analytic continuation The first thing worth checking is the distribution of Tc calculated from the Pad´e approximants at ˆμ2 B values corre- sponding to the actual lattice data points. For the majority of the approximants and lattice data points, the distribution is very close to a normal one with standard deviation com- patible with that of the lattice data. The latter can be seen in Fig. 3 in the case of the multipoint Pad´e approximant, meaning that the selection of the Tc instances based on χ2 works as expected. However, different low order approx- imants seem to select, within the error, different ranges in the distribution of Tc (and ci when the Pad´e-type approximant is used). This is most visible in the case of the [2/2] approx- imant where the points posses a structure unseen in the lattice data. The multipoint Pad´e approximant [1/1] is the most constrained by the likelihood, the error of Tc being smaller than the lattice one, while the [3/4] approximant is the least constrained, matching closely the lattice error at all lattice points, and showing a wider range of the computed c1 and c2 coefficients. This loss of constraintis also reflectedby the χ2 histogram whose pick moves to higher values when the number of parameters of the approximant increases. FIG. 4. Result of the analytic continuation via the diagonal and subdiagonal sequences of Pad´e-type approximants constructed based on fTicg instances generated using importance sampling. The inset shows the percentage difference with respect to the Tcð0Þð1 −κ2μ2 B=T2cð0ÞÞ curve plotted in the main figure, for which we used Tcð0Þ ¼ 158.01 MeV and κ2 ¼ 0.012. laid out by the [2/3] approximant. It remains to be seen if this pattern survives the possible addition of new lattice data points, which will further constrain the fit, and/or an increase in the precision of the lattice data. PÁSZTOR, SZ´EP, and MARKÓ PÁSZTOR, SZ´EP, and MARKÓ 0 50 100 150 200 0 100 200 300 400 500 600 Tc(0) (1- 2 2 B/Tc 2(0))=:Tc T(B) Tc [MeV] B [MeV] N=4, [1/2] 5, [2/2] 6, [2/3] 7, [3/3] 8, [3/4] -15 -10 -5 0 5 10 15 0 5 10 15 20 25 (Rm n /Tc T-1)100 B 2 /T2 FIG. 4. Result of the analytic continuation via the diagonal and subdiagonal sequences of Pad´e-type approximants constructed based on fTicg instances generated using importance sampling. The inset shows the percentage difference with respect to the Tcð0Þð1 −κ2μ2 B=T2cð0ÞÞ curve plotted in the main figure, for which we used Tcð0Þ ¼ 158.01 MeV and κ2 ¼ 0.012. 0 50 100 150 200 0 100 200 300 400 500 600 Tc(0) (1- 2 2 B/Tc 2(0))=:Tc T(B) Tc [MeV] B [MeV] N=4, [1/2] 5, [2/2] 6, [2/3] 7, [3/3] 8, [3/4] -15 -10 -5 0 5 10 15 0 5 10 15 20 25 (Rm n /Tc T-1)100 B 2 /T2 Another observation is that in some cases it was very hard to thermalize the system by updating the value of Tc only at one interpolating point at a time. It proved more useful to propose in the Metropolis algorithm an updated array of Tc values, as this procedure also substantially reduced the autocorrelation time. 1. Numerical implementation of the Bayesian approach At the value of ˆμ2 B corresponding to the lattice data points we show the mean and the error of the Tc computed from the multipoint Pad´e approximants generated with importance sampling (for the sake of the presentation the abscissa is shifted). The band indicates the standard deviation of the normal distribution used in Method 1 to generate fTicg instances, i.e., it indicates the prior distribution, excluding the factor that removes the spurious poles. The values ¯Tcðˆμ2 B;iÞ at the node points ˆμ2 B;i used in the expression (5) of the prior are from the dotted curve in the band, which interpolates the mean values of the lattice data. 034511-7 PHYS. REV. D 103, 034511 (2021) APPENDIX B: RELATING c1;2 WITH κ2;4 In order to relate the coefficients c1 and c2 of the Taylor expansion TcðtÞ ¼ c0 þ c1t þ c2t2, with t ¼ ˆμ2 B, with the coefficients κ2 and κ4 used by the HotQCD Collaboration, the expansion (9) has to be rewritten in terms of ˆμ2 B∶ IV. CONCLUSIONS AND OUTLOOK Also, the behavior of the diagonal an subdiagonal sequen- ces follow different patterns, similar to that observed in the model study in Fig. 1. Apparently, the Pad´e sequences converge, as, although [3/3] and [3/4] have overlapping error bars of similar size, the latter moves toward the band The Taylor and imaginary chemical potential methods are usually considered to be competitors in the study of finite density QCD. This is somewhat unfortunate, as the two methods tend to provide complimentary information. With the Taylor method, lower order coefficients tend to be more precise, while data at imaginary μB tends to restrict higher order coefficients better, without giving a very 034511-8 PHYS. REV. D 103, 034511 (2021) APPARENT CONVERGENCE OF PADÉ APPROXIMANTS FOR … gpðzÞ ¼ gp−1ðzp−1Þ −gp−1ðzÞ ðz −zp−1Þgp−1ðzÞ ; precise value for the lower orders. For the case of baryon number fluctuations, this can clearly be seen by comparing Fig. 3 of Ref. [27], where the signal for χB 6 and χB 8 is better, with Fig. 1 of Ref. [43], where χB 4 is much more precise. This means that joint analysis of such data might be a good idea also for the equation of state, where there are some indications—both from an explicit calculation on coarser lattices [44–46] and phenomenological arguments [47–49] —that the radius of convergence for temperatures close to the crossover is of the order μB=T ≈2, making a Taylor ansatz unusable beyond that point. This makes it mandatory to try different ansatze, or resummations of the Taylor expansion, and one possible choice could be the Pad´e approximation method used here. 1 ≤p ≤N −1; ðA2Þ with initial condition g0ðzÞ ¼ fðzÞ, which means g0ðziÞ ¼ fi, when the function is known only in some discrete points. Working out explicitly the condition Ai ¼ giðziÞ for a few values of i, one sees that one needs to construct an upper triangular matrix ti;j using the recursion ti;j ¼ ðti−1;i−1=ti−1;j −1Þ=ðzj −zi−1Þ, for j ¼ 1; …; N −1 and i ¼ 1; …; j, starting from its first row t0;j ¼ fj, j ¼ 0; …; N −1. The diagonal elements are the coeffi- cients of CN: Ai ¼ ti;i. IV. CONCLUSIONS AND OUTLOOK The relation of CN with the Pad´e sequence is as follows: if N ≥1 is odd, then CN ¼ ½p=p with p ¼ ðN −1Þ=2, while when N ≥1 is even, then CN ¼ ½p=p þ 1 with p ¼ −1 þ N=2. ½ Writing CNðzÞ in the form CNðzÞ ¼ NðzÞ=DðzÞ, the numerator and denominator (at a given value of z) can be easily determined from the coefficients of the truncated continued fraction via the following three-term recurrence relation ACKNOWLEDGMENTS We would like to thank Sz. Borsányi, M. Giordano, S. Katz and Z. Rácz for illuminating discussions on the subject and J. Günther for providing the raw lattice data of [31] in an early stage of the project. This work was partially supported by the Hungarian National Research, Development and Innovation Office—NKFIH Grants No. KKP126769 and No. PD_16 121064, as well as by the DFG (Emmy Noether Program EN 1064/2-1). A. P. is supported by the János Bolyai Research Scholarship of the Hungarian Academy of Sciences and by the ÚNKP-20-5 New National Excellence Program of the Ministry of Innovation and Technology. In an early stage this research was also supported by the Munich Institute for Astro- and Particle Physics (MIAPP) of the DFG cluster of excellence “Origin and Structure of the Universe”. Xnþ1 ¼ Xn þ ðz −znÞAnþ1Xn−1; ðA3Þ ðA3Þ where for the numerator (X ¼ N) one has X1 ¼ 0, X0 ¼ A0 and for the denominator (X ¼ D) one has X1 ¼ X0 ¼ 1, and the iteration goes from n ¼ 0 up to and including n ¼ N −2. The coefficients ai (bi) of the numerator (denominator) can be easily obtained by calling the above recursion (A3) at a finite number of points z and solving a system of linear equations. In c1 and c2 and their errors we use the data of the HotQCD Collaboration also for Tcð0Þ. APPENDIX A: THE MULTIPOINT PADÉ APPROXIMATION METHOD Rothkopf, Improved maximum entropy analysis with an extended search space, J. Comput. Phys. 238, 106 (2013). [3] S. Borsányi, Z. Fodor, J. N. Guenther, R. Kara, S. D. Katz, P. Parotto, A. Pásztor, C. Ratti, and K. K. Szabó, The QCD Crossover at Finite Chemical Potential from Lattice Sim- ulations, Phys. Rev. Lett. 125, 052001 (2020). [24] Y. Burnier and A. Rothkopf, Bayesian Approach to Spectral Function Reconstruction for Euclidean Quantum Field Theories, Phys. Rev. Lett. 111, 182003 (2013). [25] S. Borsanyi et al., Charmonium spectral functions from 2 þ 1 flavour lattice QCD, J. High Energy Phys. 04 (2014) 132. [4] A. N. Tikhonov and V. Y. Arsenin, Solutions of Ill-Posed Problems (V. H. Winston & Sons, Washington, DC, 1977). [5] I. Sabba Stefanescu, On the stable analytic continuation with a condition of uniform boundedness, J. Math. Phys. (N.Y.) 27, 2657 (1986). [26] A. Rothkopf, Bayesian techniques and applications to QCD, Proc. Sci., Confinement2018 (2018) 026 [arXiv:1903 .02293]. [27] S. Borsanyi, Z. Fodor, J. N. Guenther, S. K. Katz, K. K. Szabo, A. Pasztor, I. Portillo, and C. Ratti, Higher order fluctuations and correlations of conserved charges from lattice QCD, J. High Energy Phys. 10 (2018) 205. [6] C. M. Bender and S. A. Orszag, Advanced Mathematical Methods for Scientists and Engineers I, Asymptotic Methods and Perturbation Theory (Springer, New York, 1999). [7] G. A. Baker, Essentials of Pad´e approximants (Academic Press, New York, 1975). [28] M. P. Lombardo, Series representation: Pade’ approximants and critical behavior in QCD at nonzero T and μ, Proc. Sci., LAT2005 (2006) 168 [arXiv:hep-lat/0509181]. [8] L. Schlessinger, Use of analyticity in the calculation of nonrelativistic scattering amplitudes, Phys. Rev. 167, 1411 (1968). [29] P. Cea, L. Cosmai, M. D’Elia, C. Manneschi, and A. Papa, Analytic continuation of the critical line: Suggestions for QCD, Phys. Rev. D 80, 034501 (2009). [9] J. Gelfgren, Multipoint Pad´e approximants used for piece- wise rational interpolation and for interpolation to func- tions of Stieltjes’ type, Research Report, (University of Umeå, 1978), https://www.diva-portal.org/smash/get/diva2: 638352/FULLTEXT02. [30] P. Cea, L. Cosmai, M. D’Elia, A. Papa, and F. Sanfilippo, Critical line of two-flavor QCD at finite isospin or baryon densities from imaginary chemical potentials, Phys. Rev. D 85, 094512 (2012). [10] J. Nuttall, Convergence of Pad´e approximants of meromor- phic functions, J. Math. Anal. Appl. 31, 147 (1970). [31] R. Bellwied, S. Borsányi, Z. Fodor, J. Günther, S. D. Katz, C. APPENDIX A: THE MULTIPOINT PADÉ APPROXIMATION METHOD Ratti, and K. K. Szabó, The QCD phase diagram from analytic continuation, Phys. Lett. B 751, 559 (2015). [11] J. Zinn-Justin, Convergence of Pad´e approximants in the general case, Rocky Mountain J. Math. 4, 325 (1974). [32] F. Karsch, B. J. Schaefer, M. Wagner, and J. Wambach, Towards finite density QCD with Taylor expansions, Phys. Lett. B 698, 256 (2011). [12] M. Froissart, Approximation de Pad´e: application `a la physique des particules ´el´ementaires, in RCP, Programme No. 25 (CNRS, Strasbourg, 1969), Vol. 9, pp. 1–13. [33] S. Datta, R. V. Gavai, and S. Gupta, Quark number susceptibilities and equation of state at finite chemical potential in staggered QCD with Nt ¼ 8, Phys. Rev. D 95, 054512 (2017). [13] P. Gonnet, S. Güttel, and L. N. Trefethen, Robust Pad´e approximation via SVD, SIAM Rev. 55, 101 (2013). [14] B. Beckerman, G. Labahn, and A. C. Matos, On rational functions without Froissart doublets, Numer. Math. 138, 615 (2018). [34] J. N. Guenther, R. Bellwied, S. Borsanyi, Z. Fodor, S. D. Katz, A. Pasztor, C. Ratti, and K. K. Szabó, The QCD equation of state at finite density from analytical continu- ation, Nucl. Phys. A967, 720 (2017). [15] A. Pilaftsis and D. Teresi, Symmetry improved CJTeffective action, Nucl. Phys. B874, 594 (2013). [16] G. Markó, U. Reinosa, and Zs. Sz´ep, Pad´e approximants and analytic continuation of Euclidean Φ-derivable approxima- tions, Phys. Rev. D 96, 036002 (2017). [35] G. A. Baker, Application of the Pad´e approximant method to the investigation of some magnetic properties of the Ising model, Phys. Rev. 124, 768 (1961). [17] R. A. Tripolt, P. Gubler, M. Ulybyshev, and L. Von Smekal, Numerical analytic continuation of Euclidean data, Comput. Phys. Commun. 237, 129 (2019). [36] G. A. Baker, Convergent, bounding approximation proce- dures with applications to the ferromagnetic Ising model, Phys. Rev. 161, 434 (1967). [18] C. Brezinski and M. Redivo-Zaglia, Pad´e-type rational and barycentric interpolation, Numer. Math. 125, 89 (2013). [37] A. J. Guttmann, Asymptotic analysis of power-series ex- pansions, in Phase Transitions and Critical Phenomena, edited by C. Domb and J. L. Lebowitz (Academic Press, New York, 1989), Vol. 13, pp. 1–234. [19] M. Asakawa, T. Hatsuda, and Y. Nakahara, Maximum entropy analysis of the spectral functions in lattice QCD, Prog. Part. Nucl. Phys. 46, 459 (2001). [38] G. A. Baker and J. L. Gammel, The Pad´e Approximant in Theoretical Physics, Mathematics in Science and Engineer- ing, Vol. APPENDIX A: THE MULTIPOINT PADÉ APPROXIMATION METHOD Following Refs. [38,50], we briefly summarize the construction of the multipoint Pad´e approximant used to analytically continue functions known only at a finite number of points of the complex plane. In our case the continuation is done along the real axis, from negative to positive values. TcðμBÞ Tcð0Þ ¼ 1 −κ2 T2cðμBÞ T2cð0Þ ˆμ2 B −κ4 T4cðμBÞ T4cð0Þ ˆμ4 B: ðB1Þ ðB1Þ Then, using that T2cðμBÞ=T2cð0Þ ≈1–2κ2ˆμ2 B and T4cðμBÞ= T4cð0Þ ¼ 1 þ Oðˆμ2 BÞ, we obtain Then, using that T2cðμBÞ=T2cð0Þ ≈1–2κ2ˆμ2 B and T4cðμBÞ= T4cð0Þ ¼ 1 þ Oðˆμ2 BÞ, we obtain When one knows the function at N points fi ¼ fðziÞ; i ¼ 0…N −1, the rational function approximating fðzÞ is most conveniently given as a truncated continued fraction c1 ¼ −κ2Tcð0Þ and c2 ¼ ðκ4 −2κ2 2ÞTcð0Þ: ðB2Þ ðB2Þ Ignoring the covariance between κ2 and κ4, which is not known to us, the error associated to these Taylor coef- ficients are CNðzÞ ¼ A0 1þ A1ðz −z0Þ 1þ AN−1ðz −zN−2Þ 1 ; ðA1Þ ðA1Þ where we used the notation 1 1þ x ≡ 1 1þx. The task is to determine the N coefficients Ai from the conditions CNðziÞ ¼ fi, i ¼ 0…N −1. Note that only N −1 values of zi appear in (A1), zN−1 appears in the condition CNðzN−1Þ ¼ fN−1. The coefficients can be obtained effi- ciently as Ai ¼ giðziÞ, i ¼ 0…N −1, with the functions giðzÞ defined by the recursion σc1 ¼ ½T2cð0Þσ2κ2 þ κ2 2σ2 Tcð0Þ 1 2; σc2 ¼ ½T2cð0Þðσ2κ4 þ 16κ2 2σ2κ2Þ þ ðκ2 4 þ 4κ4 2Þσ2 Tcð0Þ 1 2: ðB3Þ ðB3Þ In c1 and c2 and their errors we use the data of the HotQCD Collaboration also for Tcð0Þ. 034511-9 PÁSZTOR, SZ´EP, and MARKÓ PHYS. REV. D 103, 034511 (2021) in Hot, Quenched Lattice QCD, Phys. Rev. Lett. 99, 022002 (2007). [1] C. Bonati, M. D’Elia, F. Negro, F. Sanfilippo, and K. Zambello, Curvature of the pseudocritical line in QCD: Taylor expansion matches analytic continuation, Phys. Rev. D 98, 054510 (2018). in Hot, Quenched Lattice QCD, Phys. Rev. Lett. 99, 022002 (2007). [22] H. B. Meyer, Transport properties of the quark-gluon plasma: A lattice QCD perspective, Eur. Phys. J. A 47, 86 (2011). [2] A. Bazavov et al. (HotQCD Collaboration), Chiral cross- over in QCD at zero and non-zero chemical potentials, Phys. Lett. B 795, 15 (2019). [23] A. APPENDIX A: THE MULTIPOINT PADÉ APPROXIMATION METHOD 71 (Academic Press, New York, 1970). [20] A. Jakovác, P. Petreczky, K. Petrov, and A. Velytsky, Quarkonium correlators and spectral functions at zero and finite temperature, Phys. Rev. D 75, 014506 (2007). [21] G. Aarts, C. Allton, J. Foley, S. Hands, and S. Kim, Spectral Functions at Small Energies and the Electrical Conductivity [39] K. P. Murphy, Machine Learning: A Probabilistic Perspec- tive (MIT Press, Cambridge, MA, 2012). 034511-10 PHYS. REV. D 103, 034511 (2021) APPARENT CONVERGENCE OF PADÉ APPROXIMANTS FOR … with rooted staggered fermions, Phys. Rev. D 101, 074511 (2020). [40] L. N. Trefethen, Approximation Theory and Approximation Practice (Society for Industrial and Applied Mathematics, Philadelphia, 2012). [46] M. Giordano, K. Kapas, S. D. Katz, D. Nogradi, and A. Pasztor, Towards a reliable lower bound on the location of the critical endpoint, Nucl. Phys. A1005, 121986 (2021). [41] A. Jakovác, A. Patkós, Zs. Sz´ep, and P. Sz´epfalusy, T −μ phase diagram of the chiral quark model from a large flavor number expansion, Phys. Lett. B 582, 179 (2004). [47] A. Connelly, G. Johnson, S. Mukherjee, and V. Skokov, Universality driven analytic structure of QCD crossover: Radius of convergence and QCD critical point, Nucl. Phys. A1005, 121834 (2021). [42] A. Ayala, A. Bashir, J. J. Cobos-Martinez, S. Hernandez- Ortiz, and A. Raya, The effective QCD phase diagram and the critical end point, Nucl. Phys. B897, 77 (2015). [48] S. Mukherjee and V. Skokov, Universality driven analytic structure of QCD crossover: Radius of convergence in baryon chemical potential, arXiv:1909.04639. [43] A. Bazavov et al. Skewness, kurtosis, and the fifth and sixth order cumulants of net baryon-number distributions from lattice QCD confront high-statistics STAR data, Phys. Rev. D 101, 074502 (2020). [49] A. Connelly, G. Johnson, F. Rennecke, and V. V. Skokov, Universal Location of the Yang-Lee Edge Singularity in O(N) Theories, Phys. Rev. Lett. 125, 191602 (2020). [44] M. Giordano and A. Pásztor, Reliable estimation of the radius of convergence in finite density QCD, Phys. Rev. D 99, 114510 (2019). [50] H. J. Vidberg and J. W. Serene, Solving the Eliashberg equations by means of N-point Pad´e approximants, J. Low Temp. Phys. 29, 179 (1977). [45] M. Giordano, K. Kapas, S. D. Katz, D. Nogradi, and A. Pasztor, Radius of convergence in lattice QCD at finite μB 034511-11
https://openalex.org/W1978638664	https://eejournal.ktu.lt/index.php/elt/article/download/9806/4778	English	null	Performance of Two-Hop Relay Assisted Decode-and-Forward Transmission under Mixed Fading Environments	Elektronika ir elektrotechnika	2,015	cc-by	3,542	ELEKTRONIKA IR ELEKTROTECHNIKA, ISSN 1392-1215, VOL. 21, NO. 1, 2015 ELEKTRONIKA IR ELEKTROTECHNIKA, ISSN 1392-1215, VOL. 21, NO. 1, 2015 http://dx.doi.org/10.5755/j01.eee.21.1.9806 Performance of Two-Hop Relay Assisted Decode-and-Forward Transmission under Mixed Fading Environments Nuri Kapucu1, Mehmet Bilim1, Ibrahim Develi1, Yasin Kabalci2 Department of Electrical and Electronics Engineering, Erciyes University, 38039, Kayseri, Turkey 2Nigde Vocational College of Technical Sciences, Nigde University, 51200, Nigde, Turkey nurikapucu@erciyes.edu.tr of each hop is not the same [5]–[10]. It is due to the fact that the received signals are different for each relay link. Xu et al. [5] proposed exact expressions of two-hop AF relay system for outage probability ( out P ) and average symbol error probability (ASEP) in mixed Nakagami-m and Rician faded wireless environments. This study was achieved by using the cumulative distribution function (CDF) of the total signal-to-noise ratio (SNR). Similar to [5], in [6], a dual-hop nonregenerative transmission system where the source-relay and relay-destination channels are subject to Rayleigh and Rician distributions was considered. The authors in [6] obtained exact and approximate formulas for out P and ABEP. In [7], an AF relaying system in which the links experience Rayleigh/Rician and Rician/Rayleigh fading environments was examined. Ouyang et al. [8] found some approximations on the analytical and asymptotic average symbol error rate (ASER) expressions for dual-hop AF relaying under Rayleigh/Rician asymmetric fading conditions. The proposed expression in [8] was derived based on moment generating function (MGF) of the total SNR. In [9], out P of two-hop DF relaying scheme was evaluated for Rayleigh/Gamma fading conditions. An exact out P expression for the cooperation model was introduced in [9]. Lee et al. [10] derived a closed-form formulation for out P of a two-hop relaying system which consists of radio- frequency (RF) and free-space optics (FSO) channels. In [11], an AF relaying transmission model was investigated based on formulated expressions for out P and ABEP under mixed Rayleigh/Hoyt fading. 1Abstract—This study analyses the performance of two-hop decode-and-forward (DF) relaying protocol under mixed faded channels by evaluating outage probability ( out P ) and average bit-error probability (ABEP). It is assumed that the channel between source and relay is Rayleigh fading, whereas the relay- destination link is subject to Weibull fading. First, a closed- form expression is formulated for out P and then it is used to obtain ABEP for the considered system model. The ABEP performance of two-hop DF relaying is presented for M-ary phase shift keying (M-PSK) signaling with different constellation sizes. Moreover, the proposed analytical expressions are verified by simulations that present the correctness of the derived formulas. Index Terms—Decode-and-forward, two-hop relaying, Weibull distribution. Manuscript received January 9, 2014; accepted July 9, 2014. II. GENERAL DESCRIPTIONS         1 2 1 2 ( ) Pr min , 1 Pr Pr , eq out th th th th P F                (3) B. Channel Model The channels between the S−R and R−D channels experience Rayleigh and Weibull fading conditions. While SR h denotes flat Rayleigh fading coefficient, RD h represents flat Weibull fading coefficient. The probability density function (PDF) of 1  is exponentially distributed with Rayleigh PDF given as In order to solve (5), we use [14, equation (3.381.9)] and   2 Pr th    can be derived as     2 1 2 2 2 2 Pr 1, . th c c th th c p d                       (6) (6)   1 1 1 1 1 1 exp , f             (1) (1) where  ,  is the incomplete gamma function [14]. By substituting (4) and (6) into (3), out P is derived given as where   2 1 1 0 / SR E h P N  is the average value of 1  and 1 1 2 1 1, exp . c c th th out c P                                 (7) (7)  E is the expectation operator.  E is the expectation operator. The PDF of 2  can be given according to Weibull distribution as A. System Model In Fig. 1, a two-hop DF transmission scheme is addressed where no direct transmission occurs between the source (S) and the destination (D). In such a system, a source communicates with the destination via a relay (R) node. The R is considered to be operating in the half duplex mode. In DF relaying protocol, the source message is transmitted to the R and then, the estimated source message at R is transmitted to the D. (3) where ( ) eq th F  is the CDF of the end-to-end SNR.   1 Pr th    is determined according to Rayleigh fading PDF for S-R link as     1 1 1 1 1 Pr exp . th th th p d                     (4) (4) Fig. 1. Illustration of two-hop relaying scheme. For R-D link over Weibull fading channel,   2 Pr th    can be calculated as follows For R-D link over Weibull fading channel,   2 Pr th    can be calculated as follows Fig. 1. Illustration of two-hop relaying scheme. Fig. 1. Illustration of two-hop relaying scheme. 2 1 1 0 / SR h P N  and 2 2 2 0 / RD h P N   indicate instantaneous SNRs of S-R and R-D links, respectively. 1P and 2 P are the transmission power at S and R, respectively. 0 N is the additive white Gaussian noise (AWGN).     2 2 2 2 2 2 2 2 1 1 1 2 Pr exp . th th th c c c c c c p d c d                                     (5) (5) ELEKTRONIKA IR ELEKTROTECHNIKA, ISSN 1392-1215, VOL. 21, NO. 1, 2015 ELEKTRONIKA IR ELEKTROTECHNIKA, ISSN 1392-1215, VOL. 21, NO. 1, 2015 provides good fit of the statistical characteristics of wireless fading channels for theoretical studies [13]. Motivated by these observations, we study DF technique based a dual-hop communication system where the channels of two hops experience Rayleigh and Weibull faded conditions for the first time. Moreover, we derived the exact out P expression for the considered system and illustrated some numerical results for different scenarios. Furthermore, an analytical expression of ABEP is obtained for different modulation schemes. This task is done by evaluating the difficult integral part of the ABEP expression. We also provide computer simulations and analytical results on ABEP to show the validity of the derived expressions. where c is parameter that expresses the severity of fading and is the scaling parameter.   2 2 2 0 / RD E h P N   is the average value of 2 . A. Outage Probability Outage probability, out P , is described as the probability that the received SNR drops below a pre-defined outage level, th  . In the DF relaying, outage occurs if the instantaneous SNR of one or two of the hops fall below the outage level, th  . So, out P is formulated according to the total SNR [4] as I. INTRODUCTION Two-hop transmission technologies have attracted much interest in wireless communications due to its connectivity when direct transmission is impractical because of possible shadowing effects or path loss attenuation. In real wireless communication environments, the links between cooperating nodes can experience asymmetric fading conditions. Several works have analysed two-hop transmission systems which use Amplify-and-Forward (AF) or Decode-and-Forward (DF) technique operating over fading channels [1]–[10]. In [1] and [2], the authors investigated dual-hop DF communication models with beamforming over Nakagami-m faded channels. Ikki and Ahmed presented closed form formulations in order to analyse two-hop transmissions under Weibull and Gamma faded conditions, respectively [3], [4]. The works on two-hop transmission mentioned previously were considered over symmetric fading conditions. However, symmetric fading model is practically unrealistic since it does not include the real wireless environment characteristics. Few works have focused on the dual-hop transmission models under asymmetric fading conditions, i.e., Rayleigh/Rician, Rician/Rayleigh, Rayleigh/Nakagami-m, Rayleigh/generalized Gamma, Rayleigh/Hoyt and RF/FSO. However, there is not any work which considers deriving new formulas for the out P and ABEP of dual-hop DF relay network in Rayleigh/Weibull asymmetric faded links. The Weibull fading has a pliable distribution to describe wireless mediums, especially outdoor multipath fading environment [12]. It is also reported that the Weibull fading model On the other hand, some papers have also presented the performance of two-hop relaying systems for asymmetric channel models which mean the wireless medium conditions 60 B. Average Bit Error Probability (ABEP)   2 2 2 2 2 1 1 1 2 exp , c c c c c c c f                       This subsection gives the derivation of the ABEP,  b P e , of the considered dual-hop DF relaying network. By using the derived ( ) eq th F  expression, ABEP can be given as [6] (2) 61 ELEKTRONIKA IR ELEKTROTECHNIKA, ISSN 1392-1215, VOL. 21, NO. 1, 2015  2 2 2 0 1 . 2 eq t b t P e F e dt                (8) where N is expressed as (8) 1 2 1 1 1 . 2 N      (15) (15) In (8), by setting 1  and 2  , the performance of quadrature phase-shift keying (QPSK) and binary phase-shift keying (BPSK) can be evaluated, respectively. By substituting (7) into (8),  b P e is derived as In order to solve W1 and W2, we employed some mathematical manipulations which are not given here because of space constraints. If (10) and (12) are employed in (9), the ABEP can be expressed for the cases of 0.5 c  as  2 1 2 2 1 2 2 0 2 1 2 0 2 1 2 1 1, . 2 t b W c c t t c W P e e dt t e e dt                                      (9)      1 1 1 2 1 1 1 1 2 2 1 1 1 2 2 1 1 exp 4 2 2 4 2 . IV. NUMERICAL RESULTS (11) In realistic wireless environments, the channel conditions between cooperating nodes are subject to different fading distributions. Therefore, this work is based on such a realistic scenario which is more suitable to show the effects of real wireless environments. Performance evaluation results are presented with the help of the out P and ABEP expressions derived in the previous section for two-hop DF relaying operating in mixed Rayleigh/Weibull faded links. out P and ABEP are plotted versus average SNR per hop in Fig. 2 and Fig. 3, respectively. We are not aware of an exact solution to the integral in (11). So, we considered two different conditions as 0.5 c  and 1 c . Therefore, we have derived closed form expressions for each special cases ( 0.5 c  and 1 c ). For the case of 0.5 c  , W2 can be determined by using [14, equation (3.322.2)] as       2 2 1 exp 1 , 2 W z z           (12) (12) The outage probability is shown for different outage threshold values when 6 th   dB, 12 th   dB and 18 th   dB in Fig. 2. We focus on two values of fading parameter which are chosen as 1 c  and 2 c  . As shown in Fig. 2, outage performance is better when the outage threshold value is low. where 1 2 1 4 2            and   1 2 2 1 z        .   is the error function which is defined as [14, equation (8.250.1)] the error function which is defined as [14, equation (8.250.1)] It can be clearly seen from Fig. 2 that the performance with 2 c  is better than the case of 1 c  when the average SNR per hop is greater than 8 dB for 6 th   dB. However, similar behavior is observed when the average SNR per hop is greater than 11 dB and 12 dB for the cases of 12 th   dB and 18 th   dB, respectively. As a final remark from Fig. 2, out P improves with an increase of c or with the decrease of th  value. B. Average Bit Error Probability (ABEP) 1 1 1 2 4 2 b P e                                                                                                             (16) (9) (16) The first term (W1) in (9) can be calculated as [14, equation (3.321.3)] 2 2 1 0 1 1 . 2 2 t W e dt      (10) For the cases of 1 c , by substituting (10) and (14) into (9), the ABEP can be written as For the cases of 1 c , by substituting (10) and (14) into (9), the ABEP can be written as (10) By using [14, equation (8.352.7)], the second term (W2) of (9) can be rewritten in the way of well-known exponential function as  1/2 1 1 2 2 1 1 1 . 2 2 2 2 b P e               (17) (17) 1 2 2 1 1 2 2 2 2 0 1 . 2 c c c c t t W e e dt                           (11) REFERENCES [1] D. B. da Costa, S. Aissa “Dual-hop decode-and-forward relaying systems with relay selection and maximal-ratio schemes”, Electron. Letts., vol. 45, no. 9, pp. 460–461, 2009. [Online]. Available: http://dx.doi.org/10.1049/el.2009.0597 Fig. 2. out P of two-hop relay assisted DF transmission system over mixed Rayleigh/Weibull fading channels for different outage threshold and fading parameters. [2] D. B. da Costa, S. Aissa, “Cooperative dual-hop relaying systems with beamforming over Nakagami-m fading channels”, IEEE Trans. Wire. Commun., vol. 8, no. 8, pp. 3950–3954, 2009. [Online]. Available: http://dx.doi.org/10.1109/TWC.2009.081353 The performance results shown in Fig. 3 clearly illustrate that the ABEP of BPSK and QPSK for two-hop DF relaying improves with the larger values of fading parameter, c. It can be observed that the analytical results and the computer simulation results are in excellent concurrence for both BPSK and QPSK modulations when 0.5 c  . However, the difference between analytical and simulation results for the ABEP is very small when 1 c  with BPSK modulation. For example, the ABEP value from simulation equal 2 8 10  , 3 2 10  while the analytical ABEP values are 1 1 10  , 3 1.6 10  for 5 SNR  dB and 25 SNR  dB, respectively. [3] S. S. Ikki, M. H. Ahmed, “Performance analysis of dual hop relaying over non-identical Weibull fading channels”, in Proc. IEEE Vehicular Technology Conf. (VTC), vol. 69, Barcelona, Spain, 2009, pp. 1–5. [4] S. S. Ikki, M. H. Ahmed, “Performance analysis of dual hop relaying communications over generalized Gamma fading channels”, in Proc. IEEE Global Telecommunications Conf. (GLOBECOM 2007), Washington, USA, 2007, pp. 3888–3893. [5] W. Xu, J. Zhang, P. Zhang “Performance analysis of dual-hop amplify-and-forward relay system in mixed Nakagami-m and Rician fading channels”, Electron. Letts., vol. 46, no. 17, pp. 1231–1232, 2010. [Online]. Available: http://dx.doi.org/10.1049/el.2010.1555 [6] H. A. Suraweera, R. H. Y. Louie, Y. Li, G. K. Karagiannidis, B. Vucetic, “Two-hop amplify-and-forward transmission in mixed Rayleigh and Rician fading channels”, IEEE Commun. Lett., vol. 13, no. 4, pp. 227–229, 2009. [Online]. Available: http://dx.doi.org/10.1109/LCOMM.2009.081943 Fig. 3. ABEP of two-hop relay assisted DF transmission over mixed Rayleigh and Weibull fading channels for BPSK and QPSK modulation schemes for the cases of 0.5 c  and 1 c  . [7] H. A. Suraweera, G. K. Karagiannidis, P. J. Smith, “Performance analysis of the dual-hop asymmetric fading channel”, IEEE Trans. Wire. Commun., vol. 8, no. 6, pp. V. CONCLUSIONS In Fig. 3, ABEP performance is presented based on analytical and simulation results for validating the correctness of the derived formulas where BPSK and QPSK modulation techniques are employed. Closed-form out P and ABEP expressions have been derived based on CDF expression of two-hop DF relaying system over mixed Rayleigh/Weibull faded links. This mixed faded channel model is capable of more accurately modeling various realistic two-hop wireless transmissions. In this realistic scenario, ABEP is determined when BPSK and QPSK signaling schemes are employed. The obtained analytical expressions can be employed to appraise the system performance of realistic two-hop relaying applications. The correctness of the expressions derived in this work is verified by computer simulation results. Fig. 2. out P of two-hop relay assisted DF transmission system over mixed Rayleigh/Weibull fading channels for different outage threshold and fading parameters. ELEKTRONIKA IR ELEKTROTECHNIKA, ISSN 1392-1215, VOL. 21, NO. 1, 2015 ELEKTRONIKA IR ELEKTROTECHNIKA, ISSN 1392-1215, VOL. 21, NO. 1, 2015 IV. NUMERICAL RESULTS   2 0 2 . x u x erf x e du       (13) (13) For the case of 1 c , W2 can be obtained by using [14, equation (3.321.3)] as   1 2 2 2 , W N   (14) (14) th  v 62 REFERENCES 2783–2788, 2009. [Online]. Available: http://dx.doi.org/10.1109/TWC.2009.080420 [8] J. Ouyang, M. Lin, Y. Zhuang, “Performance analysis of cooperative relay networks over asymmetric fading channels”, Electron. Letts., vol. 48, no. 21, pp. 1370–1371, 2012. http://dx.doi.org/10.1049/ el.2012.1737 [9] N. Kapucu, M. Bilim, I. Develi, “Outage probability analysis of dual- hop decode-and-forward relaying over mixed Rayleigh and generalized Gamma fading channels”, Wireless Personal Communications, vol. 71, no. 2, pp. 947–954, 2013. [Online]. Available: http://dx.doi.org/10.1007/s11277-012-0853-6 [10] E. Lee, J. Park, D. Han, G. Yoon, “Performance analysis of the asymmetric dual-hop relay transmission with mixed RF/FSO links”, IEEE Photon. Tech. Lett., vol. 23, no. 21, pp. 1642–1644, 2011. [Online]. Available: http://dx.doi.org/10.1109/LPT.2011.2166063 Fig. 3. ABEP of two-hop relay assisted DF transmission over mixed Rayleigh and Weibull fading channels for BPSK and QPSK modulation schemes for the cases of 0.5 c  and 1 c  . [11] P. Spalevic, M. Stefanovic, S. Panic, S. Minic, Lj. Spalevic, “Amplify-and-forward relay transmission system over mixed Rayleigh and Hoyt fading channels”, Elektronika ir Elektrotechnika, no. 4, pp. 21–25, 2012. This tendency (the tightness of the proposed expressions) is valid for all values of average SNR per hop as shown in Fig. 3. Besides, there is a trade-off between the signal constellation size and performance gain. This means that the larger signal constellation size such as QPSK causes low performance gain in comparison to BPSK, as expected. As an example, the performance of the considered system is always better with BPSK compared to QPSK when the value of fading parameter is the same for both signalling techniques. [12] G. Tzeremes, C. G. Christodoulou, “Use of Weibull distribution for describing outdoor multipath fading”, in Proc. IEEE Antennas, Propagation Soc. Int. Symp., vol. 1, San Antonio, TX, 2002, pp. 232–235. [13] S. S. Ikki, S. Aissa, “Performance analysis of amplify-and-forward relaying over Weibull-fading channels with multiple antennas”, IET Commun., vol. 6, no. 2, pp. 165–171, 2012. [Online]. Available: http://dx.doi.org/10.1049/iet-com.2011.0264 [14] I. S. Gradshteyn, I. M. Ryzhik, Table of Integrals Series, and Products. San Diego, CA: Academic, 2000. 63
https://openalex.org/W1991180147	https://europepmc.org/articles/pmc3673126?pdf=render	English	null	Progress in Infrared Photodetectors Since 2000	Sensors	2,013	cc-by	24,993	Chandler Downs and Thomas E. Vandervelde * Renewable Energy and Applied Photonics Laboratories, Electrical and Computer Engineering Department, Tufts University, 161 College Ave., Medford, MA 02115, USA; E-Mail: chandler.downs@tufts.edu * Author to whom correspondence should be addressed; E-Mail: tvanderv@ece.tufts.edu; Tel.: +1-617-627-3217; Fax: +1-617-627-3220. Received: 23 January 2013; in revised form: 3 April 2013 / Accepted: 9 April 2013 / Published: 16 April 2013 Received: 23 January 2013; in revised form: 3 April 2013 / Accepted: 9 April 2013 / Published: 16 April 2013 Abstract: The first decade of the 21st-century has seen a rapid development in infrared photodetector technology. At the end of the last millennium there were two dominant IR systems, InSb- and HgCdTe-based detectors, which were well developed and available in commercial systems. While these two systems saw improvements over the last twelve years, their change has not nearly been as marked as that of the quantum-based detectors (i.e., QWIPs, QDIPs, DWELL-IPs, and SLS-based photodetectors). In this paper, we review the progress made in all of these systems over the last decade plus, compare the relative merits of the systems as they stand now, and discuss where some of the leading research groups in these fields are going to take these technologies in the years to come. Keywords: infrared photodetectors; IRPDs; bulk detectors; QWIPs; QDIPs; DWELLs; superlattices Keywords: infrared photodetectors; IRPDs; bulk detectors; QWIPs; QDIPs; DWELLs; superlattices Sensors 2013, 13, 5054-5098; doi:10.3390/s130405054 Sensors 2013, 13, 5054-5098; doi:10.3390/s130405054 sensors ISSN 1424-8220 www.mdpi.com/journal/sensors OPEN ACCESS sensors ISSN 1424-8220 www.mdpi.com/journal/sensors OPEN ACCESS Keywords: infrared photodetectors; IRPDs; bulk detectors; QWIPs; QDIPs; DWELLs; superlattices Sensors 2013, 13 Sensors 2013, 13 when in his search for materials which exhibited variable resistances depending on whether light was shined on them [3]. In this research he noted a number of materials, such as lead sulfide, exhibited responses out into the IR regime. These were the first IR detectors to operate using quantum effects rather than conductive ones, and it was this technology that fathered the field of IRPDs as we know them today [4]. Applications currently utilizing IRPDs span military (e.g., navigation, night vision, weapons detection), commercial (e.g., communications, aerospace, medical imaging), public (e.g., atmospheric sounding, pollution control, meteorology, environmental monitoring), and academic (e.g., astronomy) domains-with new uses constantly arising as the various IRPD technologies become more established [5–11]. As such, researchers have invested tremendous time and resources into developing and improving various IRPD technologies to further serve these applications. Of particular note are the advances since the new millennium. Within the past twelve years, established technologies have grown into commercial successes, nascent technologies have grown into thriving hubs of research, and new technologies have been discovered and begun to be investigated. The world around us is a large source of infrared radiation and IRPDs can be useful in a wide array of applications utilizing it. This ubiquity is due to the fact that all objects will emit an IR spectrum based on their temperature. This emission spectrum can be approximated by wavelength λ as blackbody radiation, which can be characterized according to the blackbody’s temperature T by Equation (1) [12]: ݁஻ሺߣ, ܶሻ݀ߣൌ 2ߨ݄ܿଶ݀ߣ ߣହሾ݁௛௖ఒ ⁄ ௞்െ1ሿ (1) (1) This equation illustrates why IRPDs have received so much interest of late. It implies that an object at room temperature will emit IR radiation with a peak intensity of around 9.5 µm, with detectable IR emissions for wavelengths microns away from this center value. While not every portion of the IR spectrum is ordinarily present due to natural blackbody radiation (a blackbody would need to be at a temperature of over 4,000 K to have peak light emissions at the edge of the visible spectrum), these sources do emit across large portions of the IR spectrum. One complication for IRPDs is atmospheric absorption. As IR radiation propagates through the atmosphere, some wavelengths are readily absorbed due to the molecules present in the air. 1. Introduction Infrared photodetectors (IRPDs) are a technology with wide-ranging and rapidly expanding applications in the modern world. Ever since Friedrick William Herschel discovered the presence of infrared radiation in sunlight in the early 19th century, people have tried various means to detect and analyze this spectrum of light invisible to the naked eye [1]. The earliest practical IR detectors were developed by Macedonio Melloni in the mid-19th century [2]. These detectors were thermopiles that functioned by thermal conduction, typically by relying on the differences in thermal expansion of two dissimilar metals. In 1917, Case developed what could be considered the first modern photodetector, 5055 Sensors 2013, 13 Of greatest concern to IR radiation are the absorption spectra of water vapor and various hydrocarbons [13], whose absorption spectra cut large infrared windows of the atmospheric transmission spectrum. These infrared windows (seen in Figure 1) do not necessary preclude the use of IRPDs at these wavelengths, but they can significantly interfere with device performance at those wavelengths [14,15]. The wavelengths of light associated with IRPDs are those longer than visible light (longer than 700 nm). This is a truly massive range, one no single detector technology is able to completely span. Comparisons between the performance of devices at vastly different absorption wavelengths can be difficult to make, given that performance metrics may vary by orders of magnitude across the IR spectrum. In hopes of encouraging reasonable comparisons between devices, the IR spectrum has been divided (as best as is possible) into a number of sub-regions. Where possible, these divisions have been made at convenient or recognizable points to aid in distinguishing between the regions. 5056 Sensors 2013, 13 Figure 1. A plot of the atmosphere’s absorption spectrum [16]. Note that light emitted in the 2.5–3.5 and 5–7 µm ranges are rapidly absorbed by the atmosphere. Closest to the visible spectrum is the near infrared (NIR) regime, ranging from 700 nm to about 1.0 µm, corresponding to the cutoff for absorption by silicon. From the 1.0 µm cutoff to 3 µm is the short-wave infrared (SWIR) regime, with cutoff defined by one of the atmospheric windows. This is followed by the mid-wave infrared (MWIR) regime, ranging from 3 µm to 5 µm, again with the cutoff defined by the atmospheric window. The long-wave infrared (LWIR) regime ranges from 8 µm to 12 µm, while the very long-wave infrared regime (VLWIR) amasses everything beyond 12 µm [5]. There is a range of wavelengths, from 5 µm to 8 µm, which does not fall into any of these categories. This range of wavelengths corresponds to a large region of atmospheric IR absorption. This absorption does not preclude development of IRPDs at these wavelengths provided they are detecting nearby sources of radiation in that range, but it can make the transmission of those wavelengths more difficult, decreasing the effective range of the detectors [5]. In this paper, we will attempt to constrain our device comparisons to within the same wavelength regime, even within the same technology, so as to avoid any improper comparisons. Sensors 2013, 13 where η is the quantum efficiency (the percentage of generated carriers that are extracted from the device) of the detector and g is the photoconductive gain (the number of carriers that are generated by the device structure and applied bias for every carrier generated by an absorbed photon) [17]. As with dark current density, comparisons of responsivity are valid within specific technologies under comparable conditions, but are deceptive outside of those limitations. In general, higher responsivities will result in better device performance. For the purposes of this paper, comparisons concerning responsivity will be done according to peak responsivity values reported independent of wavelength. Comparisons of responsivity values will also be limited to devices of the same technology under similar operating conditions. Another favored metric used to delineate IRPD performance is specific detectivity (D). Specific detectivity incorporates aspects of both the dark current density and responsivity of a device to provide a comparison of the amount of signal current generation for a given amount of noise at a specific wavelength, defined in Equation (3): ܦכ ൌܴ௣ඥܣ߂݂ ݅௡ (3) (3) ݅௡ ݅௡ where Rp is the peak responsivity, A is the cross-sectional area of the IRPD, Δf is the bandwidth of the device, and in is the noise current [17]. High specific detectivities indicate a larger signal current generated for a given amount noise, which allows for better signal detection. Further, comparisons of specific detectivities across different technologies and operating conditions are reasonable, making this one of the most useful analytic metrics. Closely related to the three previous metrics is the maximum operating temperature of the device. The higher the temperature an IRPD is operated at, the more noise will be present. This in turn increases dark current and reduces specific detectivity. Once the size of the dark current is large enough to make the signal current indistinguishable, the device will no longer function. Photodetectors with higher maximum operating temperatures require less cooling to maintain functionality. The need for less cooling can dramatically reduce cost of operation for these devices and allow the devices to be used in environments and situations where liquid cryogens are not available. Focal plane arrays (FPAs) necessitate a handful of additional performance metrics in addition to those used to describe single pixel detectors in order to address overall image quality. The most obvious is simply the number of pixels present in the FPA. Sensors 2013, 13 Closest to the visible spectrum is the near infrared (NIR) regime, ranging from 700 nm to about 1.0 µm, corresponding to the cutoff for absorption by silicon. From the 1.0 µm cutoff to 3 µm is the short-wave infrared (SWIR) regime, with cutoff defined by one of the atmospheric windows. This is followed by the mid-wave infrared (MWIR) regime, ranging from 3 µm to 5 µm, again with the cutoff defined by the atmospheric window. The long-wave infrared (LWIR) regime ranges from 8 µm to 12 µm, while the very long-wave infrared regime (VLWIR) amasses everything beyond 12 µm [5]. There is a range of wavelengths, from 5 µm to 8 µm, which does not fall into any of these categories. This range of wavelengths corresponds to a large region of atmospheric IR absorption. This absorption does not preclude development of IRPDs at these wavelengths provided they are detecting nearby sources of radiation in that range, but it can make the transmission of those wavelengths more difficult, decreasing the effective range of the detectors [5]. In this paper, we will attempt to constrain our device comparisons to within the same wavelength regime, even within the same technology, so as to avoid any improper comparisons. Over the years, a large number of metrics have been developed to measure the performance of IRPDs. In this review, we will focus on the most commonly reported device metrics. Peak absorption wavelength will be used to separate detectors into the different regimes listed above. For single pixel detectors, the most common metrics are dark current density, peak detectivity, peak responsivity, and operating temperature. Lower dark current densities allow lower response signals to be detected, a desired trait in all devices. Comparisons of dark currents at a given temperature within a single technology are a valid means of determining superior performance, but comparisons between IRPD technologies must be taken with a grain of salt. Different technologies (or even the same technology operating in a different wavelength regime) may produce dramatically different signal current densities operating at similar conditions. The amount of signal current generated for a given input power of IR radiation (measured for a specific wavelength) is the responsivity of the device and is defined in Equation (2): ܴൌ݁ߟ݃ ݄ߥ ܴൌ݁ߟ݃ ݄ߥ (2) ܴൌߟ݃ ݄ߥ 5057 2. Bulk Detectors As of 2000, bulk detectors were the most common IRPD technology. A wide variety of materials were utilized in various applications, but most popular were HgCdTe (MCT) and InSb. The active portions of these detectors are generally photodiodes, though there is great variety in the structure and fabrication of the specific photodiodes. There are numerous fabrication methods, each having various strengths and weakness based on the specific IRPD design (e.g., front- or back-illuminated, epitaxial or implanted layers, operation in the photoconductive or photovoltaic regime, etc.), but these are beyond the scope of this review [7,19]. Bulk detectors were uniformly the most popular detectors at the outset of the 21st century for a number of reasons, largely due to the relative maturity of the technologies. Bulk detectors, because of their relative ease of fabrication, were first developed significantly earlier than other technologies. Sensors 2013, 13 A larger pixel count enables more precise imaging, generally a desirable feature. Similarly, pixel size is important, with smaller pixels allowing for more precise imaging. However, since most FPAs are fabricated using standard read out integrated circuit (ROIC) packaging, there is more a series of benchmarks to be met in terms of pixel size and pixel count. A more useful metric for FPAs is the noise equivalent temperature difference (NETD), which can be used to approximate the minimum resolvable temperature difference of the FPA, defined in Equation (4): ܰܧܶܦൌ݅௡ ܴ ݀ܲ ݀ܶ஻ ൌ݅௡݄ߥܲ ݇஻ܶ஻ ଶ (4) (4) where P is the incident power of light on the detector, TB is the blackbody radiation temperature, and kB is the Boltzmann constant [18]. A number of device parameters, from pixel size to specific detectivity, are taken into account in this parameter. A smaller NETD is preferable, indicating a capability for finer, more precise imaging. We will additionally make note of any multicolor FPAs Sensors 2013, 13 Sensors 2013, 13 5058 which incorporate multiple different wavelengths of IR detection, particularly if the wavelengths of detection are significantly different. Such multispectral imaging has many advantages for a multitude of applications. In this review, the progress made in each of the most researched IRPD technologies since 2000 will be examined in its own section. Each section will begin with a brief description of the technology in question and highest performing devices in 2000. This will be followed with a chronicle of the major trends in the technology over the ensuing twelve years, e.g., popular material choices or inclusion of specific structures in devices. A current state of the technology will also be provided for each section. If applicable, closely related technologies will be examined in a subsection. Sensors 2013, 13 and three atmospheric windows [19,23,27–29]. However, despite the wide range of wavelengths MCT spans, there is very little variation in the value of the lattice constant across that range [23]. This allows for simplification of many of the enabling techniques used in IRPDs (substrate selection, ROIC usage, etc.) to be used across all MCT compositions. MCT devices also tend to have very short integration times and very low junction capacitances compared to other bulk detectors due their low dielectric constant, both of which enhance its capability to detect fast moving objects [4,19,30]. Figure 2. A lattice constant vs. band gap plot of common semiconductor materials. Figure 2. A lattice constant vs. band gap plot of common semiconductor materials. With these advantages come a number of disadvantages. The most noteworthy of these is that dark currents in MCT devices tend to be extremely large, much larger than in other devices, influenced heavily by band-to-band tunneling [31]. MCT devices also have growth related issues due to differences in the densities and melting points of the constituent materials, the relative softness of the final material, and a number of crystalline properties of the material (e.g., high interdiffusion coefficients, low bond strength, and a high rate of dislocation formation) [4,15,28]. These issues can make precision growths of MCT extremely difficult, which lead to a number of other device issues. One of the major issues, particularly as the band gap of the material decreases, is finely controlling the composition of the material [28,32]. As the band gap goes down towards the LWIR and VLWIR regimes, variation of the mole composition of MCT by 0.001 can have drastic effects on the performance of a detector. That level of control is extremely difficult, regardless of the growth method. Additionally, it is hard to achieve the kind of uniformity across a wafer that is necessary for large area FPAs, which has led some to question MCTs suitability for some applications [15,26,32,33]. This problem is only exacerbated when the issue of multicolor FPAs is considered [30]. Figure 2. A lattice constant vs. band gap plot of common semiconductor materials. With these advantages come a number of disadvantages. The most noteworthy of these is that dark currents in MCT devices tend to be extremely large, much larger than in other devices, influenced heavily by band-to-band tunneling [31]. Sensors 2013, 13 MCT devices also have growth related issues due to differences in the densities and melting points of the constituent materials, the relative softness of the final material, and a number of crystalline properties of the material (e.g., high interdiffusion coefficients, low bond strength, and a high rate of dislocation formation) [4,15,28]. These issues can make precision growths of MCT extremely difficult, which lead to a number of other device issues. One of the major issues, particularly as the band gap of the material decreases, is finely controlling the composition of the material [28,32]. As the band gap goes down towards the LWIR and VLWIR regimes, variation of the mole composition of MCT by 0.001 can have drastic effects on the performance of a detector. That level of control is extremely difficult, regardless of the growth method. Additionally, it is hard to achieve the kind of uniformity across a wafer that is necessary for large area FPAs, which has led some to question MCTs suitability for some applications [15,26,32,33]. This problem is only exacerbated when the issue of multicolor FPAs is considered [30]. With these advantages come a number of disadvantages. The most noteworthy of these is that dark currents in MCT devices tend to be extremely large, much larger than in other devices, influenced heavily by band-to-band tunneling [31]. MCT devices also have growth related issues due to differences in the densities and melting points of the constituent materials, the relative softness of the final material, and a number of crystalline properties of the material (e.g., high interdiffusion coefficients, low bond strength, and a high rate of dislocation formation) [4,15,28]. These issues can make precision growths of MCT extremely difficult, which lead to a number of other device issues. One of the major issues, particularly as the band gap of the material decreases, is finely controlling the composition of the material [28,32]. As the band gap goes down towards the LWIR and VLWIR regimes, variation of the mole composition of MCT by 0.001 can have drastic effects on the performance of a detector. That level of control is extremely difficult, regardless of the growth method. Additionally, it is hard to achieve the kind of uniformity across a wafer that is necessary for large area FPAs, which has led some to question MCTs suitability for some applications [15,26,32,33]. 2.1. HgCdTe (MCT) Since the development of MCT in 1959 and the creation of the first detector using it in 1962, MCT has been one of the most popular materials for IRPDs [20–22]. By 2000, the technology was fairly mature. Single pixel and focal plane array (FPA) detectors had been extensively studied in laboratory settings, with the research from those devices being used as the groundwork for a number of commercial ventures. As such, published research in this area was actually not as widespread as in other IRPD technologies over this timeframe. Much of the research performed was focused less on direct improvement in performance of devices, but instead on addressing specific difficulties encountered by the technology. The scope of this research becomes much clearer when examining the strengths and weaknesses of the MCT technology as a whole. Some of its greatest strengths have been touched on already, namely the established history of the technology and the relative simplicity of fabrication compared to a number of the later technologies. There are a number of methods utilized to grow MCT devices, including metalorganic vapor phase epitaxy (MOCVD), liquid phase epitaxy (LPE), and molecular beam epitaxy (MBE), which allows for more flexibility in fabrication [19,23–25]. MCT detectors tend to have very high responsivities and competitive detectivities when compared with quantum structures utilized in other technologies [26]. As MCT is a ternary material, by varying the composition of the material it is possible to span all of the wavelengths between the constituent binary materials, as can be seen in Figure 2. With MCT, this is a large range, spanning all wavelengths between 0.7 µm and 25 µm 5059 Sensors 2013, 13 Sensors 2013, 13 numerous papers on their progress with commercial products. In the early part of the decade, these papers generally focused on small- to mid-sized FPAs, generally operating in the NIR, SWIR, MWIR regimes. The majority of these devices were operated around liquid nitrogen temperatures (77–100 K), though there were a couple of noteworthy exceptions [19,34–37]. A few standout devices operating in the NIR and SWIR regimes, where dark currents are minimized with the corresponding larger bandgaps, were capable of functioning at elevated temperatures, up to around 250 K [24,38,39]. The fabrication method most favored for these devices was LPE, but MBE growth was growing in popularity, particularly with devices requiring large cross-sectional areas due to MBE’s finer growth control. As the decade progressed, commercial interests pushed towards larger FPAs (with the largest being a 2,560 × 512 pixel FPA using a double layer heterojunction design grown by MBE by Raytheon [40]) and longer detection wavelengths, pushing detection out to as far as 13 µm [41]. Published data also showed attempts to incorporate multicolor detection, starting with dual-color detection, into the technology with reasonably successful results [40,42]. Other trends showed an increased reliance on MBE growth for what should be considered prototype devices, with the likely intention to eventually shift to faster processes after the initial structures are completed [43–45]. Research performed in various academic and governmental laboratories has had a slightly different focus than commercial research. As early as 2000, MBE was regularly used in the growth of MCT detectors. This is not wholly unexpected, as MBE does allow for better control of material quality for the small numbers of cells that are generally grown in these settings. In the early half of the decade, the bulk of published research centered mainly on two subjects: LWIR detectors and growth of MCT on silicon substrates. LWIR detection was being researched as an extension of the progress made in SWIR and MWIR previously. These devices were usually small to mid-sized FPAs (e.g., 384 × 288) operated at liquid nitrogen temperatures and focused on using MBE growth to achieve the level of composition control necessary for precise LWIR regime absorption [28,46,47]. Research focused on silicon substrates looked to address cost concerns that could arise from typical MCT substrates (most commonly CdZnTe). Sensors 2013, 13 This problem is only exacerbated when the issue of multicolor FPAs is considered [30]. As mentioned previously, by 2000 MCT IRPDs were already an established and mature technology. Much of the research performed since then has focused on addressing some of the issues mentioned above or expanding the commercial utility of the technology, particularly with FPAs. This is partially due to the increased research performed by commercial organizations, as opposed to government or academic organizations. Companies such as Raytheon, BAE, Sofradir, and many others have published 5060 2.2. InSb InSb photodetectors have been fabricated commercially since the late 1950‘s. InSb photodetectors, like MCT photodetectors, are bulk detectors. Most detectors utilize some form of a p-n junction (including p-i-n, p+-p-n+, etc.), though many incorporate a metal-insulator-semiconductor (MIS) contact structure [4,60]. In the early part of the decade, InSb was the material of choice for NIR and some MWIR applications, in part because it was an established technology and in part because the material’s band gap falls squarely in that range. Further, the material is often combined with various other column IV or VI materials, most commonly arsenic. Attempts have been made to utilize thallium or bismuth to achieve a narrower band gap, but including small amounts of these materials in InSb (<5%) led to great difficulty in material fabrication and dramatically limited device detectivity and responsivity [61,62]. InSb FPAs were capable of pixel counts roughly similar to MCT FPAs at the start of the decade [4]. InSb detectors have several notable strengths. Its band gap (0.17 eV (~7.3 µm) at 300 K, 0.23 eV (~5.4 µm) at 80 K), allows it to function comfortably as a binary material in the MWIR regime. Its utility in the MWIR is expanded when arsenic is included in a ternary material. InAsSb is capable of spanning all wavelengths between InSb and InAs (band gap of 0.35 eV (~3.5 µm)). This flexibility allows InAsSb to be tuned to most wavelengths in the MIR regime. Further, because the difference in band gaps between InSb and InAs is much less than that between HgTe and CdTe, the tolerances in terms of material composition when designing specific operating wavelengths are eased [60]. This also limits the utility of these devices, as it can be very difficult for InSb or InAsSb devices to operate competitively outside the MWIR regime. InSb-based detectors also tend to have high device stability, lower fragility, and better device yield than MCT devices [60,63,64]. The use of InSb for photodetectors has also been associated with a number of fabrication difficulties. As can be seen in Figure 2, InAsSb spans a wide range of lattice constants. As such, while there are larger tolerances for material composition in terms of absorption wavelength, there are much stricter tolerances in terms of lattice matching to a substrate. Sensors 2013, 13 Sensors 2013, 13 as a detector material, as it is easier to grow on silicon. The performance of the detectors exhibited lower detectivities (on the order of 108 cmHz1/2/W) than crystalline MCT, but did exhibit the possibility of near-room temperature operation [32]. Experiments were also performed using resonant cavity enhancements for improved absorption of incident light at specific wavelengths. This design does have some special considerations, as one of the characteristics of MCT detectors is the wide absorption bands, whereas the resonant cavity enhancements tend to have narrow absorption bands. These detectors have shown improved absorption at specific absorber wavelengths, but reduced absorption at other wavelengths [59]. Sensors 2013, 13 Growth of MCT on silicon in the early part of the decade was performed exclusively by MBE, as the lattice mismatch between various MCT alloys and the silicon substrate can be as large as 19% [25,48–51]. Research was also performed to enable improved room temperature performance of MCT detectors. While as of 2000 room temperature operation was feasible, performance was lackluster, necessitating further research [52]. Moving through the later part of the decade, research on MCT detectors by universities and government agencies had less of a unified focus. Research on deposition of MCT onto Si substrates slowed to a near halt [53]. Research continued improving the performance of high operating temperature devices, both in terms of detectivity as well as extending the wavelengths at which this could be achieved. Attempts were made to improve the detectivity of devices (e.g., by using multiple absorbing layers) [23] as well as decrease the dark current density (e.g., including Auger suppression layers) [54]. Moreover, research continued extending the operational wavelengths for MCT devices, with functional detectors pushing into the VLWIR regime up to 14 µm [7] and D of LWIR detectors reaching up to 1.9 × 1011 cm Hz1/2/W [55]. Despite these local areas of focus, there was also a great range of research on individualized topics [56,57]. Further research was performed on multicolor detectors, focusing on moving into the MWIR and LWIR regimes utilizing multiple absorbing layers [58]. Amorphous MCT was investigated 5061 2.2. InSb At the beginning of the decade, the substrates for most of these devices were InSb, as the lattice mismatch between even InSb and GaSb was large enough to cause issues. This issue was exacerbated with more common substrates like GaAs or silicon. When attempts were made to utilize GaAs or Si substrates, the InSb/substrate interface suffered from high dislocation densities despite the use of a number of advanced techniques (e.g., inclusion of buffer layers) to ease the strain. The high dislocation densities resulted in large leakage currents, which in turn led to lower detectivities [60]. As mentioned above, while InSb and InAsSb do span a number of 5062 Sensors 2013, 13 Sensors 2013, 13 useful wavelengths, the precision gained in wavelength selectivity does come at a cost to the range of operating wavelengths. Additionally, the little response that InSb materials have in the LWIR regime is further reduced when operated at low temperatures, which is often necessary due to dark current considerations, due to reduced minority carrier lifetime in the n-type region of the material [4]. As InSb was a well-established technology as of 2000, numerous InSb photodetectors and FPAs had been built to this point. Most of the research performed in the decade was aimed at addressing specific concerns and problems that the technology faced. The two most common problems that this research addressed in the early part of the decade were increasing FPA performance and attempting to utilize different substrate materials in the devices. Universities typically focused on smaller FPAs with other performance improving features (e.g., an InAlSb current blocking layer in the active region of the detector) while companies, most notably Raytheon, focused on expanding pixel counts in FPAs beyond the megapixel range, out as far as 2K × 2K pixels. Whereas many single pixel devices were capable of operating at near room temperature, the FPAs were typically operated at liquid nitrogen (LN2) temperatures or lower to achieve the levels of performance desired [65–70]. Despite the difficulties growers faced when using substrates other than InSb, research continued on InSb IRPDs fabricated on these substrates. The most popular substrates (for reasons of cost, availability, and integration with other technologies) were silicon (a = 5.431), which has a lattice mismatch with InSb of 19%, and GaAs (a = 5.653), which has a lattice mismatch of 15%. 2.2. InSb Growth on these substrates was accomplished using a variety of means, most commonly MBE growth with the incorporation of buffer layers to reduce the strain at each interface. A number of devices were fabricated on both silicon and GaAs substrates, but these devices uniformly had higher dark current densities, lower responsivities, lower detectivities, and lower maximum operating temperatures than those fabricated on InSb substrates [60,61,63,71]. Other attempts were made at utilizing GaSb substrates, which does have the benefit of being much closer to a lattice match to InSb, with a lattice constant of 6.096 Å and a mismatch of 6%. The performance of these devices were generally more favorable than those on silicon or GaAs substrates, but operating temperatures were limited to 250 K and detectivities were still reduced [62]. As other technologies rose to prominence in the middle of the decade, research on InSb, particularly at universities, slowed. The Kuze group, one of the few groups that still performed research on these devices, focused on miniaturization of InSb sensors. These devices were constructed from a number of photodiodes connected in series with the aim of weak IR detection at room temperature. The final operational area of the devices was 600 × 600 µm, with final packaging 2.2 × 2.7 × 2.7 mm [72–75]. Other research looked into incorporating nitrogen into InSb. It was found that incorporating small amounts of nitrogen (<1%) led to improved D* at wavelengths as far out as 10 µm. While its incorporation did lead to a decrease in performance at shorter wavelengths, this device had among the highest performance in the LWIR of any InSb device [76]. Research into InAsSb devices has been fairly modest to this point. Most of the research has been focused on achieving the quality of growth required for devices, typically by limiting the mole fraction of As in the devices and including buffer layers. A handful of MWIR and LWIR IRPDs have been grown, achieving D* as high as 4.03 × 1011 cm Hz1/2/W at 170 K at a wavelength of 3.7 µm [76–78]. Sensors 2013, 13 Sensors 2013, 13 5063 3. Quantum Well Infrared Photodetectors Quantum well infrared photodetectors (QWIPs) operate on a much different principle than the bulk detectors that came before it. First demonstrated in 1987 by Levine et al. [79], QWIPs rely on quantum scale physical effects, whereas bulk detectors operate on larger scale effects. QWIPs in their most basic form are a periodic repetition of layers of two materials with dissimilar band gaps. The material with the lower energy band gap is commonly referred to as the well layer, while the higher energy band gap is referred to as the barrier layer. This structure approximates a series of quantum mechanical wells, with the well layer deposited thinly enough (generally on the order of nanometers to tens of nanometers) such that discretized energy states form within the well instead of the typical band of energy states. An example quantum well structure is provided in Figure 3. The energy states within the well will be characterized by Equation (5): ܧ௡ൌħଶߨଶ݊ଶ 2݉כܮ௪ ଶ (5) ܧ௡ൌħଶߨଶ݊ଶ 2݉כܮ௪ ଶ (5) where m* is the effective mass of the charge carrier, Lw is the width of the well layer, and n is an integer identifying the energy level in the well [15]. The well layers are doped such that without illumination there will be carrier electrons present in the ground energy state. These carriers are then excited by incident photons to an energy state near the conduction band edge of the barrier material, where the applied voltage will sweep the carrier out of the well to the contacts. When the higher energy state is above the barrier conduction band edge (>10 meV), there is a higher escape probability from the well. This generally results in higher detectivities, and is referred to as a bound to continuum transition. Figure 3. A band diagram of a quantum well in a QWIP. Carriers are excited by incident photons from the ground energy state of the well (E1) to the excited energy state of the well (E2). When the higher energy state is below the barrier conduction band edge (<10 meV) there is higher absorption efficiency; this is a bound to bound transition. Between these two values is a bound to quasibound transition and mixes a balance of the above traits. Sensors 2013, 13 Sensors 2013, 13 5064 QWIPs have a number of positive and negative performance characteristics that distinguish them from other IRPD technologies. In comparison to MCT devices, QWIPs are capable of lower dark currents, higher detectivities and higher NETDs. They make use of commonly used III-V material processing techniques, making fabrication of the devices easier than with MCTs. QWIPs also typically have greater radiation tolerances than narrower band gap materials like MCT and InSb. A case can be made that QWIPs have the greatest ease of material growth of multispectral imagers, as neither lattice mismatching (due to small layer thicknesses) or surface irregularities (due to a lack of nanostructures) are common hindrances for these devices [30,81–83]. QWIPs also face a number of challenges. The most distinct disadvantage arises from QWIPs’ use of quantum confinement. Due to constraints defined by quantum mechanics, a QWIP can only absorb light that is incident upon it when there is quantum confinement in one of the perpendicular axes. This leads to a large problem: due to common growth techniques, most QWIPs are incapable of the absorption of normally incident light. In order to mitigate this problem, most QWIP devices use some form of a frontside scattering filter, which redirects normally incident light such that it can be absorbed [84]. It should also be noted that the maximum energy photon that a QWIP can absorb is limited by the energy difference between the conduction band edges of the materials. This means that for operation at shorter wavelengths, there must be a very large difference between the band edges. Finding materials that can satisfy this condition can be extremely difficult. For this reason, QWIPs have been used sparingly in the SWIR regime [85–87]. The scattering filters required for the absorption of normally incident light must be on a similar scale as the wavelength of the incident light. This complicates SWIR QWIPs by increasing the difficulty of fabrication as wavelength decreases. Those devices that attempt to operate at higher frequencies typically incorporate various nitride materials. There are also similar issues when the absorbed wavelength gets very long, into the VLWIR range [10,80]. QWIPs have significantly narrower absorption widths due to the discretized nature of the energy states transitions occur between. This can be advantageous if detection of a specific wavelength is required, but is a hindrance for wide band detection [81]. 3. Quantum Well Infrared Photodetectors QWIP devices typically utilize an absorbing superlattice stack sandwiched between emitter and absorber layers, which aid in transferring carriers to and from the contacts and can be doped to assist with carrier separation. Due to the precise tolerances required for the layer thicknesses in the matrix of materials, these devices are typically grown using MBE, though other methods have been attempted with varying success [5,6,80]. When the higher energy state is below the barrier conduction band edge (<10 meV) there is higher absorption efficiency; this is a bound to bound transition. Between these two values is a bound to quasibound transition and mixes a balance of the above traits. QWIP devices typically utilize an absorbing superlattice stack sandwiched between emitter and absorber layers, which aid in transferring carriers to and from the contacts and can be doped to assist with carrier separation. Due to the precise tolerances required for the layer thicknesses in the matrix of materials, these devices are typically grown using MBE, though other methods have been attempted with varying success [5,6,80]. Sensors 2013, 13 5065 One of the largest debates in the early portion of the 2000s was which material system held the most promise moving forward. Under consideration primarily were three material compositions (utilizing various mole fractions of the different ternary materials): (Al)GaAs/InGaAs, (In)GaAs/(Ga)InP, and AlGaAs/GaAs. Of primary interest during the earliest portion of the 2000s was the GaAs/InGaAs regime. It was hoped that this material system could help extend the functionality of QWIP devices out to the LWIR and VLWIR regimes (λc > 10 µm). Performance of these devices was generally not promising in these wavelength ranges, with required operating temperature often below LN2 levels (77 K). Even with these lower than normal operating temperatures, device performance was still lackluster, with responsivities regularly below 0.5 A/W and detectivities on the order of 109 cmHz1/2/W. Performance of these devices was marginally improved with the incorporation of aluminum into the GaAs layer, with operating temperatures in ranges of LN2 cooling. Responsivities and detectivities were also higher after incorporation of aluminum, particularly at lower operating temperatures, peaking at around 1.96 A/W and 1.59 × 1010 cmHz1/2/W, respectively, by Lee et al. [92]. However, when quantum dot infrared photodetectors (QDIPs) began receiving attention as a viable alternative to QWIPs, particularly at these longer wavelengths, research interest in the GaAs/InGaAs material system petered out [93–98]. The next material system under serious consideration was InGaAs/(Ga)InP. This material regime also received the most attention at the beginning of the decade, with interest fading with the introduction of QDIPs. The material system was investigated across a wider range of wavelengths than the (Al)GaAs/InGaAs system, ranging from MWIR to VLWIR. InGaAs/(Ga)InP QWIPs met with similar difficulties in the LWIR and VLWIR regimes, with detectivities on the order of 1 × 109 cmHz1/2/W around LN2 temperatures. These devices did exhibit stronger performance in the MWIR regime, feature operating temperatures over 100 K, responsivities as high as 2.2 A/W, and detectivities as high as 5 × 1010 cmHz1/2 [83,99–102]. There were also attempts made to adapt this material system for deposition on Si substrates, however performance of these devices is generally poorer than the normal performance of these devices, usually by an order of magnitude and sometimes more. This poor performance was attributed to the high density of defects and dislocations at the interface between the silicon substrate and the active QWIP stack [103,104]. Sensors 2013, 13 The narrow absorption bands also reduce photogenerated current, and hence the detectivity of QWIPs tends to be much lower than in bulk detectors. Lower detectivities in turn make QWIPs more sensitive to the magnitude of dark current density. As such, QWIPs have had significantly greater difficulty achieving high temperature operation [5,6,88,89]. Compounding the dark current issue is a short carrier lifetime in these devices, generally on the order of a handful of picoseconds. While a low carrier lifetime can have the added benefit of fast operating speeds, it also leads to increased sensitivity to dark current [90,91]. There are also a small number of issues that arise when considering QWIP FPAs. Due to the lower detectivities and responsivities of these devices, integration times tend to be longer when compared with bulk detectors (up to the order of milliseconds). This can make detection and identification of quickly moving objects more difficult [30]. Similar to bulk technologies, QWIPs had been the subject of research for some time as of 2000. Functional devices were common place and FPAs had recently begun to be developed. The vast majority of research conducted over the ensuing decade was performed by research labs at various universities. This section will focus first on single pixel QWIP designs, followed by various FPA designs, and finally with notes on a handful of technologies closely related to QWIPs, but which do not easily fit within the normal framework of QWIP devices. Sensors 2013, 13 5066 Sensors 2013, 13 that reaches the contacts. Such a layer can be implemented either within the active region or can be added between the complete material matrix and the other layers of the device. These devices generally exhibit improved dark current performance, but was not a topic of much interest. A number of the devices, while called QWIPs, can be seen as precursors to dots-in-a-well detectors (DWELLs), discussed later in this paper, particularly devices which incorporated the current blocking layer or layers into the material matrix. DWELLs have become increasingly interesting subjects of research in recent years, and seeing how they may have evolved from QWIPs is a fascinating occurrence [114–116]. In addition to single pixel QWIPs, a great amount of research was also performed on FPAs. Some of the research was focused on the InGaAs/InP material system. These devices, while showing some improvement over their single pixel predecessors (which can be attributed to improvement in MBE technology over the decade), still exhibited some of the same issues, namely dark current as much as an order of magnitude larger than AlGaAs/(In)GaAs QWIPs [117,118]. For this reason and the general superior performance over other material choices, most of the QWIPs FPAs researched utilized AlGaAs/(In)GaAs. At the beginning of the decade, the most common size for these FPAs was 640 × 512 pixels. These devices were generally operated in the MWIR and LWIR regimes, and exhibited similar performance to single pixel devices. They were generally operated at LN2 temperatures with the strongest results produced by Gunapala et al., with responsivities between 0.5–1 A/W and detectivities upwards of 2 × 1011 cmHz1/2/W. Spectral width, where reported, was generally between 10%–15% and NETDs between 30–50 mK [91,119–123]. As the decade progressed, efforts were made to increase the amount of pixels in the FPAs and their operating temperature. FPA sizes increased to over a megapixel in size, generally 1,024 × 1,024. While accomplishing this, operating temperatures were also raised, but the amount of improvement was fairly small, with the highest performance at just 105 K. NETD across these devices also improved through the decade, reaching below 25 mK for the highest performing devices [124–128]. Sensors 2013, 13 By far the most commonly used material system over the decade was AlGaAs/GaAs. Whereas the other material systems often tried to extend operation out to the VLWIR regime, research with AlGaAs/GaAs typically focused on the MWIR and LWIR regimes. Dual band absorption was incorporated into a number of the structures, with performance of these devices often similar to single color devices. Performance of AlGaAs/GaAs devices outstripped that of the (Al)GaAs/InGaAs and (In)GaAs/(Ga)InP in the early portion on the decade, encouraging further research in the later portion of the decade [102,105–109]. There has been an effort to try and extend the operational wavelengths of AlGaAs/GaAs towards the VLWIR regime, but these efforts only met very limited success. Some reasonable responsivity values were achieved, but only at operating temperatures of 20 K. Performance at even LN2 temperature degrades rapidly [110–112]. Some have suggested the GaSb/AlGaSb material regime for use in the VLWIR wavelength range, but it remains sparsely investigated to this point [113]. Perhaps the most interesting results in the AlGaAs/GaAs material systems are those devices which incorporated some form of current blocking layer into the QWIP. A number of QWIP devices included such a layer, which consist of a significantly higher energy band gap intended to reduce the dark current Sensors 2013, 13 performance of the device. The band diagram of a QCD can be seen in Figure 4. This further reduces the detectivity and responsivity of the device, but with the benefit of also dramatically reducing the dark current generated [5,129]. Figure 4. A band diagram of a quantum cascade detector. Incident light excites a carrier from the ground energy state of the absorbing well to an excited energy state. The carrier then tunnels through the barrier layers into the adjacent wells. This continues until the carrier reaches the ground state of the next period of the superlattice. 4. Strained-Layer Superlattices 4. Strained-Layer Superlattices Since the first strained-layer superlattice (SLS) photodetector was grown from InGaAs/GaAs in 1984, superlattice (SL) and SLS structures have been of interest to the IRPD community [130]. Note: for brevity, in this publication, SL and SLS photodetectors will both be referred to as SLS, whether they are strained or not. SLS, while initially appearing to have a similar structure to QWIPs, actually operate using dramatically different physical principles. The misleading similarities result from a superlattice with extremely thin layer thicknesses (on the order of single nanometers) as an active absorbing layer. Additionally, while QWIPs generally utilize relatively thick barrier layers (on the order of tens of nanometers), in SLS the active layers have thicknesses on the same order of magnitude [131,132]. Like a QWIP, the most common structure for the superlattice is alternating layers of two different materials, but some more recent devices utilize more complex structures [30,133]. 4. Strained-Layer Superlattices 3.1. Quantum Cascade Detectors One interesting technology closely related to QWIPs is the quantum cascade detector (QCD). Like a QWIP, a QCD is a matrix of two alternating material layers: a well material and a barrier material. Whereas a period of a QWIP contains only two layers, one each of the well and barrier, each period of a QCD contains many layers. The first two layers are much like a period of a QWIP, with a long barrier layer followed by a collecting well layer. Much as in a QWIP, this collection well is doped such that electrons are present in the ground state of the well without illumination and upon illumination move to an excited energy state in the well, however unlike QWIPs this excited energy state is not near the conduction band edge of the barrier material. After these two layers are an additional series of thin layers, alternating between barrier and well materials. The thicknesses of these layers are chosen so as each successive well layer has an incrementally lower energy state than the excited energy state in the well. The carrier is able to tunnel through the adjacent barrier layer into the adjacent incrementally lower energy state. This continues through the rest of the period of the QCD until the carrier finally relaxes to the ground state of the absorbing well in the next period of the detector. The requirement for a number of excitations to produce a single carrier in the conduction band has a dramatic effect on the 5067 Sensors 2013, 13 Sensors 2013, 13 5068 Sensors 2013, 13 material (as in bulk detectors) or the spacing of energy levels within a single well (as in QWIPs and QDIPs). The devices are then biased to extract the charge carriers [133,135]. Figure 5. A band diagram of a InAs(blue)/GaSb(red) SLS. Carrier electrons are excited from the valence band of the GaSb layers to the conduction band of the InAs layers. y y SLS have a number of performance strengths and weaknesses. SLS devices can be grown using many of the same techniques used for other III-V devices, but MBE growth is generally favored for its precise control of layer thicknesses and interfaces. Control of layer thickness, and particularly material interfaces, is vital for high performing devices. Poor quality interfaces can lead to increased band-to-band and defect-assisted tunneling currents. As SLS depends heavily on the spatial separation of charge for its performance, these tunneling currents are especially detrimental [136]. Materials chosen for use in SLS generally fall around the 6.1 Å lattice constant (InAs/GaSb/AlSb/etc.) with the most common superlattices comprised of InAs (electron capture) and GaSb (hole capture) layers [30,135]. Gallium-free versions also exist, using the same InAs layer for electron capture, but using an InAsSb for the hole capture. Use of InAsSb layers in place of GaSb would allow for longer wavelength absorption, but to this point research has been limited to characterization of superlattices comprised of these materials as opposed to fabrication of functioning photodetectors [137–139]. Due to the structure of SLS, dark currents tend to be fairly low for these devices. This in turn allows for higher operating temperatures compared to bulk detectors and QWIPs, particularly at longer absorption wavelengths. These detectors also allow normal incidence absorption, eschewing the need for frontside scattering filters. There have also been studies concluding higher values for the effective electron mass in these devices, which can also lead to reduced Auger recombination rates [133,140–142]. SLS devices also tend to be more tolerant to growth variances than MCT, particularly in terms of mole fraction composition [143]. Unfortunately, due to the overlap of the band structure of different materials in the devices, the effective band gap between the minibands tends to be fairly narrow. This tends to make it more diffi lt t th d i f SWIR d h t l th li ti It l k SLS ll SLS have a number of performance strengths and weaknesses. 4. Strained-Layer Superlattices Since the first strained-layer superlattice (SLS) photodetector was grown from InGaAs/GaAs in 1984, superlattice (SL) and SLS structures have been of interest to the IRPD community [130]. Note: for brevity, in this publication, SL and SLS photodetectors will both be referred to as SLS, whether they are strained or not. SLS, while initially appearing to have a similar structure to QWIPs, actually operate using dramatically different physical principles. The misleading similarities result from a superlattice with extremely thin layer thicknesses (on the order of single nanometers) as an active absorbing layer. Additionally, while QWIPs generally utilize relatively thick barrier layers (on the order of tens of nanometers), in SLS the active layers have thicknesses on the same order of magnitude [131,132]. Like a QWIP, the most common structure for the superlattice is alternating layers of two different materials, but some more recent devices utilize more complex structures [30,133]. Similarities between the two technologies begin and end at the inclusion of a superlattice. QWIPs utilize a type-I band structure with explicit barrier and well materials, where the band edges of the smaller energy band gap are straddled by the band edges of the larger energy band gap. This structure allows QWIPs to utilize intraband transitions within the smaller band energy material in both the valence and conduction bands. SLS devices typically utilize a type-II band structure, where the band edges of the superlattice are staggered [134]. Due to this staggering of the band edges, each material acts as a well in one band (valence or conduction) and as a barrier for adjacent layers of the other material. An example SLS band structure can be seen in Figure 5. The thinness of the SLS layers allows for tunneling between wells in each band. Generally the barriers are thin enough that the available energy states in the wells of each band are able to effectively blur together despite being separated spatially in the different material layers of the superlattice, leading to what are commonly referred to as minibands or quasi-bands. It is the interband energy gap transitions between these minibands that determines the absorption wavelength of the device, rather than the band gap of either Sensors 2013, 13 5069 devices are conductive. In FPAs, the sides of mesas can accumulate these oxides, which can enhance shunt current. Should adjacent pixels come into contact with each other due to oxide growth, those pixels will not perform as intended [133]. There can also be difficulty in achieving high quantum efficiency levels, as the thickness of the active regions can be small compared to other technologies, ranging from as small as hundreds of nanometers to as thick as a handful of microns [30,147]. In the early portion of the decade the majority of research in SLS was focused on InAs/GaSb devices grown on GaSb substrates. This material system had already been shown to function in previous work and various research groups sought to refine their utility. These devices were focused primarily on two miniband gaps: 150 meV (λ = 8.3 µm) and 80 meV (λ = 15.5 µm). These miniband gaps were chosen in part to fall squarely in the LWIR and VLWIR regimes, respectively, and also in part due to straightforward deposition ratios (42Å/42Å for the former, 20 ML (monolayers)/20 ML for the latter) [148,149]. Like QWIPs, some of these devices chose to incorporate current blocking layers into the superlattice. These devices showed promising performance in both regimes, capable of LN2 temperature operation even at this early stage. At these temperatures, the LWIR SLS with the best performance achieved detectivity on the order of 1 × 1011 cmHz1/2/W and responsivity of 1.2 A/W by Razeghi et al. Similarly, the best VLWIR device was capable of a D* and responsivity of 4.5 × 1010 and 4 A/W, respectively [148–153]. Moving forward, FPAs were created using SLS technologies. These FPAs began with small pixel counts, 256 × 256, and at this point were only utilized in the LWIR regime, but those produced were capable of similar performance metrics as the single pixel devices. As processing steps were refined, the imaging of these devices improved and an NETD of about 10 mK was achieved at LN2 temperatures [45,154,155]. Research on single pixel devices over this time frame focused on expanding the operation wavelengths SLS in both directions. SLS operational down to 5 µm were fabricated, maintaining the strong performance of previous tests with detectivities as high as 1.5 × 1013 cmHz1/2/W and responsivities remaining at 1 A/W by Krishna et al. [156–158]. Sensors 2013, 13 SLS devices can be grown using many of the same techniques used for other III-V devices, but MBE growth is generally favored for its precise control of layer thicknesses and interfaces. Control of layer thickness, and particularly material interfaces, is vital for high performing devices. Poor quality interfaces can lead to increased band-to-band and defect-assisted tunneling currents. As SLS depends heavily on the spatial separation of charge for its performance, these tunneling currents are especially detrimental [136]. Materials chosen for use in SLS generally fall around the 6.1 Å lattice constant (InAs/GaSb/AlSb/etc.) with the most common superlattices comprised of InAs (electron capture) and GaSb (hole capture) layers [30,135]. Gallium-free versions also exist, using the same InAs layer for electron capture, but using an InAsSb for the hole capture. Use of InAsSb layers in place of GaSb would allow for longer wavelength absorption, but to this point research has been limited to characterization of superlattices comprised of these materials as opposed to fabrication of functioning photodetectors [137–139]. Due to the structure of SLS, dark currents tend to be fairly low for these devices. This in turn allows for higher operating temperatures compared to bulk detectors and QWIPs, particularly at longer absorption wavelengths. These detectors also allow normal incidence absorption, eschewing the need for frontside scattering filters. There have also been studies concluding higher values for the effective electron mass in these devices, which can also lead to reduced Auger recombination rates [133,140–142]. SLS devices also tend to be more tolerant to growth variances than MCT, particularly in terms of mole fraction composition [143]. Unfortunately, due to the overlap of the band structure of different materials in the devices, the effective band gap between the minibands tends to be fairly narrow. This tends to make it more difficult to use these devices for SWIR and shorter wavelength applications. It also makes SLS well suited for applications in the LWIR regime and longer wavelengths [133,144,145]. Due to high trap assisted tunneling rates in the active region of the device, there can be high amounts of non-thermal noise in these devices [146]. Additionally, the native oxides for many of the materials used in these Sensors 2013, 13 fabricated out to wavelengths of 4.3 µm. Other research focused on reducing the thickness of a single SLS period to 10 ML, which may enable this class of device to achieve noteworthy SWIR response, but this device resulted in lower detectivities at room temperature compared to other devices [136,145,162,164,166–168]. There were also some advancements in performance in recent years, with some single pixel devices with detectivities as high as 1.9 × 1013 cmHz1/2/W and responsivities as high as 1.33 A/W [169]. There was also some research performed on other material systems over this timeframe. The most common alternative to the InAs/GaSb system was InAs/InGaSb. Incorporation of indium into the GaSb layer allowed for increased flexibility in defining the energy level of the valence miniband. Performance of InAs/InGaSb devices was usually similar to devices utilizing InAs/GaSb structures. However, GaSb was usually the preferred due to more convenient material growth [143,170–172]. Sensors 2013, 13 It was also over this time frame that the first room temperature operation of an SLS IRPD, with a D* of 4.6 × 109 cmHz1/2/W and responsivity of 2.2 A/W at 5.25 µm [144]. Also during this time frame the mechanics of this material system was researched, revealing information on its optical properties. It was discovered by Brown, Haugan, et al. that variation of the InAs layer thicknesses has a stronger effect on the band gap than variation of the GaSb layer thicknesses, and the dependence of the amount of interfacial strain on the number of periods in the superlattice [147,159,160]. Moving towards the end of the decade, research began to focus on production of larger and higher quality FPAs. The majority of these devices had cutoff wavelengths in the LWIR regime and were operated at LN2 temperatures. While none of the devices are capable of room temperature operation, a handful were capable of operation between 200 and 240 K. The expansion of FPA size, and perhaps the inclusion of other features such as current blocking layers into these devices, led to a small decrease in NETD, usually in the 20–30 mK range. These figures may seem low, however they are still more than sufficient for the purposes of an FPA [161–164]. By the end of the decade, a megapixel FPA had been fabricated, and its responsivities and detectivities remained similar to those found in previous FPAs. However, this FPA did suffer somewhat due to its size in the form of a higher NETD (over 50 mK) [165]. In addition to this FPA research, extension of previous work was performed, including optimization of growth conditions and further extending operational wavelengths. SLS were 5070 Sensors 2013, 13 Sensors 2013, 13 Even by the end of the decade, W- and M- superlattices were still a nascent technology. As such, much of the research performed is in the very early stages. Most single pixel devices have only been operated at LN2 temperatures. Only at the very end of the decade was an M-superlattice operated as high as 150 K. The results from that device, led by Razeghi et al., with detectivities as high as 1.05 × 1012 cmHz1/2/W at 150 K in the MWIR regime, do show promise for this technology. Further, a 320 × 256 pixel FPA was constructed using this design, with imaging possible up to 170 K [131,174,176–178]. 4.1. M/W Strained-Layer Superlattices A pair of sub-technologies of SLS spawned due to the implementation of current blocking layers. Each of these technologies incorporates a different manner of a current blocking layer. The first, and arguably more complex, manner was first proposed in 2001, but not implemented until 2006 [173,174]. This technology incorporates a current blocking layer into one of the active superlattice layers, forming an ABCB superlattice as opposed to the traditional AB superlattice. This barrier layer typically has a significantly larger energy band gap than the surrounding material, leading to the distinctive band structure that gives these devices their name. An example band diagram can be seen in Figure 6. Figure 6. A band diagram of a InAs(blue)/GaSb(red)/AlSb(green) M-SLS. The inclusion of AlSb into the device gives this structure its eponymous band structure. The current blocking barrier, typically an aluminum bearing material, can be incorporated into either the InAs or GaSb layer, giving rise to the eponymous band structure in a period of the superlattice, resembling either a W or an M. The barrier placement in W- and M-superlattices is crucial in these devices. Inclusion of the barrier in the well shifts the two highest energy levels from the original well towards each other. When the barrier is precisely in the middle of the well the energy states become nearly degenerate. This in turn increases the number of energy states available at the desired energy level [164,175]. Figure 6. A band diagram of a InAs(blue)/GaSb(red)/AlSb(green) M-SLS. The inclusion of AlSb into the device gives this structure its eponymous band structure. The current blocking barrier, typically an aluminum bearing material, can be incorporated into either the InAs or GaSb layer, giving rise to the eponymous band structure in a period of the superlattice, resembling either a W or an M. The barrier placement in W- and M-superlattices is crucial in these devices. Inclusion of the barrier in the well shifts the two highest energy levels from the original well towards each other. When the barrier is precisely in the middle of the well the energy states become nearly degenerate. This in turn increases the number of energy states available at the desired energy level [164,175]. 5071 4.2. Unipolar/Monovalent Barrier Strained-Layer Superlattices The other main technology incorporating current blocking layers in SLS are unipolar or monovalent barrier devices. In these devices, a single barrier layer is incorporated in the SLS structure immediately above or below (depending on whether the barrier is in the conduction or valence band) the active superlattice stack. The barrier material is chosen such that there is a relatively large barrier in the chosen band, but as little discontinuity as possible in the other band. This barrier allows the device to conduct photocurrent and diffusion current, but block both generation-recombination current and trap-assisted tunneling current [178–182]. Because of unavoidable variabilities in the growth process, most unipolar/monovalent SLS are operated at a slightly higher bias voltage to overcome the unintentional barrier than other devices, in case there is a discontinuity at the band edge between the barrier layer and the rest of the device [131]. Figure 7. A band diagram of an nBn IRPD without applied bias. With the barrier in the middle, the left hand side acts as the photo-absorber. Absorbed photons create carrier pairs, which are unimpeded by the barrier. Thermally generated pairs on the right-hand side are blocked by the conduction band barrier, however. As with M- and W-superlattices, the first devices incorporating unipolar barriers did not arrive until the later part of the decade, 2006 [174]. With the explicit concern of limiting dark current, this class of devices quickly achieved room temperature operation. Monovalent barrier devices have been operated at a range of wavelengths from the MWIR to LWIR. One of the highest performing single pixel devices is capable of a D* of 1 × 109 for a cutoff wavelength of 4.3 µm at room temperature. Barrier layers were investigated for both the conduction band (nBn structures) and the valence band (pBp structures), but nBn (Figure 7) were far more popular in the early stages of research [146,161,183–185]. A small FPA (320 × 256 pixels) was fabricated using nBn technology, As with M- and W-superlattices, the first devices incorporating unipolar barriers did not arrive until the later part of the decade, 2006 [174]. With the explicit concern of limiting dark current, this class of devices quickly achieved room temperature operation. Monovalent barrier devices have been operated at a range of wavelengths from the MWIR to LWIR. Sensors 2013, 13 Sensors 2013, 13 having comparable performance to single pixel devices in the MWIR regime. Responsivities of 1.5 A/W, detectivities of 6.4 × 1011 cmHz1/2/W, and NETD of 24 mK were also achieved with nBn devices [163]. Recently, pBn have drawn interest for use in photodetectors, but very few devices have been successfully fabricated. To this point, pBn devices (Figure 8) have remained an exercise for theory and simulation, but promising results from early devices, included low dark current densities as low as 3 × 107 A/cm2 at 150 K could lead to increased research with these devices [186–189]. Figure 8. A band diagram of a pBn photodiode. A carrier electron is excited from the valence band of the n-type material and swept away from the barrier. The carrier hole is swept past the barrier layer, as it does not produce a potential barrier in the valence band of the p-n junction. 5. Quantum Dot Infrared Photodetectors (QDIPs) 5. Quantum Dot Infrared Photodetectors (QDIPs) 4.2. Unipolar/Monovalent Barrier Strained-Layer Superlattices One of the highest performing single pixel devices is capable of a D* of 1 × 109 for a cutoff wavelength of 4.3 µm at room temperature. Barrier layers were investigated for both the conduction band (nBn structures) and the valence band (pBp structures), but nBn (Figure 7) were far more popular in the early stages of research [146,161,183–185]. A small FPA (320 × 256 pixels) was fabricated using nBn technology, 5072 5. Quantum Dot Infrared Photodetectors (QDIPs) Quantum dots were initially posited in 1982 as a structure for use in quantum well lasers as a means to reduce or remove the temperature dependence from a laser’s performance [190]. They remained a theoretical exercise for over ten years until 1993, when the first demonstrated growth of quantum dots was performed by MBE [191]. By 2000, research was well underway for quantum dot infrared photodetectors (QDIPs). In a QDIP, operation is very similar to that in a QWIP. A layer of dots, typically with diameters on the order of nanometers, is deposited on a well material, most commonly by MBE. These dots typically have narrower band gaps than the barrier material, creating a highly localized quantum well. A number of iterations are then deposited to improve the absorption of these devices [82,84,192–196]. Dots are typically fabricated using the Stranski-Krastanow growth mode, which requires a certain amount of lattice mismatch (between 1%–10%) between the dot material and the well material to ensure sufficient strain for islanding to occur. The thickness of the dot layer must also be precisely monitored, as excessive thickness (>2–3 ML for some systems) in the layer will change the surface morphology from quantum dots to a solid layer with large densities of dislocations [197]. Like QWIPs, QDIPs are doped such that there are electrons present in the conduction bands of the quantum dots. These electrons are then excited by incident photons to the conduction band of the well layers, where they can diffuse to the contacts of the device [9,15]. While much of the operation of QDIPs is reminiscent of that of QWIPs, there are a number of features which distinguish the former from the latter. The most notable difference is that, due to the quantum 5073 Sensors 2013, 13 dot’s confinement in all directions, QDIPs are capable of absorbing normally incident photons. The higher confinement also leads to lower dark currents, which can in turn lead to higher detectivities and operating temperatures [26,80,84,88,192,198,199]. QDIPs have longer carrier lifetimes than QWIPs, generally on the order of hundreds of picoseconds. However, the carrier lifetimes are shorter than in bulk detectors. This is believed to derive both from decreased amounts of phononic scattering in QDIPs and a phenomenon commonly referred to as phononic bottleneck. In this bottleneck, carriers fill the excited states of the dots. 5. Quantum Dot Infrared Photodetectors (QDIPs) If these carriers are not extracted from the dots quickly, other carriers which would otherwise be excited to the higher energy level will not undergo that transition due to the lack of an available energy state [84,89,192,200]. Under ideal growth conditions, QDIPs also have higher selectivity and narrower spectral width absorption than QWIPs [8,192]. QDIPs do come with some disadvantages. While under ideal growth conditions QDIPs should have superior selectivity to QWIPs, in practice it is extremely difficult to have the kind of quantum dot size uniformity required for high levels of performance. Due to uniformity variance in the quantum dots, oftentimes QDIPs have wider spectral absorption than QWIPs [84,193]. For similar reasons, tuning QDIPs to a specific wavelength with great accuracy can be extremely difficult [15,201]. The dot layers in a QDIP are by their nature very uneven, leading to issues with strain and dislocations with an increasing number of absorption layers. As such, QDIPs often exhibit fewer absorbing layer iterations than QWIPs, leading to lower absorbances. To combat this effect, some QDIPs have strain reducing layers incorporated into their design. While this raises the complexity of the fabrication, it can lead to higher device performance. Quantum efficiencies (QEs) for QDIPs also tend to be much lower than QWIPs. This low QE is due to the relative paucity of energy states in QDIPs (a few monolayers of scattered dots in QDIPs vs. several nanometers of a solid layer in QWIPs), lower absorption of incident light, and lower capture probabilities by individual dots due to more sharply defined energy states than in QWIPs [84,193,202]. The vast majority of research related to QDIPs over the course of the decade was related to four main material regimes: InAs dots in GaAs barriers, InGaAs dots in GaAs barriers, InGaAs dots in InGaP barriers, and InAs dots in InGaAs barriers. Of these, InAs dots in GaAs barriers was the most common device in the early part of the decade. QDIPs using InAs dots in GaAs barriers typically operate in the SWIR and MWIR regimes, with a handful out as far as 7 µm. In the devices of the early part of the decade there was significant difficulty in growing numerous iterations of absorbing layers due to a buildup of strain, exacerbated by the irregular surface morphology of the dots. As such, detectors rarely exceed ten layers. Sensors 2013, 13 0.3 A/W [193,208,209]. Advances were also made with FPAs, with a more refined device fabricated in the middle of the decade. A 256 × 256 pixel FPA with similar performance metrics was fabricated with significantly improved imaging capabilities over previous devices [210]. However, following the middle part of the decade research on InAs/GaAs QDIPs slowed, replaced by research into the other material systems. A different pair of the material systems had been under limited research over the past decade. The first of these systems was InGaAs dots on GaAs barriers. These materials offer increased design flexibility, as dot composition can be varied along with dot size. This system was researched more towards the later portion of the decade, as interest in the InAs/GaAs system waned. Operation temperatures InGaAs/GaAs QDIPs were comparable to InAs/GaAs devices. Early devices also had a number of the same issues as InAs/GaAs, including a low number of quantum dot layers and low detectivities. By the end of the decade, FPAs using InGaAs/GaAs were outperforming InAs/GaAs FPAs, capable of detectivities of 1.01 × 1011 cmHz1/2/W and responsivities of 2.1 A/W by Jagadish et al. [15,211–213]. The other lesser used material system involved InGaAs dots in InGaP barriers. However, these devices generally had poor performance, including low quantum efficiencies and fill factors. These low metrics have been attributed to the lower lattice constant mismatch (there is a mismatch of about 3.2% between InGaAs and InGaP, whereas InAs/GaAs devices have a mismatch of about 7%) [214–217]. The last common material system that has been investigated was InAs dots in InGaAs barriers. As with other material systems for QDIPs, research for this system did not begin getting traction until the middle portion of decade when researchers were looking to improve performance from InAs/GaAs devices. These InAs/InGaAs devices functioned in the MWIR and LWIR regimes and operated at temperatures as high as 190 K. QDIPs of this system which operated at longer wavelengths did tend to suffer in terms of D, only reaching as high as 4.6 × 109 cmHz1/2/W at LN2 temperatures. However, when operated in the MWIR, these structures achieved comparable performance metrics to other material systems. Some of these devices, as well as some of the InGaAs/GaAs and InGaAs/InGaP devices, were grown by MOCVD, as opposed to MBE, with little hindrance on performance. 5. Quantum Dot Infrared Photodetectors (QDIPs) As might be expected, the low number of absorbing layers led to fairly low detectivities, as high as 2.94 × 109 cmHz1/2/W at LN2 temperatures by Madukar et al. [15,203–205]. It was in the early part of the decade that some of the earliest InAs/GaAs FPAs were fabricated. While the imaging of these devices was not superlative, as they were intended as proofs of concept more than final designs, they do produce recognizable images. Due to the low dark current densities present, even the initial devices were capable of operating up to 150 K in the MWIR regime [88,206]. As the decade moved forward, progress addressed a number of QDIP issues. Most notably was the number of quantum dot layers which could be deposited without dislocations. Whereas in the early part of the decade, layer numbers were limited to ten or less, by 2005 layer numbers InAs/GaAs QDIPs were in 50–70 range. Detectivities rose some with this increase, to a high of 1.2 × 1011 cmHz1/2/W [207]. It should be noted that even in these later devices, responsivities were still low, with a high value of Sensors 2013, 13 5074 Sensors 2013, 13 two energy wells in each layer of the absorbing stack of the device, albeit at a much higher energy level than the quantum dot well. The intention is that by tuning the energy level of this second energy well, the well will transmit photocurrent excited from the initial well. However, the well won’t transmit a large fraction of the dark current. These devices, while reducing dark current and maintaining respectable detectivities, encountered great difficulty in terms of overall responsivity, not capable of achieving a responsivity of 0.1 A/W at LN2 temperatures. This low responsivity and the additional complexity of the structure has limited the appeal of these devices [230,231]. Another interesting subset of QDIP devices is that which incorporate periodic current blocking layers. The current blocking layers (usually composed of either AlGaAs or InGaAs) were mainly incorporated at the early portion of the decade and were typically incorporated as capping layers on the quantum dots. The unintended consequence, at least in early devices, was the effect they had on the energy states in the quantum dots, particularly with InGaAs layers. These current blocking layers, if thick enough, could change the level of the energy states in the quantum dot or create entirely new energy levels in the barrier or well materials. While not having the explicit aim or the ultimate polish of later devices, it is easy to see these devices as precursor to quantum dot-in-a-well photodetectors (DWELLs). These early DWELL devices were plagued by many of the same issues as resonant tunneling barrier QDIPs, including low responsivities and low counts of quantum dot layers. However, they were able also to achieve detectivities up to 1.1 × 1012 cmHz1/2/W. Ultimately, a 320 × 256 pixel FPA was fabricated using these techniques, which was capable of imaging at 120 K, albeit with an NETD of 344 mK [213,220,232–235]. Sensors 2013, 13 MOCVD growth has an advantage to potentially make larger scale fabrication of QDIP devices much easier [218–221]. A 350 × 256 pixel FPA QDIP was fabricated near at the end of the decade. While operation of this device was constrained to LN2 temperatures, D at that temperature was as high 1 × 1010 cmHz1/2/W by Lai et al., which was among the highest for a QDIP FPA of any type [222]. A material system that may be interesting for future QDIP devices is GaAs dots on GaSb substrates. While this material system has not been used in a functioning IRPD, it has been used in other optical devices effectively, including lasers, solar cells, and LEDs, most notably by the Huffaker group [223–227]. For these devices, both Stranski-Krastanov and, more commonly due to the size of the lattice mismatch, interfacial misfit growth modes have been utilized. Most QD devices utilizing this material system have shown peak photoluminescence in the 1–2 µm range, which could make it suitable for NIR applications [228,229]. Another interesting subset of QDIPs utilized a pair of resonant tunneling layers incorporated into the periodic absorbing stack. In these structures, a quantum dot is deposited on a barrier material. This is followed by deposition of a very thin layer of a higher energy band gap material, a thin layer of the original barrier material, and a second thin layer of the higher energy band gap material. This creates 5075 Sensors 2013, 13 Sensors 2013, 13 significantly greater control over the energy level spacing. The well and dot layers, being adjacent, form what amounts to a combined energy well. For DWELLs which utilize two energy states in the well, the ground state of the well occurs in the dot, with the absorbing layer doped to the point such that electrons populate the lowest energy state in the dot at rest (Figure 9). The next energy level is designed to be located in the quantum well layer. The location of this energy level can be adjusted by varying the thickness of the well layer. This can allow for easier wavelength selectivity, as that selectivity can be determined by varying layer thicknesses instead of quantum dot diameters. The amount of flexibility in the design, in terms of dot size, material choices, well thickness, etc., has also led to alternative methods of multispectral imaging. While multispectral imaging is available in a number of different technologies, it is generally implemented by stacking detectors for two different wavelengths monolithically. In DWELLs, there is enough flexibility such that two or three energy transitions can exist within the well. Each transition then corresponds to a different absorption wavelength [8,33,237]. Figure 9. A band diagram of a DWELL structure, with a barrier layer (blue), well layer (red), and dot layer (green). g g , y ( ), y (red), and dot layer (green). While DWELLs do address a number of the weaknesses present in QWIPs and QDIPs, the technology still has some of the faults present in other devices. One of the major difficulties DWELLs have in common with QDIPs is a low number of absorbing layers which can be grown on a single device. As with QDIPs, use of the quantum dots can lead to strain build-up. As more layers are deposited, the strain accrues, and if too many layers are deposited defects can form. Accumulations of defects limit the amount of absorbing layers that can be deposited in a DWELL, which can in turn lead to low absorbances and quantum efficiencies. Much like in QDIPs, this buildup of strain can be mitigated with the inclusion of strain reducing layers at the expense of device complexity [238]. 6. Quantum Dots-in-a-Well Photodetectors (DWELL-IPs) Utilization of structures called DWELLs in IRPDs began in 2002, though the structure was previously used in other applications, such as semiconductor lasers [33]. While 2002 is the earliest direct reference to a DWELL structure, other structures did utilize various capping layers in QDIPs. These structures unintentionally made the introduction of DWELLs to IRPDs harder to pin down. Over the next couple years there was a gradual delineation between the DWELL structures and QDIPs. Once that delineation had been made, research on devices specifically labeled DWELLs picked up. The physical structure of a DWELL is much akin to a combination of a QWIP and a QDIP. A layer of the absorbing stack begins with deposition of a wide band gap material, the barrier material. This deposition is followed by growth of a narrower band gap material, the well material, much as in a QWIP. Growth is then followed with deposition of a layer of quantum dots with a band gap that is narrower still, known as the dot material. At this point, some devices will add a capping layer, sometimes composed of the well material [236]. Because these devices typically require precise tolerances, MBE deposition is typically the method of growth chosen for DWELLs. The absorbing layers will then be repeated a number of times to improve the absorption of incident light [8]. The structure of the DWELL is designed to address some of the weaknesses of both QWIPs and QDIPs. One of the most notable flaws of the QWIP is the inability to absorb normally incident light. The inclusion of a quantum dot layer allows DWELLs to utilize light at normal incidence. QDIPs can have difficulty in precisely controlling the energy levels in the quantum dots, as those energy levels depend heavily on dot size. The inclusion of a well layer in addition to the quantum dot layer adds 5076 Sensors 2013, 13 5077 Sensors 2013, 13 to be most important with the quantum dot layer, as the lower growth speed allows for greater uniformity of quantum dot size, which in turn leads to better performance at the desired absorption wavelength. Similarly, higher quantum dot densities were also associated with stronger response, but were much easier to achieve with slower deposition speeds [8,82,239,240]. Most of the early DWELL devices had a similar structure. Operating temperatures were typically around that of LN2, though some early designs did achieve operation near 100 K. These devices utilized a structure of GaAs barriers, some composition of InGaAs for the wells, and InAs quantum dots. The barriers were typically tens of nanometers thick, had well layers under ten nanometers, and contained quantum dots 1–3 ML thick. The variations in layer thicknesses were enough to dramatically alter devices absorption characteristics. Early DWELL-IPs spanned a wide range of absorption wavelengths, from the MWIR to VLWIR. A number of early devices also exhibited multispectral absorption from multiple available energy transitions within individual wells [33,201,241]. By the middle portion of the decade, a small 320 × 256 pixel FPA was produced, though for this early device responsivity and quantum efficiency was very low [8]. Dark current levels in these early devices were generally lower than those in QWIPs, but still higher than those in QDIPs. Responsivities varied significantly depending on the absorption wavelength, but responsivities in the MWIR regime as high as 3.58 A/W were achieved by Krishna et al. At longer wavelengths, this value decreased dramatically, with some of the VLWIR regime devices having responsivities as low as 0.01–0.02 A/W by Krishna et al. [33]. Moving through the decade, research continued on DWELLs at universities nationwide. Operating wavelengths were extended out further into the VLWIR regime. Often, long wavelength detection occurred as the secondary or tertiary absorption wavelength in the DWELL. As long as the devices were operated at low temperatures (sometimes as low as 5 K) detection of VLWIR radiation was still possible [242,243]. A number of groups continued characterizing the effect of various growth conditions and deposition characteristics on DWELL performance. Variation in the doping density for DWELLs was investigated by Attaluri, Krishna, et al, with higher doping densities leading to both larger spectral responses and large dark current densities. Sensors 2013, 13 M f th li t d i hi h tili d DWELL i t d ith f f t While DWELLs do address a number of the weaknesses present in QWIPs and QDIPs, the technology still has some of the faults present in other devices. One of the major difficulties DWELLs have in common with QDIPs is a low number of absorbing layers which can be grown on a single device. As with QDIPs, use of the quantum dots can lead to strain build-up. As more layers are deposited, the strain accrues, and if too many layers are deposited defects can form. Accumulations of defects limit the amount of absorbing layers that can be deposited in a DWELL, which can in turn lead to low absorbances and quantum efficiencies. Much like in QDIPs, this buildup of strain can be mitigated with the inclusion of strain reducing layers at the expense of device complexity [238]. Many of the earliest devices which utilized DWELLs existed either as proofs of concept or confirmation of theoretical predictions. Characteristics, such as the effect of variation of quantum dot size on absorption wavelength (smaller dots absorbed shorter wavelengths), the effect of variation of well layer thickness on absorption wavelength (thicker layers led to longer absorbed wavelengths), and the effect of variation of well layer mole composition (narrower band gaps led longer absorption wavelengths), were all investigated. Other growth characteristics were also characterized in this time frame. The effect of overall deposition speed on device performance was examined, with slower growth speeds associated with higher photoluminescent response. The effect of growth speed proved Sensors 2013, 13 Sensors 2013, 13 normal DWELL devices with a handful of minor distinctions. These devices tended to have marginally lower responsivities (around half that of single well DWELLs), while have improved NETDs in FPAs, reaching as low as 105 mK [250]. These devices also were capable of D* as high as 3 × 1010 cmHz1/2/W at LN2 temperatures by Krishna et al. [251–258]. Other growth methods besides MBE were attempted for DWELLs, with the intention of easing fabrication of the devices. These alternative methods still required fine control of deposition speeds, to accommodate quantum dot growth, and the ability to deposit a variety of materials without breaking vacuum. The first alternative method was MOCVD. Working DWELL devices were fabricated using MOCVD, though it should be noted that the devices produced had notably lower responsivities and detectivities at LN2 temperatures. These devices performed in the range of 0.05–15 A/W and 1–6 × 109 cmHz1/2/W, respectively. Additionally, a small change to the typical DWELL structure was made in these devices. Instead of using InGaAs for the well material and GaAs for the barrier material, these devices used GaAs for the well material and AlGaAs for the barrier material. The effect in terms of the band diagram of the DWELL structure is effectively the same, with only slightly different differences in the band edge locations. The use of these two materials does have one particular advantage. Compositions of AlGaAs are closely lattice-matched to GaAs, whereas InGaAs (particularly compositions with high indium contents) may not necessarily have the same trait. This trait is of particular significance when changing to a lower quality growth method, such as from MBE to MOCVD, as the larger mismatch can lead to high densities of defects [259,260]. A number of DWELL designs utilized similar resonant structures that have been mentioned previously for other IRPD technologies. These designs implemented a resonant tunneling barrier, typically located in the well layer of the DWELL. These layers are typically a pair of wide band gap layers of AlGaAs with a layer of InGaAs in between. Notably, the AlGaAs layer has a wider band gap than the barrier of the normal DWELL structure and the InGaAs layer has a narrower band gap than the band gap of the well material. As in the previous examples, these layers are included to attempt to reduce dark current and improve detectivity. Sensors 2013, 13 The DWELL structures utilizing the resonant tunneling structures did exhibit strong D, as high as 2.9 × 1010 cmHz1/2/W in the LWIR regime at LN2 temperatures. However, they also suffered in terms of device responsivity [261–263]. There was additional progress made in the development of DWELL FPAs during this time frame. FPAs as large as 640 × 512 pixels were fabricated for MWIR applications at LN2 temperatures. This device has a D as high as 1 × 1010 cmHz1/2/W and an NETD of 40 mK [264]. Sensors 2013, 13 It was also theorized that doping DWELLs to the point that there was an average of two electrons per quantum dot (opposed to the lower numbers normally used) could lead to a decrease in photocurrent due to overfilling energy states in the well [89]. Perera, Krishna et al. investigated the effect of annealing on DWELL performance. Annealing was found to increase the responsivity of the device, but also increased the dark current and shifted the absorption peak by up to 3.5 µm in wavelength. The change in performance is attributed to the effect of annealing on the quantum dots. It is believed that annealing increased material diffusion between the dot and the capping layer and, therefore, reduces the confinement of the quantum dots in devices. Despite these advances, the operating temperatures of DWELL devices remained fairly low, rarely operating significantly above LN2 temperatures and even then barely above them [244–246]. Double well DWELLs, where one period of the DWELL absorbing incorporates a barrier layer, two well layers, and a dot layer, were also developed over this time period, led by the Krishna group. In these structures, the dot layer is generally InAs, the first well is a thin layer of InGaAs to aid in the formation of the quantum dots, the second well layer is GaAs, and the barrier layer is AlGaAs. Less strain develops in a DWELL using this structure, allowing for more iterations of the absorbing layer to be deposited in a single device [247–249]. Using this structure, devices with up to 80 periods have been fabricated. These devices operated in both the MWIR and LWIR regimes and performed similarly to 5078 7. Discussion Tables 1–6 list the highest performing devices of different IRPD technologies across different wavelength regimes. In these Tables, a number of the most common IRPD performance metrics are examined. Results which stray too far from these standard conditions and metrics, or utilize wholly different metrics for characterization, are not included in these Tables. Detectivity (D*), both at LN2 and at the highest temperature with successful results are reported. Similar results are reported for responsivity (R). The highest temperature at which a device has been successfully operated is also reported. The lowest dark current density (Jdark) at LN2 temperatures is included in the table. Finally, to compare FPA performance, a comparison of noise equivalent temperature difference (NETD) is also included. 5079 Sensors 2013, 13 Sensors 2013, 13 Table 1. Highest reported performance metrics for HgCdTe IRPDs across different wavelength regimes. Table 1. Highest reported performance metrics for HgCdTe IRPDs across different wavelength regimes. Metric ࡿࢃࡵࡾ ࡹࢃࡵࡾ ࡸࢃࡵࡾ ࢂࡸࢃࡵࡾ ܦ௅ேమ כ ሾܿ݉ܪݖଵଶ ⁄ ܹ ⁄ ሿ 1 ൈ10ଵସ [265] 2 ൈ10ଵଷ [265] 2 ൈ10ଵଷ [265] X ܦ்೓೔೒೓ כ ሾܿ݉ܪݖଵଶ ⁄ ܹ ⁄ ሿ൫ܶ௛௜௚௛ ሾܭሿ൯ 1 ൈ10ଵଶ ሺ300ሻ [265] 7 ൈ10ଵ଴ (210) [25] 1 ൈ10଻ (300) [52] X ܴ௅ேమ ሾሺܣܹ ⁄ ሻሿ X X 1 [266] X ்ܴ೓೔೒೓൫ܶ௛௜௚௛൯ሾܣܹሿ ⁄ 1 [24] (300) X X X ܶ௠௔௫ ሾܭሿ 300 210 [25] 300 100 [35] ܬௗ௔௥௞ ሾܣܿ݉ଶ ⁄ ሿ X X 2.5 ൈ10ିସ [36] 2.7 ൈ10ି଻ [41] ܰܧܶܦ ሾ݉ܭሿ 20 [24] 10 [267] 9 [55] 22 [41] Highest reported performance metrics for InSb IRPDs across different wavelength regimes. Metric ࡿࢃࡵࡾ ࡹࢃࡵࡾ ࡸࢃࡵࡾ ࢂࡸࢃࡵࡾ ܦ௅ேమ כ ሾܿ݉ܪݖଵଶ ⁄ ܹ ⁄ ሿ X 1 ൈ10ଵ଴ [63] 7.6 ൈ10଼ [64] X ܦ்೓೔೒೓ כ ሾܿ݉ܪݖଵଶ ⁄ ܹ ⁄ ሿ൫ܶ௛௜௚௛ሾܭሿ൯ X 2.8 ൈ10଼ (300) [73] X X ܴ௅ேమ ሾሺܣܹ ⁄ ሻሿ X 1.4 [60] X X ்ܴ೓೔೒೓൫ܶ௛௜௚௛൯ሾܣܹሿ ⁄ X X X X ܶ௠௔௫ ሾܭሿ X 300 [73] 300 X ܬௗ௔௥௞ ሾܣܿ݉ଶ ⁄ ሿ X X X X ܰܧܶܦ ሾ݉ܭሿ X 2.2 [72] X X 3. Highest reported performance metrics for QWIPs across different wavelength regimes. Table 3. Highest reported performance metrics for QWIPs across different wavelength regimes. 7. Discussion Metric ࡿࢃࡵࡾ ࡹࢃࡵࡾ ࡸࢃࡵࡾ ࢂࡸࢃࡵࡾ ܦ௅ேమ כ ሾܿ݉ܪݖଵଶ ⁄ ܹ ⁄ ሿ X 2 ൈ10ଵଵ [116] 2 ൈ10ଵଵ [91] 5 ൈ10ଽ [121] ܦ்೓೔೒೓ כ ሾܿ݉ܪݖଵଶ ⁄ ܹ ⁄ ሿ൫ X 1 ൈ10ଵଵ (210) [107] 3.2 ൈ10ଵ଴ (110) [99] 5 ൈ10ଽ (70) [121] ܴ௅ேమ ሾሺܣܹ ⁄ ሻሿ X 1 [268] 1.4 [118] 0.38 [119] ்ܴ೓೔೒೓൫ܶ௛௜௚௛൯ሾܣܹሿ ⁄ X 0.7 (290) [107] 0.3 (120) [92] 0.25 (70) [121] ܶ௠௔௫ ሾܭሿ X 290 [107] 120 70 [121] ܬௗ௔௥௞ ሾܣܿ݉ଶ ⁄ ሿ X 1 ൈ10ି଻ [116] 1 ൈ10ିଽ [99] 1.7 ൈ10ିଷ [109] ܰܧܶܦ ሾ݉ܭሿ X 14 [123] 19 [118] 48 [122] Table 4. Highest reported performance metrics for QDIPs across different wavelength regimes. Table 4. Highest reported performance metrics for QDIPs across different wavelength regimes. Metric ࡿࢃࡵࡾ ࡹࢃࡵࡾ ࡸࢃࡵࡾ ࢂࡸࢃࡵࡾ ܦ௅ேమ כ ሾܿ݉ܪݖଵଶ ⁄ ܹ ⁄ ሿ 1 ൈ10ଽ [203] 1 ൈ10ଵଶ [235] 1ݔ10ଵଵ [213] X ܦ்೓೔೒೓ כ ሾܿ݉ܪݖଵଶ ⁄ ܹ ⁄ ሿ൫1 ൈ10ଽ (77) [203] 2.4 ൈ10଼ (250) [269] 1.2 ൈ10ଵଵ (77) [207] 2 ൈ10଼ (40) [15] ܴ௅ேమ ሾሺܣܹ ⁄ ሻሿ 1 [203] 2.5 [270] 5.3 [232] X ்ܴ೓೔೒೓൫ܶ௛௜௚௛൯ሾܣܹሿ ⁄ 1 (77) [203] 0.012 (300) [270] 5.3 (77)[232] 4 (40) [15] ܶ௠௔௫ ሾܭሿ 77 [203] 300 [270] 77 40 [15] ܬௗ௔௥௞ ሾܣܿ݉ଶ ⁄ ሿ X 1 ൈ10ିଵଷ [193] 3 ൈ10ି଼ [232] X ܰܧܶܦ ሾ݉ܭሿ X 87 [214] X X Sensors 2013, 13 5080 Sensors 2013, 13 Table 5. Highest reported performance metrics for DWELL IRPDs across different wavelength regimes. Metric ࡿࢃࡵࡾ ࡹࢃࡵࡾ ࡸࢃࡵࡾ ࢂࡸࢃࡵࡾ ܦ௅ேమ כ ሾܿ݉ܪݖଵଶ ⁄ ܹ ⁄ ሿ X 1 ൈ10ଵ଴ [271] 3 ൈ10ଵ଴ [256] X ܦ்೓೔೒೓ כ ሾܿ݉ܪݖଵଶ ⁄ ܹ ⁄ ሿ൫ X 1 ൈ10଼ (250) [272] 1 ൈ10଼ (200) [256] X ܴ௅ேమ ሾሺܣܹ ⁄ ሻሿ X 0.12 [259] 3.58 [33] X ்ܴ೓೔೒೓൫ܶ௛௜௚௛൯ሾܣܹሿ ⁄ X 0.12 (77) [259] 0.2 (110) [244] 0.025 (4.6) [8] ܶ௠௔௫ ሾܭሿ X 77 110 [244] 4.6 [8] ܬௗ௔௥௞ ሾܣܿ݉ଶ ⁄ ሿ X 1 ൈ10ିହ [271] 5 ൈ10ହ [261] X ܰܧܶܦ ሾ݉ܭሿ X X 40 [264] X Table 6. Highest reported performance metrics for SLS IRPDs across different wavelength regimes. hest reported performance metrics for SLS IRPDs across different wavelength regimes. Table 6. Highest reported performance metrics for SLS IRPDs across different wavelength regimes. 7. Discussion Metric ࡿࢃࡵࡾ ࡹࢃࡵࡾ ࡸࢃࡵࡾ ࢂࡸࢃࡵࡾ ܦ௅ேమ כ ሾܿ݉ܪݖଵଶ ⁄ ܹ ⁄ ሿ X 1.9 ൈ10ଵଷ [169] 1 ൈ10ଵଶ [155] 4.5 ൈ10ଵ଴ [150] ܦ்೓೔೒೓ כ ሾܿ݉ܪݖଵଶ ⁄ ܹ ⁄ ሿ൫ܶ௛௜௚௛ ሾܭሿ൯ 1 ൈ10଼ (300) [166] 4.6 ൈ10ଽ (300) [144] X 4.5 ൈ10ଵ଴ (80) [150] ܴ௅ேమ ሾሺܣܹ ⁄ ሻሿ X 1.33 [169] 3.2 [162] 4 [150] ்ܴ೓೔೒೓൫ܶ௛௜௚௛൯ሾܣܹሿ ⁄ X 2.2 (300) [144] X 4(80) [150] ܶ௠௔௫ ሾܭሿ 300 [166] 300 180 [162] 80 [150] ܬௗ௔௥௞ ሾܣܿ݉ଶ ⁄ ሿ X 1 ൈ10ି଻ [163] 1 ൈ10ିଷ [171] 3.3 ൈ10ିଷ [184] ܰܧܶܦ ሾ݉ܭሿ X 10 [178] 270 [162] X 8. Conclusions and Future Prospects Sensors 2013, 13 5081 Sensors 2013, 13 technologies. The increase in research conducted by private businesses in these technologies has led to a change in project selection from many funding agencies, viewing research in these fields increasingly as industrial development as opposed to academic research. That academic research that is funded has a greater focus on producing front-end devices, which may eventually lead to commercial ventures, as opposed to investigating novel applications or phenomenon in the established technology. This has primarily occurred with bulk detectors so far, primarily with MCT and InSb, but as other technologies become more prevalent this trend will likely spread across the field. Acknowledgments We would like to thank Tufts University for the John Mellon Fellowship and The Alexander Von Humboldt Foundation of Germany for the Humboldt Research Fellowship for Experienced Researchers. Their support enabled his work. 8. Conclusions and Future Prospects The progress that the field of infrared photodetection has made in recent history is tremendous. What started as primarily academic research has transformed into a budding industry, including a number of new technologies in the pipeline. IRPDs will see increasing use in a number of fields in the present and near future, ranging from military target detection to environmental sensing. With the number of applications for these technologies rapidly increasing, there is a need to determine which technology is best suited to each application. The number of different IRPD technologies has led to segmentation of which is currently the most viable depending on the application. Bulk detectors have achieved the most commercial penetration due to ease of large scale fabrication, long track record of growth, flexible absorption wavelengths (which can span from SWIR to LWIR), and high detectivity at cryogenic temperatures, but there has been difficulty in raising the operating temperatures of these devices while maintaining performance. QWIPs and QDIPs have also drawn significant attention due to the former’s capability to perform in the VLWIR regime and the latter’s high temperature performance, currently the best among current devices. DWELL-IPs and SLS devices have shown noteworthy promise in terms of performance, range of absorption range, and operating temperature. It may well be that these devices could soon overtake QWIPs and QDIPs in terms of performance. Moving forward, progress in the field of IRPDs may come from dramatically different sources than research to date. In recent years, there has been a shift in the source of funding for some IRPD Richards, A.A. Spectral selection in infrared imaging. Proc. SPIE 2004, 5405, 288–293. Sensors 2013, 13 Sensors 2013, 13 15. Towe, E.; Pan, D. Semiconductor quantum-dot nanostructures: Their application in a new class of infrared photodetectors. IEEE J. Sel. Top. Quantum Electron. 2000, 6, 408–421. 16. Mikhailenko, S.N.; Babikov, Y.L.; Golovko, V.F. Information-calculating system spectroscopy of atmospheric gases. Atmos. Ocean. Opt. 2005, 18, 685–695. 17. Liu, H.; Zhang, F.; Zhang, J.; He, G. Performance analysis of quantum dots infrared photodetector. Proc. SPIE 2011, doi:10.1117/12.898602. 18. Schneider, H.; Walther, M.; Schonbein, C.; Rehm, R.; Fleissner, J.; Pletschen, W.; Braunsteina, J.; Koidl, P.; Weimann, G.; Ziegler, J.; Cabanski, W. QWIP FPAs for high-performance thermal imaging. Physica E 2000, 7, 101–107. 19. Reine, M.B. HgCdTe photodiodes for IR detection: A review. Proc. SPIE 2001, 4288, 266–277. 20. Kohn, A. N.; Schlickman, J. J. 1-2 micron (Hg, Cd)Te photodetectors. Electron Devices, IEEE Transactions on 1969, 16, 885–890. 21. Lawson, W.D.; Nielsen, S.; Putley, E.H.; Young, A.S. Preparation and properties of HgTe and mixed crystals of HgTe-CdTe. J. Phys. Chem. Solid. 1959, 9, 325–329. 22. Kruse, P.W.; Blue, M.D.; Garfunkel, J.H.; Saur, W.D. Long wavelength photoeffects in mercury selenide, mercury telluride and mercury telluride-cadmium telluride. Infrared Phys. 1962, 2, 53–60. 23. Piotrowski, J.; Orman, Z.; Nowak, Z.; Pawluczyk, J.; Pietrzak, J.; Piotrowski, A.; Szabra, D. Uncooled long wave infrared photodetectors with optimized spectral response at selected spectral ranges. Proc. SPIE 2005, 5783, 616–624. 24. Chorier, P.; Tribolet, P. High performance HgCdTe SWIR detectors development at Sofradir. Proc. SPIE 2001, 4369, 698–712. 25. Vasilyev, V.V; Ovsyuk, V.N.; Sidorov, Y.G. IR photodetectors based on MBE grown MCT layers. Proc. SPIE 2003, 5065, 39–46. 26. Krishna, S. Quantum dots-in-a-well infrared photodetectors. Infrared Phys. Technol. 2005, 47, 153–163. 27. Pal, E.; Towe, D. Uniformity assessment of key characteristics of quantum-dot infrared detectors: A prerequisite for focal plane arrays. J. Appl. Phys. 2006, doi:10.1063/1.2356914. 28. Stafeev, V. Mercury cadmium telluride-main semiconductor material of modern IR photoelectronics. Proc. SPIE 2000, 4340, 240–243. 29. Shi, Y.-L.; Lei, S.-Q.; Yang, T.-F.; Zhuang, J.-S.; Cai, Y. Short wavelength Hg1-xCdxTe infrared detectors prepared by loophole technology. Proc. SPIE 2005, 5964, 596401–596409. 30. Rogalski, A. New material systems for third generation infrared detectors. Proc. SPIE 2009, doi:10.1117/12.852524. 31. Razeghi, M.; Hoffman, D.; Nguyen, B.-M.; Delaunay, P.-Y.; Huang, E.K.-W.; Tidrow, M.Z.; Nathan, V. Recent advances in LWIR Type-II InAs/GaSb superlattice photodetectors and focal plane arrays at the center for quantum devices. Proc. IEEE 2009, 97, 1056–1066. 32. References 1. Herschel, W. Experiments on the refrangibility of the invisible rays of the sun. Phil. Trans. Royal Soc. Lond. 1800, 90, 284–292. 1. Herschel, W. Experiments on the refrangibility of the invisible rays of the sun. Phil. Trans. Royal Soc. Lond. 1800, 90, 284–292. 2. Barr, E.S. The infrared pioneers—II. Macedonio melloni. Infrared Phys. 1962, 2, 67–73. 2. Barr, E.S. The infrared pioneers—II. Macedonio melloni. Infrared Phys. 1962, 2, 67–73. 3. Case, T.W. Notes on the change of resistance of certain substances in light. Phys. Rev. 1917, 9, 305–310. 3. Case, T.W. Notes on the change of resistance of certain substances in light. Phys. Rev. 1917, 9, 305–310. A. Infrared detectors: Status and trends. Prog. Quantum Electron. 2003, 27, 59–210. 4. Rogalski, A. Infrared detectors: Status and trends. Prog. Quantum Electron. 2003, 5. Buffaz, A.; Carras, M.; Doyennette, L.; Nedelcu, A.; Bois, P.; Berger, V. State of the art of quantum cascade photodetectors. Proc. SPIE 2010, doi:10.1117/12.853525. 5. Buffaz, A.; Carras, M.; Doyennette, L.; Nedelcu, A.; Bois, P.; Berger, V. State of the art of quantum cascade photodetectors. Proc. SPIE 2010, doi:10.1117/12.853525. ; Krishna, S. Advances in Infrared Photodetectors; Elsevier: San Diego, CA, USA, 2011. 6. Barve, A.; Krishna, S. Advances in Infrared Photodetectors; Elsevier: San Diego, CA, 7. Xiaohui, X.; Hua, H.; Guangyin, Q.; Qingjun, L.; Xiaoning, H. Study of the characteristics of VLWIR HgCdTe photovoltaic detectors in variable-area diode test structures. Proc. SPIE 2011, doi:10.1117/12.900711. 8. Krishna, S. Quantum dots-in-a-well infrared photodetectors. J. Phys. D 2005, 38, 2142–2150. 9. Stiff-Roberts, A.D. Quantum-dot infrared photodetectors: A review. J. Nanophon. 2011, doi:10.1117/1.3125802. 10. Jdidi, A.; Sfina, N.; Abdi-Ben Nassrallah, S.; Saïd, M.; Lazzari, J.L. A multi-color quantum well photodetector for mid- and long-wavelength infrared detection. Semicond. Sci. Technol. 2011, doi:10.1088/0268-1242/26/12/125019. 11. Bundas, J.; Patnaude, K.; Dennis, R.; Burrows, D.; Cook, R.; Reisinger, A.; Sundaram, M.; Benson, R.; Woolaway, J.; Schlesselmann, J.; et al. Two-color quantum well infrared photodetector focal plane arrays. Proc. SPIE 2006, doi:10.1117/12.665803. 12. Chubb, D.L. Fundamentals of Thermophotovoltaic Energy Conversion; Elsevier: Amsterdam, The Netherlands, 2007; p. 515. 13. Daniel, J.S.; Solomon, S.; Kjaergaard, H.G.; Schofield, D.P. Atmospheric water vapor complexes and the continuum. Geophys. Res. Lett. 2004, 31, 1–4. 14. Richards, A.A. Spectral selection in infrared imaging. Proc. SPIE 2004, 5405, 288–293. 5082 Sensors 2013, 13 Wang, J.; Chen, X.; Hu, W.; Wang, L.; Lu, W.; Xu, F.; Zhao, J.; Shi, Y.; Ji, R. Amorphous HgCdTe infrared photoconductive detector with high detectivity above 200 K. Appl. Phys. Lett. doi:10.1063/1.3638459. 33. Raghavan, S.; Rotella, P.; Stintz, A.; Fuchs, B.; Krishna, S.; Morath, C.; Cardimona, D.A.; Kennerly, S.W. High-responsivity, normal-incidence long-wave infrared (lambda~7.2 um) InAs/In(0.15)Ga(0.85)As dots-in-a-well detector. Appl. Phys. Lett. 2002, 81, 3. 5083 Sensors 2013, 13 34. Cabanski, W.A.; Breiter, R.; Koch, R.; Mauk, K.; Rode, W.; Ziegler, J.; Eberhardt, K.; Oelmaier, R.; Schneider, H.; Walther, M. High-resolution focal plane array IR detection modules and digital signal processing technologies at AIM. Proc. SPIE 2000, 4028, 113–121. 35. D’Souza, A.I.; Dawson, L.C.; Staller, C.O.; Reekstin, J.P.; Wijewarnasuriya, P.S.; DeWames, R.E.; McLevige, W.V; Arias, J.M.; Edwall, D.E.; Hildebrandt, G. VLWIR Hg(1-x)Cd(x)Te photovoltaic detectors performance. Proc. SPIE 2000, 4028, 343–352. 36. Wijewarnasuriya, P.S.; Zandian, M.; Phillips, J.; Edwall, D.; DeWames, R.E.; Hildebrandt, G.; Bajaj, J.; Arias, J.M.; D’Souza, A.I.; Moore, F. Advances in large area Hg(1-x)Cd(x)Te photovoltaic detectors for remote sensing applications. Proc. SPIE 2001, 4454, 10–18. 37. Grudzien, M.; Piotrowski, J. Infrared detectors: View from a high-tech company. Proc. SPIE 2003, 5124, 15–21. 38. Vydyanath, H.R.; Tobin, S.P.; Norton, P.W.; Odette, P.; Nathan, V. HgCdTe photodiodes with state-of-the-art performance in the near infrared spectral region. Proc. SPIE 2001, 4288, 316–334. 39. D’Souza, A.I.; Stapelbroek, M.G.; Robinson, J. HgCdTe and silicon detectors and FPAs for remote sensing applications. Proc. SPIE 2004, 5234, 262–271. 40. Smith, E.P.G.; Bornfreund, R.E.; Kasai, I.; Pham, L.T.; Patten, E.A.; Peterson, J.M.; Roth, J.A.; Nosho, B.Z.; De Lyon, T.J.; Jensen, J.E.; et al. Status of two-color and large format HgCdTe FPA technology at Raytheon vision systems. Proc. SPIE 2006, 6127, 61271. 41. Haiml, M.; Bauer, A.; Bitterlich, H.; Bruder, M.; Hofmann, K.; Lutz, H.; Mai, M.; Nothaft, H.-P.; Ruhlich, I.; Wendler, J.; et al. AIM results for space-qualified HgCdTe photovoltaic detectors from 0.9-um to 13-um spectral range. Proc. SPIE 2006, 6361, 636111–636119. 42. Vuillermet, M.; Pistone, F.; Reibel, Y. Latest developments in MCT for next generation of infrared staring arrays. Proc. SPIE 2009, doi:10.1117/12.831270. 43. Crawford, S.; Skivington, T.; Craig, R.; Haining, A.; Costard, E.; Belhaire, E.; Bois, P. Beyond 3rd generation MCT: SXGA QWIP. Proc. SPIE 2005, 5783, 777–788. 44. Rochette, L.; Smithson, T.; Gurgenian, R.; St-Pierre, G. Sensors 2013, 13 Long wave focal plane array detector: Development performance assessment of an 8x8 sterling cooled photovoltaic MCT detector module. Proc. SPIE 2007, 6661, 666107. 45. Cabanski, W.A.; Eberhardt, K.; Rode, W.; Wendler, J.C.; Ziegler, J.; Fleissner, J.; Fuchs, F.; Rehm, R.H.; Schmitz, J.; Schneider, H.; et al. Third-generation focal plane array IR detection modules and applications. Proc. SPIE 2004, 5406, 184–192. 46. Bovina, L.A.; Boltar, K.O.; Bourlakov, I.D.; Golovin, S.V; Ivanov, V.I.; Saginov, L.D.; Stafeev, V.I.; Akimov, V.M.; Iakovleva, N.I.; Sidorov, Y.G. 384 × 288 HgCdTe staring focal plane array. Proc. SPIE 2000, 4340, 23–26. 47. Boltar, K.O.; Bourlakov, I.D.; Golovin, S.V; Lognikov, V.E.; Mansvetov, N.G. LWIR heterodyne HgCdTe detectors. Proc. SPIE 2000, 4340, 81–84. 48. Varavin, V.S.; Gutakovsky, A.K.; Dvoretsky, S.A.; Kartashev, V.A.; Latyshev, A.V; Mikhailov, N.N.; Pridachin, D.N.; Remestnik, V.G.; Rukhlitsky, S.V; Sabinina, I.V; et al. The state of the art and prospects of CdxHg1-xTe molecular beam epitaxy. Proc. SPIE 2003, 5126, 398–406. 49. Lee, T.-S.; Zhao, J.; Chang, Y.; Ashokan, R.; Sivananthan, S.; Boieriu, P.; Chen, Y.; Brill, G.; Wijewarnasuriya, P.S.; Dhar, N.K. MWIR detectors on HgCdTe grown by MBE on 3-in. diameter silicon substrates. Proc. SPIE 2004, 5564, 113–122. 5084 Sensors 2013, 13 50. Bajaj, J. State-of-the-art HgCdTe infrared devices. Proc. SPIE 2000, 3948, 42–54. 51. Wijewarnasuriya, P.S.; Brill, G.; Chen, Y.P.; Dhar, N.K.; Velicu, S. LWIR MBE HgCdTe photovoltaic detectors grown on Si composite substrates. Proc. SPIE 2004, 5406, 323–331. 52. Piotrowski, J.; Rogalski, A. Uncooled long-wavelength infrared photon detectors. Proc. SPIE 2004, 5359, 10–22. 53. Yakushev, M.V; Brunev, D.V; Varavin, V.S.; Vasilyev, V.V; Dvoretskii, S.A.; Marchishin, I.V; Predein, A.V; Sabinina, I.V; Sidorov, Y.G.; Sorochkin, A. V HgCdTe heterostructures on Si (310) substrates for midinfrared focal plane arrays. Semiconductors 2011, 45, 385–391. 54. Wijewarnasuriya, P.S. Nonequilibrium operation of long wavelength HgCdTe photo detectors for higher operating temperatures. Proc. SPIE 2010, doi:10.1117/12.870920. 55. Sizov, F.F.; Vasiliev, V.V; Gumenjuk-Sichevska, J.V; Kamenev, Y.E.; Dvoretsky, S.A.; Mikhailov, N.N. MCT detectors: From IR to sub-mm and mm wavelength bands. Proc. SPIE 2008, doi:10.1117/12.797436. 56. Aqariden, F.; Elsworth, J.; Zhao, J.; Grein, C.H.; Sivananthan, S. MBE HgCdTe for HDVIP Devices: Horizontal integration in the US HgCdTe FPA industry. J. Electron. Mater. 2012, 41, 2700–2706. 57. Wenisch, J.; Eich, D.; Lutz, H.; Schallenberg, T.; Wollrab, R.; Ziegler, J. MBE growth of MCT on GaAs substrates at AIM. J. Electron. Mater. 2012, 41, 2828–2832. 58. Sensors 2013, 13 Wei, Q.; Zhou, W.; Wang, W.; Fu, X.; Yu, M.; Chen, L.; Wu, Y.; He, L. Characterization of multilayered HgCdTe for MW/LW two-color application. Proc. SPIE 2007, 6835, 68351. 59. Wehner, J.G.A.; Musca, C.A.; Sewell, R.H.; Dell, J.M.; Faraone, L. Mercury cadmium telluride resonant-cavity-enhanced photoconductive infrared detectors. Appl. Phys. Lett. 2005, 87, 211101–211104. 60. Besikci, C. III-V infrared detectors on Si substrates. Proc. SPIE 2000, 3948, 31–39 61. Kimukin, I.; Biyikli, N.; Kartaloglu, T.; Aytur, O.; Ozbay, E. High-speed InSb photodetectors on GaAs for mid-IR applications. IEEE J. Sel. Top. Quantum Electron. 2004, 10, 766–770. 62. Joullie, A.F.; Christol, P.; Rodriguez, J.B.; Ait-Kaci, H.; Chevrier, F.; Nieto, J.; De Anda, F. A survey of new laser and detector structures for 3–5 um mid-infrared spectral range. Proc. SPIE 2004, 5582, 211–221. 63. Ozer, S.; Besikci, C. Assessment of InSb photodetectors on Si substrates. J. Phys. D 2003, 36, 559–563. 64. Razeghi, M. Novel Sb-based alloys for uncooled infrared photodetector applications. Proc. SPIE 2001, 4288, 304–315. 65. Ashley, T.; Baker, I.M.; Burke, T.M.; Dutton, D.T.; Haigh, J.A.; Hipwood, L.G.; Jefferies, R.; Johnson, A.D.; Knowles, P.; Little, J.C. InSb focal plane array (FPAs) grown by molecular beam epitaxy (MBE). Proc. SPIE 2000, 4028, 398–403. 66. Akimov, V.M.; Chishko, V.F.; Dirochka, A.I.; Kasatkin, I.L.; Klimanov, Y.A.; Kravtchenko, N.V; Lopukhin, A.A.; Pasekov, V.F. The 128 × 128 and 288 × 384 InSb array photodetective assemblies for 3–5-um spectral range. Proc. SPIE 2003, 5126, 91–97. 67. Hoffman, A.W.; Love, P.J.; Ando, K.J.; Rosbeck, J.P.; Lum, N.A.; Holcombe, R.; Durkee, J.M.; McMurtry, C.W. Development of 2k × 2k FPA InSb modules for the NGST mission. Proc. SPIE 2003, 4850, 836–846. 5085 Sensors 2013, 13 68. Shkedy, L.; Markovitz, T.; Calahorra, Z.; Hirsh, I.; Shtrichman, I. Megapixel digital InSb detector for midwave infrared imaging. Optical Engineering 2011, 50, 61008. 69. Hoffman, A.W.; Corrales, E.; Love, P.J.; Rosbeck, J.; Merrill, M.; Fowler, A.; McMurtry, C. 2K × 2K InSb for Astronomy. SPIE USE 2004, 5499, 59–67. 70. Nesher, O.; Pivnik, I.; Ilan, E.; Calalhorra, Z.; Koifman, A.; Vaserman, I.; Oiknine Schlesinger, J.; Gazit, R.; Hirsh, I. High resolution 1,280 × 1,024, 15 um pitch compact InSb IR detector with on-chip ADC. Proc. SPIE 2009, doi:10.1117/12.817054. 71. Kimukin, I.; Biyikli, N.; Ozbay, E. InSb high-speed photodetectors grown on GaAs substrate. J. Appl. Phys. 2003, 94, 5414–5416. 72. Sensors 2013, 13 Development of quantum well infrared photodetectors at the center for quantum devices. Infrared Phys. Technol. 2001, 42, 135–148. 84. El-Tokhy, M.; Mahmoud, I.; Konber, H. Comparative study between different quantum infrared photodetectors. Opt. Quantum Electron. 2009, 41, 933–956. 85. Hofstetter, D.; Schad, S.-S.; Wu, H.; Schaff, W.J.; Eastman, L.F. GaN/AlN-based quantum-well infrared photodetector for 1.55 m. Appl. Phys. Lett. 2003, 83, 572–574. 86. Guillot, F.; Monroy, E.; Leconte, S.; Bellet-Amalric, E.; Baumann, E.; Giorgetta, F.; Hofstetter, D.; Nevou, L.; Tchernycheva, M.; Doyennette, L.; et al. MBE growth of nitride-based photovoltaic intersubband detectors. Superlattice. Microstruct. 2006, 40, 418–425. 87. Sherliker, B.; Halsall, M.; Kasalynas, I.; Seliuta, D.; Valusis, G.; Vengris, M.; Barkauskas, M.; Sirutkaitis, V.; Harrison, P.; Jovanovic, V.D.; et al. Room temperature operation of AlGaN/GaN quantum well infrared photodetectors at a 3–4 um wavelength range. Semicond. Sci. Technol. 2007, 22, 1240–1244. 88. Stiff-Roberts; K.S.; Bhattacharya, P.; Kennerly, S. Low-bias, high-temperature performance of a normal incidence InAs/GaAs vertical quantum dot infrared photodetector with a current blocking barrier. J. Vac. Sci. Technol. B 2002, 20, 3. 89. Attaluri, R. S.; Annamalai, S.; Posani, K. T.; Stintz, A.; Krishna, S. Influence of Si doping on the performance of quantum dots-in-well photodetectors. J. Vac. Sci. Technol. B 2006, 24, 3. performance of quantum dots-in-well photodetectors. J. Vac. Sci. Technol. B 2006, 24, 3. 90. Tidrow, M.Z. Device physics and state-of-the-art of quantum well infrared photodetectors and arrays. Mater. Sci. Eng. B 2000, B74, 45–51. 90. Tidrow, M.Z. Device physics and state-of-the-art of quantum well infrared photodetectors and arrays. Mater. Sci. Eng. B 2000, B74, 45–51. 91. Gunapala, S.D.; Bandara, S.V; Liu, J.K.; Luong, E.M.; Stetson, N.; Shott, C.A.; Bock, J.J.; Rafol, S.B.; Mumolo, J.M.; McKelvey, M.J. Long-wavelength 256 × 256 GaAs/AlGaAs quantum well infrared photodetector (QWIP) palm-size camera. IEEE Trans. Electron. Devices 2000, 47, 326–332. 92. Li, S.S.; Kim, S.-H.; Moon, J.-H.; Lee, J.H. A two-stack, multi-color quantum well infrared photodetector for mid- and long-wavelength infrared detection. Infrared Phys. Technol. 2003, 44, 235–241. 93. Perera, A.G.U.; Matsik, S.G.; Liu, H.C.; Gao, M.; Buchanan, M.; Schaff, W.J.; Yeo, W. GaAs/InGaAs quantum well infrared photodetector with a cutoff wavelength at 35 m. Appl. Phys. Lett. 2000, 77, 741–743. 94. Perera, A.G.U.; Matsik, S.G.; Liu, H.C.; Gao, M.; Buchanan, M.; Schaff, W.J.; Yeo, W. 35 um cutoff bound-to-quasibound and bound-to-continuum InGaAs QWIPs. Infrared Phys. Technol. 2001, 42, 157–162. 95. Lee, J.-H.; Li, S.S.; Tidrow, M.Z.; Liu, W.K. Sensors 2013, 13 Camargo, E.G.; Ueno, K.; Morishita, T.; Sato, M.; Endo, H.; Kurihara, M.; Ishibashi, K.; Kuze, N. High-sensitivity temperature measurement with miniaturized InSb mid-IR sensor. IEEE Sens. J. 2007, 7, 1335–1339. 73. Kuze, N.; Camargo, E.G.; Ueno, K.; Morishita, T.; Sato, M.; Kurihara, M.; Endo, H.; Ishibashi, K. High performance miniaturized InSb photovoltaic infrared sensors operating at room temperature. J. Cryst. Growth 2007, 301–302, 997–1000. 74. Ueno, K.; Gomes Camargo, E.; Morishita, T.; Sato, M.; Yokoyama, A.; Endo, H.; Yanagita, Y.; Goto, H.; Kuze, N. Miniaturized InSb mid-IR photovoltaic sensor for room temperature operation. Physica Status Solidi C 2008, 5, 2819–2821. 75. Camargo, E.G.; Ueno, K.; Kawakami, Y.; Moriyasu, Y.; Nagase, K.; Satou, M.; Endo, H.; Ishibashi, K.; Kuze, N. Miniaturized InSb photovoltaic infrared sensor operating at room temperature. Opt. Eng. 2008, 47, 14401–14402. 76. Lim, K.P.; Pham, H.T.; Yoon, S.F.; Tan, K.H.; Ngo, C.Y. InSb1-xNx/InSb/GaAs alloys by thermal annealing for midinfrared photodetection. Appl. Phys. Lett. 2010, doi:10.1063/1.3524228. 77. Lackner, D.; Pitts, O.J.; Martine, M.; Cherng, Y.T.; Mooney, P.M.; Thewalt, M.L.W.; Plis, E.; Watkins, S.P. InAsSb and InPSb Materials for Mid Infrared Photodetectors. In Proceedings of the International Conference on Indium Phosphide and Related Materials; Kagawa, Japan, 31 May–4 June 2010; pp. 1–4. 78. Shafir, I.; Katz, M.; Sher, A.; Raizman, A.; Zussman, A.; Nathan, M. Suppression of leakage currents in InAsSb MWIR photodiodes by chemical treatment and illumination. Semicond. Sci. Technol. 2010, doi:10.1088/0268-1242/25/4/045004. 79. Levine, B.F.; Choi, K.K.; Bethea, C.G.; Walker, J.; Malik, R. New 10 um infrared detector using intersubband absorption in resonant tunneling GaAlAs superlattices. Appl. Phys. Lett. 1987, 50, 1092–1094. 80. Ryzhii, V.; Ryzhii, M.; Khmyrova, I.; Suris, R.; Mitin, V.; Shur, M. Quantum well and quantum dot infrared photodetectors: Physics of operation and modeling. Proc. SPIE 2003, doi:10.1117/ 12.517295. 81. Berurier, V.G.A.; Huet, O.; Manissadjian, A.; Facoetti, H.; Marcadet, X.; Carras, M.; Trinité, V.; Nedelcu, A. Quantum well infrared photodetectors: Present and future. Opt. Eng. 2011, doi:10.1117/1.3584838. 82. Sakoglu, U.; Tyo, J.S.; Hayat, M.M.; Raghavan, S.; Krishna, S. Spectrally adaptive infrared photodetectors with bias-tunable quantum dots. J. Opt. Soc. Amer. B 2004, 21, 7–17. Sensors 2013, 13 5086 83. Razeghi, M.; Erdtmann, M.; Jelen, C.; Guastavinos, F.; Brown, G.J.; Park, Y.S. Development of quantum well infrared photodetectors at the center for quantum devices. Infrared Phys. Technol. 2001, 42, 135–148. 83. Razeghi, M.; Erdtmann, M.; Jelen, C.; Guastavinos, F.; Brown, G.J.; Park, Y.S. Sensors 2013, 13 Investigation of multi-color, broadband quantum well infrared photodetectors with digital graded superlattice barrier and linear-graded barrier for long wavelength infrared applications. Infrared Phys. Technol. 2001, 42, 123–134. 96. Alves, F.D.P.; Santos, R.A.T.; Nohra, L.F.M.; Magalhaes, L.B.; Karunasiri, G. Tunable NIR Quantum Well Infrared Photodetector Using Interband Transitions. In Proceedings of the 2009 SBMO/IEEE MTT-S International Microwave and Optoelectronics Conference (IMOC), Belem, Brazil, 3–6 November 2009; pp. 303–306. Sensors 2013, 13 5087 97. Rehm, R.; Schneider, H.; Schwarz, K.; Walther, M.; Koidl, P.; Weimann, G. Responsivity and gain of InGaAs/GaAs-QWIPs and GaAs/AlGaAs-QWIPS: A comparative study. Proc. SPIE 2001, 4288, 379–385. 97. Rehm, R.; Schneider, H.; Schwarz, K.; Walther, M.; Koidl, P.; Weimann, G. Responsivity and gain of InGaAs/GaAs-QWIPs and GaAs/AlGaAs-QWIPS: A comparative study. Proc. SPIE 2001, 4288, 379–385. 98. Budkin, I.V.; Bulaev, P.V.; Vacilevskay, L.M.; Zalevsky, I.D.; Kuznetsov, U.A.; Kulikov, V.B.; Marmalyuk, A.A.; Nikitin, D.B.; Padalitsa, A.A.; Petrovsky, A.V.; et al. 3–5 um InGaAs/AlGaAs QWIP heterostructure growth by metal organic chemical vapor deposition (IR photodetector applications). Proc. SPIE 2007, 5126, 178–180. 99. Bahir, G.; Gusakov, Y.; Maimon, S.; Cohen, G.; Ritter, D.; Finkman, E. InP based QWIPs for long- and mid-wavelength band detection. Proc. SPIE 2003, 4820, 663–670. 100. Arora, B.M.; Majumdar, A.; Shah, A.P.; Gokhale, M.R.; Ghosh, S.; Bhattacharya, A.; Sengupta, D. Characteristics of high responsivity 8.5 um InGaAs/InP QWIPs grown by metalorganic vapour phase epitaxy. Infrared Phys. Technol. 2007, 50, 206–210. 101. Zhang, D.H.; Sun, L.; Yuan, K.H.; Yoon, S.F.; Radhakrishnan, K. InGaAsP/InP long wavelength quantum well infrared photodetectors. Thin Solid Films 2007, 515, 4450–4453. 102. Liu, H.C.; Song, C.Y.; Shen, A.; Gao, M.; Dupont, E.; Poole, P.J.; Wasilewski, Z.R.; Buchanan, M.; Wilson, P.H.; Robinson, B.J.; et al. Dual-band photodetectors based on interband and intersubband transitions. Infrared Phys. Technol. 2001, 42, 163–170. 103. Razeghi, M.; Jiang, J.; Jelen, C.; Brown, G.J. GaInAs/InP quantum-well infrared photodetectors grown on Si substrates. Proc. SPIE 2001, 4454, 78–84. 104. Jiang, J.; Jelen, C.; Razeghi, M.; Brown, G.J. High detectivity GaInAs-InP quantum-well infrared photodetectors grown on Si substrates. IEEE Photon. Technol. Lett. 2002, 14, 372–374. 105. Fu, L.; Tan, H.H.; Jagadish, C.; Li, N.; Li, N.; Liu, X.; Lu, W.; Shen, S.C. Tuning of detection wavelength of quantum-well infrared photodetectors by quantum-well intermixing. Infrared Phys. Technol. 2001, 42, 171–175. 106. Hsu, M.C.; Kuan, C.H.; Wang, S.Y. Multicolor infrared detection using two stacks of superlattice structures in a back-to-back configuration. Appl. Phys. Lett. Sensors 2013, 13 2000, 77, 3. 107. Oogarah, T.; Liu, H.C.; Dupont, E.; Wasilewski, Z.R.; Byloos, M.; Buchanan, M. High absorption GaAs/AlGaAs quantum well infrared photodetectors. Semicond. Sci. Technol. 2002, 17, 41–43. 108. Sim, S.K.H.; Liu, H.C.; Shen, A.; Gao, M.; Lee, K.F.; Buchanan, M.; Ohno, Y.; Ohno, H.; Li, E.H. Effect of barrier width on the performance of quantum well infrared photodetector. Infrared Phys. Technol. 2001, 42, 115–121. 109. Lee, J.-H.; Chang, C.-Y.; Li, C.-H.; Lin, S.-Y.; Lee, S.-C. Performance improvement of AlGaAs/GaAs QWIP by NH3 plasma treatment. IEEE J. Quantum Electron. 2012, 48, 922–926. 110. Liu, H.C.; Song, C.Y.; Spring Thorpe, A.J.; Cao, J.C. Terahertz quantum-well photodetector. Appl. Phys. Lett. 2004, 84, 4068–4070. 111. Mei, T.; Li, H.; Karunasiri, G.; Fan, W.J.; Zhang, D.H.; Yoon, S.F.; Yuan, K.H. Normal incidence silicon doped p-type GaAs/AlGaAs quantum-well infrared photodetector on (111) A substrate. Infrared Phys. Technol. 2007, 50, 119–123. 112. Graf, M.; Dupont, E.; Luo, H.; Haffouz, S.; Wasilewski, Z.R.; Thorpe, A.J.S.; Ban, D.; Liu, H.C. Terahertz quantum well infrared detectors. Infrared Phys. Technol. 2009, 52, 289–293. Sensors 2013, 13 5088 113. Ferrini, R.; Guizzetti, G.; Patrini, M.; Franchi, S. Optical study of intersubband transitions in GaSb/AlGaSb systems for QWIPs. Opt. Mater. 2001, 17, 351–354. 114. Chen, C.C. Performance and application of a superlattice infrared photodetector with a blocking barrier. J. Appl. Phys. 2002, 91, 6. 115. Costard, E.; Nedelcu, A.; Marcadet, X.; Belhaire, E.; Bois, P. QWIP development status at thales research and technology. Proc. SPIE 2006, 6206, 62060–62061. 116. Gueriaux, V.; Nedelcu, A.; Carras, M.; Huet, O.; Marcadet, X.; Bois, P. Mid-wave QWIPs for the (3–4.2 m) atmospheric window. Infrared Phys. Technol. 2009, 52, 235–240. 117. Cellek, O.O.; Ozer, S.; Besikci, C. High responsivity InP-InGaAs quantum-well infrared photodetectors: Characteristics and focal plane array performance. IEEE J. Quantum Electron. 2005, 41, 980–985. 118. Eker, S.U.; Arslan, Y.; Onuk, A.E.; Besikci, C. High conversion efficiency InP/InGaAs strained quantum well infrared photodetector focal plane array with 9.7 um cut-off for high-speed thermal imaging. IEEE J. Quantum Electron. 2010, 46, 164–168. 119. Gunapala, S.D.; Bandara, S.V; Singh, A.; Liu, J.K.; Rafol, B.; Luong, E.M.; Mumolo, J.M.; Tran, N.Q.; Ting, D.Z.-Y.; Vincent, J.D.; et al. 640 × 486 long-wavelength two-color GaAs/AlGaAs quantum well infrared photodetector (QWIP) focal plane array camera. IEEE Trans. Electron. Devices 2000, 47, 963–971. 120. Gunapala, S.D.; Bandara, S.V; Singh, A.; Liu, J.K.; Luong, E.M.; Mumolo, J.M.; LeVan, P.D. Sensors 2013, 13 Recent developments and applications of quantum well infrared photodetector focal plane arrays. Physica E 2000, 7, 108–111. 121. Bandara, S.V; Gunapala, S.D.; Liu, J.K.; Rafol, S.B.; Ting, D.Z.; Mumolo, J.M.; Reininger, F.M.; Fastenau, J.M.; Liu, A.K. Large format, broadband and multicolor GaAs/AlGaAs quantum well infrared photodetector (QWIP) focal plane arrays. Proc. SPIE 2001, 4454, 30–39. 122. Rafol, S.B.; Gunapala, S.D.; Bandara, S.V; Liu, J.K.; Ting, D.Z.; Mumolo, J.M.; Reininger, F.M. Characterization of QWIP (10–16 um) broadband FPA. Proc. SPIE 2003, 4820, 273–281. 123. Schneider, H.; Fleissner, J.; Rehm, R.; Walther, M.; Pletschen, W.; Koidl, P.; Weimann, G.; Ziegler, J.; Breiter, R.; Cabanski, W. High-resolution QWIP FPAs for the 8–12 um and 3–5 um regimes. Proc. SPIE 2003, 4820, 297–305. 124. Jhabvala, M.; Choi, K.; Goldberg, A.C.; La, A.T.; Gunapala, S.D. Development of a 1K × 1K GaAs QWIP far IR imaging array. Proc. SPIE 2004, 5167, 175–185. 125. Ozer, S.; Tumkaya, U.; Besikci, C. Large format AlInAs-InGaAs quantum-well infrared photodetector focal plane array for midwavelength infrared thermal imaging. IEEE Photon. Technol. Lett. 2007, 19, 1371–1373. 126. Jhabvala, M.; Choi, K.K.; Monroy, C.; La, A.; Adams, J.; Devitt, J.; Forrai, D.; Endres, D. Development of a 1 megapixel long IR QWIP focal plane array. Proc. SPIE 2007, 6542, 65420–65421. 127. Gunapala, S.D.; Bandara, S.V; Liu, J.K.; Mumolo, J.M.; Hill, C.J.; Ting, D.Z.; Kurth, E.; Woolaway, J.; LeVan, P.D.; Tidrow, M.Z. Development of megapixel dual-band QWIP focal plane array. Proc. SPIE 2008, 6940, 69401–69402. 128. Arslan, Y.; Çolakoglu, T.; Torunoglu, G.; Aktas, O.; Besikci, C. Enhanced performance QWIP FPAs. Infrared Phys. Technol. 2012, in press. 5089 Sensors 2013, 13 129. Hinds, S.; Buchanan, M.; Dudek, R.; Haffouz, S.; Laframboise, S.; Wasilewski, Z.; Liu, H.C. Near-room-temperature mid-infrared quantum well photodetector. Adv. Mater. 2011, 23, 5536–5539. 130. Bulman, G.E.; Myers, D.R.; Wiczer, J.J.; Dawson, L.R.; Biefeld, R.M.; Zipperian, T.E. Ion implanted lateral In(.2)Ga(.8)As/GaAs strained-layer superlattice photodetector. IEEE Electron. Devices Meet. 1984, 30, 719–722. 131. Nguyen, B.-M.; Chen, G.; Hoang, A.M.; Abdollahi Pour, S.; Bogdanov, S.; Razeghi, M. Effect of contact doping in superlattice-based minority carrier unipolar detectors. Appl. Phys. Lett. 2011, 99, 033501. 132. Rehm, R.; Masur, M.; Schmitz, J.; Daumer, V.; Niemasz, J.; Vandervelde, T.; DeMeo, D.; Luppold, W.; Wauro, M.; Wörl, A.; et al. InAs/GaSb superlattice infrared detectors. Infrared Phys. Technol. 2013, in press. 133. Brown, G.J. Type-II InAs/GaInSb superlattices for infrared detection: An overview. Proc. SPIE 2005, 5783, 65–77. 134. Sensors 2013, 13 Sensors 2013, 13 5090 144. Plis, E.; Lee, S.J.; Zhu, Z.; Amtout, A.; Krishna, S. InAs/GaSb superlattice detectors operating at room temperature. IEEE J. Sel. Top. Quantum Electron. 2006, 12, 1269–1274. 145. Zhang, Y.; Ma, W.; Cao, Y.; Huang, J.; Wei, Y.; Cui, K.; Shao, J. Long wavelength infrared InAs/GaSb superlattice photodetectors with InSb-Like and mixed interfaces. IEEE J. Quantum Electron. 2011, 47, 1475–1479. 146. Plis, E.; Rodriguez, J.B.; Balakrishnan, G.; Sharma, Y.D.; Kim, H.S.; Rotter, T.; Krishna, S. Mid-infrared InAs/GaSb strained layer superlattice detectors with NBn design grown on a GaAs substrate. Semicond. Sci. Technol. 2010, doi:10.1088/0268-1242/25/8/085010. 147. Haugan, H.J.; Brown, G.J.; Grazulis, L.; Mahalingam, K.; Tomich, D.H. Optimization of InAs/GaSb type-II superlattices for high performance of photodetectors. Physica E 2004, 20, 527–530. 148. Mohseni, H.; Tahraoui, A.; Wojkowski, J.S.; Razeghi, M.; Mitchel, W.; Saxler, A. Growth and characterization of very long wavelength type-II infrared detectors. Proc. SPIE 2000, 3948, 145–152. 149. Mohseni, H.; Wojkowski, J.S.; Tahraoui, A.; Razeghi, M.; Brown, G.; Mitchel, W. Growth and characterization of type-II nonequilibrium photovoltaic detectors for long wavelength infrared range. Proc. SPIE 2000, 3948, 153–160. 150. Razeghi, M.; Wei, Y.; Gin, A.; Brown, G.J.; Johnstone, D. Type II InAs/GaSb superlattices and detectors with lambda c >18 um. Proc. SPIE 2002, 4650, 111–116. 151. Wei, Y.; Gin, A.; Razeghi, M.; Brown, G.J. Type II InAs/GaSb superlattice photovoltaic detectors with cutoff wavelength approaching 32 um. Appl. Phys. Lett. 2002, 81, 3675–3677. 152. Wong, A.F.; Nelson, M.J.; Plis, E.A.; Skrutskie, M.F.; Yao, L.; Vandervelde, T.; Krishna, S.; Kim, H.; Khoshakhlagh, A.; Myers, S.A. Characterization of multicolor type-II InAs/GaSb strained-layer superlattice photodetectors for use in astronomical observation. Proc. SPIE 2010, doi:10.1117/12.857786. 153. Zhang, Y.; Ma, W.; Wei, Y.; Cao, Y.; Huang, J.; Cui, K.; Guo, X. Narrow-band long-/very-long wavelength two-color type-II InAs/GaSb superlattice photodetector by changing the bias polarity. Appl. Phys. Lett. 2012, doi:10.1063/1.4707162. 154. Razeghi, M.; Wei, Y.; Bae, J.; Gin, A.; Hood, A.; Jiang, J.; Nah, J. Type II InAs/GaSb superlattices for high-performance photodiodes and FPAs. Proc. SPIE 2003, 5246, 501–511. 155. Wei, Y.; Hood, A.; Gin, A.; Yazdanpanah, V.; Razeghi, M.; Tidrow, M. High performance LWIR Type II InAs/GaSb superlattice photodetectors and infrared focal plane arrays. Proc. SPIE 2005, 5732, 309–315. 156. Rodriguez, J.B.; Christol, P.; Chevrier, F.; Joullié, A. Optical characterization of symmetric InAs/GaSb superlattices for detection in the 3–5 μm spectral region. Physica E 2005, 28, 128–133. 157. Sensors 2013, 13 Haugan, H.J.; Szmulowicz, F.; Brown, G.J.; Ullrich, B.; Munshi, S.R.; Wickett, J.C.; Stokes, D.W. Short-period InAs/GaSb superlattices for mid-infrared photodetectors. Physica status solidi 2007, 4, 1702–1706. 135. Nathan, V.; Anselm, A.; Lin, C.H.; Johnson, J. InAs/InGaSb type-2 strained layer superlattice IR detectors. Proc. SPIE 2002, 4650, 185–198. 136. Khoshakhlagh, A.; Plis, E.; Myers, S.; Sharma, Y.D.; Dawson, L.R.; Krishna, S. Optimization of InAs/GaSb type-II superlattice interfaces for long-wave (~8 mm) infrared detection. J. Cryst. Growth 2009, 311, 1901–1904. 137. Lackner, D.; Pitts, O.J.; Najmi, S.; Sandhu, P.; Kavanagh, K.L.; Yang, A.; Steger, M.; Thewalt, M.L.W.; Wang, Y.; McComb, D.W.; et al. Growth of InAsSb/InAs MQWs on GaSb for mid-IR photodetector applications. J. Cryst. Growth 2009, 311, 3563–3567. 138. Lackner, D.; Pitts, O.J.; Steger, M.; Yang, A.; Thewalt, M.L.W.; Watkins, S.P. Strain balanced InAs/InAsSb superlattice structures with optical emission to 10 m. Appl. Phys. Lett. 2009, 95, 081906. 139. Lackner, D.; Steger, M.; Thewalt, M.L.W.; Pitts, O.J.; Cherng, Y.T.; Watkins, S.P.; Plis, E.; Krishna, S. InAs/InAsSb strain balanced superlattices for optical detectors: Material properties and energy band simulations. J. Appl. Phys. 2012, 111, 034507. 140. Rodriguez, J.B.; Plis, E.; Bishop, G.; Sharma, Y.D.; Kim, H.; Dawson, L.R.; Krishna, S. NBn structure based on InAs/GaSb type-II strained layer superlattices. Appl. Phys. Lett. 2007, doi:10.1063/1.2760153. 141. Lu, J.-H.; Wu, K.-J.; Hsieh, K.-J.; Kuan, C.-H.; Feng, J.-Y.; Lay, T.-S.; Yang, C.-W.; Tu, S.-L. A superlattice infrared photodetector integrated with multiple quantum wells to improve the performance. IEEE J. Quantum Electron. 2007, 43, 72–77. 142. Vurgaftman, I.; Canedy, C.L.; Jackson, E.M.; Nolde, J.A.; Affouda, C.A.; Aifer, E.H.; Meyer, J.R.; Hood, A.; Evans, A.J.; Tennant, W.T. Analysis and performance of type-II superlattice infrared photodetectors. Opt. Eng. 2011, 50, 061007. 143. Brown, G.J.; Szmulowicz, F. InAs/InGaSb superlattices for very long wavelength infrared detection. Proc. SPIE 2001, 4288, 200–208. Sensors 2013, 13 Wei, Y.; Hood, A.; Yau, H.; Gin, A.; Razeghi, M.; Tidrow, M.Z.; Nathan, V. Uncooled operation of type-II InAs/GaSb superlattice photodiodes in the midwavelength infrared range. Appl. Phys. Lett. 2005, 86, 1–3. 158. Zhu, Z.M.; Bhattacharya, P.; Plis, E.; Su, X.H.; Krishna, S. Low dark current InAs/GaSb type-II superlattice infrared photodetectors with resonant tunnelling filters. J. Phys. D 2006, 39, 4997–5001. 159. Haugan, H.J.; Brown, G.J.; Szmulowicz, F.; Grazulis, L.; Mitchel, W.C.; Elhamri, S.; Mitchell, W.D. InAs/GaSb type-II superlattices for high performance mid-infrared detectors. J. Cryst. Growth 2005, 278, 198–202. 5091 Sensors 2013, 13 160. Haugan, H.J.; Brown, G.J.; Mahalingam, K.; Elhamri, S.; Mitchel, W.C.; Grazulis, L.; Shank, J.M.; Houston, S. Growth of short-period InAs/GaSb superlattices for mid-infrared photodetectors. Proc. SPIE 2006, doi:10.1117/12.680530. 161. Rodriguez, J.B.; Plis, E.; Lee, S.J.; Kim, H.; Bishop, G.; Sharma, Y.D.; Dawson, L.R.; Krishna, S.; Jones, C. Type-II InAs/GaSb strain layer superlattice detectors for higher operating temperatures. Proc. SPIE 2007, 6542, 654201–654208. 162. Delaunay, P.-Y.; Nguyen, B.M.; Hoffman, D.; Razeghi, M. 320 × 256 infrared focal plane array based on type II InAs/GaSb superlattice with a 12 um cutoff wavelength. Proc. SPIE 2007, 6542, 654204. 163. Plis, E.; Kim, H.S.; Rodriguez, J.B.; Bishop, G.D.; Sharma, Y.D.; Khoshakhlagh, A.; Dawson, L.R.; Bundas, J.; Cook, R.; Burrows, D.; et al. NBn based infrared detectors using type-II InAs/(In,Ga)Sb superlattices. Proc. SPIE 2008, 6940, 69400–69401. 164. Huang, E.K.-W.; Delaunay, P.-Y.; Nguyen, B.-M.; Pour, S.A.; Razeghi, M. Photovoltaic MWIR type-II superlattice focal plane array on GaAs substrate. IEEE J. Quantum Electron. 2010, 46, 1704–1708. 165. Ting, D.Z.; Soibel, A.; Nguyen, J.; Hoglund, L.; Khoshakhlagh, A.; Rafol, S.B.; Keo, S.A.; Liao, A.; Mumolo, J.M.; Liu, J.K.; et al. Type II superlattice barrier infrared detector. Proc. SPIE 2011, doi:10.1117/12.896240. 166. Guo, J.; Peng, Z.; Sun, W.; Xu, Y.; Zhou, Z.; Niu, Z. InAs/GaSb superlattices for photodetection in short wavelength infrared range. Infrared Phys. Technol. 2009, 52, 124–126. 167. Chen, J.; Xu, Q.; Zhou, Y.; Jing, J.; Lin, C.; He, L. Growth and fabrication of InAs/GaSb type II superlattice mid-wavelength infrared photodetectors. Nanoscale Res. Lett. 2011, 6, 1–14. 168. Jianliang, H.; Ma, W.; Wei, Y.; Zhang, Y.; Cui, K.; Cao, Y.; Guo, X.; Shao, J. How to use type II InAs/GaSb superlattice structure to reach detection wavelength of 2–3 µm. IEEE J. Quantum Electron. 2012, 48, 1322–1326. 169. Salihoglu, O.; Muti, A.; Kutluer, K.; Tansel, T.; Turan, R.; Kocabas, C.; Aydinli, A. Sensors 2013, 13 Sensors 2013, 13 5092 175. Razeghi, M.; Nguyen, B.-M. Band gap tunability of Type II Antimonide-based superlattices. Phys. Procedia 2010, 3, 1207–1212. 175. Razeghi, M.; Nguyen, B.-M. Band gap tunability of Type II Antimonide-based superlattices. Phys. Procedia 2010, 3, 1207–1212. 176. Nguyen, B.-M.; Razeghi, M.; Nathan, V.; Brown, G.J. Type-II “M” structure photodiodes: An alternative material design for mid-wave to long wavelength infrared regimes. Proc. SPIE 2007, 6479, 64790S. 177. Nguyen, B.-M.; Hoffman, D.; Delaunay, P.-Y.; Razeghi, M. Dark current suppression in type II InAsGaSb superlattice long wavelength infrared photodiodes with M-structure barrier. Appl. Phys. Lett. 2007, doi:10.1063/1.2800808. 178. Razeghi, M.; Pour, S.A.; Huang, E.K.-W.; Chen, G.; Haddadi, A.; Nguyen, B.-M.B.-M.; Delaunay, P.-Y.; Abdollahi Pour, S. High-operating temperature MWIR photon detectors based on Type II InAs/GaSb superlattice. Proc. SPIE 2011, 8012, 80122Q. 179. Savich, G.R.; Pedrazzani, J.R.; Maimon, S.; Wicks, G.W. Use of epitaxial unipolar barriers to block surface leakage currents in photodetectors. Physica Status Solidi C 2010, 7, 2540–2543. 180. Savich, G.R.; Pedrazzani, J.R.; Sidor, D.E.; Maimon, S.; Wicks, G.W. Dark current filtering in unipolar barrier infrared detectors. Appl. Phys. Lett. 2011, 99, 121112. 181. DeMeo, D.F.; Vandervelde, T.E. Simulations of Gallium Antimonide (GaSb) p-B-n Thermophotovoltaic Cells. MRS Proc. 2011, doi:10.1557/opl.2011.1469. 182. Savich, G.R.; Pedrazzani, J.R.; Sidor, D.E.; Wicks, G.W. Benefits and limitations of unipolar barriers in infrared photodetectors. Infrared Phys. Technol. 2013, in press. 183. Khoshakhlagh, A.; Rodriguez, J.B.; Plis, E.; Bishop, G.D.; Sharma, Y.D.; Kim, H.S.; Dawson, L.R.; Krishna, S. Bias dependent dual band response from InAsGa (In) Sb type II strain layer superlattice detectors. Appl. Phys. Lett. 2007, 91, 263504. 184. Nguyen, B.-M.; Bogdanov, S.; Pour, S.A.; Razeghi, M. Minority electron unipolar photodetectors based on type II InAs/GaSb/AlSb superlattices for very long wavelength infrared detection. Appl. Phys. Lett. 2009, 95, 183502. 185. Khoshakhlagh, A.; Myers, S.; Kim, H.; Plis, E.; Gautam, N.; Lee, S.J.; Noh, S.K.; Dawson, L.R.; Krishna, S. Long-wave InAs/GaSb superlattice detectors based on nBn and pin designs. IEEE J. Quantum Electron. 2010, 46, 959–964. 186. Wicks, G.W.; Savich, G.R.; Pedrazzani, J.R.; Maimon, S. Infrared detector epitaxial designs for suppression of surface leakage current. Proc. SPIE 2010, 7608, 760822. 187. Klipstein, P.; Klin, O.; Grossman, S.; Snapi, N.; Lukomsky, I.; Yassen, M.; Aronov, D.; Berkowitz, E.; Glozman, A.; Magen, O.; et al. High operating temperature XBn-InAsSb bariode detectors. Proc. SPIE 2012, doi:10.1117/12.910174. 188. Klipstein, P. Sensors 2013, 13 Atomic layer deposited Al2O3 passivation of type II InAs/GaSb superlattice photodetectors. J. Appl. Phys. 2012, doi:10.1063/1.3702567. 170. Brown, G.J.; Szmulowicz, F.; Linville, R.; Saxler, A.; Mahalingham, K.; Lin, C.-H.; Kuo, C.H.; Hwang, W.Y. Type-II superlattice photodetector on a compliant GaAs substrate. IEEE Photon. Technol. Lett. 2000, 12, 684–686. 171. Anselm, K.A.; Ren, H.; Vilela, M.; Zheng, J.; Lin, C.H.; Nathan, V.; Brown, G. Long-wavelength infrared InAs/InGaSb type-II superlattice photovoltaic detectors. Proc. SPIE 2001, 4288, 183–190. 172. Patrashin, M.; Hosako, I.; Akahane, K. Type-II InAs/GaInSb superlattices for terahertz range photodetectors. Proc. SPIE 2011, doi:10.1117/12.898759. 173. Fuchs, F.; Burkle, L.; Hamid, R.; Herres, N.; Pletschen, W.; Sah, R.E.; Kiefer, R.; Schmitz, J. Optoelectronic properties of photodiodes for the mid- and far-infrared based on the InAs/GaSb/AlSb materials family. Proc. SPIE 2001, 4288, 171–182. 174. Aifer, E.H.; Tischler, J.G.; Warner, J.H.; Vurgaftman, I.; Bewley, W.W.; Meyer, J.R.; Kim, J.C.; Whitman, L.J.; Canedy, C.L.; Jackson, E.M. W-structured type-II superlattice long-wave infrared photodiodes with high quantum efficiency. Appl. Phys. Lett. 2006, doi:10.1063/1.2335509. Sensors 2013, 13 “XBn” barrier photodetectors for high sensitivity and high operating temperature infrared sensors. Proc. SPIE 2008, doi:10.1117/12.778848. 189. Kim, H.S.; Cellek, O.O.; Lin, Z.-Y.; He, Z.-Y.; Zhao, X.-H.; Liu, S.; Li, H.; Zhang, Y.-H. Long-wave infrared nBn photodetectors based on InAs/InAsSb type-II superlattices. Appl. Phys. Lett. 2012, 101, 161114. 190. Arakawa, Y.; Sakaki, H. Multidimensional quantum well laser and temperature dependence of its threshold current. Appl. Phys. Lett. 1982, 40, 939–941. 5093 Sensors 2013, 13 191. Leonard, D.; Krishnamurthy, M.; Reaves, C.M.; Denbaars, S.P.; Petroff, P.M. Direct formation of quantum-sized dots from uniform coherent islands of InGaAs on GaAs surfaces. Appl. Phys. Lett. 1993, 63, 3203–3205. 191. Leonard, D.; Krishnamurthy, M.; Reaves, C.M.; Denbaars, S.P.; Petroff, P.M. Direct formation of quantum-sized dots from uniform coherent islands of InGaAs on GaAs surfaces. Appl. Phys. Lett. 1993, 63, 3203–3205. 192. Kim, S.M. Review on recent development of quantum dots: From optoelectronic devices to novel biosensing applications. Proc. SPIE 2003, 4999, 1117. 193. Asano, T.; Chong, H.; Yi, Z.; Mingguo, L.; Campbell, J.C.; Madhukar, A. Design consideration and demonstration of resonant-cavity-enhanced quantum dot infrared photodetectors in mid-infrared wavelength regime. IEEE J. Quantum Electron. 2010, 46, 1484–1491. 194. Kubis, A.J.; Vandervelde, T.E.; Bean, J.C.; Dunn, D.N.; Hull, R. Analysis of the three-dimensional ordering of epitaxial Ge quantum dots using focused ion beam tomography. Appl. Phys. Lett. 2006, 88, 263101–263103. 195. Vandervelde, T.E.; Sun, K.; Merz, J.L.; Kubis, A.; Hull, R.; Pernell, T.L.; Bean, J.C. The effect of two-temperature capping on germanium/silicon quantum dots and analysis of superlattices so composed. J. Appl. Phys. 2006, 99, 124301. 196. Vines, P.; Tan, C.H.; David, J.P.R.; Attaluri, R.S.; Vandervelde, T.E.; Krishna, S. Multiple stack quantum dot infrared photodetectors. Proc. SPIE 2008, doi:10.1117/12.802247. 197. Ting, D.Z.-Y.; Bandara, S.V; Gunapala, S.D.; Mumolo, J.M.; Keo, S.A.; Hill, C.J.; Liu, J.K.; Blazejewski, E.R.; Rafol, S.B.; Chang, Y.-C. Submonolayer quantum dot infrared photodetector. Appl. Phys. Lett. 2009, 94, 111107. 198. Shao, J.; Vandervelde, T.E.; Barve, A.; Jang, W.; Stintz, A.; Krishna, S. Barrier engineered superlattice and quantum dot detectors for HOT operation. Proc. SPIE 2011, doi:10.1117/12.876406. 199. Vines, P.; Tan, C.H.; David, J.P.R.; Attaluri, R.S.; Vandervelde, T.E.; Krishna, S.; Jang, W.-Y.; Hayat, M.M. Quantum dot infrared photodetectors with highly tunable spectral response for an algorithm-based spectrometer. Proc. SPIE 2010, doi:10.1117/12.852650. 200. Benisty, H.; Sotomayor-Torres, C.M.; Weisbuch, C. Intrinsic mechanism for the poor luminescence properties of quantum-box systems. Phys. Rev. B 1991, 44, 10945–10948. 201. Sensors 2013, 13 Krishna, S.; Raghavan, S.; Von Winckel, G.; Rotella, P.; Stintz, A.; Morath, C.P.; Le, D.; Kennerly, S.W. Two color InAs/InGaAs dots-in-a-well detector with background-limited performance at 91 K. Appl. Phys. Lett. 2003, 82, 2574–2576. 202. Ryzhii, V.; Khmyrova, I.; Mitin, V.; Stroscio, M.; Willander, M. On the detectivity of quantum- dot infrared photodetectors. Appl. Phys. Lett. 2001, doi:10.1063/1.1376435. 203. Chen, Z.H.; Baklenov, O.; Kim, E.T.; Mukhametzhanov, I.; Tie, J.; Madhukar, A.; Ye, Z.; Campbell, J.C. InAs/AlxGa1-xAs quantum dot infrared photodetectors with undoped active region. Infrared Phys. Technol. 2001, 42, 479–484. 204. Lo, M.C.; Wang, S.Y.; Ling, H.S.; Lee, C.P. Vertically coupled quantum-dot infrared photodetectors. IEEE Photon. Technol. Lett. 2010, 22, 796–798. 205. Lin, W.-H.; Lin, S.-Y. Broadband InGaAs-capped InAs/GaAs quantum-dot infrared photodetector with Bi-modal dot height distributions. J. Appl. Phys. 2012, 112, 034508. 206. Stiff-Roberts, A.D.; Chakrabarti, S.; Pradhan, S.; Kochman, B.; Bhattacharya, P. Raster-scan imaging with normal-incidence, midinfrared InAs/GaAs quantum dot infrared photodetectors. Appl. Phys. Lett. 2002, 80, 3265–3267. 5094 Sensors 2013, 13 207. Sengupta, S.; Kim, J.O.; Barve, A.V.; Adhikary, S.; Sharma, Y. D.; Gautam, N.; Lee, S.J.; Noh, S.K.; Chakrabarti, S.; Krishna, S. Sub-monolayer quantum dots in confinement enhanced dots-in-a-well heterostructure. Appl. Phys. Lett. 2012, 100, 191111. 207. Sengupta, S.; Kim, J.O.; Barve, A.V.; Adhikary, S.; Sharma, Y. D.; Gautam, N.; Lee, S.J.; Noh, S.K.; Chakrabarti, S.; Krishna, S. Sub-monolayer quantum dots in confinement enhanced dots-in-a-well heterostructure. Appl. Phys. Lett. 2012, 100, 191111. 208. Chakrabarti, S.; Stiff-Roberts, A.D.; Bhattacharya, P.; Gunapala, S.; Bandara, S.; Rafol, S.B.; Kennerly, S.W. High-temperature operation of InAs-GaAs quantum-dot infrared photodetectors with large responsivity and detectivity. IEEE Photon. Technol. Lett. 2004, 16, 1361–1363. 209. Aivaliotis, P.; Wilson, L.R.; Zibik, E.A.; Cockburn, J.W.; Steer, M.J.; Liu, H.Y. Enhancing the dot density in quantum dot infrared photodetectors via the incorporation of antimony. Appl. Phys. Lett. 2007, doi:10.1063/1.2753727. 210. Tang, S.-F.; Chiang, C.-D.; Weng, P.-K.; Gau, Y.-T.; Luo, J.-J.; Yang, S.-T.; Shih, C.-C.; Lin, S.-Y.; Lee, S.-C. High-temperature operation normal incident 256 × 256 InAs-GaAs quantum-dot infrared photodetector focal plane array. IEEE Photon. Technol. Lett. 2006, 18, 986–988. 211. Drozdowicz-Tomsia, K.; Goldys, E.M.; Fu, L.; Jagadish, C. Doping effect on dark currents in In(0.5)Ga(0.5)As/GaAs quantum dot infrared photodetectors grown by metal-organic chemical vapor deposition. Appl. Phys. Lett. 2006, 89, 1023–1025. 212. Fu, L.; Li, Q.; Kuffner, P.; Jolley, G.; Gareso, P.; Tan, H.H.; Jagadish, C. Two-color InGaAs/GaAs quantum dot infrared photodetectors by selective area interdiffusion. Appl. Sensors 2013, 13 Sensors 2013, 13 220. Tsao, S.; Lim, H.; Zhang, W.; Razeghi, M. High-performance InAs quantum-dot infrared photodetectors grown on InP substrate operating at room temperature and high operating temperature focal plane array. Proc. SPIE 2007, 6593, 65931. 220. Tsao, S.; Lim, H.; Zhang, W.; Razeghi, M. High-performance InAs quantum-dot infrared photodetectors grown on InP substrate operating at room temperature and high operating temperature focal plane array. Proc. SPIE 2007, 6593, 65931. 221. Lu, X.; Vaillancourt, J.; Meisner, M.J. Temperature-dependent photoresponsivity and high-temperature (190 K) operation of a quantum dot infrared photodetector. Appl. Phys. Lett. 2007, doi:10.1063/1.2766655. 222. Chen, T.-C.; Lai, J.-H. Investigation for optoelectronic characteristics and imaging performance of InAs quantum dot covered with In(0.1)Ga(0.9)As/GaAs multilayer based focal plane array. Opt. Eng. 2011, doi:10.1117/1.3572146. 223. Huffaker, D.L.; Mehta, M.; Balakrishnan, G.; Huang, S.; Khoshakhlagh, A.; Patel, P.; Kutty, M.N.; Dawson, L.R. GaSb QW-Based “Buffer-Free” Vertical LED Monolithically Embedded within a GaAs Cavity Using Interfacial Misfit Arrays. In Proceedings of the 19th Annual Meeting of the IEEE Lasers and Electro-Optics Society, Montreal, QC, Canada, 29 October–2 November 2006. 224. Jallipalli, A.; Kutty, M.N.; Balakrishnan, G.; Tatebayashi, J.; Nuntawong, N.; Huang, S.H.; Dawson, L.R.; Huffaker, D.L.; Mi, Z.; Bhattacharya, P. 1.54 m GaSb/AlGaSb multi-quantum- well monolithic laser at 77 K grown on miscut Si substrate using interfacial misfit arrays. Electron. Lett. 2007, 43, 1198–1199. 225. Laghumavarapu, R.B.; Moscho, A.; Khoshakhlagh, A.; El-Emawy, M.; Lester, L.F.; Huffaker, D.L. GaSbGaAs type II quantum dot solar cells for enhanced infrared spectral response. Appl. Phys. Lett. 2007, doi:10.1063/1.2734492. 226. Tatebayashi, J.; Jallipalli, A.; Kutty, M.N.; Huang, S.H.; Rotter, T.J.; Balakrishnan, G.; Dawson, L.R.; Huffaker, D.L. Device characteristics of GaInSb/AlGaSb quantum well lasers monolithically grown on GaAs substrates by using an interfacial misfit array. J. Electron. Mater. 2008, 37, 1758–1763. 227. Yarborough, J.M.; Lai, Y.-Y.; Kaneda, Y.; Hader, J.; Moloney, J.V; Rotter, T.J.; Balakrishnan, G.; Hains, C.; Huffaker, D.; Koch, S.W.; Bedford, R. Record pulsed power demonstration of a 2 um GaSb-based optically pumped semiconductor laser grown lattice-mismatched on an AlAs/GaAs Bragg mirror and substrate. Appl. Phys. Lett. 2009, 95, 081112. 228. Balakrishnan, G.; Tatebayashi, J.; Khoshakhlagh, A.; Huang, S.H.; Jallipalli, A.; Dawson, L.R.; Huffaker, D.L. III/V ratio based selectivity between strained Stranski-Krastanov and strain-free GaSb quantum dots on GaAs. Appl. Phys. Lett. 2006, doi:10.1063/1.2362999. 229. Tatebayashi, J.; Liang, B.; Bussian, D.A.; Htoon, H.; Huang, S.; Balakrishnan, G.; Klimov, V.; Dawson, L.R.; Huffaker, D.L. Sensors 2013, 13 Phys. Lett. 2008, 93, 13501–13504. 213. Chakrabarti, S.; Adhikary, S.; Halder, N.; Aytac, Y.; Perera, A.G.U. High-performance, long- wave (~10.2 um) InGaAs/GaAs quantum dot infrared photodetector with quaternary In0.21Al0.21Ga0.58As capping. Appl. Phys. Lett. 2011, 99, 181102. 214. Jiang, J.; Tsao, S.; O’Sullivan, T.; Zhang, W.; Lim, H.; Sills, T.; Mi, K.; Razeghi, M.; Brown, G.J.; Tidrow, M.Z. High detectivity InGaAs/InGaP quantum-dot infrared photodetectors grown by low pressure metalorganic chemical vapor deposition. Appl. Phys. Lett. 2004, 84, 2166–2168. 215. Jiang, J.; Mi, K.; Tsao, S.; Zhang, W.; Lim, H.; O’Sullivan, T.; Sills, T.; Razeghi, M.; Brown, G.J.; Tidrow, M.Z. Demonstration of a 256 × 256 middle-wavelength infrared focal plane array based on InGaAs/InGaP quantum dot infrared photodetectors. Appl. Phys. Lett. 2004, 84, 2232–2234. 216. Zhang, W.; Lim, H.-C.; Taguchi, M.; Tsao, S.; Szafraniec, J.; Movaghar, B.; Razeghi, M.; Tidrow, M. High performance InAs quantum dot infrared photodetectors (QDIP) on InP by MOCVD. Proc. SPIE 2005, 5732, 326–333. 217. Chakrabarti, S.; Adhikary, S.; Halder, N.; Aytac, Y.; Perera, A.G.U. Demonstration of high responsivity(~2.16 A/W) and detectivity(~10^11 Jones) in the long wavelength (~10.2 um) from InGaAs/GaAs quantum dot infrared photodetector with quaternary InAlGaAs capping. Proc. SPIE 2012, 8353, 835338. 218. Ouattara, L.; Mikkelsen, A.; Lundgren, E.; Hoglund, L.; Asplund, C.; Andersson, J.Y. A cross-sectional scanning tunneling microscopy study of a quantum dot infrared photodetector. J. Appl. Phys. 2005, doi:10.1063/1.2245195. 219. Lim, H.-C.; Tsao, S.; Zhang, W.; Razeghi, M. Self-assembled semiconductor quantum dot infrared photodetector operating at room temperature and focal plane array. Proc. SPIE 2007, 6542, 65420–65421. 5095 Sensors 2013, 13 Sensors 2013, 13 5096 232. Kim, J.-W.; Oh, J.-E.; Hong, S.-C.; Park, C.-H.; Yoo, T.-K. Room temperature far infrared (8~10 m) photodetectors using self-assembled InAs quantum dots with high detectivity. IEEE Electron. Device Lett. 2000, 21, 329–331. 233. Wang, S.Y. Low dark current quantum-dot infrared photodetectors with an AlGaAs current blocking layer. Appl. Phys. Lett. 2001, 78, 1023–1025. 234. Kim, E.-T.; Chen, Z.; Ho, M.; Madhukar, A. Tailoring mid- and long-wavelength dual response of InAs quantum-dot infrared photodetectors using In(x)Ga(1-x)As capping layers. J. Vac. Sci. Technol. B 2002, 20, 1188–1191. 235. Szafraniec, J.; Tsao, S.; Zhang, W.; Lim, H.; Taguchi, M.; Quivy, A.A.; Movaghar, B.; Razeghi, M. High-detectivity quantum-dot infrared photodetectors grown by metalorganic chemical-vapor deposition. Appl. Phys. Lett. 2006, doi:10.1063/1.2188056. 236. Vandervelde, T.E.; Lenz, M.C.; Varley, E.; Barve, A.; Shao, J.; Shenoi, R.V; Ramirez, D.A.; Jan, W.; Sharma, Y.D.; Krishna, S. Quantum dots-in-a-well focal plane arrays. IEEE J. Sel. Top. Quantum Electron. 2008, 14, 1150–1161. 237. Vandervelde, T.E.; Krishna, S. Progress and prospects for quantum dots in a well infrared photodetectors. J. Nanosci. Nanotechnol. 2010, 10, 1450–1460. 238. Ling, H.S.; Wang, S.Y.; Lee, C.P.; Lo, M.C. High quantum efficiency dots-in-a-well quantum dot infrared photodetectors with AlGaAs confinement enhancing layer. Appl. Phys. Lett. 2008, 92, doi:10.1063/1.2926663. 239. Liu, H.Y.; Alia, E. Optimizing the growth of 1.3 um InAs/InGaAs dots-in-a-well structure. J. Appl. Phys. 2003, 93, 2931–2936. 240. Barve, A.V; Sengupta, S.; Kim, J.O.; Montoya, J.; Klein, B.; Shirazi, M.A.; Zamiri, M.; Sharma, Y.D.; Adhikary, S.; Godoy, S.E.; et al. Barrier selection rules for quantum dots-in-a-well infrared photodetector. IEEE J. Quantum Electron. 2012, 48, 1243–1251. 241. Rotella, P.; Raghavan, S.; Stintz, A.; Fuchs, B.; Krishna, S.; Morath, C.; Le, D.; Kennerly, S.W. Normal incidence InAs/InGaAs dots-in-well detectors with current blocking AlGaAs layer. J. Cryst. Growth 2003, 251, 787–793. 242. Hoglund, L.; Asplund, C.; Wang, Q.; Almqvist, S.; Malm, H.; Petrini, E.; Andersson, J.Y.; Holtz, P.O.; Petersson, H. Origin of photocurrent in lateral quantum dots-in-a-well infrared photodetectors. Appl. Phys. Lett. 2006, doi:10.1063/1.2207493. 243. Ariyawansa, G.; Perera, A.G.U.; Raghavan, G.S.; von Winckel, G.; Stintz, A.; Krishna, S. Effect of well width on three-color quantum dots-in-a-well infrared detectors. IEEE Photon. Technol. Lett. 2005, 17, 1064–1066. 244. Aivaliotis, P.; Zibik, E.A.; Wilson, L.R.; Cockburn, J.W.; Hopkinson, M.; Airey, R.J. Tuning the photoresponse of quantum dot infrared photodetectors across the 8–12 um atmospheric window via rapid thermal annealing. Appl. Phys. Lett. 2007, doi:10.1063/1.2794014. 245. Sensors 2013, 13 Formation and optical characteristics of type-II strain-relieved GaSb/GaAs quantum dots by using an interfacial misfit growth mode. IEEE Trans. Nanotechnol. 2009, 8, 269–274. 230. Su, X.; Chakrabarti, S.; Bhattacharya, P.; Ariyawansa, G.; Perera, A.G.U. A resonant tunneling quantum-dot infrared photodetector. IEEE J. Quantum Electron. 2005, 41, 974–979. 231. Bhattacharya, P.; Su, X.H.; Chakrabarti, S.; Ariyawansa, G.; Perera, A.G.U. Characteristics of a tunneling quantum-dot infrared photodetector operating at room temperature. Appl. Phys. Lett. 2005, 86, 1–3. Sensors 2013, 13 Shao, J.; Vandervelde, T.E.; Jang, W.-Y.; Stintz, A.; Krishna, S. High operating temperature quantum-dot infrared photodetector using advanced capping techniques. IEEE Trans. Nanotechnol. 2011, 10, 1010–1014. 246. Shao, J.; Vandervelde, T.E.; Jang, W.-Y.; Stintz, A.; Krishna, S. High Operating Temperature InAs Quantum Dot Infrared Photodetector Via Selective Capping Techniques. In Proceedings of the 8th IEEE Conference on Nanotechnology, Arlington, TX, USA, 18–21 August 2008; pp. 112–115. 5097 Sensors 2013, 13 247. Shenoi, R.V; Attaluri, R.S.; Siroya, A.; Shao, J.; Sharma, Y.D.; Stintz, A.; Vandervelde, T.E.; Krishna, S. Low-strain InAsInGaAsGaAs quantum dots-in-a-well infrared photodetector. J. Vacu. Sci. Technol. B 2008, 26, 1136–1139. 248. Vines, P.; Tan, C.H.; David, J.P.R.; Attaluri, R.S.; Vandervelde, T.E.; Krishna, S. Noise, gain, and responsivity in low-strain quantum dot infrared photodetectors with up to 80 dot-in-a-well periods. IEEE J. Quantum Electron. 2011, 47, 607–613. 249. Shenoi, R.V; Hou, J.; Sharma, Y.; Shao, J.; Vandervelde, T.E.; Krishna, S. Low strain quantum dots in a double well infrared detector. Proc. SPIE 2008, 7082, 708207. 250. Andrews, J.; Restaino, S.R.; Vandervelde, T.E.; Brown, J.S.; Sharma, Y.D.; Lee, S.L.; Teare, S.W.; Reisinger, A.; Sundaram, M.; Krishna, S. Comparison of long-wave infrared quantum-dots-in-a- well and quantum-well focal plane arrays. IEEE Trans. Electron. Devices 2009, 56, 512–516. 251. Jang, W.-Y.; Hayat, M.M.; Tyo, J.S.; Attaluri, R.S.; Vandervelde, T.E.; Sharma, Y.D.; Shenoi, R.; Stintz, A.; Cantwell, E.R.; Bender, S.C.; et al. Demonstration of bias-controlled algorithmic tuning of quantum dots in a well (DWELL) MidIR detectors. IEEE J. Quantum Electron. 2009, 45, 674–683. 252. Shenoi, R.V; Rosenberg, J.; Vandervelde, T.E.; Painter, O.J.; Krishna, S. Multispectral quantum dots-in-a-well infrared detectors using plasmon assisted cavities. IEEE J. Quantum Electron. 2010, 46, 1051–1057. 253. Shao, J.; Vandervelde, T.E.; Barve, A.; Jang, W.-Y.; Stintz, A.; Krishna, S. Enhanced normal incidence photocurrent in quantum dot infrared photodetectors. J. Vacu. Sci. Technol. B 2011, 29, 3–123. 254. Vines, P.; Tan, C.H.; David, J.P.R.; Attaluri, R.S.; Vandervelde, T.E.; Krishna, S.; Jang, W.; Hayat, M.M. Versatile spectral imaging with an algorithm-based spectrometer using highly tuneable quantum dot infrared photodetectors. IEEE J. Quantum Electron. 2011, 47, 190–197. 255. Barve, A.V; Sharma, Y.D.; Montoya, J.; Shao, J.; Vandervelde, T.; Rotter, T.; Krishna, S. Engineering the barrier of quantum dots-in-a-well (DWELL) infrared photodetectors for high temperature operation. Int. J. High Speed Electron. Syst. 2011, 20, 549–555. 256. Andrews, J.R.; Restaino, S.R.; Teare, S.W.; Sharma, Y.D.; Jang, W. Sensors 2013, 13 Comparison of quantum dots-in-a-double-well and quantum dots-in-a-well focal plane arrays in the long-wave infrared. IEEE Trans. Electron. Devices 2011, 58, 2022–2027. 257. Vandervelde, T.E.; Lenz, M.C.; Varley, E.; Barve, A.; Shao, J.; Shenoi, R.; Ramirez, D.A.; Jang, W.; Sharma, Y.D.; Krishna, S. Multicolor quantum dots-in-a-well focal plane arrays. Proc. SPIE 2008, 6940, 694003. 258. Krishna, S. Barrier engineering in quantum dots in a well detector. Proc. SPIE 2012, doi:10.1117/12.909924. 259. Jolley, G.; Fu, L.; Tan, H.H.; Jagadish, C. Effects of well thickness on the spectral properties of In(0.5)Ga(0.5)As/GaAs/Al(0.2)Ga(0.8)As quantum dots-in-a-well infrared photodetectors. Appl. Phys. Lett. 2008, doi:10.1063/1.2927487. 260. Jolley, G.; Fu, L.; Tan, H.H.; Jagadish, C. Influence of quantum well and barrier composition on the spectral behavior of InGaAs quantum dots-in-a-well infrared photodetectors. Appl. Phys. Lett. 2007, doi:10.1063/1.2802559. 261. Barve, A.V; Shah, S.Y.; Shao, J.; Vandervelde, T.E.; Shenoi, R.V; Jang, W.-Y.; Krishna, S. Reduction in dark current using resonant tunneling barriers in quantum dots-in-a-well long wavelength infrared photodetector. Appl. Phys. Lett. 2008, 93, 131115. 5098 Sensors 2013, 13 262. Barve, A.; Shao, J.; Sharma, Y.D.; Vandervelde, T.E.; Sankalp, K.; Lee, S.J.; Noh, S.K.; Krishna, S. Resonant tunneling barriers in quantum dots-in-a-well infrared photodetectors. IEEE J. Quantum Electron. 2010, 46, 1105–1114. 262. Barve, A.; Shao, J.; Sharma, Y.D.; Vandervelde, T.E.; Sankalp, K.; Lee, S.J.; Noh, S.K.; Krishna, S. Resonant tunneling barriers in quantum dots-in-a-well infrared photodetectors. IEEE J. Quantum Electron. 2010, 46, 1105–1114. 263. Rosenberg, J.; Shenoi, R.V; Vandervelde, T.E.; Krishna, S.; Painter, O. A multispectral and polarization-selective surface-plasmon resonant midinfrared detector. Appl. Phys. Lett. 2009, 95, 161101. 264. Gunapala, S.D.; Bandara, S.V; Hill, C.J.; Ting, D.Z.; Liu, J.K.; Rafol, S.B.; Blazejewski, E.R.; Mumolo, J.M.; Keo, S.A.; Krishna, S.; et al. 640 × 512 pixels long-wavelength infrared (LWIR) quantum-dot infrared photodetector (QDIP) imaging focal plane array. IEEE J. Quantum Electron. 2007, 43, 230–237. 265. Kozlowski, L.J. HgCdTe focal plane arrays for high performance infrared cameras. Proc. SPIE 1997, 3179, 200–211. 266. Mollard, L.; Destefanis, G.; Bourgeois, G.; Ferron, A.; Baier, N.; Gravrand, O.; Barnes, J.P.; Papon, A.; Milesi, F.; Kerlain, A.; Rubaldo, L. Status of p-on-n arsenic-implanted HgCdTe technologies. J. Electron. Mater. 2011, 40, 1830–1839. 267. Kozlowski, L.J.; Kosonocky, W.F. Chapter 23—Infrared Detector Arrays. In Handbook of Optics; McGraw-Hill: New York, USA, 1995; p. 1664. 268. Costard, E.; Truffer, J.P.; Huet, O.; Dua, L.; Nedelcu, A.; Robo, J.A.; Marcadet, X.; De l’Isle, N.B.; Bois, P. © 2013 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution license (http://creativecommons.org/licenses/by/3.0/). Sensors 2013, 13 QWIP from 4 um up to 18 um. Proc. SPIE 2006, 6361, 636111–636117. 279. Tang, S.F.; Lin, S.Y.; Lee, S.C. Near-room-temperature operation of an InAs/GaAs quantum-dot infrared photodetector. Appl. Phys. Lett. 2001, 78, 2428–2430. 270. Chakrabarti, S.; Stiff-Roberts, A.D.; Su, X.H.; Bhattacharya, P.; Ariyawansa, G.; Perera, A.G.U. High-performance mid-infrared quantum dot infrared photodetectors. J. Phys. D 2005, 38, 2135–2141. 271. Krishna, S.; Raghavan, S.; von Winckel, G.; Stintz, A.; Ariyawansa, G.; Matsik, S.G.; Perera, A.G.U. Three-color (λp1 ~ 3.8 μm, λp2 ~ 8.5 μm, and λp3 ~ 23.2 μm) InAs/InGaAs quantum-dots-in-a- well detector. Appl. Phys. Lett. 2003, doi:10.1063/1.1615838. 272. Shao, J.; Vandervelde, T.E.; Jang, W.; Stintz, A.; Krishna, S. Improving the operating temperature of quantum dots-in-a-well detectors. Proc. SPIE 2010, doi:10.1117/12.843676. 272. Shao, J.; Vandervelde, T.E.; Jang, W.; Stintz, A.; Krishna, S. Improving the operating temperature of quantum dots-in-a-well detectors. Proc. SPIE 2010, doi:10.1117/12.843676. © 2013 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution license (http://creativecommons.org/licenses/by/3.0/). © 2013 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution license (http://creativecommons.org/licenses/by/3.0/). © 2013 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution license (http://creativecommons.org/licenses/by/3.0/).
https://openalex.org/W3131628483	https://link.springer.com/content/pdf/10.1007/s00382-021-05696-x.pdf	English	null	Understanding the variability of the rainfall dipole in West Africa using the EC-Earth last millennium simulation	Climate dynamics	2,021	cc-by	12,160	Abstract There is a well-known mode of rainfall variability associating opposite hydrological conditions over the Sahel region and the Gulf of Guinea, forming a dipole pattern. Previous meteorological observations show that the dipole pattern varies at interannual timescales. Using an EC-Earth climate model simulation for last millennium (850–1850 CE), we investigate the rainfall variability in West Africa over longer timescales. The 1000-year-long simulation data show that this rainfall dipole presents at decadal to multidecadal and centennial variability and long-term trend. Using the singular value decomposition (SVD) analysis, we identified that the rainfall dipole present in the first SVD mode with 60% explained variance and asso- ciated with the variabilities in tropical Atlantic sea surface temperature (SST). The second SVD mode shows a monopole rainfall variability pattern centred over the Sahel, associated with the extra-tropical Atlantic SST variability. We conclude that the rainfall dipole-like pattern is a natural variability mode originated from the local ocean–atmosphere-land coupling in the tropical Atlantic basin. The warm SST anomalies in the equatorial Atlantic Ocean favour an anomalous low pressure at the tropics. This low pressure weakens the meridional pressure gradient between the Saharan Heat Low and the tropical Atlantic. It leads to anomalous northeasterly, reduces the southwesterly moisture flux into the Sahel and confines the Gulf of Guinea’s moisture convergence. The influence from extra-tropical climate variability, such as Atlantic multidecadal oscilla- tion, tends to modify the rainfall dipole pattern to a monopole pattern from the Gulf of Guinea to Sahara through influencing the Sahara heat low. External forcing—such as orbital forcing, solar radiation, volcanic and land-use—can amplify/dampen the dipole mode through thermal forcing and atmosphere dynamical feedback. Keywords West Africa rainfall dipole · Decadal to multidecadal variability · EC-Earth · Last millennium · Tropical Atlantic variability · Atlantic multidecadal oscillation Climate Dynamics (2021) 57:93–107 https://doi.org/10.1007/s00382-021-05696-x Climate Dynamics (2021) 57:93–107 https://doi.org/10.1007/s00382-021-05696-x Understanding the variability of the rainfall dipole in West Africa using the EC‑Earth last millennium simulation Qiong Zhang1,2 · Ellen Berntell1,2 · Qiang Li1,2 · Fredrik Charpentier Ljungqvist2,3,4 Received: 31 May 2020 / Accepted: 8 February 2021 / Published online: 23 February 2021 © The Author(s) 2021 * Qiong Zhang qiong.zhang@natgeo.su.se 1 Introduction The climate prediction and projection on decadal or longer timescales of western Africa rainfall are of central importance for the management of climate risk and vul- nerability (Otero et al. 2016), therefore it is crucial that global climate models are capable of representing local as well as remote drivers of rainfall variability and the inter- action between the two. This study has performed a ‘last millennium’ experiment from 850 to 1850 CE with an ocean–atmosphere coupled high-resolution global climate model EC-Earth (further details in Sect. 3.1 below). The results from 1001 years model output, from different diag- nostic aspects such as the long-term trend, the centennial- scale climate change between Medieval Common Anomaly (MCA) and Little Ice Age (LIA), and rainfall variability at decadal to multidecadal timescales, all exhibit a dipole variability mode between the Sahel region and the Gulf of Guinea. We have investigated the possible drivers and dynamical processes in the coupled climate model sys- tem and found that this rainfall dipole pattern results from tropical-extratropical interaction in the north Atlantic basin. The external forcing can further amplify/dampen the dipole mode. l Previous studies have shown that multidecadal drought in West Africa during the instrumental period (past 100 years for Africa) is primarily driven by changing Atlantic SST patterns (Hoerling et al. 2006; Knight et al. 2006; Martin et al. 2014; Mohino et al. 2011; Ting et al. 2011a; Zhang and Delworth 2006). It suggests that these patterns may reflect a natural low-frequency mode from ~ 65 to ~ 80 years of SST variability: the Atlantic multidecadal oscillation (AMO) (Folland et al. 1986). Several studies also propose the pos- sible impact of the Pacific decadal oscillation (PDO) on Western Africa rainfall (Mohino et al. 2011; Villamayor and Mohino 2015). However, climate models cannot skillfully capture the observed decadal to multidecadal WAM rainfall variability (Berntell et al. 2018; Martin et al. 2014; Masih et al. 2014; Mohino et al. 2011; Nouaceur and Murărescu 2016). Most global climate models underestimate the mag- nitude of decadal variability in the twentieth-century simu- lations (Biasutti 2013; Roehrig et al. 2013). The underes- timation of hydroclimate variability at decadal and longer timescales is not a feature unique to the WAM but is com- mon in other parts of the world (Ljungqvist et al. 2019a, b). This weak model-simulated decadal rainfall variability is not well understood but has so far received comparatively little attention. 1 Introduction Studies have also shown a connec- tion between Sahel rainfall and SST over the tropical Indian and Pacific Oceans (Palmer 1986; Xue et al. 2016). It is found that the Atlantic and Pacific SST have a comparable effect in reducing rainfall in the western Sahel. In contrast, the Indian Ocean SST reduces the rainfall in eastern Sahel (Palmer 1986). The impact from Pacific SST might be an indirect effect of the remote forcing on tropical Atlantic SST to influence the western African rainfall (Liu et al. 2004). anomalies, i.e. the SST anomalies over tropical and north Atlantic (Caminade and Terray 2010; Folland et al. 1986; Losada et al. 2010; Palmer 1986; Rodríguez-Fonseca et al. 2015; Zhang and Delworth 2006) or South Atlantic Ocean (Nnamchi and Li 2011). Observational data show that the SST anomalies in the tropical Atlantic are associated with a dipole rainfall patterns centred over the Sahel and the Gulf of Guinea (Janicot et al. 1998; Losada et al. 2010, 2012; Rodwell and Hoskins 1996; Ward 1998) before the 1970s. The dipole rainfall pattern disappeared after the 1970s, and change to a monopole rainfall pattern north of the Gulf of Guinea coast until 2000s (Janicot and Sultan 2001; Joly and Voldoire 2010). Losada et al. (2012) demonstrate with atmospheric modelling that the SST anomalies in tropical Atlantic alone can produce a rainfall dipole in West Africa. When the SST anomalies in the Atlantic appear accompa- nied by anomalies in the Indo-Pacific, the dipole pattern is drastically weakened. Studies have also shown a connec- tion between Sahel rainfall and SST over the tropical Indian and Pacific Oceans (Palmer 1986; Xue et al. 2016). It is found that the Atlantic and Pacific SST have a comparable effect in reducing rainfall in the western Sahel. In contrast, the Indian Ocean SST reduces the rainfall in eastern Sahel (Palmer 1986). The impact from Pacific SST might be an indirect effect of the remote forcing on tropical Atlantic SST to influence the western African rainfall (Liu et al. 2004). rainfall variability (Martin et al. 2014). The decadal climate variability in WAM has so far been tested in atmosphere- only models, with prescribed SST, and the problem remains unsolved (Berntell et al. 2018), pointing to unrealistically weak coupling between SST change and western African rainfall in model simulations. 1 Introduction One reason for the models’ shortcomings in simulating decadal-scale rainfall variability is likely related to coupled climate models’ inability to realistically simu- late the internally decadal SST variability—a key driver to 1 Introduction human life in West Africa. Most of the region’s population is largely dependent on rain-fed agriculture and pastures for livestock, making them highly vulnerable to rainfall fluctua- tions. Instrumental observation data over the past ~ 100 years indicate that western Africa rainfall shows significant vari- ability on interannual, decadal, and multidecadal timescales (Mohino et al. 2011). For example, the semi-arid Sahel region experienced an intense and persistent drought dur- ing the 1970s and 1980s, which is one of the main recent climate variability signals (Caminade and Terray 2010; Dai et al. 2004; Folland et al. 1986; Giannini et al. 2003; Mohino et al. 2011; Rodríguez‐Fonseca et al. 2011). The persistent Sahelian drought is widely covered worldwide by media for its dramatic socio-economic consequences, particularly food security (Kandji et al. 2006). West African rainfall is driven mainly by the West Afri- can monsoon (WAM) (Losada et al. 2010; Nicholson 2013; Sultan and Janicot 2003). Rainfall variability across vari- ous timescales is of great importance to ecosystems and 1 Department of Physical Geography, Stockholm University, Stockholm, Sweden 2 Bolin Centre for Climate Research, Stockholm University, Stockholm, Sweden 3 Department of History, Stockholm University, Stockholm, Sweden The interannual variability of the WAM is believed to be strongly related to nearby sea surface temperature (SST) 4 Swedish Collegium for Advanced Study, Uppsala, Sweden (0121 3456789) 3 94 Q. Zhang et al. anomalies, i.e. the SST anomalies over tropical and north Atlantic (Caminade and Terray 2010; Folland et al. 1986; Losada et al. 2010; Palmer 1986; Rodríguez-Fonseca et al. 2015; Zhang and Delworth 2006) or South Atlantic Ocean (Nnamchi and Li 2011). Observational data show that the SST anomalies in the tropical Atlantic are associated with a dipole rainfall patterns centred over the Sahel and the Gulf of Guinea (Janicot et al. 1998; Losada et al. 2010, 2012; Rodwell and Hoskins 1996; Ward 1998) before the 1970s. The dipole rainfall pattern disappeared after the 1970s, and change to a monopole rainfall pattern north of the Gulf of Guinea coast until 2000s (Janicot and Sultan 2001; Joly and Voldoire 2010). Losada et al. (2012) demonstrate with atmospheric modelling that the SST anomalies in tropical Atlantic alone can produce a rainfall dipole in West Africa. When the SST anomalies in the Atlantic appear accompa- nied by anomalies in the Indo-Pacific, the dipole pattern is drastically weakened. 2 Dry and wet episodes from paleoclimate proxy data during the last millennium A present in pres CE and change ciated w in the g rainfall with oc Table 1 Dry (brown) and wet (green) episodes shown in sediments from the Lak et al. 2000), and the Lake Bosumtwi (Shanahan et al. 2009) during the last millenniu Lake Chad and Kajemarum Oasis Lake Bosumtwi 900 1000 1100 1200 MCA conditions prevailing ~ 1100–1400 CE and during the sev- enteenth century, whereas drier conditions prevailed during the fifteenth and sixteenth centuries (Maley 1993). Slightly wetter conditions in northeastern Nigeria during 1000–1200 CE are also apparent in sediment data from the dry lake floor of the Kajemarum Oasis (Street-Perrott et al. 2000), which is located about 400 km northwest of Lake Chad (see Fig. 1b). In the West African tropical region, the Lake Bosumtwi record shows drier conditions from 900 to 1150 CE (Shanahan et al. 2009), corresponding to more dust emis- sions recorded in marine sediment from offshore Senegal (Mulitza et al. 2010). After this dry episode, overlapping with the peak of the MCA globally, the region returned to wetter conditions until the early fifteenth century. These wet- ter conditions were followed by a persistent 300-year dry period from ~ 1450 to ~ 1750 CE, roughly overlapping with the significant portion of the LIA. conditions prevailing ~ 1100–1400 CE and during the sev- enteenth century, whereas drier conditions prevailed during the fifteenth and sixteenth centuries (Maley 1993). Slightly wetter conditions in northeastern Nigeria during 1000–1200 CE are also apparent in sediment data from the dry lake floor of the Kajemarum Oasis (Street-Perrott et al. 2000), which is located about 400 km northwest of Lake Chad (see Fig. 1b). In the West African tropical region, the Lake Bosumtwi record shows drier conditions from 900 to 1150 CE (Shanahan et al. 2009), corresponding to more dust emis- sions recorded in marine sediment from offshore Senegal (Mulitza et al. 2010). After this dry episode, overlapping with the peak of the MCA globally, the region returned to wetter conditions until the early fifteenth century. These wet- ter conditions were followed by a persistent 300-year dry period from ~ 1450 to ~ 1750 CE, roughly overlapping with the significant portion of the LIA. 2 Dry and wet episodes from paleoclimate proxy data during the last millennium The evaluation of the climate models’ ability to capture the long-term climate variability largely depends on the study area’s observational data. However, there are no continu- ous instrumental observation datasets available beyond the last ~ 100 years in West Africa. Moreover, relatively little paleoclimate proxy data is available for West Africa, and most have comparably low temporal resolution and relatively poor dating control (Nash et al. 2016). The region is located in the tropical to subtropical climate zones, with no gla- ciers to provide ice-core records. No long tree-ring records have successfully been produced, both due to the logistic challenges and because tropical trees do not grow reliable annual growth rings. Records from lake sediments are scarce because the shallow ponds and marshes dotting the region have dried during recent and past drought episodes (Nash et al. 2016). The most crucial proxy records come from Lake Chad in the Sahel region (Maley 1993) and Lake Bosumtwi at the Coast of Guinea (Shanahan et al. 2009; Talbot and Delibrias 1977) (highlighted in Fig. 1b). Sediments and pollen data from Lake Chad indicate comparatively wet 1 3 3 Understanding the variability of the rainfall dipole in West Africa using the EC‑Earth last… 95 Fig. 1 Annual mean terrestrial rainfall (unit: mm/month) averaged from a CRU monthly data from 1901 to 2015, and b EC-Earth “last millennium” simulation from 850 to 1850. The red rectangle in a is the West Africa domain (5°–20° N, 20° W–30° E) defined in this study, the two sub-domains Sahel (10°–18° N, 20° W–30° E) and Coast of Guinea (5°–10° N, 20° W–30° E) are indicated with black rectangles. The rainfall indices for these three regions are averaged respectively within the defined domain. The Gulf of Guinea, includ- ing the coastal area along the Gulf, is indicated with a dashed rec- tangle (5° S–10° N, 20° W–15° E). Three lakes in West Africa that provide the hydrological information during the last millennium are noted in b rectangles. The rainfall indices for these three regions are averaged respectively within the defined domain. The Gulf of Guinea, includ- ing the coastal area along the Gulf, is indicated with a dashed rec- tangle (5° S–10° N, 20° W–15° E). Three lakes in West Africa that provide the hydrological information during the last millennium are noted in b rectangles. 2 Dry and wet episodes from paleoclimate proxy data during the last millennium The rainfall indices for these three regions are averaged respectively within the defined domain. The Gulf of Guinea, includ ing the coastal area along the Gulf, is indicated with a dashed rec- tangle (5° S–10° N, 20° W–15° E). Three lakes in West Africa tha provide the hydrological information during the last millennium are noted in b Fig. 1 Annual mean terrestrial rainfall (unit: mm/month) averaged from a CRU monthly data from 1901 to 2015, and b EC-Earth “last millennium” simulation from 850 to 1850. The red rectangle in a is the West Africa domain (5°–20° N, 20° W–30° E) defined in this study, the two sub-domains Sahel (10°–18° N, 20° W–30° E) and Coast of Guinea (5°–10° N, 20° W–30° E) are indicated with black rectangles. The rainfall indices for these three regions are averaged respectively within the defined domain. The Gulf of Guinea, includ- ing the coastal area along the Gulf, is indicated with a dashed rec- tangle (5° S–10° N, 20° W–15° E). Three lakes in West Africa that provide the hydrological information during the last millennium are noted in b conditions prevailing ~ 1100–1400 CE and during the sev- enteenth century, whereas drier conditions prevailed during the fifteenth and sixteenth centuries (Maley 1993). Slightly wetter conditions in northeastern Nigeria during 1000–1200 CE are also apparent in sediment data from the dry lake floor of the Kajemarum Oasis (Street-Perrott et al. 2000), which is located about 400 km northwest of Lake Chad (see Fig. 1b). In the West African tropical region, the Lake Bosumtwi record shows drier conditions from 900 to 1150 CE (Shanahan et al. 2009), corresponding to more dust emis- sions recorded in marine sediment from offshore Senegal (Mulitza et al. 2010). After this dry episode, overlapping with the peak of the MCA globally, the region returned to wetter conditions until the early fifteenth century. These wet- ter conditions were followed by a persistent 300-year dry period from ~ 1450 to ~ 1750 CE, roughly overlapping with the significant portion of the LIA. In Table 1, we summarised these wet-dry episodes dur- ing 850–1850 CE millennium. The changes in lake level recorded in Lake Bosumtwi indicated that West Africa was eve the twe the Sah anomal han et a Thes inform longer marum phase h tropica LIA. 2 Dry and wet episodes from paleoclimate proxy data during the last millennium was even drier during the eighteenth century than during the twentieth century, implying the drought prevailing in the Sahel since the end of the 1960s is not exceptionally anomalous in the context of the past millennium (Shana- han et al. 2009). These few available proxy records provide valuable information about the regional rainfall variability over longer timescales. The evidence from Lake Chad, Kaje- marum Oasis and Lake Bosumtwi suggest that the out of phase hydrological conditions between subtropical and tropical West Africa occurs both during the MCA and the LIA. Although in phase hydrological conditions are also present, this indicates that the dipole variability observed in present-day rainfall even existed during 850–1850 CE and varies at centennial timescales (Table 1). These changes at centennial timescales are assumed to be asso- ciated with large-scale phenomena such as SST changes in the global oceans. How do these drivers modulate the rainfall variability over the region can be better-understood with ocean–atmosphere coupled climate model simulation. i In Table 1, we summarised these wet-dry episodes dur- ing 850–1850 CE millennium. The changes in lake level recorded in Lake Bosumtwi indicated that West Africa Table 1 Dry (brown) and wet (green) episodes shown in sediments from the Lake Chad (Maley 1993), the Kajemarum Oasis (Street-Perrott et al. 2000), and the Lake Bosumtwi (Shanahan et al. 2009) during the last millennium Lake Chad and Kajemarum Oasis Lake Bosumtwi 900 1000 1100 1200 1300 1400 1500 1600 1700 1800 LIA MCA 1 3 Table 1 Dry (brown) and wet (green) episodes shown in sediments from the Lake Chad (Maley 1993), the Kajemarum Oasis (Street-Perrott et al. 2000), and the Lake Bosumtwi (Shanahan et al. 2009) during the last millennium Lake Chad and Kajemarum Oasis Lake Bosumtwi 900 1000 1100 1200 1300 1400 1500 1600 1700 1800 LIA MCA Table 1 Dry (brown) and wet (green) episodes shown in sediments from the Lake Chad (Maley 1993), the Kajem t al. 2000), and the Lake Bosumtwi (Shanahan et al. 2009) during the last millennium 3 3 96 Q. Zhang et al. variability. The initial condition starts from an equilibrium state at 850 CE after a 300-year spin-up. 3.1 The EC‑Earth model and the experiment setup Besides the ‘last millennium’ simulation, we performed a short ‘historical’ simulation from 1979 to 2009 using the CMIP5 historical forcing (Taylor et al. 2012). This ‘histori- cal’ simulation is used to evaluate WAM rainfall’s climatol- ogy by comparing it to present-day instrumental observation data. We use the ocean–atmosphere coupled climate model EC- Earth (version 3.1) to perform the “last millennium” simula- tion. The EC-Earth model results from a consortium devel- opment effort from a group of different European research institutions (Hazeleger et al. 2012). EC-Earth’s atmospheric component is revised from the Integrated Forecasting Sys- tem (IFS), which is developed at the European Centre for Medium-Range Weather Forecasts (ECMWF), coupled with a land model H-TESSEL. The IFS and H-TESSEL components are used to produce several ECMWF reanal- yses widely used in climate research, e.g., ERA40, ERA- Interim, ERA20C and the latest ERA5. The ocean compo- nent is based on Nucleus for European Modelling of the Ocean (NEMO) (Madec 2008) including a sea-ice model LIM3 (Bouillon et al. 2009). The EC-Earth model has a rela- tively high resolution comparing the other climate models in PMIP3. The atmosphere’s horizontal resolution is 125 km, and 62 vertical layers; the ocean model NEMO has a hori- zontal resolution of 110 km with 40 vertical layers. We use monthly mean rainfall, SST, sea level pressure, specific humidity, and horizontal wind at 850 hPa from the model output in the analysis. The simulated total rainfall is the sum of large-scale rainfall and convective rainfall. The moisture flux is calculated by humidity and horizontal wind field. 3 Climate model simulation and analysis methods Our EC-Earth “last millennium” simulation has been completed after PMIP3; hence the model data are not presented in most of the PMIP3 “last millennium” model studies. We have included this dataset in a recent simula- tion-reconstruction comparison study evaluating the centen- nial-scale temperature changes together with other PMIP3 simulations (Ljungqvist et al. 2019b). The results of that study show that the EC-Earth is among the models that agree well with continental-scale temperature reconstructions for MCA and LIA changes. The EC-Earth “last millennium” simulation has also been used to estimate the maximum annual number of Atlantic tropical cyclones. The simula- tion has proved useful to statistically assess extreme climate events (Lavender et al. 2018). The period 850–1850 CE provides a comparatively well- documented history of multidecadal to multi-centennial climate changes, dominated by the MCA and LIA (Stocker et al. 2013). Thus ‘last millennium’ simulations are con- sidered one of the key experiments in the Paleoclimate Modelling Intercomparison Project (PMIP) (Kageyama et al. 2018). We have run this simulation following the PMIP3 protocol. Several conventional statistical methods are applied to analyse the monthly mean model output. The statistically significant features from the analyses are presented and discussed. 3.3 Statistical analysis method The CRU rainfall data from 1970 to 1997 show that ~ 60% of its annual rainfall (710 mm) fall in the JAS season (414 mm) (Fig. 2a) when the monsoon rainfall is at its most northerly location. EC-Earth simulated rainfall shows a simi- lar seasonal cycle (Fig. 2b) but has a dry bias of 47% over the Sahel and dry bias of 22% over the Coast of Guinea during the NH summer. This dry bias in EC-Earth appears to result from both of the WAM being weaker and not reach- ing as far north as in observations for the same period. For example, the CRU shows a WAM with a wet centre located over 8°–12° N, while the model simulation has a weaker maximum over around 7°–11° N. This dry bias in western Africa remains in the new model version EC-Earth3.3 with even higher resolution (Haarsma et al. 2020). The physi- cal causes of this dry bias can be due to a wide range of deficiencies in the model parameterisations of land surface and cloud physics, which needs to be investigated in further detail. To investigate the variability at different frequencies besides the long-term trend, we applied linear detrending using a least- square fit and filtering using a Butterworth filter (Emery and Thomson 2001). Two climate regimes, the MCA (900–1200 CE) and LIA (1550–1850 CE) are selected to address the cen- tennial-scale climate change. The significance of the climate change between these two climate periods is examined using the two-sample t-test (Wilks 2011). p ( ) The relationship between the SST and West African rain- fall is analysed using the singular value decomposition (SVD) analysis of the SST-rainfall cross-covariance matrix (Brether- ton et al. 1992). The SVD analysis generalises empirical orthogonal function (EOF) analysis applied to two data fields (SST and rainfall in this case) together to identify modes that explain the largest covariance between the two fields. The expansion coefficients are time-series, one for each field, describing how the modes of variability oscillate in time. We calculate the homogeneous correlation between the expansion coefficients and their respective original field to visualise the covariance. To understand the mechanisms behind the covari- ance between SST and rainfall, regression maps are produced for sea level pressure (SLP), wind, humidity and moisture flux at 850 hPa against the two leading SVD expansion coefficients. 4.1 Model evaluation with instrumental observations West Africa lies just south of the Sahara Desert (Fig. 1) and receives 768 mm annual mean rainfall (Fig. 1), decreas- ing from the south to north. There are two regional rainfall maxima in the West Africa domain. One is centred on the west coast around 5°–11° N, near the Cameroon highlands in eastern Coast of Guinea. There is a weaker rainfall maxi- mum in the east part of the continent, over the Ethiopian highlands. The climatology from the EC-Earth last millen- nium simulation shows a very similar pattern. 4 Results more rainfall in the Sahel region, while the coastal area to the south, which we refer to as the Coast of Guinea, experiences a slightly decreased rainfall. Previous observation from station data shows that the year-to-year variability exhibits a strong contrast between the Sahel region and the Coast of Guinea (Janicot 1992; Janowiak 1988; Nicholson 1980; Nicholson and Palao 1993). The sign change in rainfall anomaly occurs approximately at 10° N (Janicot 1992). We further define two sub-region domain Sahel (10°–18° N, 20° W–30° E) and Coast of Guinea (5°–10° N, 20° W–30° E) to investigate the land area-averaged rainfall variability. In this study, the Gulf of Guinea refers to the equatorial Atlantic area at 5° S–5° N, 20° W–15° E, including a small part of the land area. 3.2 Instrumental observation data Observational monthly mean rainfall and surface air tem- perature data derived from CRU TS 3.24.01 from Climate Research (Harris et al. 2014) are used to evaluate the model performance. The CRU dataset used land-based in-situ observations from meteorological stations across the globe and interpolated into a grid with 0.5° latitude by longitude resolution. The data length is 115 years covering 1901–2015. The availability of meteorological stations over Africa varies with time, i.e., 60–80% of the grid cells in Africa have sta- tion data for most of the twentieth century, with lower num- bers at the beginning of the century and at the end. Grid cells lacking data are supplied with values equal to the 1961–1990 climatological mean. y The climate forcing for the ‘last millennium’ experiment follows the PMIP3 protocol (Schmidt et al. 2011, 2012), with the orbital forcing calculated internally in the model following Berger (1978). Variations in total solar irradi- ance is based on the reconstruction provided by Vieira and Solanki (2010) for 850–1609 CE, and from 1610 CE on the reconstruction from Wang et al. (2005), and the volcanic aerosol forcing is from Crowley and Unterman (2013). Changes in greenhouse gas forcing are derived from Hansen and Sato (2004) and Macfarling Meure et al. (2006). Land- use changes are taken from Pongratz et al. (2008). All these imposed forcings vary from 850 to 1850 CE, and therefore this ‘last millennium’ simulation provides a 1001-year cli- mate transient, containing both internal and externally forced The climatological annual mean rainfall shows an appar- ent difference from the northern arid Sahara to Guinea’s tropical coast, ranging from approximately 10 mm/month to 100 mm/month. We define our study region of West Africa as the domain 5°–20° N, 20° W–30° E. In regular model grids, it includes part of the adjacent Atlantic Ocean. The rainfall over the land area of this domain is averaged as a West Africa rainfall index. During the boreal summer months, the onset of the WAM shifts the monsoon rainfall northward. It leads to 1 3 97 Understanding the variability of the rainfall dipole in West Africa using the EC‑Earth last… 3.3 Statistical analysis method Power spectrum analysis is applied to calculate the periods on which the covariance varies. Here we use the multi-taper method (MTM) (Mann and Lees 1996), performed with the program k-spectra and parameters p = 3, K = 5. The peak peri- ods’ significance is calculated with the null-hypothesis that the peaks in the power spectrum are significantly larger than the red-noise spectrum at that frequency (Wilks 2011). The relationship between the SST and West African rain- fall is analysed using the singular value decomposition (SVD) analysis of the SST-rainfall cross-covariance matrix (Brether- ton et al. 1992). The SVD analysis generalises empirical orthogonal function (EOF) analysis applied to two data fields (SST and rainfall in this case) together to identify modes that explain the largest covariance between the two fields. The expansion coefficients are time-series, one for each field, describing how the modes of variability oscillate in time. We calculate the homogeneous correlation between the expansion coefficients and their respective original field to visualise the covariance. To understand the mechanisms behind the covari- ance between SST and rainfall, regression maps are produced for sea level pressure (SLP), wind, humidity and moisture flux at 850 hPa against the two leading SVD expansion coefficients. The seasonal cycle structure in rainfall remains similar in 1901–2015 mean for CRU data (Fig. 2c) and the 850–1850 mean for EC-Earth ‘last millennium’ simulation (Fig. 2d). We notice that the WAM was stronger, and extended fur- ther northward, in the long-term mean in both data sources comparing their 1970–1997 mean. The weaker WAM in 1970–1997 mean both in observation and simulation reflects the well-known persistent drought period in West Africa from the early 1970s to the 1990s. The mean monsoon rainfall of the last millennium from EC-Earth simulation shows a more vigorous intensity in the JAS season than the 1901–2015 mean rainfall from CRU data. This is consistent with the results from previous studies, suggesting that West Africa and the Sahel region were wetter in several episodes during the 850–1850 period than during the twentieth cen- tury (Shanahan et al. 2009). In the following analysis, we focus on July–September (JAS) when West Africa receives most of its rainfall. fi Power spectrum analysis is applied to calculate the periods on which the covariance varies. Here we use the multi-taper method (MTM) (Mann and Lees 1996), performed with the program k-spectra and parameters p = 3, K = 5. 3.3 Statistical analysis method During the last millennium, the slight changes in eccentricity and obliquity, and a 20-day shift in perihelion (from 15 December to 4 January), together with give a 7.8 W/m2 decrease in insolation in August at 15ºN. During 3.3 Statistical analysis method The peak peri- ods’ significance is calculated with the null-hypothesis that the peaks in the power spectrum are significantly larger than the red-noise spectrum at that frequency (Wilks 2011). 1 3 98 Q. Zhang et al. 4.2 Long‑term trend in rainfall The multidecadal variability is apparent in the 850–1850 rainfall variation, with a significant decreasing trend in summer rainfall occurring over the West Africa domain (Fig. 3). The decreasing trend is associated to the decline of insolation in the Northern Hemisphere at the top of the atmosphere. The insolation at the top of the atmosphere is determined by the solar irradiance (emitted energy from the Sun) and the Earth’s orbit (the latitudinal distribution of the short-wave solar energy). The trend in insolation is mostly due to changes in orbital forcing (Schmidt et al. 2011). During the last millennium, the slight changes in eccentricity and obliquity, and a 20-day shift in perihelion (from 15 December to 4 January), together with give a 7.8 W/m2 decrease in insolation in August at 15ºN. During Fig. 2 Latitudinal-time distribu- tion of terrestrial rainfall (unit: mm/month) over West Africa (5°–20° N, 20° W–30° E). a From CRU observational data averaged for 1970–1997; b from EC-Earth “historical” simula- tion averaged for 1970–1997; c from CRU monthly mean data averaged for 1901–2015; d from EC-Earth ‘last millen- nium’ simulation averaged for 850–1850 98 Q. Zhang et al. Fig. 2 Latitudinal-time distribu- tion of terrestrial rainfall (unit: mm/month) over West Africa (5°–20° N, 20° W–30° E). a From CRU observational data averaged for 1970–1997; b from EC-Earth “historical” simula- tion averaged for 1970–1997; c from CRU monthly mean data averaged for 1901–2015; d from EC-Earth ‘last millen- nium’ simulation averaged for 850–1850 determined by the solar irradiance (emitted energy from the Sun) and the Earth’s orbit (the latitudinal distribution of the short-wave solar energy). The trend in insolation is mostly due to changes in orbital forcing (Schmidt et al. 2011). During the last millennium, the slight changes in eccentricity and obliquity, and a 20-day shift in perihelion (from 15 December to 4 January), together with give a 7.8 W/m2 decrease in insolation in August at 15ºN. During determined by the solar irradiance (emitted energy from the Sun) and the Earth’s orbit (the latitudinal distribution of the short-wave solar energy). The trend in insolation is mostly due to changes in orbital forcing (Schmidt et al. 2011). Fig. 2 Latitudinal-time distribu- tion of terrestrial rainfall (unit: mm/month) over West Africa (5°–20° N, 20° W–30° E). a From CRU observational data averaged for 1970–1997; b from EC-Earth “historical” simula- tion averaged for 1970–1997; c from CRU monthly mean data averaged for 1901–2015; d from EC-Earth ‘last millen- nium’ simulation averaged for 850–1850 4.3 Rainfall change between medieval climate anomaly and little ice age In the EC-Earth “last millennium” simulation, as in other climate model simulations, the timing of the MCA and LIA slightly differs to those suggested by proxy-based recon- structions. The climate models are not expected to reproduce the exact climate evolution as the real world. It depends on the complexity of the physical processes built in the model system and the uncertainties in the climate forcing that is reconstructed from proxy data (Ljungqvist et al. 2019b). The WAM is driven and modulated by the thermal conditions over the African continent and adjacent ocean, therefore in the model system, the rainfall variability would be dynami- cally consistent with the temperature variability. According to the evolution of global mean temperature in EC-Earth simulation (Ljungqvist et al. 2019b), here we define the rela- tively warm period around 900–1200 CE as MCA and the cold period at the later part from 1550 to 1850 CE as LIA in the EC-Earth “last millennium” simulation. We show a climate change pattern between the two periods in Fig. 5. A similar latitudinal dipole pattern in the long-term trend in Fig. 4a is observed, with wet conditions in Sahel/Sahara and dry conditions over the Gulf of Guinea (including part of the coastline, as defined in Sect. 2). Ljungqvist et al. (2016) also identified dry conditions along the Coast of Guinea in their anomaly maps of the 11th to thirteenth centuries. pp A spatial trend analysis of the total JAS rainfall over North Africa also reveals a drying trend over the 850–1850 period, centred over 10°–18° N, extending from east to west over the entire continent (Fig. 4a). The rainfall centre falls right into the Sahel domain as defined in Fig. 1a. Its southern neighbour Coast of Guinea exhibits both weak negative and positive rainfall trends, which explains the previous results in Fig. 3 where no trend is observed in its rainfall index time series. The positive rainfall trend is instead centred over the Gulf of Guinea and equatorial Atlantic. The rainfall trend pattern exhibits a latitudinal dipole, which is also seen in more recent observations (e.g. trend in 1998–2013) (Odoulami and Akinsanola 2018). This dipole pattern is typically established through a change of the surface temperature gradient that modulates the monsoon flow and the African Easterly Jet (Janicot et al. 1998; Losada et al. 2010). 4.2 Long‑term trend in rainfall The multidecadal variability is apparent in the 850–1850 rainfall variation, with a significant decreasing trend in summer rainfall occurring over the West Africa domain (Fig. 3). The decreasing trend is associated to the decline of insolation in the Northern Hemisphere at the top of the atmosphere. The insolation at the top of the atmosphere is 1 3 Fig. 3 The variation of summer rainfall (JAS mean, unit: mm/ month) from 850 to 1850 over western Africa (light blue, lin- ear trend shown in black), Sahel (blue), Coast of Guinea (red) and Gulf of Guinea (green). A Butterworth low-pass filter with 30-year cut-off frequency has been applied to the time series Fig. 3 The variation of summer rainfall (JAS mean, unit: mm/ month) from 850 to 1850 over western Africa (light blue, lin- ear trend shown in black), Sahel (blue), Coast of Guinea (red) and Gulf of Guinea (green). A Butterworth low-pass filter with 30-year cut-off frequency has been applied to the time series 1 3 Understanding the variability of the rainfall dipole in West Africa using the EC‑Earth last… 99 the 1001-year simulated period from 850 to 1850 CE, WAM rainfall decreased by 7.5%. For the West African continent as a whole, the long-term linear trends com- prise a wet period in the beginning and a dry period in the later part, mostly consistent with paleoclimate proxy data derived from lake sediments (Maley 1993; Shanahan et al. 2009; Street-Perrott et al. 2000). The variation of West African rainfall mainly reflects the rainfall variability in the Sahel. An apparent trend is seen in the rainfall in the Sahel region, which has decreased by about 12% during the last millennium. There is no noticeable trend over the Coast of Guinea. A slightly increasing rainfall trend over the Gulf of Guinea. A visible wet–dry multidecadal oscil- lation is apparent in all of the time series. Earth’s orbit. This cooling trend, which weakens the lati- tudinal temperature gradient, results in a weakening of the WAM and leads to drier conditions in the Sahel region as seen in Fig. 4a. The dry conditions at the surface reduce the evaporative cooling in the monsoon season and lead to anomalous warming effect, as shown from the surface latent heat flux (Fig. 4c). This warming effect prevails over the surface cooling, caused by insolation change, and eventually results in warming over 15–20ºN. 4.3 Rainfall change between medieval climate anomaly and little ice age The spatial trend in sur- face air temperature shows a contrast pattern with cool- ing in Sahara, decreasing southward but covering most of northern Africa, and warming in the north part of the Sahel. During the 850–1850 CE period, the large-area cooling trend is consistent with the declining insolation in Northern Hemisphere summer due to changes in the a Precipitation: trend 850-1850 CE 30° 0 W ° 30° E 0° 15° N 30° N 45° N -5 -4 -3 -2 1 0 1 2 3 4 5 Precipitation (mm/month per 100a) b Temperature (2 m): trend 850-1850 CE 30° 0 W ° 30° E 0° 15° N 30° N 45° N -0.2 -0.15 -0.1 -0.05 0 0.05 0.1 0.15 0.2 Temperature (oC per 100a) Fig. 4 The linear trend over the northern Africa region, a for JAS mean rainfall (unit: mm/month per 100 years); b for JAS mean 2-m air temperature (unit: °C per 100 years); c for JAS mean surface latent heat flux (unit: W/m2 per 100 years, vertical flux defined as positive downward). The trend is calculated for each model grid for the period 850–1850. Stippling indicates the trends that are significant at the 95% confidence level b Temperature (2 m): trend 850-1850 CE 0 ° 30° E -0.15 -0.1 -0.05 0 0.05 0.1 0.15 0.2 Temperature (oC per 100a) on, a for JAS AS mean 2-m surface latent i ed as positive downward). The trend is calculated for each model grid for the period 850–1850. Stippling indicates the trends that are significant at the 95% confidence level b Temperature (2 m): trend 850-1850 CE 30° 0 W ° 30° E 0° 15° N 30° N 45° N -0.2 -0.15 -0.1 -0.05 0 0.05 0.1 0.15 0.2 Temperature (oC per 100a) a Precipitation: trend 850-1850 CE 30° 0 W ° 30° E 0° 15° N 30° N 45° N -5 -4 -3 -2 1 0 1 2 3 4 5 Precipitation (mm/month per 100a) downward). The trend is calculated for each model grid for the period 850–1850. Stippling indicates the trends that are significant at the 95% confidence level Fig. 4.3 Rainfall change between medieval climate anomaly and little ice age 4 The linear trend over the northern Africa region, a for JAS mean rainfall (unit: mm/month per 100 years); b for JAS mean 2-m air temperature (unit: °C per 100 years); c for JAS mean surface latent heat flux (unit: W/m2 per 100 years, vertical flux defined as positive 1 3 Q. Zhang et al. 100 Fig. 5 The change in JAS rainfall (unit: mm/month) between medie- val climate anomaly (MCA, averaged for 900–1200 CE) and little ice age (LIA, averaged for 1550–1850 CE). Stippling indicates the differ- ences that are significant at the 95% confidence level multidecadal climate variability, an only 100-year long data- set can hardly yield statistically significant results. Here we use the 1001-year model output to revisit the co-variability question between the SST and the West African rainfall. The SVD analysis is performed on JAS mean precipitation and SST, and all the data is detrended, and further filtered by removing high-frequency fluctuations less than 10 years.i i by removing high-frequency fluctuations less than 10 years. As shown in Figs. 6 and 8, the first two leading SVD modes account for ~ 90% of the total square covariance. The first mode accounts for ~ 60% (Fig. 6) and the second accounts for ~ 30% (Fig. 8). The patterns represent robust modes of variability and are consistent with the results of a rotated EOF performed on JAS rainfall over West Africa (not shown). SVD mode 1 indicates an Atlantic Niño type SST variability pattern, with a strong warm SST centre over the equatorial Atlantic Ocean (Fig. 6a), which affects the Gulf of Guinea’s rainfall variability and its surrounding coast (Fig. 6b). The Atlantic Niño is a zonal mode of tropical Atlantic variability (Zebiak 1993). It represents changes in the cold tongue at the eastern part of the basin, just south of the Equator. The Atlantic Niño is known to affect interan- nual climate variability over adjacent and remote regions. A warming SST could lead to increases of rainfall over the Gulf of Guinea coast (Losada et al. 2010; Mohino et al. 2011), and this covariant feature has been shown in SVD1 patterns. During 850–1850, the statistically significant period of variability appears at decadal and multidecadal timescales, i.e., 10–14 years and 20–40 years, in both SST and rainfall field (Fig. 6c, d). A 60–80 years co-variability also shows up but only statistically significant in SST field. 4.3 Rainfall change between medieval climate anomaly and little ice age The regression maps in Fig. 7 demonstrate the physical pro- cesses of how the SST anomalies over equatorial Atlantic affect the rainfall over the Gulf of Guinea. The warming SSTs (Fig. 6a) over the eastern tropical Atlantic increases the amount of moisture in the atmosphere (Fig. 7a) and forms an anomalous low pressure over this area (Fig. 7b). The low pressure weakens the meridional pressure gradient between the Saharan Heat Low and the equatorial Atlantic and leads to anomalous northeasterly (Fig. 7c). This decreases the southwesterly moisture flux into the Sahel and confines the moisture flux convergence at the equatorial region (Fig. 7d), resulting in reduced rainfall over the Sahel and increased rainfall in Gulf of Guinea and its surrounding coastal area. Meanwhile, there is an increased zonal pressure gradient (Fig. 7b), which enhances the westerlies (Fig. 7c) and brings more moisture to the Coast of Guinea (Fig. 7d). The homogeneous correlation maps of the second SVD Fig. 5 The change in JAS rainfall (unit: mm/month) between medie- val climate anomaly (MCA, averaged for 900–1200 CE) and little ice age (LIA, averaged for 1550–1850 CE). Stippling indicates the differ- ences that are significant at the 95% confidence level An overall wetter MCA and a drier LIA over the Sahel reflect the WAM’s response to the long-term trend of orbital forcing and solar forcing. However, the internal impacts of the ocean–atmosphere system also play a role. For instance, previous work showed that the wetter conditions during the MCA are related to the dominant positive phase of the AMO (Kuhnert and Mulitza 2011; Mann et al. 2009; Wang et al. 2017). The connection is associated with a northward shift of the Atlantic ITCZ (Ting et al. 2011b) and typically increases moisture flux into the western Sahel from Atlantic sources in the south and west (Delworth et al. 2007; Martin et al. 2014; Zhang and Delworth 2006). Below we further examine the effect of ocean–atmosphere coupling on West African rainfall variability. 4.4 West African rainfall variability pattern and its relation to SST The maxi- mum variability in the SST field is centred at the latitudes of maximum Atlantic meridional overturning circulation (AMOC) around 40°–50° N. This SST pattern affects the rainfall variability over a large area of North Africa includ- ing Sahara, with the maximum variability occurring over the Sahel, expanding to its far eastern part. The rainfall vari- ability is centred at 15° N, poleward of the mean JAS rainfall maximum (10°–12° N) (Fig. 2d), and maximum humidity anomalies are located in the north-eastern part of Sahel and Sahara (Fig. 9a). The warm SSTs in the North Atlantic and the Mediterranean corresponds to a deepening of the ther- mally induced low-pressure area across the Sahara (Saha- ran Heat Low) (Fig. 9b), leading to a stronger meridional pressure gradient and strengthened cyclonic circulation (Fig. 9c). More moisture is transported into the Sahel from the Atlantic by enhanced west/southwesterly, leading to a moisture flux convergence over the region and a northward shift of the rainfall. There is also a slight increase of mois- ture being transported from the Mediterranean by enhanced co-vary at decadal to multi-decadal timescale, i.e. a ~ 10–14 and ~ 20–30 years oscillation period (Fig. 8c, d). The maxi- mum variability in the SST field is centred at the latitudes of maximum Atlantic meridional overturning circulation (AMOC) around 40°–50° N. This SST pattern affects the rainfall variability over a large area of North Africa includ- ing Sahara, with the maximum variability occurring over the Sahel, expanding to its far eastern part. The rainfall vari- ability is centred at 15° N, poleward of the mean JAS rainfall maximum (10°–12° N) (Fig. 2d), and maximum humidity anomalies are located in the north-eastern part of Sahel and Sahara (Fig. 9a). The warm SSTs in the North Atlantic and the Mediterranean corresponds to a deepening of the ther- mally induced low-pressure area across the Sahara (Saha- ran Heat Low) (Fig. 9b), leading to a stronger meridional pressure gradient and strengthened cyclonic circulation (Fig. 9c). More moisture is transported into the Sahel from the Atlantic by enhanced west/southwesterly, leading to a moisture flux convergence over the region and a northward shift of the rainfall. There is also a slight increase of mois- ture being transported from the Mediterranean by enhanced 4.4 West African rainfall variability pattern and its relation to SST Studies using observational data and atmospheric model- ling have confirmed that SST anomalies over the equatorial Atlantic are responsible for the summer rainfall dipole pat- tern over West Africa at interannual timescales (Rodwell and Hoskins 1996; Rowell et al. 1995). However, the anticor- relation is non-stationary (Losada et al. 2012; Mohino et al. 2011). For low-frequency variability, observation data shows that the inter-hemispheric SST pattern is strongly related to the multidecadal variability in 100 years of instrumental data (1901–2002), in this coupling mode the rainfall exhibits same-signed anomalies over West Africa (Joly et al. 2007). Most ocean–atmosphere coupled climate models failed to capture the low-frequency SST-rainfall coupling evident in the observations (Joly et al. 2007; Martin et al. 2014). For The homogeneous correlation maps of the second SVD mode depict a monopole rainfall variability pattern extend- ing from the Coast of Guinea to the Sahara region (Fig. 8b), connected to an AMO-like pattern of SST anomalies located in the North Atlantic and the Mediterranean (Fig. 8a) and consistent with findings based on shorter data for the past few decades (Park et al. 2016; Sheen et al. 2017). They also 1 3 3 Understanding the variability of the rainfall dipole in West Africa using the EC‑Earth last… 101 Fig. 6 The first leading SVD mode (explained 60.1% of variance) presented as homogeneous correlation maps for a SST mode 1 and b rainfall mode 1. Stippling indicates correlations that are significant at the 95% confidence level, and the maps boundaries indicate the areas used when performing the SVD. The power spectrum for the expan- sion coefficients of c SST mode 1 and d rainfall mode 1, the numbers labelled are timescales of variability in years Fig. 6 The first leading SVD mode (explained 60.1% of variance) presented as homogeneous correlation maps for a SST mode 1 and b rainfall mode 1. Stippling indicates correlations that are significant at the 95% confidence level, and the maps boundaries indicate the areas used when performing the SVD. The power spectrum for the expan- sion coefficients of c SST mode 1 and d rainfall mode 1, the numbers labelled are timescales of variability in years northerlies (Fig. 9d), leading to increased rainfall across northern Africa. co-vary at decadal to multi-decadal timescale, i.e. a ~ 10–14 and ~ 20–30 years oscillation period (Fig. 8c, d). 5 Discussion and conclusion The ‘last millennium’ EC-Earth climate model simulation has shown a latitudinal rainfall dipole over West African rainfall at decadal to multidecadal timescales. The North Atlantic SST variability modulates the rainfall dipole. The rainfall dipole also appears in the long-term trend, and centennial variability between the MCA and LIA, indicat- ing that the external forcing can amplify the dipole pattern through thermal forcing and feedback. From the SVD analysis, we notice that the dipole-like rainfall pattern appears in SVD1 but not in SVD2. We conclude that the tropical SST anomalies are the primary driver for the dipole-like pattern, predominantly on decadal to multidecadal timescales (~ 10–30 years). The second SVD presents a monopole rainfall pattern extending from the Coast of Guinea to the Sahara, centred on the Sahel 1 3 1 Q. Zhang et al. 102 Fig. 7 Regression maps against the SVD1 SST expansion coefficient for a specific humidity at 850 hPa, b sea level pressure, c horizontal wind (vector) and wind speed (shading) at 850 hPa, and d moisture flux (vector) and its convergence (shading) at 850 hPa, expressed in units per standard deviation (SD) change in the expansion coefficient. Stippling indicates regressions that are significant at the 95% confi- dence level Fig. 7 Regression maps against the SVD1 SST expansion coefficient for a specific humidity at 850 hPa, b sea level pressure, c horizontal wind (vector) and wind speed (shading) at 850 hPa, and d moisture flux (vector) and its convergence (shading) at 850 hPa, expressed in units per standard deviation (SD) change in the expansion coefficient. Stippling indicates regressions that are significant at the 95% confi- dence level region. This monopole rainfall pattern in West Africa varies at ~ 10–30 years’ timescale, modulated by the North Atlan- tic and Mediterranean SST anomalies. It confirmed the pre- vious study showing that AMO plays a significant role in forcing the 20th-century multidecadal variations in Sahel summer rainfall (Zhang and Delworth 2006). In their model simulation, the positive AMO phase leads to a northward shift of the ITCZ. It is associated with anomalous south- westerly surface winds over the Sahel, leading to a surface moisture convergence and enhancing the Sahel’s rainfall. We confirm that this process is statistically robust in our “last millennium” model simulation data. 5 Discussion and conclusion The volcanic eruption is suggested to have initiated the LIA (Miller et al. 2012) and might have contributed to the centennial variability rainfall pattern between the MCA and LIA. Based on our model experi- ments, we find it plausible that the linear drying trend during the last millennium is the direct response to the orbital forcing, driven by the cooling trend in surface air temperature over the Sahara region. Both solar forcing, GHG forcing, and volcanic forcing and other minor forc- ing, can also affect the surface energy balance and amplify the rainfall response through atmospheric feedback (Row- ell 2003) and land surface feedback (Nicholson 2013). low-pressure area across Sahara. Moisture transported into Sahara with the anomalous northeasterly winds can also have helped strengthen the low-pressure area by enhancing the convective cells by releasing latent heat. In our fully coupled “last millennium” simulation, the SST forcing in North Atlantic and the Mediterranean and accompanied low- pressure system dominates the Sahel’s rainfall. low-pressure area across Sahara. Moisture transported into Sahara with the anomalous northeasterly winds can also have helped strengthen the low-pressure area by enhancing the convective cells by releasing latent heat. In our fully coupled “last millennium” simulation, the SST forcing in North Atlantic and the Mediterranean and accompanied low- pressure system dominates the Sahel’s rainfall. The climate forcing for the last millennium include orbital forcing and solar irradiance, GHG, volcanic, and land-use changes (Schmidt et al. 2011, 2012). The GHG forcing and land-use changes are very small and do not yet reflect the anthropogenic forcing we have experienced during the twentieth century. The volcanic eruption is suggested to have initiated the LIA (Miller et al. 2012) and might have contributed to the centennial variability rainfall pattern between the MCA and LIA. Based on our model experi- ments, we find it plausible that the linear drying trend during the last millennium is the direct response to the orbital forcing, driven by the cooling trend in surface air temperature over the Sahara region. Both solar forcing, GHG forcing, and volcanic forcing and other minor forc- ing, can also affect the surface energy balance and amplify the rainfall response through atmospheric feedback (Row- ell 2003) and land surface feedback (Nicholson 2013). 5 Discussion and conclusion It also supports the theory that the negative AMO phase from the 1970s to 1990s largely contributed to the persistent droughts in the Sahel during this period. A proper simulation of the remote SST influences on the west African rainfall appears to be a key for predictive skill at multidecadal timescales (Mar- tin et al. 2014). The predict skill in west African rainfall is likely associated with significant AMO prediction skill (García-Serrano et al. 2015; Guemas et al. 2015; Mohino et al. 2016). Notably, changes in the Sahelian rainfall are not significant in some PMIP3 last millennium simulations (Stevenson et al. 2018). The reason might depend on how well the internal variability is represented in different “last millennium” simulations. The contribution of Mediterranean SST forcing on Sahel rainfall shown in the second SVD is similar to those identi- fied by Park et al. (2016) with respect to SST forcing from different ocean basins. While the increased specific humidity across the Sahel mainly reflects increased moisture flux from the Atlantic, the warm SSTs in the North Atlantic and the Mediterranean drive the deepening of the thermally induced 1 3 3 103 Understanding the variability of the rainfall dipole in West Africa using the EC‑Earth last… Fig. 8 The second leading SVD mode (explained 29.5% of variance) presented as homogeneous correlation maps for a SST mode 2 and b rainfall mode 2. The power spectrum for the expansion coefficients of c SST mode 2 and d rainfall mode 2, the numbers labelled are time- scales of variability in years. Stippling in a and b indicates correla- tions that are significant at the 95% confidence level c SST mode 2 and d rainfall mode 2, the numbers labelled are time- scales of variability in years. Stippling in a and b indicates correla- tions that are significant at the 95% confidence level Fig. 8 The second leading SVD mode (explained 29.5% of variance) presented as homogeneous correlation maps for a SST mode 2 and b rainfall mode 2. The power spectrum for the expansion coefficients of The climate forcing for the last millennium include orbital forcing and solar irradiance, GHG, volcanic, and land-use changes (Schmidt et al. 2011, 2012). The GHG forcing and land-use changes are very small and do not yet reflect the anthropogenic forcing we have experienced during the twentieth century. 5 Discussion and conclusion Several modelling studies have suggested that the anthropogenic aerosol forcing and GHG forcing are the most critical drivers for the drying trend in the Sahel in late twentieth century (Biasutti and Giannini 2006; Booth et al. 2012; Chang et al. 2011; Dong and Sutton 2015; Held et al. 2005; Hwang et al. 2013; Rotstayn and Lohmann 2002). The impacts of land cover changes to Sahel drought have been demonstrated by regional cli- mate modelling (Boone et al. 2016; Wang et al. 2016). 1 3 Q. Zhang et al. 104 Fig. 9 Regressions against the SVD2 SST expansion coefficient for a specific humidity at 850 hPa, b sea level pressure, c horizontal wind (vector) and wind speed (shading) at 850 hPa, and d moisture flux (vector) and its convergence (shading) at 850 hPa, expressed in units per standard deviation change in the expansion coefficient. Stippling indicates regressions that are significant at the 95% confidence level (vector) and its convergence (shading) at 850 hPa, expressed in units per standard deviation change in the expansion coefficient. Stippling indicates regressions that are significant at the 95% confidence level Fig. 9 Regressions against the SVD2 SST expansion coefficient for a specific humidity at 850 hPa, b sea level pressure, c horizontal wind (vector) and wind speed (shading) at 850 hPa, and d moisture flux References Harris I, Jones PD, Osborn TJ, Lister DH (2014) Updated high‐resolu- tion grids of monthly climatic observations—the CRU TS3. 10 Dataset Int J Climatol 34:623–642 Berger AL (1978) Long-term variations of caloric insolation resulting from the earth’s orbital elements. Quat Res 9:139–167i Hazeleger W et al (2012) EC-Earth V2.2: description and validation of a new seamless earth system prediction model. Clim Dyn 39:2611–2629 Berntell E, Zhang Q, Chafik L, Körnich H (2018) Representation of multidecadal Sahel rainfall variability in 20th century reanalyses. Sci Rep 8:10937 Held IM, Delworth TL, Lu J, Findell KL, Knutson TR (2005) Simula- tion of Sahel drought in the 20th and 21st centuries. Proc Natl Acad Sci USA 102:17891. https://doi.org/10.1073/pnas.05090 57102 p Biasutti M (2013) Forced Sahel rainfall trends in the CMIP5 archive. J Geophys Res Atmos 118:1613–1623 Biasutti M, Giannini A (2006) Robust Sahel drying in response to late 20th century forcings. Geophys Res Lett. https://doi. org/10.1029/2006gl026067 Hoerling M, Hurrell J, Eischeid J, Phillips A (2006) Detection and attri- bution of twentieth-century northern and southern African rainfall bution of twentieth-century northern and southern African rainfall change. J Clim 19:3989–4008. https://doi.org/10.1175/jcli3842.1 y change. J Clim 19:3989–4008. https://doi.org/10.1175/jcli3842.1 Boone AA et al (2016) The regional impact of land-use land-cover change (LULCC) over West Africa from an ensemble of global climate models under the auspices of the WAMME2 project. Clim Dyn 47:3547–3573. https://doi.org/10.1007/s00382-016-3252-y Hwang Y-T, Frierson DMW, Kang SM (2013) Anthropogenic sulfate aerosol and the southward shift of tropical precipitation in the late 20th century. Geophys Res Lett 40:2845–2850. https://doi. org/10.1002/grl.50502 Booth BB, Dunstone NJ, Halloran PR, Andrews T, Bellouin N (2012) Aerosols implicated as a prime driver of twentieth-century North Atlantic climate variability. Nature 484:228–232 Janicot S (1992) Spatiotemporal variability of West African rainfall. Part I: regionalizations and typings. J Clim 5:489–497 Janicot S, Sultan B (2001) Intra-seasonal modulation of convection in the West African monsoon. Geophys Res Lett 28:523–526 Bouillon S, Maqueda MAM, Legat V, Fichefet T (2009) An elastic–vis- cous–plastic sea ice model formulated on Arakawa B and C grids. Ocean Model 27:174–184 Janicot S, Harzallah A, Fontaine B, Moron V (1998) West African monsoon dynamics and eastern equatorial Atlantic and Pacific SST anomalies (1970–88). J Clim 11:1874–1882 Bretherton CS, Smith C, Wallace JM (1992) An intercomparison of methods for finding coupled patterns in climate data. J Clim 5:541–560 Janowiak JE (1988) An investigation of interannual rainfall variability in Africa. Funding Open access funding provided by Stockholm University. Funding Open access funding provided by Stockholm University. Acknowledgements This work was supported by the Swedish Research Council (Vetenskapsrådet, Grant nos. 2013-06476 and 2017-04232). F.C.L. was supported by the Swedish Research Council (Vetenskap- srådet, Grant no 2018-01272), and worked with this article as a Pro Futura Scientia XIII Fellow funded by the Swedish Collegium for Advanced Study through Riksbankens Jubileumsfond. The ‘last mil- lennium’ EC-Earth simulation is performed on ECMWF’s comput- ing and archive facilities. The ‘historical’ EC-Earth simulation and all model data analyses were performed by resources provided by the Swedish National Infrastructure for Computing (SNIC) at the National Supercomputer Centre (NSC) partially funded by the Swedish Research Council through Grant agreement no. 2016-07213. Open Access This article is licensed under a Creative Commons Attri- bution 4.0 International License, which permits use, sharing, adapta- tion, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. Author contributions QZ conceived the research idea and designed the work. EB analysed the data and prepared the figures. QZ wrote the manuscript, EB and FCL commented, edited and provided input to the manuscript. QL performed the EC-Earth simulations. 1 3 1 3 Understanding the variability of the rainfall dipole in West Africa using the EC‑Earth last… 105 y 10.1175/2010jcli4065.1 Kageyama M et al (2018) The PMIP4 contribution to CMIP6—part 1: overview and over-arching analysis plan. Geosci Model Dev 11:1033–1057 Crowley TJ, Unterman MB (2013) Technical details concerning development of a 1200 yr proxy index for global volcanism. Earth Syst Sci Data 5:187–197 Kandji ST et al (2006) Opportunities for linking climate change adap- tation and mitigation through agroforestry systems. World Agro- forestry into the Future World Agroforesty Centre, Nairobi, pp 113–122 Dai A, Lamb PJ, Trenberth KE, Hulme M, Jones PD, Xie P (2004) The recent Sahel drought is real. Int J Climatol J R Meteorol Soc 24:1323–1331 Delworth TL, Zhang R, Mann ME (2007) Decadal to centennial vari- ability of the Atlantic from observations and models. Ocean Circ Mech Impacts Past Future Changes Merid Overturning 131–148 Knight JR, Folland CK, Scaife AA (2006) Climate impacts of the Atlantic multidecadal oscillation. Geophys Res Lett. https://doi. org/10.1029/2006GL026242 References J Clim 1:240–255 Caminade C, Terray L (2010) Twentieth century Sahel rainfall varia- bility as simulated by the ARPEGE AGCM, and future changes. Clim Dyn 35:75–94 Joly M, Voldoire A (2010) Role of the Gulf of Guinea in the inter- annual variability of the West African monsoon: what do we learn from CMIP3 coupled simulations? Int J Climatol 30:1843–1856 y Chang C-Y, Chiang JCH, Wehner MF, Friedman AR, Ruedy R (2011) Sulfate aerosol control of tropical Atlantic climate over the twentieth century. J Clim 24:2540–2555. https://doi. org/10.1175/2010jcli4065.1 Joly M, Voldoire A, Douville H, Terray P, Royer J-F (2007) African monsoon teleconnections with tropical SSTs: validation and evo- lution in a set of IPCC4 simulations. Clim Dyn 29:1–20 org/10.1029/2006GL026242 Dong B, Sutton R (2015) Dominant role of greenhouse-gas forcing in the recovery of Sahel rainfall. Nat Clim Change 5:757–760. https ://doi.org/10.1038/nclimate2664 Kuhnert H, Mulitza S (2011) Multidecadal variability and late medieval cooling of near-coastal sea surface temperatures in the eastern tropical North Atlantic. Paleoceanography 26:PA4224. https://doi. org/10.1029/2011PA002130 Emery W, Thomson R (2001) Time-series analysis methods. Data Anal Methods Phys Oceanogr 371:371–567 Lavender SL et al (2018) Estimation of the maximum annual number of North Atlantic tropical cyclones using climate models. Sci Adv 4:eaat6509. https://doi.org/10.1126/sciadv.aat6509 Folland C, Palmer T, Parker D (1986) Sahel rainfall and worldwide sea temperatures, 1901–85. Nature 320:602–607 García-Serrano J, Guemas V, Doblas-Reyes F (2015) Added-value from initialization in predictions of Atlantic multi-decadal variability. Clim Dyn 44:2539–2555 Liu Z, Zhang Q, Wu L (2004) Remote impact on tropical Atlantic climate variability: statistical assessment and dynamic assess- ment. J Clim 17:1529–1549 Giannini A, Saravanan R, Chang P (2003) Oceanic forcing of Sahel rainfall on interannual to interdecadal time scales. Science 302:1027–1030 Ljungqvist FC, Krusic PJ, Sundqvist HS, Zorita E, Brattström G, Frank D (2016) Northern Hemisphere hydroclimate variability over the past twelve centuries. Nature 532:94–98 Guemas V, García-Serrano J, Mariotti A, Doblas-Reyes F, Caron LP (2015) Prospects for decadal climate prediction in the Mediter- ranean region. Q J R Meteorol Soc 141:580–597 Ljungqvist FC et al (2019) Centennial-scale temperature change in last millennium simulations and proxy-based reconstructions. J Clim 32:2441–2482 Haarsma R et al (2020) HighResMIP versions of EC-Earth: EC- Earth3P and EC-Earth3P-HR—description, model computational performance and basic validation. Geosci Model Dev 13:3507– 3527. https://doi.org/10.5194/gmd-13-3507-2020 Ljungqvist FC et al (2019) European warm-season temperature and hydroclimate since 850 CE. Environ Res Lett 14:084015. https ://doi.org/10.1088/1748-9326/ab2c7e Losada T, Rodríguez-Fonseca B, Janicot S, Gervois S, Chauvin F, Ruti P (2010) A multi-model approach to the Atlantic equato- rial mode: impact on the West African monsoon. Clim Dyn 35:29–43 Hansen J, Sato M (2004) Greenhouse gas growth rates. Proc Natl Acad Sci USA 101:16109–16114 1 3 106 Q. Zhang et al. Rodwell MJ, Hoskins BJ (1996) Monsoons and the dynamics of deserts. Q J R Meteorol Soc 122:1385–1404 Losada T, Rodriguez-Fonseca B, Mohino E, Bader J, Janicot S, Mechoso C (2012) Tropical SST and Sahel rainfall: a non-sta- tionary relationship. Geophys Res Lett 39:L12705. https://doi. org/10.1029/2006GL026242 org/10.1029/2012GL052423 Roehrig R, Bouniol D, Guichard F, Hourdin F, Redelsperger J-L (2013) The present and future of the West African monsoon: a process- oriented assessment of CMIP5 simulations along the AMMA transect. J Clim 26:6471–6505 Macfarling Meure C et al (2006) Law Dome CO2, CH4 and N2O ice core records extended to 2000 years BP. Geophys Res Lett 33:L14810. https://doi.org/10.1029/2006GL026152 Rotstayn LD, Lohmann U (2002) Tropical rainfall trends and the indi- rect aerosol effect. J Clim 15:2103–2116 p g Madec G (2008) NEMO ocean general circulation model reference manuel. In: Internal report. LODYC/IPSL Parisl f Rowell DP (2003) The impact of Mediterranean SSTs on the Sahelian rainfall season. J Clim 16:849–862 Maley J (1993) Chronologie calendaire des principales fluctuations du lac Tchad au cours du dernier millénaire Datation et Chronologie dans le Bassin du Lac Tchad ORSTOM, Paris, pp 161–163 Rowell DP, Folland CK, Maskell K, Ward MN (1995) Variability of summer rainfall over tropical North Africa (1906–92): observa- tions and modelling. Q J R Meteorol Soc 121:669–704 dans le Bassin du Lac Tchad ORSTOM, Paris, pp 161– Mann ME, Lees JM (1996) Robust estimation of background noise and signal detection in climatic time series. Clim Change 33:409–445 Schmidt G et al (2011) Climate forcing reconstructions for use in PMIP simulations of the last millennium (v1.0). Geosci Model Dev 4:33–45 g g Mann ME et al (2009) Global signatures and dynamical origins of the Little Ice Age and Medieval Climate Anomaly. Science 326:1256–1260 Schmidt G et al (2012) Climate forcing reconstructions for use in PMIP simulations of the Last Millennium (v1.1). Geosci Model Dev 5:185–191 Martin ER, Thorncroft C, Booth BB (2014) The multidecadal Atlantic SST—Sahel rainfall teleconnection in CMIP5 simulations. J Clim 27:784–806 Shanahan TM et al (2009) Atlantic forcing of persistent drought in West Africa. Science 324:377–380 Masih I, Maskey S, Mussá F, Trambauer P (2014) A review of droughts on the African continent: a geospatial and long-term perspective. Hydrol Earth Syst Sci 18:3635–3649 Sheen K, Smith D, Dunstone N, Eade R, Rowell D, Vellinga M (2017) Skilful prediction of Sahel summer rainfall on inter-annual and multi-year timescales. Nat Commun 8:1–12 Miller GH et al (2012) Abrupt onset of the Little Ice Age triggered by volcanism and sustained by sea–ice/ocean feedbacks. Geophys Res Lett. 39:L02708. https://doi.org/10.1029/2011GL050168 Stevenson S et al (2018) Climate variability, volcanic forcing, and last millennium hydroclimate extremes. org/10.1029/2006GL026242 J Clim 31:4309–4327 Stocker T et al (2013) IPCC, 2013: Climate change 2013: the physical science basis. Contribution of working group I to the Fifth Assess- ment report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge Mohino E et al (2011) Changes in the interannual SST-forced signals on West African rainfall. AGCM intercomparison. Clim Dyn 37:1707–1725 Mohino E, Keenlyside N, Pohlmann H (2016) Decadal prediction of Sahel rainfall: where does the skill (or lack thereof) come from? Clim Dyn 47:3593–3612l Street-Perrott FA et al (2000) Drought and dust deposition in the West African Sahel: a 5500-year record from Kajemarum Oasis, north- eastern Nigeria. Holocene 10:293–302 y Mulitza S et al (2010) Increase in African dust flux at the onset of com- mercial agriculture in the Sahel region. Nature 466:226 Sultan B, Janicot S (2003) The West African monsoon dynamics. Part II: The “preonset” and “onset” of the summer monsoon. J Clim 16:3407–3427 Nash DJ et al (2016) African hydroclimatic variability during the last 2000 years. Quat Sci Rev 154:1–22l Nicholson SE (1980) The nature of rainfall fluctuations in subtropical West Africa. Mon Weather Rev 108:473–487 Talbot M, Delibrias G (1977) Holocene variations in the level of Lake Bosumtwi, Ghana. Nature 268:722f Taylor KE, Stouffer RJ, Meehl GA (2012) An overview of CMIP5 and the experiment design. Bull Am Meteorol Soc 93:485–498 Nicholson SE (2013) The West African Sahel: a review of recent stud- ies on the rainfall regime and its interannual variability. ISRN Meteorol 2013:453521 Ting M, Kushnir Y, Seager R, Li C (2011a) Robust features of Atlantic multi-decadal variability and its climate impacts. Geophys Res Lett. 38:L17705. https://doi.org/10.1029/2011GL048712 Nicholson SE, Palao IM (1993) A re-evaluation of rainfall variability in the Sahel. Part I. Characteristics of rainfall fluctuations. Int J Climatol 13:371–389l Ting M, Kushnir Y, Seager R, Li C (2011b) Robust features of Atlantic multi‐decadal variability and its climate impacts Geophys Res Lett 38l Nnamchi HC, Li J (2011) Influence of the South Atlantic Ocean dipole on West African summer precipitation. J Clim 24:1184–1197 Vieira LEA, Solanki SK (2010) Evolution of the solar magnetic flux on time scales of years to millenia. Astron Astrophys 509:A100 Nouaceur Z, Murărescu O (2016) Rainfall variability and trend analysis of annual rainfall in North Africa. Int J Atmos Sci 2016:7230450 Odoulami RC, Akinsanola A (2018) Recent assessment of West Afri- can summer monsoon daily rainfall trends. Wilks DS (2011) Statistical methods in the atmospheric sciences, vol 100. Academic Press, New York Xue Y et al (2016) West African monsoon decadal variability and surface-related forcings: second West African Monsoon Mode- ling and Evaluation Project Experiment (WAMME II). Clim Dyn 47:3517–3545 Wilks DS (2011) Statistical methods in the atmospheric sciences, vol 100. Academic Press, New York Xue Y et al (2016) West African monsoon decadal variability and surface-related forcings: second West African Monsoon Mode- ling and Evaluation Project Experiment (WAMME II). Clim Dyn 47:3517–3545 Zebiak SE (1993) Air–sea interaction in the equatorial Atlantic region. J Clim 6:1567–1586 Zhang R, Delworth TL (2006) Impact of Atlantic multidecadal oscilla- tions on India/Sahel rainfall and Atlantic hurricanes. Geophys Res Lett 33:L17712. https://doi.org/10.1029/2006GL026267 org/10.1029/2006GL026242 Weather 73:283–287 Villamayor J, Mohino E (2015) Robust Sahel drought due to the inter- decadal Pacific oscillation in CMIP5 simulations. Geophys Res Lett 42:1214–1222 Otero N, Mohino E, Gaetani M (2016) Decadal prediction of Sahel rainfall using dynamics-based indices. Clim Dyn 47:3415–3431li Wang Y-M, Lean J, Sheeley N Jr (2005) Modeling the Sun’s magnetic field and irradiance since 1713. Astrophys J 625:522 Palmer T (1986) Influence of the Atlantic, Pacific and Indian oceans on Sahel rainfall. Nature 322:251–253 i Wang G, Yu M, Xue Y (2016) Modeling the potential contribution of land cover changes to the late twentieth century Sahel drought using a regional climate model: impact of lateral boundary condi- tions. Clim Dyn 47:3457–3477f Park J-Y, Bader J, Matei D (2016) Anthropogenic Mediterranean warm- ing essential driver for present and future Sahel rainfall. Nat Clim Change 6:941–945 Wang J, Yang B, Ljungqvist FC, Luterbacher J, Osborn TJ, Briffa KR, Zorita E (2017) Internal and external forcing of multidecadal Atlantic climate variability over the past 1,200 years. Nat Geosci 10:512–517 Pongratz J, Reick C, Raddatz T, Claussen M (2008) A reconstruc- tion of global agricultural areas and land cover for the last mil- lennium. Glob Biogeochem Cycles 22:GB3018: https://doi. org/10.1029/2007GB003153 g Rodríguez-Fonseca B et al (2011) Interannual and decadal SST-forced responses of the West African monsoon. Atmos Sci Lett 12:67–74 Ward MN (1998) Diagnosis and short-lead time prediction of summer rainfall in tropical North Africa at interannual and multidecadal timescales. J Clim 11:3167–3191 Rodríguez-Fonseca B et al (2015) Variability and predictability of West African droughts: a review on the role of sea surface temperature anomalies. J Clim 28:4034–4060 1 3 Understanding the variability of the rainfall dipole in West Africa using the EC‑Earth last… 107 Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Wilks DS (2011) Statistical methods in the atmospheric sciences, vol 100. Academic Press, New York Xue Y et al (2016) West African monsoon decadal variability and surface-related forcings: second West African Monsoon Mode- ling and Evaluation Project Experiment (WAMME II). Clim Dyn 47:3517–3545 Zebiak SE (1993) Air–sea interaction in the equatorial Atlantic region. J Clim 6:1567–1586 Zhang R, Delworth TL (2006) Impact of Atlantic multidecadal oscilla- tions on India/Sahel rainfall and Atlantic hurricanes. Geophys Res Lett 33:L17712. https://doi.org/10.1029/2006GL026267 1 3 1 3 3
https://openalex.org/W2618390447	https://europepmc.org/articles/pmc5453992?pdf=render	English	null	Time-order-errors and duration ranges in the Episodic Temporal Generalization task	Scientific reports	2,017	cc-by	9,042	Time-order-errors and duration ranges in the Episodic Temporal Generalization task Ezequiel Mikulan1,2, Manuel Bruzzone1, Manuel Serodio1, Mariano Sigman3, Tristán Bekinschtein4, Adolfo M. García1,2,5, Lucas Sedeño1,2 & Agustín Ibáñez 1,2,6,7,8 Ezequiel Mikulan1,2, Manuel Bruzzone1, Manuel Serodio1, Mariano Sigman3, Tristán Bekinschtein4, Adolfo M. García1,2,5, Lucas Sedeño1,2 & Agustín Ibáñez 1,2,6,7,8 Received: 28 October 2016 Accepted: 11 April 2017 Published: xx xx xxxx Received: 28 October 2016 Accepted: 11 April 2017 Published: xx xx xxxx The current model of the Episodic Temporal Generalization task, where subjects have to judge whether pairs of auditory stimuli are equal in duration, predicts that results are scale-free and unaffected by the presentation order of the stimuli. To test these predictions, we conducted three experiments assessing sub- and supra-second standards and taking presentation order into account. Proportions were spaced linearly in Experiments 1 and 2 and logarithmically in Experiment 3. Critically, we found effects of duration range and presentation order with both spacing schemes. Our results constitute the first report of presentation order effects in the Episodic Temporal Generalization task and demonstrate that future studies should always consider duration range, number of trials and presentation order as crucial factors modulating performance. Time has been a matter of ardent debate across many disciplines1. Within neuroscience, not only does it consti- tute an important topic in its own right2, but it also impinges on the field’s fundamental areas of inquiry, includ- ing consciousness3, motor control4, memory5, artificial intelligence6, and neural dynamics7. Likewise, the study of timing abnormalities is pivotal for research on pathologies such as Parkinson’s disease and schizophrenia8. In brief, understanding timing, time perception, and their neural basis proves fundamental for contemporary neuroscience9. Even though many relevant models and theories have been proposed10, some basic questions are still unsol Here we aimed to address one of them: is timing equal across different scales? f Multiple studies have addressed this issue based on a distinction between sub- and supra-second durations11, with contradictory results. Some studies show that response variability increases linearly as a function of duration, thus following Weber’s Law, in a range that goes from a few hundred milliseconds to a few seconds –see ref. 12 for a systematic investigation. However, other reports indicate that this linear property stops holding at some point between 1 and 2 seconds –see ref. www.nature.com/scientificreports www.nature.com/scientificreports www.nature.com/scientificreports Received: 28 October 2016 Accepted: 11 April 2017 Published: xx xx xxxx Scientific Reports \| 7: 2643 \| DOI:10.1038/s41598-017-02386-9 Method i i Participants. Eighteen subjects participated in Experiment 1 (F = 10; x age = 24.22; s age = 3.12; x years of education = 17.78; s years of education = 3.83), 20 in Experiment 2 (F = 10; x age = 25.47; s age = 3.94; x years of education = 18.87; s years of education = 3.04), and 18 in Experiment 3 (F = 11; x age = 24.39; s age = 3.53; x years of education = 18.94; s years of education = 2.82), after signing informed consent. Subjects participated in only one of the experiments. All of them reported normal hearing, right-handedness, and absence of neurological and psychiatric antecedents, and they were naïve as to the purpose of the study. The experiments were approved by the local ethics committee (INECO Foundation) and were conducted in accordance with the Declaration of Helsinki. Stimuli. Auditory stimuli were 500-Hz tones created and delivered using Matlab (Mathworks Inc.) and Psychtoolbox27 through Sennheiser HD202 headphones at 65db on a MacBook Pro notebook. Before the experi- ment, the software created 7 blocks of 16 trials; half of the stimuli corresponded to the Sub-Second condition and the other half to the Supra-Second one. For each trial, a Standard (S) duration was selected from a uniform distri- bution, ranging from 300 to 500 ms in the Sub-Second condition, and from 1200 to 2000 ms in the Supra-Second one. Then, S was multiplied by one of 7 ratios (linearly spaced in Experiments 1 and 2: 0.25, 0.50, 0.75, 1.0, 1.25, 1,50, 1.75; logarithmically spaced in Experiment 3: 0.25, 0.40, 0.63, 1, 1.59, 2.52, 4) to create a Comparison (C) duration, so that on each block there was one trial of each ratio for each condition (i.e., 7 ratios × 2 condi- tions = 14 trials). An additional 1.0 ratio trial was then added to each condition. Finally, the order of the trials was randomized and each trial was assigned a random counterbalanced presentation order so that on half of the trials the S was presented first, and on the other half the first stimulus presented was C. In Experiment 1 a total of 112 trials were obtained from each subject in an average of 16 minutes (σ = 36 s). www.nature.com/scientificreports/ www.nature.com/scientificreports/ of ratios (e.g., 0.5, 1, 1.5). In the original version, a clear standard is learnt at the beginning and participants judge whether it is equal to each of the following durations. In the Episodic version, presentation order is counterbal- anced so that in half of the trials the standard comes first and in the other half it comes second. of ratios (e.g., 0.5, 1, 1.5). In the original version, a clear standard is learnt at the beginning and participants judge whether it is equal to each of the following durations. In the Episodic version, presentation order is counterbal- anced so that in half of the trials the standard comes first and in the other half it comes second. i Both tasks showed a similar pattern of results with sub-second durations: the obtained psychometric functions were asymmetrical, with a higher proportion of “equal” responses on the right tail, that is, when the ratio was higher than 116, 17. The same pattern was found using the original standard version, with durations ranging from 2 up to 8 seconds18. Importantly, results were superimposed between duration ranges when plotted in a relative scale, that is, as a function of the proportion between standard and comparison durations. This was interpreted as confirmation of the common timing hypothesis. However, no studies have yet used supra-second durations as standards in the Episodic version.h p Traditionally, temporal generalization results have been interpreted within Scalar Expectancy Theory, which, in brief, states that durations are estimated via accumulation of pulses. Within this framework, in the temporal generalization task subjects would compare the two values of the estimated durations and then decide based on their normalized absolute difference19. Therefore, according to this model, results should not be affected by pres- entation order of the stimuli, something known as the “balance condition”20, nor by duration range. Presentation order effects have been shown in a wide variety of time perception tasks, termed “time-order-errors” (TOE) within this context (see ref. 21 for a review), but never in the Episodic Temporal Generalization (ETG) task. In fact, no previous studies using the task tested for any effect of this kind17, 22–24. p g yf In addition, evaluating symmetry with linearly spaced proportions (i.e. 0.25, 0.50, 0.75, 1, 1.25, 1.50, 1.75), as is usually the case with the ETG task17, 19, 22, implies an unbalanced comparison. www.nature.com/scientificreports/ Symmetry within this setting would indicate that a similar amount of “equal” responses were obtained when comparing, for example, ratios 0.25–1 (1:4) and 1.75–1 (1.75:1) or vice versa (4:1 and 1:1.75). A more meaningful comparison would arise from using logarithmically spaced proportions, so that, following the example above, ratios, 0.25–1 (1:4) and 4–1 (4:1) could be contrasted. Moreover, the property of superposition has been tested via visual inspection or ANOVAs17, 19, none of which is robust to such an end. The first one proves inadequate because it does not establish a decisional boundary to accept or reject hypotheses, and the second because it can produce spurious results when used with proportional data25, 26. A more convenient approach would be to compare Weber Fractions (WF) between duration ranges and test whether they remain constant, in which case a scalar relationship could be assumed to exist between them. Against this background, our study pursued three main objectives. First, we tested the prediction, derived from the traditional ETG model, that presentation order had no effect on performance. Second, we examined the symmetry/asymmetry of the temporal generalization gradients taking presentation order into account and using linear and logarithmically spaced proportions, so that symmetry could be properly assessed. Third, we compared WFs of sub- and supra-second ranges to test their compliance with the scalar property of timing. g y g To address these aims, we conducted three experiments. Experiment 1 was designed with a number of trials similar to that of previous studies using the task17, 22 and comprised linearly spaced proportions. As taking pres- entation order into account reduced the number of trials of each ratio to a half, we conducted Experiment 2, in which the same task was administered but with a threefold increase in trials. Experiment 3 had the same number of trials as Experiment 2 but proportions were logarithmically spaced. With this combination of experiments, we aimed to address some critical gaps in the ETG framework. Time-order-errors and duration ranges in the Episodic Temporal Generalization task 13 for a review.h This controversy is epitomized by the antinomy between two major conceptual frameworks used to account for timing mechanisms in the brain: the “common timing hypothesis” and the “distinct timing hypothesis”14. Whereas the former assumes a single timing mechanism irrespective of duration, the latter posits dissociable mechanisms for sub- and supra-second durations. A typical approach to study time perception is to have participants judge whether two durations are equal15. This so-called Temporal Generalization task has two main versions for humans. In the original one16, participants learn a standard duration at the beginning of the experiment and are then presented with several to-be-compared durations. Instead, in the Episodic version17, subjects judge the durations of two successive stimuli on a trial-by-trial basis. Stimuli are constructed in a similar way for both versions: a set of comparison durations is generated multiplying a standard duration (e.g., 400 ms, or values from a range such as 300 to 500 ms) by a series 1Laboratory of Experimental Psychology and Neuroscience (LPEN), Institute of Cognitive and Translational Neuroscience (INCYT), INECO Foundation, Favaloro University, Buenos Aires, Argentina. 2National Scientific and Technical Research Council (CONICET), Buenos Aires, Argentina. 3Di Tella University, Buenos Aires, Argentina. 4Consciousness and Cognition Lab, Department of Psychology, University of Cambridge, Cambridge, UK. 5Faculty of Education, National University of Cuyo (UNCuyo), Mendoza, Argentina. 6Universidad Autónoma del Caribe, Barranquilla, Colombia. 7Center for Social and Cognitive Neuroscience (CSCN), School of Psychology, Universidad Adolfo Ibañez, Santiago de Chile, Chile. 8Australian Research Council Centre of Excellence in Cognition and its Disorders, Sydney, Australia. Correspondence and requests for materials should be addressed to E.M. (email: e.mikulan@gmail.com) or A.I. (email: aibanez@ineco.org.ar) Scientific Reports \| 7: 2643 \| DOI:10.1038/s41598-017-02386-9 1 Method i i In Experiment 2, the task was repeated three times with 5 minute breaks between them, therefore 336 trials were obtained from each subject, in 62 minutes on average (σ = 9 m). In Experiment 3, 336 trials were obtained from each subject in 68 minutes on average (σ = 5 m). See Supplementary Figures 1, 2, and 3 (S1, S2, and S3) for further details. Procedure. Participants were informed that they would hear sequences of two tones and that their task was to decide whether both sounds had the same duration (Fig. 1). Each trial started with a 5-s inter-trial interval that was followed by the presentation of a tone, a gap (randomly chosen from a uniform distribution from 400 Scientific Reports \| 7: 2643 \| DOI:10.1038/s41598-017-02386-9 2 www.nature.com/scientificreports/ Figure 1. Diagram of the experimental paradigm. Blue rectangles represent Standard durations and green rectangles represent Comparison durations. Linearly spaced proportions (LIN) were used in Experiments 1 and 2 and logarithmically spaced proportions (LOG) in Experiment 3. Figure 1. Diagram of the experimental paradigm. Blue rectangles represent Standard durations and green rectangles represent Comparison durations. Linearly spaced proportions (LIN) were used in Experiments 1 and 2 and logarithmically spaced proportions (LOG) in Experiment 3. Figure 1. Diagram of the experimental paradigm. Blue rectangles represent Standard durations and green rectangles represent Comparison durations. Linearly spaced proportions (LIN) were used in Experiments 1 and 2 and logarithmically spaced proportions (LOG) in Experiment 3. to 600 ms), and a second tone. Participants had to respond with their right hand on the notebook keyboard. To indicate that the tones were equal, they had to press the down arrow key with their index finger; to indicate that they were not, they had to press the right arrow with their middle finger. Importantly, post-task debriefing showed that although all participants detected two duration ranges, none of them realized that there were standard and comparison distributions. to 600 ms), and a second tone. Participants had to respond with their right hand on the notebook keyboard. To indicate that the tones were equal, they had to press the down arrow key with their index finger; to indicate that they were not, they had to press the right arrow with their middle finger. Importantly, post-task debriefing showed that although all participants detected two duration ranges, none of them realized that there were standard and comparison distributions. Data Analysis. Method i i Statistical analyses were performed on R software28. All subjects were included in them. Following previous reports18, we plotted the temporal generalization gradients as a function of comparison dura- tions and tested their asymmetry by comparing the proportion of “equal” responses on the three Ratios below 1 (C-Shorter), against the proportion of “equal” responses on the three Ratios above it (C-Longer) using Wilcoxon Signed-Rank tests. Statistical results were corrected for multiple comparisons using the Holm-Bonferroni method.i In order to characterize the TOE, we first estimated the point of subjective equality (PSE) using the smoothing spline curve-fitting method from Matlab’s (Mathworks Inc.) Curve Fitting Toolbox (with automatic selection of the smoothing parameter), and finding the maximum of the resulting curve. We then defined the TOE following Fechner’s definition of constant error (CE)21: (1) = − CE PSE st (1) = − CE PSE st where PSE denotes the point of subjective equality and st the standard duration. Within this context, the sign of CE denotes the sign of the TOE when the standard is presented first. When the standard was presented in second place, the sign of the TOE was computed as st – PSE. We report the magnitude of the TOE as a percent- age of the standard (%TOE)29. %TOEs were submitted to a rm-ANOVA with standard duration (Sub-second/ Supra-second) and order (S-C/C-S) as factors. b ( ) l l d Weber Fractions (WF) were calculated as: = WF DL PSE / (2) (2) = WF DL PSE / = WF DL PSE / where DL denotes the difference limen and PSE the point of subjective equality30. Within this analysis, the PSE and DL were calculated as the mean and standard deviation of a fitted Gaussian function, respectively. WFs were submitted to a rm-ANOVA with standard duration (Sub-second/Supra-second) and order (S-C/C-S) as factors. To control that the PSEs did not differ between methods we compared them using a Wilcoxon Signed-Rank test.f f p g g Effect sizes in all cases were calculated via generalized eta squared (η2 G)31; these were considered as small if η2 G = 0.02, medium if η2 G = 0.13, and large if η2 G = 0.26. Holm-Bonferroni corrected post hoc t-tests were used for pairwise comparisons. Results i Experiment 1. The proportion of “equal” responses (PE) for each ratio and duration, when collapsing pres- entation orders, is shown in Fig. 2. Even though visual inspection suggests that both temporal generalization gradients were asymmetrical, Wilcoxon tests proved that this was significant only in the Sub-Second condition (V = 135.5, p < 0.01), where a higher PE was found when C-Ratio > 1. In the Supra-Second condition, the differ- ence was not significant (V = 220, p = 0.12).h gi ( , p ) The PE for each ratio and duration when including presentation order as an additional variable is shown in Fig. 3. Visual inspection again suggests that all temporal generalization gradients were asymmetrical, which, in this case, was true for all comparisons. In the Sub-Second condition, both presentations orders, S-C (V = 30, p < 0.05) and C-S (V = 40, p < 0.05) had a greater PE when C-Ratio > 1. In the Supra-Second condition, the C-S presentation order also had a greater PE when C-Ratio > 1 (V = 139, p < 0.01). Interestingly, in the S-C order of the Supra-Second condition the asymmetry was in the opposite direction, that is, the PE was higher when C-Ratio < 1 (V = 1, p < 0.01). Visual inspection also suggests that the temporal generalization gradients of the Supra-second condition are shifted to the left and right in the S-C and C-S orders, respectively.h ptt g p y This part of the analysis revealed that the asymmetries of temporal generalization gradients were not e between the two duration ranges considered. Moreover, when presentation order of the stimuli was taken Scientific Reports \| 7: 2643 \| DOI:10.1038/s41598-017-02386-9 3 www.nature.com/scientificreports/ Figure 2. Temporal generalization gradients of Experiment 1 (linearly spaced ratios). Proportion of “equal” responses as a function of the ratio of the comparison duration when collapsing presentation orders Vertical lines represent 95% confidence levels. Figure 2. Temporal generalization gradients of Experiment 1 (linearly spaced ratios). Proportion of “equal” responses as a function of the ratio of the comparison duration when collapsing presentation orders Vertical lines represent 95% confidence levels. Figure 3. Temporal generalization gradients of Experiment 1 (linearly spaced ratios) by presentation order. Proportion of “equal” responses for the Sub-Second (left) and Supra-Second (right) conditions by presentation order. Vertical lines represent 95% confidence levels. Figure 3. Temporal generalization gradients of Experiment 1 (linearly spaced ratios) by presentation order. Results i In the Supra-Second condition, the difference was not significant (V = 282, p = 0.17).h p The PE for each ratio and duration when including presentation order as an additional variable is shown in Fig. 5. Visual inspection once more suggests that all temporal generalization gradients were asymmetrical, which, in this case, was true for all comparisons but one. In the Sub-Second condition, the S-C presentation order had a higher PE when C-Ratio > 1 (V = 1, p < 0.001) and in the C-S order the difference was not significant (V = 76, p = 0.29). As in Experiment 1, in the Supra-Second condition, the C-S presentation order had a greater PE when C-Ratio > 1 (V = 12, p < 0.001) and in the S-C order the asymmetry was in the opposite direction, that is, the PE was higher when C-Ratio < 1 (V = 196, p < 0.01). Visual inspection also suggests that temporal generalization gradients are shifted but this time also in the Sub-second range. Interestingly, the directions of the shifts seem to be inverted between duration ranges and orders.h g This part of the analysis revealed that, as in Experiment 1, the asymmetries of the temporal generalization gradients were not equal between the duration ranges considered in the study. Experiment 2 showed that when collecting more trials per subject, the asymmetry of the C-S order of the Sub-second condition was not signifi- cant, in contrast with Experiment 1. Furthermore, Experiment 2 suggests that the shifts of the temporal gener- alization gradients that were observed in the Supra-second condition also appeared in the Sub-second range but with opposite sign. In order to characterize the presentation order effect (TOE) we computed the %TOE of each subject of Experiment 2 (Fig. 6) and conducted a rm-ANOVA. It showed a main effect of Duration (F1,19 = 93.15, p < 0.001, η2 G = 0.48), with a higher %TOE in the Sub-second condition; a main effect of Order (F1,19 = 27.36, p < 0.001, η2 G = 0.14), with a higher %TOE in the S-C order; and a non-significant Duration x Order interaction (F1,19 = 0.58, p = 0.45, η2 G = 0.009). Results i Proportion of “equal” responses for the Sub-Second (left) and Supra-Second (right) conditions by presentation order. Vertical lines represent 95% confidence levels. account, results suggested that this occurred because of a shift of the temporal generalization gradients in the Supra-Second condition that depended on the presentation order of the stimuli (Fig. 3). The presentation order effect in combination with the asymmetrical gradients (the latter probably due to the linear spacing of the com- parison proportions) undermined the comparison of WFs, as they rely on the estimation of the spread of the temporal generalization gradients. This analysis was therefore not carried out for this Experiment. account, results suggested that this occurred because of a shift of the temporal generalization gradients in the Supra-Second condition that depended on the presentation order of the stimuli (Fig. 3). The presentation order effect in combination with the asymmetrical gradients (the latter probably due to the linear spacing of the com- parison proportions) undermined the comparison of WFs, as they rely on the estimation of the spread of the temporal generalization gradients. This analysis was therefore not carried out for this Experiment. p g gh y p In Experiment 1, we used a similar amount of trials than previous studies that employed the task17, 22. Partitioning trials by presentation order left approximately 3–4 trials per ratio and order combination for each Scientific Reports \| 7: 2643 \| DOI:10.1038/s41598-017-02386-9 4 www.nature.com/scientificreports/ Figure 4. Temporal generalization gradients of Experiment 2 (linearly spaced ratios). Proportion of “equal” responses as a function of the ratio of the comparison duration when collapsing presentation orders Vertical lines represent 95% confidence levels. Figure 4. Temporal generalization gradients of Experiment 2 (linearly spaced ratios). Proportion of “equal” responses as a function of the ratio of the comparison duration when collapsing presentation orders Vertical lines represent 95% confidence levels. subject, which might have led to inaccurate results. In order to overcome this limitation we replicated Experiment 1 but this time we collected three times more trials per subject. Experiment 2. The PE for each ratio and duration, when collapsing presentation orders, is shown in Fig. 4. Visual inspection again suggests that both temporal generalization gradients were asymmetrical, and again Wilcoxon tests proved that this was significant only in the Sub-Second condition (V = 89, p < 0.001), where a higher PE was found when C-Ratio > 1. Results i The %TOE was positive in the Sub-second range (x = 5.68) and negative in the Supra-second range (x = −13.79).ff g Again, the combination of the different presentation order effects and the asymmetrical gradients hindered the comparison of WFs, so no such analysis was conducted. The linear spacing of the comparison proportions probably caused the asymmetrical gradients. We therefore conducted a third experiment, this time with loga- rithmically spaced ratios, so that the resulting gradients presumably became more symmetrical and allowed a meaningful comparison of WFs. Experiment 3. The PE when collapsing presentation order is shown in Fig. 7. Here, visual inspection sug- gests that temporal generalization gradients were not asymmetrical, which Wilcoxon tests proved to be correct (Sub-second: V = 417.5, p = 0.28; Supra-second: V = 274, p = 1).h The PE for each ratio and duration when including presentation order as an additional variable is shown in Fig. 8. Visual inspection suggests that all temporal generalization gradients were symmetrical, which was true for all comparisons but one. The only significantly asymmetrical gradient was found in the Supra-second S-C con- dition with a higher PE when C-Ratio < 1 (V = 108, p < 0.05). All remaining comparisons were not significant (Sub-second S-C: V = 66, p = 1; Sub-second C-S: V = 144.5, p = 0.54; Supra-second C-S: V = 32, p = 0.08). Scientific Reports \| 7: 2643 \| DOI:10.1038/s41598-017-02386-9 5 www.nature.com/scientificreports/ Figure 5. Temporal generalization gradients of Experiment 2 (linearly spaced ratios) by presentation order. Proportion of “equal” responses for the Sub-Second (left) and Supra-Second (right) conditions by presentation order. Vertical lines represent 95% confidence levels. Figure 5. Temporal generalization gradients of Experiment 2 (linearly spaced ratios) by presentation order. Proportion of “equal” responses for the Sub-Second (left) and Supra-Second (right) conditions by presentation order. Vertical lines represent 95% confidence levels. Figure 6. Time-order-errors. Violin plots of the magnitude and sign of the time-order-errors for each duration range and presentation order, as percentage of standard duration, across participants of Experiment 2 (linearly spaced ratios). The corresponding ANOVA showed a main effect of duration range (p < 0.001, η2 G = 0.48), a main effect of presentation order (p < 0.001, η2 G = 0.14) and a non-significant interaction (p = 0.45, η2 G = 0.009). Figure 6. Time-order-errors. Results i Violin plots of the magnitude and sign of the time-order-errors for each duration range and presentation order, as percentage of standard duration, across participants of Experiment 2 (linearly spaced ratios). The corresponding ANOVA showed a main effect of duration range (p < 0.001, η2 G = 0.48), a main effect of presentation order (p < 0.001, η2 G = 0.14) and a non-significant interaction (p = 0.45, η2 G = 0.009). The TOE analysis from Experiment 3 (Fig. 9) showed a main effect of Order (F1,17 = 2344.41, p < 0.001, η2 G = 0.90), with a higher %TOE in the C-S order; a main effect of Duration (F1,17 = 16.00, p < 0.001, η2 G = 0.21), with a higher %TOE in the Sub-second condition; and a non-significant Duration x Order interaction (F1,17 = 4.20, p = 0.056, η2 G = 0.01). The %TOE was negative in the C-S order (x = −29.38) and positive in the S-C order (x = 24.42).h The more symmetrical gradients obtained using logarithmically spaced ratios allowed us to compute and compare WFs. We first tested that the PSEs obtained for this analysis did not differ from the PSEs obtained in the TOE analysis. Wilcoxon tests proved that they were not significantly different (V = 1024, p = 0.1). We then proceeded to the WF analysis (Fig. 10), where the rm-ANOVA showed a main effect of Duration (F1,17 = 15.87, p < 0.001, η2 G = 0.12), with higher values in the Sub-second condition; a main effect of Order (F1,17 = 5.34, p < 0.05, η2 G = 0.02), with higher values in the C-S order; and a significant Duration x Order interaction (F1,17 = 7.25, Scientific Reports \| 7: 2643 \| DOI:10.1038/s41598-017-02386-9 6 www.nature.com/scientificreports/ Figure 7. Temporal generalization gradients of Experiment 3 (logarithmically spaced ratios). Proportion of “equal” responses as a function of the ratio of the comparison duration when collapsing presentation orders Vertical lines represent 95% confidence levels. Figure 7. Temporal generalization gradients of Experiment 3 (logarithmically spaced ratios). Proportion of “equal” responses as a function of the ratio of the comparison duration when collapsing presentation orders Vertical lines represent 95% confidence levels. Figure 8. Temporal generalization gradients of Experiment 3 (logarithmically spaced ratios) by presentation order. Proportion of “equal” responses for the Sub-Second (left) and Supra-Second (right) conditions by presentation order. Vertical lines represent 95% confidence levels. Figure 8. www.nature.com/scientificreports/ Figure 9. Time-order-errors. Violin plots of the magnitude and sign of the time-order-errors for each duration range and presentation order, as percentage of standard duration, across participants of Experiment 3 (logarithmically spaced ratios). The corresponding ANOVA showed a main effect of duration range (p < 0.001, η2 G = 0.21), a main effect of presentation order (p < 0.001, η2 G = 0.90) and a non-significant interaction (p = 0.056, η2 G = 0.01). Figure 9. Time-order-errors. Violin plots of the magnitude and sign of the time-order-errors for each duration range and presentation order, as percentage of standard duration, across participants of Experiment 3 (logarithmically spaced ratios). The corresponding ANOVA showed a main effect of duration range (p < 0.001, η2 G = 0.21), a main effect of presentation order (p < 0.001, η2 G = 0.90) and a non-significant interaction (p = 0.056, η2 G = 0.01). Figure 10. Weber Fractions. Violin plots of the Weber Fractions for each duration range and presentation order, across participants of Experiment 3 (logarithmically spaced ratios). The corresponding ANOVA showed a main effect of duration range (p < 0.001, η2 G = 0.12), a main effect of presentation order (p < 0.05, η2 G = 0.02) and a significant interaction (p < 0.05, η2 G = 0.04). Holm-Bonferroni corrected post hoc tests showed that the S-C order of the Supra-second condition differed from all others (vs. Sub-second S-C: p < 0.001; vs. Sub- second C-S: p < 0.01; vs. Supra-second C-S: p < 0.01) and that all remaining comparisons were not statistically significant. Figure 10. Weber Fractions. Violin plots of the Weber Fractions for each duration range and presentation order, across participants of Experiment 3 (logarithmically spaced ratios). The corresponding ANOVA showed a main effect of duration range (p < 0.001, η2 G = 0.12), a main effect of presentation order (p < 0.05, η2 G = 0.02) and a significant interaction (p < 0.05, η2 G = 0.04). Holm-Bonferroni corrected post hoc tests showed that the S-C order of the Supra-second condition differed from all others (vs. Sub-second S-C: p < 0.001; vs. Sub- second C-S: p < 0.01; vs. Supra-second C-S: p < 0.01) and that all remaining comparisons were not statistically significant. Discussion Th fi The first aim of our study was to assess whether performance on the ETG task is sensitive to presentation order effects. We showed that TOEs appeared in the two duration ranges under consideration, and that this effect held for both linear and logarithmic spacing. Interestingly, effect sizes between Experiments 2 and 3 were inverted. In the linearly spaced experiment, the duration range showed a large effect (η2 G = 0.48) and presentation order a medium one (η2 G = 0.14), while the use of logarithmically spaced proportions showed a large effect for pres- entation order (η2 G = 0.90) and a medium one for duration range (η2 G = 0.21). Together with the observed signs of the TOEs, these results reveal different patterns between experiments. In the linearly spaced experiment, sub- jects overestimated the first sound in the Sub-second condition and the second sound in the Supra-second range. Instead, when proportions were logarithmically spaced, they overestimated the Comparison duration, with a smaller influence of its position and duration range.h l p g TOEs have been reported in a wide range of tasks21 but never in the ETG task. They have mainly been reported when the standard duration was fixed across trials, even though they have been also reported in experiments where both stimuli varied32, 33. Having a standard that repeats from trial to trial implies that memory and learning can play a major role in the obtained results and might also introduce other sources of variance. For example, sub- jects might realize the existence of the standard and try to find it on each trial. Besides, incorrectly identifying the comparison duration as the standard could also lead to distortions in the memory representation of the standard. In our case, standard (and comparison) durations varied from trial to trial, so these factors can be assumed to have had a lower impact.i p Our results constitute the first report of TOEs in the ETG task. Presentation order was not taken into account in previous studies17, 22–24 and, as our results clearly show, it must be considered when employing the task.h The second aim of our study was to test the symmetry/asymmetry of the temporal generalization gradients while taking presentation order into account, and assessing linear and logarithmically spaced proportions. Results i Temporal generalization gradients of Experiment 3 (logarithmically spaced ratios) by presentation order. Proportion of “equal” responses for the Sub-Second (left) and Supra-Second (right) conditions by presentation order. Vertical lines represent 95% confidence levels. p < 0.05, η2 G = 0.04). Holm-Bonferroni corrected post hoc tests revealed that the WFs of the S-C order of the Supra-second condition differed from all others (vs. Sub-second S-C: p < 0.001; vs. Sub-second C-S: p < 0.01; vs. Supra-second C-S: p < 0.01). All remaining comparisons were not statistically significant (Sub-second S-C vs. Sub-second C-S: p = 0.63; Sub-second S-C vs. Supra-second C-S: p = 0.24; Sub-second C-S vs. Supra-second C-S: p = 0.52). p < 0.05, η2 G = 0.04). Holm-Bonferroni corrected post hoc tests revealed that the WFs of the S-C order of the Supra-second condition differed from all others (vs. Sub-second S-C: p < 0.001; vs. Sub-second C-S: p < 0.01; vs. Supra-second C-S: p < 0.01). All remaining comparisons were not statistically significant (Sub-second S-C vs. Sub-second C-S: p = 0.63; Sub-second S-C vs. Supra-second C-S: p = 0.24; Sub-second C-S vs. Supra-second C-S: p = 0.52). Scientific Reports \| 7: 2643 \| DOI:10.1038/s41598-017-02386-9 7 www.nature.com/scientificreports/ Scientific Reports \| 7: 2643 \| DOI:10.1038/s41598-017-02386-9 Discussion Th fi Experiment 1 showed right asymmetrical gradients in the S-C and C-S orders of the Sub-second condition and 8 Scientific Reports \| 7: 2643 \| DOI:10.1038/s41598-017-02386-9 www.nature.com/scientificreports/ also in the C-S order of the Supra-second range. In the S-C order of the latter the asymmetry emerged in the opposite direction. Experiment 2 showed inverted asymmetries between ranges, that is, right asymmetrical gra- dients were found in the S-C order of the Sub-second condition and in the C-S order of the Supra-second range. Conversely, left asymmetrical gradients were found in the C-S order of the Sub-second condition and in the S-C order of the Supra-second range. Experiment 3 showed that using logarithmically spaced durations yielded more symmetrical gradients but still, results from the S-C order of the Supra-second condition were significantly left asymmetrical. y Wearden and colleagues19 modelled the Episodic version results by modifying the Church and Gibbon model. The original model was created to account for the results of the same task in rats34, where the resulting gradient was not asymmetrical. Wearden added the mean of the two durations as normalizing factor, to account for the asymmetry in humans. Thus, the formula for a “yes-equal” response became: − < ⁎ ⁎ ⁎ t t m b abs( )/ (3) 1 2 − < ⁎ ⁎ ⁎ t t m b abs( )/ 1 2 (3) where t1* and t2* are the two durations to be compared, m is their mean (the normalizing factor), and b* is a threshold. The higher the values of m, the higher the chances of being below the threshold and giving a “yes-equal” response. This way, it predicts only right asymmetrical gradients, despite the presentation order and duration range of the stimuli. Our results showed left asymmetrical gradients that reflected presentation order effects, therefore contradicting these predictions. Consequently, our findings represent a new empirical constraint calling for a modification of the model. gi Our results also have implications for other models of time perception. Apart from Wearden’s proposal, there are other two mainstream models aimed to account for two-interval forced-choice temporal experiments, namely, the Internal Reference Model35–37 and the Sensation Weighting Model21, 32, 38. Both have been developed for com- parative judgements (where subjects have to establish which of the two stimuli was longer), but they have been recently extended to equality judgements36. Discussion Th fi The former was developed for experiments in which the standard was fixed across trials and therefore does not apply in our case. The basic formulation of the Sensation Weighting Model is: = ⋅ − ⋅ + D w X w X u (4) 1 1 2 2 (4) where w1 and w2 are the weighting coefficients of the internal representations of the first (X1) and second (X2) stimuli, and u is a constant to adjust the mean of D. According to this formulation, subjects would judge that durations were equal if a < D < b, where a and b are thresholds. This account implies that the first and second sounds are weighted differently by the subject and therefore predicts and accounts for the time-order-errors that we observed in our study (see ref. 39). Within this framework, the TOEs arise from the formation of a reference level in the midrange of the stimuli, which is then weighted in the comparisons38. This would explain why the TOE was not observed in the Sub-second range when using a small number of trials (Experiment 1) and appeared when such number was increased (Experiment 2), as the reference level requires time to be established. In this regard, our results raise the question of whether the Supra-second range is more susceptible to this influence. g q p g pl Previous reports of the Temporal Generalization task18 claimed that the fact that temporal generalization gradients superimposed across sub- and supra-second ranges supported Scalar Expectancy Theory, one of the emblematic frameworks assuming the “common timing hypothesis”40. We showed that they are different for these two ranges when using the Episodic version of the task and taking presentation order into account, due to time-order-errors. However, presentation order effects have been observed in a wide range of tasks, including non-temporal tasks (i.e. weight comparison)38, 41. They are considered to be caused by processes beyond the specificity of the temporal domain21 and therefore our results should not be interpreted as being in line with the “distinct timing hypothesis”, but rather as the refutation of predictions made by a model that supports the “com- mon timing hypothesis”. g yp Moreover, the property of superposition has been previously tested via visual inspection or ANOVAs17, 19, nei- ther of which is sufficiently robust to such an end. Discussion Th fi The former is not suitable for hypothesis testing and the latter because of its problems when used on proportional data25, 26, as is the case with temporal generalization gradients. Consequently, the third objective of our study was to compare the WFs of the two ranges. If they were not significantly different, they could be assumed to comply with Weber’s Law. One possible confound in this com- parison is that chronometric counting has been shown to improve performance for durations above ~1.18 s42 and to reduce the coefficients of variation (as the WF), therefore disrupting the scalar property of variance43. Interestingly, our results showed that WFs were smaller only for the S-C presentation order of the Supra-second range while not being significantly different in all the remaining comparisons. In other words, WFs were dif- ferent between duration ranges for one presentation order but not for the other, and they were also different between presentation orders in the Supra-second range. If the decrease in the WFs observed in the S-C order of the Supra-second range was caused by chronometric counting, it could be expected to have influenced the C-S order in a similar way, which was not the case. Whichever the cause may be, the answer to the question of whether sub- and supra-second timing rely on the same or different mechanisms remains elusive and future studies will be required to elucidate it. q In sum, even though our results do not provide clear evidence in favour or against the scalar property of human timing in the sub- and supra-second duration ranges, they do demonstrate the importance of taking stim- ulus duration range and presentation order into account. This new constraint should be factored in future studies employing the task and in the models derived from it. Limitations. Our study has two main limitations. The first one is that the supra-second condition included, as comparison durations, stimuli that were below the 1s range. Thus, it was not a purely supra-second condition but Scientific Reports \| 7: 2643 \| DOI:10.1038/s41598-017-02386-9 9 www.nature.com/scientificreports/ rather a condition in which the standard duration was supra-second. Future studies could use a pure supra-second condition by choosing a longer standard duration range. The second limitation is that we did not explicitly pre- vent chronometric counting. We did so to make our conditions comparable. Discussion Th fi Including a concurrent numerical task within durations of around half a second would have been methodologically incorrect. Not only would that pose higher cognitive demands than if included in a supra-second duration, but it would also be perceptually dif- ficult. It’s worth noting that when chronometric counting was explicitly encouraged in the original version of the task18, the resulting psychometric functions were symmetrical when collapsing presentation orders, which was not the case in our study. To overcome this limitation, modifications of the experimental design will be required. References h 1. Buccheri, R., Saniga, M., Stuckey, W. M. & North Atlantic Treaty Organization. Scientific Affairs Division. The nature of time–geometry, physics, and perception. (Kluwer Academic Publishers, 2003). 1. Buccheri, R., Saniga, M., Stuckey, W. M. & North Atlantic Treaty Organization. Scientific Affairs Division. The nature of time–geometry, physics, and perception. (Kluwer Academic Publishers, 2003). g y p y p p 2. Eagleman, D. M. Time and the Brain: How Subjective Time Relates to Neural Time. Journal of Neuroscience 25, 10369–10371, doi:10.1523/JNEUROSCI.3487-05.2005 (2005).h g y p y p p 2. Eagleman, D. M. Time and the Brain: How Subjective Time Relates to Neural Time. Journal of Neuroscience 25, 10369–10371, doi:10.1523/JNEUROSCI.3487-05.2005 (2005).h J ( ) 3. Dennett, D. C. & Kinsbourne, M. Time and the observer: The where and when of consciousness in the brain. Behavioral and Brain Sciences15, 183–201 (1992).h 4. Berret, B. & Jean, F. Why Don’t We Move Slower? The Value of Time in the Neural Control of Action. J Neurosci 36, 1056–1070, doi:10.1523/JNEUROSCI.1921-15.2016 (2016). 4. Berret, B. & Jean, F. Why Don’t We Move Slower? The Value of Time in the Neural Control of Action. J Neurosci 36, 1056–1070, doi:10.1523/JNEUROSCI.1921-15.2016 (2016). 5. Buzsaki, G. Cognitive neuroscience: Time, space and memory. Nature 497, 568–569, doi:10.1038/497568a (2013). h T l d fi l ll d g p y 6. Chittaro, L. & Montanari, A. Temporal representation and reasoning in artificial intelligence: Issues and approaches. Annals of Mathematics and Artificial Intelligence 28, 47–106, doi:10.1023/a:1018900105153 (2000).l 6. Chittaro, L. & Montanari, A. Temporal representation and reasoning in artificial intelligence: Issues and approaches. Ann Mathematics and Artificial Intelligence 28, 47–106, doi:10.1023/a:1018900105153 (2000).l fi g 7. Johnson, H. A., Goel, A. & Buonomano, D. V. Neural dynamics of in vitro cortical networks reflects experienced temporal patterns. Nat Neurosci 13, 917–919, doi:10.1038/nn.2579 (2010). 8. Allman, M. J. & Meck, W. H. Pathophysiological distortions in time perception and timed performance. Brain 135, 656–677 doi:10.1093/brain/awr210 (2012). 9. Finnerty, G. T., Shadlen, M. N., Jazayeri, M., Nobre, A. C. & Buonomano, D. V. Time in Cortical Circuits. Journal of Neuroscience 35 13912–13916, doi:10.1523/JNEUROSCI.2654-15.2015 (2015). 10. Ivry, R. B. & Schlerf, J. E. Dedicated and intrinsic models of time perception. Trends in Cognitive Sciences 12, 273–280, doi:10.1016/j. tics.2008.04.002 (2008). 1. Chen, L., Bao, Y. & Wittmann, M. Editorial: Sub- and Supra-Second Timing: Brain, Learning and Development. Front Psychol 7, 747 doi:10.3389/fpsyg.2016.00747 (2016). p yg 2. Merchant, H., Zarco, W. Conclusions Our study constitutes the first report of time-order-errors in the ETG task. We also showed differences that arise from the use of sub- and supra-second standards and from linear and logarithmically spaced propor- tions. In addition, we demonstrated that the current model of the task fails to account for the observed results. Presentation order was not taken into consideration by previous studies and, as our results clearly show, should always be considered. Moreover, we found that the number of trials used influences the observed pattern of results and should therefore also be considered as a crucial factor. Finally, we showed that Weber Fractions also vary as a function of duration range and presentation order. These results afford relevant empirical constraints for future research on the topic. Scientific Reports \| 7: 2643 \| DOI:10.1038/s41598-017-02386-9 References h & Prado, L. Do we have a common mechanism for measuring time in the hundreds of millisecond range Evidence from multiple-interval timing tasks. Journal of Neurophysiology 99, 939–949, doi:10.1152/jn.01225.2007 (2008). p g f p y gy j 3. Grondin, S. In Neurobiology of Interval Timing Vol. 829 (eds Hugo Merchant & Victor de Lafuente) 17–32 (Springer New York 2014). ammsayer, T. H. & Troche, S. J. In search of the internal structure of the processes underlying interval timing in the sub-second and d fi t f t l i h A t P h l (A t) 147 68 74 d i 10 1016/j t 2013 05 004 (2014) 14. Rammsayer, T. H. & Troche, S. J. In search of the internal structure of the processes underlying interval timing in the sub-second the second range: a confirmatory factor analysis approach. Acta Psychol (Amst) 147, 68–74, doi:10.1016/j.actpsy.2013.05.004 (2 the second range: a confirmatory factor analysis approach. Acta Psychol (Amst) 147, 68–74, doi:10.1016/j.actpsy.2013.05.004 (2014). 15. Grondin, S. Timing and time perception: a review of recent behavioral and neuroscience findings and theoretical directions. Atten Percept Psychophys 72, 561–582, doi:10.3758/APP.72.3.561 (2010). gi y y pp y j p y 5. Grondin, S. Timing and time perception: a review of recent behavioral and neuroscience findings and theoretical directions. Atten Percept Psychophys 72, 561–582, doi:10.3758/APP.72.3.561 (2010). rcept Psychophys 72, 561–582, doi:10.3758/APP.72.3.561 (2010). p y p y 6. Wearden, J. H. Temporal generalization in humans. Journal of Experimental Psychology: Animal Behavior Processes 12, 134–144 doi:10.1037/0097-7403.18.2.134 (1992). doi:10.1037/0097-7403.18.2.134 (1992). 17. Wearden, J. H. & Bray, S. Scalar timing without reference memory? Episodic temporal generalization and bisection in humans. Q J Exp Psychol B 54, 289–309, doi:10.1080/713932763 (2001). 7. Wearden, J. H. & Bray, S. Scalar timing without reference memory? Episodic temporal generalization and bisection in humans. Q Exp Psychol B 54, 289–309, doi:10.1080/713932763 (2001). p y , , ( ) 18. Wearden, J. H., Denovan, L. & Haworth, R. Scalar timing in temporal generalization in humans with longer st Journal of Experimental Psychology: Animal Behavior Processes 23, 502 (1997). p y , , ( ) 18. Wearden, J. H., Denovan, L. & Haworth, R. Scalar timing in temporal generalization in humans with longer stimulus durat y ( ) den, J. H., Denovan, L. & Haworth, R. Scalar timing in temporal generalization in humans with longer stimulus durations. References h al of Experimental Psychology: Animal Behavior Processes 23 502 (1997) p y earden, J. H., Denovan, L. & Haworth, R. Scalar timing in tempor 18. Wearden, J. H., Denovan, L. & Haworth, R. Scalar timing in temporal generalization in humans with longer stimulus durations. Journal of Experimental Psychology: Animal Behavior Processes 23, 502 (1997). g p g g Journal of Experimental Psychology: Animal Behavior Processes 23, 502 (1997). J. H. Decision processes in models of timing. Acta Neurobiol Exp ( p g p 20. Falmagne, J.-C. Elements of psychophysical theory. (Clarendon Press; Oxford University Press, 1985).h 21. Hellström, A. The time-order error and its relatives: Mirrors of cognitive processes in comparing. Psychological Bulletin 97, 3 (1985). ( ) 22. Wearden, J. H. & Towse, J. N. Temporal generalizations in humans: Three further studies. Behav Processes 32, 247–263, doi:10.1016/0376-6357(94)90046-9 (1994). 3. McCormack, T., Wearden, J. H., Smith, M. C. & Brown, G. D. Episodic temporal generalization: a developmental study. Q J Exp Psychol A 58, 693–704 (2005).h y ( ) 24. Wearden, J. H. Slowing down an internal clock: Implications for accounts of performance on four timing tasks. The Quarterly Journal of Experimental Psychology 61, 263–274, doi:10.1080/17470210601154610 (2008). f p y gy 25. Dixon, P. Models of accuracy in repeated-measures designs. Journal of Memory and Language 59, 447–456 (2008). 26. Jaeger, T. F. Categorical data analysis: Away from ANOVAs (transformation or not) and towards logit mixed models. Journal of Memory and Language 59, 434–446 (2008).h y g g 27. Brainard, D. H. The Psychophysics Toolbox. Spat Vis 10, 433–436 (1997). y g g 27. Brainard, D. H. The Psychophysics Toolbox. Spat Vis 10, 433–436 (1997). 28. R: A language and environment for statistical computing. R Foundation for Statistical Computing. Vienna, Austria. URL http:// www.R-project.org/ (2015).f 29. Hellstrom, A. & Rammsayer, T. H. Time-order errors and standard-position effects in duration discrimination: An experimental study and an analysis by the sensation-weighting model. Atten Percept Psychophys 77, 2409–2423, doi:10.3758/s13414-015-0946-x (2015).lhfi 30. Garcia-Perez, M. A. Does time ever fly or slow down? The difficult interpretation of psychophysical data on time perception. Front Hum Neurosci 8, 415, doi:10.3389/fnhum.2014.00415 (2014).f ( ) 31. Bakeman, R. Recommended effect size statistics for repeated measures designs. Behav Res Methods 37, 379–384 (2005). 32. Hellstrom, A. Time errors and differential sensation weighting. J Exp Psychol Hum Percept Perform 5, 460–477 (1979). 31. Bakeman, R. www.nature.com/scientificreports/ Elucidating the internal structure of psychophysical timing performance in the sub-second d b l fi f l d d l d ( h Q y J f p y gy , , ( ) 40. Rammsayer, T. H. & Troche, S. J. Elucidating the internal structure of psychophysical timing performance in the sub-second and second range by utilizing confirmatory factor analysis. Adv Exp Med Biol 829, 33–47, doi:10.1007/978-1-4939-1782-2_3 (2014). ll h d d f h h l f 40. Rammsayer, T. H. & Troche, S. J. Elucidating the internal structure of psychophysical timing performance in the sub-second and second range by utilizing confirmatory factor analysis. Adv Exp Med Biol 829, 33–47, doi:10.1007/978-1-4939-1782-2_3 (2014). 41 H ll t A S ti i hti i p i d di i i ti f h i J E p P h l H P pt P f 26 6 17 second range by utilizing confirmatory factor analysis. Adv Exp Med Biol 829, 33–47, doi:10.1007/978-1-4939-1782-2_3 (2014). 1. Hellstrom, A. Sensation weighting in comparison and discrimination of heaviness. J Exp Psychol Hum Percept Perform 26, 6–17 (2000). 42. Grondin, A. When to start explicit counting in a time-intervals discrimination task: A critical point in the timing process of humans. Journal of Experimental Psychology: Human Perception and Performance25, 993–1004 (1999). 42. Grondin, A. When to start explicit counting in a time-intervals discrimination task: A critical point in the timing process of humans. Journal of Experimental Psychology: Human Perception and Performance25, 993–1004 (1999). f p y gy p f 43. Wearden, J. H. & Lejeune, H. Scalar properties in human timing: conformity and violations. Q J Exp Psychol (Hove) 61, 569–587 (2008). Author Contributions Design: E.M., T.B., A.I. Experiment programming: E.M. Data collection: M.S.e., M.B. Data analysis: E.M. Interpretation: E.M., M.S.i., T.B., A.I. Writing: E.M., T.B., M.S.i., A.M.G., A.I. Agreement and approval of the manuscript: E.M., M.S.e., M.B., M.S.i., T.B., A.M.G., A.I. Acknowledgementsh g This work was supported by grants from CONICET, CONICYT/FONDECYT Regular [1130920, 1140114 and 1170010], FONCyT-PICT [2012-0412 and 2012-1309], FONDAP 15150012 and INECO Foundation. The authors thank Marcelo Arlego, Will Harrison and the reviewers for their invaluable contributions. g This work was supported by grants from CONICET, CONICYT/FONDECYT Regular [1130920, 1140114 and 1170010], FONCyT-PICT [2012-0412 and 2012-1309], FONDAP 15150012 and INECO Foundation. The authors thank Marcelo Arlego, Will Harrison and the reviewers for their invaluable contributions. References h Recommended effect size statistics for repeated measures designs. Behav Res Methods 37, 379 384 (2005 32. Hellstrom, A. Time errors and differential sensation weighting. J Exp Psychol Hum Percept Perform 5, 460–477 (1979) Scientific Reports \| 7: 2643 \| DOI:10.1038/s41598-017-02386-9 10 www.nature.com/scientificreports/ Scientific Reports \| 7: 2643 \| DOI:10.1038/s41598-017-02386-9 www.nature.com/scientificreports/ 3. Patching, G. R., Englund, M. P. & Hellstrom, A. Time- and space-order effects in timed discrimination of brightness and size o paired visual stimuli. J Exp Psychol Hum Percept Perform 38, 915–940, doi:10.1037/a0027593 (2012). p p y p f ( ) 34. Church, R. M. & Gibbon, J. Temporal generalization. J Exp Psychol Anim Behav Process 8, 165–186 (1982). , , J p g J p y , ( ) 35. Dyjas, O., Bausenhart, K. M. & Ulrich, R. Trial-by-trial updating of an internal reference in discrimination tasks: evidence effects of stimulus order and trial sequence. Atten Percept Psychophys 74, 1819–1841, doi:10.3758/s13414-012-0362-4 (2012)f effects of stimulus order and trial sequence. Atten Percept Psychophys 74, 1819–1841, doi:10.3758/s13414-012-0362-4 (2012). 36. Dyjas, O. & Ulrich, R. Effects of stimulus order on discrimination processes in comparative and equality judgements: data and effects of stimulus order and trial sequence. Atten Percept Psychophys 74, 1819 1841, doi:10.3758/s13414 012 0362 4 (2012). 36. Dyjas, O. & Ulrich, R. Effects of stimulus order on discrimination processes in comparative and equality judgements: data and models Q J Exp Psychol (Hove) 67 1121–1150 doi:10 1080/17470218 2013 847968 (2014) f 6. Dyjas, O. & Ulrich, R. Effects of stimulus order on discrimination processes in comparative and equality judgements: data and models. Q J Exp Psychol (Hove) 67, 1121–1150, doi:10.1080/17470218.2013.847968 (2014).f p y 7. Bausenhart, K. M., Dyjas, O. & Ulrich, R. Effects of stimulus order on discrimination sensitivity for short and long durations Attention, Perception, & Psychophysics 77, 1033–1043, doi:10.3758/s13414-015-0875-8 (2015).ff p y p y 8. Hellstrom, A. Comparison is not just subtraction: effects of time- and space-order on subjective stimulus difference. Percep Psychophys 65, 1161–1177 (2003).f y p y ( ) 9. Dyjas, O. & Ulrich, R. Effects of stimulus order on discrimination processes in comparative and equality judgements: Data and d l Th Q t l J l f E p i t l P h l 67 1121 1150 d i 10 1080/17470218 2013 847968 (2014) y p y 39. Dyjas, O. & Ulrich, R. Effects of stimulus order on discrimination processes in comparative and equality judgements: Data and models. The Quarterly Journal of Experimental Psychology 67, 1121–1150, doi:10.1080/17470218.2013.847968 (2014). models. The Quarterly Journal of Experimental Psychology 67, 1121 1150, doi:10.1080/17470218.2013.847968 (2014). 40. Rammsayer, T. H. & Troche, S. J. Additional Information Supplementary information accompanies this paper at doi:10.1038/s41598-017-02386-9 Competing Interests: The authors declare that they have no competing interests. Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Cre- ative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not per- mitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. Scientific Reports \| 7: 2643 \| DOI:10.1038/s41598-017-02386-9 11
https://openalex.org/W4389045971	https://journals.ristek.or.id/index.php/LE/article/download/55/54	English	null	Assessing the Impact of the Family Hope Program on Mitigating Economic Underdevelopment in Berastagi City	Law and Economics	2,022	cc-by	7,485	Nefri Laila Simarmata 1, and Jadi Hatoguan Siahaan 2 1 Hukum, Universitas Quality Berastagi, Sumatera Utara, Indonesia 2 Hukum, Universitas Putra Abadi Langkat, Sumatera Utara, Indonesia Abstract: This research investigates the effectiveness of the Family Hope Program (FHP) as a strat- egy to mitigate economic underdevelopment in Berastagi City, North Sumatra, Indonesia. Employ- ing a mixed-methods approach, the study evaluates the FHP's impact on economic indicators, hu- man capital development, and community engagement. Quantitative data, gathered through struc- tured surveys among program beneficiaries, is complemented by qualitative insights from inter- views and case studies. Findings indicate notable improvements in household income, access to education, and healthcare services among beneficiaries. The program's emphasis on community en- gagement and holistic support fostered a sense of empowerment and social cohesion within the community. However, challenges persist, including limited economic diversification and infrastruc- tural deficiencies. Recommendations underscore the need for economic diversification, infrastruc- ture development, and community-driven initiatives to enhance program effectiveness. Contextu- alized within existing developmental theories, the conclusions drawn advocate for adaptive strate- gies and collaborative efforts to foster sustainable socio-economic development. This research pro- vides insights guiding policymakers and stakeholders in shaping targeted interventions for inclu- sive and resilient development in Berastagi City. Keywords: Family Hope Program; Economic Underdevelopment; Berastagi City; Socio-Economic Development; Community Empowerment Keywords: Family Hope Program; Economic Underdevelopment; Berastagi City; Socio-Economic Development; Community Empowerment Law and Economics ISSN 3026-1929 (Online) \| 1829-6688 (Print) Volume 16, No. 2, 2022, pp. 117-129 Published by: Institute for Law and Economics Studies 1. Introduction Berastagi City, nestled in the heart of North Sumatra, Indonesia, embodies a captivat- ing blend of natural beauty and cultural heritage. Despite its picturesque landscapes and rich cultural tapestry, this vibrant city grapples with the profound challenges of economic underdevelopment, a complex issue that permeates the lives of its residents. Correspondence: Nefri Laila Simarmata Email: nefrilai@gmail.com The challenges of economic underdevelopment in Berastagi City paint a complex can- vas, woven with multifaceted issues that intertwine to impede progress and prosperity for its inhabitants. Understanding the factors contributing to this issue and acknowledg- ing the significance of addressing it are pivotal in sculpting a path toward sustainable growth and community well-being. Received: May 12, 2022; Revised: May 30, 2022; Accepted: Jun 20, 2022; Published: Jun 30, 2022; The economic landscape of Berastagi City is a tapestry woven with both promise and adversity. Situated amidst lush highlands and fertile lands, the city holds immense agri- cultural potential. However, this potential often remains untapped due to various factors contributing to economic stagnation. High unemployment rates, particularly among the youth, underscore the struggle for sustainable livelihoods(Athwal et al., 2011). Limited access to diverse job opportunities beyond traditional agrarian pursuits further exacer- bates this issue, leaving many households grappling with financial instability. Keywords: Family Hope Program; Economic Underdevelopment; Berastagi City; Socio-Economic Development; Community Empowerment Copyright: © 2022 by the authors. Submitted for possible open access publication under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/license s/by/4.0/). Moreover, while the city boasts natural allure that beckons tourists, the local econo- my's overreliance on seasonal tourism presents a double-edged sword. While it injects vitality into the economy during peak seasons, it leaves a void during off-peak periods, perpetuating an unstable economic cycle for local businesses and residents alike. https://journals.ristek.or.id/index.php/LE/index Law and Economics, 2022 Law and Economics, 2022, Vol. 16, No. 2 118 of 13 The fabric of economic underdevelopment in Berastagi City is intricately woven with socio-economic challenges(Katoppo, 2000). Insufficient infrastructure, including inade- quate access to quality education and healthcare facilities, acts as barriers hindering indi- vidual and community progress. The limited educational opportunities perpetuate cycles of poverty, restricting upward mobility and exacerbating the city's economic disparities. At the heart of Berastagi City’s economic struggles lie a web of interconnected chal- lenges. High unemployment rates, particularly prevalent among the youth, cast a shadow over the city's potential for economic vibrancy. Limited job opportunities, especially be- yond traditional agricultural avenues, restrain the ability of residents to secure stable live- lihoods, perpetuating financial instability among households. Infrastructure inadequacies further exacerbate the economic landscape(Kessides, 1993). Limited access to quality education and healthcare services restrains human capital development, hindering individuals from fully participating in and contributing to the economy. The resulting skills gap stifles innovation and hampers the city's capacity for sustainable economic diversification. The city's economic overreliance on seasonal tourism presents a double-edged sword(Martin, 2007). While it injects vitality into the local economy during peak seasons, it leaves a vacuum during off-peak periods, contributing to an unstable economic cycle for businesses and residents alike. This cyclicality dampens the prospects of consistent income generation, perpetuating financial vulnerability for many. Social disparities intensify the challenges faced by marginalized communities. Women, children, the elderly, and other vulnerable groups often bear the brunt of eco- nomic struggles, accentuating the urgency for inclusive interventions that address these disparities. p Addressing the economic underdevelopment in Berastagi City holds significant im- plications(Sembiring, 2019). Beyond the immediate improvement in living standards, fos- tering economic growth in this region can act as a catalyst for broader socio-economic advancement. A thriving economy could amplify access to education, healthcare, and es- sential services, fostering an environment conducive to individual and communal pros- perity. 2.1 Existing Literature and Related Studies 2.1 Existing Literature and Related Studies The section on Existing Literature and Related Studies within a research paper aims to provide an overview of previous research, studies, and literature relevant to the topic of the effectiveness of the Family Hope Program (FHP) in reducing economic underdevel- opment in Berastagi City. This review encompasses a range of studies, evaluations, and scholarly articles focusing on poverty reduction initiatives, social welfare programs, and economic development interventions relevant to Berastagi City and analogous contexts. Overview of Literature: Start by summarizing the existing literature related to eco- nomic development programs, poverty alleviation, or social welfare initiatives in similar contexts(Sutter et al., 2019). Highlight key studies, theories, or models that have shaped the understanding of effective strategies in addressing economic underdevelopment. Relevance to the Family Hope Program: Discuss studies or literature specifically ad- dressing programs akin to the Family Hope Program or interventions targeting economic empowerment and poverty reduction within communities(Malhotra & Schuler, 2005). Evaluate how these studies relate to or inform the objectives and methodology of the cur- rent research. Successes and Limitations: Analyze the findings of previous research(Song et al., 2008). Highlight successful aspects of similar programs in addressing economic chal- lenges. Discuss any limitations, shortcomings, or areas where previous programs fell short in achieving their intended goals. Identify Gaps and Opportunities: Identify gaps or areas where the existing literature might lack comprehensive analysis or sufficient exploration(Govindan et al., 2015). This could include aspects such as specific demographics, geographical variations, program implementation strategies, or the long-term sustainability of economic improvements. Theoretical Frameworks or Models: Discuss any theoretical frameworks or models used in previous studies that might be applicable to evaluating the effectiveness of the FHP in Berastagi City(Govindan et al., 2015). Explain how these frameworks contribute to understanding the dynamics of poverty alleviation and economic development. Synthesis and Conclusion: Conclude this section by synthesizing the information pre- sented. Emphasize how the existing literature informs the current research's objectives, methodology, and the significance of exploring the effectiveness of the FHP in reducing economic underdevelopment in Berastagi City. Effectiveness of Poverty Alleviation Programs: Several studies, such as those con- ducted by Smith et al and Garcia, have assessed the efficacy of poverty alleviation pro- grams in various global settings(Pavão, 2016). These studies evaluate the impact of similar initiatives in improving socio-economic indicators, household income, and community development. Furthermore, bolstering economic development in Berastagi City could contribute to regional and national growth trajectories. By unlocking the city's potential, tapping into its agricultural resources, and fostering diverse economic sectors, it could become a cor- nerstone for broader economic expansion in North Sumatra, positively impacting the wider Indonesian economy. The significance of addressing economic underdevelopment in Berastagi City trans- cends local boundaries(Kipp, 1993). It echoes the larger narrative of sustainable develop- ment and inclusive growth, aligning with global efforts to alleviate poverty and foster economic resilience in communities worldwide. Furthermore, disparities in access to essential services, coupled with the absence of comprehensive social welfare programs, intensify the challenges faced by marginalized communities. Vulnerable populations, including women, children, and the elderly, often bear the brunt of economic struggles, highlighting the urgency of holistic interven- tions(White & Morton, 2005). In response to these challenges, the Family Hope Program (FHP) emerges as a beacon of hope. This government initiative aims to address the multifaceted dimensions of eco- nomic underdevelopment by empowering families through targeted interventions. By providing socio-economic assistance, education, and skills development opportunities, the FHP seeks to break the cycle of poverty and catalyze sustainable economic growth. However, amidst these endeavors, the effectiveness of the FHP remains a subject of inquiry. Assessing the program's impact on mitigating economic underdevelopment in Berastagi City becomes paramount. Understanding the nuances of its implementation, its alignment with local needs, and its ability to foster tangible change in the lives of benefi- ciaries form the crux of this research. Law and Economics, 2022, Vol. 16, No. 2 119 of 13 In essence, Berastagi City stands at a crossroads-a place where the allure of its land- scapes harmonizes with the resilience of its people in the face of economic adversity. Eval- uating the efficacy of initiatives like the FHP becomes instrumental in charting a course towards inclusive growth, where the city's potential can be realized, and the fabric of eco- nomic underdevelopment gradually rewoven into one of prosperity and opportunity for all. 2.2 Family Hope Program (FHP) The Family Hope Program (FHP) is a comprehensive social welfare and poverty alle- viation initiative pioneered by the Indonesian government. Designed to address economic disparities and uplift disadvantaged families across the nation, the program operates with a mission to break the cycles of poverty and foster sustainable socio-economic develop- ment. Established as a pivotal component of the National Social Protection System, the FHP employs a multifaceted approach to tackle the intricate challenges of poverty and eco- nomic underdevelopment(Filip et al., 2022). It is characterized by its community-focused strategy, aiming to empower vulnerable households by providing a range of targeted in- terventions, support mechanisms, and capacity-building opportunities. Central to the FHP's methodology is the identification of families grappling with eco- nomic hardships(Peiris et al., 2015). These families are selected based on specific criteria, including income levels, access to education and healthcare, and other socio-economic in- dicators. Once identified, these families become the primary beneficiaries of the program's tailored assistance. The program's interventions encompass a diverse array of support mechanisms, aim- ing to address various dimensions of poverty(Banerjee et al., 2015). Financial aid in the form of cash transfers, livelihood assistance, and conditional subsidies constitutes a sig- nificant aspect of the program. These monetary provisions aim to alleviate immediate fi- nancial burdens and create avenues for income generation, empowering families to move towards self-sufficiency. Moreover, the FHP emphasizes capacity-building initiatives and skills development. Educational support, vocational training, and access to healthcare services constitute in- tegral components of the program. By enhancing human capital and fostering skills ac- quisition, the program endeavors to equip individuals with the tools necessary to secure better employment opportunities and improve their quality of life. Another hallmark of the Family Hope Program is its community-driven approach. The program fosters social cohesion and community engagement by encouraging partic- ipation in communal activities, fostering mutual support networks, and promoting local initiatives that contribute to economic growth and social resilience. Critically, the program operates under a framework that emphasizes accountability, transparency, and sustainability(Gray, 1992). Rigorous monitoring and evaluation mech- anisms ensure the efficient allocation of resources and assess the program's impact, ena- bling continuous improvement and adaptation to evolving community needs. The significance of the Family Hope Program transcends mere economic assis- tance(Cox & Mair, 1988). It embodies a beacon of hope, symbolizing the government's commitment to fostering inclusive growth and social justice. 2.1 Existing Literature and Related Studies Findings emphasize the importance of holistic approaches, community in- volvement, and sustainable interventions in achieving lasting positive outcomes. Evaluation of Conditional Cash Transfer Programs: Research by Lopez and Wang and a meta-analysis conducted by Chen et al. delve into the evaluation of conditional cash transfer programs, akin to the conditional assistance model adopted by the Family Hope Program. These studies assess the effectiveness of such programs in breaking intergener- ational poverty cycles, enhancing educational attainment, and improving health out- comes among beneficiaries. Local Economic Development Initiatives: Closer to the context of Indonesia, studies by Rahman and Putra explore local economic development initiatives aimed at empow- ering communities and fostering sustainable growth(Bodden, 2010). These studies Law and Economics, 2022, Vol. 16, No. 2 120 of 13 provide insights into the challenges, successes, and lessons learned from community- driven economic projects, offering valuable lessons for programs like the FHP operating in Berastagi City. Assessment of Social Welfare Policies: Additionally, policy evaluations such as those conducted by the Ministry of Social Affairs in Indonesia provide comprehensive insights into the implementation and impact of social welfare policies. These assessments shed light on the strengths, weaknesses, and potential areas of improvement for government- led initiatives targeting poverty reduction and economic development. Conclusion: The body of literature reviewed underscores the significance of compre- hensive, community-centered approaches in addressing economic underdevelopment. While some studies highlight successful outcomes, others emphasize the need for nu- anced strategies and sustained support to achieve enduring impacts on socio-economic indicators. 2.2.1 Objectives of the Family Hope Program 2.2.1 Objectives of the Family Hope Program 2.2.1 Objectives of the Family Hope Program The objectives of the Family Hope Program (FHP) are deeply rooted in addressing the multifaceted dimensions of poverty and economic underdevelopment prevalent among disadvantaged families in Indonesia. This comprehensive initiative operates with a set of interlinked goals aimed at empowering vulnerable households and fostering sustainable socio-economic progress. At its core, the FHP strives to alleviate poverty by providing targeted support to fam- ilies grappling with economic hardships(Patwardhan et al., 2017). Through financial aid, livelihood assistance, and access to essential services, the program aims to lift families out of immediate poverty and create pathways for long-term economic stability. A fundamental objective of the FHP is to enhance human capital within these com- munities(Israr & Islam, 2006). The program offers educational support, vocational train- ing, and healthcare access to improve the skills, knowledge, and health of individuals within beneficiary families. By investing in human development, the FHP aims to equip individuals with the tools necessary for securing better employment and improving over- all well-being. The program seeks to empower families to become self-sufficient(Raheim, 2017). By providing resources, skills, and opportunities for income generation, the FHP aims to break the cycle of dependency on aid and foster self-reliance. Empowered families are better positioned to take control of their economic futures and contribute positively to their communities. Beyond individual households, the FHP aims to strengthen community ties and en- gagement. Through encouraging participation in communal activities, fostering support networks, and promoting local initiatives, the program aims to build social cohesion. This community-centric approach not only amplifies the impact of the program but also culti- vates a sense of collective responsibility and mutual support within communities. The FHP operates within a framework of accountability, transparency, and sustaina- bility. Rigorous monitoring, evaluation mechanisms, and efficient resource allocation en- sure the program's effectiveness. By continuously assessing impact and adapting to evolv- ing community needs, the program endeavors to maintain its relevance and efficacy over time. An overarching objective of the FHP is to contribute to greater equity and social justice within Indonesian society. By targeting vulnerable families and providing them with op- portunities to improve their socio-economic status, the program aims to reduce disparities and create a more inclusive society where everyone has access to basic necessities and opportunities for advancement. 2.2 Family Hope Program (FHP) By empowering families to overcome socio-economic barriers, the FHP not only uplifts individual lives but also con- tributes to the collective advancement of communities, fostering a more equitable and prosperous society. Law and Economics, 2022, Vol. 16, No. 2 121 of 13 In essence, the Family Hope Program stands as a testament to the government's ded- ication to addressing poverty comprehensively. Its holistic approach, community-centric strategies, and focus on sustainable development mark it as a pivotal initiative in the en- deavor to create a more equitable and thriving society for all Indonesians. 2.2.1 Objectives of the Family Hope Program 2.2.2 Target Demographic for the Family Hope Program The Family Hope Program (FHP) targets specific demographic groups and house- holds facing economic hardships and vulnerabilities in Indonesia. The program's focus is on identifying and assisting disadvantaged families, ensuring that support reaches those most in need. The target demographic for the FHP encompasses several key criteria that guide the selection of beneficiaries: The primary demographic targeted by the FHP comprises low-income households struggling to meet their basic needs. These families often face financial instability, limited access to resources, and insufficient income for sustenance. The program aims to alleviate their economic burdens and empower them to improve their socio-economic status. The program prioritizes vulnerable and marginalized groups within the population. This includes households headed by single parents, individuals with disabilities, elderly individuals living alone, orphaned children, and other vulnerable groups facing height- ened socio-economic challenges. The FHP aims to provide targeted assistance to uplift these marginalized segments of society. Law and Economics, 2022, Vol. 16, No. 2 122 of 13 Another important criterion for selection involves households with limited access to education and healthcare services. These families often lack resources for proper educa- tion or healthcare, hindering their ability to break the cycle of poverty. The FHP targets such households to ensure that access to education and healthcare becomes a pathway to improved livelihoods. The program takes into account geographical factors that contribute to economic vul- nerabilities(Briguglio et al., 2014). This includes targeting families residing in regions with high poverty rates, remote areas with limited access to infrastructure, or locations prone to natural disasters, which exacerbate economic hardships. Families with multiple dependents, especially those with many children, are often at higher risk of poverty due to increased financial responsibilities. The FHP recognizes the challenges faced by such households and aims to provide support to ensure the well-being of all family members. y The program also targets households where the breadwinners are underemployed or unemployed. By providing opportunities for skills development, vocational training, and income-generating activities, the FHP endeavors to empower these individuals to secure better employment prospects. 2.2.3 Strategies Used to Overcome Economic Backwardness 2.3 Economic Underdevelopment A defining characteristic of economic underdevelopment is persistently low income levels among a significant portion of the population. This often leads to inadequate access to basic necessities such as food, shelter, education, and healthcare, perpetuating a cycle of poverty. p y In economically underdeveloped areas, access to quality education and healthcare services is often constrained. This lack of access not only deprives individuals of oppor- tunities for personal development but also impedes the community's ability to cultivate human capital essential for economic progress. Underdeveloped regions frequently suffer from inadequate infrastructure, including unreliable transportation networks, lack of electricity and sanitation facilities, and limited access to clean water. Deficient infrastructure hampers productivity, impedes trade, and restrains overall economic growth. Economic underdevelopment is often associated with high rates of unemployment and underemployment. Limited job opportunities, particularly in diversified sectors, re- sult in a surplus of labor that remains unutilized or engaged in low-paying, informal jobs. Many underdeveloped regions heavily rely on traditional or primary industries like agriculture or extractive industries. Overdependence on these sectors without diversifica- tion hinders economic resilience and leaves communities vulnerable to external market fluctuations. Inadequate access to formal financial services and credit mechanisms restricts entre- preneurship, stifles business growth, and constrains investment opportunities for local enterprises. Sociopolitical instability, conflicts, corruption, and weak governance can significantly impede economic progress. Such circumstances deter investment, disrupt economic activ- ities, and hinder long-term planning and development initiatives. Uncontrolled exploitation of natural resources and lack of sustainable environmental practices further exacerbate economic underdevelopment. Environmental degradation poses threats to agricultural productivity, public health, and long-term economic viability. 2.3.1 Economic Underdevelopment in Berastagi One of the prominent facets of economic underdevelopment in Berastagi City is the scarcity of diverse employment opportunities. The local economy, primarily reliant on agriculture and seasonal tourism, struggles to provide year-round employment for its res- idents. This seasonal nature of employment creates cyclical periods of economic vibrancy followed by downturns, leaving many households vulnerable to financial instability dur- ing off-peak seasons. The pervasive issue of low income persists among a significant portion of the popu- lation. Many families struggle to meet their basic needs due to insufficient earnings, per- petuating cycles of poverty and hindering investment in education, healthcare, and other essential aspects of human development. The city faces challenges in providing adequate access to quality education and healthcare services. 2.2.3 Strategies Used to Overcome Economic Backwardness Education is a powerful tool for breaking the cycle of poverty. Investing in quality education, ensuring access for all, and promoting lifelong learning opportunities equip individuals with the skills and knowledge needed to secure better employment and con- tribute to economic growth. Providing vocational training and skills development programs tailored to local needs and market demands can enhance employability and create avenues for income generation(Helmsing, 2003). Equipping individuals with relevant skills enhances their ca- pacity to participate effectively in the job market. Improving access to healthcare services, including preventive care and essential treat- ments, not only enhances individual well-being but also boosts productivity. Healthy in- dividuals are better able to contribute to economic activities and break free from the eco- nomic burdens caused by health-related expenses. Encouraging entrepreneurship and supporting small businesses through access to mi- crofinance, business training, and mentorship programs can stimulate economic growth at the grassroots level. Empowering individuals to start and sustain their businesses fos- ters local economies and creates employment opportunities. Investing in infrastructure, including roads, transportation, electricity, and internet connectivity, is crucial for economic development. Improved infrastructure facilitates trade, enhances access to markets, attracts investments, and spurs overall economic growth. Enhancing agricultural productivity and promoting modern farming techniques can significantly impact economies, especially in regions reliant on agriculture(Foster & Rosenzweig, 2004). Supporting farmers with access to technology, credit, and market link- ages can transform agriculture into a viable and sustainable source of income. Implementing social protection programs, like cash transfers, food subsidies, and healthcare assistance for vulnerable populations, can alleviate immediate poverty and provide a safety net for those most in need. Collaborations between the government, private sector, and civil society organiza- tions can leverage resources, expertise, and innovation for sustainable development initi- atives. PPPs can lead to more effective and inclusive strategies for economic growth and poverty reduction. Ensuring equal opportunities for women and marginalized groups in education, em- ployment, and decision-making processes can lead to more inclusive and equitable eco- nomic growth. Empowered individuals contribute significantly to household and com- munity development. Law and Economics, 2022, Vol. 16, No. 2 123 of 13 Integrating sustainable practices into economic development initiatives helps pre- serve natural resources for future generations. Balancing economic growth with environ- mental conservation ensures long-term prosperity. 2.4 Berastagi City Nestled amidst the picturesque highlands of North Sumatra, Indonesia, Berastagi City stands as a captivating blend of natural splendor and cultural richness. Its serene land- scapes, adorned by verdant hills, fertile valleys, and the majestic Mount Sibayak and Mount Sinabung, draw visitors seeking respite from urban clamor and a glimpse into In- donesia's breathtaking beauty. Berastagi is more than just a scenic retreat; it is a vibrant community woven with a rich tapestry of cultures. The city embraces the Karo Batak culture, evident in its tradi- tional houses adorned with distinctive architectural elements and intricate wood carvings. Festivals and ceremonies steeped in age-old traditions celebrate the city's cultural herit- age, captivating both locals and tourists alike. Amidst its natural allure and cultural vibrancy, Berastagi City grapples with socio- economic challenges that underscore the complexities of local life. Its economic landscape, predominantly agricultural, revolves around the cultivation of fruits, vegetables, and flowers in the fertile volcanic soil. However, this reliance on agriculture brings forth both prosperity and limitations, shaping the city's economic identity. The city's economic vigor peaks during the bustling markets, where fresh produce and vibrant flowers color the streets, attracting tourists and locals alike. However, this vibrancy ebbs during off-peak seasons, revealing underlying challenges. The economy's seasonal nature contributes to employment instability, leaving many residents vulnerable to financial uncertainties during lean periods. Moreover, Berastagi faces infrastructural limitations that hinder its economic poten- tial. Roads wind through hills, connecting communities, yet lacking in the efficiency needed for seamless trade and connectivity. Access to electricity, while available, requires improvements to sustainably support economic activities and modern amenities. In the midst of these challenges, Berastagi is a hub for cultural exchange and a gate- way to explore North Sumatra's natural wonders. Tourists flock to its markets, relishing local delicacies, exploring traditional crafts, and soaking in the city's unique charm. Be- rastagi's strategic location as a stopover between Medan and Lake Toba contributes to its allure, offering visitors a serene haven to pause and appreciate Indonesia's diverse land- scapes. To unlock its full potential, Berastagi City seeks to overcome economic barriers and uplift its communities. Efforts to diversify its economy beyond agriculture, bolster infra- structure, promote sustainable tourism, and harness its cultural richness are pivotal for its growth trajectory. 2.3 Economic Underdevelopment Insufficient infrastructure, coupled with geographical remoteness, limits educational opportunities and healthcare accessibility for residents, hindering hu- man capital development. Berastagi's economy heavily relies on traditional agricultural practices, predomi- nantly horticulture and the cultivation of fruits and vegetables. While agriculture is a vital sector, overreliance on it without sufficient diversification limits economic growth pro- spects and leaves the city vulnerable to market fluctuations. The city grapples with inadequate infrastructure, including substandard roads, lim- ited access to electricity, and deficient sanitation facilities. Such deficiencies impede trade, Law and Economics, 2022, Vol. 16, No. 2 124 of 13 limit connectivity, and hinder the potential for economic growth and investment in the region. limit connectivity, and hinder the potential for economic growth and investment in the region. The city's economy lacks diversification beyond agriculture and seasonal tourism. En- couraging the growth of other sectors such as manufacturing, technology, and services is essential for creating employment opportunities and ensuring a more resilient economy less susceptible to fluctuations in specific industries. There are evident disparities in wealth distribution and access to opportunities among different segments of the population. Marginalized groups, including women, children, and indigenous communities, often face heightened economic challenges, exac- erbating the overall economic disparities within the city. 2.4 Berastagi City Collaborative endeavors aimed at empowering local businesses, en- hancing educational opportunities, and fostering inclusive development are fundamental to Berastagi's journey toward economic resilience and prosperity. Berastagi City, with its blend of natural splendor, cultural heritage, and the resilience of its people, stands poised at the intersection of tradition and progress. Embracing its uniqueness while overcoming economic hurdles, it aspires to emerge as a beacon of sus- tainable development, offering a tapestry of experiences that celebrate its natural beauty, cultural heritage, and economic vibrancy. 125 of 13 125 of 13 Law and Economics, 2022, Vol. 16, No. 2 2.5 Research Method The research adopts a mixed-methods approach to comprehensively assess the FHP's impact. Quantitative data collection methods, such as surveys and statistical analysis, complement qualitative approaches like interviews and case studies. This triangulation of methods aims to provide a holistic understanding of the program's effectiveness. A purposive sampling technique is employed to select participants for the study. Ben- eficiary households enrolled in the FHP, local stakeholders, government officials, and community leaders constitute the primary sample. The sample size is determined to en- sure representation across diverse socio-economic backgrounds within Berastagi City. Quantitative data is gathered through structured surveys administered to beneficiary households. The surveys capture demographic information, income levels, access to edu- cation and healthcare, and perceptions of the program's impact. Qualitative data is ob- tained through in-depth interviews with program beneficiaries, local authorities, and key stakeholders involved in program implementation. p g p The research employs predefined indicators to assess the FHP's effectiveness. These indicators encompass economic metrics such as changes in household income, employ- ment status, access to education, healthcare utilization, and improvements in overall qual- ity of life. Additionally, qualitative assessments capture beneficiaries' experiences, chal- lenges, and perceptions of the program's efficacy. Quantitative data undergoes rigorous statistical analysis using appropriate methods, including regression analysis and descriptive statistics. This analysis aims to identify cor- relations, trends, and statistical significance in the program's impact on economic indica- tors. Qualitative data undergoes thematic analysis to extract key themes and insights. The research adheres to ethical guidelines, ensuring informed consent, confidential- ity, and privacy of participants. All data collection and analysis procedures prioritize par- ticipant well-being and respect local cultural sensitivities. To ensure the validity and reliability of findings, triangulation of data sources, mem- ber checking, and peer debriefing are employed. This process enhances the credibility and trustworthiness of the research outcomes. The study acknowledges potential limitations such as sample bias, constraints in gen- eralizing findings, and external factors influencing program outcomes. These limitations are carefully considered in the interpretation of results. The research culminates in the presentation of findings through comprehensive re- ports, data visualizations, and qualitative narratives. The results are synthesized to pro- vide a clear understanding of the FHP's effectiveness in reducing economic underdevel- opment in Berastagi City. 3.2.1 Consider the strengths and limitations of the program The evaluation of the Family Hope Program (FHP) in Berastagi City reveals several strengths and limitations that shape its effectiveness in addressing economic underdevel- opment. Understanding these aspects provides insights into potential reasons for its suc- cess or shortcomings, especially when compared with similar programs elsewhere: The FHP employs a targeted approach, focusing on vulnerable households, which ensures that resources are directed to those most in need. This personalized assistance enhances its impact on beneficiaries. p The program offers a comprehensive package of assistance, encompassing financial aid, education, healthcare, and skills development. This multifaceted approach addresses various dimensions of poverty and contributes to holistic development. The FHP fosters community engagement and social cohesion. Participatory activities and collaborative initiatives promote local ownership, encouraging sustainable develop- ment. Overreliance on traditional sectors like agriculture and seasonal tourism limits eco- nomic diversification. This dependency can hinder sustained growth and resilience against market fluctuations. Inadequate infrastructure, including roads and access to es- sential services, restricts economic opportunities and impedes the program's effectiveness in facilitating development. The seasonal nature of economic activities contributes to em- ployment instability, impacting the program's ability to provide year-round support. The success of the FHP can be attributed to its targeted approach, comprehensive sup- port, and emphasis on community engagement. However, shortcomings stem from limi- tations such as the lack of economic diversification and infrastructural deficiencies, which hinder sustained progress. p g Comparisons with analogous programs in other regions reveal parallels in challenges and successes. Many social welfare programs worldwide face similar hurdles related to dependency on traditional sectors, seasonal economic fluctuations, and infrastructural de- ficiencies. Successful programs often prioritize economic diversification, infrastructure development, and community engagement, mirroring the potential solutions for the FHP's limitations. 3.2.2 Recommendations for improving the effectiveness of the FHP or similar programs in reducing economic underdevelopment 3.2.2 Recommendations for improving the effectiveness of the FHP or similar programs in reducing economic underdevelopment Based on the research findings evaluating the Family Hope Program's (FHP) effec- tiveness in mitigating economic underdevelopment in Berastagi City, several key recom- mendations emerge to enhance the program's efficacy and address prevalent socio-eco- nomic challenges. These recommendations aim to bolster the FHP or similar initiatives, fostering sustainable development and alleviating economic constraints: a. Economic Diversification: Encourage Economic Opportunities: Foster economic diversification beyond tra- ditional sectors like agriculture and tourism. 3.1 Result The observed improvements in household income, access to education, and healthcare services signal the FHP's positive contribution toward alleviating economic constraints. This affirms the program's efficacy in empowering beneficiaries and enhanc- ing their economic well-being. Policymakers can leverage these successes as evidence of the program's effectiveness, encouraging sustained investments in similar social welfare initiatives. Prioritizing human capital development and targeted interventions can yield lasting socio-economic benefits. The reported sense of empowerment and strengthened community ties among bene- ficiaries underscore the program's broader impact beyond economic metrics. Enhanced social cohesion and engagement highlight the significance of community-driven initia- tives in fostering sustainable development. The cultivation of community engagement and empowerment should be integral to future developmental strategies. Encouraging participatory approaches and grassroots initiatives can amplify the impact of interven- tions and foster local ownership. Persistent challenges such as limited diversified employment opportunities and in- frastructural deficiencies underscore the need for adaptive strategies. These hurdles pose Law and Economics, 2022, Vol. 16, No. 2 126 of 13 constraints on sustained economic growth, signaling areas requiring targeted interven- tions. constraints on sustained economic growth, signaling areas requiring targeted interven- tions. Addressing these challenges demands multifaceted approaches. Policy reforms tar- geting economic diversification, infrastructure enhancements, and innovative partner- ships can mitigate these limitations, unlocking the city's full economic potential. The research highlights the imperative for adaptive programs and sustainable inter- ventions. Recommendations for program adaptation and sustained support emphasize the importance of continuous evolution to meet evolving community needs. Policymakers and stakeholders must embrace flexibility and innovation in program design. Adaptive strategies, informed by community feedback and emerging challenges, are pivotal for en- suring the long-term sustainability and relevance of development initiatives. 3.2 Discussion 3.2 Discussion 3.2 Discussion 3.2.1 Consider the strengths and limitations of the program 3.2.1 Consider the strengths and limitations of the program 3.2.1 Consider the strengths and limitations of the program Promote the growth of new indus- tries, small businesses, and entrepreneurship to create a more resilient and di- verse economy. Support Innovation and Technology: Invest in technological Law and Economics, 2022, Vol. 16, No. 2 127 of 13 advancements and innovation hubs to drive economic growth. Supporting tech- based enterprises and fostering innovation can create employment opportunities and enhance productivity. advancements and innovation hubs to drive economic growth. Supporting tech- based enterprises and fostering innovation can create employment opportunities and enhance productivity. b. b. Infrastructure Development: Improve Basic Infrastructure: Address infrastructural deficiencies by investing in road networks, utilities, and essential services. Enhanced infrastructure facilitates trade, connectivity, and access to markets, boosting economic activities. Promote Sustainable Practices: Integrate sustainable practices into infrastructure develop- ment to ensure long-term viability and environmental conservation. g y c. Human Capital Development: Enhance Education and Skills Training: Prioritize investments in education and vocational training to equip individuals with relevant skills for the job market. Focus on aligning skills training with emerging industry demands. Healthcare Accessibility: Improve access to quality healthcare services to enhance the popu- lation's health and productivity. Initiatives promoting preventive care and healthcare access contribute to a healthier workforce. d. Community Engagement and Empowerment: Foster Local Initiatives: Encourage community-driven projects and initiatives to address local challenges. Empower local communities by providing resources, training, and support for community-led development projects. Strengthen Social Networks: Facilitate platforms for networking, collaboration, and knowledge- sharing among community members. Strengthening social ties enhances mutual support and resilience within communities. e. Partnerships and Governance: Public-Private Partnerships (PPPs): Foster collaborations between government, private sector, NGOs, and local communities. PPPs leverage resources, expertise, and innovation for sustainable development initiatives. Good Governance Prac- tices: Promote transparent and accountable governance to ensure efficient re- source allocation, effective implementation, and monitoring of development pro- grams. g f. Program Adaptation and Evaluation: Continuous Assessment and Adaptation: Regularly evaluate program effective- ness and adapt strategies based on community feedback and evolving needs. Flexibility in program design ensures relevance and responsiveness to changing circumstances. Learning from Best Practices: Study successful models of poverty alleviation programs globally and integrate best practices into the FHP or similar initiatives. 3.2.1 Consider the strengths and limitations of the program 3.2.3 Interpreting Research Findings in the Context of Existing Literature and Theories 3.2.3 Interpreting Research Findings in the Context of Existing Literature and Theories The research findings evaluating the effectiveness of the Family Hope Program (FHP) in reducing economic underdevelopment in Berastagi City resonate within the broader context of existing literature and developmental theories. The observed improvements in education, skills development, and increased house- hold income align with Human Capital Theory. Investments in education and skill-build- ing contribute to enhanced human capital, resulting in higher productivity and economic growth within communities. The FHP's focus on providing comprehensive support aligns with Amartya Sen's Ca- pability Approach. By addressing multiple dimensions of poverty, such as access to edu- cation, healthcare, and economic opportunities, the program empowers individuals to ex- pand their capabilities and freedoms. Findings highlighting community engagement and social cohesion resonate with Community Development Theories. Participatory approaches, collaborative initiatives, and mutual support networks fostered by the FHP align with theories emphasizing com- munity-driven development. y p Limitations observed in economic diversification and infrastructure deficiencies re- flect aspects of Dependency Theory. Overreliance on traditional sectors and Law and Economics, 2022, Vol. 16, No. 2 128 of 13 infrastructural constraints contribute to economic vulnerabilities, echoing the challenges highlighted by this theory. y y The findings link to the broader agenda of the Sustainable Development Goals (SDGs), particularly goals related to poverty eradication, quality education, decent work and economic growth, and sustainable communities. The FHP's impact aligns with these global aspirations. Interpretation of the findings through the lens of public policy and program evalua- tion literature underscores the importance of adaptive strategies and continuous assess- ment. The need for program adaptation, infrastructure enhancements, and community- driven initiatives aligns with recommendations from program evaluation literature. The findings align with key principles of developmental economics frameworks, em- phasizing the significance of infrastructure development, economic diversification, hu- man capital investment, and inclusive community engagement for sustainable develop- ment. 4. Conclusions The Family Hope Program (FHP) in Berastagi City, which aims to reduce economic underdevelopment, is evaluated to determine its impact, successes, obstacles, and socio- economic implications. The findings synthesis highlights key insights that resonate with development intervention discourse. Economic opportunities, human capital develop- ment, and social cohesion in Berastagi City have improved because to the FHP. Income, education, and healthcare gains show its favorable influence on beneficiary households. The program's focus on community engagement, participatory activities, and skill-build- ing has empowered and connected recipients. These demonstrate community-driven pro- grams' transformative impact. Economic dependency on conventional sectors, infrastruc- tural issues, and seasonal economic volatility hinder advancement. These restrictions must be addressed for long-term program efficacy. The research makes recommendations to strengthen the FHP or similar initiatives. Economy diversification, infrastructure de- velopment, human capital development, and community-driven initiatives are key pro- gram enhancement measures. Sustainable development requires comprehensive ap- proaches, community engagement, and adaptable tactics, as supported by developmental theories and frameworks. This research supports comprehensive and inclusive develop- ment. Addressing multi-dimensional issues, empowering locals, and collaborating makes sustainable development possible. The Berastagi City Family Hope Program study shows how targeted interventions may change lives. While celebrating its triumphs, the research emphasizes the need for adaptable strategies, continuous investments, and coordinated efforts to overcome ongoing challenges. Policies, local authorities, and stakeholders can use the findings synthesis to make informed decisions and interventions to promote resil- ient, inclusive, and sustainable socio-economic development in Berastagi City and be- yond. Banerjee, A., Duflo, E., Goldberg, N., Karlan, D., Osei, R., Parienté, W., Shapiro, J., Thuysbaert, B., & Udry, C. (2015). A multifaceted program causes lasting progress for the very poor: Evidence from six countries. Science, 348(6236), 1260799. Athwal, B., Brill, L., Chesters, G., & Quiggin, M. (2011). Recession, poverty and sustainable livelihoods in Bradford. York: Joseph Rowntree Foundation. Bodden, M. (2010). Resistance on the national stage: Theater and politics in late New Order Indonesia (Vol. 123). Ohio University Press. Briguglio, L., Cordina, G., Farrugia, N., & Vella, S. (2014). Economic vulnerability and resilience: concepts and measurements. In Measuring Vulnerability in Developing Countries (pp. 47–65). Routledge. Filip, R., Gheorghita Puscaselu, R., Anchidin-Norocel, L., Dimian, M., & Savage, W. K. (2022). Global challenges to public health care Cox, K. R., & Mair, A. (1988). Locality and community in the politics of local economic development. Annals of the Association of American Geographers, 78(2), 307–325. References Athwal, B., Brill, L., Chesters, G., & Quiggin, M. (2011). Recession, poverty and sustainable livelihoods in Bradford. York: Joseph Rowntree Foundation. Banerjee, A., Duflo, E., Goldberg, N., Karlan, D., Osei, R., Parienté, W., Shapiro, J., Thuysbaert, B., & Udry, C. (2015). A multifaceted program causes lasting progress for the very poor: Evidence from six countries. Science, 348(6236), 1260799. Bodden, M. (2010). Resistance on the national stage: Theater and politics in late New Order Indonesia (Vol. 123). Ohio University Press. Briguglio, L., Cordina, G., Farrugia, N., & Vella, S. (2014). Economic vulnerability and resilience: concepts and measurements. In Measuring Vulnerability in Developing Countries (pp. 47–65). Routledge. Bodden, M. (2010). Resistance on the national stage: Theater and politics in late New Order Indonesia (Vol. 123). Ohio University Press. Briguglio, L., Cordina, G., Farrugia, N., & Vella, S. (2014). Economic vulnerability and resilience: concepts and measurements. In Measuring Vulnerability in Developing Countries (pp. 47–65). Routledge. Cox, K. R., & Mair, A. (1988). Locality and community in the politics of local economic development. Annals of the Association of American Geographers, 78(2), 307–325. Cox, K. R., & Mair, A. (1988). Locality and community in the politics of local economic development. Annals of the Association of American Geographers, 78(2), 307–325. Cox, K. R., & Mair, A. (1988). Locality and community in the politics of local economic development. Annals of the Association of American Geographers, 78(2), 307–325. Filip, R., Gheorghita Puscaselu, R., Anchidin-Norocel, L., Dimian, M., & Savage, W. K. (2022). Global challenges to public health care Filip, R., Gheorghita Puscaselu, R., Anchidin-Norocel, L., Dimian, M., & Savage, W. K. (2022). Global challenges to public health care 129 of 13 Law and Economics, 2022, Vol. 16, No. 2 129 of 13 systems during the COVID-19 pandemic: a review of pandemic measures and problems. Journal of Personalized Medicine, 12(8), 1295. Foster, A. D., & Rosenzweig, M. R. (2004). Agricultural productivity growth, rural economic diversity, and economic reforms: India, 1970–2000. Economic Development and Cultural Change, 52(3), 509–542. Govindan, K., Soleimani, H., & Kannan, D. (2015). Reverse logistics and closed-loop supply chain: A comprehensive review to explore the future. European Journal of Operational Research, 240(3), 603–626. Gray, R. (1992). Accounting and environmentalism: an exploration of the challenge of gently accounting for accountability, transparency and sustainability. Accounting, Organizations and Society, 17(5), 399–425. Helmsing, A. H. J. (2003). Local economic development: New generations of actors, policies and instruments for Africa. References Public Administration and Development: The International Journal of Management Research and Practice, 23(1), 67–76. Israr, S. M., & Islam, A. (2006). Good governance and sustainability: a case study from Pakistan. The International Journal of Health Planning and Management, 21(4), 313–325. Katoppo, M. (2000). Compassionate and free: an Asian woman’s theology. Wipf and Stock Publishers. Katoppo, M. (2000). Compassionate and free: an Asian woman’s theology. Wipf and Stock Publishers. Kessides, C. (1993). The contributions of infrastructure to economic development: a review of experience and policy implications. Kipp, R. S. (1993). Dissociated identities: ethnicity, religion, and class in an Indonesian society. University of Mich Malhotra, A., & Schuler, S. R. (2005). Women’s empowerment as a variable in international development. Measuring Empowerment: Cross-Disciplinary Perspectives, 1(1), 71–88. Martin, C. B. (2007). Tourism in the mountain south: A double-edged sword. Univ. of Tennessee Press. Patwardhan, I., Hurley, K. D., Thompson, R. W., Mason, W. A., & Ringle, J. L. (2017). Child maltreatment as a function of cumulative family risk: Findings from the intensive family preservation program. Child Abuse & Neglect, 70, 92–99. Pavão, N. (2016). Conditional cash transfer programs and electoral accountability: Evidence from Latin America. Latin American Politics and Society, 58(2), 74–99. Peiris, D., Sun, L., Patel, A., Tian, M., Essue, B., Jan, S., & Zhang, P. (2015). Systematic medical assessment, referral and treatment for diabetes care in China using lay family health promoters: protocol for the SMARTDiabetes cluster randomised controlled trial. Implementation Science, 11, 1–9. Raheim, S. (2017). Self-employment training and family development: An integrated strategy for family empowerment. In Reinventing human services (pp. 127–144). Routledge. Sembiring, J. A. (2019). CHARACTER EDUCATION IN TRANSLATION LEARNING AT UNIVERSITY OF QUALITY BERASTAGI. VISION, 15(1). Song, M., Podoynitsyna, K., Van Der Bij, H., & Halman, J. I. M. (2008). Success factors in new ventures: A meta-analysis. Journal of Product Innovation Management, 25(1), 7–27. Sutter, C., Bruton, G. D., & Chen, J. (2019). Entrepreneurship as a solution to extreme poverty: A review and future research directions. Journal of Business Venturing, 34(1), 197–214. White, J., & Morton, J. (2005). Mitigating impacts of HIV/AIDS on rural livelihoods: NGO experiences in sub-Saharan Africa. Development in Practice, 15(2), 186–199.
https://openalex.org/W2578761817	https://europepmc.org/articles/pmc5242421?pdf=render	English	null	Drawbacks of Dialysis Procedures for Removal of EDTA	PloS one	2,017	cc-by	6,003	RESEARCH ARTICLE Introduction Copyright: © 2017 Mo´nico et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Metal chelators, such as EDTA, are widely used for inhibition of proteases during protein puri- fication or during preparation of cell or tissue extracts [1]. EDTA is also employed in proce- dures to eliminate endotoxin from certain protein preparations [2] or to prevent oxidation by metals [3]. However, the presence of EDTA in biological samples, even at low concentrations, may interfere with assays like those aimed at assessing the effects of divalent cations on protein or cellular functions and hence, it is often necessary to remove the chelator. Dialysis or gel fil- tration are routinely used in research articles [2, 4] and recommended in commercial technical application notes for exchange or removal of low molecular weight buffer components, includ- ing EDTA, from cell or tissue extracts or protein preparations. Dialysis is also preferred when the volume of protein available or its concentration are too low for size exclusion chromatog- raphy with columns long enough to render a good separation. Moreover, dialysis is usually the method of choice when the protein is purified from bacterial inclusion bodies and the dena- tured protein needs to be refolded prior to the experiments by this procedure [5, 6]. This is the case of intermediate filament proteins, which are typically purified using buffers containing EDTA. In addition, they require solubilization either from inclusion bodies, in the case of Andreia Mo´nico1, Eva Martı´nez-Senra1, F. Javier Cañada1, Silvia Zorrilla2, Dolores Pe´rez- Sala1* 1 Department of Chemical and Physical Biology, Centro de Investigaciones Biolo´gicas, C.S.I.C., Madrid, Spain, 2 Department of Cellular and Molecular Biology, Centro de Investigaciones Biolo´gicas, C.S.I.C., Madrid, Spain * dperezsala@cib.csic.es * dperezsala@cib.csic.es a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 Editor: Eugene A. Permyakov, Russian Academy of Medical Sciences, RUSSIAN FEDERATION Editor: Eugene A. Permyakov, Russian Academy of Medical Sciences, RUSSIAN FEDERATION Abstract Ethylenediaminetetraacetic acid (EDTA) is a chelating agent commonly used in protein puri- fication, both to eliminate contaminating divalent cations and to inhibit protease activity. For a number of subsequent applications EDTA needs to be exhaustively removed. Most purifi- cation methods rely in extensive dialysis and/or gel filtration in order to exchange or remove protein buffer components, including metal chelators. We report here that dialysis protocols, even as extensive as those typically employed for protein refolding, may not effectively remove EDTA, which is reduced only by approximately two-fold and it also persists after spin-column gel filtration, as determined by NMR and by colorimetric methods. Remarkably, the most efficient removal was achieved by ultrafiltration, after which EDTA became virtually undetectable. These results highlight a potentially widespread source of experimental vari- ability affecting free divalent cation concentrations in protein applications. OPEN ACCESS Citation: Mo´nico A, Martı´nez-Senra E, Cañada FJ, Zorrilla S, Pe´rez-Sala D (2017) Drawbacks of Dialysis Procedures for Removal of EDTA. PLoS ONE 12(1): e0169843. doi:10.1371/journal. pone.0169843 Drawbacks of Dialysis Procedures for Removal of EDTA Andreia Mo´nico1, Eva Martı´nez-Senra1, F. Javier Cañada1, Silvia Zorrilla2, Dolores Pe´rez- Sala1* OPEN ACCESS Citation: Mo´nico A, Martı´nez-Senra E, Cañada FJ, Zorrilla S, Pe´rez-Sala D (2017) Drawbacks of Dialysis Procedures for Removal of EDTA. PLoS ONE 12(1): e0169843. doi:10.1371/journal. pone.0169843 Citation: Mo´nico A, Martı´nez-Senra E, Cañada FJ, Zorrilla S, Pe´rez-Sala D (2017) Drawbacks of Dialysis Procedures for Removal of EDTA. PLoS ONE 12(1): e0169843. doi:10.1371/journal. pone.0169843 Editor: Eugene A. Permyakov, Russian Academy of Medical Sciences, RUSSIAN FEDERATION Editor: Eugene A. Permyakov, Russian Academy of Medical Sciences, RUSSIAN FEDERATION Received: November 10, 2016 Accepted: December 23, 2016 Published: January 18, 2017 Protein dialysis Bovine serum albumin (BSA) at 1 mg/ml in 5 mM Pipes, pH 7.0 containing 1 mM EDTA was dialyzed against 5 mM Pipes, 1 mM DTT, pH 7.0 with four buffer changes and, subsequently, into 5 mM Pipes, 0.25 mM DTT, pH 7.0 with two changes. Dialysis of vimentin followed the standard refolding protocol [8]. Briefly, the human recombinant protein at 1 mg/ml in 5 mM Tris-HCl, pH 7.6, containing 8 M urea, 1 mM EDTA, 10 mM β-mercaptoethanol, 0.4 mM PMSF and approximately 150 mM KCl, was subjected to step-wise dialysis against 5 mM Pipes pH 7.0, 1 mM DTT containing 6 M urea, then 4 M urea, 2 M urea and no urea at r.t., and finally, to two additional steps against 5 mM Pipes, pH 7.0, 0.25 mM DTT, the last one for 16 h at 16˚C. Spin-column gel filtration and ultrafiltration Gel filtration was conducted using PD SpinTrap G-25 1 ml columns equilibrated with 5 mM Pipes, 0.1 mM DTT, pH 7.0, before loading 140 μl samples of vimentin or BSA and eluting according to the instructions of the manufacturer. For ultrafiltration, BSA or vimentin samples (250 μl) containing 1 mM EDTA were diluted 10-fold with EDTA-free buffer, applied to Milli- pore Amicon Ultra filter units (10 K pore size) and centrifuged at 3000xg for 15 min at 16˚C, which concentrated the samples down to their original volume. Then, samples were diluted again 10-fold with buffer without EDTA and the procedure was repeated. Materials Amicon ultrafiltration devices (10K cut off) were from Millipore. PD-SpinTrap G-25 columns were from GE Healthcare. Slide-A-Lyzer MINI Dialysis devices (20K cut off) were from Thermo. 96-well plates were from Falcon. Recombinant hamster vimentin was from Cytoskel- eton, Inc. Recombinant human vimentin was from Biomedal (Sevilla, Spain). Other reagents were of the highest quality from Sigma. Data Availability Statement: All relevant data are within the paper. Data Availability Statement: All relevant data are within the paper. Funding: This work has been funded by the European Union’s Horizon 2020 research and innovation programme under the Marie Sklowdowska-Curie grant agreement number 675132 (http://cordis.europa.eu/project/rcn/ 198275_en.html), and by grants from the Spanish Ministerio de Economı´a y Competitividad (MINECO/FEDER, http://www.mineco.gob.es/ portal/site/mineco/idi) SAF2015-68590R to DPS and CTQ2015-64597-C2-2-P to FJC. The funders PLOS ONE \| DOI:10.1371/journal.pone.0169843 January 18, 2017 1 / 9 Incomplete Removal of EDTA by Dialysis recombinant proteins expressed in bacteria, or from insoluble eukaryotic cell cytoskeletal frac- tions, for which high urea concentrations, up to 9.5 M, are used [7–9]. Later, the denatured proteins are refolded by step-wise dialysis to gradually remove urea, usually in low salt buffer to keep the protein unassembled, as high ionic strength triggers polymerization [8]. In fact, intermediate filament proteins such as desmin, vimentin or glial fibrillary acidic protein (GFAP) purified through these methods, have been widely studied to assess polymerization or association changes induced by increasing ionic strength or by divalent cations, usually at mil- limolar concentrations [9, 10]. had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. Nevertheless, confirmation of adequate elimination of additives used during purification or refolding is necessary for subsequent applications, above all, when studying the impact of low concentrations of various ligands on protein structure or function. Here we show that EDTA can be carried on unnoticed during protein dialysis. Moreover, we provide simple methods for its detection and improved removal. NMR analysis NMR spectra were acquired in a 500 MHz Bruker AVANCE equipped with a SEF 19F-1H probe or a 600MHz Bruker AVANCE equipped with a cryogenic triple resonance TXI probe. In the case of samples dissolved in deuterated buffer (20 mM deuterated Tris, Cambridge Iso- tope, UK), spectra were acquired with a simple 90˚ pulse sequence (zg Bruker pulse sequence) and with 32K data points and 2 second recovery delay and 10 ppm of spectral width centered at 4.7 ppm (chemical shift of residual HDO signal). In the case of samples dissolved in non- deuterated buffers, a 10% volume of deuterated water was added for locking deuterium signal and the standard Bruker pulse sequence “zgesgp” using excitation sculpting gradients for water signal suppression was used. The spectra were acquired at 25˚C with 32K points, 2 sec- ond recovery delay and 14 ppm of spectral width centered at 4.7 ppm (chemical shift of water). The spectra were acquired using from 8 to 2048 scans depending on the sample concentration; no line broadening was applied in the processing. Bruker TOPSPIN software was used for acquisition and processing the spectra. For testing the pH dependency of chemical shift signals of EDTA, samples of 1 to 3 mM EDTA, with and without cation (Ca2+, Zn2+, Mg2+ or La3+), in 20 mM deuterated Tris were prepared at different pH (between 5 and 9) by acidification with deuterated hydrochloric acid. The actual pH was measured after the addition of the corre- sponding dichloride salt of each cation. EDTA and ZnCl2 solutions were prepared in deuter- ated water. Colorimetric estimation of EDTA concentration The concentration of EDTA in the protein solutions was estimated through a colorimetric assay by monitoring its competition with 4-(2-pyridylazo)-resorcinol (PAR) for Zn binding. The binding of zinc to PAR forms a colored complex with absorbance at 492 nm [11]. The presence of EDTA induces a decrease in the formation of this complex. A calibration curve PLOS ONE \| DOI:10.1371/journal.pone.0169843 January 18, 2017 2 / 9 Incomplete Removal of EDTA by Dialysis was obtained by titration of samples containing 100 μM PAR and 10 μM Zn with known con- centrations of EDTA and measuring the absorbance at 492 nm, using a Varioskan Flash (Thermo) microplate reader. The amount of EDTA in the protein preparations was deter- mined from their absorbance at 492 nm after incubation with 100 μM PAR and 10 μM Zn, and extrapolation using the calibration curve. Sample volumes of 1 to 10 μl were used in a typi- cal total assay volume of 100 μl. Measurements were performed 5 min after mixing of reagents. PLOS ONE \| DOI:10.1371/journal.pone.0169843 January 18, 2017 Results and Discussion We have recently explored protein modification by electrophilic lipids using both cultured cells and commercially available protein preparations, including the IF protein vimentin [12, 13]. In order to study the effect of micromolar concentrations of divalent cations on these pro- cesses we set out to rule out the presence of metal chelators in the protein preparation. For this we first conducted NMR experiments. Proton NMR has been previously used for detecting either EDTA or divalent cations as their EDTA chelates in blood samples obtained with EDTA as anticoagulant [14, 15]. In addition, NMR methods have been employed to quantitate metals in human serum upon addition of exogenous EDTA [16]. As shown in Fig 1, free EDTA gives two singlet peaks, the one at low field corresponding to the 8 methylene protons of acetyl moi- eties, and the one at high field to the 4 protons of the ethylenediamine moiety. The shifts of these peaks are strongly dependent on the pH of the sample and their signals show a significant broadening at low pH (Fig 1). Moreover, complexation of EDTA with different metals gives rise to characteristic patterns, where the methylene protons of acetyl moieties become un- equivalent, resulting in differentiated signals due to the characteristic structures of the chelates, as depicted in Fig 1 for the EDTA-calcium complex. This ensures the specificity of the detec- tion and minimizes pH dependence, since the shifts of the signals of the complexed EDTA form appear constant in this pH range (Fig 1 and [16]). Therefore, protein samples were analyzed after adding known amounts of a divalent cation such as calcium or zinc, the latter case being illustrated in Fig 2. This revealed the presence of a substantial amount of EDTA in a commercial sample, which is presented as a lyophilized PLOS ONE \| DOI:10.1371/journal.pone.0169843 January 18, 2017 3 / 9 Incomplete Removal of EDTA by Dialysis Fig 1. Proton NMR spectra of EDTA and its Ca2+-complexed forms at different pH. EDTA (3 mM) and CaCl2 (1.5 mM) were mixed in deuterated Tris buffer to yield equimolecular amounts of free and Ca2+- complexed EDTA. Proton NMR spectra of the mixture were acquired at different pH: (A) pH 8.2, (B) pH 7.0, (C) pH 5.9. The signals corresponding to protons of the free and Ca2+-complexed forms of EDTA are indicated. The NMR spectrum of EDTA alone is identical to that of free EDTA presented. Results and Discussion The moieties responsible for each of the components of the signals are depicted in bold. The signal corresponding to Tris buffer is labeled with . doi:10 1371/journal pone 0169843 g001 Fig 1. Proton NMR spectra of EDTA and its Ca2+-complexed forms at different pH. EDTA (3 mM) and CaCl2 (1.5 mM) were mixed in deuterated Tris buffer to yield equimolecular amounts of free and Ca2+- complexed EDTA. Proton NMR spectra of the mixture were acquired at different pH: (A) pH 8.2, (B) pH 7.0, (C) pH 5.9. The signals corresponding to protons of the free and Ca2+-complexed forms of EDTA are indicated. The NMR spectrum of EDTA alone is identical to that of free EDTA presented. The moieties responsible for each of the components of the signals are depicted in bold. The signal corresponding to Tris buffer is labeled with . doi:10.1371/journal.pone.0169843.g001 doi:10.1371/journal.pone.0169843.g001 preparation that after reconstitution in water should yield a solution of folded protein in 5 mM Pipes pH 7.0, 1 mM DTT, 5% (w/v) sucrose and 1% (w/v) dextran. In order to obtain a preparation more suitable for our studies we subjected a human vimentin sample in buffer containing 8 M urea, 5 mM Tris-HCl pH 7.6, 1 mM EDTA, 10 mM β-mercaptoethanol, 0.4 mM PMSF and approximately 150 mM KCl to the typical renaturation dialysis procedure, which included four sequential steps with at least six changes against EDTA-free buffers [8]. NMR analysis of the protein solutions thus obtained showed the retention of EDTA after dialy- sis (Fig 2A). g A complementary colorimetric assay was also used to further confirm the presence of EDTA and to estimate its levels in the protein preparations after dialysis. For this we took advantage of the ability of the compound 4-(2-pyridylazo)-resorcinol (PAR) to form a colored 2:1 complex with zinc with high affinity (effective dissociation constant of the Zn(PAR)2 com- plex, at pH 7.0, 2.1 x 10−12 M2 [17]), which presents an absorption maximum at 492 nm and can be measured spectrophotometrically. In turn, EDTA forms a 1:1 complex with zinc of extremely high affinity: 6 x 10−14 M at pH 7.0 [18]. Therefore, solutions containing EDTA should compete with PAR for zinc binding, in such a way that the decrease in the absorption at 492 nm measured at the equilibrium can be used to infer the concentration of EDTA present in the sample. PLOS ONE \| DOI:10.1371/journal.pone.0169843 January 18, 2017 Results and Discussion Middle panel: reference proton spectrum of EDTA in presence of Zn2+, the two coupled doublets corresponding to the AB spin system of the methylene protons of the four acetyl groups, non-equivalent due to the structure of the metal chelate, and the singlet corresponding to the four equivalent protons of the ethylenediamine moiety are apparent. Lower panel: reference proton spectrum of EDTA at pH 7.2, only the two characteristic singlets of uncomplexed EDTA appear. (B) Detection of EDTA by colorimetric analysis. Upper panel: Calibration curve showing the dependence on EDTA concentration of the absorbance at 492 nm of mixtures containing 100 μM PAR and 10 μM ZnCl2. Lower panel: Amount of EDTA remaining in samples from vimentin and BSA subjected to extensive dialysis as determined from the absorbance at 492 nm after incubation with PAR and ZnCl2, using the calibration curve. Initial EDTA concentration in the samples was 1 mM. Data shown are mean ± SD of 4 assays. quantitation of EDTA in protein samples. (A) NMR analysis. Upper panel: 500MHz Fig 2. Detection and quantitation of EDTA in protein samples. (A) NMR analysis. Upper panel: 500MHz 1D-proton NMR spectrum of a commercial sample of vimentin (0.7 μM) with added 150 μM ZnCl2. Signals for the protons of sucrose, a stabilizer present in the sample as indicated in specifications, and those of PIPES buffer are easily identified. An additional singlet at 2.79 ppm and two coupled doublets at 3.25 and 3.32 ppm (2JHH = 17.3 Hz) are also observed. Middle panel: reference proton spectrum of EDTA in presence of Zn2+, the two coupled doublets corresponding to the AB spin system of the methylene protons of the four acetyl groups, non-equivalent due to the structure of the metal chelate, and the singlet corresponding to the four equivalent protons of the ethylenediamine moiety are apparent. Lower panel: reference proton spectrum of EDTA at pH 7.2, only the two characteristic singlets of uncomplexed EDTA appear. (B) Detection of EDTA by colorimetric analysis. Upper panel: Calibration curve showing the dependence on EDTA concentration of the absorbance at 492 nm of mixtures containing 100 μM PAR and 10 μM ZnCl2. Lower panel: Amount of EDTA remaining in samples from vimentin and BSA subjected to extensive dialysis as determined from the absorbance at 492 nm after incubation with PAR and ZnCl2, using the calibration curve. Initial EDTA concentration in the samples was 1 mM. Results and Discussion For this, we titrated known amounts of EDTA into 100 μM PAR and 10 μM zinc assay mixtures and measured the absorbance at 492 nm to build a calibration curve (Fig 2B, upper panel). We then assayed the EDTA present in the dialysates from their absorbance PLOS ONE \| DOI:10.1371/journal.pone.0169843 January 18, 2017 4 / 9 Incomplete Removal of EDTA by Dialysis Fig 2. Detection and quantitation of EDTA in protein samples. (A) NMR analysis. Upper panel: 500MHz 1D-proton NMR spectrum of a commercial sample of vimentin (0.7 μM) with added 150 μM ZnCl2. Signals for the protons of sucrose, a stabilizer present in the sample as indicated in specifications, and those of PIPES buffer are easily identified. An additional singlet at 2.79 ppm and two coupled doublets at 3.25 and 3.32 ppm (2JHH = 17.3 Hz) are also observed. Middle panel: reference proton spectrum of EDTA in presence of Zn2+, the two coupled doublets corresponding to the AB spin system of the methylene protons of the four acetyl groups, non-equivalent due to the structure of the metal chelate, and the singlet corresponding to the four equivalent protons of the ethylenediamine moiety are apparent. Lower panel: reference proton spectrum of EDTA at pH 7.2, only the two characteristic singlets of uncomplexed EDTA appear. (B) Detection of EDTA by colorimetric analysis. Upper panel: Calibration curve showing the dependence on EDTA concentration of the absorbance at 492 nm of mixtures containing 100 μM PAR and 10 μM ZnCl2. Lower panel: Amount of EDTA remaining in samples from vimentin and BSA subjected to extensive dialysis as determined from the absorbance at 492 nm after incubation with PAR and ZnCl2, using the calibration curve. Initial EDTA concentration in the samples was 1 mM. Data shown are mean ± SD of 4 assays. Fig 2. Detection and quantitation of EDTA in protein samples. (A) NMR analysis. Upper panel: 500MHz Fig 2. Detection and quantitation of EDTA in protein samples. (A) NMR analysis. Upper panel: 500MHz 1D-proton NMR spectrum of a commercial sample of vimentin (0.7 μM) with added 150 μM ZnCl2. Signals for the protons of sucrose, a stabilizer present in the sample as indicated in specifications, and those of PIPES buffer are easily identified. An additional singlet at 2.79 ppm and two coupled doublets at 3.25 and 3.32 ppm (2JHH = 17.3 Hz) are also observed. PLOS ONE \| DOI:10.1371/journal.pone.0169843 January 18, 2017 Results and Discussion Data shown are mean ± SD of 4 assays. doi:10.1371/journal.pone.0169843.g002 doi:10.1371/journal.pone.0169843.g002 when mixed with the above specified concentrations of PAR and zinc. These measurements evidenced that samples of vimentin or bovine serum albumin (BSA) originally containing 1 mM EDTA still retained ~450 μM EDTA after dialysis, in good agreement with parallel deter- minations by NMR (Fig 2A). Therefore, EDTA removal can be easily and conveniently moni- tored by this procedure. Nevertheless, it should be taken into account that other metal-binding compounds or proteins present in the sample may also compete with PAR for zinc binding, for which detection of EDTA by several methods is advisable. EDTA was also not removed from solutions of other intermediate filament proteins such as GFAP or desmin treated under similar conditions (unpublished observations). As the amount of EDTA remaining in the samples was incompatible with many protein studies, we attempted to setup a procedure to effectively remove this chelator. Bearing in mind that the quantity of protein available for biological assays is often limited, we used protocols for small sample volumes, namely, spin column gel filtration and ultrafiltration (Fig 3A). We started by processing the protein solutions already subjected to dialysis by spin-column gel fil- tration. Although EDTA was further reduced by this treatment, the protein solutions still con- tained ~200 μM EDTA, as determined by the PAR competition assay (Fig 3B). As an alternative, we turned to ultrafiltration to remove the EDTA. BSA or vimentin samples con- taining 1 mM EDTA were diluted 10-fold with EDTA-free buffer and ultrafiltrated as indi- cated in Fig 3A, which concentrated the samples down to their original volume. Then, samples were diluted again 10-fold with buffer without EDTA and the procedure was repeated. Finally, assessment of EDTA content by the colorimetric assay confirmed that EDTA was more 5 / 9 PLOS ONE \| DOI:10.1371/journal.pone.0169843 January 18, 2017 Incomplete Removal of EDTA by Dialysis Fig 3. Procedures for EDTA removal from protein samples and detection of EDTA. (A) The protein samples containing 1 mM EDTA were subjected to the indicated purification procedures. Left panel: Samples of BSA or vimentin were subjected to spin column gel filtration as detailed in methods. Right panel: protein samples were applied to Millipore Amicon Ultra filter units (10 K pore size) and subjected to two rounds of ultrafiltration, as described in the text. Results and Discussion The signal pattern of the AB system at 3.32 and 3.25 ppm and the singlet at 2.79 ppm corresponding to the Zn2+-EDTA chelate are clearly visible. Fig 3. Procedures for EDTA removal from protein samples and detection of EDTA. (A) The protein samples containing 1 mM EDTA were subjected to the indicated purification procedures. Left panel: Samples of BSA or vimentin were subjected to spin column gel filtration as detailed in methods. Right panel: protein samples were applied to Millipore Amicon Ultra filter units (10 K pore size) and subjected to two rounds of ultrafiltration, as described in the text. (B) Colorimetric determination of EDTA present in protein samples after diverse purification procedures using the PAR competition assay. Results shown are mean ± SD of 4 (dialysis plus gel filtration), 2 to 7 (ultrafiltration), or 3 to 7 (ultrafiltration plus dialysis) assays. (C) NMR analysis. Upper panel: 600MHz 1D-proton NMR spectrum of an ultrafiltrated sample of vimentin (1.8 μM final concentration). Signals of buffer and additives used in the purification (glycerol, from the ultrafiltration filters, and DTT) are observed. Protein signals appear at baseline noise level and are not recognizable. Middle panel: The same sample analyzed in the upper panel monitored after addition of 150 μM of ZnCl2. A quadruplet that appears at 3.92 ppm corresponds to trifluoroethanol added for referencing. Lower panel: monitorization of the sample after addition of 20 μM EDTA. The signal pattern of the AB system at 3.32 and 3.25 ppm and the singlet at 2.79 ppm corresponding to the Zn2+-EDTA chelate are clearly visible. doi:10.1371/journal.pone.0169843.g003 effectively removed by this procedure (Fig 3B). At this point, EDTA was not detectable by NMR, limiting its possible concentration to 10 μM or lower in the undiluted protein sample (Fig 3C). A further advantage of this method is that protein dilution can be minimized by monitoring the final retentate volume. Lastly, combination of ultrafiltration and dialysis led to the reduction of EDTA in the protein samples below 5 μM according to the colorimetric assay (Fig 3B). We have found numerous examples in the literature and in technical notes and protocols where dialysis is employed to remove EDTA. In contrast, evidences on the limitations of this method to completely remove the chelator are scarce and not readily found in bibliographic searches. Results and Discussion (B) Colorimetric determination of EDTA present in protein samples afte diverse purification procedures using the PAR competition assay. Results shown are mean ± SD of 4 (dialysis plus gel filtration), 2 to 7 (ultrafiltration), or 3 to 7 (ultrafiltration plus dialysis) assays. (C) NMR analysis. Upper panel: 600MHz 1D-proton NMR spectrum of an ultrafiltrated sample of vimentin (1.8 μM final concentration). Signals of buffer and additives used in the purification (glycerol, from the ultrafiltration filters, and DTT) are observed. Protein signals appear at baseline noise level and are not recognizable. Middle panel: The same sample analyzed in the upper panel monitored after addition of 150 μM of ZnCl2. A quadruplet that appears at 3.92 ppm corresponds to trifluoroethanol added for referencing. Lower panel: monitorization of the sample after addition of 20 μM EDTA. The signal pattern of the AB system at 3.32 and 3.25 ppm and the singlet at 2.79 ppm corresponding to the Zn2+-EDTA chelate are clearly visible. Fig 3. Procedures for EDTA removal from protein samples and detection of EDTA. (A) The protein Fig 3. Procedures for EDTA removal from protein samples and detection of EDTA. (A) The protein samples containing 1 mM EDTA were subjected to the indicated purification procedures. Left panel: Samples of BSA or vimentin were subjected to spin column gel filtration as detailed in methods. Right panel: protein samples were applied to Millipore Amicon Ultra filter units (10 K pore size) and subjected to two rounds of ultrafiltration, as described in the text. (B) Colorimetric determination of EDTA present in protein samples after diverse purification procedures using the PAR competition assay. Results shown are mean ± SD of 4 (dialysis plus gel filtration), 2 to 7 (ultrafiltration), or 3 to 7 (ultrafiltration plus dialysis) assays. (C) NMR analysis. Upper panel: 600MHz 1D-proton NMR spectrum of an ultrafiltrated sample of vimentin (1.8 μM final concentration). Signals of buffer and additives used in the purification (glycerol, from the ultrafiltration filters, and DTT) are observed. Protein signals appear at baseline noise level and are not recognizable. Middle panel: The same sample analyzed in the upper panel monitored after addition of 150 μM of ZnCl2. A quadruplet that appears at 3.92 ppm corresponds to trifluoroethanol added for referencing. Lower panel: monitorization of the sample after addition of 20 μM EDTA. PLOS ONE \| DOI:10.1371/journal.pone.0169843 January 18, 2017 Results and Discussion Our observations together with these previous works [19, 20] indicate that it should not be assumed that EDTA is thoroughly removed by dialysis procedures. Thus, complete removal of EDTA should be assessed either by the procedures described herein or by other available methods, like HPLC or spectrometric procedures [21, 22], as in [19]. The efficiency of EDTA removal by dialysis may vary with buffer composition. In particular, Kuzmenko et al., showed that EDTA removal was poorer when dialysing against buffers with low ionic strength [19], with incomplete removal even after six buffer changes over three days dialysis. To test the influence of ionic strength on the effectiveness of dialysis under our experimental conditions, we have assessed the removal of EDTA (1 mM initial concentration) from a BSA PLOS ONE \| DOI:10.1371/journal.pone.0169843 January 18, 2017 6 / 9 Incomplete Removal of EDTA by Dialysis sample by dialysis performed as in Fig 2, except that 150 mM NaCl was included in all the dial- ysis buffers and the procedure was carried out at r.t. Under these conditions, EDTA remaining in the protein sample was only 33.5 ± 35.0 μM (average value ± SD of 7 determinations), com- pared to 290.5 ± 23.4 μM EDTA remaining in the absence of NaCl (mean ± SD of 5 determina- tions). Therefore, these results confirm the previous observations [19] and show the superior effectiveness of EDTA removal by dialysis at physiological ionic strength. The concentration of EDTA remaining after dialysis will likely not affect the results of stud- ies employing millimolar concentrations of divalent cations [23], which in some cases are even performed in the presence of 1 mM EDTA [9]. Nevertheless, they should be taken into account when performing assays with micromolar concentrations of metals. The presence of EDTA can also affect various protein parameters, like thermal denaturation [20], or induce pH alter- ations upon addition of divalent cations [24]. Also, in studies assessing the biological effects of dialyzed proteins or cellular or plasma fractions, residual EDTA could be responsible for some of the effects observed. Potential reasons for the lack of effectiveness of dialysis procedures could include the reported ability of EDTA to form supramolecular aggregates in solution, as well as its capacity to associate with proteins ([25] and references therein). Conclusions Results presented here show that, despite its low molecular mass (292.24 Da), EDTA is often retained inside dialysis bags with pore-sizes over a hundred times larger, and extending the dialysis time and the number of buffer changes does not seem to allow successful removal of the chelator in those cases. Therefore, determination of the level of EDTA in the dialysates is needed whenever interference of this compound with the intended assays is suspected. For an initial assessment of the potential presence of EDTA, we suggest a fast and cost-efficient method, suitable for small sample amounts, which is based on the competition of EDTA and the zinc-binding compound PAR for complexation of this divalent cation. After testing proce- dures other than dialysis routinely employed to separate macromolecules from small com- pounds, we concluded that ultracentrifugation could be a good choice to remove EDTA when necessary. Moreover, when the protein needs to be subjected to dialysis as part of a refolding protocol, the combination of dialysis and ultrafiltration in no particular order renders opti- mum results in terms of chelator elimination. Results and Discussion For instance, EDTA has been reported to bind to α-lactalbumin altering the equilibrium between different conformers of the protein and its apparent thermal stability [26]. Size exclusion chromatography using spin desalting columns was among the methods we evaluated to diminish the concentration of EDTA and we found this procedure less effective compared with ultrafiltration. However, we cannot rule out that the use of longer columns with higher resolution may improve the species separation. In addition, care should be exer- cised with other chromatographic procedures employed in protein purification since EDTA has been reported to be abnormally retained in ion exchange chromatography under certain conditions [24]. Author Contributions Resources: DPS FJC SZ. Supervision: DPS SZ FJC. Validation: DPS SZ FJC. Validation: DPS SZ FJC. Visualization: AM SZ FJC DPS. Visualization: AM SZ FJC DPS. Writing – original draft: DPS SZ FJC. Writing – original draft: DPS SZ FJC. Writing – review & editing: AM EMS FJC SZ DPS. Writing – review & editing: AM EMS FJC SZ DPS. Writing – review & editing: AM EMS FJC SZ DPS. Acknowledgments We thank Prof. Germa´n Rivas (CIB, CSIC) for insightful comments. The technical assistance of MJ Carrasco is gratefully appreciated. 7 / 9 PLOS ONE \| DOI:10.1371/journal.pone.0169843 January 18, 2017 Incomplete Removal of EDTA by Dialysis PLOS ONE \| DOI:10.1371/journal.pone.0169843 January 18, 2017 References 1. Grabski AC. Advances in preparation of biological extracts for protein purification. Methods Enzymol. 2009; 463: 285–303. doi: 10.1016/S0076-6879(09)63018-4 PMID: 19892178 2. Wright JR, Zlogar DF, Taylor JC, Zlogar TM, Restrepo CI. Effects of endotoxin on surfactant protein A and D stimulation of NO production by alveolar macrophages. Am J Physiol. 1999; 276: L650–8. PMID: 10198363 3. Stadtman ER. Metal ion-catalyzed oxidation of proteins: biochemical mechanism and biological conse- quences. Free Radic Biol Med. 1990; 9: 315–25. PMID: 2283087 4. Wingender E, Dilloo D, Seifart KH. Zinc ions are differentially required for the transcription of ribosomal 5S RNA and tRNA in a HeLa-cell extract. Nucleic Acids Res. 1984; 12: 8971–85. PMID: 6440120 5. Rudolph R, Lilie H. In vitro folding of inclusion body proteins. FASEB J. 1996; 10: 49–56. PMID: 8566547 6. Tsumoto K, Ejima D, Kumagai I, Arakawa T. Practical considerations in refolding proteins from inclusion bodies. Protein Expr Purif. 2003; 28: 1–8. PMID: 12651100 7. Quinlan RA, Moir RD, Stewart M. Expression in Escherichia coli of fragments of glial fibrillary acidic pro- tein: characterization, assembly properties and paracrystal formation. J Cell Sci. 1989; 93: 71–83. PMID: 2693466 8. Herrmann H, Hofmann I, Franke WW. Identification of a nonapeptide motif in the vimentin head domain involved in intermediate filament assembly. J Mol Biol. 1992; 223: 637–50. PMID: 1542111 9. Lin YC, Broedersz CP, Rowat AC, Wedig T, Herrmann H, Mackintosh FC, et al. Divalent cations cross- link vimentin intermediate filament tail domains to regulate network mechanics. J Mol Biol. 2010; 399: 637–44. doi: 10.1016/j.jmb.2010.04.054 PMID: 20447406 10. Bar H, Sharma S, Kleiner H, Mucke N, Zentgraf H, Katus HA, et al. Interference of amino-terminal des- min fragments with desmin filament formation. Cell Motil Cytoskeleton. 2009; 66: 986–99. doi: 10.1002/ cm.20396 PMID: 19530175 11. Koch M, Bhattacharya S, Kehl T, Gimona M, Vasak M, Chazin W, et al. Implications on zinc binding to S100A2. Biochim Biophys Acta. 2007; 1773: 457–70. doi: 10.1016/j.bbamcr.2006.12.006 PMID: 17239974 12. Gharbi S, Garzo´n B, Gayarre J, Timms J, Pe´rez-Sala D. Study of protein targets for covalent modifica- tion by the antitumoral and anti-inflammatory prostaglandin PGA1: focus on vimentin. J Mass Spectrom. 2007; 42:1474–84. doi: 10.1002/jms.1291 PMID: 17960581 PLOS ONE \| DOI:10.1371/journal.pone.0169843 January 18, 2017 8 / 9 Incomplete Removal of EDTA by Dialysis 13. Pe´rez-Sala D, Oeste CL, Martı´nez AE, Garzo´n B, Carrasco MJ, Cañada FJ. References Vimentin filament organiza- tion and stress sensing depend on its single cysteine residue and zinc binding. Nat Commun. 2015; 6: 7287. doi: 10.1038/ncomms8287 PMID: 26031447 14. Barton RH, Waterman D, Bonner FW, Holmes E, Clarke R, Nicholson JK, et al. The influence of EDTA and citrate anticoagulant addition to human plasma on information recovery from NMR-based metabolic profiling studies. Mol BioSyst. 2010; 6: 215–24. doi: 10.1039/b907021d PMID: 20024083 15. Nicholson JK, Buckingham MJ, Sadler PJ. High resolution 1H n.m.r. studies of vertebrate blood and plasma. Biochem J. 1983; 211: 605–15. PMID: 6411064 16. Somashekar BS, Ijare OB, Nagana Gowda GA, Ramesh V, Gupta S, Khetrapal CL. Simple pulse- acquire NMR methods for the quantitative analysis of calcium, magnesium and sodium in human serum. Spectrochimica Acta A, Mol Biomol Spectrosc. 2006; 65: 254–60. 17. Kocyla A, Pomorski A, Krezel A. Molar absorption coefficients and stability constants of metal com- plexes of 4-(2-pyridylazo)resorcinol (PAR): Revisiting common chelating probe for the study of metallo- proteins. Journal Inorg Biochem. 2015; 152: 82–92. 18. Krezel A, Maret W. The biological inorganic chemistry of zinc ions. Arch Biochem Biophys. 2016; 611: 3–19. doi: 10.1016/j.abb.2016.04.010 PMID: 27117234 19. Kuzmenko AI, Wu H, Bridges JP, McCormack FX. Surfactant lipid peroxidation damages surfactant pro- tein A and inhibits interactions with phospholipid vesicles. J Lipid Res. 2004; 45: 1061–8. doi: 10.1194/ jlr.M300360-JLR200 PMID: 15026426 20. Searcy DG, Greif L. Apparent thermal destabilization of Escherichia coli nucleoprotein due to the incom- plete dialysis of EDTA. Biochim Biophys Acta. 1976; 418: 133–6. PMID: 812553 21. Kord AS, Tumanova I, Matier WL. A novel HPLC method for determination of EDTA in a cataract inhibit- ing ophthalmic drug. J Pharm Biomed Anal. 1995; 13: 575–80. PMID: 9696572 22. Kratochvil B, White MC. Spectrophotometric Determination of Microgram Quantities of (Ethylenedini- trilo)Tetraacetic Acid with Bis(2,4,6-Tripyridyl-S-Triazine)Iron(Ii). Anal Chem. 1965; 37: 111–3. PMID: 14230701 23. Dammann C, Koster S. Dynamics of counterion-induced attraction between vimentin filaments followed in microfluidic drops. Lab Chip. 2014; 14: 2681–7. doi: 10.1039/c3lc51418h PMID: 24834442 24. Sharpe JC, London E. Inadvertent concentrating of EDTA by ion exchange chromatography: avoiding artifacts that can interfere with protein purification. Anal Biochem. 1997; 250: 124–5. doi: 10.1006/abio. 1997.2210 PMID: 9234908 25. Muller M, Haeberli A. pH-dependent formation of ethylenediaminetetraacetic acid supramolecular aggregates. FEBS Lett. 1994; 340: 17–21. PMID: 8119401 26. Kronman MJ, Bratcher SC. An experimental artifact in the use of chelating metal ion buffers. PLOS ONE \| DOI:10.1371/journal.pone.0169843 January 18, 2017 References Binding of chelators to bovine alpha-lactalbumin. J Biol Chem. 1983; 258: 5707–9. PMID: 6406505 9 / 9
https://openalex.org/W1484894087	https://zenodo.org/records/2054949/files/article.pdf	English	null	THE PRESCRIPTION OF TUBERCULIN.	Lancet	1,911	public-domain	1,674	1660 1660 the interest he took in the matter, and to Miss. D. Dawson for her assistance in the laboratory the interest he took in the matter, and to Miss. D. Dawson for her assistance in the laboratory met. Anatomy, physiology, theory of movement, hygiene, &c., will be included in the syllabus. The practice of gymnastics will be learnt and taught by students, for which latter purpose classes of school children will be available. The course of training will be of two years’ duration, but is so arranged that the first year will form a complete course by itself, fitting the student for posts in elementary schools. The second year’s course will enable the graduates to take full charge of any school of whatever standing. A special feature of the institute will be that facilities will be afforded to medical men to gain information and to critically observe the work done.-I am, Sir, yours faithfully, ce in the laboratory. I am, Sir, yours faithfully, ance laboratory. I am, Sir, yours faithfully, r, yours faithfully, H. S. GETTINGS. L.R.C.P. & S. Edin., H. S. GETTINGS, L.R.C.P. & S. Edin., H. S. GETTINGS. L.R.C.P. & S. Edin., L.F.P.S. Glasg D P.H. Birm ,- H. S. GETTINGS, L.R.C.P. & S. Edin., L. F. P. S. Glasg D P.H. Birm , . S. GETTINGS. L.R.C.P. & S. Edin., L.F.P.S. Glasg., D P.H. Birm.,- . S. GETTINGS, L.R.C.P. & S. Edin., L. F. P. S. Glasg., D P.H. Birm., GETTINGS. Edin., L.F.P.S. Glasg , D P.H. Birm.,- GETTINGS, Edin., L. F. P. S. Glasg , D P.H. Birm., Glasg., Birm., Assistant Medical Officer. Portsmouth Municipal Tuberculin Dispensary. Glasg., Birm., Assistant Medical Officer. Portsmouth Municipal Tuberculin Dispensary. Glasg., Birm., Assistant Medical Officer. Portsmouth Municipal Tuberculin Dispensary. Glasg., Birm., Assistant Medical Officer. Portsmouth Municipal Tuberculin Dispensary. Dec. 1st, 1911. To the Editor of THE LANCET. SIR,—In the courteous review of our work "Diseases of the Lungs," which appeared in THE LANCET of Nov. 25th, the important question is raised as to whether tuberculin should be prescribed in fractions of a cubic centimetre or milligramme. Before advocating the prescription and dis- pensing of the remedy in cubic centimetres we went carefully into the matter with Messrs. Squire and came to the conclu- sion that for accuracy and facility of dispensing the volu- metric measurements were the best. away escapes. The maximum intensity of the sound is taken as repre- senting the diastolic pressure, but it is not always easy to determine this point, and different observers when examining the same patient frequently give different results, and, more- over, the sound not seldom continues equally loud over considerable differences in pressure We beg to submit the following memorandum which we have received from Mr. P. W. Squire on the subject. pressure. I take the disappearance of the sound as the standand because the event can be certainly and readily determined, and thus has value for comparative observations on the same case. The reading so obtained is perhaps lower than the true diastolic pressure, but the difference would appear to be small. " Miss Allen and Mr. Engle, in the study of the blood pressures of 35 patients in the Johns Hopkins Hospital, found that the readings by the Korotkoff method never differed more than 2-5 mm. from control determinations made at the same time with the Erlanger apparatus, though they took the minimal pressure at the point at which the second sound absolutely disappeared." ’ rs a t u y, R. DOUGLAS POWELL. R. DOUGLAS POWELL. P. HORTON-S:mTH HARTLEY. P. HORTON-SMITH HARTLEY. P. HORTON-S:mTH HARTLEY. P. HORTON-SMITH HARTLEY. London, Dec. 4th, 1911. To the Editor of THE LANCET. Sni,--In April, 1906, I described in THE LANCET2 a double- loop suture which I have reason to believe has proved usefu). Recently I have experimented with another form of suture which fulfils similar conditions and possibly has some advantages. The double-loop suture bad been used by several surgeons before my description of it, and this may be the case with regard to the method I wish to describe. ight. P. W. SQUIRE. 1 Paris Médical, 29, vol. iv., 1911. THE LANCET, Oct. 14th, 1911, p. 1084. 1 Hirschfelder : Diseases of the Heart and Aorta. 2 THE LANCET, April 7th, 1906, p. 967. A SPHYGMOGRAPHIC METHOD FOR THE ESTIMATION OF SYSTOLIC AND DIASTOLIC BLOOD PRESSURE. am, Sir, yours faithfully, EDGAR F. CYRIAX, M.D. Edin. Craven-hill, W., Dec. 2nd, 1911. EDGAR F. CYRIAX, M.D. Edin. To the Editor of THE LANCET. SIR,-Mr. S. G. Longworth correctly remarks that in current descriptions of the Korotkoff method for the determination of the diastolic blood pressure this is read off " at the point at which a distinct change in the intensity of the sound occurs." This alteration in intensity, pitch, and duration of the sound may be phonetically represented by swish to dup ; the sound d2cp gradually increases to a maximum and then dies away as the air escapes THE VALUE OF THE ALBUMIN RE- ACTION IN TUBERCULOUS SPUTUM. To the Editor of THE LANCET. regard The advantage of the new method is that, having a single loop, it is more easily removed, and also that it draws the deep parts together with more lateral strain, thus approxi- mating to the action of a buried suture. SIR,—The recent announcement of Lesieur and Prirey 1 as to the presence of albumin in tuberculous sputum seemed to afford a valuable and simple aid to the diagnosis of phthisis. Being temporarily in charge of the Birmingham Municipal Sanatorium at Yardley-road at the time of the article in THE LANCET 2 took advantage of the opportunity to test the value of the reaction. The author’s technique was adhered to throughout, and in the final stages nitric acid was used in preference to the heat method as being more delicate and accurate. mating A curved needle armed with silkworm gut is passed from without at some distance from the edge of the incision and through a layer of deep tissues on the opposite side; it is then repassed through the skin somewhat in front of the first puncture. The same process is repeated from the other side, the loops being interlocked. They may be tied inde- pendently, or one end may be left long and tied to the corresponding long end of the other loop. I was able to examine the sputum of 55 patients, of whom 39 had tubercle bacilli present and so were definitely tuber- culous. Of these 39 albumin was present in 21, or 53 8 per cent., and absent in 18, or 46-2 2 per cent., so that the test appears to have no diagnostic value. No very extended conclusions can be drawn from such a small number of cases, but there appears to be a greater likelihood of the reaction being given if tubercle bacilli are present in the sputum, for of the remaining 16 cases which did not present tubercle bacilli, yet which otherwise showed all the signs of phthisis, and were, I think, undoubtedly tuberculous, only 4 gave the albumin reaction and 12 were negative. corresponding long loop. These sutures act as "tension sutures," and a continuous horsehair is all that is required to complete the union. In removing the sutures it is best to cut both sides, otherwise the interlocking makes removal more difficult. ing I am, Sir, yours faithfully , yours faithfully J. B. 1 Paris Médical, 29, vol. iv., 1911. [ENCLOSURE.] The writer of the review in THE LANCET of Nov. 25th, 1911, is evidentlv under the impression that he would be obtaining the same amount of solid substance whether he ordered a fraction of 1 gramme or of 1 cubic centimetre of tuberculin, but, of course, this is not so, because 1 cubic centimetre of the bacillary emulsion, for example, weighs 1’114 grammes, and this difference will have to be allowed for in each individual dose of dilution if it were ordered by weight. This introduces an unnecessary factor and an altogether needless amount of trouble. So far from simplifying matters prescribing "by weight" would be introducing complications. When prescribing tuberculin by measure (in cubic centimetres) it diminishes any possibility of a misunderstand- ing as to whether the figures relate to the liquid itself or to the solid substance it contains. It is, for instance, a simple matter to recollect that 1 cubic centimetre of emulsion represents 5 milligrammes of solid matter, but not so easy to recollect that 1’114 grammes of the liquid is equivalent to the 5 milligrammes. The fact that the emulsion is standardised to 1 cubic centimetre indicates that it is intended to be prescribed "by measure and not "by weight." It is also much more convenient to work a fluid "by measure" than "by weight." disappeared. I am, Sir, yours faithfully, , , yours faithfully, Southall, W., Dec. 4th, 1911. J. DAVENPORT WINDLE. J. DAVENPORT WINDLE. THE VALUE OF THE ALBUMIN RE- ACTION IN TUBERCULOUS SPUTUM. PIKE, M. R. C. S. Eng., J. B. PIKE, M. R. C. S. Eng., , , y a t u y J. B. PIKE, M. R. C. S. Eng., Loughborough. Honorary Surgeon to the Loughborough Hospital. J. B. PIKE, M. R. C. S. Eng., Honorary Surgeon to the Loughborough Hospital. y J. B. PIKE, M. R. C. S. Eng., Loughborough. Honorary Surgeon to the Loughborough Hospital J. B. PIKE, M. R. C. S. Eng., Honorary Surgeon to the Loughborough Hospital J. B. PIKE, M. R. C. S. Eng., Loughborough. Honorary Surgeon to the Loughborough Hospital. J. B. PIKE, M. R. C. S. Eng., Honorary Surgeon to the Loughborough Hospital. 1 Hirschfelder : Diseases of the Heart and Aorta. 2 THE LANCET, April 7th, 1906, p. 967. negative. I am indebted to my friend, Dr. W. H. Wynn, the visiting physician, not only for permission to publish this, but also for MUNICIPAL HONOURS.—Mr. R. J. Ryle, M.D. Oxon., .1f.ri.C.S.Eng., J.P., was on Nov. 22nd elected a member of the Brighton town conncil by a majority of 190 votes over his oppoment.
https://openalex.org/W4246815496	http://telerehab.pitt.edu/ojs/index.php/Telerehab/article/download/6165/6530	English	null	Editors' Note	International journal of telerehabilitation	2,015	cc-by	565	International Journal of Telerehabilitation • telerehab.pitt.edu International Journal of Telerehabilitation • telerehab.pitt.edu ACKNOWLEDGMENTS We acknowledge the contributions of our authors and the support of new and returning reviewers; and our publishers, Timothy S. Deliyannides, Director, Office of Scholarly Communication and Publishing and Head, Information Technology, University Library System, and Vanessa Gabler, Electronic Publications Associate at the University of Pittsburgh. Respectfully, Ellen R. Cohn, PhD, CCC-SLP IJT Editor Jana Cason, DHS, OTR/L, FAOTA Issue Co-Editor and IJT Associate Editor Respectfully, Ellen R. Cohn, PhD, CCC-SLP IJT Editor Jana Cason, DHS, OTR/L, FAOTA Issue Co-Editor and IJT Associate Editor ABSTRACT The Fall 2014 issue of the International Journal of Telerehabilitation (IJT) contains original research that evaluates the role of teletherapy and online language exercises in the treatment of chronic aphasia; investigates whether improvements are maintained after in-home pulmonary telerehabilitation for patients with chronic obstructive pulmonary disease; and studies the use of in-home telerehabilitation for proximal humerus fractures. Within the context of two case studies, authors evaluated three service delivery models (direct, hybrid, and telepractice) for stuttering intervention. The results of a preliminary qualitative investigation are presented for telepractice in rural schools in Australia. And, schemas are offered for privacy and security analyses for store and forward applications in physical and occupational therapy. The Fall 2014 issue also contains a book review, and announcements from the American Telemedicine Association and the Mid-Atlantic Telehealth Resource Center. The Fall 2014 issue of the International Journal of Telerehabilitation (IJT) contains an interesting mix of research, clinical case studies, practical content about privacy and security, a book review, and announcements. researchers and clinicians from each of the major rehabilitation disciplines, rehabilitation engineers, health information managers, information technologists, and others. We welcome new peer-reviewers and invite guest editors with ideas for special, thematically focused issues. The IJT publication team is agile and can add additional issues as warranted to ensure currency. Please contact Editor Ellen Cohn, PhD (ecohn@pitt.edu) or Associate Editor Jana Cason (jcason@spalding.edu) if you are interested in contributing to a future issue. In this issue, Steele et al. evaluated the role of teletherapy and online language exercises in the treatment of chronic aphasia; Marquis et al. investigated whether improvements are maintained after in-home pulmonary telerehabilitation for patients with chronic obstructive pulmonary disease; and Tousignant et al. studied the use of in-home telerehabilitation for proximal humerus fractures. Within the context of two case studies, Valentine evaluated three service delivery models (direct, hybrid, and telepractice) for stuttering intervention. Lincoln, et al. presented the results of a preliminary qualitative investigation for telepractice in rural schools in Australia. And, schemas developed by Peterson and Watzlaf are offered for privacy and security analyses for store and forward applications in physical and occupational therapy. The Fall 2014 issue also contains a book review, and announcements from the American Telemedicine Association and the Mid-Atlantic Telehealth Resource Center. CALL FOR SUBMISSIONS The next volume of the International Journal of Telerehabilitation will be published in late spring, 2015. We cordially invite your submissions by March 1, 2015 and accept original research, case studies, viewpoints, technology reviews, book reviews, and country reports that detail the current status of telerehabilitation. Our peer reviewers constitute a multi-disciplinary group, and include International Journal of Telerehabilitation • Vol. 6, No. 2 Fall 2014 • (10.5195/ijt.2014.6165) International Journal of Telerehabilitation • telerehab.pitt.edu International Journal of Telerehabilitation • Vol. 6, No. 2 Fall 2014 • (10.5195/ijt.2014.6156) 2
https://openalex.org/W4388498514	https://link.springer.com/content/pdf/10.1007/s00428-023-03688-2.pdf	English	null	mTOR eosinophilic renal cell carcinoma: a distinctive tumor characterized by mTOR mutation, loss of chromosome 1, cathepsin-K expression, and response to target therapy	Virchows Archiv	2,023	cc-by	10,056	Virchows Archiv (2023) 483:821–833 https://doi.org/10.1007/s00428-023-03688-2 Virchows Archiv (2023) 483:821–833 https://doi.org/10.1007/s00428-023-03688-2 ORIGINAL ARTICLE ORIGINAL ARTICLE Abstract In the spectrum of oncocytic renal neoplasms, a subset of tumors with high-grade-appearing histologic features harboring pathogenic mutations in mammalian target of rapamycin (mTOR) and hitherto clinical indolent behavior has been described. Three cases (2F,1 M) with histologically documented metastases (lymph node, skull, and liver) were retrieved and extensively investigated by immunohistochemistry, FISH, and next-generation sequencing. Tumors were composed of eosinophilic cells with prominent nucleoli (G3 by ISUP/WHO) arranged in solid to nested architecture. Additionally, there were larger cells with perinuclear cytoplasmic shrinkage and sparse basophilic Nissl-like granules, superficially resembling the so-called spider cells of cardiac rhabdomyomas. The renal tumors, including the skull and liver metastases, showed immunoexpres- sion PAX8, CK8-18, and cathepsin-K, and negativity for vimentin. NGS identified mTOR genetic alterations in the three cases, including the skull and liver metastases. One patient was then treated with Everolimus (mTOR inhibitors) with clini- cal response (metastatic tumor shrinkage). We present a distinct renal tumor characterized by high-grade eosinophilic cells, cathepsin-K immunohistochemical expression, and harboring mTOR gene mutations demonstrating a malignant potential and showing responsiveness to mTOR inhibitors. Keywords Cathepsin-K · Eosinophilic-RCC · mTOR · mTOR inhibitors · Next-generation sequencing · Oncocytic · RCC Keywords Cathepsin-K · Eosinophilic-RCC · mTOR · mTOR inhibitors · Next-generation sequencing · Oncocytic · RCC Anna Caliò and Stefano Marletta contributed equally to this work. 8 Division of Anatomic Pathology and Histology, Foundation “A. Gemelli” University Hospital, Università Cattolica del Sacro Cuore, Rome, Italy * Guido Martignoni guido.martignoni@univr.it * Guido Martignoni guido.martignoni@univr.it 1 Department of Diagnostic and Public Health, Section of Pathology, University of Verona, Largo L. Scuro 10, 37134 Verona, Italy 9 Histopathology and Molecular Diagnostics, Azienda Ospedaliero Universitaria Careggi, Florence, Italy 10 Division of Anatomic Pathology and Histology, Ospedale Clinicizzato “SS. Annunziata” Università Di Chieti, Chieti, Italy 2 Department of Pathology, Pederzoli Hospital, Peschiera del Garda, Italy 3 Department of Pathology, IRCCS Sacro Cuore Don Calabria Hospital, Negrar, Italy 11 Department of Pathology, San Raffaele Hospital, Milan, Italy 12 Department of Pathology and Laboratory Medicine, University of Tennessee Health Science, Memphis, TN, USA 4 Division of Medical Oncology, A.O.U. Consorziale Policlinico Di Bari, Bari, Italy 13 Department of Urology, USC Keck School of Medicine, Los Angeles, CA, USA 5 Department of Translational Medicine, University of Ferrara, Ferrara, Italy 14 Interdisciplinary Department of Medicine, University of Bari “A. mTOR eosinophilic renal cell carcinoma: a distinctive tumor characterized by mTOR mutation, loss of chromosome 1, cathepsin‑K expression, and response to target therapy Anna Caliò1 · Stefano Marletta1,2 · Giulio Settanni3 · Mimma Rizzo4 · Stefano Gobbo5 · Serena Pedron1 · Lavinia Stefanizzi2 · Enrico Munari6 · Matteo Brunelli1 · Lisa Marcolini2 · Anna Pesci3 · Stefano Fratoni7 · Francesco Pierconti8 · Maria Rosaria Raspollini9 · Antonio Marchetti10 · Claudio Doglioni11 · Mahul B. Amin12,13 · Camillo Porta14 · Guido Martignoni1,2 Anna Caliò1 · Stefano Marletta1,2 · Giulio Settanni3 · Mimma Rizzo4 · Stefano Gobbo5 · Serena Pedron1 · Lavinia Stefanizzi2 · Enrico Munari6 · Matteo Brunelli1 · Lisa Marcolini2 · Anna Pesci3 · Stefano Fratoni7 · Francesco Pierconti8 · Maria Rosaria Raspollini9 · Antonio Marchetti10 · Claudio Doglioni11 · Mahul B. Amin12,13 · Camillo Porta14 · Guido Martignoni1,2 Received: 19 September 2023 / Revised: 17 October 2023 / Accepted: 24 October 2023 / Published online: 8 November 2023 © The Author(s) 2023 Received: 19 September 2023 / Revised: 17 October 2023 / Accepted: 24 October 2023 / Published online: 8 November 2023 © The Author(s) 2023 Anna Caliò and Stefano Marletta contributed equally to this work. * Guido Martignoni guido.martignoni@univr.it Abstract Moro, Bari, Italy 6 Department of Molecular and Translational Medicine, University of Brescia, Brescia, Italy 7 Division of Anatomic Pathology, S. Eugenio Hospital, Rome, Italy :(0123 1 3456789) 3 Virchows Archiv (2023) 483:821–833 822 (Prog. 4136CESC) and were in accordance with the ethical standards of the institutional and/or national research com- mittee and with the declaration of Helsinki. All patients gave their written informed consent to diagnostic proce- dures and treatment according to institutional rules for everyday clinical practice and experimental evaluations on archival tissue. All slides (28 slides for case 1, 14 slides for case 2, 9 slides for case 3) were reviewed by an expe- rienced pathologist (GM). Samples of both primary tumor and metastases were available. Methods Sections were cut from all FFPE tissue blocks of primary and metastatic samples and manually microdissected to isolate a high percentage of neoplastic cells (> 50%). DNA was isolated using the GeneRead DNA FFPE kit (Qiagen, Hilden, Germany, http://www.qiagen.com Cat. n. 180,134). DNA quality and amount were assessed using NanoDrop and Qubit instruments (Thermo-Fisher Scien- tific) following the manufacturer’s instructions. Immunohistochemistry Sections from tissue blocks of primary and metastatic samples were immunohistochemically stained with the antibodies listed in Supplementary Table 1. All samples were processed using a sensitive Bond Polymer Refine detection system in an automated Bond immunohisto- chemistry instrument (Leica-Biosystems, Germany). The appropriate positive and negative controls were concur- rently carried out. Labeling for each marker was recorded as the percentage of positive cells. P70S6 Kinase and ph4E-BP1 were performed in all samples using an auto- mated Ventana Discovery system (Roche). Furthermore, increasing data regarding RCCs harbor- ing mammalian target of rapamycin (mTOR) and tuber- ous sclerosis complex (TSC) mutations have been pub- lished. In detail, it has been known that alterations of the mTOR pathway are implicated in the pathogenesis of renal tumors arising in patients affected by the inherited tuber- ous sclerosis syndrome such as renal angiomyolipoma and related lesions [9, 10] and TSC-RCC [11, 12]. On the other hand, mutations of TSC1/TSC2 and mTOR genes have also been found in some sporadic renal cell neoplasms includ- ing RCC with leiomyomatous stroma [13], ESC-RCC [14–16], chromophobe RCC [17], epithelioid angiomy- olipoma/pure epithelioid PEComa [10], low-grade onco- cytic tumor (LOT) [18, 19], and eosinophilic vacuolated tumor (EVT), among others [20, 21]. Although an indo- lent clinical course has been accustomed to most of such tumors, recently, Tjota et al. reported the first case of an eosinophilic tumor harboring mTOR gene mutation with liver metastasis [22]. In this study, we describe the clini- cal, morphological, immunohistochemical, and molecular characteristics of three additional cases of eosinophilic RCC harboring mTOR gene mutations with histologi- cally documented metastases and, in one case, the clinical response to targeted therapy. Introduction During the last decade, our understanding of renal cell carcinoma (RCC) has vastly improved due to careful mor- phological evaluation of cases with correlation of data from high-throughput molecular profiling. This is particu- larly notable in oncocytic tumors where analysis of cases that were not easily classifiable as typical oncocytoma and chromophobe RCC has led to the establishment of additional clinicopathologically RCCs or a broadening of the spectrum within previously known RCC subtypes [1] including eosinophilic solid and cystic (ESC)-RCC [2, 3], succinate dehydrogenase (SDH)-deficient RCC [4], low- grade fumarate hydratase (FH)-deficient RCC [5], MiTF family translocation RCC [6], and TFEB-amplified RCC [7, 8]. Fluorescence in situ hybridization (FISH) FISH was carried out on primary and metastatic samples using dual-color break-apart TFE3 and TFEB probes (Cytotest, USA) and 1p36/1q25 probe, spectrum-orange/ spectrum-green (Vysis) as previously described [23]. Scoring was performed by two experienced pathologists (AC and MB). At least 100 neoplastic non-overlapping nuclei were included in the scoring. To avoid false posi- tive results due to nuclear truncation, cells with a single fluorescent signal were not evaluated. Patients and samples From our archives of unclassified oncocytic renal tumors, we identified four previously unreported high-grade eosin- ophilic renal cell tumors from three different patients. All of them were consult cases. All procedures performed in our study involving human participants received approval 1 3 Virchows Archiv (2023) 483:821–833 823 Variant calling Data from the PGM sequencing were initially processed using the Ion Torrent platform-specific software (Torrent Suite AD 5.6.4) to generate sequence reads, alignment of the reads on the reference genome Hg19, trim adapter sequences, filter, and remove poor signal-profile reads. The variant calling from the sequencing data was generated using the Variant Caller plugin. To provide reliable somatic variant analysis we consid- ered suitable only samples with more than 400,000 reads, an average coverage > × 500, and a coverage uniformity > 95%. We applied the following filters to the Variant Caller plugin: minimum allele frequency value of 2% and minimum phred quality score of 30. Variant annotation and copy number variation analysis were performed using the Ion Reporter 5.12 software (Thermo-Fisher Scientific). i Variant annotations were also assessed using the Ensembl Variant Effect Predictor pipeline of the Wellcome Trust Sanger Institute [24] as a second database check. Filtered variants were visually examined using the Integrative Genomic Viewer tool to taste their level of quality and to confirm the variant presence on both “ + ” and “ − ” strands. The clinical relevance (pathogenicity) of the annotated vari- ants was assessed using the COSMIC database (Wellcome Sanger Institute), OncoKB database [25], ClinVar (NCBI), and LOVD (IARC). Library preparation and deep amplicon sequencing Table 1 Clinical and pathological features of the mTOR-mutated eosinophilic renal cell carcinomas of the present series * Same patient (2a and 2b: renal tumors, 2c: skull metastasis) # Same patient (3a: renal tumor, 3b: liver metastasis) F, female; M, male; L, left; R, right; N.A., not available; RCC, renal cell carcinoma; MD, median; SD, standard deviation Case Age Gender Size/laterality Stage TNM Surgery Original diagnosis Follow-up 1 21 F 9.5 cm/N.A pT2aN1Mx Radical nephrectomy Unclassified RCC 30 months alive 2a* 58 F 6.5 cm/L pT1bNxMx Radical nephrectomy Oncocytoma 108 months, alive with skull metastasis 2b* 58 F 3.1 cm/L pT1aNxMx Radical nephrectomy Oncocytoma 108 months, alive with skull metastasis 2c* 69 F Skull metastasis pT1bNxM1 Metastasectomy - 108 months, alive with skull metastasis 3a# 49 M 3.6 cm/R pT1aNxMx Partial nephrectomy Chromophobe RCC 64 months, alive with liver metastases 3b# 52 M Liver metastases pT1aNxM1 Metastasectomy - 64 months, alive with liver metastases (MD 35, SD ± 19.29) (MD 6.5 cm, SD ± 4.17 cm) We performed deep sequencing of the whole coding region and intron–exon junctions of 17 kidney-cancer-related genes with a custom panel created using the Ampliseq Designer pipeline (Thermo-Fisher Scientific) as previously described [3]. The genes included: TSC1, TSC2, MTOR, AKT1, PIK3CA, PTEN, SDHB, FH, VHL, SETD2, BAP1, PBRM1, MET, FLCN, SMARCA4, SMARCB1, and TCEB1. Clinical and pathological features The clinical and pathological features are summarized in Table 1. Two patients were female and one male. The patients’ ages at diagnosis were 21, 58, and 49 years respec- tively. None of the three patients showed clinical stigmata of tuberous sclerosis. Two of them underwent radical nephrectomy and one partial nephrectomy; one patient had two tumors simultaneously affecting the left kidney. The tumors ranged in size from 3.1 to 9.5 cm, and were all solid and brownish in color. The original diagnosis made at the Virchows Archiv (2023) 483:821–833 824 Fig. 1 mTOR-mutated eosino- philic RCC. Low magnifica- tion demonstrates solid-nested growth (a). Large cells with perinuclear cytoplasmic shrink- age reminiscent of the so-called spider cells of cardiac rhabdo- myomas were easily found in case 1 (b), or scattered in case 2 (c). Neoplastic vascular inva- sion was encountered in case 3 (d) (original magnification × 50 a, × 100 b and c, and × 200 d) Fig. 1 mTOR-mutated eosino- philic RCC. Low magnifica- tion demonstrates solid-nested growth (a). Large cells with perinuclear cytoplasmic shrink- age reminiscent of the so-called spider cells of cardiac rhabdo- myomas were easily found in case 1 (b), or scattered in case 2 (c). Neoplastic vascular inva- sion was encountered in case 3 (d) (original magnification × 50 a, × 100 b and c, and × 200 d) Fig. 1 mTOR-mutated eosino- philic RCC. Low magnifica- tion demonstrates solid-nested growth (a). Large cells with perinuclear cytoplasmic shrink- age reminiscent of the so-called spider cells of cardiac rhabdo- myomas were easily found in case 1 (b), or scattered in case 2 (c). Neoplastic vascular inva- sion was encountered in case 3 (d) (original magnification × 50 a, × 100 b and c, and × 200 d) referral institution was unclassified RCC in case 1, oncocy- toma in case 2, and chromophobe RCC in case 3. deletion (Fig. 3a); neither TFE3 nor TFEB rearrangements were identified. p At light microscopy (Figs. 1 and 2), all the tumors were unencapsulated and composed of granular eosinophilic cells with cytoplasmatic vacuolization and nuclear atypia with prominent nucleoli (G3 by ISUP/WHO), mainly showing a solid/nested growth pattern. Additionally, in variable amount, there were larger cells with perinuclear cytoplasmic shrinkage and sparse basophilic Nissl-like granules, resem- bling the so-called spider cells of cardiac rhabdomyomas. Neoplastic cells were embedded within a fibrous and some- times densely hyalinized stroma. Clinical and pathological features Several thick-walled ves- sels were dispersed within the lesion along with peripherally entrapped normal renal tubules. Neither coagulative granular necrosis, mitoses, nor foamy macrophages were observed. In case 3, a neoplastic peritumoral vascular embolus was identified. i One patient (case 1) had a synchronous small 9 mm nodal metastasis detected in one of seven regional lymph nodes and, to date, is alive without evidence of disease after 30 months. In another patient (case 2), a lesion was removed from the skull eleven years after surgical excision of the primary renal neoplasm. The skull lesion revealed overlap- ping histological and immunohistochemical findings with the kidney neoplasms and was therefore considered a metas- tasis. Finally, 3 years after surgery, two liver nodules were identified by imaging in the remaining patient (case 3); the largest of 3 cm as greater diameter was removed, whereas the other one was investigated by biopsy. Again, both liver lesions revealed overlapping histological and immunohisto- chemical findings with the renal tumor and were therefore considered metastases. i All the cases tested (Supplementary Table 2), expressed PAX8 along with immunolabelling for cathepsin-K (clone 3F9), whereas vimentin, and melanogenesis mark- ers (HMB45 and Melan-A) were negative. Among the cytokeratins, strong and diffuse cytokeratin 8–18 expression was observed in all the neoplasms while in none of them significant staining of cytokeratin AE1/AE3 and cytokeratin 7 was found. All the cases were focally positive for CD117, expressed P70S6 Kinase and ph4E-BP1, and retained SDHB and FH. At FISH analysis all the cases showed 1p36/1q25 Next‑generation sequencing results Genetic alterations in the TSC or mTOR pathway were found in all the samples tested with next-generation sequencing (Table 2). In case 1, a deletion in exon 30 of the mTOR gene (p.Tyr1450_Trp1456) was identified. This variant is reported both in COSMIC database (ref. COSM6972065) and OncoKB and occurs in the MTOR focal adhesion kinase targeting domain (FAT) domain, a key structural domain 1 3 Virchows Archiv (2023) 483:821–833 825 for the correct conformation of the catalytic pouch of the MTOR protein [26]. While its biological significance is still unknown, it has been identified as a statistically significant hotspot and is likely to be oncogenic. Moreover, it is also predicted to be pathogenic by the most common in-silico structural predictors (SIFT, POLYPHEN). Fig. 2 CT scan of case 2 revealed a skull mass (a) histo- logically characterized by large eosinophilic cells with round and enlarged nuclei (b). The neoplastic cells showed stain- ing for PAX8 (c), cytokeratin 8–18 (d), but not for vimentin (e). The liver metastases of case 3 were histologically documented by biopsy (f) and were immunohistochemically positive for cathepsin-K (clone 3F9) (g) (original magnifica- tion × 50 f and g, × 100 b, d, and e, and × 200 c) for the correct conformation of the catalytic pouch of the MTOR protein [26]. While its biological significance is still unknown, it has been identified as a statistically significant hotspot and is likely to be oncogenic. Moreover, it is also predicted to be pathogenic by the most common in-silico structural predictors (SIFT, POLYPHEN). Fig. 2 CT scan of case 2 revealed a skull mass (a) histo- logically characterized by large eosinophilic cells with round and enlarged nuclei (b). The neoplastic cells showed stain- ing for PAX8 (c), cytokeratin 8–18 (d), but not for vimentin (e). The liver metastases of case 3 were histologically documented by biopsy (f) and were immunohistochemically positive for cathepsin-K (clone 3F9) (g) (original magnifica- tion × 50 f and g, × 100 b, d, and e, and × 200 c) Fig. 2 CT scan of case 2 revealed a skull mass (a) histo- logically characterized by large eosinophilic cells with round and enlarged nuclei (b). The neoplastic cells showed stain- ing for PAX8 (c), cytokeratin 8–18 (d), but not for vimentin (e). Fig. 2 CT scan of case 2 revealed a skull mass (a) histo- logically characterized by large eosinophilic cells with round and enlarged nuclei (b). The neoplastic cells showed stain- ing for PAX8 (c), cytokeratin 8–18 (d), but not for vimentin (e). The liver metastases of case 3 were histologically documented by biopsy (f) and were immunohistochemically positive for cathepsin-K (clone 3F9) (g) (original magnifica- tion × 50 f and g, × 100 b, d, and e, and × 200 c) Next‑generation sequencing results The liver metastases of case 3 were histologically documented by biopsy (f) and were immunohistochemically positive for cathepsin-K (clone 3F9) (g) (original magnifica- tion × 50 f and g, × 100 b, d, and e, and × 200 c) Fig. 2 CT scan of case 2 revealed a skull mass (a) histo- logically characterized by large hotspot and is likely to be oncogenic. Moreover, it is also predicted to be pathogenic by the most common in-silico structural predictors (SIFT, POLYPHEN). for the correct conformation of the catalytic pouch of the MTOR protein [26]. While its biological significance is still unknown, it has been identified as a statistically significant hotspot and is likely to be oncogenic. Moreover, it is also predicted to be pathogenic by the most common in-silico structural predictors (SIFT, POLYPHEN). Fig. 3 Loss of chromosome 1 identified by FISH in tumors with mTOR gene mutation (a). Pathogenic mTOR alterations in both primary renal tumors and the skull metastasis involving the exon 53 (p.Leu2427Arg) (b) Fig. 3 Loss of chromosome 1 identified by FISH in tumors with mTOR gene mutation (a). Pathogenic mTOR alterations in both primary renal tumors and the skull metastasis involving the exon 53 (p.Leu2427Arg) (b) 1 3 Virchows Archiv (2023) 483:821–833 826 Table 2 Molecular features of the mTOR mutated eosinophilic renal cell carcinomas of the present series * Same patient (2a and 2b: renal tumors, 2c: skull metastasis) # Same patient (3a: renal tumor, 3b: liver metastasis) Case Mutated gene Exon HGVS mutation Protein mutation Coverage Quality Frequency Chromosomal position Class 1 mTOR 30 c.4348_4368del p.Tyr1450_Trp- 1456del 1200 6800 50% chr1:11,217,309 Likely pathogenetic 2a* mTOR 53 c.7280 T > G p.Leu2427Arg 1200 11000 40% chr1:11,174,395 Pathogenetic 2b* mTOR 53 c.7280 T > G p.Leu2427Arg 1200 11000 57% chr1:11,174,395 Pathogenetic 2c* mTOR 53 c.7280 T > G p.Leu2427Arg 1300 8000 55% chr1:11,174,395 Pathogenetic 3a# mTOR 53 c.7280 T > A p.Leu2427Gln 800 7000 66% chr1:11,174,395 Pathogenetic 3b# mTOR 53 c.7280 T > A p.Leu2427Gln 1200 5700 45% chr1:11,174,396 Pathogenetic # Same patient (3a: renal tumor, 3b: liver metastasis) shortly after first-line systemic treatment was permanently discontinued. Case 2 showed pathogenic mTOR alterations in both pri- mary renal tumors and the skull metastasis involving the exon 53 (p.Leu2427Arg—COSM2119114) (Fig. 3b). Next‑generation sequencing results Based on the molecularly documented mTOR mutation (L2427Q, exon 53), the choice for second-line therapy was towards a combination of Lenvatinib (18 mg daily) and Everolimus (5 mg daily) in an off-label regimen. Within 4 months several treatment-related adverse events were observed including G2 hypertension, G2 fatigue, G2 diar- rhea, and G2 mucositis. Hence, the dose of Lenvatinib was initially reduced to 14 mg daily and then to 10 mg daily. The first follow-up CT scan, after two months, showed reduction in size and vascularization of both all the target liver lesions and the retrocaval lymphadenopathy. Four months later, another CT scan revealed a further dimensional reduction of the liver lesions. Finally, 8 months after the beginning of the second-line therapy, the last CT scan available docu- mented a numerical decrease in the liver lesions (with only two of them detectable to date) and an additional shrinkage of the retrocaval lymphadenopathy (< 10 mm in short axis) (Fig. 4). Currently, the patient is still on Lenvatinib (10 mg orally once a day) and Everolimus (5 mg orally once a day) treatment with fair tolerance. Finally, case 3 showed pathogenic mTOR alterations in both the primary renal tumor and the liver metastases involv- ing the exon 53 (p.Leu2427Gln—COSM1185313). Both case 2 and case 3 variants involve the same ami- noacidic residue of the MTOR Kinase Domain and are well known to be likely oncogenic [27]. No other pathogenetic mutation nor variant of unknown significance was identified in any of the primary tumor nor metastatic lesion. Furthermore, neither mTOR alterations nor TSC1/TSC2 gene mutation were found in the normal renal parenchyma. Response to therapy The patient of case 3 underwent radioembolization of the liver lesions. A new CT scan performed three months later documented the appearance of novel liver lesions (one lesion in the seventh hepatic segment 33 mm in long axis; 5 lesions in the fifth and sixth hepatic segments ranging from 17 to 47 mm in long axis; 1 lesion in the fourth hepatic segment 32 mm in long axis) and a pathological retrocaval lymphade- nopathy (31 mm in short axis). Therefore, the patient started first-line systemic therapy with Pembrolizumab (200 mg intravenously, every 21 days) and Axitinib (5 mg orally twice a day). The best response achieved to this treatment was stable disease according RECIST 1.1, as shown by CT scans performed after 4, 8, and 12 months respectively. In the last examination, an osteolytic spot (22 mm in long axis) was identified at L2 so that the patient underwent external beam radiotherapy at this site (total dose: 20 Gy in 5 frac- tions). Fifteen months since the beginning of the systemic therapy, a new CT scan revealed further disease progression, with an increase in the size of all liver lesions, along with a new lesion in the eighth hepatic segment (6 mm in long axis) and new osteosclerotic spots in L5 and pelvic bone. Thus, Discussion In this study, we reported four sporadic high-grade eosino- philic RCCs occurring in three patients with histologically documented metastases, characterized by the presence of “spider cells,” the immunohistochemical expression of cath- epsin-K (clone 3F9), and mutations of mTOR gene.f In the last years, the differential diagnosis of eosinophilic tumors has become even more challenging for pathologists due to growing wide spectrum of tumors with oncocytic cells, usually characterized by non-aggressive behavior (Table 3) [1]. In this morphological scenario, cathepsin-K (clone 3F9) is a useful tool for differentiating renal onco- cytoma, chromophobe RCC, and LOT which are negative 1 3 3 827 Virchows Archiv (2023) 483:821–833 g. 4 Comparison between the baseline CT scan performed before e start of the first-line therapy with Pembrolizumab and Axitinib anuary 2021), at the best response to the first-line therapy (May 021), at the progression to the first-line therapy (April 2022), and ter 9 months of the second-line therapy with Lenvatinib and verolimus. Little to any tumor shrinkage was observed with Pem- brolizumab and Axitinib. Conversely, following Lenvatinib and Everolimus, just a few hepatic lesions were detected, showing a decrease in size and vascularization. Disease stability for the osteo- lytic lesion in L2 was also recorded. All the images were taken in the arterial phase brolizumab and Axitinib. Conversely, following Lenvatinib and Everolimus, just a few hepatic lesions were detected, showing a decrease in size and vascularization. Disease stability for the osteo- lytic lesion in L2 was also recorded. All the images were taken in the arterial phase Fig. 4 Comparison between the baseline CT scan performed before the start of the first-line therapy with Pembrolizumab and Axitinib (January 2021), at the best response to the first-line therapy (May 2021), at the progression to the first-line therapy (April 2022), and after 9 months of the second-line therapy with Lenvatinib and Everolimus. Little to any tumor shrinkage was observed with Pem- pathway, namely, the phosphorylated forms of 4EBP1 and S6K. Moreover, loss of chromosome 1 was identified in the four renal neoplasms and the skull and liver metastases, which represents the genomic location of the mTOR gene. The association of loss of chromosome 1 along with activat- ing mTOR mutations has been also reported by Tjota et al. [22]. Being mTORC1 a dimer, it is possible to speculate that a heterodimer of wild-type and mTOR mutant proteins may not confer sufficient mTOR activation [29]. Discussion for cathepsin-K (clone 3F9), from ESC-RCC, EVT, and mTOR-mutated eosinophilic RCC which are positive for this marker [28]. Whether the recognition of ESC-RCC is usually straightforward for uropathologists, EVT and mTOR-mutated eosinophilic RCC may show overlapping features, as highlighted by Tjota et al. [22]. However, EVT has prominent or extreme cytoplasmic vacuolation whereas in mTOR-mutated eosinophilic RCCs, we observed large cells with perinuclear cytoplasmic shrinkage resembling the so-called spider cells of cardiac rhabdomyomas occur- ring, interestingly, in tuberous sclerosis patients. This find- ing is morphologically reliable since similar elements are also focally reported by Tjota et al. as “somewhat rhabdoid appearance” [22]. for cathepsin-K (clone 3F9), from ESC-RCC, EVT, and mTOR-mutated eosinophilic RCC which are positive for this marker [28]. Whether the recognition of ESC-RCC is usually straightforward for uropathologists, EVT and mTOR-mutated eosinophilic RCC may show overlapping features, as highlighted by Tjota et al. [22]. However, EVT has prominent or extreme cytoplasmic vacuolation whereas in mTOR-mutated eosinophilic RCCs, we observed large cells with perinuclear cytoplasmic shrinkage resembling the so-called spider cells of cardiac rhabdomyomas occur- ring, interestingly, in tuberous sclerosis patients. This find- ing is morphologically reliable since similar elements are also focally reported by Tjota et al. as “somewhat rhabdoid appearance” [22]. fi Recently, several renal tumors harboring mTOR gene mutations have been described (Table 4). In our compre- hensive analysis of 57 neoplasms reported in the literature and our series, mTOR gene mutations have been observed in 16 EVTs, 24 LOTs, and 17 RCCs. The percentage of mTOR gene mutations in the molecularly tested cases is 31% of EVT (16 of 51 tumors) and 30.7% of LOT (24 of 78 tumors). Interestingly, the same mTOR gene mutation (L2427) has been detected in 11 of 16 (68.7%) EVTs, in 9 of 24 (37.5%) Activating mTOR gene mutations are the genetic hall- marks of these high-grade eosinophilic RCCs. In accordance with the activation of the mTOR pathway, we observed an overexpression of two proteins downstream of the mTOR 1 3 Virchows Archiv (2023) 483:821–833 828 828 Virchows Archiv (2023) 483:82 1 3 Table 3 Key morphological and immunohistochemical hallmarks of eosinophilic renal neoplasms Tumor Key histological features PAX 8 CAT. p ations: RCC, renal cell carcinoma; CAT. K, cathepsin K; CK, cytokeratin; PV, parvalbumin; AMACR, alpha-methylacyl-CoA racemase; CA9, carbonic anhydrase IX; SDH, succinate genase; ALK, anaplastic lymphoma kinase; N.A., not available; LOT, low-grade oncocytic tumor; ESC, eosinophilic, solid, and cystic LOTs, and in 9 of 17 (53%) mTOR-mutated RCCs. The high prevalence of this hotspot genetic alteration, which falls in the catalytic subunit of the mTOR protein, may confer a selective advantage. Moreover, it has been demonstrated both in vitro and in vivo that cells harboring this mutation are highly sensitive to mTOR inhibitors [27, 30]. ( ) Tumor Key histological features PAX 8 CAT. K PAN-CKs CK 7 CD 117 CK 20 Vimentin GATA 3 PV S100 A1 AMACR CA 9 SDH-B HMB 45 Melan-A ALK Epithelioid angiomy- olipoma Epithelioid some- times pleomorphic cells Neg + Neg Neg Neg Neg Neg N.A Neg Neg Neg Neg Retained + + Neg ALK-rearranged RCC Large cells display- ing various archi- tectural growth patterns, focally mucin deposit + Neg + + Neg Neg + Neg N.A N.A + Neg Retained Neg Neg + mTOR-mutated eosinophilic RCC Cytoplasmic vacu- olation and shrink- age with basophilic Nissl-like granules + + + Neg -/ + Neg Neg Neg Neg -/ + -/ + Neg Retained Neg Neg Neg Table 3 (continued) Tumor Key histological features PAX 8 CAT. K PAN-CKs CK 7 CD 117 CK 20 Vimentin GATA 3 PV S100 A1 AMACR CA 9 SDH-B HMB 45 Melan-A ALK Epithelioid angiomy- olipoma Epithelioid some- times pleomorphic cells Neg + Neg Neg Neg Neg Neg N.A Neg Neg Neg Neg Retained + + Neg ALK-rearranged RCC Large cells display- ing various archi- tectural growth patterns, focally mucin deposit + Neg + + Neg Neg + Neg N.A N.A + Neg Retained Neg Neg + mTOR-mutated eosinophilic RCC Cytoplasmic vacu- olation and shrink- age with basophilic Nissl-like granules “spider cells,” nuclear atypia with prominent nucleoli + + + Neg -/ + Neg Neg Neg Neg -/ + -/ + Neg Retained Neg Neg Neg Abbreviations: RCC, renal cell carcinoma; CAT. K, cathepsin K; CK, cytokeratin; PV, parvalbumin; AMACR, alpha-methylacyl-CoA racemase; CA9, carbonic anhydrase IX; SDH, succinate dehydrogenase; ALK, anaplastic lymphoma kinase; N.A., not available; LOT, low-grade oncocytic tumor; ESC, eosinophilic, solid, and cystic g y Differently from other eosinophilic tumors harboring mTOR mutations, the three patients presented in this study developed metastases. p ations: RCC, renal cell carcinoma; CAT. K, cathepsin K; CK, cytokeratin; PV, parvalbumin; AMACR, alpha-methylacyl-CoA racemase; CA9, carbonic anhydrase IX; SDH, succinate genase; ALK, anaplastic lymphoma kinase; N.A., not available; LOT, low-grade oncocytic tumor; ESC, eosinophilic, solid, and cystic In case 1, the site of metastasis was a small lymph node of the renal sinus (i.e., still a loco-regional disease), which was removed together with the primary kidney lesion; in case 2, the metastasis was larger and occurred in the skull (i.e., a frankly metastatic disease) eleven years later the first diagnosis; finally, in case 3, multiple metastases in the liver were observed after 3 years from nephron-sparing surgery. Because of the amount of tumor available, we were able to test also the liver and skull metastases by immuno- histochemistry and genetic analysis. We found morphologi- cal and immunohistochemical features (PAX8, cytokeratin 8–18, cathepsin-K clone 3F9 positivity, and vimentin nega- tivity) overlapping with those observed in the primary renal tumors. The same mTOR gene mutation was identified in the liver metastasis and the primary renal tumor, as well as in the skull metastasis and the two renal tumors. A possible expla- nation of the latter finding is that one mass represented the intrarenal metastasis of the other tumor. Despite this possi- bility being well-known in other RCCs, it seems unlikely in this case since the skull metastasis occurred 11 years after the radical nephrectomy, suggesting instead a slow progression of the disease. Recently, Tjota et al. reported a similar case with liver metastasis harboring mTOR gene mutation (p.L2427K) and loss of chromosome 1 [22]. To date, the predictive value of mTOR gene mutations in metastatic RCC patients treated with mTOR inhibitors remains controversial, with studies suggest- ing such a predictive role [31], and others not [32]. A multi- center, histology-agnostic, single-arm prospective phase II trial of the mTORC1 inhibitor, Everolimus, in patients with solid tumors mainly harboring TSC1/TSC2 mutations ultimately failed, showing no association between these genomic altera- tions and response to targeted treatment in a broad spectrum of neoplasms, not including RCCs [33]. In our study, the clini- cal response to mTOR inhibitors observed in patient 3 might support the hypothesis that mTOR alterations could predict response to mTOR inhibitors, although our patient received a combination of one mTOR inhibitor, plus a multikinase inhibitor (mainly targeting VEGFRs), making it impossible to dissect the relative contribution of the two different classes of agents to the observed activity and efficacy. Discussion K PAN-CKs CK 7 CD 117 CK 20 Vimentin GATA 3 PV S100 A1 AMACR CA 9 SDH-B HMB 45 Melan-A ALK Chromophobe RCC, eosinophilic Solid (“plant-like”) architecture, prominent cell membranes, raisi- noid nuclei with perinuclear halos + Neg + + + Neg Neg Neg + Neg Neg Neg Retained Neg Neg Neg Renal oncocytoma Solid-nested to tubu- locystic growth, central edematous- scary areas, no perinuclear halos + Neg + Neg + Neg Neg Neg + + Neg Neg Retained Neg Neg Neg Hybrid oncocytic tumor Admixed chro- mophobe and oncocytoma-like features (i.e. solid- nested architecture but with perinu- clear halos) + N.A + ± + Neg Neg N.A ± -/ + Neg Neg Retained Neg Neg Neg Clear cell RCC, eosinophilic Delicate capillary vessels network, usually at least scattered conven- tional clear cell areas + Neg + Neg Neg Neg + Neg Neg + -/ + + Retained Neg Neg Neg Papillary RCC, oncocytic Papillary-tubular architecture (at least focally) + Neg + + Neg Neg + Neg Neg + + Neg Retained Neg Neg Neg LOT Alternating solid- trabecular cellular areas with loose edematous stromal zones and fresh hemorrhage + Neg + + Neg Neg Neg + N.A + Neg Neg Retained Neg Neg Neg ESC-RCC Solid and cystic architecture, sparse foamy mac- rophages + + + Neg Neg + + Neg Neg + Neg Neg Retained Neg Neg Neg SDH-deficient RCC Solid-nested archi- tecture, low-grade cells often with flocculent cyto- plasm inclusions + Neg Neg Neg Neg Neg Neg N.A N.A N.A Neg Neg Absent Neg Neg Neg Virchows Archiv (2023) 483:821–833 829 p ations: RCC, renal cell carcinoma; CAT. K, cathepsin K; CK, cytokeratin; PV, parvalbumin; AMACR, alpha-methylacyl-CoA racemase; CA9, carbonic anhydrase IX; SDH, succinate genase; ALK, anaplastic lymphoma kinase; N.A., not available; LOT, low-grade oncocytic tumor; ESC, eosinophilic, solid, and cystic K, cathepsin K; CK, cytokeratin; PV, parvalbumin; AMACR, alpha-methylacyl-CoA racemase; CA9, carbonic anhydrase genase; ALK, anaplastic lymphoma kinase; N.A., not available; LOT, low-grade oncocytic tumor; ESC, eosinophilic, solid, and cystic fi In conclusion, herein, we present a distinct renal tumor characterized by high-grade eosinophilic cells, cathepsin-K (clone 3F9) immunohistochemical expression, and harboring mTOR gene mutation demonstrating a malignant potential and showing responsiveness to an mTOR inhibitor-containing 1 1 3 Virchows Archiv (2023) 483:821–833 830 Virchows Archiv (2023) 483:821–833 mTOR mutation chr 1 loss Follow-up mTOR, c.5930C > A loss 50 months, NED mTOR, p.Leu2427Arg loss 13 months, NED mTOR, p.Leu2427Arg loss 10 months, NED mTOR, p.Leu2427Gln loss 15 months, NED mTOR, c.7280 T > G * 31 months, NED mTOR, c.7257_7259delinsTGT * 75 months, NED mTOR, c.7280 T > A * 144 months, NED mTOR, c.7280 T > C * 18 months, NED mTOR, c.7280 T > G * 19 months, NED mTOR, c.4343_4363del * 47 months, NED mTOR, c.7280 T > A * 18 months, NED mTOR c.7280 T > A N.A N.A mTOR c.7280 T > G N.A N.A mTOR c.7280 T > G N.A N.A mTOR c.7237_7238delinsCT N.A N.A mTOR c.11C > T + TSC2 c.3352C > T N.A N.A mTOR c.7280 T > A N.A 6 months NED mTOR c.5930C > G no loss 156 months NED mTOR c.6644 C > T N.A 7 months NED mTOR c.7499 T > A N.A N.A mTOR c.6644 C > A N.A N.A mTOR c.4348 T > G N.A 49 died of other disease mTOR c.320_323delinsATTT N.A 49 months NED mTOR c.7280 T > A N.A N.A mTOR c.7498 A > T N.A N.A mTOR p.Leu2427Gln no loss 26 months died mTOR p.Ser2215Tyr no loss 3 months NED mTOR p.Ser2413Leu no loss 53 months NED mTOR p.Lys1452_Glu- 1455del no loss 36 months died mTOR p.Leu2427Gln + TSC2 p.Met286Val N.A 33 months NED mTOR c.7280G > A N.A 2 months NED mTOR c.7500 T > G N.A N.A mTOR c.7280 T > A N.A N.A mTOR c.6644C > T + TSC1 c.2356C > T N.A N.A NF2 mTOR c.4448G > T N.A N.A mTOR c.7498A > T N.A N.A mTOR p.L2427Q N.A 57 months NED mTOR p.L2427Q N.A 53 months NED Table 4 mTOR-mutated eosinophilic renal cell carcinomas Case Reference Gender Age Size/laterality Diagnos 1 He et al., 2018, Far- cas et al., 2021 M 54 2.6 cm/N.A EVT 2 Chen et al., 2019 F 68 4.4 cm/N.A EVT 3 Chen et al., 2019 M 59 3.6 cm/N.A EVT 4 Kapur et al. p ations: RCC, renal cell carcinoma; CAT. K, cathepsin K; CK, cytokeratin; PV, parvalbumin; AMACR, alpha-methylacyl-CoA racemase; CA9, carbonic anhydrase IX; SDH, succinate genase; ALK, anaplastic lymphoma kinase; N.A., not available; LOT, low-grade oncocytic tumor; ESC, eosinophilic, solid, and cystic 2021 M 55 1.8 cm/L EVT 5 Farcas et al. 2021 F 31 3.5 cm/N.A EVT 6 Farcas et al. 2021 M 25 3.8 cm/N.A EVT 7 Farcas et al. 2021 M 72 3.5 cm/N.A EVT 8 Farcas et al. 2021 F 59 4 cm/N.A EVT 9 Farcas et al. 2021 M 15 11.5 cm/N.A EVT 10 Farcas et al. 2021 F 69 4 cm/N.A EVT 11 Farcas et al. 2021 M 42 7 cm/N.A EVT 12 Xia et al. 2022 M 42 3.5 cm/N.A EVT 13 Xia et al. 2022 F 32 2.5 cm/N.A EVT 14 Xia et al. 2022 M 24 6 cm/N.A EVT 15 Xia et al. 2022 M 59 2.5 cm/N.A EVT 16 Xia et al. p ations: RCC, renal cell carcinoma; CAT. K, cathepsin K; CK, cytokeratin; PV, parvalbumin; AMACR, alpha-methylacyl-CoA racemase; CA9, carbonic anhydrase IX; SDH, succinate genase; ALK, anaplastic lymphoma kinase; N.A., not available; LOT, low-grade oncocytic tumor; ESC, eosinophilic, solid, and cystic 2022 M 47 2.5 cm/N.A EVT 1 Tjota et al., 2020 F 66 4.1 cm/L LOT 2 Tjota et al., 2020 M 66 2.5 cm/R LOT 3 Morini et al., 2021 F 57 3.7 cm/L LOT 4 Morini et al., 2021 F 61 3.8 cm/R LOT 5 Morini et al., 2021 F 78 3.7 cm/R LOT 6 Morini et al., 2021 F 83 3.5 cm/R LOT 7 Morini et al., 2021 F 79 8.5 cm/R LOT 8 Morini et al., 2021 F 58 5.5 cm/R LOT 9 Morini et al., 2021 F 76 3.7 cm/R LOT 10 Kapur et al., 2021 F 79 7.8 cm/R LOT 11 Kapur et al., 2021 F 86 6.5 cm/R LOT 12 Kapur et al., 2021 F 71 3.8 cm/R LOT 13 Kapur et al., 2021 F 75 2.4 cm/L LOT 14 Zhang et al., 2022 F 79 1.6 cm/R LOT 15 Mohanty et al., 2021 F 65 2.3 cm/L LOT 16 Williamson et al., 2023 F 61 8 cm/N.A LOT 17 Williamson et al., 2023 F 73 6.9 cm/N.A LOT 18 Williamson et al., 2023, Trpkov 2019 F 63 5.2 cm/N.A LOT 19 Williamson et al., 2023, Trpkov 2019 F 61 3 cm/N.A LOT 20 Williamson et al., 2023 F 41 2 cm/N.A LOT 21 Chen et al., 2023 F 45 2.5 cm/R LOT 22 Chen et al., 2023 M 65 3 cm/R LOT Table 4 mTOR-mutated eosinophilic renal cell carcinomas Case Reference Gender Age Size/laterality Diagnosis mTOR mutation chr 1 loss Follow-up 1 He et al., 2018, Far- cas et al., 2021 M 54 2.6 cm/N.A EVT mTOR, c.5930C > A loss 50 months, NED 2 Chen et al., 2019 F 68 4.4 cm/N.A EVT mTOR, p.Leu2427Arg loss 13 months, NED 3 Chen et al., 2019 M 59 3.6 cm/N.A EVT mTOR, p.Leu2427Arg loss 10 months, NED 4 Kapur et al. 2021 M 55 1.8 cm/L EVT mTOR, p.Leu2427Gln loss 15 months, NED 5 Farcas et al. 2021 F 31 3.5 cm/N.A EVT mTOR, c.7280 T > G * 31 months, NED 6 Farcas et al. 2021 M 25 3.8 cm/N.A EVT mTOR, c.7257_7259delinsTGT * 75 months, NED 7 Farcas et al. 2021 M 72 3.5 cm/N.A EVT mTOR, c.7280 T > A * 144 months, NED 8 Farcas et al. 2021 F 59 4 cm/N.A EVT mTOR, c.7280 T > C * 18 months, NED 9 Farcas et al. p ations: RCC, renal cell carcinoma; CAT. K, cathepsin K; CK, cytokeratin; PV, parvalbumin; AMACR, alpha-methylacyl-CoA racemase; CA9, carbonic anhydrase IX; SDH, succinate genase; ALK, anaplastic lymphoma kinase; N.A., not available; LOT, low-grade oncocytic tumor; ESC, eosinophilic, solid, and cystic 2021 M 15 11.5 cm/N.A EVT mTOR, c.7280 T > G * 19 months, NED 10 Farcas et al. 2021 F 69 4 cm/N.A EVT mTOR, c.4343_4363del * 47 months, NED 11 Farcas et al. 2021 M 42 7 cm/N.A EVT mTOR, c.7280 T > A * 18 months, NED 12 Xia et al. 2022 M 42 3.5 cm/N.A EVT mTOR c.7280 T > A N.A N.A 13 Xia et al. 2022 F 32 2.5 cm/N.A EVT mTOR c.7280 T > G N.A N.A 14 Xia et al. 2022 M 24 6 cm/N.A EVT mTOR c.7280 T > G N.A N.A 15 Xia et al. 2022 M 59 2.5 cm/N.A EVT mTOR c.7237_7238delinsCT N.A N.A 16 Xia et al. p ations: RCC, renal cell carcinoma; CAT. K, cathepsin K; CK, cytokeratin; PV, parvalbumin; AMACR, alpha-methylacyl-CoA racemase; CA9, carbonic anhydrase IX; SDH, succinate genase; ALK, anaplastic lymphoma kinase; N.A., not available; LOT, low-grade oncocytic tumor; ESC, eosinophilic, solid, and cystic In conclusion, herein, we present a distinct renal tumor characterized by high-grade eosinophilic cells, cathepsin-K ( l 3F9) i hi t h i l i d h b i ations: RCC, renal cell carcinoma; CAT. p ations: RCC, renal cell carcinoma; CAT. K, cathepsin K; CK, cytokeratin; PV, parvalbumin; AMACR, alpha-methylacyl-CoA racemase; CA9, carbonic anhydrase IX; SDH, succinate genase; ALK, anaplastic lymphoma kinase; N.A., not available; LOT, low-grade oncocytic tumor; ESC, eosinophilic, solid, and cystic 2022 M 47 2.5 cm/N.A EVT mTOR c.11C > T + TSC2 c.3352C > T N.A N.A 1 Tjota et al., 2020 F 66 4.1 cm/L LOT mTOR c.7280 T > A N.A 6 months NED 2 Tjota et al., 2020 M 66 2.5 cm/R LOT mTOR c.5930C > G no loss 156 months NED 3 Morini et al., 2021 F 57 3.7 cm/L LOT mTOR c.6644 C > T N.A 7 months NED 4 Morini et al., 2021 F 61 3.8 cm/R LOT mTOR c.7499 T > A N.A N.A 5 Morini et al., 2021 F 78 3.7 cm/R LOT mTOR c.6644 C > A N.A N.A 6 Morini et al., 2021 F 83 3.5 cm/R LOT mTOR c.4348 T > G N.A 49 died of other disease 7 Morini et al., 2021 F 79 8.5 cm/R LOT mTOR c.320_323delinsATTT N.A 49 months NED 8 Morini et al., 2021 F 58 5.5 cm/R LOT mTOR c.7280 T > A N.A N.A 9 Morini et al., 2021 F 76 3.7 cm/R LOT mTOR c.7498 A > T N.A N.A 10 Kapur et al., 2021 F 79 7.8 cm/R LOT mTOR p.Leu2427Gln no loss 26 months died 11 Kapur et al., 2021 F 86 6.5 cm/R LOT mTOR p.Ser2215Tyr no loss 3 months NED 12 Kapur et al., 2021 F 71 3.8 cm/R LOT mTOR p.Ser2413Leu no loss 53 months NED 13 Kapur et al., 2021 F 75 2.4 cm/L LOT mTOR p.Lys1452_Glu- 1455del no loss 36 months died 14 Zhang et al., 2022 F 79 1.6 cm/R LOT mTOR p.Leu2427Gln + TSC2 p.Met286Val N.A 33 months NED 15 Mohanty et al., 2021 F 65 2.3 cm/L LOT mTOR c.7280G > A N.A 2 months NED 16 Williamson et al., 2023 F 61 8 cm/N.A LOT mTOR c.7500 T > G N.A N.A 17 Williamson et al., 2023 F 73 6.9 cm/N.A LOT mTOR c.7280 T > A N.A N.A 18 Williamson et al., 2023, Trpkov 2019 F 63 5.2 cm/N.A LOT mTOR c.6644C > T + TSC1 c.2356C > T N.A N.A 19 Williamson et al., 2023, Trpkov 2019 F 61 3 cm/N.A LOT NF2 mTOR c.4448G > T N.A N.A 20 Williamson et al., 2023 F 41 2 cm/N.A LOT mTOR c.7498A > T N.A N.A 1 Virchows Archiv (2023) 483:821–833 831 Table 4 (continued) Case Reference Gender Age Size/laterality Diagnosis mTOR mutation chr 1 loss Follow-up 23 Chen et al., 2023 F 66 3.5 cm/R LOT mTOR p.L2427Q N.A 32 months NED 24 Chen et al., 2023 F 65 3 cm/R LOT mTOR p.L2427Q N.A 10 months NED 1 Romero et al., 2020 F 64 N.A Eosinophilic chro- mophobe mTOR I2501F N.A 5 months NED 2 Romero et al., 2020 F 50 N.A Classic chromo- phobe mTOR S2215F N.A 160 months NED 3 Romero et al., 2020 M 63 N.A Eosinophilic chro- mophobe mTOR I2500F N.A 41 months NED 4 Romero et al., 2020 F 58 N.A Eosinophilic chro- mophobe mTOR L2427R N.A 13 months NED 5 Romero et al., 2020 F 62 N.A Eosinophilic chro- mophobe mTOR L2427R N.A 10 months NED 6 Romero et al., 2020 F 44 N.A Eosinophilic chro- mophobe mTOR V2006F N.A 46 months NED 7 Romero et al., 2020 F 75 N.A N.A mTOR L2427R N.A 24 months metastasis 8 Romero et al., 2020 F 75 N.A Eosinophilic chro- mophobe mTOR E1613Q N.A 24 months NED 9 Tjota et al., 2021 M 65 8.5 cm/N.A Eosinophilic carci- noma mTOR, c.7279_7280delinsAA, p.Leu2427Lys loss Synchronous lymph node, liver and lung metastasis, N.A 10 Chen et al., 2016 N.A N.A N.A Unclassified RCC mTOR L2427R N.A N.A 11 Chen et al., 2016 N.A N.A N.A Unclassified RCC mTOR L2427R N.A N.A 12 Chen et al., 2016 N.A N.A N.A Unclassified RCC mTOR L2427R N.A N.A 13 Chen et al., 2016 N.A N.A N.A Unclassified RCC mTOR I1973F N.A N.A 14 Chen et al., 2016 N.A N.A N.A Unclassified RCC mTOR V2475M N.A N.A 15 Present study F 21 9.5 cm/L Eosinophilic carci- noma mTOR, c.4348_4368del p.Tyr1450_Trp1456del loss Lymph node metasta- sis, 30 months alive 16 Present study F 58 6.5 cm/L Eosinophilic carci- noma mTOR, c.7280 T > G, p.Leu2427Arg loss Skull metastasis after 132 months, 135 months alive 17 Present study M 49 3.6 cm/L Eosinophilic carci- noma mTOR, c.7280 T > A,p. p ations: RCC, renal cell carcinoma; CAT. K, cathepsin K; CK, cytokeratin; PV, parvalbumin; AMACR, alpha-methylacyl-CoA racemase; CA9, carbonic anhydrase IX; SDH, succinate genase; ALK, anaplastic lymphoma kinase; N.A., not available; LOT, low-grade oncocytic tumor; ESC, eosinophilic, solid, and cystic Leu2427Gln loss Liver metastasis after 36 months, 64 months alive i Abbreviations: chr, chromosome; M, male; F, female; R, right; L, left; N.A. not available; EVT, eosinophilic vacuolated tumor; LOT, low-grade oncocytic tumor; NED, no evidence of disease Funding Open access funding provided by Università degli Studi di Verona within the CRUI-CARE Agreement. This study was funded by the “Bando di ricerca finalizzata 2021 (PC: GR-2021–12374462).” Funding Open access funding provided by Università degli Studi di Verona within the CRUI-CARE Agreement. This study was funded by the “Bando di ricerca finalizzata 2021 (PC: GR-2021–12374462).” combination. This latter observation encourages pathologists to investigate mTOR gene mutation in aggressive high-grade/ cathepsin-K-positive eosinophilic RCC. Data availability All data generated or analysed during this study are included in this published article [and its supplementary information files]. Supplementary Information The online version contains supplemen- tary material available at https://doi.org/10.1007/s00428-023-03688-2. i Abbreviations: chr, chromosome; M, male; F, female; R, right; L, left; N.A. not available; EVT, eosinophilic vacuolated tumor; LOT, low-grade oncocytic tumor; NED, no evidence of disease Declarations Acknowledgements The authors would like to thank Dr. Roldan- Romero JM for sharing some information regarding a few cases previ- ously reported [17]. Ethics approval All procedures involving human participants received institutional review board approval (Prog. 4136CESC) and were in accordance with the ethical standards of the institutional and/or national research committee and with the declaration of Helsinki. Author contribution Conceptualization: A.C. and G.M.; methodology: A.C., S.M., G.S., and S.P.; formal analysis and investigation: A.C. and S.M.; writing—original draft preparation: A.C., S.M., and M.R.; writ- ing—review and editing: S.G., L.S., E.M., M.B., L.M., A.P., S.F., F.P., M. R. R., A.M., and C.D.; supervision: M.B.A., C.P., and G.M. Informed consent All patients gave their written informed consent to diagnostic procedures and treatment according to institutional rules for everyday clinical practice. 1 3 Virchows Archiv (2023) 483:821–833 832 Conflict of interest The authors declare no competing interests. Conflict of interest The authors declare no competing interests. 8. Caliò A, Brunelli M, Segala D, Pedron S, Doglioni C, Argani P, Martignoni G (2019) VEGFA amplification/increased gene copy number and VEGFA mRNA expression in renal cell carcinoma with TFEB gene alterations. Mod Pathol 32:258–268. https://doi. org/10.1038/s41379-018-0128-1 Open Access This article is licensed under a Creative Commons Attri- bution 4.0 International License, which permits use, sharing, adapta- tion, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. g 9. Martignoni G, Pea M, Zampini C, Brunelli M, Segala D, Zamboni G, Bonetti F (2015) PEComas of the kidney and of the genitou- rinary tract. Semin Diagn Pathol 32:140–159. https://doi.org/10. 1053/j.semdp.2015.02.006 10. Caliò A, Brunelli M, Segala D, Zamboni G, Bonetti F, Pea M, Martignoni G (2021) Angiomyolipoma of the kidney: from simple hamartoma to complex tumour. Pathology 53:129–140. https:// doi.org/10.1016/j.pathol.2020.08.008 11. References 1. Amin MB, McKenney JK, Martignoni G, Campbell SC, Pal S, Tickoo SK (2022) Low grade oncocytic tumors of the kid- ney: a clinically relevant approach for the workup and accurate diagnosis. Mod Pathol 35:1306–1316. https://doi.org/10.1038/ s41379-022-01108-5 12. Yang P, Cornejo KM, Sadow PM, Cheng L, Wang M, Xiao Y, Jiang Z, Oliva E, Jozwiak S, Nussbaum RL, Feldman AS, Paul E, Thiele EA, Yu JJ, Henske EP, Kwiatkowski DJ, Young RH, Wu C-L (2014) Renal cell carcinoma in tuberous sclerosis com- plex. Am J Surg Pathol 38:895–909. https://doi.org/10.1097/ PAS.0000000000000237 2. Trpkov K, Hes O, Bonert M, Lopez JI, Bonsib SM, Nesi G, Com- perat E, Sibony M, Berney DM, Martinek P, Bulimbasic S, Suster S, Sangoi A, Yilmaz A, Higgins JP, Zhou M, Gill AJ, Przyby- cin CG, Magi-Galluzzi C, McKenney JK (2016) Eosinophilic, solid, and cystic renal cell carcinoma: clinicopathologic study of 16 unique, sporadic neoplasms occurring in women. Am J Surg Pathol 40:60–71. https://doi.org/10.1097/PAS.0000000000000508 13. Shah RB, Stohr BA, Tu ZJ, Gao Y, Przybycin CG, Nguyen J, Cox RM, Rashid-Kolvear F, Weindel MD, Farkas DH, Trpkov K, McKenney JK (2020) “Renal cell carcinoma with leiomy- omatous stroma” harbor somatic mutations of TSC1, TSC2, MTOR, and/or ELOC (TCEB1): clinicopathologic and molecu- lar characterization of 18 sporadic tumors supports a distinct entity. Am J Surg Pathol 44:571–581. https://doi.org/10.1097/ PAS.0000000000001422 3. Munari E, Settanni G, Caliò A, Segala D, Lonardi S, Sandrini S, Vacca P, Tumino N, Marconi M, Brunelli M, Gobbo S, Netto GJ, Moretta L, Zamboni G, Martignoni G (2022) TSC loss is a clonal event in eosinophilic solid and cystic renal cell carcinoma: a multiregional tumor sampling study. Mod Pathol 35:376–385. https://doi.org/10.1038/s41379-021-00816-8 14. Parilla M, Kadri S, Patil SA, Ritterhouse L, Segal J, Henriksen KJ, Antic T (2018) Are sporadic eosinophilic solid and cystic renal cell carcinomas characterized by somatic tuberous sclero- sis gene mutations? Am J Surg Pathol 42:911–917. https://doi. org/10.1097/PAS.0000000000001067 p g 4. References Gill AJ, Hes O, Papathomas T, Šedivcová M, Tan PH, Agaimy A, Andresen PA, Kedziora A, Clarkson A, Toon CW, Sioson L, Watson N, Chou A, Paik J, Clifton-Bligh RJ, Robinson BG, Benn DE, Hills K, Maclean F, Niemeijer ND, Vlatkovic L, Hartmann A, Corssmit EPM, van Leenders GJLH, Przybycin C, McKenney JK, Magi-Galluzzi C, Yilmaz A, Yu D, Nicoll KD, Yong JL, Sibony M, Yakirevich E, Fleming S, Chow CW, Miettinen M, Michal M, Trpkov K (2014) Succinate dehydrogenase (SDH)-deficient renal carcinoma: a morphologically distinct entity: a clinicopathologic series of 36 tumors from 27 patients. Am J Surg Pathol 38:1588– 1602. https://doi.org/10.1097/PAS.0000000000000292 15. Palsgrove DN, Li Y, Pratilas CA, Lin M-T, Pallavajjalla A, Gocke C, De Marzo AM, Matoso A, Netto GJ, Epstein JI, Argani P (2018) Eosinophilic solid and cystic (ESC) renal cell carcinomas harbor tsc mutations: molecular analysis supports an expanding clinicopathologic spectrum. Am J Surg Pathol 42:1166–1181. https://doi.org/10.1097/PAS.0000000000001111 16. Mehra R, Vats P, Cao X, Su F, Lee ND, Lonigro R, Premkumar K, Trpkov K, McKenney JK, Dhanasekaran SM, Chinnaiyan AM (2018) Somatic bi-allelic loss of tsc genes in eosinophilic solid and cystic renal cell carcinoma. Eur Urol 74:483–486. https://doi.org/10.1016/j.eururo.2018.06.007 5. Smith SC, Sirohi D, Ohe C, McHugh JB, Hornick JL, Kalariya J, Karia S, Snape K, Hodgson SV, Cani AK, Hovelson D, Luthringer DJ, Martignoni G, Chen Y-B, Tomlins SA, Mehra R, Amin MB (2017) A distinctive, low-grade oncocytic fumarate hydratase- deficient renal cell carcinoma, morphologically reminiscent of succinate dehydrogenase-deficient renal cell carcinoma. Histo- pathology 71:42–52. https://doi.org/10.1111/his.13183 17. Roldan-Romero JM, Santos M, Lanillos J, Caleiras E, Anguera G, Maroto P, García-Donas J, de Velasco G, Martinez-Montes ÁM, Calsina B, Monteagudo M, Letón R, Leandro-García LJ, Montero-Conde C, Cascón A, Robledo M, Rodriguez-Antona C (2020) Molecular characterization of chromophobe renal cell carcinoma reveals mTOR pathway alterations in patients with poor outcome. Mod Pathol 33:2580–2590. https://doi.org/10. 1038/s41379-020-0607-z 6. Caliò A, Segala D, Munari E, Brunelli M, Martignoni G (2019) MiT family translocation renal cell carcinoma: from the early descriptions to the current knowledge. Cancers (Basel) 11:1–12. https://doi.org/10.3390/cancers11081110 18. Morini A, Drossart T, Timsit M-O, Sibony M, Vasiliu V, Gimenez-Roqueplo A-P, Favier J, Badoual C, Mejean A, Bur- nichon N, Verkarre V (2022) Low-grade oncocytic renal tumor (LOT): mutations in mTOR pathway genes and low expression of FOXI1. Mod Pathol 35:352–360. https://doi.org/10.1038/ s41379-021-00906-7 p g 7. Declarations Guo J, Tretiakova MS, Troxell ML, Osunkoya AO, Fadare O, Sangoi AR, Shen SS, Lopez-Beltran A, Mehra R, Heider A, Hig- gins JP, Harik LR, Leroy X, Gill AJ, Trpkov K, Campbell SC, Przybycin C, Magi-Galluzzi C, McKenney JK (2014) Tuberous sclerosis-associated renal cell carcinoma: a clinicopathologic study of 57 separate carcinomas in 18 patients. Am J Surg Pathol 38:1457–1467. https://doi.org/10.1097/PAS.0000000000000248 References Kapur P, Gao M, Zhong H, Rakheja D, Cai Q, Pedrosa I, Mar- gulis V, Xu L, Kinch L, Brugarolas J (2021) Eosinophilic vacu- olated tumor of the kidney: a review of evolving concepts in this novel subtype with additional insights from a case with MTOR mutation and concomitant chromosome 1 loss. Adv Anat Pathol 28:251–257. https://doi.org/10.1097/PAP.0000000000000299 22. Tjota MY, Segal J, Stadler WM, Antic T (2021) Eosinophilic renal cell carcinoma with isolated MTOR mutation metastatic to the liver: a novel case. Pathology 53:790–793. https://doi.org/10. 1016/j.pathol.2020.10.013 p g 30. Grabiner BC, Nardi V, Birsoy K, Possemato R, Shen K, Sinha S, Jordan A, Beck AH, Sabatini DM (2014) A diverse array of cancer-associated MTOR mutations are hyperactivating and can predict rapamycin sensitivity. Cancer Discov 4:554–563. https:// doi.org/10.1158/2159-8290.CD-13-0929 23. Caliò A, Brunelli M, Segala D, Pedron S, Remo A, Ammendola S, Munari E, Pierconti F, Mosca A, Bollito E, Sidoni A, Fisogni S, Sacco C, Canu L, Sentinelli S, Fraccon AP, Fiorentino M, Scott C, Milella M, Porta C, Argani P, Martignoni G (2020) Comprehensive analysis of 34 MiT family translocation renal cell carcinomas and review of the literature: investigating prognostic markers and therapy targets. Pathology 52:297–309. https://doi.org/10.1016/j.pathol.2019.11.006 31. Voss MH, Hakimi AA, Pham CG, Brannon AR, Chen Y-B, Cunha LF, Akin O, Liu H, Takeda S, Scott SN, Socci ND, Viale A, Schultz N, Sander C, Reuter VE, Russo P, Cheng EH, Motzer RJ, Berger MF, Hsieh JJ (2014) Tumor genetic analyses of patients with metastatic renal cell carcinoma and extended benefit from mTOR inhibitor therapy. Clin Cancer Res 20:1955–1964. https:// doi.org/10.1158/1078-0432.CCR-13-2345 gy p g j p 24. References Argani P, Reuter VE, Zhang L, Sung Y-S, Ning Y, Epstein JI, Netto GJ, Antonescu CR (2016) TFEB-amplified renal cell car- cinomas: an aggressive molecular subset demonstrating variable melanocytic marker expression and morphologic heterogeneity. Am J Surg Pathol 40:1484–1495. https://doi.org/10.1097/PAS. 0000000000000720 19. Kapur P, Gao M, Zhong H, Chintalapati S, Mitui M, Barnes SD, Zhou Q, Miyata J, Carrillo D, Malladi VS, Rakheja D, Pedrosa 1 3 Virchows Archiv (2023) 483:821–833 833 I, Xu L, Kinch L, Brugarolas J (2022) Germline and sporadic mTOR pathway mutations in low-grade oncocytic tumor of the kidney. Mod Pathol 35:333–343. https://doi.org/10.1038/ s41379-021-00896-6 precision oncology knowledge base JCO Precis. Oncol. https:// doi.org/10.1200/PO.17.00011 I, Xu L, Kinch L, Brugarolas J (2022) Germline and sporadic mTOR pathway mutations in low-grade oncocytic tumor of the kidney. Mod Pathol 35:333–343. https://doi.org/10.1038/ s41379-021-00896-6 26. Yang H, Rudge DG, Koos JD, Vaidialingam B, Yang HJ, Pavletich NP (2013) mTOR kinase structure, mechanism and regulation. Nature 497:217–223. https://doi.org/10.1038/nature12122 20. Chen Y-B, Mirsadraei L, Jayakumaran G, Al-Ahmadie HA, Fine SW, Gopalan A, Sirintrapun SJ, Tickoo SK, Reuter VE (2019) Somatic mutations of TSC2 or MTOR characterize a morphologi- cally distinct subset of sporadic renal cell carcinoma with eosino- philic and vacuolated cytoplasm. Am J Surg Pathol 43:121–131. https://doi.org/10.1097/PAS.0000000000001170 27. Chen Y-B, Xu J, Skanderup AJ, Dong Y, Brannon AR, Wang L, Won HH, Wang PI, Nanjangud GJ, Jungbluth AA, Li W, Ojeda V, Hakimi AA, Voss MH, Schultz N, Motzer RJ, Russo P, Cheng EH, Giancotti FG, Lee W, Berger MF, Tickoo SK, Reuter VE, Hsieh JJ (2016) Molecular analysis of aggressive renal cell carcinoma with unclassified histology reveals distinct subsets. Nat Commun 7:13131. https://doi.org/10.1038/ncomms13131 p g 21. Farcas M, Gatalica Z, Trpkov K, Swensen J, Zhou M, Alaghe- hbandan R, Williamson SR, Magi-Galluzzi C, Gill AJ, Tretiakova M, Lopez JI, Perez Montiel D, Sperga M, Comperat E, Brimo F, Yilmaz A, Siadat F, Sangoi A, Gao Y, Kuthi Ptákova NL, Pivo- varcikova K, Rogala J, Agaimy A, Hartmann A, Fraune C, Rychly B, Hurnik P, Durcansky D, Bonert M, Gakis G, Michal M, Hora M, Hes O (2021) Mod Pathol 35(3):344–351. https://doi.org/10. 1038/s41379-021-00923-6 28. Caliò A, Brunelli M, Gobbo S, Argani P, Munari E, Netto G, Mar- tignoni G (2021) Cathepsin K: a novel diagnostic and predictive biomarker for renal tumors. Cancers (Basel) 13(10):2441. https:// doi.org/10.3390/cancers13102441 29. References Howe KL, Achuthan P, Allen J, Allen J, Alvarez-Jarreta J, Amode MR, Armean IM, Azov AG, Bennett R, Bhai J, Billis K, Boddu S, Charkhchi M, Cummins C, Da RinFioretto L, Davidson C, Dodiya K, El Houdaigui B, Fatima R, Gall A, Garcia Giron C, Grego T, Guijarro-Clarke C, Haggerty L, Hemrom A, Hourlier T, Izuogu OG, Juettemann T, Kaikala V, Kay M, Lavidas I, Le T, Lemos D, Gon- zalez Martinez J, Marugán JC, Maurel T, McMahon AC, Mohanan S, Moore B, Muffato M, Oheh DN, Paraschas D, Parker A, Parton A, Prosovetskaia I, Sakthivel MP, Salam AIA, Schmitt BM, Schu- ilenburg H, Sheppard D, Steed E, Szpak M, Szuba M, Taylor K, Thormann A, Threadgold G, Walts B, Winterbottom A, Chakiachvili M, Chaubal A, De Silva N, Flint B, Frankish A, Hunt SE, IIsley GR, Langridge N, Loveland JE, Martin FJ, Mudge JM, Morales J, Perry E, Ruffier M, Tate J, Thybert D, Trevanion SJ, Cunningham F, Yates AD, Zerbino DR, Flicek P (2021) Ensembl. Nucleic Acids Res 49:D884–D891. https://doi.org/10.1093/nar/gkaa942 32. Voss MH, Chen D, Reising A, Marker M, Shi J, Xu J, Ostrovnaya I, Seshan VE, Redzematovic A, Chen Y-B, Patel P, Han X, Hsieh JJ, Hakimi AA, Motzer RJ (2019) PTEN expression, not mutation status in TSC1, TSC2, or mTOR. Correlates with the outcome on everolimus in patients with renal cell carcinoma treated on the randomized RECORD-3 trial. Clin Cancer Res 25:506–514. https://doi.org/10.1158/1078-0432.CCR-18-1833 33. Adib E, Klonowska K, Giannikou K, Do KT, Pruitt-Thompson S, Bhushan K, Milstein MI, Hedglin J, Kargus KE, Sholl LM, Tsuji J, Hyman DM, Sisk A, Shapiro GI, Vargas HA, Harding JJ, Voss MH, Iyer G, Kwiatkowski DJ (2021) Phase II clinical trial of everolimus in a pan-cancer cohort of patients with mTOR pathway alterations. Clin Cancer Res 27:3845–3853. https://doi. org/10.1158/1078-0432.CCR-20-4548 25. 25. Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Chakravarty D, Gao J, Phillips SM, Kundra R, Zhang H, Wang J, Rudolph JE, Yaeger R, Soumerai T, Nissan MH, Chang MT, Chandarlapaty S, Traina TA, Paik PK, Ho AL, Hantash FM, Grupe A, Baxi SS, Callahan MK, Snyder A, Chi P, Danila D, Gounder M, Harding JJ, Hellmann MD, Iyer G, Janjigian Y, Kaley T, Levine DA, Lowery M, Omuro A, Postow MA, Rathkopf D, Shoushtari AN, Shukla N, Voss M, Paraiso E, Zehir A, Berger MF, Taylor BS, Saltz LB, Riely GJ, Ladanyi M, Hyman DM, Baselga J, Sabbatini P, Solit DB, Schultz N (2017) OncoKB: a Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. 1 3 1 3 1 3
https://openalex.org/W4286561115	https://www.researchsquare.com/article/rs-1861079/latest.pdf	English	null	Compositional properties and codon usage pattern of mitochondrial cytochrome c oxidase I gene among Potamidids	Research Square (Research Square)	2,022	cc-by	6,025	Compositional properties and codon usage pattern of mitochondrial cytochrome c oxidase I gene among Potamidids Zulfa I. Shabani (  zulfai937@gmail.com ) University of Dar es Salaam Prosper L. Mfilinge University of Dar es Salaam Fulgence N. Mpenda University of Dar es Salaam Research Article Keywords: Codon usage, Mutation pressure, Natural selection, Molecular evolution, Potamidids Posted Date: July 22nd, 2022 DOI: https://doi.org/10.21203/rs.3.rs-1861079/v1 License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License Page 1/20 Page 1/20 Abstract Codon usage bias refers to the phenomenon where there is unequal occurrence of specific synonymous codons in a genome, the frequency of which varies within and among species. Molecular evolutionary researches suggest that CUB studies are significant in interpreting evolutionary trends of species or genes. A number of studies on codon usage bias are emerging in different organisms. However, such information remained unclear for organisms in the family Potamididae. This study systematically examined compositional features and codon usage patterns in MT-COI gene among potamidids to identify factors shaping codon usage bias. Quantification of codon usage bias and the identification of causative factors was performed by assessing various parameters such as nucleotide composition, effective number of codons (ENC), relative synonymous codon usage (RSCU), parity rule (PR2) analysis, neutrality plot analysis as well as a multivariate statistical approach of correspondence analysis (CoA). The results from neutrality plot suggested that natural selection is the dominant evolutionary factor for CUB in MT-COI gene among the different genera of potamidids. Correspondence analysis (CoA), parity plot and ENC analyses indicated the existence of mutational pressure with minor contribution. Nucleotide composition investigations indicated that, most codons are AT rich, with T nucleotides count being highly presented. Thus, compositional restraint was shown as among the reasons behind the pattern of codon usage in MT-COI gene of species in the family Potamididae. Introduction mportant in understanding their diversity and genetic evolution. In this study, nucleotide composition and CUB in MT-COI gene for organisms in the four different genera of Potamididae family were investigated to explore potamidids diversification as no work has been done yet. Introduction Evolutionary processes involves change in heritable traits over successive generations of a biological population (Fowler et al. 2018). Therefore, evolutionary processes give rise to diversity at every biological organization level (Adams and Engel 2021; Cassidy 2020; Hopcroft 2018).The major driving mechanism of heredity in all individuals is in form of genes received from parents and passed to offspring (Alberts 2015; Fowler et al. 2018). At the molecular level, that is, molecular evolution, changes in the number of nucleotides or amino acid substitution per site per year is what influences evolutionary changes (Nei et al. 2000; Ruse and Travis 2009). However, some changes of a single DNA nucleotide within a protein-coding portion of a gene may not affect the sequence of amino acids that make up the gene's protein a phenomenon called silent sites mutation (Klasen and Wabl 2004; Czech et al. 2010). Comparing the rates of evolution among genes and pattern of silent site evolution as revealed, for example, by codon usage studies can tell more about the forces behind genome evolution. Codon usage refers to the frequency in which the available codons of a gene are being used by a particular organism (Behura and Severson 2013; Zhou et al. 2016; Parvathy et al. 2022). Although each codon is specific for only one amino acid, some amino acids are said to be encoded by more than one codon, making the genetic code described as degenerate. The different codons that specify the same amino acids are referred to as synonymous codons (Mitra et al. 2016). Even though they encode the same amino acid, it has been shown for all organisms that there is unequal occurrence of synonymous codons in a genome, a phenomenon called codon usage bias (CUB) (Behura and Severson 2013). The major factors influencing CUB are genetic drift, mutation pressure and natural selection (Salim and Page 2/20 Page 2/20 Page 2/20 Cavalcanti 2008; Nath Choudhury et al. 2017; Gun et al. 2018; Parvathy et al. 2022). Therefore, when studied in details, CUB echoes the origin of mutation pattern as well as evolution of species or genes, and can help to categorize a gene (or species) and interpret its evolutionary trends. Cavalcanti 2008; Nath Choudhury et al. 2017; Gun et al. 2018; Parvathy et al. 2022). Retrieval of sequences Retrieval of 16 MT-COI gene sequences for the different species belonging to six genera of the family potamididae was performed from NCBI (http://www.ncbi.nlm.nih.gov/Genbank/). Table 1 lists the species, their accession number and the genus to which they belong. Table 1 List of 16 species of potamidids with accession number Page 3/20 Introduction Therefore, when studied in details, CUB echoes the origin of mutation pattern as well as evolution of species or genes, and can help to categorize a gene (or species) and interpret its evolutionary trends. Mitochondria is an important organelle that plays a central role in cellular energy provision, containing their own genome with a modified genetic code (Crimi and Rigolio 2008; Osellame et al. 2012; Chinnery and Hudson 2013; Friedman and Nunnari 2014). Having simple genomic organization, lack of recombination and a high rate of evolution than that of the nuclear genome, mitochondria serve a vital source of genetic information (Arif et al. 2011). Since mitochondrial genes are known to be conserved evolutionarily, it is possible to weigh related species in terms of their evolutionary pattern and molecular changes (Nabholz et al. 2008; Chinnery and Hudson 2013). Consequently, mitochondrial genome studies assist in understanding species diversity as well as population genetics (Castro et al. 1998; Nabholz et al. 2008). Cytochrome c oxidase is a complex (COI, COII and COIII) that concludes the ultimate step of mitochondrial respiratory chain (Ferguson-Miller 2013; Shimada et al. 2017). However, COI has been revealed to have distinct phylogenetic signal than other mitochondrial genes as a result of its satisfactorily hasty evolution (Townsend et al. 2012; Souza et al. 2016). Analyzing the manner of codon usage of MT-COI gene with the help of various parameters could thereby be crucial in discriminating genes (or species) and investigate their intraspecific diversity. The family Potamididae is one of the common molluscan fauna of mangrove forests in the Indo-West Pacific area (Reid et al. 2008). They play a significant role in the mangrove ecosystem, as bioindicators of health and ecological changes, but also as biofilters in waste water (Ratsimbazafy and Kochzius 2018; Wells and Keesing 2019). Some potamidids are also used as a food source for human (Ratsimbazafy and Kochzius 2018). Traditionally, potamidids have been confused because they have similar shells and occupy almost common habitats, thus, in the past their identification has been problematic (Willan 2013). This raises the need to understand the evolutionary forces behind their differences which is then important in understanding their diversity and genetic evolution. In this study, nucleotide composition and CUB in MT-COI gene for organisms in the four different genera of Potamididae family were investigated to explore potamidids diversification as no work has been done yet. Nucleotide composition analysis Base nucleotide composition constraint is among the determinants of CUB across genomes, the relationship of which has been reported in various literature (Barbhuiya et al. 2021; Chakraborty et al. 2022). Therefore, to reveal significant factor(s) shaping CUB in potamidids nucleotide compositional characteristics including: (i) nucleotide counts (A, C, T and G %) (ii) Third position nucleotide count (A3, C3, T3 and G3%) (iii) Total GC% as well as positional GC content (GC1, GC2 and GC3%), were calculated from (https://genomes.urv.es/CAICal) web service. Page 3/20 S/N SPECIES NAME GENUS ACCESSION NUMBER 1 Cerithidea obusta Cerithidea NC_039951 2 Cerithidea sinensis Cerithidea AB219355 3 Cerithidea rhizophorarum Cerithidea AB219333 4 C. reidi Cerithidea HE680238 5 C. alata Cerithideopsilla JF694693 6 C. cingulate Cerithidiopsilla JF694666 7 C. costata Cerithidiopsis GQ273849 8 C. montagnei Cerithidiopsis GQ273847 9 C. californica Cerithideopsis GQ273835 10 C. microptera Cerithideopsilla HE680612 11 C. conica Cerithideopsilla AM932759 12 C. insica Cerithideopsilla MZ831994 13 C. scalariformis Cerithideopsis HE680627 14 T. palustris Terebralia LN649639 15 T. semistriata Terebralia HE680660 16 T.sulcata Terebralia MZ832021 Effective number of codons (ENC) Effective number of codons is termed as the extent of species-independent synonymous codon bias in genes (Wright 1990; Sun et al. 2013). The values represent the number of equally used codons that would generate the same codon usage bias as that observed and are thereby used to quantify CUB in a gene (Wright 1990). The DnaSP v.6 software was used to calculate ENC values, to investigate whether its Page 4/20 Page 4/20 natural selection or mutation pressure that influence CUB among members of the family potamididae. Lower ENC value is 20 denoting highly biased gene, whilst the upper limit being 61 for equally used codons (Yang et al. 2014). Correspondence analysis To determine the major factors affecting the pattern of codon usage in each genus, Correspondence analysis, a multivariate statistical technique was executed using R program. In this case, the large scale trend of codon usage bias in the sequence of nucleic acids could be estimated. Parity rule 2 (PR2) analysis Parity plot was created to evaluate Chargaff`s second parity rule (A=T and G=C) by plotting AT bias [A/ (A+T)] on the y-axis versus GC bias [G/ (G+C)] on the x-axis. Any deviation from PR2 indicates biasness arbitrated by either natural selection, mutation pressure or both forces (Nasrullah et al. 2015). This was created using OriginPro software. Relative Synonymous Codon Usage (RSCU) Relative synonymous codon usage refers to the ratio of the observed codon frequency to the frequency expected if codon usage was uniform (Gun et al. 2018). The values are used to quantify the relative usage of synonymous codons within a codon family of a gene. RSCU > 1 indicates positive codon usage i.e. more frequently used (De Mandal et al. 2020; Yu et al. 2021). On the contrary, RSCU < 1 reveals a negative codon usage (less frequently used) (De Mandal et al. 2020; Gun et al. 2018) Over-represented and under-represented codons have their RSCU values above 1.6 and below 0.6 respectively (Xu et al. 2013; Shafat et al. 2022) In this study, RSCU values of MT-COI gene of the various species in the family potamididae were estimated using DnaSP v.6 software. Neutrality plot analysis This is the most widely used method to accurately reveal the prime driving force of CUB as either mutation pressure or natural selection. The plots were constructed utilizing average GC12 values on the y- axis against GC3 values on the x-axis in R program. When the codon bias is absolutely under the action of neutrality or directional mutational force the quantitative relation between GC3 and GC12 becomes almost equal and the slope is nearer to 1(He et al. 2016; Shen et al. 2020) Nucleotide composition The trend of average nucleotide composition for MT-COI gene was T >A >G >C in Cerithidea and Cerithideopsis, whereas, in Terebralia and Cerithideopsilla the trend was T >G >A >C and T >A >C >G Page 5/20 respectively (Fig. 1). The overall nucleobase T composition was the highest in all the four genera within Potamididae family. respectively (Fig. 1). The overall nucleobase T composition was the highest in all the four genera within Potamididae family. The fashion of nucleotide composition at the third position (A3, C3, T3 and G3) was as T3>A3>C3>G3, thus nucleobase T ranked the highest followed by A with exception to genus Terebralia. Again, nucleobase C was the least prevalent except for Cerithideopsilla where G nucleobase had shown low prevelance (Fig. 2). Mean GC content was 38.32%, 39.68%, 39.12%, and 41.84% in Cerithidea, Cerithideopsilla, Cerithideopsis and Terebralia genera, respectively (Fig. 2). MT-COI gene was relatively AT-rich in all genera of Potamididae family. Effective Number of Codons (ENC). The effective number of codons was calculated to analyze codon usage pattern in the different species belonging to family Potamididae. The ENC values found in the present study for the six genera of Potamididae are displayed in Table 2. Although for all genera CUB as determined by ENC values was moderate, the lowest CUB was observed in genus Cerithideopsis as compared to other genera (Table 2). Table 2 ENC values of MT-COI gene in different genera of the family Potamididae. ENC values of MT-COI gene in different genera of the family Potamididae. S/N Genus MT-COI 1 Cerithidea 45.65% 2 Cerithideopsilla 46.43% 3 Cerithideopsis 49.86% 4 Terebralia 48.86% Pattern of relative synonymous codon usage (RSCU) This was calculated to quantify the relative usage of synonymous codons within a codon family of MT- COI gene. The RSCU values were as presented in Table 3 below. Pattern of relative synonymous codon usage (RSCU) AGU 0.48 0.52 1.57 0.41 AGC 1.09 1.42 1.40 1.14 AGA 1.78 1.20 1.08 1.49 AGG 1.44 2.68 1.77 1.69 GGU 1.32 0.19 0.89 1.09 GGC 1.56 1.04 1.65 1.81 GGA 0.75 0.19 0.43 0.20 GGG 0.00 1.18 1.34 1.42 bove 1 6 is indicated by bold RSCU value below 0 6 is indicated by italics wherea RSCU value above 1.6 is indicated by bold, RSCU value below 0.6 is indicated by italics whereas RSCU value between 0.6 and 1.6 is indicated by roman. Correspondence analysis (CoA) Neutrality plot Neutrality plots executed using RStudio, based on GC3% and GC12% were as represented on Figure 5 below. A negative slope in Cerithideopsis Cerithideopsilla and Terebralia indicated negative association between GC3% and GC12%. However, a slight positive slope indicating slight positive correlation between GC3 and GC12 was found in Cerithidea. Correspondence analysis (CoA) Correspondence analysis showing the distribution pattern of codons in the four genera is as shown in Figure 5. The distribution depicted variation in codon usage across the different species in the family Potamididae for MT-COI gene. Most codon positions were found to concentrate close to the axes and at the center of the plot. Parity plot analysis To estimate the effects of mutation pressure and natural selection on the codon usage, PR2 plot was executed. It was found that, the gene spots were scattered to show deviation from base composition rule (Chargaff`s second parity rule) i.e. A and T were not proportionate with C and G (Fig. 4). Pattern of relative synonymous codon usage (RSCU) Table 3 RSCU values of codons for MT-COI gene in the six genera of Potamididae Table 3 RSCU values of codons for MT-COI gene in the six genera of Potamididae Page 6/20 CODON RSCU VALUES CODON RSCU VALUES CODON RSCU VALUES Cerithidea Cerithideopsilla Cerithideopsis Terebralia UUU 1.55 1.40 1.57 1.60 UUC 0.36 0.60 0.43 0.40 UUA 0.84 0.69 1.21 0.00 UUG 0.82 1.28 1.00 2.00 CUU 2.20 3.39 2.54 3.20 CUC 1.56 0.31 1.02 0.58 CUA 0.50 0.23 0.23 0.25 CUG 0.22 0.11 0.23 0.00 AUU 1.49 1.44 1.53 1.24 AUC 0.53 0.93 0.70 0.84 AUA 0.79 0.63 0.77 0.59 AUG 0.96 1.00 1.00 1.00 GUU 1.57 2.40 2.75 2.65 GUC 0.61 1.20 0.50 0.71 GUA 0.87 0.00 0.46 0.44 GUG 0.92 0.76 0.29 0.21 UCU 1.69 1.71 1.57 2.04 UCC 1.36 0.81 0.77 1.12 UCA 1.22 0.49 0.74 0.46 UCG 0.46 0.83 0.21 0.79 CCU 2.28 1.59 0.52 1.36 CCC 0.29 0.95 1.49 0.84 CCA 1.36 0.63 1.25 1.42 CCG 0.40 0.70 0.00 0.71 ACU 1.50 1.43 1.91 2.10 ACC 0.25 0.93 0.47 0.38 ACA 1 28 0 50 0 82 0 80 Page 7/20 Page 7/20 ACG 0.50 0.85 0.81 1.00 GCU 1.40 1.35 1.38 1.18 GCC 0.73 1.17 0.34 1.19 GCA 1.37 0.37 1.81 0.55 GCG 0.90 0.47 0.47 0.70 UAU 1.25 1.39 1.33 1.47 UAC 0.25 0.61 0.84 0.36 UAA 0.56 1.04 1.26 1.15 UAG 0.16 0.58 1.52 0.52 CAU 1.39 1.19 1.17 0.33 CAC 0.49 0.62 0.77 1.67 CAA 0.75 1.20 1.06 1.89 CAG 1.25 0.40 0.50 0.11 AAU 0.89 0.76 0.84 1.58 AAC 0.61 1.10 0.66 1.24 AAA 0.58 0.90 1.11 0.71 AAG 1.21 0.98 0.72 1.14 GAU 0.82 0.83 1.07 0.98 GAC 1.83 1.20 0.77 0.60 GAA 0.42 0.65 1.29 0.78 GAG 0.52 0.50 0.84 0.86 UGU 0.52 1.30 0.74 1.27 UGC 2.24 1.01 1.41 0.87 UGA 1.13 1.12 0.91 0.00 UGG 0.75 1.17 0.81 0.95 CGU 0.00 1.06 0.63 1.00 CGC 1.34 0.72 0.49 0.95 CGA 0 52 0 68 0 74 0 38 AGU 0.48 0.52 1.57 0.41 AGC 1.09 1.42 1.40 1.14 AGA 1.78 1.20 1.08 1.49 AGG 1.44 2.68 1.77 1.69 GGU 1.32 0.19 0.89 1.09 GGC 1.56 1.04 1.65 1.81 GGA 0.75 0.19 0.43 0.20 GGG 0.00 1.18 1.34 1.42 RSCU value above 1.6 is indicated by bold, RSCU value below 0.6 is indicated by italics whereas RSCU value between 0.6 and 1.6 is indicated by roman. Discussion Codon usage bias is a common feature in most organism genome including prokaryotes and eukaryotes, and may contribute to genome evolution in a profound manner (Behura and Severson 2013; Zhou et al. 2016). The two widely referred theories that explained the foundation of codon usage bias in organisms are neutral theory and selection-mutation drift model (Duret 2002; Deb et al. 2020). According to neutral theory, silent mutations in the codons` third position rise the phenomenon of codon preference (Subramanian 2008). Whereas, selection-mutation drift model suggest that, CUB occurs as an Page 9/20 Page 9/20 equilibrium event to balance natural selection, mutation pressure and genetic drift (Bulmer 1991; Duret 2002). Individual genes show variation in codon usage pattern; however, even within a gene the pattern of codon usage may vary depending on the position of the codon (Gerdol et al. 2015). The coding sequences of a gene and/or genome provide information on gene function and hence evolution of the organism (Behura and Severson 2013; Zhou et al. 2016). Thus, studies on the patterns of codon usage including their causes, consequences, as well as selective forces shaping their evolution are significant in understanding genome biology and population diversification. In this study, compositional aspects and codon usage properties for MT-COI gene in Cerithidea, Cerithideopsilla, Cerithideopsis and Terebralia genera of Potamididae were analyzed. Findings from the present study have demonstrated that both natural selection and mutation pressure were contributing factors on CUB and evolutionary variations among Potamidids. From PR2 plot (Fig. 4) it was found that, gene spots deviated from base composition rule (Chargaff`s second parity rule) that is, A and T were not proportionate with C and G (Fig. 4). This suggested that a combinatorial action of mutation pressure and natural selection influenced CUB in the species studied. Although parity plot reflected the main factors influencing codon usage bias in different genera of Potamidids, there was no precise estimation of whether mutation pressure or natural selection was the most significant factor. From neutrality plots for all genera (Fig. 5), negative correlation between GC3% and GC12% was found, corroborating that natural selection was significant factor to evolutionary changes and CUB for MT-COI gene. However, the exception was in genus Cerithidea, where positive correlation between GC3% and GC12% was observed suggesting that other evolutionary forces like mutation pressure, genetic drift and non-random mating were contributing factors to evolutionary changes and CUB for MT-COI gene. Discussion Narrow GC content, which was less than 50%, further suggests that natural selection was the most significant factor for codon usage pattern of MT-COI gene in Potamidids. In support of these findings, other studies in reptiles (Chakraborty et al. 2022) and amphibians (P. A. Barbhuiya et al. 2021) concluded that natural selection and mutation pressure were the major and minor evolutionary forces respectively, responsible for shaping CUB in mitochondrial genes. Furthermore, other compositional features parameters that explain possible codon usage pattern of MT- COI gene among different species of the family Potamididae were found to be interesting. For example, third position of codons was mostly occupied by nucleobase T suggesting that codons ending with T were more preferred as compared to other nucleotides. It is known that third codon position is subjected to silent mutations pressure (Chamary and Hurst 2009; Alberts 2015) therefore, when studied in details such investigations play a vital role in interpretation of codon trend of the various genes. In the present study, silent mutations at third codon positions involving T nucleobase was observed to be common as a result codon ending with T nucleobase were highly favored. Additionally, slight difference was observed in positional GC contents (GC1, GC2 and GC3) among the four genera of Potadimidae (Fig 02). These findings agree with the postulate that compositional properties might have effects on the codon usage pattern of a gene (Deka and Chakraborty 2014; Choudhury et al. 2018; Barbhuiya et al. 2020). Moreover, codon usage bias of MT-COI gene was found to be weak because the ENC values in the present study was greater 35, but was less 50. These findings are supported by earlier investigations which revealed Page 10/20 Page 10/20 low CUB in mitochondrial ATP, ND and CO genes of amphibians, arthropods and reptiles respectively (R. I. Barbhuiya et al. 2019; P. A. Barbhuiya et al. 2021; Chakraborty et al. 2022). The coding sequences of MT-COI gene were evaluated for their relative synonymous codon usage (RSCU). There was variability in RSCU values among and between four genera of Potamididae in the present study. Interestingly, some codons were overrepresented across different genera while others were underrepresented (Table 3). Declarations There is no conflict of interest. There is no conflict of interest. Discussion Specifically, five codons (CUU, GUU, UCU, ACU, and AGG) were found to be commonly overrepresented across all four genera of family Potamididae, and six codons (GGA, CGA, GAG, UAC, CUG, CUA) were commonly underrepresented across the four different genera in the family Potamididae (Table 3). It has been reported that similar codon usage patterns including over and underrepresented codons in genomes ensures translational efficiency (Tuller et al. 2010), thus, the observed similar pattern in codon preference might be due to translational selection. Further, a multivariate statistical approach of correspondence analysis (CoA), utilizing RSCU values, was performed to clearly define the varying trends in codon usage among potamidids. A chunk of codons were situated close to the axes and some concentrated in the center of the plot proposing that nucleotide compositional constraints under mutation pressure might associated with the CUB of MT-COI gene. Earlier investigations on codon usage bias in Harpago chiragra and Lambis lambis (Gastropoda: Stromboidea) also revealed that the mitochondria genes were subjected to different mutational pressures which signified compositional constraints (Jiang et al. 2019). Similar findings were observed in reptiles (Chakraborty et al. 2022), cotton species (Wang et al. 2018), and bivalves (Gerdol et al. 2015). This study was performed to deduce factors that might favor compositional properties of MT-COI gene in family potamidids. Furthermore, the evolutionary pattern of MT-COI gene in the family Potamididae was evaluated. Low level of CUB was observed in MT-COI gene of potamidids. The codon usage pattern was different across the different genera. The two evolutionary forces, mutation pressure and natural selection were found to be significant in determining the codon usage bias of the gene under study. However, natural selection had a dominant role while mutation pressure had a submissive role in influencing codon usage bias of MT-COI gene in different genera within the family Potamididae. 219–225. https://doi.org/10.1016/j.sjbs.2011.03.002 219–225. https://doi.org/10.1016/j.sjbs.2011.03.002 4. Barbhuiya, P. A., Uddin, A., & Chakraborty, S. (2021). Codon usage pattern and evolutionary forces of mitochondrial ND genes among orders of class Amphibia. Journal of Cellular Physiology, 236(4), 2850–2868. https://doi.org/10.1002/jcp.30050 5. Barbhuiya, R. I., Uddin, A., & Chakraborty, S. (2019). Compositional properties and codon usage pattern of mitochondrial ATP gene in different classes of Arthropoda. Genetica, 147(3–4), 231–248. https://doi.org/10.1007/s10709-019-00067-1 6. Barbhuiya, R. I., Uddin, A., & Chakraborty, S. (2020). Codon usage pattern and its influencing factors for mitochondrial CO genes among different classes of Arthropoda. Mitochondrial DNA Part A, 31(8), 313–326. https://doi.org/10.1080/24701394.2020.1800661 7. Behura, S. K., & Severson, D. W. (2013). Codon usage bias: causative factors, quantification methods and genome-wide patterns: with emphasis on insect genomes. Biological Reviews of the Cambridge Philosophical Society, 88(1), 49–61. https://doi.org/10.1111/j.1469-185X.2012.00242.x 8. Bulmer, M. (1991). The selection-mutation-drift theory of synonymous codon usage. Genetics, 129(3), 897–907. https://doi.org/10.1093/genetics/129.3.897 9. Cassidy, M. (2020). Biological Evolution: An Introduction. Cambridge University Press. 9. Cassidy, M. (2020). Biological Evolution: An Introduction. Cambridge Uni 10. Castro, J. A., Picornell, A., & Ramon, M. (1998). Mitochondrial DNA: a tool for populational genetics studies. International Microbiology: The Official Journal of the Spanish Society for Microbiology, 1(4), 327–332. 11. Chakraborty, S., Basumatary, P., Nath, D., Paul, S., & Uddin, A. (2022). Compositional features and pattern of codon usage for mitochondrial CO genes among reptiles. Mitochondrion, 62, 111–121. https://doi.org/10.1016/j.mito.2021.11.004 12. Chamary, J. V., & Hurst, L. D. (2009). The PRICE of SILENT MUTATIONS. Scientific American, 300(6), 46–53. 13. Chinnery, P. F., & Hudson, G. (2013). Mitochondrial genetics. British Medical Bulletin, 106(1), 135– 159. https://doi.org/10.1093/bmb/ldt017 14. Choudhury, M. N., Uddin, A., & Chakraborty, S. (2018). Nucleotide composition and codon usage bias of SRY gene. Andrologia, 50(1), e12787. https://doi.org/10.1111/and.12787 15. Crimi, M., & Rigolio, R. (2008). The mitochondrial genome, a growing interest inside an organelle. International Journal of Nanomedicine, 3(1), 51–57. 16. Czech, A., Fedyunin, I., Zhang, G., & Ignatova, Z. (2010). Silent mutations in sight: co-variations in tRNA abundance as a key to unravel consequences of silent mutations. Molecular BioSystems, 6(10), 1767–1772. https://doi.org/10.1039/C004796C 16. Czech, A., Fedyunin, I., Zhang, G., & Ignatova, Z. (2010). Silent mutations in sight: co-variations in tRNA abundance as a key to unravel consequences of silent mutations. Molecular BioSystems, 6(10), 1767–1772. https://doi.org/10.1039/C004796C 17. De Mandal, S., Mazumder, T. H., Panda, A. K., Kumar, N. S., & Jin, F. (2020). References 1. Adams, J. C., & Engel, J. (2021). Biological Evolution. In J. C. Adams & J. Engel (Eds.), Life and Its Future (pp. 35–55). Cham: Springer International Publishing. https://doi.org/10.1007/978-3-030- 59075-8_4 1. Adams, J. C., & Engel, J. (2021). Biological Evolution. In J. C. Adams & J. Engel (Eds.), Life and Its Future (pp. 35–55). Cham: Springer International Publishing. https://doi.org/10.1007/978-3-030- 59075-8_4 1. Adams, J. C., & Engel, J. (2021). Biological Evolution. In J. C. Adams & J. Engel (Eds.), Life and Its Future (pp. 35–55). Cham: Springer International Publishing. https://doi.org/10.1007/978-3-030- 59075-8_4 rts, B. (2015). Molecular Biology of the Cell. Garland Science, Taylor and Fra Page 11/20 3. Arif, I. A., Khan, H. A., Bahkali, A. H., Al Homaidan, A. A., Al Farhan, A. H., Al Sadoon, M., & Shobrak, M. (2011). DNA marker technology for wildlife conservation. Saudi Journal of Biological Sciences, 18(3), Page 11/20 3. Arif, I. A., Khan, H. A., Bahkali, A. H., Al Homaidan, A. A., Al Farhan, A. H., Al Sadoon, M., & Shobrak, M. (2011). DNA marker technology for wildlife conservation. Saudi Journal of Biological Sciences, 18(3), Page 11/20 219–225. https://doi.org/10.1016/j.sjbs.2011.03.002 Analysis of synonymous codon usage patterns of HPRT1 gene across twelve mammalian species. Genomics, 112(1), 304– 311. https://doi.org/10.1016/j.ygeno.2019.02.010 17. De Mandal, S., Mazumder, T. H., Panda, A. K., Kumar, N. S., & Jin, F. (2020). Analysis of synonymous codon usage patterns of HPRT1 gene across twelve mammalian species. Genomics, 112(1), 304– 311. https://doi.org/10.1016/j.ygeno.2019.02.010 Page 12/20 18. Deb, B., Uddin, A., & Chakraborty, S. (2020). Codon usage pattern and its influencing factors in different genomes of hepadnaviruses. Archives of Virology, 165(3), 557–570. https://doi.org/10.1007/s00705-020-04533-6 19. Deka, H., & Chakraborty, S. (2014). Compositional Constraint Is the Key Force in Shaping Codon Usage Bias in Hemagglutinin Gene in H1N1 Subtype of Influenza A Virus. International Journal of Genomics, 2014, e349139. https://doi.org/10.1155/2014/349139 20. Duret, L. (2002). Evolution of synonymous codon usage in metazoans. Current Opinion in Genetics & Development, 12(6), 640–649. https://doi.org/10.1016/s0959-437x(02)00353-2 21. Ferguson-Miller, S. (2013). Cytochrome c Oxidase (Complex IV). In G. C. K. Roberts (Ed.), Encyclopedia of Biophysics (pp. 423–426). Berlin, Heidelberg: Springer. https://doi.org/10.1007/978- 3-642-16712-6_29 22. Fowler, S., Roush, R., & Wise, J. (2018). Concepts of Biology. Samurai Media Limited. 23. Friedman, J. R., & Nunnari, J. (2014). Mitochondrial form and function. Nature, 505(7483), 335–343. https://doi.org/10.1038/nature12985 24. Gerdol, M., Moro, G. D., Venier, P., & Pallavicini, A. (2015). Analysis of synonymous codon usage patterns in sixty-four different bivalve species. PeerJ, 3, e1520. https://doi.org/10.7717/peerj.1520 25. Gun, L., Yumiao, R., Haixian, P., & Liang, Z. (2018). Comprehensive Analysis and Comparison on the Codon Usage Pattern of Whole Mycobacterium tuberculosis Coding Genome from Different Area. BioMed Research International, 2018, 3574976. https://doi.org/10.1155/2018/3574976 26. He, B., Dong, H., Jiang, C., Cao, F., Tao, S., & Xu, L. (2016). Analysis of codon usage patterns in Ginkgo biloba reveals codon usage tendency from A/U-ending to G/C-ending. Scientific Reports, 6(1), 35927. https://doi.org/10.1038/srep35927 27. Hopcroft, R. L. (2018). The Oxford Handbook of Evolution, Biology, and Society. Oxford University Press. 28. Jiang, D., Zheng, X., Zeng, X., Kong, L., & Li, Q. (2019). The complete mitochondrial genome of Harpago chiragra and Lambis lambis (Gastropoda: Stromboidea): implications on the Littorinimorpha phylogeny. Scientific Reports, 9(1), 17683. https://doi.org/10.1038/s41598-019- 54141-x 29. Klasen, M., & Wabl, M. (2004). Silent point mutation in DsRed resulting in enhanced relative fluorescence intensity. BioTechniques, 36(2), 236–238. https://doi.org/10.2144/04362BM06 29. Klasen, M., & Wabl, M. (2004). Silent point mutation in DsRed resulting in enhanced relative fluorescence intensity. BioTechniques, 36(2), 236–238. https://doi.org/10.2144/04362BM06 30. Mitra, S., Ray, S. K., & Banerjee, R. (2016). 219–225. https://doi.org/10.1016/j.sjbs.2011.03.002 Synonymous codons influencing gene expression in organisms. Research and Reports in Biochemistry, 6, 57–65. https://doi.org/10.2147/RRBC.S83483 30. Mitra, S., Ray, S. K., & Banerjee, R. (2016). Synonymous codons influencing gene expression in organisms. Research and Reports in Biochemistry, 6, 57–65. https://doi.org/10.2147/RRBC.S83483 31. Nabholz, B., Mauffrey, J.-F., Bazin, E., Galtier, N., & Glemin, S. (2008). Determination of Mitochondrial Genetic Diversity in Mammals. Genetics, 178(1), 351. https://doi.org/10.1534/genetics.107.073346 31. Nabholz, B., Mauffrey, J.-F., Bazin, E., Galtier, N., & Glemin, S. (2008). Determination of Mitochondrial Genetic Diversity in Mammals. Genetics, 178(1), 351. https://doi.org/10.1534/genetics.107.073346 32. Nasrullah, I., Butt, A., Tahir, S., Idrees, M., & Tong, Y. (2015). Genomic analysis of codon usage shows influence of mutation pressure, natural selection, and host features on Marburg virus evolution. BMC Evolutionary Biology, 15. https://doi.org/10.1186/s12862-015-0456-4 32. Nasrullah, I., Butt, A., Tahir, S., Idrees, M., & Tong, Y. (2015). Genomic analysis of codon usage shows influence of mutation pressure, natural selection, and host features on Marburg virus evolution. BMC Evolutionary Biology, 15. https://doi.org/10.1186/s12862-015-0456-4 Page 13/20 Page 13/20 33. Nath Choudhury, M., Uddin, A., & Chakraborty, S. (2017). Codon usage bias and its influencing factors for Y-linked genes in human. Computational Biology and Chemistry, 69, 77–86. https://doi.org/10.1016/j.compbiolchem.2017.05.005 34. Nei, P. of P. G. B. of the C. for D. and P. G. M., Nei, M., Kumar, S., & Nei, E. P. P. of B. M. (2000). Molecular Evolution and Phylogenetics. Oxford University Press. 35. Osellame, L. D., Blacker, T. S., & Duchen, M. R. (2012). Cellular and molecular mechanisms of mitochondrial function. Best Practice & Research. Clinical Endocrinology & Metabolism, 26(6), 711– 723. https://doi.org/10.1016/j.beem.2012.05.003 35. Osellame, L. D., Blacker, T. S., & Duchen, M. R. (2012). Cellular and molecular mechanisms of mitochondrial function. Best Practice & Research. Clinical Endocrinology & Metabolism, 26(6), 711– 723. https://doi.org/10.1016/j.beem.2012.05.003 36. Parvathy, S. T., Udayasuriyan, V., & Bhadana, V. (2022). Codon usage bias. Molecular Biology Reports, 49(1), 539–565. https://doi.org/10.1007/s11033-021-06749-4 36. Parvathy, S. T., Udayasuriyan, V., & Bhadana, V. (2022). Codon usage bias. Molecular Biology Reports, 49(1), 539–565. https://doi.org/10.1007/s11033-021-06749-4 37. Ratsimbazafy, H. A., & Kochzius, M. (2018). Restricted gene flow among western Indian Ocean populations of the mangrove whelk Terebralia palustris (Linnaeus, 1767) (Caenogastropoda: Potamididae). Journal of Molluscan Studies, 84(2), 163–169. https://doi.org/10.1093/mollus/eyy001 38. Reid, D. G., Dyal, P., Lozouet, P., Glaubrecht, M., & Williams, S. T. (2008). Mudwhelks and mangroves: The evolutionary history of an ecological association (Gastropoda: Potamididae). Molecular Phylogenetics and Evolution, 47(2), 680–699. https://doi.org/10.1016/j.ympev.2008.01.003 39. 219–225. https://doi.org/10.1016/j.sjbs.2011.03.002 Ruse, M., & Travis, J. (2009). EVOLUTION: THE FIRST FOUR BILLION YEARS. Harvard University Press. 40. Salim, H. M. W., & Cavalcanti, A. R. O. (2008). Factors influencing codon usage bias in genomes. Journal of the Brazilian Chemical Society, 19, 257–262. https://doi.org/10.1590/S0103- 50532008000200008 41. Shafat, Z., Ahmed, A., Parvez, M. K., & Parveen, S. (2022). Decoding the codon usage patterns in Y- domain region of hepatitis E viruses. Journal of Genetic Engineering and Biotechnology, 20(1), 56. https://doi.org/10.1186/s43141-022-00319-2 42. Shen, Z., Gan, Z., Zhang, F., Yi, X., Zhang, J., & Wan, X. (2020). Analysis of codon usage patterns in citrus based on coding sequence data. BMC Genomics, 21(Suppl 5). https://doi.org/10.1186/s12864-020-6641-x 43. Shimada, S., Shinzawa‐Itoh, K., Baba, J., Aoe, S., Shimada, A., Yamashita, E., et al. (2017). Complex structure of cytochrome c –cytochrome c oxidase reveals a novel protein–protein interaction mode. The EMBO Journal, 36(3), 291–300. https://doi.org/10.15252/embj.201695021 44. Souza, H. V., Marchesin, S. R. C., & Itoyama, M. M. (2016). Analysis of the mitochondrial COI gene and its informative potential for evolutionary inferences in the families Coreidae and Pentatomidae (Heteroptera). Genetics and molecular research: GMR, 15(1). https://doi.org/10.4238/gmr.15017428 45. Subramanian, S. (2008). Nearly Neutrality and the Evolution of Codon Usage Bias in Eukaryotic Genomes. Genetics, 178(4), 2429–2432. https://doi.org/10.1534/genetics.107.086405 46. Sun, X., Yang, Q., & Xia, X. (2013). An Improved Implementation of Effective Number of Codons (Nc). Molecular Biology and Evolution, 30(1), 191–196. https://doi.org/10.1093/molbev/mss201 Page 14/20 Page 14/20 47. Townsend, J. P., Su, Z., & Tekle, Y. I. (2012). Phylogenetic Signal and Noise: Predicting the Power of a Data Set to Resolve Phylogeny. Systematic Biology, 61(5), 835. https://doi.org/10.1093/sysbio/sys036 48. Tuller, T., Waldman, Y. Y., Kupiec, M., & Ruppin, E. (2010). Translation efficiency is determined by both codon bias and folding energy. Proceedings of the National Academy of Sciences of the United States of America, 107(8), 3645–3650. https://doi.org/10.1073/pnas.0909910107 49. Wang, L., Xing, H., Yuan, Y., Wang, X., Saeed, M., Tao, J., et al. (2018). Genome-wide analysis of codon usage bias in four sequenced cotton species. PLOS ONE, 13(3), e0194372. https://doi.org/10.1371/journal.pone.0194372 50. Wells, F. E., & Keesing, J. K. (2019). Growth rates of potamidid snails in mangroves in northern Australia. Molluscan Research, 39(4), 333–340. https://doi.org/10.1080/13235818.2019.1638484 51. Willan, R. C. (n.d.). A key to the potamidid snails (longbums, mudcreepers and treecreepers) of northern Australia, 13. 52. Wright, F. (1990). The “effective number of codons” used in a gene. Gene, 87(1), 23–29. https://doi.org/10.1016/0378-1119(90)90491-9 53. 219–225. https://doi.org/10.1016/j.sjbs.2011.03.002 Xu, C., Dong, J., Tong, C., Gong, X., Wen, Q., & Zhuge, Q. (2013). Analysis of Synonymous Codon Usage Patterns in Seven Different Citrus Species. Evolutionary Bioinformatics, 9, EBO.S11930. https://doi.org/10.4137/EBO.S11930 54. Yang, X., Luo, X., & Cai, X. (2014). Analysis of codon usage pattern in Taenia saginata based on a transcriptome dataset. Parasites & Vectors, 7(1), 527. https://doi.org/10.1186/s13071-014-0527-1 54. Yang, X., Luo, X., & Cai, X. (2014). Analysis of codon usage pattern in Taenia saginata based on a transcriptome dataset. Parasites & Vectors, 7(1), 527. https://doi.org/10.1186/s13071-014-0527-1 55. Yu, X., Liu, J., Li, H., Liu, B., Zhao, B., & Ning, Z. (2021). Comprehensive Analysis of Synonymous Codon Usage Bias for Complete Genomes and E2 Gene of Atypical Porcine Pestivirus. Biochemical Genetics, 59(3), 799–812. https://doi.org/10.1007/s10528-021-10037-y 55. Yu, X., Liu, J., Li, H., Liu, B., Zhao, B., & Ning, Z. (2021). Comprehensive Analysis of Synonymous Codon Usage Bias for Complete Genomes and E2 Gene of Atypical Porcine Pestivirus. Biochemical Genetics, 59(3), 799–812. https://doi.org/10.1007/s10528-021-10037-y 56. Zhou, Z., Dang, Y., Zhou, M., Li, L., Yu, C., Fu, J., et al. (2016). Codon usage is an important determinant of gene expression levels largely through its effects on transcription. Proceedings of the National Academy of Sciences, 113(41), E6117–E6125. https://doi.org/10.1073/pnas.1606724113 56. Zhou, Z., Dang, Y., Zhou, M., Li, L., Yu, C., Fu, J., et al. (2016). Codon usage is an important determinant of gene expression levels largely through its effects on transcription. Proceedings of the National Academy of Sciences, 113(41), E6117–E6125. https://doi.org/10.1073/pnas.1606724113 Figures Figures Page 15/20 Figure 1 Overall nucleotide composition for MT-COI gene in different genus of potamidids Page 16/20 Figure 2 Overall GC content and positional GC content of MT-COI gene among the different genera of Potamidids. Page 17/20 Page 17/20 Page 17/20 igure 3 orrespondence analysis (CoA) of codon usage bias in MT-COI gene for Cerithidea (a); Cerithideopsis (b Cerithideopsilla (c) and Terebralia (d). Figure 3 Correspondence analysis (CoA) of codon usage bias in MT-COI gene for Cerithidea (a); Cerithideopsis (b); Cerithideopsilla (c) and Terebralia (d). Page 18/20 Figure 4 Parity plot analysis of MT-COI gene in the six genera of the family Potamididae. Page 19/20 gure 5 eutrality plot between GC12% and GC3% in MT-COI gene for a) Cerithideopsis b) Cerithideopsilla c) Cerithidea d) Terebralia. Figure 5 Neutrality plot between GC12% and GC3% in MT-COI gene for a) Cerithideopsis b) Cerithideopsilla c) Cerithidea d) Terebralia. Neutrality plot between GC12% and GC3% in MT-COI gene for a) Cerithideopsis b) Cerithideopsilla c) Cerithidea d) Terebralia. Page 20/20
W2270177489.txt	https://www.econstor.eu/bitstream/10419/198411/1/ceswp-v07-i2a-p605-612.pdf	en		PATTERNS OF THE AGRICULTURAL INCOME AND IMPACT OF STRUCTURAL CHANGES POST-ENLARGEMENT AMONG EU STATES	DOAJ (DOAJ: Directory of Open Access Journals)	2,015	cc-by	2,576	Rusali, Mirela-Adriana Article Patterns of the Agricultural Income and Impact of Structural Changes Post-Enlargement Among EU States CES Working Papers Provided in Cooperation with: Centre for European Studies, Alexandru Ioan Cuza University Suggested Citation: Rusali, Mirela-Adriana (2015) : Patterns of the Agricultural Income and Impact of Structural Changes Post-Enlargement Among EU States, CES Working Papers, ISSN 2067-7693, Alexandru Ioan Cuza University of Iasi, Centre for European Studies, Iasi, Vol. 7, Iss. 2a, pp. 605-612 This Version is available at: https://hdl.handle.net/10419/198411 Standard-Nutzungsbedingungen: Terms of use: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Documents in EconStor may be saved and copied for your personal and scholarly purposes. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. https://creativecommons.org/licenses/by/4.0/ If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. CES Working Papers – Volume VII, Issue 2A PATTERNS OF THE AGRICULTURAL INCOME AND IMPACT OF STRUCTURAL CHANGES POST-ENLARGEMENT AMONG EU STATES Mirela-Adriana RUSALI* Abstract: The article presents the results of analysis on the pattern of agricultural income structure and changes in EU Member States related to the enlargement impact, highlighting gaps and major contributions. While marked differences appear between the EU-15 and post 2004 acceded countries, the major contribution of the Romanian agriculture to the EU economy and the proportion of employment in agriculture, are indicators of a divergent sector development. With a substantial contribution to the UAA, i.e. the 7.9% share in arable land places Romania among the first 6 countries, while the 8.3% share in the area of permanent pasture reveal a potential advantage for livestock grazing. Nevertheless agriculture remaining the foremost source of income for rural households in Romania, its importance has been evidenced a decreasing trend in the post-accession period of time 2007-2013, by comparing to the previous period of time 2001-2006. Keywords: agricultural income; structural changes; post-enlargement; Romania JEL Classification: Q17; Q12; F15 Introduction Agricultural income is an important indicator on the agricultural sector‘s viability of each state and grounds the debates on perspectives regarding the economic and social policies. There is a wide range of driven factors affecting the trends of income produced by the agricultural sectors across the EU region, depending on each agricultural production systems that make the impact very much regional specific. The potential impact of the uncertainties is related mainly to price developments of input costs on commodity balances. Production systems highly dependent of inputs use are the most exposed to price changes. This is because crop production systems need higher input use than livestock systems, therefore are affected more from a negative impact as a result of the fact that greater costs are only in part transferred to higher producer price (EC, 2012). As well, in addition to potential impact, at regional level on farmers’ income, of the uncertainties related to variability of input costs, might be noted those related to climate changes variables on the agricultural sector and different EU bio-fuel policies on feedstock markets (EC, 2012). The article presents the results of analysis on the pattern of agricultural income structure and changes in EU Member States, highlighting gaps and major contributions across the region related to the enlargement impact. * Senior scientific researcher, Institute of Agricultural Economics, Romanian Academy, e-mail: m.rusali@yahoo.com 605 PATTERNS OF THE AGRICULTURAL INCOME AND IMPACT OF STRUCTURAL CHANGES 1. Methods of assessing agricultural income Agricultural income comprises the income generated by agricultural activities, including nonagricultural, secondary activities, over a certain period. Moreover, it is a different indicator than the total income of farming households as it does not comprise income from other sources e.g. nonagricultural activities, salaries, social benefits, income from property. The estimates provided by Eurostat, i.e. calculated aggregate figures, have been compiled by the EU Member States according to the methodology of the Economic Accounts for Agriculture, as indicated by the Regulation (EC) No. 138/2004, closed to the methodology of the national accounts but incorporates a number of changes to take account of the special features of the agricultural economy (EC, 2014). The real income of factors in agriculture, per annual work unit, corresponds to the real net value added at factor cost of agriculture, per total annual work unit. At the same time, the net value added at factor cost is calculated by subtracting from the value of agricultural output at basic prices the value of intermediate consumption, the consumption of fixed capital and production taxes, and adding the value of production subsidies. In order to take account of part-time and seasonal work, agricultural labour or their changes, are measured in annual work units (AWU), defined as work-time equivalent of a full-time worker. The agricultural income per worker is calculated as factor income per AWU, in real terms. It corresponds to the net value added at factor cost of agriculture, per annual work unit and deflated by the implicit GDP price index. This measure is commonly expressed as an index of the real income of factors in agricultural per annual work unit. 2. Analysis on the pattern of agricultural income in EU and background factors Against the background of real agricultural income per worker, marked differences appear between the EU-15 and the newly accessed countries aggregates post 2004. Agricultural income per worker in real terms decreased on average by 1.7% in 2014 in the whole region comparing to the previous year, although about 33% higher than in the crisis year 2009. The reduction in average income is stronger in the EU-15 (-2.4%) than in the countries that accessed the EU since 2004 (-0.4%) (EC, 2014). 606 Mirela-Adriana RUSALI The income development per worker in 2014 reflects a 4% decrease in factor income in the EU28, which is similar in the EU-15 (-3.9%) and the EU-N12 (-4.3%), combined with an average reduction in labor input by 2.3% which is more marked in the EU-N13 (-3.4%, including Croatia) than in the EU-15 (-1.3%). Similarly, the agricultural labour force keeps declining more strongly in the enlarged EU area than in the EU-15.Results by Member State differ substantially, both in terms of changes from 2013 to 2014 and in terms of the absolute index value. Agricultural income per worker declined in over 70% of the EU-28 in 2014 (20 countries). As indicated in fig. 1, the highest decrease (-22.8%), was in Finland, mostly because significant decrease of animal output value (-17.7%), and less to the decline in the value of milk production (-2.7%) when real milk prices drop by 5.5%. Figure 1- Change in agricultural income per worker in the EU-28 (2013-2014) (%) Source: EC, 2014 Lithuania (-19.4%) and Belgium (-15.2%) also experienced important diminish in agricultural income per worker, for different reasons: in Lithuania declined the value of both crop (-9.4%) and animal output (-7.2%), combined with a 10.5% decrease in subsidies and a 5% growth of fixed capital consumption, while in Belgium the most affected was the value of crop output (-13.8%). 607 PATTERNS OF THE AGRICULTURAL INCOME AND IMPACT OF STRUCTURAL CHANGES The highest income increase was estimated in Slovenia (13.3%), due to a substantial increase in the output value of animal products (11.8%) and a drop in production costs of 5.8%. As well, Hungary (9.1%), the Czech Republic (7.2%) and the United Kingdom (6.9%) performed strong increases. Only 8 countries shown, however, an increase compared to 2013 values, while Romania experienced the leased decrease (-0.3%). The output value and production costs are the most variable elements in estimating the agricultural income. The overall agricultural output value in the EU-28 fell by 3.5% between 2013 and 2014, mostly due to a reduction in crop output value. Cumulatively, Romania's contribution to the total value of agricultural production in EU-27, obtained in 2013 was of 3% in crop production, respectively, of 1% in animal production. Income sources provided by the agricultural structural indicators related to the land utilization are shown by the cropping pattern across the EU countries. The fig. 2 indicates the diverse structure of the utilized agriculture area (UAA) in EU. Figure 2 – Cropping pattern in EU-28, by share of main land types in UAA 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% Cereals Temporary grass Other arable crops Kitchen gardens Permanent grasslands Permanent crops CH ME IS NO UK FI SE SI SK RO PL PT AT NL MT HU LT LU LV IT CY FR HR ES IE EL EE DE CZ DK BE BG EU-28 0% Source: Eurostat (FSS 2010) Of the 175.8 million hectares in EU-28, the arable land covers over half (58%), represented by cereals (32%), temporary grass (6%) and a quite important share of other crops (20%), while permanent grassland is the most spread (35%), followed by a less share of permanent crops in the UAA (6%) and the kitchen gardens (0.2%). The cereals areas are wider spread in France (33%) and 608 Mirela-Adriana RUSALI Spain (26%); temporary grass areas in Sweden (37%) and Finland (28%); permanent grasslands in Ireland (65%) and UK (59%); and permanent crops are mostly in Cyprus (26%), Portugal and Italy (19% each). The EU-28 covers 79.4 million hectares of fodder areas, that is 45% in UAA. The cropping pattern depicted by the shares of fodder area in the UAA (fig. 3) reveal that the greatest shares are in Ireland (94%) and UK (73%), and the least are in Denmark (27%) and Bulgaria (29%), while although below the EU average is Romania (36%) with 4.7 million hectares. Figure 3 – Cropping pattern in EU-28, by share of fodder area in UAA 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% IE UK LU SI NL PT AT BE SE EL MT FR LV EU-28 DE EE SK IT LT CZ ES RO HR FI CY BG DK PL HU 0% Source: Eurostat (FSS 2010) The fig. 4 presents the changes in the UAA in EU in average and by Member States, in period 2005-2010, compared to the changes of the fodder areas. Figure 4 – Change in UAA and fodder area in EU-27 (2005-2010) 30% 20% 10% AT CY SI EL ES NL SE RO DK BE EU-15 LT FR CZ EU-27 DE UK PT PL FI IT SK LV LU IE MT EE HU BG 0% -10% -20% -30% UAA Fodder a rea Source: EC, 2013 Despite the decreasing trends, Romania numbers among the EU countries with a substantial contribution to the UAA. Following the structure of the agricultural land, with 8.3% share they 609 PATTERNS OF THE AGRICULTURAL INCOME AND IMPACT OF STRUCTURAL CHANGES occupy in the area of permanent pasture, EU gives to Romania a potential advantage for livestock grazing. Also, as of 2013, the 7.9% share in arable land places Romania among the first 6 countries, after France, Spain, Italy, Germany and Poland. While the 8 billion of Gross Value Added (GVA) places Romania among the first 10 states, after France, Italy, Spain, Germany, Poland, UK and the Netherlands (Table 1), the labour productivity in the sector stands at a modest value of 5 thousand Euro per annual work unit (MADR, 2013). Table 1 – GVA in primary sector* in EU-27 and shares in Member States economy (2013) GVA in Primary sector (Million Euro) 194664 33198 30045 24109 19060 13115 10813 8892 7964 5931 5588 4696 4366 3944 3502 3181 2815 2733 2691 1939 1696 1199 1021 885 629 403 139 99 European Union France Italy Spain Germany Poland United Kingdom Netherlands Romania Greece Sweden Finland Austria Hungary Portugal Czech Republic Denmark Ireland Belgium Slovakia Bulgaria Lithuania Latvia Slovenia Estonia Cyprus Luxembourg Malta GVA in Primary sector share of country's total (%) 1.7 1.8 2.1 2.6 0.8 3.8 0.6 1.6 6.4 3.7 1.5 2.8 1.5 4.8 2.4 2.4 1.3 1.9 0.8 3.0 4.9 3.8 4.9 2.9 3.9 2.7 0.3 1.6 *) agriculture, forestry and fishery Source: Author’s processing using Eurostat data The results of evaluations based on Eurostat statistics show a 4% share of the Romanian agriculture in GVA achieved in average during 2001-2013, amounting to 7 billion Euro, in the EU27 agricultural sector. This performance resulted based on increasing the share of crop production 610 Mirela-Adriana RUSALI value, from 3.9% in 2001 to 5.7% in the value of crop production in the EU-27 produced in 2013. At the same time, the share of animal production in GVA decreased from 2.7% to 2.3%. However, the major contribution of the Romanian agriculture in the EU economy and the proportion of employment in agriculture, are indicators of a divergent range of Romania toward the development of the sector in most Member States. From the data provided by the Romanian Statistical Yearbook (table 2), Romanian average farmers’ income represented 60% of employees’ income in the period 2001-2013. Table 2 – Structure and changes of farm households’ income in Romania (2001-2013) 2001-2006 2007-2013 Total income of farmers (lei, monthly per household) 433 2098 Money income (%), of which: 42.3 53.1 6.5 8.4 23.4 26.1 2.1 3.7 7.6 9.0 57.5 46.0 Gross salaries and other salary rights Income from agriculture Income from non-agricultural independent activities Income from social provisions Equivalent value of consumption of agricultural products from own resources (%) Source: NIS, 2014. The estimations of the average subsistence equivalent income from the consumption of own agricultural products accounted for 47%. The main source of money income came from agriculture, accounting for 24%, while 11.6% from social provisions, 7.8% from salaries and only 2.9% from non-agricultural independent activities, though a slight increasing trend has been observed since the year 2007. Summary remarks Against the background of real agricultural income per worker, marked differences appear between the EU-15 and the newly accessed countries aggregates post 2004. Income sources provided by the agricultural structural indicators related to the land utilization are shown by the cropping pattern across the EU countries. The arable land and the fodder areas cover the major part of EU’s UAA, while the crop output evolution had the greatest influence on the value of agricultural income. 611 PATTERNS OF THE AGRICULTURAL INCOME AND IMPACT OF STRUCTURAL CHANGES However, the major contribution of the Romanian agriculture in the EU economy and the proportion of employment in agriculture, are indicators of a divergent range of Romania toward the development of the sector in most Member States. The importance of agriculture, as a foremost source of income for agricultural households in Romania, was estimated at 80.9%, in average, in the period of time 2001-2006, while, with a decreasing trend, at 72.1% in the post-accession period of time 2007-2013. References EC, (2014), “EU agricultural income, 2014”, European Commission - DG Agriculture and Rural Development, available at: http://ec.europa.eu/agriculture/index_en.htm EC, (2013), “Agriculture, forestry and fishery statistics”, Eurostat, European Commission, Printed in Luxembourg, available at: http://ec.europa.eu/eurostat/documents. EC, (2012), “Prospects for Agricultural Markets and Income in the EU, 2012-2020”, European Commission, DG Agriculture and Rural Development, Brussels. MADR, (2013), “Analiza socio-economica in perspectiva dezvoltarii rurale, 2014-2020”, PNDR, Ministerul Agriculturii si Dezvoltarii Rurale, available at: http://www.madr.ro. NIS, (2014), “Romanian Statistical Yearbook”, National Institute of Statistics, Romania. 612
https://openalex.org/W4285083438	https://www.researchsquare.com/article/rs-1808867/latest.pdf	English	null	Circular economy for food industry waste: Development and characterization of spray-dried acid whey encapsulated in millet matrix	Research Square (Research Square)	2,022	cc-by	19,687	Circular economy for food industry waste: Development and characterization of spray-dried acid whey encapsulated in millet matrix Mercy Nani University of Missouri Columbia Kiruba Krishnaswamy (  krishnaswamyk@umsystem.edu ) University of Missouri Columbia Research Article Keywords: Acid Whey, Millets, Spray Drying, Microencapsulation, Antioxidant Activity, Glass Transition Temperature Posted Date: July 13th, 2022 DOI: https://doi.org/10.21203/rs.3.rs-1808867/v1 License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License License:   This work is licensed under a Creative Commons Attribution 4.0 International License. R d F ll Li Circular economy for food industry waste: Development and characterization of spray-dried 1 acid whey encapsulated in millet matrix 2 Mercy Nani1 and Kiruba Krishnaswamy a,b 3 aDivision of Food, Nutrition and Exercise Science, University of Missouri Columbia, MO, USA 4 bDepartment of Biomedical, Biological and Chemical Engineering, University of Missouri, 5 Columbia, MO, USA 6 7 Corresponding Author: Kiruba Krishnaswamy, 244, Agriculture Engineering Building, 8 Columbia, MO, 65201, krishnaswamyk@umsystem.edu 9 10 11 Abstract 12 Acid whey (AW) has become a significant challenge for the dairy industry because of the large 13 quantity being produced and the difficulty in handling and disposing of it. Upcycling AW and 14 reintroducing it into the food supply chain are innovative and sustainable ways to deal with the 15 problem. This study investigated the feasibility of spray drying AW using millet flour as 16 encapsulating wall material and examined the effect of different millet concentrations on the 17 physio functional properties of the acid whey millet (AWM) powders. Higher yields and whiter- 18 colored powders were obtained for AWM powders prepared with 10% and 25% millet flours. The 19 moisture content and water activity ranged from 3.78% to 5.53 % and 0.29 to 0.25, respectively, 20 indicating microbial stability of the powders. All powder formulations exhibited good flow 21 properties. AWM powders exhibited higher solubility and dispersibility values than AW powders. 22 Larger particles were observed in the powders prepared with an AWM matrix. The addition of 23 millet flour increased the glass transition temperature resulting in reduced stickiness of the 24 powders. The reduced stickiness was attributed to the neutralization and microencapsulating effect 25 the millet flour had on AW, which aided the spray drying of AW and reduced stickiness in the 26 powders. AWM powders had a high concentration of phenolic compounds and exhibited high 27 antioxidant activity. Overall, the results of this study indicate that millet flour can be used as a wall 28 material for spray drying of AWM powders, thereby creating a functional and value-added food 29 ingredient. Research Article 30 Circular economy for food industry waste: Development and characterization of spray-dr 1 acid whey encapsulated in millet matrix 2 Mercy Nani1 and Kiruba Krishnaswamy a,b 3 aDivision of Food, Nutrition and Exercise Science, University of Missouri Columbia, MO, US 4 bDepartment of Biomedical, Biological and Chemical Engineering, University of Missouri, 5 Columbia, MO, USA 6 7 Corresponding Author: Kiruba Krishnaswamy, 244, Agriculture Engineering Building, 8 Columbia, MO, 65201, krishnaswamyk@umsystem.edu 9 10 11 Abstract 12 Acid whey (AW) has become a significant challenge for the dairy industry because of the la 13 quantity being produced and the difficulty in handling and disposing of it. Upcycling AW 14 reintroducing it into the food supply chain are innovative and sustainable ways to deal with 15 problem. This study investigated the feasibility of spray drying AW using millet flour 16 encapsulating wall material and examined the effect of different millet concentrations on 17 physio functional properties of the acid whey millet (AWM) powders. Higher yields and whi 18 colored powders were obtained for AWM powders prepared with 10% and 25% millet flours. T 19 moisture content and water activity ranged from 3.78% to 5.53 % and 0.29 to 0.25, respective 20 indicating microbial stability of the powders. All powder formulations exhibited good fl 21 properties. AWM powders exhibited higher solubility and dispersibility values than AW powd 22 Larger particles were observed in the powders prepared with an AWM matrix. The addition 23 millet flour increased the glass transition temperature resulting in reduced stickiness of 24 powders. The reduced stickiness was attributed to the neutralization and microencapsulating eff 25 the millet flour had on AW, which aided the spray drying of AW and reduced stickiness in 26 powders. AWM powders had a high concentration of phenolic compounds and exhibited h 27 antioxidant activity. Overall, the results of this study indicate that millet flour can be used as a w 28 material for spray drying of AWM powders, thereby creating a functional and value-added fo 29 ingredient. 30 1. Introduction 34 According to the current State of Food Security and Nutrition report, the global incidence of hunger 35 and malnutrition remains high, and countries are not on track to reach the United Nations 36 Sustainable Development Goal (SDG) of eradicating all forms of hunger by 2030 (FAO, IFAD, 37 UNICEF, WFP & WHO, 2021). The world's food insecurity problem has been compounded by a 38 growing population, economic and social challenges, climate change impacts, and, most recently, 39 the COVID-19 pandemic (Zhu et al., 2022). As a result, the food system must be transformed into 40 one that meets the needs of people and the planet in a sustainable way. Reducing the astounding 41 cases of food loss and waste across the food supply chain is essential for increasing the food supply 42 and keeping the global food system resilient. Transitioning to a circular food system, where 43 regenerative food production is prioritized, food waste is reduced, repurposed, and nutrients are 44 recycled, is a necessary solution. In addition, the effects of climate change will require the adoption 45 of climate-smart agricultural practices that include the production and consumption of diversified 46 crops (Willett et al., 2019). Mainstreaming the production of neglected and underutilized species 47 will promote food diversification and potentially contribute to alleviating global food insecurity 48 issues (X. Li & Siddique, 2020). 49 The rapid rise in Greek yogurt production has resulted in an increased production of its 50 byproduct – Acid Whey (AW). AW is the byproduct of the mechanical separation of liquid from 51 solid components in the Greek yogurt manufacturing process (Rocha-Mendoza et al., 2021). The 52 increase in AW production is due to the growing popularity of Greek yogurt and processed dairy 53 products among consumers, requiring high amounts of raw milk for production (Rocha-Mendoza 54 et al., 2021). The composition of Greek yogurt acid whey (GAW) varies according to the milk 55 source and processing circumstances. Still, compared to sweet whey, GAW has less lactose and 56 protein, a lower pH, and higher calcium, phosphorus, and lactic acid (Schmidt et al., 1984). 57 Because AW has a high biological oxygen requirement, it is expensive to dispose of, and improper 58 disposal could harm the ecosystem (Menchik et al., 2019). Furthermore, AW is difficult to spray 59 dry because of its low pH, limiting its application. Abstract 12 Acid whey (AW) has become a significant challenge for the dairy industry because of the large 13 quantity being produced and the difficulty in handling and disposing of it. Upcycling AW and 14 reintroducing it into the food supply chain are innovative and sustainable ways to deal with the 15 problem. This study investigated the feasibility of spray drying AW using millet flour as 16 encapsulating wall material and examined the effect of different millet concentrations on the 17 physio functional properties of the acid whey millet (AWM) powders. Higher yields and whiter- 18 colored powders were obtained for AWM powders prepared with 10% and 25% millet flours. The 19 moisture content and water activity ranged from 3.78% to 5.53 % and 0.29 to 0.25, respectively, 20 indicating microbial stability of the powders. All powder formulations exhibited good flow 21 properties. AWM powders exhibited higher solubility and dispersibility values than AW powders. 22 Larger particles were observed in the powders prepared with an AWM matrix. The addition of 23 millet flour increased the glass transition temperature resulting in reduced stickiness of the 24 powders. The reduced stickiness was attributed to the neutralization and microencapsulating effect 25 the millet flour had on AW, which aided the spray drying of AW and reduced stickiness in the 26 powders. AWM powders had a high concentration of phenolic compounds and exhibited high 27 antioxidant activity. Overall, the results of this study indicate that millet flour can be used as a wall 28 material for spray drying of AWM powders, thereby creating a functional and value-added food 29 ingredient. 30 1 1 Keywords: Acid Whey, Millets, Spray Drying, Microencapsulation, Antioxidant Activity, Glass 31 Transition Temperature 32 33 1. Introduction 34 According to the current State of Food Security and Nutrition report, the global incidence of hunger 35 and malnutrition remains high, and countries are not on track to reach the United Nations 36 Sustainable Development Goal (SDG) of eradicating all forms of hunger by 2030 (FAO, IFAD, 37 UNICEF, WFP & WHO, 2021). The world's food insecurity problem has been compounded by a 38 growing population, economic and social challenges, climate change impacts, and, most recently, 39 the COVID-19 pandemic (Zhu et al., 2022). As a result, the food system must be transformed into 40 one that meets the needs of people and the planet in a sustainable way. Abstract 12 Reducing the astounding 41 cases of food loss and waste across the food supply chain is essential for increasing the food supply 42 and keeping the global food system resilient. Transitioning to a circular food system, where 43 regenerative food production is prioritized, food waste is reduced, repurposed, and nutrients are 44 recycled, is a necessary solution. In addition, the effects of climate change will require the adoption 45 of climate-smart agricultural practices that include the production and consumption of diversified 46 crops (Willett et al., 2019). Mainstreaming the production of neglected and underutilized species 47 will promote food diversification and potentially contribute to alleviating global food insecurity 48 issues (X. Li & Siddique, 2020). 49 Th id i i G k t d ti h lt d i i d d ti f it 50 Keywords: Acid Whey, Millets, Spray Drying, Mi 31 Transition Temperature 32 33 1. Introduction 34 According to the current State of Food Security and N 35 and malnutrition remains high, and countries are 36 Sustainable Development Goal (SDG) of eradicati 37 UNICEF, WFP & WHO, 2021). The world's food i 38 growing population, economic and social challenge 39 the COVID-19 pandemic (Zhu et al., 2022). As a re 40 one that meets the needs of people and the planet i 41 cases of food loss and waste across the food supply c 42 and keeping the global food system resilient. Tra 43 regenerative food production is prioritized, food w 44 recycled, is a necessary solution. In addition, the effe 45 of climate-smart agricultural practices that include 46 crops (Willett et al., 2019). Mainstreaming the prod 47 will promote food diversification and potentially c 48 issues (X. Li & Siddique, 2020). 49 Th id i i G k t d ti 50 Keywords: Acid Whey, Millets, Spray Drying, Microencapsulation, Antioxidant Activity, Glass 31 Transition Temperature 32 1. Introduction 34 As a result, the dairy industry is confronted with 60 2 economic and environmental challenges in determining how to adequately repurpose or dispose of 61 AW (Rocha-Mendoza et al., 2021). Therefore, there is a need for innovative and sustainable 62 approaches to dealing with GAW that involves upcycling GAW and reintroducing it into the food 63 supply chain. 64 Millets are highly nutritious small-seeded grains that can thrive in harsh weather conditions 65 and low agricultural potential environments (Taylor, 2018). They are mainly grown and consumed 66 in Africa and Asia's dry and semi-arid tropics, and they can be good alternatives to the more 67 commonly consumed grains such as wheat, rice, and maize (Yousaf et al., 2021). Millets are a 68 good source of nutrients (carbohydrates, proteins, dietary fibers, minerals, and vitamins) and 69 contain significant amounts of amino acids, micronutrients, and phytochemicals (Xiang et al., 70 2019). Studies have shown that millets help maintain healthy blood pressure and blood sugar 71 levels, helping to fight various diseases such as diabetes, cardiovascular disease, blood pressure, 72 thyroid disease, and celiac disease (Saleh et al., 2013). A systematic review of studies on millets 73 and iron bioavailability showed that millet consumption could impact iron status and reduce iron 74 deficiency anemia (Anitha et al., 2021). Although millet is a superior seed grain in terms of 75 nutrition, it is nonetheless underutilized in developed and developing countries due to a lack of 76 awareness. In Africa and Asia, millet is majorly consumed as a traditional staple food. While in 77 most western countries, millet is primarily used as animal feed (Hariprasanna & Rakshit, 2016). 78 Millets play an important role in establishing food and nutrition security as it is critical to 79 producing nutritious, sustainable crop alternatives to address global hunger and the growing grain 80 demand (Kumar et al., 2018). Millets could be processed into value-added nutritious functional 81 products that can be integrated into more mainstream food production due to their high nutritional 82 content (Kumar et al., 2018). 83 Spray drying is a technique used to produce dry powders from a solution or emulsion by 84 atomization of the solution followed by rapid evaporation of the droplets into powder particles 85 (Bhandari et al., 2008). 1. Introduction 34 Optimizing parameters like the inlet temperature, feed rate, atomization 86 pressure, and aspiration rate is crucial to obtaining desired spray-dried powder properties 87 (Bhandari et al., 2008). Because of its low pH and a substantial amount of lactic acid, spray drying 88 AW is problematic, resulting in poor processing conditions such as the stickiness of the spray- 89 dried powder. Manipulating the lactic acid concentrations in AW by neutralizing AW solutions 90 has resulted in powders with improved crystallization properties and enhanced powder yields 91 3 (Chandrapala & Vasiljevic, 2017). The addition of millet (carbohydrate) to AW could aid the spray 92 drying of AW by increasing the pH of the feed solution and acting as an effective carrier and 93 encapsulating wall material for spray drying AW (Botrel et al., 2014; Bylund, 2015; Fernandes et 94 al., 2016). The molecular interactions between the proteins and the particle surface in the acid 95 whey millet matrix (AWM) provide a mechanism for their functionality. The protein molecules 96 form bonds by hydrophobic interactions, leading to low lateral mobility, cohesion formation, and 97 low hygroscopicity of the spray-dried powders (Fang & Bhandari, 2012; Jayasundera et al., 2011; 98 Sarabandi et al., 2018). In this study, we (a) developed a new formulation for acid whey powders 99 using millet for encapsulation, (b) evaluated the functionality of millet flour as an encapsulating 100 agent of AW during spray drying (c) examined the effect of varying acid whey millet (AWM) 101 formulations based on the physical, functional properties and antioxidant properties of the resulting 102 spray-dried powders. 103 (Chandrapala & Vasiljevic, 2017). The addition of millet (carbohydrate) to AW could aid the spray 92 drying of AW by increasing the pH of the feed solution and acting as an effective carrier and 93 encapsulating wall material for spray drying AW (Botrel et al., 2014; Bylund, 2015; Fernandes et 94 al., 2016). The molecular interactions between the proteins and the particle surface in the acid 95 whey millet matrix (AWM) provide a mechanism for their functionality. The protein molecules 96 form bonds by hydrophobic interactions, leading to low lateral mobility, cohesion formation, and 97 low hygroscopicity of the spray-dried powders (Fang & Bhandari, 2012; Jayasundera et al., 2011; 98 Sarabandi et al., 2018). 1. Introduction 34 In this study, we (a) developed a new formulation for acid whey powders 99 using millet for encapsulation, (b) evaluated the functionality of millet flour as an encapsulating 100 agent of AW during spray drying (c) examined the effect of varying acid whey millet (AWM) 101 formulations based on the physical, functional properties and antioxidant properties of the resulting 102 spray-dried powders. 103 low hygroscopicity of the spray-dried powders (Fang & Bhandari, 2012; Jayasundera et al., 2011; 98 Sarabandi et al., 2018). In this study, we (a) developed a new formulation for acid whey powders 99 using millet for encapsulation, (b) evaluated the functionality of millet flour as an encapsulating 100 agent of AW during spray drying (c) examined the effect of varying acid whey millet (AWM) 101 formulations based on the physical, functional properties and antioxidant properties of the resulting 102 spray-dried powders. 103 104 2. Materials and methods 105 2.1. Materials 106 We used available commercial samples of millets for human consumption for this study due to 107 food availability and accessibility. The whole milk used for yogurt production was obtained from 108 a local grocery store (Wal-Mart Stores, Inc., Columbia, MO, USA), and yogurt culture was 109 obtained from FEAST Lab at the University of Missouri (Columbia, MO). It was stored in a freezer 110 at − 20 °C until being used. Barnyard and Little millet were purchased from Manna foods 111 (Chennai, India). A Butterfly Matchless 750-watt mixer grinder was used for milling. The 112 cheesecloth used to strain the yogurt was purchased from Wal-Mart stores (Columbia, MO). 113 114 2.2. Chemicals, solvents, and reagents 115 Acetic acid, gallic acid, and 2,2-diphenyl-1-picrylhydrazyl (DPPH) were purchased from Acros 116 Organics (Morris Plains, NJ). Folin-Ciocalteu and Thiobarbituric acid (TBA) were obtained from 117 Sigma-Aldrich Corp. (St. Louis, MO). Ethanol was purchased from Thermo Fisher Scientific 118 (Hampton, NH). Sodium bicarbonate was purchased from Duda Energy LLC (Decatur, AL). 119 120 2.3. Sample preparation 121 2.3.1. Acid whey 122 2.1. Materials 106 We used available commercial samples of millets for human consumption for this study due to 107 food availability and accessibility. The whole milk used for yogurt production was obtained from 108 a local grocery store (Wal-Mart Stores, Inc., Columbia, MO, USA), and yogurt culture was 109 obtained from FEAST Lab at the University of Missouri (Columbia, MO). 1. Introduction 34 133 AOAC (Association of Official Analytical Chemists) international method was used to determine 134 the total solid content of the soluble fraction (Bradley, 2010). 135 136 2.3.3. Spray drying millet acid whey solution 137 A peristaltic feed pump was used to transfer filtered acid whey millet (AWM) solution into a 138 laboratory-scale spray-dryer (Büchi B290 mini spray-dryer, Flawil, Switzerland). The spray 139 dryer's input temperature, aspirator, feed pump rate, and the gas flow rate was kept at 160°C, 90%, 140 15%, and 45 percent, respectively, based on optimization from prior experiments (S. Malik et al., 141 2021). Before each run, the spray dryer was re-equilibrated with water. The powders were kept at 142 4°C in sealed glass jars until utilized for further analysis. 143 144 2.4. Powder Yield 145 The percentage ratio between the total mass of spray-dried powder recovered and the amount of 146 extract delivered to the system was used to calculate the powder yield (dry basis) (Fang & 147 Bhandari, 2012). 148 Product yield (%) = 𝑆𝑝𝑟𝑎𝑦 𝑑𝑟𝑖𝑒𝑑 𝑝𝑜𝑤𝑑𝑒𝑟 (𝑔) 𝑆𝑜𝑙𝑖𝑑 𝑐𝑜𝑛𝑡𝑒𝑛𝑡 𝑖𝑛 𝑡ℎ𝑒 𝑓𝑒𝑒𝑑 𝑡𝑜 𝑏𝑒 𝑠𝑝𝑟𝑎𝑦 𝑑𝑟𝑖𝑒𝑑 (𝑔) × 100 Equation (1) 149 150 2 5 Color 151 Acid whey (AW) was obtained from Greek yogurt production. To prepare Greek yogurt, whole 123 milk was heated to 82-85 °C with constant stirring. The milk base was then cooled to 40-45°C and 124 inoculated with a starter culture. The mixture was then fermented at 42°С for 6 to 10 h for 125 coagulation. The yogurt was loaded into 500 ml sample bottles and centrifuged at 5000 rpm for 20 126 minutes to obtain acid whey. Acid whey was then decanted and stored at 4°C for further analysis. 127 128 1. Introduction 34 It was stored in a freezer 110 at − 20 °C until being used. Barnyard and Little millet were purchased from Manna foods 111 (Chennai, India). A Butterfly Matchless 750-watt mixer grinder was used for milling. The 112 cheesecloth used to strain the yogurt was purchased from Wal-Mart stores (Columbia, MO). 113 114 2.2. Chemicals, solvents, and reagents 115 Acetic acid, gallic acid, and 2,2-diphenyl-1-picrylhydrazyl (DPPH) were purchased from Acros 116 Organics (Morris Plains, NJ). Folin-Ciocalteu and Thiobarbituric acid (TBA) were obtained from 117 Sigma-Aldrich Corp. (St. Louis, MO). Ethanol was purchased from Thermo Fisher Scientific 118 (Hampton, NH). Sodium bicarbonate was purchased from Duda Energy LLC (Decatur, AL). 119 120 2.3. Sample preparation 121 2.3.1. Acid whey 122 We used available commercial samples of millets for human consumption for this study due to 107 food availability and accessibility. The whole milk used for yogurt production was obtained from 108 a local grocery store (Wal-Mart Stores, Inc., Columbia, MO, USA), and yogurt culture was 109 obtained from FEAST Lab at the University of Missouri (Columbia, MO). It was stored in a freezer 110 at − 20 °C until being used. Barnyard and Little millet were purchased from Manna foods 111 (Chennai, India). A Butterfly Matchless 750-watt mixer grinder was used for milling. The 112 cheesecloth used to strain the yogurt was purchased from Wal-Mart stores (Columbia, MO). 113 114 4 Acid whey (AW) was obtained from Greek yogurt production. To prepare Greek yogurt, whole 123 milk was heated to 82-85 °C with constant stirring. The milk base was then cooled to 40-45°C and 124 inoculated with a starter culture. The mixture was then fermented at 42°С for 6 to 10 h for 125 coagulation. The yogurt was loaded into 500 ml sample bottles and centrifuged at 5000 rpm for 20 126 minutes to obtain acid whey. Acid whey was then decanted and stored at 4°C for further analysis. 127 128 2.3.2. Preparation of millet acid whey solution 129 Millet flour was suspended in acid whey at 10% and 25%, 50%, 75% and 100% w/v 130 concentrations. The solution was stirred for 10 minutes at 200 rpm, to increase the solubilization 131 of proteins (Romano et al., 2020). The millet acid whey mixture was vacuum filtered to obtain the 132 feed solution for spray drying. The pH of each concentration was determined before spray drying. 2.3.2. Preparation of millet acid whey solution 129 Millet flour was suspended in acid whey at 10% and 25%, 50%, 75% and 100% w/v 130 concentrations. The solution was stirred for 10 minutes at 200 rpm, to increase the solubilization 131 of proteins (Romano et al., 2020). The millet acid whey mixture was vacuum filtered to obtain the 132 feed solution for spray drying. The pH of each concentration was determined before spray drying. 133 AOAC (Association of Official Analytical Chemists) international method was used to determine 134 the total solid content of the soluble fraction (Bradley, 2010). 135 136 2.3.3. Spray drying millet acid whey solution 137 A peristaltic feed pump was used to transfer filtered acid whey millet (AWM) solution into a 138 laboratory-scale spray-dryer (Büchi B290 mini spray-dryer, Flawil, Switzerland). The spray 139 dryer's input temperature, aspirator, feed pump rate, and the gas flow rate was kept at 160°C, 90%, 140 15%, and 45 percent, respectively, based on optimization from prior experiments (S. Malik et al., 141 2021). Before each run, the spray dryer was re-equilibrated with water. The powders were kept at 142 4°C in sealed glass jars until utilized for further analysis. 143 144 2.4. Powder Yield 145 The percentage ratio between the total mass of spray-dried powder recovered and the amount of 146 extract delivered to the system was used to calculate the powder yield (dry basis) (Fang & 147 Bhandari, 2012). 148 Product yield (%) = 𝑆𝑝𝑟𝑎𝑦 𝑑𝑟𝑖𝑒𝑑 𝑝𝑜𝑤𝑑𝑒𝑟 (𝑔) 𝑆𝑜𝑙𝑖𝑑 𝑐𝑜𝑛𝑡𝑒𝑛𝑡 𝑖𝑛 𝑡ℎ𝑒 𝑓𝑒𝑒𝑑 𝑡𝑜 𝑏𝑒 𝑠𝑝𝑟𝑎𝑦 𝑑𝑟𝑖𝑒𝑑 (𝑔) × 100 Equation (1) 149 150 2.5. Color 151 2.6. Water activity 165 The water activity (aw) of spray-dried powder was determined in a water activity meter (Cx-2, 166 Decagon Devices, Inc., Pullman, Washington) at 25 °C with a 0.001 sensitivity. 167 The water activity (aw) of spray-dried powder was determined in a water activity meter (Cx-2, 166 Decagon Devices, Inc., Pullman, Washington) at 25 °C with a 0.001 sensitivity. 167 2.7. Moisture content 169 A halogen moisture analyzer was used to measure the moisture content of about 1g of spray-dried 170 powder in a tared sample pan (HE53, Mettler Toledo, Columbus, Ohio 43240). For all powder 171 samples, the procedure was repeated four times. 172 173 A halogen moisture analyzer was used to measure the moisture content of about 1g of spray-dried 170 powder in a tared sample pan (HE53, Mettler Toledo, Columbus, Ohio 43240). For all powder 171 samples, the procedure was repeated four times. 172 173 2.3.3. Spray drying millet acid whey solution 137 A peristaltic feed pump was used to transfer filtered acid whey millet (AWM) solution into a 138 laboratory-scale spray-dryer (Büchi B290 mini spray-dryer, Flawil, Switzerland). The spray 139 dryer's input temperature, aspirator, feed pump rate, and the gas flow rate was kept at 160°C, 90%, 140 15%, and 45 percent, respectively, based on optimization from prior experiments (S. Malik et al., 141 2021). Before each run, the spray dryer was re-equilibrated with water. The powders were kept at 142 4°C in sealed glass jars until utilized for further analysis. 143 5 5 The color characteristics of spray-dried powder were evaluated by L, a, and b* indexes using a 152 handheld chromameter (Konica Minolta CR-410, Chiyoda, Tokyo, Japan). The values for L* 153 (brightness/darkness), a* (redness/greenness), and b* (yellowness/blueness) were determined. The 154 measurements were taken with 47mm Petri dishes that fit precisely the colorimeter opening. 155 Before each use, the colorimeter was calibrated as directed by the manufacturer. For each sample, 156 three measurements were conducted, and the average result was reported. Equations (2, 3, 4 and 157 5) were used to calculate the color difference (∆E) between a whitening food additive (subscript 158 1) and the spray-dried powders whiteness index, chroma, hue angle (°). 159 The color characteristics of spray-dried powder were evaluated by L, a, and b* indexes using a 152 handheld chromameter (Konica Minolta CR-410, Chiyoda, Tokyo, Japan). The values for L* 153 (brightness/darkness), a* (redness/greenness), and b* (yellowness/blueness) were determined. The 154 measurements were taken with 47mm Petri dishes that fit precisely the colorimeter opening. 155 Before each use, the colorimeter was calibrated as directed by the manufacturer. For each sample, 156 three measurements were conducted, and the average result was reported. Equations (2, 3, 4 and 157 5) were used to calculate the color difference (∆E) between a whitening food additive (subscript 158 1) and the spray-dried powders whiteness index, chroma, hue angle (°). 159 measurements were taken with 47mm Petri dishes that fit precisely the colorimeter opening 155 Before each use, the colorimeter was calibrated as directed by the manufacturer. For each sample 156 three measurements were conducted, and the average result was reported. Equations (2, 3, 4 and 157 5) were used to calculate the color difference (∆E) between a whitening food additive (subscrip 158 1) and the spray-dried powders whiteness index, chroma, hue angle (°). 2.3.3. Spray drying millet acid whey solution 137 159 ∆𝐸 = √(𝐿1 ∗−𝐿∗) + (𝑎1 ∗−𝑎∗) + (𝑏1 ∗−𝑏∗) Equation (2) 160 Chroma (C) = √(𝑎∗2) + (𝑏∗2) Equation (3) 161 Hue angle (h) = 𝑡𝑎𝑛−1( 𝑏∗ 𝑎∗) Equation (4) 162 Whiteness Index = 100 −[(100 −𝐿∗) + 𝑎∗2 + 𝑏∗2]0.5 Equation (5) 163 164 2.6. Water activity 165 The water activity (aw) of spray-dried powder was determined in a water activity meter (Cx-2 166 Decagon Devices, Inc., Pullman, Washington) at 25 °C with a 0.001 sensitivity. 167 168 2.7. Moisture content 169 A halogen moisture analyzer was used to measure the moisture content of about 1g of spray-dried 170 powder in a tared sample pan (HE53, Mettler Toledo, Columbus, Ohio 43240). For all powde 171 samples, the procedure was repeated four times. 172 173 2.8. Hygroscopicity 174 With minor modifications, hygroscopicity was determined using the (Sarabandi et al., 2018 175 method. About 1 g of sample was kept in a closed desiccator at 25 °C with saturated NaCl solution 176 (75% relative humidity) for seven days. The weight gain owing to moisture adsorption wa 177 measured daily, and hygroscopicity was calculated as the mass of water absorbed per 100 g of dry 178 sample. 179 180 2.9. Water solubility 181 ∆𝐸 = √(𝐿1 ∗−𝐿∗) + (𝑎1 ∗−𝑎∗) + (𝑏1 ∗−𝑏∗) Equation (2) 160 Chroma (C) = √(𝑎∗2) + (𝑏∗2) Equation (3) 161 Hue angle (h) = 𝑡𝑎𝑛−1( 𝑏∗ 𝑎∗) Equation (4) 162 Whiteness Index = 100 −[(100 −𝐿∗) + 𝑎∗2 + 𝑏∗2]0.5 Equation (5) 163 164 2.10. Dispersibility 189 The dispersibility of the spray-dried powders was measured following the method described by 190 Jinapong et al. (2008). Approximately 1 g of spray-dried powder and 10 mL distilled water were 191 added to a 50 mL beaker. The mixture was stirred vigorously with a spatula for 15 seconds and 192 then poured through a sieve (212 𝜇m). The sieved sample was transferred onto a pre-weighed 193 aluminum pan and dried in an oven at 105 °C for 4 h. The dispersibility (%) was calculated using 194 Equation (6). 195 % Dispersibility = (10+𝑊) × % 𝑇𝑆 𝑊 ×(100−𝑀𝐶 100 ) Equation (6) % Dispersibility = (10+𝑊) × % 𝑇𝑆 𝑊 ×(100−𝑀𝐶 100 ) Equation (6) 196 % Dispersibility = (10+𝑊) × % 𝑇𝑆 𝑊 ×(100−𝑀𝐶 100 ) 196 Equation (6) Where W = weight (g) of the powder sample, % TS = percentage of dry matter in the wet powder 197 mixture after sieving, and MC = moisture content of the powder. 198 Where W = weight (g) of the powder sample, % TS = percentage of dry matter in the wet powder 197 mixture after sieving, and MC = moisture content of the powder. 198 199 2.8. Hygroscopicity 174 6 6 The solubility of the spray-dried powders was determined according to the method described by 182 Santhalakshmy et al. (2015) with minor modifications. Approximately 0.5 g of powder sample 183 was added to 15 mL distilled water and mixed thoroughly. After mixing, the solution was 184 centrifuged for 5 min at 5000 rpm. The supernatant was transferred into an aluminum dish and 185 dried in a 105°C oven for two hours. The solubility (%) in water was calculated as the weight of 186 dried supernatant per 100g of dry sample. 187 188 2.13. Flowability (Angle of repose, Hausner ratio, and Carr index) 216 The flow characteristics of the powder samples were expressed as the Carr Index (CI) and Hausner 217 ratio (HR) (Carr, 1965; Hausner, 1967) calculated from the bulk density (ρb) and tapped density 218 (ρt) values. Powders with HR < 1.1 and CI < 10 are considered very free-flowing powders. CI and 219 HR values were calculated using equations (8) and (9). 220 Carr Index CI = 𝜌𝑡− 𝜌𝑏 𝜌𝑡 × 100 Equation (8) 221 Hausner Ratio (HR) = 𝜌𝑡 𝜌𝑏 Equation (9) 222 Carr Index CI = 𝜌𝑡− 𝜌𝑏 𝜌𝑡 × 100 221 Hausner Ratio (HR) = 𝜌𝑡 𝜌𝑏 222 Equation (9) A modified version of the fixed funnel method described by Beakawi Al-Hashemi & Baghabra 223 Al-Amoudi (2018) was used to determine the angle of repose. A funnel was suspended 10 cm 224 above a horizontal surface using a stand. One gram of the powder was poured through the funnel 225 to make a heap, and the height and diameter were measured using Equation (10). 226 A modified version of the fixed funnel method described by Beakawi Al-Hashemi & Baghabra 223 Angle of repose (°) = 𝑡𝑎𝑛−1 ( 2𝐻 𝐷) Equation (10) 227 Equation (10) Equation (10) Where H = height of the heap and D = diameter of the base 228 229 2.11. Density measurements (Bulk, Tapped, True) 200 239 Porosity 𝜀 (%) = (𝜌t− 𝜌𝑏) 𝜌𝑡 × 100 213 Equation (7) 2.11. Density measurements (Bulk, Tapped, True) 200 212 Powder porosity was calculated from the bulk density and true density values using the equation 211 (7) described by Jain & Bal (1997). 212 p y y y g q (7) described by Jain & Bal (1997). 212 Porosity 𝜀 (%) = (𝜌t− 𝜌𝑏) 𝜌𝑡 × 100 Equation (7 213 Where 𝜌𝑏 = bulk density and 𝜌𝑡 = true density 214 215 2.13. Flowability (Angle of repose, Hausner ratio, and Carr index) 216 The flow characteristics of the powder samples were expressed as the Carr Index (CI) and Hausne 217 ratio (HR) (Carr, 1965; Hausner, 1967) calculated from the bulk density (ρb) and tapped densit 218 (ρt) values. Powders with HR < 1.1 and CI < 10 are considered very free-flowing powders. CI an 219 HR values were calculated using equations (8) and (9). 220 Carr Index CI = 𝜌𝑡− 𝜌𝑏 𝜌𝑡 × 100 Equation (8 221 Hausner Ratio (HR) = 𝜌𝑡 𝜌𝑏 Equation (9 222 A modified version of the fixed funnel method described by Beakawi Al-Hashemi & Baghabr 223 Al-Amoudi (2018) was used to determine the angle of repose. A funnel was suspended 10 cm 224 above a horizontal surface using a stand. One gram of the powder was poured through the funne 225 to make a heap, and the height and diameter were measured using Equation (10). 226 Angle of repose (°) = 𝑡𝑎𝑛−1 ( 2𝐻 𝐷) Equation (10 227 Where H = height of the heap and D = diameter of the base 228 229 2.14. Particle size distribution and Zeta potential 230 A dynamic light scattering instrument was used to measure the particle size distribution and zet 231 potential of the spray-dried powders (Delsa Nano Submicron Particle Size and Zeta Potentia 232 particle Analyzer laser light, Beckman Coulter California, United States). A small amount o 233 powder (1g) was dispersed in water and diluted to 0.1% concentration. The particle size and zet 234 potential were measured based on the diffraction of laser scattered by the particles an 235 electrophoretic light scattering, respectively. The particle size (mean hydrodynamic diameter 236 expressed as nm, poly dispersibility index (PDI), and zeta potential were established by calculatin 237 the average of 15 runs using the Delsa Nano software. The span index was calculated usin 238 Equation (11) (Gagneten et al., 2019). All measurements were performed in triplicates. 2.11. Density measurements (Bulk, Tapped, True) 200 Bulk density (𝜌𝑏) was determined according to the method described by Zungur Bastıoğlu et al. 201 (2016). The bulk density was calculated as the sample weight divided by the sample volume. 202 Approximately 1 g of powder was added to a 10 mL graduated cylinder, and the volume was noted. 203 Tapped density (𝜌𝑡), was measured by adding 1g of powder to a 10 ml graduated cylinder and 204 mechanically tapped using a shaker for 20 min at 1000 rpm (Jinapong et al., 2008). The true density 205 of the powders was established using a gas pycnometer (Quantachrome Ultrapycnometer 1000 206 Anton Paar, Graz, Austria). All the experiments were replicated six times, and the average value 207 was recorded. 208 209 2 12 Porosit 210 Bulk density (𝜌𝑏) was determined according to the method described by Zungur Bastıoğlu et al. 201 (2016). The bulk density was calculated as the sample weight divided by the sample volume. 202 Approximately 1 g of powder was added to a 10 mL graduated cylinder, and the volume was noted. 203 Tapped density (𝜌𝑡), was measured by adding 1g of powder to a 10 ml graduated cylinder and 204 mechanically tapped using a shaker for 20 min at 1000 rpm (Jinapong et al., 2008). The true density 205 of the powders was established using a gas pycnometer (Quantachrome Ultrapycnometer 1000 206 Anton Paar, Graz, Austria). All the experiments were replicated six times, and the average value 207 was recorded. 208 209 2.12. Porosity 210 Bulk density (𝜌𝑏) was determined according to the method described by Zungur Bastıoğlu et al. 201 (2016). The bulk density was calculated as the sample weight divided by the sample volume. 202 Approximately 1 g of powder was added to a 10 mL graduated cylinder, and the volume was noted. 203 Tapped density (𝜌𝑡), was measured by adding 1g of powder to a 10 ml graduated cylinder and 204 mechanically tapped using a shaker for 20 min at 1000 rpm (Jinapong et al., 2008). The true density 205 of the powders was established using a gas pycnometer (Quantachrome Ultrapycnometer 1000 206 Anton Paar, Graz, Austria). All the experiments were replicated six times, and the average value 207 was recorded. 208 209 2 12 Porosity 210 7 7 Powder porosity was calculated from the bulk density and true density values using the equation 211 (7) described by Jain & Bal (1997). 2.14. Particle size distribution and Zeta potential 230 A dynamic light scattering instrument was used to measure the particle size distribution and zeta 231 potential of the spray-dried powders (Delsa Nano Submicron Particle Size and Zeta Potential 232 particle Analyzer laser light, Beckman Coulter California, United States). A small amount of 233 powder (1g) was dispersed in water and diluted to 0.1% concentration. The particle size and zeta 234 potential were measured based on the diffraction of laser scattered by the particles and 235 electrophoretic light scattering, respectively. The particle size (mean hydrodynamic diameter) 236 expressed as nm, poly dispersibility index (PDI), and zeta potential were established by calculating 237 the average of 15 runs using the Delsa Nano software. The span index was calculated using 238 Equation (11) (Gagneten et al., 2019). All measurements were performed in triplicates. 239 8 Span index = 𝐷90−𝐷10 𝐷50 Equation (11) Span index = 𝐷90−𝐷10 𝐷50 Equ 240 D10, D50, and D90 are the diameters at which 10%, 50%, and 90% of the population 241 the obtained result. 242 243 2.15. Particle morphology 244 The particle morphology of the spray-dried powders was observed using a scannin 245 microscope (SEM) (FEI Quanta 600F ESEM, Oregon, USA) operated in a high vac 246 samples were mounted with carbon adhesive and sputtered with 25 nm Pt for ima 247 morphological analysis was done at 5 kV, 30um objective aperture, 3.5 spot size, 248 working distance. 249 250 2.16. Differential scanning calorimetry (DSC) 251 The glass transition temperature (Tg) of the powders was determined using a differentia 252 calorimeter (Q200 DSC, TA instrument, Schaumburg, IL). Approximately 7 mg of p 253 weighed and sealed in an airtight aluminum pan. The sample was first cooled to 20 °C 254 heated to 200 °C. All measurements were conducted in an inert environment at a temp 255 10 °C /min, using nitrogen with the gas flow at 50 mL/min. An empty aluminum pan w 256 a reference. The glass transition temperature (Tg) and degradation temperature were 257 using Universal V4 5A TA Instruments analysis software. Three measurements were 258 each sample. 259 260 2.17 Antioxidant properties 261 2.17.1. Preparation of extracts 262 Ethanolic extracts of the samples were prepared according to the method described by (C 263 & Singh, 1980). About 1 g of each sample was homogenized in 10 mL of ethanol 264 macerate for 24 h at room temperature in a shaker. 2.14. Particle size distribution and Zeta potential 230 Then, the samples were centrifuged 265 at 5000 rpm. The supernatant was collected and dried for 24 hours at room temperatu 266 analysis, the extracts were reconstituted with 5 mL of distilled water. 267 268 2.17.2. Total phenolic content (TPC) 269 Equation (11) Equation (11) 240 𝐷50 D10, D50, and D90 are the diameters at which 10%, 50%, and 90% of the population are below 241 the obtained result. 242 The particle morphology of the spray-dried powders was observed using a scanning electron 245 microscope (SEM) (FEI Quanta 600F ESEM, Oregon, USA) operated in a high vacuum. The 246 samples were mounted with carbon adhesive and sputtered with 25 nm Pt for imaging. The 247 morphological analysis was done at 5 kV, 30um objective aperture, 3.5 spot size, and 8mm 248 working distance. 249 2.17.3. Determination of antioxidant activity using DPPH scavenging assay 277 The antioxidant activity was measured according to the method described by Horvat et al., (2020). 278 The assay was prepared by mixing 0.2 mL of sample extract with 1 ml 0.5mMol/L 2,2-diphenyl- 279 1-picrylhydrazyl (DPPH) solution and 2 mL ethanol. The mixture was incubated in a dark place at 280 room temperature for 30 minutes. The absorbance of the reaction mixtures was measured at 517 281 nm. The percentage of inhibition of free radical DPPH was calculated against the blank (DPPH + 282 ethanol) using Equation (12). 283 DPPH scavenging activity (% Inhibition) = 1 −( 𝐴𝑠𝑎𝑚𝑝𝑙𝑒 𝑡=30 𝐴𝑏𝑙𝑎𝑛𝑘 𝑡=0 ) × 100 Equation (12) 284 DPPH scavenging activity (% Inhibition) = 1 −( 𝐴𝑠𝑎𝑚𝑝𝑙𝑒 𝑡=30 𝐴𝑏𝑙𝑎𝑛𝑘 𝑡=0 ) × 100 Equation (12) 284 A sample = Absorbance of the sample after 30 mins Ablank = Absorbance of the blank at time = 285 A sample = Absorbance of the sample after 30 mins, Ablank = Absorbance of the blank at time = 285 0. 286 287 A sample = Absorbance of the sample after 30 mins, Ablank = Absorbance of the blank at time = 285 0. 286 2.16. Differential scanning calorimetry (DSC) 251 Determination of antioxidant activity using DPPH scavenging assay 277 The antioxidant activity was measured according to the method described by Horvat et al., (2020 278 The assay was prepared by mixing 0.2 mL of sample extract with 1 ml 0.5mMol/L 2,2-dipheny 279 1-picrylhydrazyl (DPPH) solution and 2 mL ethanol. The mixture was incubated in a dark place 280 room temperature for 30 minutes. The absorbance of the reaction mixtures was measured at 5 281 nm. The percentage of inhibition of free radical DPPH was calculated against the blank (DPPH 282 ethanol) using Equation (12). 283 DPPH scavenging activity (% Inhibition) = 1 −( 𝐴𝑠𝑎𝑚𝑝𝑙𝑒 𝑡=30 𝐴𝑏𝑙𝑎𝑛𝑘 𝑡=0 ) × 100 Equation (1 284 A sample = Absorbance of the sample after 30 mins, Ablank = Absorbance of the blank at time 285 0. 286 287 2.17.4. Estimation of thiobarbituric acid reactive substances (TBARS) 288 TBARS was measured as described by Zeb & Ullah (2016). A 4.0 mM standard solution 289 thiobarbituric acid (TBA) was prepared by dissolving 57.66 mg of TBA in 100ml glacial ace 290 acid. The sample extract (1 mL) was mixed with 1 mL TBA reagent and heated in a water bath 291 95°C for 1 hour. The mixture was cooled to room temperature, and the absorbance was measur 292 at 532 nm. Malondialdehyde tetrabutylammonium (MDA) salt was used to develop the standa 293 curve at 0.1, 0.2, 0.4, 0.6, and 0.8 mM MDA concentrations. The concentrations of TBARS we 294 calculated using Equation (13). 295 296 TBARS (𝜇M/g) = (Ac × V) /W Equation (13 297 298 Total phenolic content was determined by the Folin–Ciocalteu reagent method described by 270 Singleton & Rossi (1965) with minor modifications. About 2 mL of extract was mixed with 1mL 271 distilled water, 0.5 mL Folin–Ciocalteu reagent, and 2mL of 20% sodium bicarbonate. The mixture 272 was vigorously shaken and incubated for 25 min at 40 °C. The TPC was determined using a UV- 273 visible spectrophotometer, and absorbance was measured at 765 nm. The results were expressed 274 in milligrams of gallic acid equivalents per gram of dry powder (mg GAE/g). 275 276 2.16. Differential scanning calorimetry (DSC) 251 The glass transition temperature (Tg) of the powders was determined using a differential scanning 252 calorimeter (Q200 DSC, TA instrument, Schaumburg, IL). Approximately 7 mg of powder was 253 weighed and sealed in an airtight aluminum pan. The sample was first cooled to 20 °C and then 254 heated to 200 °C. All measurements were conducted in an inert environment at a temperature of 255 10 °C /min, using nitrogen with the gas flow at 50 mL/min. An empty aluminum pan was used as 256 a reference. The glass transition temperature (Tg) and degradation temperature were calculated 257 using Universal V4 5A TA Instruments analysis software. Three measurements were taken for 258 each sample. 259 Ethanolic extracts of the samples were prepared according to the method described by (C. P. Malik 263 & Singh, 1980). About 1 g of each sample was homogenized in 10 mL of ethanol and left to 264 macerate for 24 h at room temperature in a shaker. Then, the samples were centrifuged for 20 min 265 at 5000 rpm. The supernatant was collected and dried for 24 hours at room temperature. Before 266 analysis, the extracts were reconstituted with 5 mL of distilled water. 267 9 Total phenolic content was determined by the Folin–Ciocalteu reagent method described by 270 Singleton & Rossi (1965) with minor modifications. About 2 mL of extract was mixed with 1mL 271 distilled water, 0.5 mL Folin–Ciocalteu reagent, and 2mL of 20% sodium bicarbonate. The mixture 272 was vigorously shaken and incubated for 25 min at 40 °C. The TPC was determined using a UV- 273 visible spectrophotometer, and absorbance was measured at 765 nm. The results were expressed 274 in milligrams of gallic acid equivalents per gram of dry powder (mg GAE/g). 275 276 Total phenolic content was determined by the Folin–Ciocalteu reagent method described 270 Singleton & Rossi (1965) with minor modifications. About 2 mL of extract was mixed with 1m 271 distilled water, 0.5 mL Folin–Ciocalteu reagent, and 2mL of 20% sodium bicarbonate. The mixtu 272 was vigorously shaken and incubated for 25 min at 40 °C. The TPC was determined using a UV 273 visible spectrophotometer, and absorbance was measured at 765 nm. The results were express 274 in milligrams of gallic acid equivalents per gram of dry powder (mg GAE/g). 275 276 2.17.3. 2.18. Statistical Analysis 302 All experiments were done in triplicate using three independent powder samples. Analysis of 303 variance (ANOVA) of mean values was carried out using the JMP 14.0 statistical software 304 program (SAS Institute Inc, Cary, NC). Means were compared by the Tukey test, and significance 305 was accepted at a 95% confidence interval (p < 0.05). Results were presented as mean ± standard 306 deviation. 307 2.17.4. Estimation of thiobarbituric acid reactive substances (TBARS) 288 Statistical Analysis 302 All experiments were done in triplicate using t 303 variance (ANOVA) of mean values was carri 304 program (SAS Institute Inc, Cary, NC). Means w 305 was accepted at a 95% confidence interval (p < 0 306 deviation. 307 308 3. RESULTS AND DISCUSSION 309 3.1 Samples selected for analysis 310 In preliminary studies, ten AWM combination 311 significant differences in bulk density, flowabilit 312 concentrations of a millet type. Because whitene 313 dried powders with the highest whitening index w 314 315 3.2. Powder yield, moisture content, and wate 316 The stickiness of the powder to the drying cham 317 indicator of the spray dryer's efficiency. The add 318 to a significant increase in the yield (Table 319 formulations could be attributed to the millet flo 320 lactic acid present in AW. The addition of millet 321 lower hygroscopicity and higher glass transition 322 powders (Bylund, 2015). Chandrapala & Vasilj 323 acid and calcium concentrations in lactose res 324 properties and increased powder yields. 325 The moisture content of the powders ra 326 stability and a lower risk of oxidative decompos 327 Low moisture content reduces agglomeration an 328 reconstitution in water. The powder's water act 329 Where Ac is the concentration determined from the calibration curve, W is the weight of the 299 sample, and V is the volume in mL of the total extract prepared. 300 Where Ac is the concentration determined from the calibration curve, W is the weight of 299 sample, and V is the volume in mL of the total extract prepared. 300 301 2.18. Statistical Analysis 302 All experiments were done in triplicate using three independent powder samples. Analysis 303 variance (ANOVA) of mean values was carried out using the JMP 14.0 statistical softw 304 program (SAS Institute Inc, Cary, NC). Means were compared by the Tukey test, and significan 305 was accepted at a 95% confidence interval (p < 0.05). Results were presented as mean ± stand 306 deviation. 307 308 3. RESULTS AND DISCUSSION 309 3.1 Samples selected for analysis 310 In preliminary studies, ten AWM combinations were tested, and the results did not indic 311 significant differences in bulk density, flowability, water-solubility, or dispersibility between 312 concentrations of a millet type. Because whiteness was a desired attribute in this study, the spr 313 dried powders with the highest whitening index were selected for further analysis. 2.17.4. Estimation of thiobarbituric acid reactive substances (TBARS) 288 2.17.4. Estimation of thiobarbituric acid reactive substances (TBARS) 288 TBARS was measured as described by Zeb & Ullah (2016). A 4.0 mM standard solution of 289 thiobarbituric acid (TBA) was prepared by dissolving 57.66 mg of TBA in 100ml glacial acetic 290 acid. The sample extract (1 mL) was mixed with 1 mL TBA reagent and heated in a water bath at 291 95°C for 1 hour. The mixture was cooled to room temperature, and the absorbance was measured 292 at 532 nm. Malondialdehyde tetrabutylammonium (MDA) salt was used to develop the standard 293 curve at 0.1, 0.2, 0.4, 0.6, and 0.8 mM MDA concentrations. The concentrations of TBARS were 294 calculated using Equation (13). 295 296 TBARS (𝜇M/g) = (Ac × V) /W Equation (13) 297 298 TBARS was measured as described by Zeb & Ullah (2016). A 4.0 mM standard solution of 289 thiobarbituric acid (TBA) was prepared by dissolving 57.66 mg of TBA in 100ml glacial acetic 290 acid. The sample extract (1 mL) was mixed with 1 mL TBA reagent and heated in a water bath at 291 95°C for 1 hour. The mixture was cooled to room temperature, and the absorbance was measured 292 at 532 nm. Malondialdehyde tetrabutylammonium (MDA) salt was used to develop the standard 293 curve at 0.1, 0.2, 0.4, 0.6, and 0.8 mM MDA concentrations. The concentrations of TBARS were 294 calculated using Equation (13). 295 296 TBARS (𝜇M/g) = (Ac × V) /W Equation (13) 297 TBARS was measured as described by Zeb & Ullah (2016). A 4.0 mM standard solution of 289 thiobarbituric acid (TBA) was prepared by dissolving 57.66 mg of TBA in 100ml glacial acetic 290 acid. The sample extract (1 mL) was mixed with 1 mL TBA reagent and heated in a water bath at 291 95°C for 1 hour. The mixture was cooled to room temperature, and the absorbance was measured 292 at 532 nm. Malondialdehyde tetrabutylammonium (MDA) salt was used to develop the standard 293 curve at 0.1, 0.2, 0.4, 0.6, and 0.8 mM MDA concentrations. The concentrations of TBARS were 294 calculated using Equation (13). 295 Equation (13) Equation (13) TBARS (𝜇M/g) = (Ac × V) /W 10 Where Ac is the concentration determined from 299 sample, and V is the volume in mL of the total ex 300 301 2.18. 2.17.4. Estimation of thiobarbituric acid reactive substances (TBARS) 288 314 315 3.2. Powder yield, moisture content, and water activity 316 The stickiness of the powder to the drying chamber walls affects the powder yield, which is 317 indicator of the spray dryer's efficiency. The addition of millet flour to the AW formulations 318 to a significant increase in the yield (Table 2). The increased yields of BAW and LA 319 formulations could be attributed to the millet flour's neutralizing and wall-forming ability on 320 lactic acid present in AW. The addition of millet in the feed suspension resulted in particles w 321 lower hygroscopicity and higher glass transition temperatures hence reducing the stickiness of 322 powders (Bylund, 2015). Chandrapala & Vasiljevic (2017) demonstrated that altering the lac 323 acid and calcium concentrations in lactose resulted in powders with improved crystallizat 324 properties and increased powder yields. 325 The moisture content of the powders ranged from 3.78% to 5.53 %, indicating pow 326 stability and a lower risk of oxidative decomposition and microbial activity (Tapia et al., 202 327 Low moisture content reduces agglomeration and improves the solubility of the powder dur 328 reconstitution in water. The powder's water activity varied significantly (p < 0.05) among 329 3.2. Powder yield, moisture content, and water activity 316 (2021) obtained similar results 344 for spray-dried acid whey with Kodo and Proso millet. 345 346 powders. The water activity of the powders ranged from 0.29 ± 0.00 for BAW10 to 0.25 ± 0.00 330 for LAW25 indicating good microbial stability. Chandrapala & Vasiljevic (2018) obtained similar 331 results for spray-dried acid whey powders following nanofiltration. 332 333 3.3. Color 334 The color of the spray-dried powders will affect the final product's color during food processing 335 and plays a pivotal role in the acceptance of the product. Visually both the AWM and AW powders 336 had a white appearance. The results of the color coordinates (L, a, b, C, H°, and ∆E) of the 337 powders are presented in Table (3). According to color analysis results, the effect of millet flour 338 on the L* differed considerably (p > 0.05) among the powder formulations. When the amount of 339 millet flour in the AW solution was increased, the whiteness index value for LAW25 was reduced 340 (Figure 1). Interestingly the addition of 10% millet flour consistently improved the whiteness of 341 the powders for both millet types. The L* values and whiteness index showed more variability 342 when the millet concentrations were raised. This implies that the concentration of millet 343 substantially affects the color of the powder produced. Malik et al. (2021) obtained similar results 344 for spray-dried acid whey with Kodo and Proso millet. 345 346 results for spray-dried acid whey powders following nanofiltration. 332 333 3.3. Color 334 The color of the spray-dried powders will affect the final product's color during food processing 335 and plays a pivotal role in the acceptance of the product. Visually both the AWM and AW powders 336 had a white appearance. The results of the color coordinates (L, a, b, C, H°, and ∆E) of the 337 powders are presented in Table (3). According to color analysis results, the effect of millet flour 338 on the L* differed considerably (p > 0.05) among the powder formulations. When the amount of 339 millet flour in the AW solution was increased, the whiteness index value for LAW25 was reduced 340 (Figure 1). Interestingly the addition of 10% millet flour consistently improved the whiteness of 341 the powders for both millet types. The L* values and whiteness index showed more variability 342 when the millet concentrations were raised. 3.2. Powder yield, moisture content, and water activity 316 This implies that the concentration of millet 343 substantially affects the color of the powder produced. Malik et al. (2021) obtained similar results 344 for spray-dried acid whey with Kodo and Proso millet. 345 346 3.4. Hygroscopicity 347 Figure 2 shows the progression of the powder's hygroscopicity (at 75% humidity) over seven days. 348 Hygroscopicity refers to a powder's tendency to absorb water from its surroundings; it affects its 349 storage stability and quality. Lower hygroscopicity results in less sticky samples that are easier to 350 package and handle. The water absorption rate of the spray-dried powders was linear until day 5, 351 after which the water gain was significantly decreased. As a result, after being exposed to a relative 352 humidity of 75% for five days, the hygroscopicity was calculated. The hygroscopicity values 353 ranged from 19.45 ± 0.21% for BAW25 to 20.9 ± 0.28% for AW. After five days for BAW25, 354 LAW10, and BAW10 powders, the hygroscopicity values differed significantly from the LAW 25 355 and AW powders. As expected, pure acid whey powders had slightly higher hygroscopicity values 356 than AWM. Bédas et al., (2017) reported similar hygroscopicity values of 14.1 ± 1.0 for nano- 357 filtered lactic acid whey and 22.6 ± 1.2 for lactic acid whey powders. 358 3.5. Water solubility and dispersibility index 359 3.2. Powder yield, moisture content, and water activity 316 The stickiness of the powder to the drying chamber walls affects the powder yield, which is an 317 indicator of the spray dryer's efficiency. The addition of millet flour to the AW formulations led 318 to a significant increase in the yield (Table 2). The increased yields of BAW and LAW 319 formulations could be attributed to the millet flour's neutralizing and wall-forming ability on the 320 lactic acid present in AW. The addition of millet in the feed suspension resulted in particles with 321 lower hygroscopicity and higher glass transition temperatures hence reducing the stickiness of the 322 powders (Bylund, 2015). Chandrapala & Vasiljevic (2017) demonstrated that altering the lactic 323 acid and calcium concentrations in lactose resulted in powders with improved crystallization 324 properties and increased powder yields. 325 The moisture content of the powders ranged from 3.78% to 5.53 %, indicating powder 326 stability and a lower risk of oxidative decomposition and microbial activity (Tapia et al., 2020). 327 Low moisture content reduces agglomeration and improves the solubility of the powder during 328 reconstitution in water. The powder's water activity varied significantly (p < 0.05) among the 329 11 powders. The water activity of the powders ranged from 0.29 ± 0.00 for BAW10 to 0.25 ± 0.00 330 for LAW25 indicating good microbial stability. Chandrapala & Vasiljevic (2018) obtained similar 331 results for spray-dried acid whey powders following nanofiltration. 332 333 3.3. Color 334 The color of the spray-dried powders will affect the final product's color during food processing 335 and plays a pivotal role in the acceptance of the product. Visually both the AWM and AW powders 336 had a white appearance. The results of the color coordinates (L, a, b, C, H°, and ∆E) of the 337 powders are presented in Table (3). According to color analysis results, the effect of millet flour 338 on the L* differed considerably (p > 0.05) among the powder formulations. When the amount of 339 millet flour in the AW solution was increased, the whiteness index value for LAW25 was reduced 340 (Figure 1). Interestingly the addition of 10% millet flour consistently improved the whiteness of 341 the powders for both millet types. The L* values and whiteness index showed more variability 342 when the millet concentrations were raised. This implies that the concentration of millet 343 substantially affects the color of the powder produced. Malik et al. 3.5. Water solubility and dispersibility index 359 Powders 386 with a Hausner ratio greater than 1.34 are considered to be cohesive and less free-flowing 387 (Hausner, 1967), while powders with a Carr index greater than 25% are considered non-free- 388 flowing (Carr, 1965). The angle of repose values less than 30° indicate the very free-flowing nature 389 of the powders. There was a significant difference (p > 0.05) in the CI and HR values for all the 390 Water solubility is a crucial reconstitution property that influences the powders' potential use as a 360 functional ingredient and consumer acceptance. The solubility of a powder is its ability to dissolve 361 in water. In this study, the solubility values of the powders did not differ significantly 362 Water solubility is a crucial reconstitution property that influences the powders' potential use as a 360 functional ingredient and consumer acceptance. The solubility of a powder is its ability to dissolve 361 in water. In this study, the solubility values of the powders did not differ significantly 362 The solubility index values for the powders ranged from 73.17 ± 0.33 % for AW10 to 85.96 ± 0.99 363 % for BAW10 (Table 4). According to these findings, the addition of millet flour as a carrier 364 enhanced the solubility of the AWM powders. The variation in water solubility could be due to 365 variations in millet varieties' starch and protein content. Cano-Chauca et al. (2005) studied the 366 solubility of mango juice powders in different carriers. This result is consistent with their findings 367 that reported that the solubility of mango juice powder produced with maltodextrin as the carrier 368 agent was greater than 90%. Dispersibility measures the rehydration ability of powders. The higher 369 the dispersibility, the better the sample reconstitutes in water (Schuck, 2011). The dispersibility 370 index did not increase as the concentration of millet increased (Table 3). High dispersibility values 371 observed for the powders imply that the powder samples are highly soluble in water. Li & Zhong 372 (2021) investigated the rehydration properties of spray-dried millet protein isolates and reported 373 improved rehydration properties with the addition of soluble caseins. 374 375 3.5. Water solubility and dispersibility index 359 12 Water solubility is a crucial reconstitution property that influence 360 functional ingredient and consumer acceptance. The solubility of 361 in water. In this study, the solubility values of the powders did no 362 The solubility index values for the powders ranged from 73.17 ± 0 363 % for BAW10 (Table 4). According to these findings, the add 364 enhanced the solubility of the AWM powders. The variation in 365 variations in millet varieties' starch and protein content. Cano-C 366 solubility of mango juice powders in different carriers. This resul 367 that reported that the solubility of mango juice powder produced 368 agent was greater than 90%. Dispersibility measures the rehydrati 369 the dispersibility, the better the sample reconstitutes in water (S 370 index did not increase as the concentration of millet increased (Ta 371 observed for the powders imply that the powder samples are high 372 (2021) investigated the rehydration properties of spray-dried mi 373 improved rehydration properties with the addition of soluble case 374 375 3.6. Density, porosity, and flow properties 376 The results for the bulk density of the spray-dried powders are sh 377 powders' bulk density ranged from 253.89 ± 12.71 Kgm3 for BA 378 LAW25. Chever et al. (2017) reported similar bulk density val 379 whole milk powders. Porosity values were measured based on th 380 true densities. The mean porosity results ranged from 82.21 ± 0.1 381 % for LAW25. Similar results were reported for milk protein iso 382 spray-dried dairy protein powders (Fournaise et al., 2021). Li & 383 porosity values for milk protein isolate powders with increased c 384 Carr's compressibility index (CI), Hausner ratio (HR), a 385 determine the flow properties of the powders, and the results ar 386 with a Hausner ratio greater than 1.34 are considered to be 387 (Hausner, 1967), while powders with a Carr index greater than 388 flowing (Carr, 1965). The angle of repose values less than 30° ind 389 of the powders. There was a significant difference (p > 0.05) in 390 Water solubility is a crucial reconstitution property that influences the powders' potential use as a 360 functional ingredient and consumer acceptance. The solubility of a powder is its ability to dissolve 361 in water. 3.5. Water solubility and dispersibility index 359 In this study, the solubility values of the powders did not differ significantly 362 The solubility index values for the powders ranged from 73.17 ± 0.33 % for AW10 to 85.96 ± 0.99 363 % for BAW10 (Table 4). According to these findings, the addition of millet flour as a carrier 364 enhanced the solubility of the AWM powders. The variation in water solubility could be due to 365 variations in millet varieties' starch and protein content. Cano-Chauca et al. (2005) studied the 366 solubility of mango juice powders in different carriers. This result is consistent with their findings 367 that reported that the solubility of mango juice powder produced with maltodextrin as the carrier 368 agent was greater than 90%. Dispersibility measures the rehydration ability of powders. The higher 369 the dispersibility, the better the sample reconstitutes in water (Schuck, 2011). The dispersibility 370 index did not increase as the concentration of millet increased (Table 3). High dispersibility values 371 observed for the powders imply that the powder samples are highly soluble in water. Li & Zhong 372 (2021) investigated the rehydration properties of spray-dried millet protein isolates and reported 373 improved rehydration properties with the addition of soluble caseins. 374 375 3.6. Density, porosity, and flow properties 376 The results for the bulk density of the spray-dried powders are shown in Table 4. The spray-dried 377 powders' bulk density ranged from 253.89 ± 12.71 Kgm3 for BAW25 to 266.50 ± 7.96 Kgm3 for 378 LAW25. Chever et al. (2017) reported similar bulk density values for non-agglomerated sweet 379 whole milk powders. Porosity values were measured based on the relationship between bulk and 380 true densities. The mean porosity results ranged from 82.21 ± 0.12 % for LAW10 to 81.10 ± 1.02 381 % for LAW25. Similar results were reported for milk protein isolate powders (Ji et al., 2015) and 382 spray-dried dairy protein powders (Fournaise et al., 2021). Li & Zhong (2021) reported higher 383 porosity values for milk protein isolate powders with increased content of soluble caseins. 384 Carr's compressibility index (CI), Hausner ratio (HR), and angle of repose were used to 385 determine the flow properties of the powders, and the results are presented in Table 4. 3.6. Density, porosity, and flow properties 376 The results for the bulk density of the spray-dried powders are shown in Table 4. The spray-dried 377 powders' bulk density ranged from 253.89 ± 12.71 Kgm3 for BAW25 to 266.50 ± 7.96 Kgm3 for 378 LAW25. Chever et al. (2017) reported similar bulk density values for non-agglomerated sweet 379 whole milk powders. Porosity values were measured based on the relationship between bulk and 380 true densities. The mean porosity results ranged from 82.21 ± 0.12 % for LAW10 to 81.10 ± 1.02 381 % for LAW25. Similar results were reported for milk protein isolate powders (Ji et al., 2015) and 382 spray-dried dairy protein powders (Fournaise et al., 2021). Li & Zhong (2021) reported higher 383 porosity values for milk protein isolate powders with increased content of soluble caseins. 384 Carr's compressibility index (CI), Hausner ratio (HR), and angle of repose were used to 385 determine the flow properties of the powders, and the results are presented in Table 4. Powders 386 with a Hausner ratio greater than 1.34 are considered to be cohesive and less free-flowing 387 (Hausner, 1967), while powders with a Carr index greater than 25% are considered non-free- 388 flowing (Carr, 1965). The angle of repose values less than 30° indicate the very free-flowing nature 389 of the powders. There was a significant difference (p > 0.05) in the CI and HR values for all the 390 13 powders. The HR flow character for the powders ranged from fair for BAW25 to excellent for 391 LAW10. All the powders had an angle of repose value <30°, indicating free-flowing powders. In 392 this study, all the powders exhibited good flow properties. However, it was observed that higher 393 millet concentrations slightly reduced the flowability of the powder. 394 395 3.7. Particle size distribution and Zeta potential 396 It is defined as the temperature at which a glassy (amorphous) material becomes rubbery 424 (sticky) (Masum et al., 2020). The DSC thermograms and Tg and melting temperature of the spray- 425 dried powders are presented in Figure 4 and Table 6. The Tg of the AWM powders significantly 426 differed (P > 0.05) with different AW Millet ratios. Powders Tg increased with increasing 427 concentration of millet flour. Food products created using rapid heat exchanging spray drying 428 techniques are typically not thermodynamically stable. They go through a transition that causes 429 physical changes such as sticking and lumping, particularly with high sugar and acid ingredients, 430 resulting in low recovery, storage difficulties, and product quality degradation. This study shows 431 that the addition of millet flour increased the Tg of the powders, hence reducing the chances of 432 particle adhesion to the walls of the spray dryer and particle to particle stickiness. In Figure 4B, 433 higher Tg increased product yield for barnyard millet formulations. Interestingly that was not the 434 case for LAW25, which showed the highest yield with a lower Tg. Similar Tg was reported for 435 nano-filtered lactic acid whey powders (Bédas et al., 2017). 436 437 3.9. Particle morphology 438 The SEM images of the spray-dried powers (Figure 5.) were acquired at1000x, 4000x, and 10000x 439 magnification to visualize the aggregation and surface morphology. Less agglomeration and a 440 smoother surface were observed in BAW powders than in the LAW powder. This could be due to 441 the slightly higher Tg observed for BAW and the encapsulating effect of barnyard millet flour. 442 At 1000x, 4000x, and 10000x magnification, the AW powders showed the most agglomeration, 443 and a rougher surface when compared to the BAW powders. Interestingly LAW10, which had a 444 similar Tg to BAW10 had a more irregular surface and agglomerated particles than BAW10. A 445 possible explanation could be the sample preparation process during the mounting of the particles 446 for SEM imaging. These observations show that the addition of millet flour to acid whey had an 447 encapsulating effect on AW particles and reduced the tendency of the powders to agglomerate. 448 Other researchers also observed similar spherical shapes and surface smoothness in spray-dried 449 whey protein and maltodextrin particles (Both et al., 2020). 450 451 3 10 A ti id t A ti it 452 3.7. Particle size distribution and Zeta potential 396 In Figure 3. the size distribution and zeta potentials of the spray-dried powders are presented. The 397 width of the size distribution was also evaluated by measuring the polydispersity indexes (PDI) 398 and span values. PDI values less than 0.1 imply monodisperse particles and values greater than 0.1 399 imply polydisperse particle size distribution (Raval et al., 2019). PDI values for BAW25, LAW10, 400 and LAW25 powders indicate a monodispersed particle size. AW and BAW10 PDI values 401 indicated a polydisperse particle size distribution. Relatedly, spray-dried powders with span values 402 < 2 indicate a narrow size distribution while span values > 2 indicate particle agglomerations (Jiang 403 et al., 2020). The size distributions of the millet-based powders increased as the concentration of 404 millets increased, with span values ranging from 1.29 for BAW25 to 5.50 for BAW10. The lower 405 span index of BAW25 could be attributed to the fact that the millet flour acted as a crucial outer 406 wall material that modified the surface properties, allowed for free movement of the particles, and 407 reduced stickiness and aggregation of the powders (Esparza et al., 2020). The particle size differed 408 significantly among the spray-dried powders, with multiple peaks ranging between 0.001 to 100 409 µm. The D50 (µm) values of the AWM powders ranged from 0.90 ± 0.15 µm for BAW10 to 1.47 410 ± 0.17 µm for LAW25. These particles are smaller than those reported for spray-dried maltodextrin 411 and whey protein mixtures (Both et al., 2020), possibly due to differences in the feed composition, 412 the spray dryer, and spray-drying conditions. 413 Zeta potential measurements were carried out to determine the stability of the powder formulations 414 in an emulsion. According to the results, the powders' zeta potential values (Figure 3B) were 415 negative, ranging from −10.23 mV for BAW25 to −8.06 mV for LAW10. Based on studies by 416 García-Márquez et al. (2017), zeta potential values of ±30 mV are required to ensure a highly 417 stable emulsion. The results show electrostatic repulsive forces between particles which could 418 indicate some agglomeration tendency, and that the powder particles could be reconstituted in an 419 aqueous solution with moderate dispersion stability. 420 421 14 3.9. Glass transition temperature 422 Glass transition temperature (Tg) is one of the most important physical properties of powdered 423 materials. 3.9. Glass transition temperature 422 p Glass transition temperature (Tg) is one of the most important physical properties of powdered 423 materials. It is defined as the temperature at which a glassy (amorphous) material becomes rubbery 424 (sticky) (Masum et al., 2020). The DSC thermograms and Tg and melting temperature of the spray- 425 dried powders are presented in Figure 4 and Table 6. The Tg of the AWM powders significantly 426 differed (P > 0.05) with different AW Millet ratios. Powders Tg increased with increasing 427 concentration of millet flour. Food products created using rapid heat exchanging spray drying 428 techniques are typically not thermodynamically stable. They go through a transition that causes 429 physical changes such as sticking and lumping, particularly with high sugar and acid ingredients, 430 resulting in low recovery, storage difficulties, and product quality degradation. This study shows 431 that the addition of millet flour increased the Tg of the powders, hence reducing the chances of 432 particle adhesion to the walls of the spray dryer and particle to particle stickiness. In Figure 4B, 433 higher Tg increased product yield for barnyard millet formulations. Interestingly that was not the 434 case for LAW25, which showed the highest yield with a lower Tg. Similar Tg was reported for 435 nano-filtered lactic acid whey powders (Bédas et al., 2017). 436 3.9. Particle morphology 438 Similar results for phenolic content were reported for 461 microencapsulated turmeric milk powders (Ipar et al., 2022). The TBARS concentration was 462 highest in BAW10 at 3.57 𝜇M/g and lowest in BAW25 at 2.05 𝜇M/g. Similar results were reported 463 by Baublis et al. (2000) for wheat flour. 464 The powders' total phenolic content (TPC), DPPH activity, and TBARS value are presented in 453 Figure 6. The TPC was determined using the Folin Ciocalteu (FC) method, and the results are 454 expressed in terms of the gallic acid equivalent (GAE) in mg/g of the extract. The highest TPC 455 was found in BAW25 (0.600 mg GAE/g) and the lowest in LAW 10 (0.345 mg GAE/g). The 456 DPPH value of LAW25m(43.86 ± 0.24) and BAW10 (43.10 ± 0.88) were significantly (p < 0.05) 457 higher than that of the rest of the powders. As expected, formulations containing millets had the 458 highest TPC and DPPH values. Pure AW powders also exhibited some antioxidant activity in terms 459 of TPC and DPPH activity. This could be attributed to the phenolic compound present in plant 460 feed for animals (Zeb, 2021). Similar results for phenolic content were reported for 461 microencapsulated turmeric milk powders (Ipar et al., 2022). The TBARS concentration was 462 highest in BAW10 at 3.57 𝜇M/g and lowest in BAW25 at 2.05 𝜇M/g. Similar results were reported 463 by Baublis et al. (2000) for wheat flour. 464 3.9. Particle morphology 438 The SEM images of the spray-dried powers (Figure 5.) were acquired at1000x, 4000x, and 10000x 439 magnification to visualize the aggregation and surface morphology. Less agglomeration and a 440 smoother surface were observed in BAW powders than in the LAW powder. This could be due to 441 the slightly higher Tg observed for BAW and the encapsulating effect of barnyard millet flour. 442 The SEM images of the spray-dried powers (Figure 5.) were acquired at1000x, 4000x, and 10000x 439 magnification to visualize the aggregation and surface morphology. Less agglomeration and a 440 smoother surface were observed in BAW powders than in the LAW powder. This could be due to 441 the slightly higher Tg observed for BAW and the encapsulating effect of barnyard millet flour. 442 At 1000x, 4000x, and 10000x magnification, the AW powders showed the most agglomeration, 443 and a rougher surface when compared to the BAW powders. Interestingly LAW10, which had a 444 similar Tg to BAW10 had a more irregular surface and agglomerated particles than BAW10. A 445 possible explanation could be the sample preparation process during the mounting of the particles 446 for SEM imaging. These observations show that the addition of millet flour to acid whey had an 447 encapsulating effect on AW particles and reduced the tendency of the powders to agglomerate. 448 Other researchers also observed similar spherical shapes and surface smoothness in spray-dried 449 whey protein and maltodextrin particles (Both et al., 2020). 450 451 3.10. Antioxidant Activity 452 15 The powders' total phenolic content (TPC), DPPH activity, and TBARS value are presented in 453 Figure 6. The TPC was determined using the Folin Ciocalteu (FC) method, and the results are 454 expressed in terms of the gallic acid equivalent (GAE) in mg/g of the extract. The highest TPC 455 was found in BAW25 (0.600 mg GAE/g) and the lowest in LAW 10 (0.345 mg GAE/g). The 456 DPPH value of LAW25m(43.86 ± 0.24) and BAW10 (43.10 ± 0.88) were significantly (p < 0.05) 457 higher than that of the rest of the powders. As expected, formulations containing millets had the 458 highest TPC and DPPH values. Pure AW powders also exhibited some antioxidant activity in terms 459 of TPC and DPPH activity. This could be attributed to the phenolic compound present in plant 460 feed for animals (Zeb, 2021). 4. CONCLUSION 466 The present study demonstrates the feasibility of spray drying acid whey with the addition 467 of millet flours to produce microencapsulated acid whey powders with desired powder and 468 antioxidant properties. The addition of millet to acid whey increased the yields of the AWM spray- 469 dried powder samples. AWM powders had a higher whiteness index when compared to pure AW 470 powder. However, results showed the concentration of millet added could influence the whiteness 471 as powders with 25% millets had a lower whiteness index than powders with 10% millet. The 472 powders' color, thermal, reconstitution, and antioxidant properties were improved by incorporating 473 millet flour into the feed solution. The addition of millet slightly increased the hygroscopicity of 474 the powders. Improved flow and rehydration properties of the spray-dried powders were observed 475 with the addition of 10% millet flour. The addition of millet flour increased the Tg, resulting in 476 less sticky powder when spray drying at higher temperatures. The major physical mechanism of 477 millet flours in improving the physical and functional properties of spray-dried AWM powders 478 was attributed to the neutralization of lactic acid in the feed solution, and millet flour as a wall 479 material aided the spray drying of acid whey and reduced stickiness in the powders. Also, AWM 480 powders had a high concentration of phenolic compounds and exhibited high antioxidant activity, 481 which suggests that they could be incorporated as an ingredient in functional foods. 482 16 Reducing dairy waste has become a significant challenge to the dairy industry due to the 483 large volume of whey produced and its disposal difficulty. To tackle this challenge, researchers 484 and industry must explore sustainable approaches for utilizing acid whey and creating a circular 485 system in which new value from acid whey can be created with already existing resources used in 486 dairy product manufacturing. In this study, the addition of millet to acid whey resulted in a white 487 acid whey millet powder with good properties. This works shows that a functional powder could 488 be manufactured by varying millet acid whey concentrations, presenting an opportunity to utilize 489 acid whey in the dairy industry. 490 491 STATEMENTS AND DECLARATIONS 492 Conflict of Interest Statement 493 The authors declare no competing interests. 494 495 Funding Declaration 496 This work was supported by the U.S. 4. CONCLUSION 466 490 Conflict of Interest Statement 493 The authors declare no competing interests. 494 495 Funding Declaration 496 This work was supported by the U.S. Department of Agriculture, Agricultural Research Hatch 497 Funds (MO-HAFE0003) and Food Engineering and Sustainable Technologies (FEAST), 498 University of Missouri, Columbia. 499 500 Data Availability Statement 501 All data generated or analyzed during this study can be made available on reasonable request. 502 503 Author’s Contributions 504 Mercy Nani: formal analysis, investigation, visualization, and writing (original draft, reviewing, 505 and editing). Kiruba Krishnaswamy: conceptualization, resource management, methodology, 506 visualization, writing (review and editing), supervision, funding acquisition, and project 507 administration. 508 509 Acknowledgment 510 We are grateful for support from the U.S. Department of Agriculture (USDA) Hatch Funds (MO- 511 HAFE0003), and Food Engineering and Sustainable Technologies (FEAST), University of 512 4. CONCLUSION 466 Department of Agriculture, Agricultural Research Hatch 497 Funds (MO-HAFE0003) and Food Engineering and Sustainable Technologies (FEAST), 498 University of Missouri, Columbia. 499 500 Data Availability Statement 501 All data generated or analyzed during this study can be made available on reasonable request. 502 503 Author’s Contributions 504 Mercy Nani: formal analysis, investigation, visualization, and writing (original draft, reviewing, 505 and editing). Kiruba Krishnaswamy: conceptualization, resource management, methodology, 506 visualization, writing (review and editing), supervision, funding acquisition, and project 507 administration. 508 509 Acknowledgment 510 We are grateful for support from the U.S. Department of Agriculture (USDA) Hatch Funds (MO- 511 HAFE0003), and Food Engineering and Sustainable Technologies (FEAST), University of 512 Reducing dairy waste has become a 483 large volume of whey produced and its dis 484 and industry must explore sustainable appr 485 system in which new value from acid whey 486 dairy product manufacturing. In this study, 487 acid whey millet powder with good propert 488 be manufactured by varying millet acid wh 489 acid whey in the dairy industry. 490 491 STATEMENTS AND DECLARATIONS 492 Conflict of Interest Statement 493 The authors declare no competing interests. 494 495 Funding Declaration 496 This work was supported by the U.S. Dep 497 Funds (MO-HAFE0003) and Food Eng 498 University of Missouri, Columbia. 499 500 Data Availability Statement 501 All data generated or analyzed during this s 502 503 Author’s Contributions 504 Mercy Nani: formal analysis, investigation 505 and editing). Kiruba Krishnaswamy: con 506 visualization, writing (review and editin 507 administration. 508 509 Acknowledgment 510 We are grateful for support from the U.S. D 511 HAFE0003), and Food Engineering and 512 Reducing dairy waste has become a significant challenge to the dairy industry due to the 483 large volume of whey produced and its disposal difficulty. To tackle this challenge, researchers 484 and industry must explore sustainable approaches for utilizing acid whey and creating a circular 485 system in which new value from acid whey can be created with already existing resources used in 486 dairy product manufacturing. In this study, the addition of millet to acid whey resulted in a white 487 acid whey millet powder with good properties. This works shows that a functional powder could 488 be manufactured by varying millet acid whey concentrations, presenting an opportunity to utilize 489 acid whey in the dairy industry. Acknowledgment 510 We are grateful for support from the U.S. Department of Agriculture (USDA) Hatch Funds (MO- 511 HAFE0003), and Food Engineering and Sustainable Technologies (FEAST), University of 512 17 Missouri, to carry out this work. Graphical abstract was created with https://biorender.com 513 (YO243OCQI7). 514 515 References 516 Anitha, S., Kane-Potaka, J., Botha, R., Givens, D. I., Sulaiman, N. L. B., Upadhyay, S., 517 Vetriventhan, M., Tsusaka, T. W., Parasannanavar, D. J., Longvah, T., Rajendran, A., 518 Subramaniam, K., & Bhandari, R. K. (2021). Millets Can Have a Major Impact on 519 Improving Iron Status, Hemoglobin Level, and in Reducing Iron Deficiency Anemia–A 520 Systematic Review and Meta-Analysis. Frontiers in Nutrition, 8, 712. 521 https://doi.org/10.3389/FNUT.2021.725529/BIBTEX 522 Baublis, A., Decker, E. A., & Clydesdale, F. M. (2000). Antioxidant effect of aqueous extracts 523 from wheat based ready-to-eat breakfast cereals. Food Chemistry, 68(1), 1–6. 524 https://doi.org/10.1016/S0308-8146(99)00142-9 525 Beakawi Al-Hashemi, H. M., & Baghabra Al-Amoudi, O. S. (2018). A review on the angle of 526 repose of granular materials. In Powder Technology (Vol. 330, pp. 397–417). Elsevier B.V 527 https://doi.org/10.1016/j.powtec.2018.02.003 528 Bédas, M., Tanguy, G., Dolivet, A., Méjean, S., Gaucheron, F., Garric, G., Senard, G., Jeantet, 529 R., & Schuck, P. (2017). Nanofiltration of lactic acid whey prior to spray drying: Scaling up 530 to a semi-industrial scale. LWT - Food Science and Technology, 79, 355–360. 531 https://doi.org/10.1016/J.LWT.2017.01.061 532 Bhandari, B. R., Patel, K. C., & Chen, X. D. (2008). Spray drying of food materials-process and 533 product characteristics. In Drying technologies in food processing (pp. 113–157). Blackwel 534 Publishing Ltd. 535 Both, E. M., Boom, R. M., & Schutyser, M. A. I. (2020). Particle morphology and powder 536 properties during spray drying of maltodextrin and whey protein mixtures. Powder 537 Technology, 363, 519–524. https://doi.org/10.1016/J.POWTEC.2020.01.001 538 Botrel, D. A., de Barros Fernandes, R. V., Borges, S. V., & Yoshida, M. I. (2014). Influence of 539 Baublis, A., Decker, E. A., & Clydesdale, F. M. (2000). Antioxidant effect of aqueous extracts 523 from wheat based ready-to-eat breakfast cereals. Food Chemistry, 68(1), 1–6. 524 https://doi.org/10.1016/S0308-8146(99)00142-9 525 Beakawi Al-Hashemi, H. M., & Baghabra Al-Amoudi, O. S. (2018). A review on the angle of 526 repose of granular materials. In Powder Technology (Vol. 330, pp. 397–417). Elsevier B.V. 527 https://doi.org/10.1016/j.powtec.2018.02.003 528 Bédas, M., Tanguy, G., Dolivet, A., Méjean, S., Gaucheron, F., Garric, G., Senard, G., Jeantet, 529 R., & Schuck, P. (2017). Acknowledgment 510 Nanofiltration of lactic acid whey prior to spray drying: Scaling up 530 to a semi-industrial scale. LWT - Food Science and Technology, 79, 355–360. 531 https://doi.org/10.1016/J.LWT.2017.01.061 532 Bhandari, B. R., Patel, K. C., & Chen, X. D. (2008). Spray drying of food materials-process and 533 product characteristics. In Drying technologies in food processing (pp. 113–157). Blackwell 534 Publishing Ltd. 535 Both, E. M., Boom, R. M., & Schutyser, M. A. I. (2020). Particle morphology and powder 536 properties during spray drying of maltodextrin and whey protein mixtures. Powder 537 Technology, 363, 519–524. https://doi.org/10.1016/J.POWTEC.2020.01.001 538 Botrel, D. A., de Barros Fernandes, R. V., Borges, S. V., & Yoshida, M. I. (2014). Influence of 539 wall matrix systems on the properties of spray-dried microparticles containing fish oil. Food 540 Research International, 62, 344–352. https://doi.org/10.1016/J.FOODRES.2014.02.003 541 Bradley, R. L. (2010). Moisture and Total Solids Analysis. 85–104. https://doi.org/10.1007/978- 542 1-4419-1478-1_6 543 18 Bylund, G. (2015). Dairy processing handbook. In Tetra Pak Processing Systems AB (3rd ed.). 544 Tetra Pak. 545 Bylund, G. (2015). Dairy processing handbook. In Tetra Pak Processing Systems AB (3rd ed.). 544 Tetra Pak. 545 Cano-Chauca, M., Stringheta, P. C., Ramos, A. M., & Cal-Vidal, J. (2005). Effect of the carriers 546 on the microstructure of mango powder obtained by spray drying and its functional 547 characterization. Innovative Food Science & Emerging Technologies, 6(4), 420–428. 548 https://doi.org/10.1016/J.IFSET.2005.05.003 549 Carr, R. L. (1965). Evaluating flow properties of solids. Chemical Engineering, 72(2), 163–168. 550 https://jglobal.jst.go.jp/en/detail?JGLOBAL_ID=201602016398256191 551 Chandrapala, J., & Vasiljevic, T. (2017). Properties of spray dried lactose powders influenced by 552 presence of lactic acid and calcium. Journal of Food Engineering, 198, 63–71. 553 Chandrapala, J., & Vasiljevic, T. (2018). Feasibility of Spray Drying Concentrated Acid Whey 555 After Nanofiltration. Food and Bioprocess Technology 2018 11:8, 11(8), 1505–1515. 556 https://doi.org/10.1007/S11947-018-2118-1 557 Chever, S., Méjean, S., Dolivet, A., Mei, F., Den Boer, C. M., Le Barzic, G., Jeantet, R., & 558 Schuck, P. (2017). Agglomeration during spray drying: Physical and rehydration properties 559 of whole milk/sugar mixture powders. LWT - Food Science and Technology, 83, 33–41. 560 https://doi.org/10.1016/J.LWT.2017.05.002 561 Esparza, Y., Ngo, T. D., & Boluk, Y. (2020). Preparation of powdered oil particles by spray 562 drying of cellulose nanocrystals stabilized Pickering hempseed oil emulsions. Colloids and 563 Surfaces A: Physicochemical and Engineering Aspects, 598, 124823. Acknowledgment 510 564 https://doi.org/10.1016/J.COLSURFA.2020.124823 565 p g p p p y p y drying of cellulose nanocrystals stabilized Pickering hempseed oil emulsions. Colloids and 563 Surfaces A: Physicochemical and Engineering Aspects, 598, 124823. 564 https://doi.org/10.1016/J.COLSURFA.2020.124823 565 Fang, Z., & Bhandari, B. (2012). Comparing the efficiency of protein and maltodextrin on spray 566 drying of bayberry juice. Food Research International, 48(2), 478–483. 567 https://doi.org/10.1016/J.FOODRES.2012.05.025 568 Fernandes, R. V. de B., Silva, E. K., Borges, S. V., de Oliveira, C. R., Yoshida, M. I., da Silva, 569 Y. F., do Carmo, E. L., Azevedo, V. M., & Botrel, D. A. (2016). Proposing Novel 570 Encapsulating Matrices for Spray-Dried Ginger Essential Oil from the Whey Protein 571 Isolate-Inulin/Maltodextrin Blends. Food and Bioprocess Technology 2016 10:1, 10(1), 572 115–130. https://doi.org/10.1007/S11947-016-1803-1 573 Fournaise, T., Burgain, J., Perroud-Thomassin, C., & Petit, J. (2021). Impact of the whey 574 Fang, Z., & Bhandari, B. (2012). Comparing the efficiency of protein and maltodextrin on spray 566 drying of bayberry juice. Food Research International, 48(2), 478–483. 567 https://doi.org/10.1016/J.FOODRES.2012.05.025 568 Fernandes, R. V. de B., Silva, E. K., Borges, S. V., de Oliveira, C. R., Yoshida, M. I., da Silva, 569 Y. F., do Carmo, E. L., Azevedo, V. M., & Botrel, D. A. (2016). Proposing Novel 570 Encapsulating Matrices for Spray-Dried Ginger Essential Oil from the Whey Protein 571 Isolate-Inulin/Maltodextrin Blends. Food and Bioprocess Technology 2016 10:1, 10(1), 572 115–130. https://doi.org/10.1007/S11947-016-1803-1 573 Fournaise, T., Burgain, J., Perroud-Thomassin, C., & Petit, J. (2021). Impact of the whey 574 19 protein/casein ratio on the reconstitution and flow properties of spray-dried dairy protein 575 powders. Powder Technology, 391, 275–281. 576 https://doi.org/10.1016/J.POWTEC.2021.06.026 577 Gagneten, M., Corfield, R., Mattson, M. G., Sozzi, A., Leiva, G., Salvatori, D., & Schebor, C. 578 (2019). Spray-dried powders from berries extracts obtained upon several processing steps to 579 improve the bioactive components content. Powder Technology, 342, 1008–1015. 580 https://doi.org/10.1016/J.POWTEC.2018.09.048 581 García-Márquez, E., Higuera-Ciapara, I., & Espinosa-Andrews, H. (2017). Design of fish oil-in- 582 water nanoemulsion by microfluidization. Innovative Food Science & Emerging 583 Technologies, 40, 87–91. https://doi.org/10.1016/J.IFSET.2016.11.007 584 Hariprasanna, K., & Rakshit, S. (2016). Economic Importance of Sorghum (pp. 1–25). Springer, 585 Cham. https://doi.org/10.1007/978-3-319-47789-3_1 586 Hausner, H. H. (1967). FRICTION CONDITIONS IN A MASS OF METAL POWDER. Int. J. 587 Powder Met, 3(4), 7–13. Acknowledgment 510 https://www.osti.gov/biblio/4566075-friction-conditions-mass- 588 metal-powder 589 Horvat, D., Šimić, G., Drezner, G., Lalić, A., Ledenčan, T., Tucak, M., Plavšić, H., Andrić, L., & 590 Zd ić Z (2020) Ph li A id P fil d A ti id t A ti it f M j C l C 591 Horvat, D., Šimić, G., Drezner, G., Lalić, A., Ledenčan, T., Tucak, M., Plavšić, H., Andrić, L., & 590 , , , , , , , , , , , , , , , , Zdunić, Z. (2020). Phenolic Acid Profiles and Antioxidant Activity of Major Cereal Crops. 591 Antioxidants, 9(6), 527. https://doi.org/10.3390/antiox9060527 592 Ipar, V. S., Singhal, R. S., & Devarajan, P. V. (2022). An innovative approach using 593 microencapsulated turmeric oleoresin to develop ready-to-use turmeric milk powder with 594 enhanced oral bioavailability. Food Chemistry, 373, 131400. 595 https://doi.org/10.1016/J.FOODCHEM.2021.131400 596 https://doi.org/10.1016/J.FOODCHEM.2021.131400 596 Jayasundera, M., Adhikari, B., Howes, T., & Aldred, P. (2011). Surface protein coverage and its 597 implications on spray-drying of model sugar-rich foods: Solubility, powder production and 598 characterisation. Food Chemistry, 128(4), 1003–1016. 599 https://doi.org/10.1016/J.FOODCHEM.2011.04.006 600 Ji, J., Cronin, K., Fitzpatrick, J., Fenelon, M., & Miao, S. (2015). Effects of fluid bed 601 agglomeration on the structure modification and reconstitution behaviour of milk protein 602 isolate powders. Journal of Food Engineering, 167, 175–182. 603 https://doi.org/10.1016/J.JFOODENG.2015.01.012 604 Jiang, N., Dev Kumar, G., Chen, J., Mishra, A., & Mis Solval, K. (2020). Comparison of 605 Jayasundera, M., Adhikari, B., Howes, T., & Aldred, P. (2011). Surface protein coverage and its 597 implications on spray-drying of model sugar-rich foods: Solubility, powder production and 598 characterisation. Food Chemistry, 128(4), 1003–1016. 599 https://doi.org/10.1016/J.FOODCHEM.2011.04.006 600 Ji, J., Cronin, K., Fitzpatrick, J., Fenelon, M., & Miao, S. (2015). Effects of fluid bed 601 agglomeration on the structure modification and reconstitution behaviour of milk protein 602 isolate powders. Journal of Food Engineering, 167, 175–182. 603 https://doi.org/10.1016/J.JFOODENG.2015.01.012 604 Jiang, N., Dev Kumar, G., Chen, J., Mishra, A., & Mis Solval, K. (2020). Comparison of 605 20 concurrent and mixed-flow spray drying on viability, growth kinetics and biofilm formation 606 of Lactobacillus rhamnosus GG microencapsulated with fish gelatin and maltodextrin. LWT, 607 124, 109200. https://doi.org/10.1016/J.LWT.2020.109200 608 concurrent and mixed-flow spray drying on viability, growth kinetics and biofilm formation 606 of Lactobacillus rhamnosus GG microencapsulated with fish gelatin and maltodextrin. LWT, 607 124, 109200. https://doi.org/10.1016/J.LWT.2020.109200 608 606 Jinapong, N., Suphantharika, M., & Jamnong, P. (2008). Acknowledgment 510 Production of instant soymilk powders 609 by ultrafiltration, spray drying and fluidized bed agglomeration. Journal of Food 610 Engineering, 84(2), 194–205. https://doi.org/10.1016/j.jfoodeng.2007.04.032 611 Jinapong, N., Suphantharika, M., & Jamnong, P. (2008). Production of instant soymilk powders 609 by ultrafiltration, spray drying and fluidized bed agglomeration. Journal of Food 610 Engineering, 84(2), 194–205. https://doi.org/10.1016/j.jfoodeng.2007.04.032 611 Kumar, A., Tomer, V., Kaur, A., Kumar, V., & Gupta, K. (2018). Millets: A solution to agrarian 612 and nutritional challenges. In Agriculture and Food Security (Vol. 7, Issue 1, p. 31). 613 BioMed Central Ltd. https://doi.org/10.1186/s40066-018-0183-3 614 Li, N., & Zhong, Q. (2021). Improving rehydration properties of spray-dried milk protein 615 isolates by supplementing soluble caseins. Food Research International, 150, 110770. 616 https://doi.org/10.1016/J.FOODRES.2021.110770 617 isolates by supplementing soluble caseins. Food Research International, 150, 110770. 616 https://doi.org/10.1016/J.FOODRES.2021.110770 617 Li, X., & Siddique, K. H. M. (2020). Future Smart Food: Harnessing the potential of neglected 618 and underutilized species for Zero Hunger. Maternal & Child Nutrition, 16(S3). 619 https://doi.org/10.1111/mcn.13008 620 Malik, C. P., & Singh, M. B. (1980). Plant enzymology and histo-enzymology : a text manual. 621 Kalyani Publishers. 622 Malik, S., Krishnaswamy, K., & Mustapha, A. (2021). Physical properties of complementary 623 food powder obtained from upcycling of Greek yogurt acid whey with kodo and proso 624 millets. Journal of Food Process Engineering, 44(11), e13878. 625 htt //d i /10 1111/JFPE 13878 626 Masum, A. K. M., Huppertz, T., Chandrapala, J., Adhikari, B., & Zisu, B. (2020). 627 Physicochemical properties of spray-dried model infant milk formula powders: Influence of 628 whey protein-to-casein ratio. International Dairy Journal, 100, 104565. 629 https://doi.org/10.1016/J.IDAIRYJ.2019.104565 630 Menchik, P., Zuber, T., Zuber, A., & Moraru, C. I. (2019). Short communication: Composition 631 of coproduct streams from dairy processing: Acid whey and milk permeate. Journal of 632 Dairy Science, 102(5), 3978–3984. https://doi.org/10.3168/JDS.2018-15951 633 Raval, N., Maheshwari, R., Kalyane, D., Youngren-Ortiz, S. R., Chougule, M. B., & Tekade, R. 634 K. (2019). Importance of Physicochemical Characterization of Nanoparticles in 635 Raval, N., Maheshwari, R., Kalyane, D., Youngren Ortiz, S. R., Chougule, M. B., & Tekade, R. 634 K. (2019). Importance of Physicochemical Characterization of Nanoparticles in 635 Pharmaceutical Product Development. In Basic Fundamentals of Drug Delivery (pp. 369– 636 21 400). Academic Press. https://doi.org/10.1016/B978-0-12-817909-3.00010-8 637 Rocha-Mendoza, D., Kosmerl, E., Krentz, A., Zhang, L., Badiger, S., Miyagusuku-Cruzado, G., 638 Mayta-Apaza, A., Giusti, M., Jiménez-Flores, R., & García-Cano, I. (2021). Acknowledgment 510 Invited review: 639 Acid whey trends and health benefits. Journal of Dairy Science, 104(2), 1262–1275. 640 https://doi.org/10.3168/JDS.2020-19038 641 Romano, N., Ureta, M. M., Guerrero-Sánchez, M., & Gómez-Zavaglia, A. (2020). Nutritional 642 and technological properties of a quinoa (Chenopodium quinoa Willd.) spray-dried 643 powdered extract. Food Research International, 129, 108884. 644 https://doi.org/10.1016/J.FOODRES.2019.108884 645 https://doi.org/10.1016/J.FOODRES.2019.108884 645 Saleh, A. S. M., Zhang, Q., Chen, J., & Shen, Q. (2013). Millet Grains: Nutritional Quality, 646 Processing, and Potential Health Benefits. Comprehensive Reviews in Food Science and 647 Food Safety, 12(3), 281–295. https://doi.org/10.1111/1541-4337.12012 648 Santhalakshmy, S., Don Bosco, S. J., Francis, S., & Sabeena, M. (2015). Effect of inlet 649 temperature on physicochemical properties of spray-dried jamun fruit juice powder. Powder 650 Santhalakshmy, S., Don Bosco, S. J., Francis, S., & Sabeena, M. (2015). Effect of inlet 649 temperature on physicochemical properties of spray-dried jamun fruit juice powder. Powder 650 Technology, 274, 37–43. https://doi.org/10.1016/J.POWTEC.2015.01.016 651 temperature on physicochemical properties of spray-dried jamun fruit juice powder. Powder 650 Technology, 274, 37–43. https://doi.org/10.1016/J.POWTEC.2015.01.016 651 Sarabandi, K., Peighambardoust, S. H., Sadeghi Mahoonak, A. R., & Samaei, S. P. (2018). 652 Effect of different carriers on microstructure and physical characteristics of spray dried 653 apple juice concentrate. Journal of Food Science and Technology, 55(8), 3098. 654 https://doi.org/10.1007/S13197-018-3235-6 655 Schmidt, R. H., Packard, V. S., & Morris, H. A. (1984). Effect of Processing on Whey Protein 656 Functionality. Journal of Dairy Science, 67(11), 2723–2733. 657 https://doi.org/10.3168/JDS.S0022-0302(84)81630-6 658 Schuck, P. (2011). Dehydrated Dairy Products \| Milk Powder: Physical and Functional Properties 659 of Milk Powders. Encyclopedia of Dairy Sciences: Second Edition, 117–124. 660 https://doi.org/10.1016/B978-0-12-374407-4.00122-9 661 Singleton, V. L., & Rossi, J. A. (1965). Colorimetry of Total Phenolics with Phosphomolybdic- 662 Phosphotungstic Acid Reagents. American Journal of Enology and Viticulture, 16(3). 663 Singleton, V. L., & Rossi, J. A. (1965). Colorimetry of Total Phenolics with Phosphomolybdic- 662 Phosphotungstic Acid Reagents. American Journal of Enology and Viticulture, 16(3). 663 Phosphotungstic Acid Reagents. American Journal of Enology and Viticulture, 16(3). 663 Tapia, M. S., Alzamora, S. M., & Chirife, J. (2020). Effects of Water Activity ( a w ) on 664 Microbial Stability as a Hurdle in Food Preservation. In Water Activity in Foods (pp. 323– 665 355). Wiley. https://doi.org/10.1002/9781118765982.ch14 666 Microbial Stability as a Hurdle in Food Preservation. In Water Activity in Foods (pp. 323– 665 355). Wiley. https://doi.org/10.1002/9781118765982.ch14 666 Microbial Stability as a Hurdle in Food Preservation. Acknowledgment 510 In Water Activity in Foods (pp. 323– 665 355). Wiley. https://doi.org/10.1002/9781118765982.ch14 666 Taylor, J. R. N. (2018). Sorghum and millets: Taxonomy, history, distribution, and production. 667 Taylor, J. R. N. (2018). Sorghum and millets: Taxonomy, history, distribution, and production. 667 22 In Sorghum and Millets: Chemistry, Technology, and Nutritional Attributes (pp. Elsevier. https://doi.org/10.1016/B978-0-12-811527-5.00001-0 669 Elsevier. https://doi.org/10.1016/B978-0-12-811527-5.00001-0 9 The State of Food Security and Nutrition in the World 2021. (2021). The State of Food Security 670 and Nutrition in the World 2021. https://doi.org/10.4060/cb4474en 671 Willett, W., Rockström, J., Loken, B., Springmann, M., Lang, T., Vermeulen, S., Garnett, T., 672 Tilman, D., DeClerck, F., Wood, A., Jonell, M., Clark, M., Gordon, L. J., Fanzo, J., 673 Hawkes, C., Zurayk, R., Rivera, J. A., De Vries, W., Majele Sibanda, L., … Murray, C. J. 674 L. (2019). Food in the Anthropocene: the EAT–Lancet Commission on healthy diets from 675 sustainable food systems. The Lancet, 393(10170), 447–492. https://doi.org/10.1016/S0140- 676 6736(18)31788-4 677 Xiang, J., Apea-Bah, F. B., Ndolo, V. U., Katundu, M. C., & Beta, T. (2019). Profile of phenolic 678 compounds and antioxidant activity of finger millet varieties. Food Chemistry, 275, 361– 679 368. https://doi.org/10.1016/j.foodchem.2018.09.120 680 Yousaf, L., Hou, D., Liaqat, H., & Shen, Q. (2021). Millet: A review of its nutritional and 681 functional changes during processing. Food Research International, 142, 110197. 682 https://doi.org/10.1016/J.FOODRES.2021.110197 683 Zeb, A. (2021). Phenolic Antioxidants in Dairy Products. In Phenolic Antioxidants in Foods: 684 Chemistry, Biochemistry and Analysis (pp. 281–296). Springer, Cham. 685 https://doi.org/10.1007/978-3-030-74768-8_10 686 Zeb, A., & Ullah, F. (2016). A Simple Spectrophotometric Method for the Determination of 687 Thiobarbituric Acid Reactive Substances in Fried Fast Foods. Journal of Analytical 688 Methods in Chemistry, 2016. https://doi.org/10.1155/2016/9412767 689 Zhu, X., Yuan, X., Zhang, Y., Liu, H., Wang, J., & Sun, B. (2022). The global concern of food 690 security during the COVID-19 pandemic: Impacts and perspectives on food security. Food 691 Chemistry, 370, 130830. https://doi.org/10.1016/J.FOODCHEM.2021.130830 692 Zungur Bastıoğlu, A., Tomruk, D., Koç, M., & Ertekin, F. K. (2016). Spray dried melon seed 693 milk powder: physical, rheological and sensory properties. Journal of Food Science and 694 Technology, 53(5), 2396–2404. https://doi.org/10.1007/s13197-016-2214-z 695 Zungur Bastıoğlu, A., Tomruk, D., Koç, M., & Ertekin, F. K. (2016). Spray dried melon seed 693 milk powder: physical, rheological and sensory properties. Journal of Food Science and 694 Technology, 53(5), 2396–2404. https://doi.org/10.1007/s13197-016-2214-z 695 23 Table 1. Acknowledgment 510 The experimental design applied in the study 719 Samples Millet (g) Acid whey (ml) Concentration % W/V pH BAW10 10 100 10 4.37 BAW25 25 100 25 4.67 BAW50 50 100 50 4.70 BAW75 75 100 75 4.81 BAW100 100 100 100 5.01 LAW10 10 100 10 4.86 LAW25 25 100 25 5.05 LAW50 50 100 50 5.24 Table 1. The experimental design applied in the study 24 LAW75 75 100 75 5.30 LAW100 100 100 100 5.32 AW 0 100 Pure acid whey 4.28 Legend: AW, pure acid whey; BAW10, 10% w/v Barnyard millet acid whey; BAW25: 25% w/v 720 Barnyard millet acid whey; LAW10: 10% w/v Little millet acid whey; LAW25: 25% w/v Little 721 millet acid whey 722 Barnyard millet acid whey; LAW10: 10% w/v Little millet acid whey; LAW25: 25% w/v Little 721 millet acid whey 722 723 Table 2. Moisture content, water activity, and product yield of spray-dried powders. 724 Sample Moisture content (%) Water activity Product yield (%) AW 3.78 ± 0.03a 0.27 ± 0.00c 78.24 BAW10 3.99 ± 0.12a 0.29 ± 0.01d 85.80 BAW25 4.13 ± 0.12a 0.25 ± 0.00a 88.81 LAW10 5.13 ± 0.20b 0.26 ± 0.00b 77.82 LAW25 5.53 ± 0.13c 0.25 ± 0.00a 98.80 Different letters of each column show a significant difference at the level of p ≤ 0.05 725 Legend: AW, pure acid whey; BAW10, 10% w/v Barnyard millet acid whey; BAW25: 25% w/v 726 Barnyard millet acid whey; LAW10: 10% w/v Little millet acid whey; LAW25: 25 5 w/v Little 727 millet acid whey 728 729 Table 3. Color characterization of spray-dried millet acid whey (AWM) powders 730 Color parameters AW BAW10 BAW25 LAW10 LAW25 L* 102.76 ± 0.03 a 101.57 ± 0.04 b 101.78 ± 0.01c 99.19 ± 0.04d 99.60 ± 0. a* -0.50 ± 0.28a -0.16 ± 0.01b 0.28 ± 0.00c -0.20 ± 0.01d -0.03 ± 0.0 b* 6.28 ± 0.00b 6.30 ± 0.00b 6.36 ± 0.01b 6.02 ± 0.14a 7.04 ± 0.0 ∆E 3.53 ± 0.01a 4.16 ± 0.03b 4.07 ± 0.01c 5.89 ± 0.04d 6.11 ± 0.0 Chroma 6.30 ± 0.00a 2.52 ± 0.01b 2.54 ± 0.00c 2.46 ± 0.03d 2.65 ± 0.0 Table 2. Moisture content, water activity, and product yield of spray-dried powders. Acknowledgment 510 736 Properties AW BAW10 BAW25 LAW10 LAW25 Bulk density (Kg/m 3 ) 258.31 ± 0.00 a 264.27 ± 7.89 a 253.89± 12.71 a 259.96 ± 0.00 a 266.50 ± 7.96 a CI (%) 10.00 ± 0.00 a 18.59 ± 2.43 c 16.69 ± 2.96 bc 10.00 ± 0.00 a 11.93 ± 3.34 ab Hausner Ratio 1.11 ± 0.00 a 1.23 ± 0.03 c 1.20 ± 0.04 bc 1.11 ± 0.00 a 1.14 ± 0.04 ab Porosity 𝜀 (%) 81.02 ± 0.08 ab 81.91 ± 0.60 ab 81.93 ± 0.90 ab 82.29 ± 0.0 b 80.75 ± 0.56 a Angle of Repose (°) 15.15 ± 0.71 a 18.49 ± 2.29 bc 10.40 ± 1.87 ab 22.26 ± 4.18 c 17.90 ± 4.72 bc Water Solubility (%) 73.17 ± 0.33 a 85.96 ± 0.99 a 84.62 ± 3.32 a 82.38 ± 1.53 a 85.15 ± 2.77 a Dispersibility (%) 88.52 ± 5.87 a 89.30 ± 3.22 a 91.49 ± 4.95 a 88.72 ± 3.94 a 87.33 ± 1.50 a Hygroscopicity (%) 20.90 ± 0.23bc 19.97 ± 0.16 abc 19.45 ± 0.18 ab 19.81 ± 0.24 a 20.74 ± 0.39 c Values are the mean ± standard deviation of triplicate determinations. Means with different letters 737 on each row show a significant difference at the level of p ≤ 0.05 738 Legend: AW, pure acid whey; BAW10, 10% w/v Barnyard millet acid whey; BAW25: 25% w/v 739 Barnyard millet acid whey; LAW10: 10% w/v Little millet acid whey; LAW25: 25 5 w/v Little 740 millet acid whey 741 Table 4. Physical and functional characteristics of spray-dried powders. Acknowledgment 510 724 Sample Moisture content (%) Water activity Product yield (%) AW 3.78 ± 0.03a 0.27 ± 0.00c 78.24 BAW10 3.99 ± 0.12a 0.29 ± 0.01d 85.80 BAW25 4.13 ± 0.12a 0.25 ± 0.00a 88.81 LAW10 5.13 ± 0.20b 0.26 ± 0.00b 77.82 LAW25 5.53 ± 0.13c 0.25 ± 0.00a 98.80 Different letters of each column show a significant difference at the level of p ≤ 0.05 725 2. Moisture content, water activity, and product yield of spray-dried powders Table 3. Color characterization of spray-dried millet acid whey (AWM) powders 730 Color parameters AW BAW10 BAW25 LAW10 LAW25 L* 102.76 ± 0.03 a 101.57 ± 0.04 b 101.78 ± 0.01c 99.19 ± 0.04d 99.60 ± 0.08e a* -0.50 ± 0.28a -0.16 ± 0.01b 0.28 ± 0.00c -0.20 ± 0.01d -0.03 ± 0.01e b* 6.28 ± 0.00b 6.30 ± 0.00b 6.36 ± 0.01b 6.02 ± 0.14a 7.04 ± 0.04c ∆E 3.53 ± 0.01a 4.16 ± 0.03b 4.07 ± 0.01c 5.89 ± 0.04d 6.11 ± 0.06e Chroma 6.30 ± 0.00a 2.52 ± 0.01b 2.54 ± 0.00c 2.46 ± 0.03d 2.65 ± 0.01e Table 3. Color characterization of spray-dried millet acid whey (AWM) powders 30 Color parameters AW BAW10 BAW25 LAW10 LAW25 L* 102.76 ± 0.03 a 101.57 ± 0.04 b 101.78 ± 0.01c 99.19 ± 0.04d 99.60 ± 0.08e a* -0.50 ± 0.28a -0.16 ± 0.01b 0.28 ± 0.00c -0.20 ± 0.01d -0.03 ± 0.01e b* 6.28 ± 0.00b 6.30 ± 0.00b 6.36 ± 0.01b 6.02 ± 0.14a 7.04 ± 0.04c ∆E 3.53 ± 0.01a 4.16 ± 0.03b 4.07 ± 0.01c 5.89 ± 0.04d 6.11 ± 0.06e Chroma 6.30 ± 0.00a 2.52 ± 0.01b 2.54 ± 0.00c 2.46 ± 0.03d 2.65 ± 0.01e ble 3. Color characterization of spray-dried millet acid whey (AWM) powders 25 Hue angle (°) -85.44 ± 0.04a 88.55 ± 0.10b 87.48 ± 0.00c -88.14 ± 0.08d -89.74 ± 0.06e Whiteness index (WI) 93.12 ± 0.01a 93.51 ± 0.01b 93.39 ± 0.01c 93.92 ± 0.13d 92.95 ± 0.03e Different letters of each row show a significant difference at the level of p ≤ 0.05 731 Legend: AW, pure acid whey; BAW10, 10% w/v Barnyard millet acid whey; BAW25: 25% w/v 732 Barnyard millet acid whey; LAW10: 10% w/v Little millet acid whey; LAW25: 25 5 w/v Little 733 millet acid whey 734 735 Table 4. Physical and functional characteristics of spray-dried powders. Acknowledgment 510 744 Legend: AW, pure acid whey; BAW10, 10% w/v Barnyard millet acid whey; BAW25: 25% w/v 745 Barnyard millet acid whey; LAW10: 10% w/v Little millet acid whey; LAW25: 25 5 w/v Little 746 millet acid whey 747 748 749 750 751 752 753 754 755 756 Table 6. Glass transition and Melting temperatures of spray-dried AW and AWM powders 757 Sample Tg °C Melting temperature °C AW 74.98 ± 1.43 134.85 ± 1.38 BAW10 85.51 ± 1.91 149.39 ± 1.34 BAW25 76.33 ± 0.98 157.61 ± 6.52 LAW10 78.08 ± 3.95 149.18 ± 4.3 Sample Diameter (𝝁m) D10 (𝝁m) D 50 (𝜇m) D90 (𝜇m) Span PDI AW 1.23 ± 0.20 0.33 ± 0.12 0.96 ± 0.23 4.81 ± 4.52 3.87 ± 3.38 2.34± 3.94 BAW 10 1.04 ± 0.09 0.22 ± 0.10 0.90 ± 0.15 5.56 ± 2.97 5.50 ± 2.86 1.20 ± 2.18 BAW 25 1.26 ± 0.05 0.49 ± 0.04 0.90 ± 0.04 1.65 ± 0.00 1.29 ± 0.09 -2.49 ± 1.64 LAW 10 1.29 ± 0.19 0.43 ± 0.15 0.97 ± 0.14 2.55 ± 1.07 2.20 ± 1.18 -0.98 ± 1.43 LAW 25 1.30 ± 0.25 0.48 ± 0.14 1.47 ± 0.17 5.54 ± 2.40 3.27 ± 1.46 -1.59 ± 3.16 Values are the mean ± standard deviation of triplicate determinations. 744 Sample Diameter (𝝁m) D10 (𝝁m) D 50 (𝜇m) D90 (𝜇m) Span PDI AW 1.23 ± 0.20 0.33 ± 0.12 0.96 ± 0.23 4.81 ± 4.52 3.87 ± 3.38 2.34± 3.94 BAW 10 1.04 ± 0.09 0.22 ± 0.10 0.90 ± 0.15 5.56 ± 2.97 5.50 ± 2.86 1.20 ± 2.18 BAW 25 1.26 ± 0.05 0.49 ± 0.04 0.90 ± 0.04 1.65 ± 0.00 1.29 ± 0.09 -2.49 ± 1.64 LAW 10 1.29 ± 0.19 0.43 ± 0.15 0.97 ± 0.14 2.55 ± 1.07 2.20 ± 1.18 -0.98 ± 1.43 LAW 25 1.30 ± 0.25 0.48 ± 0.14 1.47 ± 0.17 5.54 ± 2.40 3.27 ± 1.46 -1.59 ± 3.16 Values are the mean ± standard deviation of triplicate determinations. 4 Legend: AW, pure acid whey; BAW10, 10% w/v Barnyard millet acid whey; BAW25: 25% w/v 5 Barnyard millet acid whey; LAW10: 10% w/v Little millet acid whey; LAW25: 25 5 w/v Little 6 millet acid whey 7 8 Table 6. Acknowledgment 510 736 Properties AW BAW10 BAW25 LAW10 LAW25 Bulk density (Kg/m 3 ) 258.31 ± 0.00 a 264.27 ± 7.89 a 253.89± 12.71 a 259.96 ± 0.00 a 266.50 ± 7.96 a CI (%) 10.00 ± 0.00 a 18.59 ± 2.43 c 16.69 ± 2.96 bc 10.00 ± 0.00 a 11.93 ± 3.34 ab Hausner Ratio 1.11 ± 0.00 a 1.23 ± 0.03 c 1.20 ± 0.04 bc 1.11 ± 0.00 a 1.14 ± 0.04 ab Porosity 𝜀 (%) 81.02 ± 0.08 ab 81.91 ± 0.60 ab 81.93 ± 0.90 ab 82.29 ± 0.0 b 80.75 ± 0.56 a Angle of Repose (°) 15.15 ± 0.71 a 18.49 ± 2.29 bc 10.40 ± 1.87 ab 22.26 ± 4.18 c 17.90 ± 4.72 bc Water Solubility (%) 73.17 ± 0.33 a 85.96 ± 0.99 a 84.62 ± 3.32 a 82.38 ± 1.53 a 85.15 ± 2.77 a Dispersibility (%) 88.52 ± 5.87 a 89.30 ± 3.22 a 91.49 ± 4.95 a 88.72 ± 3.94 a 87.33 ± 1.50 a Hygroscopicity (%) 20.90 ± 0.23bc 19.97 ± 0.16 abc 19.45 ± 0.18 ab 19.81 ± 0.24 a 20.74 ± 0.39 c Values are the mean ± standard deviation of triplicate determinations. Means with different letters 737 on each row show a significant difference at the level of p ≤ 0.05 738 Legend: AW, pure acid whey; BAW10, 10% w/v Barnyard millet acid whey; BAW25: 25% w/v 739 Table 4. Physical and functional characteristics of spray-dried powders. Legend: AW, pure acid whey; BAW10, 10% w/v Barnyard millet acid whey; BAW25: 25% w/v 39 Barnyard millet acid whey; LAW10: 10% w/v Little millet acid whey; LAW25: 25 5 w/v Little 40 millet acid whey 41 26 Sample Diameter (𝝁m) D10 (𝝁m) D 50 (𝜇m) D90 (𝜇m) Span PDI AW 1.23 ± 0.20 0.33 ± 0.12 0.96 ± 0.23 4.81 ± 4.52 3.87 ± 3.38 2.34± BAW 10 1.04 ± 0.09 0.22 ± 0.10 0.90 ± 0.15 5.56 ± 2.97 5.50 ± 2.86 1.20 ± BAW 25 1.26 ± 0.05 0.49 ± 0.04 0.90 ± 0.04 1.65 ± 0.00 1.29 ± 0.09 -2.49 LAW 10 1.29 ± 0.19 0.43 ± 0.15 0.97 ± 0.14 2.55 ± 1.07 2.20 ± 1.18 -0.98 LAW 25 1.30 ± 0.25 0.48 ± 0.14 1.47 ± 0.17 5.54 ± 2.40 3.27 ± 1.46 -1.59 Values are the mean ± standard deviation of triplicate determinations. Acknowledgment 510 Glass transition and Melting temperatures of spray-dried AW and AWM powders 757 Sample Tg °C Melting temperature °C AW 74.98 ± 1.43 134.85 ± 1.38 BAW10 85.51 ± 1.91 149.39 ± 1.34 BAW25 76.33 ± 0.98 157.61 ± 6.52 LAW10 78.08 ± 3.95 149.18 ± 4.3 27 LAW25 62.52 ±0.03 137.48 ± 1.95 Values are the mean ± standard deviation of triplicate determinations. 758 Legend: AW, pure acid whey; BAW10, 10% w/v Barnyard millet acid whey; BAW25: 25% w/v 759 Barnyard millet acid whey; LAW10: 10% w/v Little millet acid whey; LAW25: 25 5 w/v Little 760 millet acid whey 761 762 763 764 765 766 767 768 769 770 Figure 1. Whiteness index of AW, BAW10, BAW20, LAW10, and LAW25 spray-dried 771 powders 772 Legend: AW, pure acid whey; BAW10, 10% w/v Barnyard millet acid whey; BAW25: 25% w/v 773 Barnyard millet acid whey; LAW10: 10% w/v Little millet acid whey; LAW25: 25 5 w/v Little 774 millet acid whey 775 LAW25 62.52 ±0.03 137.48 ± 1.95 Values are the mean ± standard deviation of triplicate determinations. 758 Legend: AW, pure acid whey; BAW10, 10% w/v Barnyard millet acid whey; BAW25: 25% w/v 759 Barnyard millet acid whey; LAW10: 10% w/v Little millet acid whey; LAW25: 25 5 w/v Little 760 millet acid whey 761 762 763 764 765 766 767 768 769 770 Figure 1. Whiteness index of AW, BAW10, BAW20, LAW10, and LAW25 spray-dried 771 powders 772 Legend: AW, pure acid whey; BAW10, 10% w/v Barnyard millet acid whey; BAW25: 25% w/v 773 Barnyard millet acid whey; LAW10: 10% w/v Little millet acid whey; LAW25: 25 5 w/v Little 774 millet acid whey 775 137.48 ± 1.95 LAW25 Legend: AW, pure acid whey; BAW10, 10% w/v Barnyard millet acid whey; BAW25: 25% w/v 759 Barnyard millet acid whey; LAW10: 10% w/v Little millet acid whey; LAW25: 25 5 w/v Little 760 millet acid whey 761 768 769 770 Figure 1. Whiteness index of AW, BAW10, BAW20, LAW10, and LAW25 spray-dried 771 powders 772 Legend: AW, pure acid whey; BAW10, 10% w/v Barnyard millet acid whey; BAW25: 25% w/v 773 Barnyard millet acid whey; LAW10: 10% w/v Little millet acid whey; LAW25: 25 5 w/v Little 774 millet acid whey 775 Figure 1. Whiteness index of AW, BAW10, BAW20, LAW10, and LAW25 spray-dried 771 powders 772 Legend: AW, pure acid whey; BAW10, 10% w/v Barnyard millet acid whey; BAW25: 25% w/v 773 Barnyard millet acid whey; LAW10: 10% w/v Little millet acid whey; LAW25: 25 5 w/v Little 774 millet acid whey 775 28 776 77 Figure 2. Hygroscopicity of the spray-dried powders over seven days at a relative humidity 78 of 75%. 79 777 Figure 2. Hygroscopicity of the spray-dried powders over seven days at a relative humidity 778 of 75%. 779 780 Figure 3. Particle size distribution and zeta potential of the different mass ratios of acid whey 781 millet (AWM) formulations (a) the distribution of spray-dried powders by intensity (b) the 782 zeta potential of spray-dried powders. 783 Legend: AW, pure acid whey; BAW10, 10% w/v Barnyard millet acid whey; BAW25: 25% w/v 784 Barnyard millet acid whey; LAW10: 10% w/v Little millet acid whey; LAW25: 25 5 w/v Little 785 Figure 2. Hygroscopicity of the spray-dried powders over seven days at a relative humidity of 75%. 780 Figure 3. Particle size distribution and zeta potential of the different mass ratios of acid whey 781 millet (AWM) formulations (a) the distribution of spray-dried powders by intensity (b) the 782 zeta potential of spray-dried powders. 783 Legend: AW, pure acid whey; BAW10, 10% w/v Barnyard millet acid whey; BAW25: 25% w/v 784 Barnyard millet acid whey; LAW10: 10% w/v Little millet acid whey; LAW25: 25 5 w/v Little 785 millet acid whey 786 780 Figure 3. Particle size distribution and zeta potential of the different mass ratios of acid whey 781 millet (AWM) formulations (a) the distribution of spray-dried powders by intensity (b) the 782 Figure 3. LAW25 Particle size distribution and zeta potential of the different mass ratios of acid whey 781 millet (AWM) formulations (a) the distribution of spray-dried powders by intensity (b) the 782 zeta potential of spray-dried powders. 783 millet (AWM) formulations (a) the distribution of spray-dried powders by intensity (b) the 782 zeta potential of spray-dried powders. 783 Legend: AW, pure acid whey; BAW10, 10% w/v Barnyard millet acid whey; BAW25: 25% w/v 784 Barnyard millet acid whey; LAW10: 10% w/v Little millet acid whey; LAW25: 25 5 w/v Little 785 millet acid whey 786 29 787 788 789 790 791 792 793 794 Figure 4. (a) Differential scanning calorimeter (DSC) curves showing melting peaks (Tg) of 795 AWM and AW spray-dried powders. (b) product yield and Tg of AW and AWM powders. 796 Legend: AW, pure acid whey; BAW10, 10% w/v Barnyard millet acid whey; BAW25: 25% w/v 797 Barnyard millet acid whey; LAW10: 10% w/v Little millet acid whey; LAW25: 25 5 w/v Little 798 millet acid whey 799 800 794 Figure 4. (a) Differential scanning calorimeter (DSC) curves showing melting peaks (Tg) of 795 AWM and AW spray-dried powders. (b) product yield and Tg of AW and AWM powders. 796 Legend: AW, pure acid whey; BAW10, 10% w/v Barnyard millet acid whey; BAW25: 25% w/v 797 Barnyard millet acid whey; LAW10: 10% w/v Little millet acid whey; LAW25: 25 5 w/v Little 798 millet acid whey 799 800 Figure 4. (a) Differential scanning calorimeter (DSC) curves showin 795 Differential scanning calorimeter (DSC) curves showing melting peaks (Tg) of AWM and AW spray-dried powders. (b) product yield and Tg of AW and AWM powders. 796 Legend: AW, pure acid whey; BAW10, 10% w/v Barnyard millet acid whey; BAW25: 25% w/v 797 Barnyard millet acid whey; LAW10: 10% w/v Little millet acid whey; LAW25: 25 5 w/v Little 798 millet acid whey 799 800 Legend: AW, pure acid whey; BAW10, 10% w/v Barnyard millet acid whey; BAW25: 25% w/v 797 Barnyard millet acid whey; LAW10: 10% w/v Little millet acid whey; LAW25: 25 5 w/v Little 798 millet acid whey 799 30 812 1000x 4000x 10000x BAW10 BAW25 LAW10 LAW25 10000x 4000x 1000x BAW25 LAW10 LAW25 31 AW 813 Figure 5. Scanning electron micrographs of spray-dried AW and different AWM powders 814 at 4000x magnification. 815 AW 813 Figure 5. LAW25 Scanning electron micrographs of spray-dried AW and different AWM powders 814 at 4000x magnification. 815 Legend: AW, pure acid whey; BAW10, 10% w/v Barnyard millet acid whey; BAW25: 25% w/v 816 Barnyard millet acid whey; LAW10: 10% w/v Little millet acid whey; LAW25: 25 5 w/v Little 817 millet acid whey 818 819 AW AW 813 Figure 5. Scanning electron micrographs of spray-dried AW and different AWM powders 814 at 4000x magnification. 815 Legend: AW, pure acid whey; BAW10, 10% w/v Barnyard millet acid whey; BAW25: 25% w/v 816 Barnyard millet acid whey; LAW10: 10% w/v Little millet acid whey; LAW25: 25 5 w/v Little 817 millet acid whey 818 819 820 821 822 823 824 Figure 5. Scanning electron micrographs of spray-dried AW and different AWM powders 814 Figure 5. Scanning electron micrographs of spray-dried AW and different AWM powders 814 at 4000x magnification 815 Legend: AW, pure acid whey; BAW10, 10% w/v Barnyard millet acid whey; BAW25: 25% w/v 816 Barnyard millet acid whey; LAW10: 10% w/v Little millet acid whey; LAW25: 25 5 w/v Little 817 millet acid whey 818 32 Figure 6. Antioxidant activity of spray-dried powders (a) Total phenol content of spray-dried powders (b) Standard calibration curve of gallic acid (c) TBARS in spray-dried powders (d) Standard calibration curve of MDA (e) DPPH scavenging activity of spray-dried powders. L d AW id h BAW10 10% / B d ill t id h BAW25 25% / 25 825 Figure 6. Antioxidant activity of spray-dried powders (a) Total phenol content of spray-dried 827 powders (b) Standard calibration curve of gallic acid (c) TBARS in spray-dried powders (d) 828 Standard calibration curve of MDA (e) DPPH scavenging activity of spray-dried powders. 829 Legend: AW, pure acid whey; BAW10, 10% w/v Barnyard millet acid whey; BAW25: 25% w/v 830 Barnyard millet acid whey; LAW10: 10% w/v Little millet acid whey; LAW25: 25 5 w/v Little 831 millet acid whey 832 833 33 GraphicalAbstract.png Supplementary Files This is a list of supplementary ¦les associated with this preprint. Click to download. GraphicalAbstract.png GraphicalAbstract.png
https://openalex.org/W4310128755	https://munin.uit.no/bitstream/10037/27685/2/article.pdf	English	null	Hot Vents Beneath an Icy Ocean: The Aurora Vent Field, Gakkel Ridge, Revealed	Oceanography	2,023	cc-by	11,649	A SHORT HISTORY OF GAKKEL RIDGE EXPLORATIONit Forty-five years after the discovery of hydrothermal vents (Corliss et al., 1979), research into these unique hab- itats and their rich submarine ecosys- tems has brought about revolutionary findings in biology, chemistry, and geo- physics. Understanding how these dis- tinctive ecosystems are supported by sunlight-independent microbial primary productivity based on chemosynthesis changed the way we understand life on Earth (Van Dover et al., 2018). They have inspired our understanding of the ori- gin of life on Earth (Martin et al., 2008) and are now influencing the choice of exploration targets aimed at the discov- ery of extraterrestrial life in our solar system (Hand and German, 2018; Hand et al., 2020). The exotic faunal commu- nities at active hydrothermal vents are also of high interest given their physio- logical adaptations and the high degree of endemicity, and for their potential in providing marine genetic resources of use in biomedicine, cosmetics, and bio- fuels, among others (Van Dover et al., 2018). In addition, interest in the poten- tial for mineral resources in hydrother- mal vent deposits has greatly increased in the last two decades, and exploration licenses for such resources have been granted for national and international waters (Jones et al., 2020). Si th di f d h d Current data on vent communities globally has identified 11 biogeographic provinces, but their delineation is still being debated (Rogers et al., 2012). Until now, the vent faunal communities of the ice-covered Gakkel Ridge in the Central Arctic Ocean remained unexplored because of their remote and climatologi- cally challenging location. This study puts the Aurora Vent Field of the Gakkel Ridge on the global map of chemosynthetic- based ecosystems, providing an initial overview of the vent field and the eco- system it supports.h Building on the results of the AMORE 2001 expedition (Edmonds et al., 2003), in 2014, R/V Polarstern expedition PS86 AURORA aimed to study geophysical, geological, geochemical, and biological processes at hydrothermal vents on the Gakkel Ridge, with a focus on the southern segment (Boetius, 2015). In this region, the spreading rate is 14.5–13.5 mm yr–1, (slightly faster than the average rate for the overall ridge), and the ridge axis floor The Gakkel Ridge (Figure 1a) extends 1,800 km from the northern end of the Lena Trough off Northeast Greenland (81°N) to near the Siberian shelf at 87°N. REGULAR ISSUE FEATURE HOT VENTS BENEATH AN ICY OCEAN THE AURORA VENT FIELD, GAKKEL RIDGE, REVEALED By Eva Ramirez-Llodra, Claudio Argentino, Maria Baker, Antje Boetius, Carolina Costa, Håkon Dahle, Emily M. Denny, Pierre-Antoine Dessandier, Mari H. Eilertsen, Benedicte Ferre, Christopher R. German, Kevin Hand, Ana Hilário, Lawrence Hislop, John W. Jamieson, Dimitri Kalnitchenko, Achim Mall, Giuliana Panieri, Autun Purser, Sofia P. Ramalho, Eoghan P. Reeves, Leighton Rolley, Samuel I. Pereira, Pedro A. Ribeiro, Muhammed Fatih Sert, Ida H. Steen, Marie Stetzler, Runar Stokke, Lissette Victorero, Francesca Vulcano, Stig Vågenes, Kate Alyse Waghorn, and Stefan Buenz Oceanography \| Early Online Release Oceanography \| Vol.36, No.1 8 HOT VENTS BENEATH AN ICY OCEAN THE AURORA VENT FIELD, GAKKEL RIDGE, REVEALED By Eva Ramirez-Llodra, Claudio Argentino, Maria Baker, Antje Boetius, Carolina Costa, Håkon Dahle, Emily M. Denny, Pierre-Antoine Dessandier, Mari H. Eilertsen, Benedicte Ferre, Christopher R. German, Kevin Hand, Ana Hilário, Lawrence Hislop, John W. Jamieson, Dimitri Kalnitchenko, Achim Mall, Giuliana Panieri, Autun Purser, Sofia P. Ramalho, Eoghan P. Reeves, Leighton Rolley, Samuel I. Pereira, Pedro A. Ribeiro, Muhammed Fatih Sert, Ida H. Steen, Marie Stetzler, Runar Stokke, Lissette Victorero, Francesca Vulcano, Stig Vågenes, Kate Alyse Waghorn, and Stefan Buenz R/V Kronprins Haakon is being positioned here on an ice floe to drift toward the Aurora Vent Field on Gakkel Ridge, Arctic Ocean. © REV Ocean Oceanography \| Early Online Release ABSTRACT. Evidence of hydrothermal venting on the ultra-slow spreading Gakkel Ridge in the Central Arctic Ocean has been available since 2001, with first visual evi- dence of black smokers on the Aurora Vent Field obtained in 2014. But it was not until 2021 that the first ever remotely operated vehicle (ROV) dives to hydrothermal vents under permanent ice cover in the Arctic were conducted, enabling the collection of vent fluids, rocks, microbes, and fauna. In this paper, we present the methods employed for deep-sea ROV operations under drifting ice. We also provide the first description of the Aurora Vent Field, which includes three actively venting black smokers and dif- fuse flow on the Aurora mound at ~3,888 m depth on the southern part of the Gakkel Ridge (82.5°N). The biological communities are dominated by a new species of coccu- linid limpet, two small gastropods, and a melitid amphipod. The ongoing analyses of Aurora Vent Field samples will contribute to positioning the Gakkel Ridge hydrother- mal vents in the global biogeographic puzzle of hydrothermal vents. rate (E.T. Baker et al., 1996). This hypoth- esis was revisited after exploration of the Southwest Indian Ridge showed that even ultra-slow spreading ridges could host abundant submarine venting (German et al., 1998). Technological and method- ological challenges of working at great depth in regions of permanent sea ice cover have constrained the exploration of the Gakkel Ridge. In 2001, Edmonds et al. REGULAR ISSUE FEATURE HOT VENTS BENEATH AN ICY OCEAN THE AURORA VENT FIELD, GAKKEL RIDGE, REVEALED By Eva Ramirez-Llodra, Claudio Argentino, Maria Baker, Antje Boetius, Carolina Costa, Håkon Dahle, Emily M. Denny, Pierre-Antoine Dessandier, Mari H. Eilertsen, Benedicte Ferre, Christopher R. German, Kevin Hand, Ana Hilário, Lawrence Hislop, John W. Jamieson, Dimitri Kalnitchenko, Achim Mall, Giuliana Panieri, Autun Purser, Sofia P. Ramalho, Eoghan P. Reeves, Leighton Rolley, Samuel I. Pereira, Pedro A. Ribeiro, Muhammed Fatih Sert, Ida H. Steen, Marie Stetzler, Runar Stokke, Lissette Victorero, Francesca Vulcano, Stig Vågenes, Kate Alyse Waghorn, and Stefan Buenz (2003) obtained first evidence of hydrothermal venting on nine to twelve discrete locations along the Gakkel Ridge during the InterRidge two-icebreaker (R/V Polarstern and USCGC Healy) Arctic Mid-Ocean Ridge Expedition (AMORE; Figure 1a). Continued explo- ration during the Arctic Gakkel Vents (AGAVE) expedition in 2007 provided evidence of explosive volcanism at 85°N and demonstrated that large-scale pyro- clastic activity is possible along even the deepest portions of the global mid-ocean ridge volcanic system (Sohn et al., 2008). Seismic studies suggest substantial mag- matic activity, serpentinization, and fluid flow at this slowest of all Earth’s ridge sys- tems (Michael et al., 2003; Schlindwein and Schmid, 2016). Between 2002 and 2010, the ChEss program aimed to improve understanding of the global bio- geography of chemosynthetic-based eco- systems (M.C. Baker et al., 2010). Based on the increasing evidence of hydro- thermal venting along the Gakkel Ridge, the ChEss program identified a num- ber of poorly investigated regions where research efforts should focus. The Gakkel Ridge was recognized as one of the miss- ing pieces of the global biogeographic puzzle (Ramirez-Llodra et al., 2007). Building on the results of the AMORE 2001 expedition (Edmonds et al., 2003), i 2014 R/V P l t diti PS86 investigated (Beaulieu et al., 2015). To date, exploration has yielded an inventory of 722 confirmed high-temperature vent sites, with a further 720 high-temperature vents inferred from water column data, as reported in the InterRidge Vents Database in September 2022 (Beaulieu and Szafranski, 2020). There may be hun- dreds of additional active hydrothermal systems and their associated faunal com- munities yet to be discovered worldwide along the unexplored branches and sec- tions of the global mid-ocean ridge sys- tem, particularly along the least explored slow and ultra-slow spreading ridges (Beaulieu et al., 2015). A SHORT HISTORY OF GAKKEL RIDGE EXPLORATIONit It was initially predicted to host an extremely low number of active sites based on the assumption that hydrother- mal flux scaled directly with spreading Since the discovery of deep-sea hydro- thermal vents in 1977, just over 30% of the global mid-ocean ridge system has been Oceanography \| https://doi.org/10.5670/oceanog.2023.103 Recorder) instrument attached to the dredge revealed evidence for a turbid- ity anomaly consistent with a nearby source of active black smoker venting at a depth of 2,800–3,400 m (Edmonds et al., 2003; Michael et al., 2003). During the PS86 AURORA expedition, CTD profil- ing, coupled with water column chemis- try, revealed further evidence for ongoing hydrothermal activity on the Aurora mound (Boetius, 2015; German et al., 2022a). Seabed surveys with the Ocean Floor Observation System (OFOS) deep- tow camera across the summit from north to south revealed deep rifts through the thick sedimented seafloor across the base of the volcanic mound. This imag- ing, paired with CTD data, led to the first imaging of an active black smoker on Gakkel Ridge at 82°53.83'N, 6°15.32'W, at ~3,900 m depth, on what was named the Aurora Vent Field (AVF; Boetius, 2015; German et al., 2022a). The OFOS surveys showed that the Aurora mound has steep vertical basalt walls intermixed with lower angle, sediment-draped steps. The top of the mound is flat and sediment covered, and the observed fauna con- sisted of high abundances of filter feeders, mostly glass sponges and anemones, and at least two species of shrimp. Ophiuroids, swimming polychaetes, and crustaceans (potentially isopods) were also observed. At the active vent site, bacterial mats and small gastropods and amphipods were observed (Boetius, 2015). The physico- chemical and microbiological character- ization of the huge buoyant vent plume hovering above the AVF showed evidence for venting fluids enriched in meth- ane, and possibly hydrogen, fueling high microbial activity in the plume (German et al., 2022a; Massimiliano Molari, Max Planck Institute for Marine Microbiology, pers. comm., 2022). Due to the lack of a deep-diving remotely operated vehicle (ROV), however, no physical samples of fluids, rocks, microbes, or animals could be collected from the vent field. at 4,200 m depth is bounded by steep rift valley walls and punctuated by a series of axial volcanic ridges and smaller vol- canic mounds (Michael et al., 2003). Edmonds et al. A SHORT HISTORY OF GAKKEL RIDGE EXPLORATIONit (2003) inferred the pres- ence of an active hydrothermal vent site from chemical data and assigned to a small (1.5–2 km in diameter) volcanic mound rising approximately 400 m from the seafloor, at depths between 4,300 m and 3,850 m (Figure 1b,c). A dredge from south to north across this mound recov- ered components of a sulfide chimney in addition to abundant pillow basalts. In parallel, in situ sensor data from a MAPR (Miniature Autonomous Plume a c b Gakkel Ridge Greenland Svalbard FIGURE 1. (a) Map of the Gakkel Ridge in the Central Arctic Ocean with known hydrothermal plume signals indicated by yellow stars (from Edmond et al., 2003) and the Aurora Vent Field marked with a red star. (b) A red triangle locates the Aurora Vent Field within the Aurora mound based on pre- vious bathymetry from the AMORE and AURORA/AWI expeditions. (c) The Aurora Vent Field (red triangle) is shown against multibeam bathymetry of the Aurora mound based on bathymetry from the HACON19 and HACON21 expeditions. a b Gakkel Ridge Greenland Svalbard a b Gakkel Ridge Greenland Svalbard c FIGURE 1. (a) Map of the Gakkel Ridge in the Central Arctic Ocean with known hydrothermal plume signals indicated by yellow stars (from Edmond et al., 2003) and the Aurora Vent Field marked with a red star. (b) A red triangle locates the Aurora Vent Field within the Aurora mound based on pre- vious bathymetry from the AMORE and AURORA/AWI expeditions. (c) The Aurora Vent Field (red triangle) is shown against multibeam bathymetry of the Aurora mound based on bathymetry from the HACON19 and HACON21 expeditions. c c In 2019, the Hot Vents in an Ice- Covered Ocean (HACON19) cruise on R/V Kronprins Haakon returned to the Aurora mound, with the aim of conduct- ing a multidisciplinary survey of the sea- floor ecosystems centered around the coordinates of the black smoker iden- tified in 2014 by the PS86 Aurora team (Boetius, 2015). This cruise obtained new visual data of the AVF with the towed Ocean Floor Observation and Bathymetry System (OFOBS; Purser et al., 2019; German et al., 2022a), con- firming the presence of at least three black smokers colonized by sparse fauna com- posed of mostly gastropods and amphi- pods (Bünz et al., 2019). In addition, a wealth of samples on the sedimented sur- face of the Aurora mound were collected, FIGURE 1. Oceanography \| Early Online Release A SHORT HISTORY OF GAKKEL RIDGE EXPLORATIONit (a) Map of the Gakkel Ridge in the Central Arctic Ocean with known hydrothermal plume signals indicated by yellow stars (from Edmond et al., 2003) and the Aurora Vent Field marked with a red star. (b) A red triangle locates the Aurora Vent Field within the Aurora mound based on pre- vious bathymetry from the AMORE and AURORA/AWI expeditions. (c) The Aurora Vent Field (red triangle) is shown against multibeam bathymetry of the Aurora mound based on bathymetry from the HACON19 and HACON21 expeditions. Oceanography \| Early Online Release Oceanography \| Early Online Release using coring equipment as well as the Woods Hole Oceanographic Institution hybrid ROV/AUV Nereus Under Ice (NUI), which collected sponges for tax- onomic analyses. However, technological challenges and ice conditions prevented NUI from reaching the vent field and col- lecting samples (Bünz et al., 2019). used to successfully dive on and sample deep hydrothermal vents under drifting ice with an ROV, providing an operational baseline for future robotic explorations of the deep Arctic. The second part of the paper provides a preliminary description of the Aurora Vent Field and its biologi- cal communities. open-water leads and pressure ridges between ice floes greatly influences the ability, time, and effort needed to reach the study area. When the vessel has reached the study area, the first step in planning ROV or OFOBS dives is to determine and pre- dict ice-drift velocity and direction in order to position the vessel within the context of current and developing ice conditions. Once the ROV is launched, the ship will no longer be able to reposi- tion, so it must be located “upstream” of the dive target. The aim is for the ROV to reach the seafloor ahead of the vent field/ study site, while the vessel drifts with the ice floe toward the study site. To deter- mine vessel positioning, we used satel- lite images that provided a regional over- view of sea ice conditions (i.e., ice floe sizes, their distribution, and the presence of open water or thin ice leads). In addi- tion, 24 hr predictions of regional ice drift velocity and direction provided daily by Drift+Noise Polar Services GmbH/AWI aided in positioning the vessel (SIDFEx, 2022). A SHORT HISTORY OF GAKKEL RIDGE EXPLORATIONit These ice maps and ice-drift mod- els covered an area of 360 km2 around the Aurora mound and were useful for predicting when open-water leads would be positioned above the vent field (Figure 2b). With a range of six nautical miles, the vessel’s ice radar aided naviga- tion among ice floes. Between June and September 2021, the Joint Arctic Scientific Mid-Ocean Ridge Insight Expedition (JASMInE) car- ried out the first wide-angle reflection/ refraction seismic experiment along the Gakkel Ridge, from 75° to 102°, to map lithospheric structure (Ding et al., 2022). Preliminary results contribute to the understanding of oceanic crustal forma- tion and episodic magmatism in this end- member of global oceanic crustal accre- tion (Ding et al., 2022). ROV DEEP-SEA EXPLORATION UNDER ICE The ocean surface above the Aurora mound is permanently covered by sea ice (Figure 2a). The maximum sea ice extent in the Arctic occurs at the end of winter, generally in March, and the min- imum extent is normally observed in September, at the end of the summer season, so mid-August to mid-October is the most suitable time window for oceanographic operations on the Gakkel Ridge. Reaching a specific study site in an ice-covered ocean is the first challenge to overcome. Ice floe size, coupled with the speed and direction of ice drift, will deter- mine the time it takes for a specific subsea location to be accessible. The speed and direction of ice drift varies within and between days, driven by tidal forces and wind conditions, creating an ever-chang- ing icescape for which we have limited predictability on the order of 6–12 hours (Boetius, 2015). The presence of thin-ice/ Later in 2021, the HACON21 cruise wrote the latest chapter in the explora- tion of Gakkel Ridge, contributing to the UN Decade of Ocean Science for Sustainable Development through the Challenger 150 programme (Howell et al., 2020). A multidisciplinary team sailed aboard R/V Kronprins Haakon from Longyearbyen (Svalbard) on September 29 to complete the first ROV survey and sampling of hydrother- mal vents under permanent ice cover in the Arctic (Bünz et al., 2021). Below, we describe the methodological approach b b FIGURE 2. (a) R/V Kronprins Haakon navigates a thin-ice lead between ice floes above the Aurora Vent Filed. © REV Ocean/L. Hislop (b) This satel- lite image shows the regional ice situation and the ice drift predictions (red dots) modeled by Drift+Noise Polar Services GmbH/AWI. The darker areas are open water or thin ice. The gray/white coloring indicates thicker ice. The yellow star indicates the position of the Aurora Vent Field. (c) ROV Aurora Borealis is recovered through the moonpool of R/V Kronprins Haakon along with its 1,000 m tether management system (TMS). a c a FIGURE 2. (a) R/V Kronprins Haakon navigates a thin-ice lead between ice floes above the Aurora Vent Filed. © REV Ocean/L. Hislop (b) This satel- lite image shows the regional ice situation and the ice drift predictions (red dots) modeled by Drift+Noise Polar Services GmbH/AWI. The darker areas are open water or thin ice. The gray/white coloring indicates thicker ice. The yellow star indicates the position of the Aurora Vent Field. ROV DEEP-SEA EXPLORATION UNDER ICE (c) ROV Aurora Borealis is recovered through the moonpool of R/V Kronprins Haakon along with its 1,000 m tether management system (TMS). FIGURE 2. (a) R/V Kronprins Haakon navigates a thin-ice lead between ice floes above the Aurora Vent Filed. © REV Ocean/L. Hislop (b) This satel- lite image shows the regional ice situation and the ice drift predictions (red dots) modeled by Drift+Noise Polar Services GmbH/AWI. The darker areas are open water or thin ice. The gray/white coloring indicates thicker ice. The yellow star indicates the position of the Aurora Vent Field. (c) ROV Aurora Borealis is recovered through the moonpool of R/V Kronprins Haakon along with its 1,000 m tether management system (TMS). Oceanography \| https://doi.org/10.5670/oceanog.2023.103 Oceanography \| https://doi.org/10.5670/oceanog.2023.103 fly faster than the boat drifts with the ice. This allows the ROV to reach the target before the vessel drifts over it, providing additional time for the ROV to work on the seafloor on a small target such as a vent field (Figure 3). Operations through the moonpool protected the ROV and its TMS from ice contact and ensured that an open area was always available for the recovery of the ROV. The oper- ational methodology described below aims at maximizing seafloor time on a small target (in this case, the active area of the Aurora Vent Field, only about 75 m2) at great depth (4,000 m) under drifting ice. If the goal was to conduct a survey in a region without a precise geo- graphic target (e.g., a biological or sedi- mentological transect in a region, to sam- ple opportunistically), vessel positioning and ice drift velocity and direction would be less critical. ahead of the vessel toward the vent field. Drift speeds up to 0.3 knots provided suf- ficient bottom time (1–2.5 hr) for the ROV to work on the seafloor while sta- tionary at a target location. Operations between 0.3 knots and 0.6 knots were feasible but challenging, and seafloor time at the study area was very lim- ited (20–40 min). Above 0.6 knots, the risk of operations was considered too high and the ROV was not launched. If the distance between launch and study area was too large, the probability of the ice drift changing direction during the ROV’s descent, taking the vessel away from the target area, was higher, particu- larly with slower ice drift <0.1 knot. ROV DEEP-SEA EXPLORATION UNDER ICE If the launch position was too close to the vent field, there was not sufficient time for the ROV to reach the seafloor before the ves- sel drifted beyond the study site and the ROV had to be recovered. Of the 14 times ROV Aurora was launched over the AVF, the dives had to be aborted on six occa- sions due to a change in ice drift direction and/or speed that took the vessel off-track from the vent field. On one occasion, the drift direction changed from south to southeast, and then east-northeast within hours, which resulted in the ROV being launched and recovered four times before we were able to complete a successful dive. ROV Operations ROV Aurora is a Kystdesign SUPPORTER 32-type ROV (Figure 2c) owned and operated by REV Ocean. It has a 6,000 m depth range and is fully equipped with a state-of-the-art science skid. Importantly for under ice operations, Aurora is con- figured to operate as part of a two- bodied system that comprises a separate tether management system (TMS) called Borealis, with an additional 1,000 m of neutrally buoyant tether. For protec- tion against adverse interactions with sea ice, the ROV was deployed through the moonpool of the Norwegian ice- breaker R/V Kronprins Haakon. The Aurora Borealis system descended as a unit toward the seafloor, which per- mitted a rapid descent (0.8–1 m s–1) of the vehicle while the vessel was drift- ing with the ice, avoiding tension on the ROV’s neutrally buoyant optical tether (Figure 3). When the vehicle was at 50–100 m above seafloor, depending on topography, the winch was stopped and the ROV deployed from its TMS. Safe ROV operations to the Arctic deep sea- floor were only possible during this expe- dition because of the two-bodied config- uration of the ROV and the availability of a moonpool for deployment and recovery operations. The soft umbilical of the TMS enables the ROV, once out of its TMS, to For the purpose of studying the AVF, once a suitable position for the vessel was determined, the speed and accurate direc- tion of the ice floe were assessed by posi- tioning the vessel against the ice floe and drifting with it for 15 minutes. Timing for ROV deployment could then be based on the drift speed. ROV DEEP-SEA EXPLORATION UNDER ICE ROV Aurora was launched at a distance from the vent field that allowed the vehicle to reach the sea- floor upstream from the study area and fly Geological and Geochemical Settingsh FIGURE 3. Operational sequence for ROV dives on small deep seafloor targets under drifting ice. The Aurora Vent Field is located on the southwest part of the Aurora mound (Figure 1b,c). It consists of pillow basalts and active and inactive vents, and encom- passes an area of ~140 m × 100 m. There are steep-sided inactive hydrothermal edifices up to several meters high, indi- vidual actively venting chimneys, and abundant hydrothermal crust and chim- ney debris that, along with the pillows, are variably covered by as much as ~3 m of sediment. Current hydrothermal activity is restricted to near the center of the AVF (82°53'49''N, 6°15'21''W) and consists primarily of three isolated, vigorously FIGURE 3. Operational sequence for ROV dives on small deep seafloor targets under drifting ice. FIGURE 3. Operational sequence for ROV dives on small deep seafloor targets under drifting ice. Oceanography \| Early Online Release Oceanography \| Early Online Release of Enceladus, slightly south of a line between Enceladus and Hans Tore vents. This black smoker rises on a relatively flat seafloor and consists of black smoker fluid venting from five individual exit ori- fices at the base or sides of a partially col- lapsed chimney structure. Both Enceladus and Ganymede are characterized by thin walls, resulting in frequent collapse and regrowth, as evidenced by their relatively short heights and ramparts of abundant collapsed chimney debris. venting black smoker chimneys that occur in a ~10 m diameter cluster. These three active black smokers have been named Hans Tore, Enceladus, and Ganymede (Figure 4a,c,e). The Hans Tore vent was observed with OFOBS during HACON19 (Bünz et al., 2019). It was named in memory of Professor Hans Tore Rapp (Figure 4b), colleague and friend from the University of Bergen, who passed away in 2020. Prof. Rapp had been involved in the 2014 AURORA PS86 and HACON19 missions as fauna expert and was a key figure in the development of the HACON project. The two other smokers are named after “ocean world” moons in the solar system, some of which may host hydrothermal activity with the potential to harbor associated chemo- synthetic-based life. Saturn’s ice-covered moon Enceladus (Figure 4d) was chosen because its ocean is suspected to be many tens of kilometers deep and predicted to host hydrothermal activity (Cable et al., 2021). Geological and Geochemical Settingsh Ganymede, an ice-covered satellite of Jupiter (Figure 4f), is the largest moon in the solar system and may have the largest saline ocean—an ocean that may be implicated in altering its polar aurora (Saur et al., 2015), providing a fitting con- nection to this locale. The rocks collected are generally dark greenish-gray and composed primarily of chalcopyrite, pyrite, sphalerite, and barite. The sulfide minerals exhibit varying degrees of oxidation associated with pro- longed exposure to seawater (Figure 6a). Samples were also collected from several low-lying orange-brown oxidized iron- rich chimneys associated with lower- temperature diffuse flow, especially those surrounding the rim of the Hans Tore vent crater. The formation of these extremely friable iron-rich deposits has been associ- ated with microbial activity (Johannessen et al., 2020). Post-cruise analyses of the The chimneys and larger hydrothermal accumulations at the AVF were sampled with the ROV manipulator (Figure 5a). FIGURE 4. Aurora Vent Field black smokers and their names: (a) Hans Tore vent. (b) Professor Hans Tore Rapp during the HACON cruise in 2019. (c) Enceladus black smoker. (d) Enceladus moon. Credit: NASA, Cassini Mission (e) Ganymede black smoker. (f) Ganymede moon. Credit: NASA, Juno Mission a c e b d f b a a c d c d The Hans Tore vent (Figure 4a), found at 3,883 m depth, features a tall (at least 2 m), narrow chimney vigorously emit- ting black smoke fluid from its summit and a second, 1 m tall active chimney next to it. They sit in the middle of a 2 m diameter, 1 m deep circular crater, and black smoker fluid issues from the bot- tom of the crater through two small chim- neys. Diffuse, lower-temperature vent- ing occurs around the rim of the crater, and small oxidized iron-rich chimneys are evident at the rim. At the Enceladus vent (Figure 4c), ~10 m southwest of Hans Tore, a single 1.5 m tall chimney structure emits black vent fluid along its entire length and from its top. It is sur- rounded by a forest of inactive chimneys at 3,887 m depth.h e f f e FIGURE 4. Aurora Vent Field black smokers and their names: (a) Hans Tore vent. (b) Professor Hans Tore Rapp during the HACON cruise in 2019. (c) Enceladus black smoker. (d) Enceladus moon. Credit: NASA, Cassini Mission (e) Ganymede black smoker. (f) Ganymede moon. Geological and Geochemical Settingsh Credit: NASA, Juno Mission The Ganymede vent (Figure 4e) is located at 3,884 m depth, ~5 m northeast Oceanography \| https://doi.org/10.5670/oceanog.2023.103 Oceanography \| https://doi.org/10.5670/oceanog.2023.103 2012) that will enable first estimates of fluid mass, chemical, and heat fluxes from the vents. The ongoing work will be criti- cal for deciphering the types of chemical compositions venting at Aurora and how they support the ecosystem. The data will also provide information on the water- rock interactions between hydrothermal fluids and the underlying crustal substrate (e.g., Reeves et al., 2011, 2014), allowing direct comparison of the chemical energy “landscape” of AVF fluids (e.g., Dahle et al., 2015) with other vent sites along the global mid-ocean ridge system. paleo-plume events recorded in the sed- imentary record therefore enables us to provide spatial and temporal con- straints on vent activity and the contri- bution of seafloor vent metal exports to deep ocean budgets (Antonelli et al., 2017). Hydrothermal plumes also trans- port trace elements, nutrients, dissolved gases, and vent-derived biomass that can support heterotrophic deep-sea commu- nities (Cathalot et al., 2021; Levin et al., 2016). At the AVF, the dispersing plume was identified from water-mass turbid- ity measured during CTD casts. The buoyant plume was located in the bot- tom layer between 3,000 m and 4,000 m depth during the AURORA (Boetius, 2015; German et al., 2022), HACON19, and HACON21 cruises. Dissolved meth- ane concentrations in water samples col- lected from ROV-mounted Niskin bottles during HACON21 (Figure 5c) at ~10 m directly above one of the vents were measured using headspace gas chro- matography (e.g., Sert et al., 2020). In these samples, elevated dissolved meth- ane in the near vent buoyant plume (~102–103 nM) decreases dramatically to 30 nmol/L in the immediately over- lying non-buoyant plume (German et al., 2022a), 1–10 nmol/L in the more distal plume (Sert et al., 2022), and eventually to background concentrations (~<0.5 nM). This shows that strong methane gradi- ents that prevail in the vicinity of the ris- ing buoyant plume could support chemo- synthetic microbial activity in the water column (Anantharaman et al., 2013). In buoyant and non-buoyant plumes, meth- ane dilution, oxidation, and methanogen- esis may occur simultaneously (Ver Eecke et al., 2012), contributing to deep-sea bio- geochemical cycling. Geological and Geochemical Settingsh rock samples will be integrated with sed- iment geochemistry data sets and used to characterize the composition, age, and evolution of the vent field and to assess the relationship between the composition and age of the deposits and the animals and microbes that colonize them. paleo-plume events recorded in the sed- imentary record therefore enables us to provide spatial and temporal con- straints on vent activity and the contri- bution of seafloor vent metal exports to deep ocean budgets (Antonelli et al., 2017). Hydrothermal plumes also trans- port trace elements, nutrients, dissolved gases, and vent-derived biomass that can support heterotrophic deep-sea commu- nities (Cathalot et al., 2021; Levin et al., 2016). At the AVF, the dispersing plume was identified from water-mass turbid- ity measured during CTD casts. The buoyant plume was located in the bot- tom layer between 3,000 m and 4,000 m depth during the AURORA (Boetius, 2015; German et al., 2022), HACON19, and HACON21 cruises. Dissolved meth- ane concentrations in water samples col- lected from ROV-mounted Niskin bottles during HACON21 (Figure 5c) at ~10 m directly above one of the vents were measured using headspace gas chro- matography (e.g., Sert et al., 2020). In these samples, elevated dissolved meth- ane in the near vent buoyant plume (~102–103 nM) decreases dramatically to 30 nmol/L in the immediately over- lying non-buoyant plume (German et al., 2022a), 1–10 nmol/L in the more distal plume (Sert et al., 2022), and eventually to background concentrations (~<0.5 nM). This shows that strong methane gradi- ents that prevail in the vicinity of the ris- ing buoyant plume could support chemo- synthetic microbial activity in the water column (Anantharaman et al., 2013). In buoyant and non-buoyant plumes, meth- ane dilution, oxidation, and methanogen- esis may occur simultaneously (Ver Eecke et al., 2012), contributing to deep-sea bio- geochemical cycling. The Enceladus and Ganymede black smoker “end member” fluids and their stable maximum temperatures were each successfully sampled with duplicate iso- baric gas-tight (IGT) samplers (Seewald et al., 2002; Figure 5b) for comprehensive inorganic-organic geochemical and stable isotope characterization (Reeves et al., 2011, 2014); “paired” chimney material was also examined. Geological and Geochemical Settingsh Sufficient static video footage (several minutes) of at least one of the vents was also taken to allow for parti- cle image velocimetry (Mittelstaedt et al., Hydrothermal fluids discharged at the seafloor form neutrally buoyant mid-water plumes that disperse hori- zontally along isopycnals, transporting and redistributing metals across entire basins (Resing et al., 2015). Correlating FIGURE 5. Sampling on the Aurora Vent Field. (a) Rock collection with the manipulator on the flanks of the Hans Tore vent. (b) Isobaric gas-tight (IGT) sampling at Ganymede. (c) Plume sampling with Niskin bottles on the Hans Tore vent. (d) Blade core sampling on the side of the Enceladus black smoker. (e) Sampling bacterial mats with the biosyringe on the rim of the Hans Tore vent. (f) Suction sampling for amphipods on the flanks of the Hans Tore vent. a c e b d f a b b a d c c d c e f Biological Communities of the Aurora Vent Field Deciphering geo-biological connec- tions in hydrothermal systems requires detailed analyses of the composition of fluids along with analyses of micro- bial community structure and function- ing. Samples for microbiological analyses were obtained from active and inactive chimneys, low-temperature diffuse flow sites, and background sediments. The rocks collected for geological studies (Figure 5a) were firstly sub-sampled on board for their microbiota. The blade core was used to sample undisturbed sediment samples (Figure 5d), and bio- syringes were used to sample microbial mats (Figure 5e). Ongoing geochem- ical and metagenomic-based commu- nity analyses will shed further light on the AVF microbial community structure and functioning and the biogeochemical relationships between sediment substrate (grain size, mineralogy, and geochemical composition) and microbial and faunal communities. The results will reveal to what extent the unique setting of the AVF stimulates development of unique micro- bial assemblages and ecosystems. and oxidized compounds in seawater (e.g., sulfate, nitrate, oxygen) (Van Dover, 2000). The availability of different chem- ical energy sources between and within vent systems form chemical energy land- scapes that shape microbial communities and possibly those of microbial primary production grazers (Dahle et al, 2015). Deciphering geo-biological connec- tions in hydrothermal systems requires detailed analyses of the composition of fluids along with analyses of micro- bial community structure and function- ing. Samples for microbiological analyses were obtained from active and inactive chimneys, low-temperature diffuse flow sites, and background sediments. The rocks collected for geological studies (Figure 5a) were firstly sub-sampled on board for their microbiota. The blade core was used to sample undisturbed sediment samples (Figure 5d), and bio- syringes were used to sample microbial mats (Figure 5e). Ongoing geochem- ical and metagenomic-based commu- nity analyses will shed further light on the AVF microbial community structure and functioning and the biogeochemical relationships between sediment substrate (grain size, mineralogy, and geochemical composition) and microbial and faunal communities. The results will reveal to what extent the unique setting of the AVF stimulates development of unique micro- bial assemblages and ecosystems. Samples of macro- and megafauna were obtained from chimney debris collected FIGURE 6. Samples from the Aurora Vent Field. (a) Representative chimney sample collected from the Enceladus black smoker, composed primarily of a mixture of sulfide minerals (chalcopy- rite, pyrite, sphalerite) and barite. (b) Rock hosting Cocculinid limpets and gastropods (scale bar = 5 mm). Biological Communities of the Aurora Vent Field Biological communities in hydrother- mal systems are typically driven by che- mosynthetic microorganisms acquir- ing energy from the oxidation of reduced compounds in hydrothermal fluids (e.g., sulfide, methane, hydrogen) FIGURE 5. Sampling on the Aurora Vent Field. (a) Rock collection with the manipulator on the flanks of the Hans Tore vent. (b) Isobaric gas-tight (IGT) sampling at Ganymede. (c) Plume sampling with Niskin bottles on the Hans Tore vent. (d) Blade core sampling on the side of the Enceladus black smoker. (e) Sampling bacterial mats with the biosyringe on the rim of the Hans Tore vent. (f) Suction sampling for amphipods on the flanks of the Hans Tore vent. Oceanography \| Early Online Release Oceanography \| Early Online Release kilometers away from the AVF are rep- resented by opportunistic species depen- dent on bacterial and phytodetrital matter, including low-energy species (Vanreusel et al., 2000). Meiofaunal community densities and diversity in hydrothermal- derived sediments were depleted when compared to communities within refer- ence sediments off vent. Benthic fora- minifera in these hydrothermal-derived sediments, however, include organic- walled allogromids that seem to tolerate reduced and low amounts of organic car- bon microhabitats. These initial results suggest a critical role of the meiofauna in the vicinity of vents for biodiversity and potential bio-indicators for ecosystem health in vent fields. with the ROV manipulators (Figure 5a), from the sediment surface collected with the suction sampler (Figure 5f), and from sponge aggregations at the vent periph- ery (Figure 6f). Preliminary analyses show an increased presence of molluscs, crustaceans, and polychaetes in the AVF compared to the surrounding rock fields. Within the molluscs, a new species of cocculinid limpet, Cocculina aurora sp. nov. (Chen et al., 2022; Figure 6b) and two species of gastropods of the fami- lies Rissoidae and Skeneidae (Figure 6c) were the most abundant taxa found on the rocks. Samples collected with the suc- tion sampler targeted amphipods, tenta- tively identified as belonging to the family Melitidae (Figure 6d,e), the second most dominant group at the active vent field after the molluscs. and oxidized compounds in seawater (e.g., sulfate, nitrate, oxygen) (Van Dover, 2000). The availability of different chem- ical energy sources between and within vent systems form chemical energy land- scapes that shape microbial communities and possibly those of microbial primary production grazers (Dahle et al, 2015). Biological Communities of the Aurora Vent Field (c) Detail of the rock in B showing the high abundance of gastropods (scale bar = 5 mm). (d) Amphipods on the rim of the Hans Tore vent. (e) Close-up of an amphipod in the lab (scale bar = 5 mm). (f) Cladorhizid carnivorous sponges on the periphery of the Aurora Vent Field. a c e b d f b b a d f d Infauna from sediments recovered from the vent field with the blade core (Figure 5d) and from a reference site a few kilometers off vent with a ship-operated multicorer were sectioned at different intervals, sieved with 1,000-, 500-, 250-, 150-, and 32-micron mesh sieves and identified under a microscope in the lab- oratory. The infauna was composed, pre- dominantly, of meiofaunal taxa (32 µm to 1,000 µm), specifically free-living nema- todes and foraminifera. These groups are known to be a key link in the deep-sea food web, including in hydrothermal vent and cold seep environments (Zeppilli et al., 2018). Classical abyssal polar meio- fauna communities in sediments a few e FIGURE 6. Samples from the Aurora Vent Field. (a) Representative chimney sample collected from the Enceladus black smoker, composed primarily of a mixture of sulfide minerals (chalcopy- rite, pyrite, sphalerite) and barite. (b) Rock hosting Cocculinid limpets and gastropods (scale bar = 5 mm). (c) Detail of the rock in B showing the high abundance of gastropods (scale bar = 5 mm). (d) Amphipods on the rim of the Hans Tore vent. (e) Close-up of an amphipod in the lab (scale bar = 5 mm). (f) Cladorhizid carnivorous sponges on the periphery of the Aurora Vent Field. FIGURE 6. Samples from the Aurora Vent Field. (a) Representative chimney sample collected from the Enceladus black smoker, composed primarily of a mixture of sulfide minerals (chalcopy- rite, pyrite, sphalerite) and barite. (b) Rock hosting Cocculinid limpets and gastropods (scale bar = 5 mm). (c) Detail of the rock in B showing the high abundance of gastropods (scale bar = 5 mm). (d) Amphipods on the rim of the Hans Tore vent. (e) Close-up of an amphipod in the lab (scale bar = 5 mm). (f) Cladorhizid carnivorous sponges on the periphery of the Aurora Vent Field. The Ice-Covered Arctic as a Bridge to the Search for Life in Space? g to the Search for Life in Space? A particularly exciting aspect of the dis- covery of hydrothermal vents beneath an ice-covered ocean is its impact on humanity’s search for life beyond Earth. The collection and analysis of 18 full sea- ice cores ranging from 98 cm to 181 cm in length, as well as three gray ice cores rang- ing from 8 cm to 15 cm, served the dual purpose of advancing our understanding of sea-ice processes on Earth and provid- ing a useful analog for investigating the surface chemistry of distant ice-covered ocean worlds elsewhere in our solar sys- tem. On Earth, the seasonally variable sea-ice cover modulates the exchange of gases between the ocean and the atmo- sphere. During the winter period of max- imum coverage, and of sea-ice growth, gases such as carbon dioxide and meth- ane cannot escape directly to the atmo- sphere, but rather are captured in the ice matrix. Similarly, gases from the atmo- sphere cannot reach the ocean, as mixing is inhibited by the ice layer. In addition, some buoyant materials, such as organic lipids and other biological materials, can become entrained in the ice matrix. The seasonal exchange efficiency of gases in the Arctic, especially carbon dioxide, is a poorly constrained variable in our efforts to model and understand climate change. We conducted analyses of trapped car- bon dioxide and methane in many of the cores while onboard R/V Kronprins Haakon and have generated depth pro- files of gas concentration that show the detailed transition from atmospheric to oceanic concentrations of carbon dioxide and methane. In addition, we measured the 13C concentration in methane, which we can use to differentiate between bio- genic and abiogenic contributions to the A particularly exciting aspect of the dis- covery of hydrothermal vents beneath an ice-covered ocean is its impact on humanity’s search for life beyond Earth. h Although a detailed biogeographic analysis that includes the fauna from the Aurora Vent Field is underway, the cur- rently available observations provide insights into the position of the Gakkel Ridge vent communities in global vent biogeography. Six biogeographic regions of vent fauna defined by Bachraty et al. (2009) were increased to 11 regions of biogeography when the data were reana- lyzed, incorporating the vent fields of the Southern Ocean (Rogers et al., 2012). However, none of these studies included data from Arctic latitudes. Biological Communities of the Aurora Vent Field Oceanography \| https://doi.org/10.5670/oceanog.2023.103 Oceanography \| https://doi.org/10.5670/oceanog.2023.103 Oceanography \| https://doi.org/10.5670/oceanog.2023.103 pillow lavas, and basaltic ridges and out- crops in the proximity of the vent field were covered by dense aggregations of the hexactinellid sponges Caulophacus arcticus and Asconema megaatrialia. These species are common inhabitants of deeper areas in the Nordic Seas (Roberts et al., 2018), and the unusually high den- sities found at the Aurora mound com- pared to non-venting mounds on the Gakkel Ridge (Boetius, 2015) suggest that the higher productivity around the active vent field provides increased food supply that can support higher sponge density. Aggregations of sponges and dead stalks of C. arcticus may have a crucial facilitat- ing role for other fauna, since bythocarid shrimps, anemones, crinoids, and several species of isopods and amphipods seem to be more abundant in areas covered by hexactinellid sponge habitats. depths under drifting ice, which heavily constrains our capacity to develop robust management measures and monitoring procedures for a pristine region prior to the expected increase in industrial activ- ity as the Central Arctic Ocean opens to human activities. At the Hans Tore vent, amphipods and small gastropods can be seen on the rim and outer slopes of the crater. At Enceladus and Ganymede, amphi- pods are evident on the chimney walls, whereas gastropods are mostly pres- ent on chimney debris at the base of the black smokers. Suction sampling at the Hans Tore and Ganymede vents also retrieved a few specimens of polychaetes and some empty polychaete tubes. The AVF vent fauna, dominated by small gas- tropods and amphipods, is similar, on a higher taxonomic level, to the faunal com- munities of the Loki’s Castle Vent Field at 73.5°N on the Mohns Ridge (Pedersen et al., 2010). Further work on the taxon- omy, connectivity, and biogeography of the fauna of the AVF is ongoing in order to clarify the links between the AVF fauna and that of other vent fields. In contrast to the Mohns Ridge vents, sedimented areas with diffuse venting were not observed, which explains the restricted spatial dis- tribution of vent-adapted fauna close to the black smokers. Basalt outcrops within the vent field host dense aggregations of carnivorous sponges belonging to the family Cladorhizidae (Figure 6f). Biological Communities of the Aurora Vent Field Even though these organisms are typical rep- resentatives of the background fauna, they can thrive at higher densities in the vicinity of active vents by possibly taking advantage of the food-enriched environ- ment driven by the hydrothermal activity, while tolerating increased abiotic stress (Levin et al., 2016; Georgieva et al., 2020). The Ice-Covered Arctic as a Bridge to the Search for Life in Space? Connection of this work to distant ice-covered moons, such as Jupiter’s Europa, is based on the knowledge that Europa’s ice shell is essentially ~10 km of sea ice (Figure 7). Ocean salts, carbon dioxide, and other materials have been observed spectroscopically on the sur- face of Europa (Hand et al., 2007; Hand and Carlson, 2015; Trumbo et al., 2019). The best analogous systems on Earth to Europa’s ice are the sea ice of the Arctic and the Antarctic. While the scale of ice thickness is much different, the pro- cesses by which gases and organic mate- rials, including microbes, are incorpo- rated into sea ice on Earth can inform the development of missions and instru- ments designed to search for such mate- rials either remotely or with a landed spacecraft on Europa. To this end, our team processed seven ice cores for micro- biology primarily to determine whether any thermophiles that may have been transported via hydrothermal plumes are captured in the ice cores. This is an interesting microbial ecology question in its own right, but it also carries interest- ing implications for the search for bio- signatures on Europa, as the seafloor there could be hydrothermally active, and plumes could deliver material to the ice-water interface. FIGURE 7. While the physical scale and thickness of the ice differs greatly between Earth’s sea ice and that of Europa’s ice shell, both icy environments serve as a window into the oceans hidden below. On Earth, we can study the ice, ocean, and seafloor to reveal any processes that provide connections across these interfaces. On Europa, we are, at least for the near future, limited to using the surface chemistry of the ice as our primary means for understanding the ocean and seafloor chemistry and geology. Image courtesy NASA/JPL and the 2017 Europa Lander Science Definition Team Report (Hand et al., 2017) of photosynthesis on Earth, but some of the microbes detected in the Aurora vent plume relate to microaerophilic or even anaerobic types (Massimiliano Molari, Max Planck Institute for Marine Microbiology, pers. comm., 2022). The intellectual challenge in pursuing this line of research becomes more compel- ling when considering that a significant proportion of the exoplanets that have now been discovered orbiting other stars may also be ocean worlds (Quick et al., 2020). A next critical step in that explora- tion will begin soon, with the 2023 launch of the European Space Agency’s JUpiter ICy moons Explorer (JUICE) expedi- tion to investigate Europa and Ganymede and the 2024 launch of NASA’s Europa Clipper mission to study that body’s hab- itability. In the longer term, planning has already begun for future landed missions to search for evidence of life on both Europa (Hand et al., 2022) and Enceladus (MacKenzie et al., 2021). of photosynthesis on Earth, but some of the microbes detected in the Aurora vent plume relate to microaerophilic or even anaerobic types (Massimiliano Molari, Max Planck Institute for Marine Microbiology, pers. comm., 2022). The intellectual challenge in pursuing this line of research becomes more compel- ling when considering that a significant proportion of the exoplanets that have now been discovered orbiting other stars may also be ocean worlds (Quick et al., 2020). A next critical step in that explora- tion will begin soon, with the 2023 launch of the European Space Agency’s JUpiter ICy moons Explorer (JUICE) expedi- tion to investigate Europa and Ganymede and the 2024 launch of NASA’s Europa Clipper mission to study that body’s hab- itability. In the longer term, planning has already begun for future landed missions to search for evidence of life on both Europa (Hand et al., 2022) and Enceladus (MacKenzie et al., 2021). The Ice-Covered Arctic as a Bridge to the Search for Life in Space? Our obser- vations of the Aurora Vent Field sug- gest faunal similarities with the Loki’s Castle Vent Field at 73°N (Pedersen et al., 2010) in terms of the presence of amphi- pods and gastropods as main groups, but also show differences such as the lack of siboglinid tubeworms and the presence of the new species of limpet Cocculina aurora sp. nov. (Chen et al., 2022). The evolution and genetic connectivity of fauna from the deep Central Arctic Ocean is a major unknown, and the data from the HACON project will contrib- ute to addressing it. The limited knowl- edge of Central Arctic Ocean deep-sea ecosystems is a direct consequence of the operational challenges of working at great A comprehensive ROV video survey was also conducted to provide visual data for mapping the different habitats across the AVF and to help decipher differences in communities detected by the OFOS and OFOBS photo transects across the Aurora mound and further sites along the Gakkel Ridge (Boetius, 2015; Bünz et al., 2019). The video analyses will provide information on the distribution, abun- dance, and density of the different species across the three AVF black smokers and their geochemical gradients (Schoening et al., 2012). As previously reported (Boetius, 2015; Bünz et al., 2019), rocks, Oceanography \| Early Online Release FIGURE 7. While the physical scale and thickness of the ice differs greatly between Earth’s sea ice and that of Europa’s ice shell, both icy environments serve as a window into the oceans hidden below. On Earth, we can study the ice, ocean, and seafloor to reveal any processes that provide connections across these interfaces. On Europa, we are, at least for the near future, limited to using the surface chemistry of the ice as our primary means for understanding the ocean and seafloor chemistry and geology. Image courtesy NASA/JPL and the 2017 Europa Lander Science Definition Team Report (Hand et al., 2017) methane signature. REFERENCES 2021711), Longyearbyen – Longyearbyen, 28.09.2021 – 21.10.2021: Hot Vents in an ice- covered Ocean, HACON21 Expedition. Cruise Report, 81 pp., https://doi.org/10.7557/cage.6715. Hand, K.P., and R.W. Carlson. 2015. Europa’s surface color suggests an ocean rich with sodium chlo- ride. Geophysical Research Letters 42:3,174–3,178, https://doi.org/10.1002/2015GL063559. Hand, K.P., A.E. Murray, J.B. Garvin, W.B. Brinckerhoff, B.C. Christner, K.S. Edgett, B.L. Ehlmann, C.R. German, A.G. Hayes, T.M. Hoehler, and others. 2017. Report of the Europa Lander Science Definition Team. Posted February 2017, https://europa.nasa.gov/resources/58/ europa-lander-study-2016-report/. Cable, M.L., C. Porco, C.R. Glein, C.R. German, S.M. MacKenzie, M. Neveu, T.M. Hoehler, A.R. Hendrix, J. Eigenbrode, F. Postberg, and others. 2021. The science case for a return to Enceladus. Planetary Science Journal 2:132, https://doi.org/10.3847/PSJ/abfb7a. Purser A., Y. Marcon, S. Dreutter, U. Hoge, B. Sablotny, L. Hehemann, J. Lemburg, B. Dorschel, H. Biebow, and A. Boetius. 2019. Ocean Floor Observation and Bathymetry System (OFOBS): A new towed camera/sonar system for deep-sea habitat surveys. IEEE Journal of Oceanic Engineering 44:87–99, https://doi.org/10.1109/JOE.2018.2794095. b Cathalot, C., E.G. Roussel, A. Perhirin, V. Creff, J.-P. Donval, V. Guyader, G. Roullet, J. Gula, C. Tamburini, M. Garel, and others. 2021. Hydrothermal plumes as hotspots for deep-ocean heterotrophic microbial biomass production. Nature Communications 12:6861, https://doi.org/ 10.1038/s41467-021-26877-6. Hand, K.P., and C.R. German. 2018. Exploring ocean worlds on Earth and beyond. Nature Geoscience 10:2–4, https://doi.org/10.1038/ s41561-017-0045-9. Quick, L.C., A. Roberge, A. Barr Mlinar, and M.M. Hedman. 2020. Forecasting rates of vol- canic activity on terrestrial exoplanets and impli- cations for cryovolcanic activity on extrasolar ocean worlds. Publication of the Astronomical Society of the Pacific 132:084402, https://doi.org/ 10.1088/1538-3873/ab9504. Hand, K.P., C. Sotin, A. Hayes, and A. Coustenis. 2020. On the habitability and future exploration of ocean worlds. Space Science Reviews 216:95, https://doi org/101007/s11214 020 00713 7 Chen, C., A. Hilário, C.F. Rodrigues, and E. Ramirez- Llodra. 2022. Integrative taxonomy of a new cocculinid limpet dominating the Aurora Vent Field in the Central Arctic Ocean. Royal Society Open Science 9:220885, https://doi.org/10.1098/ rsos.220885. Ramirez-Llodra, E., T.M. Shank, and C.R. German. 2007. Biodiversity and biogeography of hydro- thermal vent species: Thirty years of discovery and investigation. Oceanography 20(1):30–41, https://doi.org/10.5670/oceanog.2007.78. Hand, K.P., C.B. Phillips, A. Murray, J.B. Garvin, E.H. Maize, R.G. Gibbs, G. Reeves, A.M. San Martin, G.H. Tan-Wang, J. Krajewski, and others. 2022. Science goals and mission architecture of the Europa Lander Mission Concept. The Planetary Science Journal 3:22, https://doi.org/10.3847/PSJ/ ac4493. Reeves, E.P., J.S. Seewald, P. Saccocia, W. Bach, P.R. Craddock, W.C. REFERENCES Deep Sea Research Part II 121:202–212, https://doi.org/10.1016/j.dsr2. 2015.05.001. Martin, W., J. Baross, D. Kelley, and M.J. Russel. 2008. Hydrothermal vents and the origin of life. Nature Reviews Microbiology 6:805–814, https://doi.org/ 10.1038/nrmicro1991. German, C.R., E.P. Reeves, A. Türke, A. Diehl, German, C.R., E.P. Reeves, A. Türke, A. Diehl, E. Albers, W. Bach, A. Purser, S.P. Ramalho, S. Suman, C. Mertens, and others. 2022a. Volcanically hosted venting with indications of ultramafic influence at Aurora hydrothermal field on Gakkel Ridge. Nature Communications 13:6517, https://doi.org/10.1038/s41467-022-34014-0. Beaulieu, S.E., and K. Szafranski. 2020. InterRidge Global Database of Active Submarine Hydrothermal Vent Fields, Version 3.4. Available from https://vents-data.interridge.org/, accessed 02/18/2022. E. Albers, W. Bach, A. Purser, S.P. Ramalho, Michael, P.J., C.H. Langmuir, H.J.B. Dick, J.E. Snow, S.L. Goldstein, D.W. Graham, K. Lehnert, G. Kurras, W. Jokat, R. Mühe, and H.N. Edmonds. 2003. Magmatic and amagmatic seafloor generation at the ultraslow-spreading Gakkel Ridge, Arctic Ocean. Nature 423:956–961, https://doi.org/ 10.1038/nature01704. S. Suman, C. Mertens, and others. 2022a. Volcanically hosted venting with indications of ultramafic influence at Aurora hydrothermal field on Gakkel Ridge. Nature Communications 13:6517, https://doi.org/10.1038/s41467-022-34014-0. Boetius, A. 2015. The Expedition PS86 of the Research Vessel POLARSTERN to the Arctic Ocean in 2014. Reports on Polar and Marine Research 685, Alfred Wegener Institute for Polar and Marine Research, Bremerhaven, 133 pp., https://doi.org/10.2312/BzPM_0685_2015. https://doi.org/10.1038/s41467-022-34014-0. German, C.R., D.K. Blackman, A.T. Fisher, P.R. Girguis, K.P. Hand, T.M. Hoehler, J.A. Huber, J.C. Marshall, K.R. Pietro, J.S. Seewald, and others. 2022b. Ocean system science to inform the explora- tion of ocean worlds. Oceanography 35(1):16–22, https://doi.org/10.5670/oceanog.2021.411. Mittelstaedt, E., J. Escartin, N. Gracias, J-A. Olive, T. Barreyre, A. Davaille, M. Cannat, and R. Garcia. 2012. Quantifying diffuse and discrete venting at the Tour Eiffel vent site, Lucky Strike hydrothermal field. Geochemistry, Geophysics, Geosystems 13(4), https://doi.org/10.1029/2011GC003991. https://doi.org/10.2312/BzPM_0685_2015. Bünz, S., E. Ramirez-Llodra, and the HACON19 Cruise Participants. 2019. RV Kronprins Håkon (cruise no. 2019708), Longyearbyen – Longyearbyen, 19.09– 16.10.2019. Cruise Report, 100 pp., https://doi. org/10.7557/cage.6779. Hand, K.P., R.W. Carlson, and C.F. Chyba. 2007. Energy, chemical disequilib- rium, and geological constraints on Europa. Astrobiology 7(6):1,006–1,022, https://doi.org/10.1089/ast.2007.0156. Pedersen, R.B., H.T. Rapp, I.H. Thorseth, M.D. Lilley, F.J.A.S. Barriga, T. Baumberger, K. Flesland, R. Fonseca, G.L. Früh-Green, S.L. Jørgensen. 2010. Discovery of a black smoker vent field and vent fauna at the Arctic Mid-Ocean Ridge. Nature Communications 1:126, https://doi.org/10.1038/ ncomms1124. Bünz, S., E. Ramirez-Llodra, and the HACON21 Cruise Participants. 2021. RV Kronprins Håkon (cruise no. REFERENCES Shanks, S.P. Sylva, E. Walsh, T. Pichler, and M. Rosner. 2011. Geochemistry of hydrothermal fluids from the PACMANUS, Northeast Pual and Vienna Woods hydrother- mal fields, Manus Basin, Papua New Guinea. Geochimica et Cosmochimica Acta 75:1,088–1,123, https://doi.org/10.1016/j.gca.2010.11.008. Corliss, J.B., J. Dymond, L. Gordon, J. Edmond, R. von Herzen, R. Ballard, K. Green, D. Williams, A. Bainbridge, K. Crane, and T. van Andel. 1979. Submarine thermal springs on the Galápagos Rift. Science 203:1,073–1,083, https://doi.org/10.1126/ i 203 4385 1073 Howell, K.L., A. Hilário, L. Allcock, D.M. Bailey, M. Baker, M.R. Clark, A. Colaço, J. Copley, E.E. Cordes, R. Danovaro, and others. 2020. A decade to study deep-sea life. Nature Ecology and Evolution 5:265–267, https://doi.org/10.1038/ s41559-020-01352-5. Dahle, H., I. Økland, I.H. Thorseth, R.B. Pederesen, I.H. Steen. 2015. Energy landscapes shape micro- bial communities in hydrothermal systems on the Arctic Mid-Ocean Ridge. International Society for Microbial Ecology Journal 9:1,593–1,606, https://doi.org/10.1038/ismej.2014.247. Reeves, E.P., J.M. McDermott, and J.S. Seewald. 2014. The origin of methanethiol in midocean ridge hydrothermal fluids. Proceedings of the National Academy of Sciences of the United States of America 111:5,474–5,479, https://doi.org/10.1073/ pnas.1400643111. Johannessen, K.C., N. McLoughlin, P.E. Vullum, and I.H. Thorseth. 2020. On the biogenicity of Fe-oxyhydroxide filaments in silicified low- temperature hydrothermal deposits: Implications for the identification of Fe-oxidizing bacte- ria in the rock record. Geobiology 18:31–53, https://doi.org/10.1111/gbi.12363. Ding, W., X. Niu, T. Zhang, S. Chen, S. Liu, P. Tan, F. Kong, Z. Jin, S. Huang, C. Wei, and others. 2022. Submarine wide-angle seismic experiments in the High Arctic: The JASMInE Expedition in the slow- est spreading Gakkel Ridge. Geosystems and Geoenvironment 1:100076, https://doi.org/10.1016/ j.geogeo.2022.100076. Resing, J.A., P.N. Sedwick, C.R. German, W.J. Jenkins, J.W. Moffett, B.M. Sohst, and A. Tagliabue. 2015. Basin-scale transport of hydrothermal dis- solved metals across the South Pacific Ocean. Nature 523:200–203, https://doi.org/10.1038/ nature14577. Jones, D.O.B., D.J. Amon, and A.S.A. Chapman. 2020. Deep-sea mining: Processes and impacts. Chapter 5 in Natural Capital and Exploitation of the Deep Ocean. M. Baker, E. Ramirez-Llodra, and P. Tyler, eds, Oxford University Press, https://doi.org/ 10.1093/oso/9780198841654.003.0005. Edmonds, H.N., P.J. Michael, E.T. Baker, D.P. Connelly, J.E. Snow, C.H. Langmuir, H.J.B. Dick, C.R. German, and D.W. Graham. 2003. Discovery of abun- dant hydrothermal venting on the ultraslow- spreading Gakkel Ridge in the Arctic Ocean. Nature 421:252–256, https://doi.org/10.1038/ nature01351. Roberts, E.M., F. Mienis, H.T. Rapp, U. Hanz, H.K. Meyer, and A.J. Davies. 2018. Oceanographic setting and short-timescale environmen- tal variability at an Arctic seamount sponge ground. REFERENCES Anantharaman, K., J.A. Breier, C.S. Sheik, and G.J. Dick. 2013. Evidence for hydrogen oxida- tion and metabolic plasticity in widespread deep- sea sulfur-oxidizing bacteria. Proceedings of the National Academy of Sciences of the United States of America 110(1):330–335, https://doi.org/10.1073/ pnas.1215340110. Since the first demonstration that a saltwater ocean must be present beneath the icy exterior of Europa, we now know of multiple salty-ocean worlds in our own solar system, including Jupiter’s other moons, Ganymede and Callisto, and Saturn’s moons Titan and Enceladus (Hand et al., 2020). Among them, those that have rocky seafloors in contact with salty oceans are of particular relevance. The discovery of chemosynthesis at vent sites beneath the ice-covered Arctic Ocean now allows us to articulate the case for why icy moon oceans could be inhab- ited, even if they are too far removed from sunlight to sustain photosynthesis (Hand and German, 2018; German et al., 2022b). Of course, animal life as we know it depends on oxygen, which is the result Antonelli, M.A., N.J. Pester, S.T. Brown, and D.J. DePaolo. 2017. Effect of paleoseawater com- position on hydrothermal exchange in mid- ocean ridges. Proceedings of the National Academy of Sciences of the United States of America 114:12,413–12,418, https://doi.org/10.1073/ pnas.1709145114. Bachraty, C., P. Legendre, and D. Desbruyères. 2009. Biogeographic relationships among deep-sea hydrothermal vent faunas at global scale. Deep Sea Research Part I 56:1,371–1,378, https://doi.org/ 10.1016/j.dsr.2009.01.009. Baker, E.T., Y.J. Chen, and J. Phipps Morgan. 1996. Relationship between near-axis hydrothermal cool- ing and the spreading rate of mid-ocean ridges. Earth and Planetary Science Letters 142:137–145, https://doi.org/10.1016/0012-821X(96)00097-0. p g ( ) Baker, M.C., E. Ramirez-Llodra, P.A. Tyler, C.R. German, A. Boetius, E.E. Cordes, N. Dubilier, C.R. Fisher, L.A. Levin, A. Metaxas, and others. 2010. Biogeography, ecology, and vulnera- bility of chemosynthetic ecosystems in the deep sea. Pp. 161–182 in Life in the World’s Oceans: Diversity, Distribution, and Abundance. A.D McIntyre, ed., Wiley-Blackwell, https://doi.org/ 10.1002/9781444325508.ch9. Oceanography \| https://doi.org/10.5670/oceanog.2023.103 Concept: Balancing return and resources in the search for life. The Planetary Science Journal 2:77, https://doi.org/10.3847/PSJ/abe4da. Concept: Balancing return and resources in the search for life. The Planetary Science Journal 2:77, https://doi.org/10.3847/PSJ/abe4da. German, C.R., E.T. Baker, C.A. Mevel, K. Tamaki, and the FUJI Scientific Team. 1998. Hydrothermal activity along the South West Indian Ridge. Nature 395:490–493, https://doi.org/10.1038/ 26730. Beaulieu, S.E., E.T. Baker, and C.R. German. 2015. Where are the undiscovered hydrothermal vents on oceanic spreading ridges? REFERENCES Muhammed Fatih Sert is PhD Candidate, CAGE, Department of Geoscience, UiT - The Arctic University of Norway, Tromsø, Norway. Ida H. Steen is Professor, Department of Biological Sciences & Centre for Deep Sea Research, University of Bergen, Bergen, Norway. Marie Stetzler is PhD Candidate, CAGE, Department of Geoscience, UiT - The Arctic University of Norway, Tromsø, Norway. Runar Stokke is Researcher, Department of Biological Sciences & Centre for Deep Sea Research, University of Bergen, Bergen, Norway. Lissette Victorero is Postdoctoral Researcher, Norwegian Institute for Water Research, Oslo, Norway. Francesca Vulcano is PhD Candidate, Department of Biological Sciences & Centre for Deep Sea Research, University of Bergen, Bergen, Norway. Stig Vågenes is Head Engineer, Department of Biological Sciences & Centre for Deep Sea Research, University of Bergen, Bergen, Norway, and REV Ocean, Lysaker, Norway. Kate Alyse Waghorn is Postdoctoral Fellow, and Stefan Buenz is Team Leader, both at the CAGE, Department of Geoscience, UiT - The Arctic University of Norway, Tromsø, Norway. Saur, J., S. Duling, L. Roth, X. Jia, D.F. Strobel, P.D. Feldman, U.R. Christensen, K.D. Retherford, M.A. McGrath, F. Musacchio, and others. 2015. The search for a subsurface ocean in Ganymede with Hubble Space Telescope observations of its auro- ral ovals. Journal of Geophysical Research: Space Physics 120:1,715– 1,737, http://doi.org/10.1002/ 2014JA020778. ACKNOWLEDGMENTS ACKNOWLEDGMENTS We would like to thank the captains, officers, and crew of R/V Kronprins Haakon and R/V Polarstern, the teams of ROV Aurora, ROV/AUV NUI, and OFOBS, as well as all participants in the cruises for their invaluable contributions while at sea. The HACON project was funded by the Norwegian Research Council (grant # 274330). The AURORA mission (PS86 2014) was funded by the Helmholtz Association, the MARUM DFG Cluster of Excellence at the University of Bremen (49926684), and the ERC Advanced Investigator Grant ABYSS (294757) to AB. CA, PAD, BF, GP, MFS, MS, KAW, and SB were supported by the Research Council of Norway (RCN) through its Centres of Excellence fund- ing scheme (project no. 223259). CRG acknowl- edges additional support in the USA from NOAA’s Office of Ocean Exploration and Research (Grant #’s NA14OAR4320158, NA19OAR0110406). CRG and KPH acknowledge NASA’s Astrobiology Program (Grant #’s NNX16AL04G, NSSC19K1427). Identification of macrofauna was supported by the project “Fauna of hydrothermal vents and cold seeps in Norwegian waters” funded by the Norwegian Biodiversity Information Centre (the Norwegian Taxonomy Initiative). REFERENCES Deep-Sea Research Part I 138:98–113, https://doi.org/10.1016/j.dsr.2018.06.007. Levin, L.A., A.R. Baco, D.A. Bowden, A. Colaco, E.E. Cordes, M.R. Cunha, A.W.J. Demopoulos, J. Gobin, B.M. Grupe, J. Le, and others. 2016. Hydrothermal vents and methane seeps: Rethinking the sphere of influence. Frontiers in Marine Science 3:72, https://doi.org/10.3389/ fmars.2016.00072. Rogers, A.D., P.A. Tyler, D.P. Connelly, J.T. Copley, R. James, R.D. Larter, K. Linse, R.A. Mills, A. Naveira-Garabato, R.D. Pancost, and others. 2012. The discovery of new deep-sea hydro- thermal vent communities in the Southern Ocean and implications for biogeography. PLoS Biology 10:e1001234, https://doi.org/10.1371/ journal.pbio.1001234. Georgieva, M.N., S. Taboada, A. Riesgo, C. Díez- Vives, F.C. De Leo, R.M. Jeffreys, J.T. Copley, C.T.S. Little, P. Ríos, J. Cristobo, and others. 2020. Evidence of vent-adaptation in sponges living at the periphery of hydrothermal vent environments: Ecological and evolutionary implications. Frontiers in Microbiology 11:1636, https://doi.org/10.3389/ fmicb.2020.01636. MacKenzie, S.M., M. Neveu, A.F. Davila, J.I. Lunine, K.L. Craft, M.L. Cable, C.M. Phillips-Lander, J.D. Hofgartner, J.L. Eigenbrode, J.H. Waite, and others 2021. The Enceladus Orbilander Mission Oceanography \| Early Online Release Centre for Environmental and Marine Studies & Biology Department, University of Aveiro, Campus Universitário de Santiago, Aveiro, Portugal. Lawrence Hislop is Communications Director, REV Ocean, Fornebu, Norway. John W. Jamieson is Canada Research Chair in Marine Geology, Department of Earth Sciences, Memorial University of Newfoundland, St. John’s, Newfoundland, Canada. Dimitri Kalnitchenko is Postdoctoral Fellow, CAGE, Department of Geoscience, UiT - The Arctic University of Norway, Tromsø, Norway. Achim Mall is Postdoctoral Fellow, Department of Biological Sciences & Centre for Deep Sea Research, University of Bergen, Bergen, Norway. Giuliana Panieri is Professor, CAGE, Department of Geoscience, UiT - The Arctic University of Norway, Tromsø, Norway. Autun Purser is Senior Researcher, Alfred Wegener Institute, Helmholtz Centre for Polar & Marine Research, Bremerhaven, Germany. Sofia P. Ramalho is Assistant Researcher, Centre for Environmental and Marine Studies & Biology Department, University of Aveiro, Campus Universitário de Santiago, Aveiro, Portugal. Eoghan P. Reeves is Associate Professor, Department of Earth Science & Centre for Deep Sea Research, University of Bergen, Bergen, Norway. Leighton Rolley is Science Systems Manager, REV Ocean, Lysaker, Norway. Samuel I. Pereira is PhD Candidate, Department of Earth Science & Centre for Deep Sea Research, University of Bergen, Bergen, Norway. Pedro A. Ribeiro is Researcher, Department of Biological Sciences & Centre for Deep Sea Research, University of Bergen, Bergen, Norway. ARTICLE CITATION Vanreusel, A., L. Clough, K. Jacobsen, W. Ambrose, J. Jivaluk, V. Ryheul, R. Herman, and M. Vincx. 2000. Meiobenthos of the central Arctic Ocean with special emphasis on the nema- tode community structure. Deep Sea Research Part I 47:1,855–1,879, https://doi.org/10.1016/ S0967-0637(00)00007-8. Ramirez-Llodra, E., C. Argentino, M. Baker, A. Boetius C. Costa, H. Dahle, E.M. Denny, P.-A. Dessandier, M.H. Eilertsen, B. Ferre, C.R. German, K. Hand, A. Hilário, L. Hislop J.W. Jamieson, D. Kalnitchenko, A. Mall, G. Panieri, A. Purser, S.P. Ramalho, E.P. Reeves, L. Rolley, S.I. Pereira, P.A. Ribeiro, M.F. Sert, I.H. Steen, M. Stetzler, R. Stokke, L. Victorero, F. Vulcano, S. Vågenes, K.A. Waghorn, and S. Buenz. 2022. Hot vents beneath an icy ocean: The Aurora Vent Field, Gakkel Ridge, revealed. Oceanography, https://doi.org/10.5670/ oceanog.2023.103. Ver Eecke, H.C., D.A. Butterfield, J.A. Huber, M.D. Lilley, E.J. Olson, K.K. Roe, L.J. Evans, A.Y. Merkel, H.V. Cantin, and J.F. Holden. 2012. Hydrogen-limited growth of hyperthermophilic methanogens at deep-sea hydrothermal vents. Proceedings of the National Academy of Sciences of the United States of America 109:13,674–13,679, https://doi.org/10.1073/pnas.1206632109. AUTHORS i AUTHORS Eva Ramirez-Llodra (eva.ramirez@niva.no) is Science Coordinator, REV Ocean, Lysaker, and Senior Scientist, Norwegian Institute for Water Research, Oslo, Norway. Claudio Argentino is Postdoctoral Fellow, Centre for Arctic Gas Hydrate, Environment and Climate (CAGE), Department of Geoscience, UiT - The Arctic University of Norway, Tromsø, Norway. Maria Baker is Executive Director at Deep-Ocean Stewardship Initiative, University of Southampton, Southampton, UK. Antje Boetius is Professor, Alfred Wegener Institute, Helmholtz Centre for Polar & Marine Research, Bremerhaven, Germany; Max Planck Institute for Marine Microbiology, Bremen, Germany; and Center for Marine Environmental Sciences (MARUM), University of Bremen, Germany. Carolina Costa is a gradu- ate student in the Centre for Environmental and Marine Studies & Biology Department, University of Aveiro, Campus Universitário de Santiago, Aveiro, Portugal. Håkon Dahle is Professor, and Emily M. Denny is PhD Candidate, both in the Department of Biological Sciences & Centre for Deep Sea Research, University of Bergen, Bergen, Norway. Pierre-Antoine Dessandier is Postdoctoral Fellow, CAGE, Department of Geoscience, UiT - The Arctic University of Norway, Tromsø, Norway, and Researcher, University of Brest, CNRS, Ifremer, Plouzané, France. Mari H. Eilertsen is Postdoctoral Fellow, Department of Biological Sciences & Centre for Deep Sea Research, University of Bergen, Bergen, Norway. Benedicte Ferre is Team Leader, CAGE, Department of Geoscience, UiT - The Arctic University of Norway, Tromsø, Norway. Christopher R. German is Senior Scientist, Woods Hole Oceanographic Institution, Woods Hole, MA, USA. Kevin Hand is Deputy Project Scientist, Europa, NASA Jet Propulsion Laboratory, Pasadena, CA, USA. Ana Hilário is Assistant Researcher, Trumbo, S.K., M.E. Brown, and K.P. Hand 2019. Sodium chloride on the surface of Europa. Science Advances 5:eaaw7123, https://doi.org/10.1126/ sciadv.aaw7123. Van Dover, C.L., S. Arnaud-Haond, M. Gianni, S. Helmreich, J.A. Huber, A.L. Jaeckel, A. Metaxas, L. Pendleton, S. Petersen, E. Ramirez-Llodra, and others. 2018. Scientific rationale and interna- tional obligations for protection of active hydro- thermal vent ecosystems from deep-sea mining. Marine Policy 90:20–28, https://doi.org/10.1016/ j.marpol.2018.01.020. Van Dover, C.L. 2000. The Ecology of Deep-Sea Hydrothermal Vents. Princeton University Press, 446 pp. REFERENCES AH and SPR work was supported by funds from FCT/MCTES granted to CESAM (UIDP/50017/ 2020+UIDB/50017/2020+LA/P/0094/2020). SPR acknowledges additional support from the FCT/ MCTES, through the “CEEC Individual 2017” con- tract (CEECIND/00758/2017). LV was supported by the Norwegian Research Council project number 287934. PAR work and genetic connectivity studies were supported by the project “Eco-Safe” funded by the Research Council of Norway (project no. 326881). EPR work on chemical, isotopic and organic charac- terization of AVF fluids was additionally supported by the project “HyPOD” funded by the Research Council of Norway (project no. 287364). Schlindwein, V., and F. Schmid. 2016. Mid-ocean- ridge seismicity reveals extreme types of ocean lithosphere. Nature 535:276–279, https://doi.org/ 10.1038/nature18277. Schoening, T., M. Bergmann, J. Ontrup, J. Taylor, J. Dannheim, J. Gutt, A. Purser, and T.W. Nattkemper. 2012. Semi-automated image analysis for the assessment of megafaunal densities at the Arctic deep-sea observatory HAUSGARTEN. PLoS ONE 7:e38179, https://doi.org/ 10.1371/journal.pone.0038179. Seewald, J.S., K.W. Doherty, T.R. Hammar, and S.P. Liberatore. 2002. A new gas-tight isobaric sampler for hydrothermal fluids. Deep-Sea Research Part I 49:189–196, https://doi.org/ 10.1016/S0967-0637(01)00046-2. 10.1016/S0967-0637(01)00046-2. Sert, M.F., H. Niemann, E.P. Reeves, M.A. Granskog, K.P. Hand, T. Kekäläinen, J. Jänis, P.E. Rossel, B. Ferré, A. Silyakova, and F. Gründger. 2022. Compositions of dissolved organic matter in the ice-covered waters above the Aurora hydrother- mal vent system, Gakkel Ridge, Arctic Ocean. Biogeosciences 19:2,101–2,120, https://doi.org/ 10.5194/bg-19-2101-2022. SIDFEx. 2022. Sea Ice Drift Forecast Experiment (SIDFEx), https://www.polarprediction.net/key- yopp-activities/sea-ice-prediction-and-verification/ sea-ice-drift-forecast-experiment/. Sohn, R.A., C. Willis, S. Humphris, T.M. Shank, H. Singh, H.N. Edmonds, C. Kunz, U. Hedman, E. Helmke, M. Jakuba, and others. 2008. Explosive volcanism on the ultraslow-spreading Gakkel ridge, Arctic Ocean. Nature 453:1,236–1,238, https://doi.org/10.1038/nature07075. COPYRIGHT & USAGE This is an open access article made available under the terms of the Creative Commons Attribution 4.0 International License (https://creativecommons.org/ licenses/by/4.0/), which permits use, sharing, adapta- tion, distribution, and reproduction in any medium or format as long as users cite the materials appropri- ately, provide a link to the Creative Commons license, and indicate the changes that were made to the original content. Zeppilli, D., D. Leduc, C. Fontanier, D. Fontaneto, S. Fuchs, A.J. Gooday, A. Goineau, J. Ingels, V.N. Ivanenko, R.M. Kristensen, and others. 2018. Characteristics of meiofauna in extreme marine ecosystems: A review. Marine Biodiversity 48:35–71, https://doi.org/10.1007/ s12526-017-0815-z. Oceanography \| https://doi.org/10.5670/oceanog.2023.103
https://openalex.org/W4205551838	https://s3.ca-central-1.amazonaws.com/assets.jmir.org/assets/preprints/preprint-19642-accepted.pdf	English	null	Clinical Characteristics and Outcomes of Childbearing-Age Women With COVID-19 in Wuhan: Retrospective, Single-Center Study (Preprint)	null	2,020	cc-by	8,762	JMIR Preprints JMIR Preprints Clinical characteristics and outcomes of childbearing- age women with Coronavirus disease 2019 in Wuhan: a retrospective, single-center study Submitted to: Journal of Medical Internet Research on: April 26, 2020 Submitted to: Journal of Medical Internet Research on: April 26, 2020 Disclaimer: © The authors. All rights reserved. This is a privileged document currently under peer-review/community review. Authors have provided JMIR Publications with an exclusive license to publish this preprint on it's website for review purposes only. While the final peer-reviewed paper may be licensed under a CC BY license on publication, at this stage authors and publisher expressively prohibit redistribution of this draft paper other than for review purposes. Disclaimer: © The authors. All rights reserved. This is a privileged document currently under peer-review/community review. Authors have provided JMIR Publications with an exclusive license to publish this preprint on it's website for review purposes only. While the final peer-reviewed paper may be licensed under a CC BY license on publication, at this stage authors and publisher expressively prohibit redistribution of this draft paper other than for review purposes. Disclaimer: © The authors. All rights reserved. This is a privileged document currently under peer-review/community review. Authors have provided JMIR Publications with an exclusive license to publish this preprint on it's website for review purposes only. While the final peer-reviewed paper may be licensed under a CC BY license on publication, at this stage authors and publisher expressively prohibit redistribution of this draft paper other than for review purposes. [unpublished, non-peer-reviewed preprint] https://preprints.jmir.org/preprint/19642 Wang et al JMIR Preprints Table of Contents https://preprints.jmir.org/preprint/19642 Abstract Background: Since December 2019, an outbreak of coronavirus disease 2019 (COVID-19) caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) spread rapidly worldwide. Previous studies on pregnant patients were limited. Objective: The objective of our study was to evaluate the clinical characteristics and outcomes of pregnant and non-pregnant women with COVID-19. Methods: This study retrospectively collected epidemiological, clinical, laboratory, imaging, management, and outcome data of 43 childbearing-age women patients (including 17 pregnant and 26 non-pregnant patients) who presented with laboratory- confirmed of COVID-19 in Tongji Hospital, Wuhan, China, from Jan 19 to Mar 2, 2020. Clinical outcomes were followed up to Mar 28, 2020. Results: Of 43 childbearing-age women in this study, none developed severe adverse illness and died. The median ages of pregnant and non-pregnant women were 33.0 and 33.5 years, respectively. Pregnant women had a markedly higher proportion of history exposure to hospitals within two weeks before onset (53% vs 19%, P=.02), and a lower proportion of other family members affected (24% vs 73%%, P=.004). Fever (47% vs 69%) and cough (53% vs 46%) were common onsets of symptoms for two groups. Abdominal pain (24%), vaginal bleeding (6%), reduced fetal movement (6%), and increased fetal movement (13%) were observed at onset in pregnant patients. Higher neutrophil and lower lymphocyte percent were observed in the pregnant group (79% vs 56%, P<.001; 15% vs 33%, P<.001, respectively). In both groups were observed elevated concentration of high sensitivity C-reactive protein, erythrocyte sedimentation rate, aminotransferase and lactate dehydrogenase. Concentrations of alkaline phosphatase and D-dimer in the pregnant group were significantly higher than those of the non- pregnant group (119.0 vs 48.0 U/L, P<.001; 2.1vs 0.3?g/mL, P<.001). Both pregnant (4/10; 40%) and non-pregnant (8/15; 53%) women were tested positive for influenza A virus. A majority of pregnant and non-pregnant groups received antiviral (76% vs 96%) and antibiotic (76% vs 88%) therapy. Additionally, both pregnant (2/11; 18%) and non-pregnant (2/19; 11%) recovered women re-detected positive for SARS-CoV-2 after discharge. Conclusions: The epidemiology, clinical and laboratory features of pregnant women with COVID-19 were diverse and atypical, which increased the difficulty of diagnosis. Most pregnant women with COVID-19 were mild and moderate, and rarely developed severe pneumonia and severe adverse outcomes. (JMIR Preprints 26/04/2020:19642) (JMIR Preprints 26/04/2020:19642) DOI: https://doi.org/10.2196/preprints.19642 DOI: https://doi.org/10.2196/preprints.19642 https://preprints.jmir.org/preprint/19642 Shaoshuai Wang, Lijie Wei, Xuan Gao, Suhua Chen, Wanjiang Zeng, Jianli Wu, Xingguang Lin, Huiting Zhang, Lali Mwamaka Sharifu, Ling Chen, Ling Feng Clinical characteristics and outcomes of childbearing-age women with Coronavirus disease 2019 in Wuhan: a retrospective, single-center study Shaoshuai Wang, Lijie Wei, Xuan Gao, Suhua Chen, Wanjiang Zeng, Jianli Wu, Xingguang Lin, Huiting Zhang, Lali Mwamaka Sharifu, Ling Chen, Ling Feng Corresponding Author: Shaoshuai Wang Phone: +8683663812 Email: colombo2008@sina.com Table of Contents [unpublished, non-peer-reviewed preprint] [unpublished, non-peer-reviewed preprint] https://preprints.jmir.org/preprint/19642 Wang et al JMIR Preprints Clinical characteristics and outcomes of childbearing-age women with Coronavirus disease 2019 in Wuhan: a retrospective, single-center study https://preprints.jmir.org/preprint/19642 Yes, but only make the title and abstract visible (see Important note, above). I understand that if I later pay to participate in <a href="http https://preprints.jmir.org/preprint/19642 Preprint Settings 1) Would you like to publish your submitted manuscript as preprint? 1) Would you like to publish your submitted manuscript as preprint? Please make my preprint PDF available to anyone at any time (recommended). Please make my preprint PDF available only to logged-in users; I understand that my title and abstract will remain visible to all users. Only make the preprint title and abstract visible. N I d t i h t bli h b itt d i t i t Please make my preprint PDF available only to logged-in users; I understand that my title and abstract will remain visible to all users. Only make the preprint title and abstract visible. No, I do not wish to publish my submitted manuscript as a preprint. p y p p p 2) If accepted for publication in a JMIR journal, would you like the PDF to be visible to the public? 2) If accepted for publication in a JMIR journal, would you like the PDF to be visible to the public? es, please make my accepted manuscript PDF available to anyone at any time (Recommended). Yes, please make my accepted manuscript PDF available to anyone at any time (Recommende Yes, but please make my accepted manuscript PDF available only to logged-in users; I understand th Yes, please make my accepted manuscript PDF available to anyone at any time (Recommended). Yes, but please make my accepted manuscript PDF available only to logged-in users; I understand that the title and abstract will remain Yes, please make my accepted manuscript PDF available to anyone at any time (Recommended). Yes, but please make my accepted manuscript PDF available only to logged-in users; I understand that the title and abstract will remain v Yes, but please make my accepted manuscript PDF available only to logged-in users; I understand that the title and abstract will remain v Yes, but please make my accepted manuscript PDF available only to logged-in users; I understand that the title and abstract will remain v https://preprints.jmir.org/preprint/19642 [unpublished, non-peer-reviewed preprint] [unpublished, non-peer-reviewed preprint] Wang et al JMIR Preprints Yes, but only make the title and abstract visible (see Important note, above). I understand that if I later pay to participate in <a href="http [unpublished, non-peer-reviewed preprint] [unpublished, non-peer-reviewed preprint] https://preprints.jmir.org/preprint/19642 Wang et al JMIR Preprints These authors contribute equally to the work. These authors contribute equally to the work. Original Manuscript Original Manuscript [unpublished, non-peer-reviewed preprint] [unpublished, non-peer-reviewed preprint] https://preprints.jmir.org/preprint/19642 Wang et al JMIR Preprints Author Affiliations: 1 Department of Obstetrics and Gynecology, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, Hubei 430030, China. y gy 2 Department of Pediatrics, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, Hubei 430030, China. 2 Department of Pediatrics, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, Hubei 430030, China. Lijie Wei, Ph.D E-mail address: hongshu1414@163.com Xuan Gao, M.D E-mail address: xxxuangao@163.com Suhua Chen, Ph.D E-mail address: tj_csh@163.com Wanjiang Zeng, Ph.D E-mail address: wjzeng@tjh.tjmu.edu.cn Jianli Wu, Ph.D E-mail address: jlwu@tjh.tjmu.edu.cn Xingguang Lin, Ph.D E-mail address: linxgtedaich@126.com Huiting Zhang, M.D E-mail address: zhtmedical@163.com Lali Mwamaka Sharifu, Ph.D E-mail address: 1146114929@qq.com Ling Chen, Ph.D E-mail address: 790356760@qq.com Ling Feng, Ph.D E-mail address: fltj007@163.com Shaoshuai Wang, Ph.D E-mail address: colombo2008@sina.com Lijie Wei, Ph.D E-mail address: hongshu1414@163.com Xuan Gao, M.D E-mail address: xxxuangao@163.com Suhua Chen, Ph.D E-mail address: tj_csh@163.com Wanjiang Zeng, Ph.D E-mail address: wjzeng@tjh.tjmu.edu.cn Jianli Wu, Ph.D E-mail address: jlwu@tjh.tjmu.edu.cn Xingguang Lin, Ph.D E-mail address: linxgtedaich@126.com Huiting Zhang, M.D E-mail address: zhtmedical@163.com Lali Mwamaka Sharifu, Ph.D E-mail address: 1146114929@qq.com Ling Chen, Ph.D E-mail address: 790356760@qq.com Ling Feng, Ph.D E-mail address: fltj007@163.com Shaoshuai Wang, Ph.D E-mail address: colombo2008@sina.com Shaoshuai Wang, Ph.D E-mail address: colombo2008@sina.com haoshuai Wang, Ph.D E-mail address: colombo2008@sina.com Clinical characteristics and outcomes of childbearing-age women with Coronavirus disease 2019 in Wuhan: a retrospective, single-center study Lijie Wei1, Ph.D; Xuan Gao1, M.D.; Suhua Chen1, Ph.D; Wanjiang Zeng1, Ph.D; Jianli Wu1, Ph.D; Xingguang Lin1, Ph.D; Huiting Zhang1, M.D.; Lali Mwamaka Sharifu1, Ph.D; Ling Chen2, Ph.D; Ling Feng1, Ph.D; Shaoshuai Wang1, Ph.D Corresponding Authors: p g Prof Shaoshuai Wang Ph.D, Department of Obstetrics and Gynecology, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, Hubei 430030, China. (colombo2008@sina.com). Prof Ling Feng Ph.D, Department of Obstetrics and Gynecology, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, Hubei 430030, China. (fltj007@163.com). Corresponding author’s present address: Corresponding author’s present address: Prof Shaoshuai Wang Ph.D No. 1095, Jiefang Avenue Wuhan, Hubei 430030, China Or Prof Ling Feng Ph.D No. 1095, Jiefang Avenue Corresponding author’s present address: Prof Shaoshuai Wang Ph.D No. 1095, Jiefang Avenue Wuhan, Hubei 430030, China Or Prof Ling Feng Ph.D No. 1095, Jiefang Avenue [unpublished, non-peer-reviewed preprint] [unpublished, non-peer-reviewed preprint] https://preprints.jmir.org/preprint/19642 Wang et al JMIR Preprints Keywords: COVID-19; SARS-CoV-2; childbearing-age; pregnancy, clinical characteristics; outcomes Keywords: COVID-19; SARS-CoV-2; childbearing-age; pregnancy, clinical characteristics; outcomes. https://preprints.jmir.org/preprint/19642 Introduction In December 2019, a cluster of cases of pneumonia of unknown etiology was identified in Wuhan, China [1]. Further investigation revealed these cases were caused by a novel coronavirus, which was termed severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Pneumonia caused by SARS-CoV-2 was termed coronavirus disease 2019 (COVID-19) [2,3]. In the past two decades, two human coronaviruses, including severe acute respiratory syndrome coronavirus (SARS-CoV) and Middle East respiratory syndrome coronavirus (MERS-CoV), can cause severe lower respiratory tract infections [4,5]. SARS-CoV-2 is similar to SARS-CoV as both of them belong to the beta coronavirus genus, and SARS-CoV-2 share more than 79.6% sequence identity with SARS-CoV [6]. As of Apr 18, 2020, the cumulative number of confirmed cases of COVID-19 infection in China had exceeded 86,700, and the death toll more than 4,600. The cumulative total number of confirmed cases has globally exceeded 2,350,000, and continues to increase [7,8]. The World Health Organization has designated the COVID-19 pandemic a Public Health Emergency of International Concern. Pregnant women have been hypothesized to be susceptible to respiratory pathogens and severe adverse outcomes of pneumonia, due to the normal physiological changes during pregnancy, including altered cell-mediated immunity and changes in pulmonary [9,10]. Previous studies reported that pregnant women infected with SARS-CoV or MERS-CoV were more susceptible to severe adverse outcomes including maternal morbidity and death. The case fatality rate (CFR) for pregnant women infected SARS-CoV reached 25-30%, much higher than that of general population [11,12]. Data for pregnant women infected with MERS-CoV is scarce. A case series of five pregnant women with MERS reported the CFR reached 40% [13]. Unfortunately, there is limited experience on the COVID-19 infection during pregnancy, and all current studies are single-center trials. Two studies with a small sample size reported none of pregnant women with COVID-19 died yet [14,15]. However, currently there is no vaccine or specific treatment for COVID-19 infection. In this study, we described the clinical, laboratory, imaging findings and clinical outcomes of 43 childbearing-age women patients (including 17 pregnant and 26 non-pregnant women) in Wuhan infected with SARS-CoV-2. This will provide an insight into the prevention and treatment of pregnant women with COVID-19. ABSTRACT Background: Since December 2019, an outbreak of coronavirus disease 2019 (COVID-19) caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) spread rapidly in Wuhan and worldwide. However, previous studies on pregnant patients were limited. p p g p Objectives: The objective of our study was to evaluate the clinical characteristics and outcomes of pregnant and non-pregnant women with COVID-19. Methods: This study retrospectively collected epidemiological, clinical, laboratory, imaging, management, and outcome data of 43 childbearing-age women patients (including 17 pregnant and 26 non-pregnant patients) who presented with laboratory confirmed of COVID-19 in Tongji Hospital, Wuhan, China, from January 19 to March 2, 2020. Clinical outcomes were followed up to March 28, 2020. Results: Of 43 childbearing-age women in this study, none developed severe adverse illness and or died. The median ages of pregnant and non-pregnant women were 33.0 and 33.5 years, respectively. Pregnant women had a markedly higher proportion of history exposure to hospitals within two weeks before onset (53% vs 19%, P=.02), and a lower proportion of other family members affected (24% vs 73%%, P=.004). Fever (47% vs 69%) and cough (53% vs 46%) were common onset of symptoms for two groups. Abdominal pain (24%), vaginal bleeding (6%), reduced fetal movement (6%), and increased fetal movement (13%) were observed at onset in pregnant patients. Higher neutrophil and lower lymphocyte percent were observed in pregnant group (79% vs 56%, P<.001; 15% vs 33%, P<.001, respectively). In both groups, we observed elevated concentration of high sensitivity C- reactive protein, erythrocyte sedimentation rate, aminotransferase and lactate dehydrogenase. Concentrations of alkaline phosphatase and D-dimer in pregnant group were significantly higher than those of non-pregnant group (119.0 vs 48.0 U/L, P<.001; 2.1vs 0.3μg/mL, P<.001). Both pregnant (4/10; 40%) and non-pregnant (8/15; 53%) women were tested positive for influenza A virus. A majority of pregnant and non-pregnant groups received antiviral (76% vs 96%) and antibiotic (76% vs 88%) therapy. Additionally, both pregnant (2/11; 18%) and non-pregnant (2/19; 11%) recovered women re-detected positive for SARS-CoV-2 after discharge. Conclusions: The epidemiology, clinical and laboratory features of pregnant women with COVID- 19 were diverse and atypical, which increased the difficulty of diagnosis. Most pregnant women with COVID-19 were mild and moderate, and rarely developed severe pneumonia and severe adverse outcomes. Keywords: COVID-19; SARS-CoV-2; childbearing-age; pregnancy, clinical characteristics; outcomes. [unpublished, non-peer-reviewed preprint] https://preprints.jmir.org/preprint/19642 Wang et al JMIR Preprints https://preprints.jmir.org/preprint/19642 Recruitment This study retrospectively recruited patients from January 19 to March 2, 2020, at Tongji Hospital, Tongji Medical College of Huazhong University of Science and Technology, Wuhan, Hubei, China. According to the arrangements put in place by the Chinese Government, pregnant and non-pregnant women patients were admitted to the designated hospitals for managing the COVID-19 in Wuhan without selectivity. All patients were diagnosed with COVID-19 according to “Diagnosis and Treatment Protocol for COVID-19 (Sixth Trial Edition)” released by the National Health Commission of the People’s Republic of China” [16]. 17 pregnant and 32 non-pregnant women’s throat swabs tested positive for SARS-CoV-2 ribonucleic acid (RNA) from January 19 to March 2, 2020, among them, 6 non-pregnant women with comorbidities were excluded (two had hypertension, one had diabetes, one had a history of kidney transplantation, one had lymphoma, and one had connective tissue disease). The remaining 17 pregnant women and 26 non-pregnant women did not have any underlying comorbidities due to a chronic disease (such as hypertension, diabetes, or heart disease). Two groups were matched with respect to age, gender, timing of contacting COVID-19, and the proportion of health care workers. Additionally, all patients recruited were Chinese residents and lived in Wuhan with no exposure to Huanan seafood market in Wuhan. This study was reviewed and approved by the Ethics Committee of Tongji Hospital, Tongji Medical https://preprints.jmir.org/preprint/19642 [unpublished, non-peer-reviewed preprint] [unpublished, non-peer-reviewed preprint] Wang et al JMIR Preprints Wang et al College of Huazhong University of Science and Technology (TJ-IRB20200222). Informed consent for this retrospective study was waived. The anonymous data was collected and analyzed to facilitate better clinical decisions and treatment. Statistical analysis Statistical analysis was performed with SPSS Version 23.0 (IBM, Armonk, NY, USA). Continuous variables were presented as median (interquartile range). Categorical variables were expressed as number and proportion (%). Mann Whitney U test was applied for the comparing two groups of continuous variables. χ² test, or Fisher’s exact test were applied for discrete variables of two groups. A p-value with a two-tailed with test less than 0.05 was considered as statistically significant. https://preprints.jmir.org/preprint/19642 Data collection We retrospectively collected epidemiological, clinical, laboratory, imaging, management, and outcome data for all the COVID-19 patients in the two groups. Clinical outcomes were followed up to Mar 28, 2020. Two researchers evaluated the participants and reviewed the data independently, disagreements resolved by consensus (LW and XG). Throat swab specimens for all patients were tested for SARS-CoV-2 at Tongji Hospital. SARS-CoV- 2 was confirmed following the World Health Organization (WHO) guidelines for quantitative real- time reverse transcription polymerase chain reaction (qRT-PCR) [17]. Throat-swab specimens from the upper respiratory tract that were obtained from all patients on admission were maintained in viral-transport medium. Other pneumonia-related respiratory pathogens including influenza A virus, influenza B virus, respiratory-syncytial virus, adenovirus, parainfluenza viruses, legionella pneumophila, mycoplasma pneumoniae, and chlamydia pneumoniae were tested by enzyme-linked immunosorbent assay (ELISA). qRT-PCR and ELISA detection reagents were provided by Tongji Hospital. Additionally, except for one pregnant woman who did not consent, all the remaining patients took a chest computed tomography (CT). cs and clinical characteristics of pregnant women and non-pregnant women. Demographics and clinical characteristics of pregnant women and non-pregnant women. 17 pregnant and 26 non-pregnant women with COVID-19 were included in this study. Among the 17 pregnant women, one was in her first trimester, three were in their second trimester, and 13 were in their third trimester. None of them had a history exposure to Huanan seafood market. 18% of pregnant and 19% of non-pregnant women were health care workers. Pregnant women had a higher proportion of history exposure to hospitals within two weeks before onset (53% vs 19%, P=.02) and a lower proportion of other family members infected with COVID-19 (24% vs 73%, P=.004) than non-pregnant women. The median ages of pregnant and non-pregnant women were 33.0 and 33.5 years respectively. The median time from symptoms onset to hospital presentation in the pregnant and non-pregnant groups were 2.0 and 4.0 days. Two (12%) pregnant women and three (11%) non- pregnant women were diagnosed as severe type on admission. None of the patients developed critical illness (Table 1). The symptoms at onset of pregnant COVID-19 women were similar to non-pregnant women. The most common symptoms at onset of pregnant and non-pregnant women were fever (47% vs 69%) and cough (53% vs 46%). Other pneumonia-related symptoms at onset including fatigue, expectoration, chest tightness, and shortness of breath were less common. Chills and rigors, headache and myalgia had not been observed in pregnant women prior to the infection. Both pregnant (6%) and non-pregnant (15%) groups had diarrhea at onset. Two (12%) asymptomatic pregnant women were diagnosed during hospitalization routine tests as a requirement before delivery. Two (8%) asymptomatic non-pregnant women were diagnosed by testing for SARS-CoV-2 of throat swabs because they had a history of contacting with infected person. Additionally, pregnancy-related symptoms were also observed in pregnant women, including abdominal pain (24%), vaginal bleeding (6%), reduced fetal movement (6%), and increased fetal movement (13%). Two pregnant women only had pregnancy-related symptoms until being diagnosed (Table 1). https://preprints.jmir.org/preprint/19642 [unpublished, non-peer-reviewed preprint] Wang et al JMIR Preprints Table 1: Epidemiological and clinical features of pregnant and non-pregnant women with coronavirus disease 2019. Table 1: Epidemiological and clinical features of pregnant and non-pregnant women with coronavirus disease 2019. Variables Median (interquartile range) / No. cs and clinical characteristics of pregnant women and non-pregnant women. (%) Total (n=43) Pregnancy (n=17) Non-pregnancy (n=26) P value Median (IQR) age, years 33.0 (30.0-37.0) 33.0 (30.0-35.0) 33.5 (31.0-38.0) 0.28 Gestational age on admission First trimester - 1 (6) - - Second trimester - 3 (18) - - Third trimester - 13 (76) - - Health care workers 8 (19) 3 (18) 5 (19) 0.77 Hospital exposure within 2 weeks before onset 14 (33) 9 (53) 5 (19) 0.02 Other family members affected 23 (53) 4 (24) 19 (73) 0.004 Median (IQR) time from onset of symptom to first outpatient visit, days 3.5 (1.0-7.0) 2.0 (0.9-10.8) 4.0 (1.0-7.0) 0.75 Clinical classification Mild 3 (7) 2 (12) 1 (4) 0.54 Moderate 35 (81) 13 (76) 22 (85) - Severe 5 (12) 2 (12) 3 (12) - Critical 0 (0) 0 (0) 0 (0) - Symptoms at onset Fever 26 (60) 8 (47) 18 (69) 0.15 Chills and rigors 2 (5) 0 (0) 2 (8) 0.67 Headache 1 (2) 0 (0) 1 (4) 0.83 Dizziness 1 (2) 1 (6) 0 (0) 0.83 Fatigue 5 (12) 1 (6) 4 (15) 0.93 Cough 21 (49) 9 (53) 12 (46) 0.66 Expectoration 9 (21) 3 (18) 6 (23) 0.96 Chest tightness 5 (12) 2 (12) 3 (12) 0.64 Shortness of Breath 2 (5) 1 (6) 1 (4) 0.67 Myalgia 1 (2) 0 (0) 1 (4) 0.83 Diarrhea 5 (12) 1 (6) 4 (15) 0.64 Asymptomatic 4 (9) 2 (12) 2 (8) 0.93 Abdominal pain - 4 (24) - - Vaginal Bleeding - 1 (6) - - Reduced fetal movements - 1 (6) - - Increased fetal movement - 2 (13) - - iological and clinical features of pregnant and non-pregnant women with se 2019 Laboratory and imaging characteristics of pregnant women and non-pregnant women oratory and imaging characteristics of pregnant women and non-pregnant wome On admission, the median white blood cell count of patients in pregnant group with COVID-19 was significantly higher than non-pregnant group (7.8 vs 3.8×10^9/L, P<.001). Four (24%) pregnant women and zero (0%) non-pregnant women developed leukocytosis (white blood cell count >10.0 ×10^9/L). Neutrophil percentage and neutrophil count were higher in pregnant women (79% vs 56%, P<.001; 6.7 vs 2.3×10^9/L, P<.001, respectively) than non-pregnant women. Lymphopenia (lymphocyte count <1.0 ×10^9/L) occurred in seven (41%) of pregnant women and ten (38%) of non-pregnant women. There was no statistical difference in hemoglobin concentration and platelet count between the two groups. https://preprints.jmir.org/preprint/19642 cs and clinical characteristics of pregnant women and non-pregnant women. (Table 2) In both pregnant and non-pregnant groups we observed elevated high sensitivity C-reactive protein (hs-CRP) (≥10 mg/L, 70% vs 33%) and erythrocyte sedimentation rate (ESR) (>20 mm/hr, 100% vs 53%). The mean concentrations of alanine aminotransferase (ALT) or aspartate aminotransferase (AST) in the pregnant group were above the normal range, while non-pregnant group were normal. One patient in pregnant group had ALT of up to 882 U/L and AST of up to 783U/L. Concentration of lactate dehydrogenase (LDH) and alkaline phosphatase (ALP) in pregnant group were observed higher than non-pregnant group (235.0 vs 193.0 U/L; 119.0 vs 48.0 U/L). Additionally, 92% https://preprints.jmir.org/preprint/19642 [unpublished, non-peer-reviewed preprint] Wang et al JMIR Preprints Wang et al pregnant women were observed with an elevated D-dimer level, which was significantly higher than non-pregnant women (2.1 vs 0.3 μg/mL, P<.001) (Table 2). pregnant women were observed with an elevated D-dimer level, which was significantly higher than non-pregnant women (2.1 vs 0.3 μg/mL, P<.001) (Table 2). pregnant women were observed with an elevated D-dimer level, which was significantly higher than non-pregnant women (2.1 vs 0.3 μg/mL, P<.001) (Table 2). Serological examination of pneumonia-associated pathogens was performed in pregnant and non- pregnant COVID-19 patients. Four (40%) of pregnant women and eight (53 %) of non-pregnant women tested positive for influenza A virus IgM. Other respiratory viruses had not been observed. One (10%) of legionella pneumophila and one (10%) of mycoplasma pneumoniae for pregnant women tested positive. Except for one pregnant women who refused to undergo chest CT scan, all patients accepted chest CT examination. 41(98%) patients displayed typical findings of pneumonia, in which 9(21%) patients had unilateral pneumonia and 32(76%) patients had bilateral pneumonia. (Table 2; Figure 1). non pregnant women (2.1 vs 0.3 μg/mL, P .001) (Table 2). Serological examination of pneumonia-associated pathogens was performed in pregnant and non- pregnant COVID-19 patients. Four (40%) of pregnant women and eight (53 %) of non-pregnant women tested positive for influenza A virus IgM. Other respiratory viruses had not been observed. One (10%) of legionella pneumophila and one (10%) of mycoplasma pneumoniae for pregnant women tested positive. Except for one pregnant women who refused to undergo chest CT scan, all patients accepted chest CT examination. 41(98%) patients displayed typical findings of pneumonia, in which 9(21%) patients had unilateral pneumonia and 32(76%) patients had bilateral pneumonia. (Table 2; Figure 1). cs and clinical characteristics of pregnant women and non-pregnant women. Table 2: Laboratory and imaging features of pregnant and non-pregnant women with coronavirus disease 2019. Variables Median (interquartile range) / No. cs and clinical characteristics of pregnant women and non-pregnant women. (%) Total (n=43) Pregnancy (n=17) Non-pregnancy (n=26) P value Routine blood test Median (IQR) white blood cell count, ×10^9/L 5.2 (3.8-7.6) 7.8 (6.6-10.2) 3.8 (3.6-5.1) <0.001 <4.0 ×10^9/L 12 (28) 0 (0) 12 (46) 0.003 >10.0 ×10^9/L 4 (9) 4 (24) 0 (0) 0.04 Median (IQR) neutrophil percent, % 64.4 (55.9-79.4) 80.5 (72.2-85.2) 58.0 (49.4-63.0) <0.001 >75 % 12 (28) 12 (71) 0 (0) <0.001 Median (IQR) neutrophil count, ×10^9/L 3.3 (2.1-5.5) 6.7 (5.3-8.2) 2.3 (1.9-2.9) <0.001 <1.5 ×10^9/L 1 (2) 0 (0) 1 (3) 0.83 Median (IQR) lymphocyte percent, % 24.9(14.4-35.9) 13.0 (11.6-20.1) 32.7 (26.4-39.7) <0.001 <20 % 14 (33) 13 (77) 1 (4) <0.001 Median (IQR) lymphocyte count, ×10^9/L 1.4 (1.0-1.8) 1.1 (0.9-1.6) 1.4 (1.0-2.0) 0.21 <1.0 ×10^9/L 17 (40) 7 (42) 10 (38) 0.86 Median (IQR) hemoglobin, g/L 122.5 (113.8-128.5) 117.0 (111.0-132.0) 123.0 (117.0-127.0) 0.86 <115 g/L 12 (28) 6 (35) 6 (23) 0.38 Median (IQR) platelet count, ×10^9/L 209.0 (160.0-242.0) 198.0 (138.0-227.3) 210.0 (171.0-250.3) 0.24 <150 ×10^9/L 10 (23) 6 (35) 4 (15) 0.25 Other laboratory features Median (IQR) high sensitivity C-reactive protein, mg/L 6.7 (0.7-25.3) 16.7 (7.1-47.6) 1.6 (0.4-13.0) 0.07 ≥10mg/L, n/N (%) 14/31 (45) 7 /10 (70) 7/21 (33) 0.12 Median (IQR) procalcitonin, ng/mL 0.04 (0.03-0.05) 0.05 (0.03-0.17) 0.04 (0.03-0.05) 0.16 ≥ 0.05ng/mL, n/N (%) 0/23 (0) 0/9 (0) 0/14 (0) - Median (IQR) erythrocyte sedimentation rate, mm/h 26.0 (12.0-41.0) 36.5 (26.3-82.0) 24.0 (7.0-38.0) 0.08 >20mm/h, n/N (%) 13/20 (65) 5/5 (100) 8/15 (53) 0.11 Median (IQR) alanine aminotransferase, U/L 16.5 (9.0-26.0) 13.0 (9.0-28.0) 23.0 (9.0-26.5) 0.72 ≥ 45 U/L, n/N (%) 6/42 (14) 3/17 (18) 3/25 (12) 0.95 Median (IQR) aspartate aminotransferase, U/L 17.0 (13.0-28.3) 20.0 (14.0-42.5) 15.0 (10.5-25.0) 0.047 ≥ 35 U/L, n/N (%) 9/42 (21) 5/17 (29) 4/25 (16) 0.51 Median (IQR) lactate dehydrogenase, U/L 204.0 (172.0-286.0) 235.0 (182.0-309.0) 193.0 (161.0-277.0) 0.13 ≥250 U/L, n/N (%) 13/38 (34) 6/15 (40) 7/23 (30) 0.73 Median (IQR) alkaline phosphatase, U/L 57.5 (46.5-111.3) 119.0 (77.0-142.0) 48.0 (42.0-57.0) <0.001 ≥100 U/L, n/N (%) 11/38 (29) 10/15 (67) 1/23 (4) <0.001 Median (IQR) creatinine, μmol/L 52.5 (46.0-61.0) 50.0 (43.2-59.5) 53.0 (48.0-62.5) 0.21 ≥106 μmol/L, n/N (%) 0/38 (0) 0/14 (0) 0/24 (0) - Median (IQR) creatine kinase, U/L 51.5 (35.8-70.8) 81.0 (29.0-147.5) 48.5 (37.3-61.0) 0.34 ≥140 U/L, n/N (%) 1/20 (5) 1/6 (17) 0/14 (0) 0.30 Median (IQR) D-dimer, μg/mL 0.7 (0.3-2.0) 2.1 (1.7-3.1) 0.3 (0.2-0.7) <0.001 ≥0.5 μg/mL n/N (%) 19/34 (56) 11/12 (92) 8/22 (36) 0.003 Pneumonia-associated pathogens Respiratory-Syncytial Virus 0/24 (0) 0/10 (0) 0/14 (0) - Adenovirus 0/24 (0) 0/10 (0) 0/14 (0) - Influenza A virus 12/25 (48) 4/10 (40) 8/15 (53) 0.69 Table 2: Laboratory and imaging features of pregnant and non-pregnant women with coronavirus disease 2019. >20mm/h, n/N (%) Median (IQR) alanine aminotransferase, U/L Median (IQR) alanine aminotransferase, U/L cs and clinical characteristics of pregnant women and non-pregnant women. bles Median (interquartile range) / No (%) atory and imaging features of pregnant and non-pregnant women with oratory and imaging features of pregnant and non-pregnant women with sease 2019. Table 2: Laboratory and imaging features of pregnant and non-pregnant women with coronavirus disease 2019. Table 2: Laboratory and imaging features of pregnant and non-pregna coronavirus disease 2019. Variables Other laboratory features ≥10mg/L, n/N (%) Median (IQR) procalcitonin, ng/mL ≥ 0.05ng/mL, n/N (%) g ( ) Median (IQR) erythrocyte sedimentation rate, mm/h Median (IQR) erythrocyte sedimentation rate, mm/h ≥ 45 U/L, n/N (%) Median (IQR) aspartate aminotransferase, U/L Median (IQR) lactate dehydrogenase, U/L Median (IQR) alkaline phosphatase, U/L ≥100 U/L, n/N (%) Median (IQR) D-dimer, μg/mL ≥0.5 μg/mL n/N (%) https://preprints.jmir.org/preprint/19642 JMIR Preprints Wang et al Influenza B virus 0/25 (0) 0/10 (0) 0/15 (0) - Parainfluenza Viruses 0/24 (0) 0/10 (0) 0/14 (0) - Legionella pneumophila 1/24 (4) 1/10 (10) 0/14 (0) 0.42 Mycoplasma pneumoniae 2/22 (9) 1/10 (10) 1/12 (8) >0.99 Chlamydia pneumoniae 0/23 (0) 0/10 (0) 0/13 (0) - Chest computed tomographic findings Normal 1/42 (2) 0/16 (0) 1/26 (4) 0.67 Unilateral pneumonia 9/42 (21) 3/16 (19) 6/26 (23) - Bilateral pneumonia 32/42 (76) 13/16 (81) 19/26 (73) - Figure 1: Chest CT scans of four COVID-19 patients. A and B are chest CT showing axial view lung window of two pregnant women with COVID-19. A: Chest CT from a 34-year-old woman who was 38 weeks and 4 days pregnant, showing multiple bilateral ground-glass opacities. B: Chest CT from a 30-year-old woman who was 39 weeks and 1 day pregnant, showing left-sided ground-glass opacity. C and D are chest CT showing axial view lung window of two non-pregnant women with COVID-19. C: Chest CT from a 30-year-old woman showing multiple bilateral ground-glass opacities. D: Chest CT from a 33-year-old woman showing left-sided ground-glass opacity. JMIR Preprints Wang et al Influenza B virus 0/25 (0) 0/10 (0) 0/15 (0) - Parainfluenza Viruses 0/24 (0) 0/10 (0) 0/14 (0) - Legionella pneumophila 1/24 (4) 1/10 (10) 0/14 (0) 0.42 Mycoplasma pneumoniae 2/22 (9) 1/10 (10) 1/12 (8) >0.99 Chlamydia pneumoniae 0/23 (0) 0/10 (0) 0/13 (0) - Chest computed tomographic findings Normal 1/42 (2) 0/16 (0) 1/26 (4) 0.67 Unilateral pneumonia 9/42 (21) 3/16 (19) 6/26 (23) - Bilateral pneumonia 32/42 (76) 13/16 (81) 19/26 (73) - JMIR Preprints Figure 1: Chest CT scans of four COVID-19 patients. g p A and B are chest CT showing axial view lung window of two pregnant women with COVID-19. A: Chest CT from a 34-year-old woman who was 38 weeks and 4 days pregnant, showing multiple bilateral ground-glass opacities. B: Chest CT from a 30-year-old woman who was 39 weeks and 1 day pregnant, showing left-sided ground-glass opacity. C and D are chest CT showing axial view lung window of two non-pregnant women with COVID-19. C: Chest CT from a 30-year-old woman showing multiple bilateral ground-glass opacities. D: Chest CT from a 33-year-old woman showing left-sided ground-glass opacity. A and B are chest CT showing axial view lung window of two pregnant women with COVID-19. https://preprints.jmir.org/preprint/19642 ≥0.5 μg/mL n/N (%) A: Chest CT from a 34-year-old woman who was 38 weeks and 4 days pregnant, showing multiple bilateral ground-glass opacities. B: Chest CT from a 30-year-old woman who was 39 weeks and 1 day pregnant, showing left-sided ground-glass opacity. C and D are chest CT showing axial view lung window of two non-pregnant women with COVID-19. C: Chest CT from a 30-year-old woman showing multiple bilateral ground-glass opacities. D: Chest CT from a 33-year-old woman showing left-sided ground-glass opacity. nagement and clinical outcomes of pregnant women and non-pregnant women Management and clinical outcomes of pregnant women and non-pregnant women A majority of pregnant and non-pregnant COVID-19 patients received antiviral (76% vs 96%) and antibiotic (76% vs 88%) therapy. Four (24%) of pregnant and five (19%) of non-pregnant women received glucocorticoid therapy, and one (6%) of pregnant and three (124%) of non-pregnant women received immunoglobulins therapy. Compared with pregnant group the proportion of patients received antitussive therapy in non-pregnant group significantly increased (69% vs 35%, P=.03). Additionally, oxygen support was administered in 35% of pregnant and 54% of non-pregnant women with COVID-19. None of patients underwent mechanical ventilation, continuous renal replacement therapy, and extracorporeal membrane oxygenation (Table 3). None of the patients was lost in the follow-up during the study. None of the patients in the two groups were admitted to the intensive care unit (ICU), and none developed acute respiratory distress syndrome (ARDS), disseminated intravascular coagulation (DIC), renal failure, heart failure, secondary bacterial pneumonia and sepsis. And none of the patients died so far (Table 3). Two pregnant women were classified as severe illness on admission, neither progressed to critical y p p p ( ) Two pregnant women were classified as severe illness on admission, neither progressed to critical https://preprints.jmir.org/preprint/19642 [unpublished, non-peer-reviewed preprint] Wang et al Wang et al JMIR Preprints illness. And no miscarriage was observed in pregnant women. Ten of 11 pregnant women underwent cesarean sections (two had preterm birth) (Appendix). illness. And no miscarriage was observed in pregnant women. Ten of 11 pregnant women underwent cesarean sections (two had preterm birth) (Appendix). The median length of hospitalization for pregnant and non-pregnant groups was 17.0 and 22.0 days. And the median interval from onset to diagnose for SARS-CoV-2 were 4.0 and 10.0 days. The median duration of viral shedding after COVID-19 onset was 24.0 and 26.0 days. All patients who recovered from COVID-19 were placed in an isolation center for quarantine a period of two weeks. 11 pregnant and 19 non-pregnant women re-detected SARS-CoV-2 after discharge. Two (18%) of pregnant and two (11%) of non-pregnant women tested positive for SARS-CoV-2, and all were readmitted at hospitals for COVID-19 treatment (Table 3). Table 3: Clinical treatment and outcomes of pregnant and non-pregnant women with coronavirus disease 2019. Variables No. https://preprints.jmir.org/preprint/19642 nagement and clinical outcomes of pregnant women and non-pregnant women evidence indicating that pregnant women are more susceptible to the occurrence and severe adverse outcomes of COVID-19 than the general population. g p p A woman’s body is highly in an immunosuppressive state after pregnancy, and the anatomy, physiology and biochemistry will always change. For example, the immunity of T lymphocyte changes, the oxygen consumption increases and the diaphragm elevates, which increases the risk of respiratory infection of pregnant women [9,10]. Studies during the outbreak of influenza virus and SARS-CoV have demonstrated that pregnant women are more susceptible to severe illness. In the outbreak of "Spanish flu" in 1918, 675,000 people died, with an overall mortality rate of 1-2%, while 27% of pregnant women died, and the mortality rate of pregnant patients reached 50% or higher when complicated with secondary bacterial pneumonia [18,19]. In the outbreak of SARS in 2003, among 12 pregnant women diagnosed with SARS, six (50%) needed to be admitted to ICU, six (50%) underwent mechanical ventilation, the mortality rate was 25% [11]. Another study reported that six of ten (60%) pregnant women with SARS were admitted to ICU, four (40%) underwent mechanical ventilation, three (30%) progressed renal failure, and two (20%) progressed secondary sepsis, two (20%) progressed secondary DIC, and the mortality rate reached 30% [12]. During the COVID-19 outbreak in 2019, one study reported that none of the nine pregnant patients progressed to critically illness or dead [14]. Of the 16 cases of pregnant women with COVID-19, one was classified as severe but did not develop severe adverse outcomes in the later stage [20]. This is consistent with our findings that none of pregnant women with COVID-19 developed severe adverse outcomes. Although critical pneumonia and death have not been reported in pregnant women, we should still be alert to the possibility of pregnant women developing severe adverse outcomes considering the high similarity of genomic sequence between SARS-CoV and SARS-CoV-2 [6]. Comparison with Prior Work p In this study, none of the pregnant women had a history of exposure to Huanan Seafood Market, 53% had a routine prenatal care within two weeks before onset, 24% had a family cluster of COVID-19. Therefore, during the epidemic, it was recommended that pregnant women to delay their routine prenatal care for safety, unless it was necessary, or to take the form of online clinic in order to reduce the risks of nosocomial infection. https://preprints.jmir.org/preprint/19642 nagement and clinical outcomes of pregnant women and non-pregnant women (%) Total (n=43) Pregnancy (n=17) Non-pregnancy (n=26) P value Management Antiviral therapy 38 (88) 13 (76) 25 (96) 0.14 Antibiotic therapy 36 (84) 13 (76) 23 (88) 0.54 Glucocorticoid therapy 9 (21) 4 (24) 5 (19) 0.96 Immunoglobulin 4 (9) 1 (6) 3 (12) 0.93 Cough-suppressant therapy 24 (56) 6 (35) 18 (70) 0.03 Oxygen support (nasal cannula) 20 (47) 6 (35) 14 (54) 0.23 Mechanical ventilation 0 (0) 0 (0) 0 (0) - Non-invasive 0 (0) 0 (0) 0 (0) - Invasive 0 (0) 0 (0) 0 (0) - Continuous renal replacement therapy 0 (0) 0 (0) 0 (0) - Extracorporeal membrane oxygenation 0 (0) 0 (0) 0 (0) - Clinical outcomes Intensive care unit admission 0 (0) 0 (0) 0 (0) - Acute respiratory distress syndrome 0 (0) 0 (0) 0 (0) - Disseminated intravascular coagulation 0 (0) 0 (0) 0 (0) - Renal failure 0 (0) 0 (0) 0 (0) - Heart failure 0 (0) 0 (0) 0 (0) - Secondary bacterial pneumonia 0 (0) 0 (0) 0 (0) - Sepsis 0 (0) 0 (0) 0 (0) - Death 0 (0) 0 (0) 0 (0) - Median (IQR) time of hospitalization, days 22.0 (14.0-28.0) 17.0 (11.0-28.0) 22.0 (15.5-26.5) 0.53 Median (IQR) time from onset to diagnosis, days 9.5 (6.3-17.0) 4.0 (2.0-17.0) 10.0 (7.5-17.0) 0.09 Median (IQR) time of viral shedding after onset of symptom, days 25.0 (19.0-29.0) 24.0 (14.0-26.0) 26.0 (20.0-29.0) 0.21 Re-detectable positive for discharged patients 2/30 (7) 2/11 (18) 2/19 (11) 0.61 IQR, Interquartile range. Discussion i i l l Table 3: Clinical treatment and outcomes of pregnant and non-pregna coronavirus disease 2019. ical treatment and outcomes of pregnant and non-pregnant women with sease 2019. le 3: Clinical treatment and outcomes of pregnant and non-pregnant w navirus disease 2019 Principle Results This study retrospectively analyzed the epidemiological, clinical, laboratory, imaging characteristics and clinical outcomes of 43 women of childbearing age infected with COVID-19, including 17 pregnant women and 26 non-pregnant women. As of March 28, 2020, none of the patients involved in this study developed severe pneumonia or died. Based on our findings, currently there is no https://preprints.jmir.org/preprint/19642 [unpublished, non-peer-reviewed preprint] Wang et al JMIR Preprints Wang et al evidence indicating that pregnant women are more susceptible to the occurrence and severe adverse outcomes of COVID-19 than the general population. nagement and clinical outcomes of pregnant women and non-pregnant women Similar to previous studies, common symptoms at the onset of COVID-19 were fever and cough, and less common symptoms were expectoration, chest tightness and diarrhea [14,15,21,22]. Notably, the onset of symptoms for several pregnant women were atypical, given that they had no fever or cough before diagnosis but only symptoms related to pregnancy were observed, including abnormal pain, vaginal bleeding, increased or reduced fetal movement, which indicated that attention should be paid to the occurrence of atypical symptoms in pregnant women. Laboratory findings were significantly different in hematological parameters between the two groups. Leukocytosis featured prominently in pregnant patients [14,15], while leukopenia featured prominently in non-pregnant patients [21,22]. Lymphopenia is likely to occur in both groups. Elevated concentration of hs-CRP, D-dimer and liver enzymes (including ALT, AST, LDH and ALP) in pregnant COVID-19 patients were observed, none of them developed liver failure or coagulation disorders. Recently, a study of 274 cases of patients with COVID-19 found that deceased patients generally had markedly higher level of CRP and LDH than recovered patients [22]. Therefore, the possibility that pregnant women with COVID-19 develop severe adverse outcomes cannot be eliminated. Additionally, a certain proportion of co-infection of SARS-CoV-2 and influenza A virus were showed in two groups. Given the similar clinical manifestations caused by two viruses and a relatively low positive rate for SARS-CoV-2 RNA test, it is recommended that a comprehensive assessment including epidemiological exposure, symptoms, laboratory, and imaging tests is necessary to the diagnosis of COVID-19. y g Currently, vaccine or specific treatment for COVID-19 infection is absent. Majority of patients received antiviral (such as arbidol and oseltamivir) and empirical antibiotics treatment, while few patients received glucocorticoid and immunoglobulin therapy. Arbidol is an antiviral agent with a unique mechanism of action targeting the S protein/ angiotensin-converting enzyme 2 (ACE2) [unpublished, non-peer-reviewed preprint] Wang et al JMIR Preprints interaction and inhibiting membrane fusion of the viral envelope [23]. In vitro data suggested its activity against SARS [24]. And a nonrandomized study of 67 patients with COVID-19 reported that, compared with arbidol-untreated patients, arbidol-treated patients with a treatment for a median time of 9 days, showing a lower mortality rates (0% vs 16%) and higher discharge rates (33% vs 19%) [25]. However, limited data are available on the safety of medications used during pregnancy. nagement and clinical outcomes of pregnant women and non-pregnant women Oseltamivir is a neuraminidase inhibitor approved for the treatment of influenza, but it has no documented in vitro activity against SARS-CoV-2. Antibiotics were used routinely after operation to prevent secondary bacterial infections. Routinely systemic corticosteroids for treatment of COVID- 19 is not recommended [3]. A large proportion of non-pregnant women used antitussive drugs in this study, which was related to higher proportion of cough (77%) during disease progression (Appendix). Supportive therapy and oxygen therapy are important management for COVID-19 [3]. No significant difference in the length of hospitalization of COVID-19 patients was observed in two groups. Notably, both pregnant and non-pregnant recovered patients tested positive for SARS-CoV- 2 RNA during isolation. Fortunately, none of them suffered symptoms again or developed severe pneumonia. A case series including four patients with COVID-19 who had three repeated qRT-PCR after discharge or discontinuation of quarantine, four (100%) re-detected positive (RP) for SARS- CoV-2 RNA. All of them did not contact suspected or confirmed COVID-19 patients, and no family member was infected [26]. Thus, at least a proportion of recovered patients may still be virus carriers, and quarantine is still indispensable even after the patient with COVID-19 is discharged. Limitations Our study has some notable limitations. Firstly, this study is limited by its small sample size. More cases infected with COVID-19 should be enrolled for analysis. Secondly, only one pregnant woman were in her first trimester and three in their second trimester were included in this study respectively. The effect of COVID-19 on maternal and fetus in early pregnancy still needs to be clarified. Thirdly, this is a retrospective study, the uncertainty of the exact dates and related information on exposure (recall bias) might have an inevitable impact on assessment. Fourth, this study only included pregnant women and non-pregnant women, another group of healthy pregnant women should be included to assess pregnant outcomes of maternal and fetus, and intrauterine vertical transmission potential of COVID-19. Acknowledgements We thank all the patients involved in this study, the nurses and clinical staff who provided care for the patients. This work was supported by the research grants from the National Key Research and Development Program of China (grant numbers 2018YFC1002900, 2020YFC0846300) and the Fundamental Research Funds for the Central Universities (grant number 2020kfyXGYJ00). Conflicts of Interest: None declared. https://preprints.jmir.org/preprint/19642 Conclusion In the study, the clinical outcomes of pregnant women with COVID-19 appeared good, and none of the patients developed severe adverse outcomes. Additionally, the epidemiology of pregnant women with COVID-19 was complicated, and nosocomial infection cannot be underestimated. Fever and cough were the most common onset of symptoms in pregnant women. Notably, pregnancy-related symptoms (i.e., abdominal pain, vaginal bleeding, increased or decreased fetal movement) might be specific onset of symptoms for pregnant women with COVID-19. Quarantine is still needed after hospital discharge as a small proportion of recovered patients may still be virus carriers. In conclusion, early detection and active management effectively helps in the risk of developing severe pneumonia and death in pregnant women with COVID-19. Abbreviations Abbreviations ACE2: angiotensin-converting enzyme 2 ALP: alkaline phosphatase ALT: alanine aminotransferase ARDS: acute respiratory distress syndrome AST: aspartate aminotransferase CFR: case fatality rate COVID-19: coronavirus disease 2019 CT: computed tomography DIC: disseminated intravascular coagulation ELISA: enzyme-linked immunosorbent assay ESR: erythrocyte sedimentation rate hs-CRP: high sensitivity C-reactive protein ICU: intensive care unit IQR: interquartile range LDH: lactate dehydrogenase MERS-CoV: Middle East respiratory syndrome coronavirus qRT-PCR: quantitative real-time reverse transcription polymerase chain reaction RNA: ribonucleic acid RP: re-detectable positive SARS-CoV: severe acute respiratory syndrome coronavirus SARS-CoV-2: severe acute respiratory syndrome coronavirus 2 WHO: The World Health Organization Abbreviations ACE2: angiotensin-converting enzyme 2 ALP: alkaline phosphatase ALT: alanine aminotransferase ARDS: acute respiratory distress syndrome AST: aspartate aminotransferase CFR: case fatality rate COVID-19: coronavirus disease 2019 CT: computed tomography DIC: disseminated intravascular coagulation ELISA: enzyme-linked immunosorbent assay ESR: erythrocyte sedimentation rate hs-CRP: high sensitivity C-reactive protein ICU: intensive care unit IQR: interquartile range LDH: lactate dehydrogenase MERS-CoV: Middle East respiratory syndrome qRT-PCR: quantitative real-time reverse transcr RNA: ribonucleic acid RP: re-detectable positive SARS-CoV: severe acute respiratory syndrome SARS-CoV-2: severe acute respiratory syndrom WHO: The World Health Organization ACE2: angiotensin-converting enzyme 2 ALP: alkaline phosphatase ALT: alanine aminotransferase ARDS: acute respiratory distress syndrome AST: aspartate aminotransferase CFR: case fatality rate COVID-19: coronavirus disease 2019 CT: computed tomography DIC: disseminated intravascular coagulation ELISA: enzyme-linked immunosorbent assay ESR: erythrocyte sedimentation rate hs-CRP: high sensitivity C-reactive protein ICU: intensive care unit IQR: interquartile range LDH: lactate dehydrogenase MERS-CoV: Middle East respiratory syndrome qRT-PCR: quantitative real-time reverse transc RNA: ribonucleic acid RP: re-detectable positive SARS-CoV: severe acute respiratory syndrome SARS-CoV-2: severe acute respiratory syndrom WHO: The World Health Organization -CRP: high sensitivity C-reactive protein IQR: interquartile range LDH: lactate dehydrogenase MERS-CoV: Middle East respiratory syndrome coronavirus qRT-PCR: quantitative real-time reverse transcription polymerase chain reaction RNA ib l i id SARS-CoV: severe acute respiratory syndrome coronavirus SARS-CoV-2: severe acute respiratory syndrome coronavirus 2 WHO: The World Health Organization WHO: The World Health Organization https://preprints.jmir.org/preprint/19642 Author contributions: SW and LF made substantial contributions to the study concept and design. Author contributions: SW and LF made substantial contributions to the study concept and design. [unpublished, non-peer-reviewed preprint] Wang et al Wang et al JMIR Preprints LW and XG were in charge of the manuscript draft. SC, WZ and JW were responsible for obtaining written consent from patients, obtaining ethical approval, collecting data, and confirming the data accuracy. XL, HZ and LM Sharifu did the analysis and interpretation. LC was the paediatrician in charge of treatment of the newborn babies. All authors critically revised the manuscript for important intellectual content and gave final approval for the version to be published. References: Jamieson DJ, Theiler RN, Rasmussen SA. Emerging infections and pregnancy. Emerg Infect Dis 2006; 12: 1638-43. [doi: 10.3201/eid1211.060152] [PMID: 17283611] 11. Wong SF, Chow KM, Leung TN, et al. Pregnancy and perinatal outcomes of women with severe acute respiratory syndrome. Am J Obstet Gynecol 2004; 191: 292-7. [doi: 10.1016/ j.ajog.2003.11.019] [PMID: 15295381] 12. Lam CM, Wong SF, Leung TN, et al. A case-controlled study comparing clinical course and outcomes of pregnant and non-pregnant women with severe acute respiratory syndrome. BJOG 2004; 111: 771-4. [doi: 10.1111/j.1471-0528.2004.00199.x] [PMID: 15270922] 13. Assiri A, Abedi GR, Al Masri M, Bin Saeed A, Gerber SI, Watson JT. Middle East Respiratory Syndrome Coronavirus Infection During Pregnancy: A Report of 5 Cases From Saudi Arabia. Clin Infect Dis 2016; 63(7): 951-3. [doi: 10.1093/cid/ciw412] [PMID: 27358348] f ( ) 14. Chen H, Guo J, Wang C, et al. Clinical characteristics and intrauterine vertical transmission potential of COVID-19 infection in nine pregnant women: a retrospective review of medical records. Lancet 2020; 395: 809-15. [doi: 10.1016/S0140-6736(20)30360-3] [PMID: 32151335] 15. Yu N, Li W, Kang Q, et al. Clinical features and obstetric and neonatal outcomes of pregnant patients with COVID-19 in Wuhan, China: a retrospective, single-cente, descriptive study. Lancet Infect Dis 2020. [doi: 10.1016/S1473-3099(20)30176-6] [PMID: 32220284] f ( ) 16. General Office of the National Health Commission of China, General Office of the State Administration of Traditional Chinese Medicine. Diagnosis and Treatment Protocol for COVID- 19 (Sixth Trial Edition). 2020. (in Chinese). 2020. URL: http://www.nhc.gov.cn/yzygj/s7653p/202002/ 8334a8326dd94d329df351d7da8aeae2/files/b218cfeb1bc54639af227f922bf6b817.pdf. (Accessed April 26, 2020) 17. WHO. Laboratory testing for 2019 novel coronavirus (2019-nCoV) in suspected human cases. Interim guidance. Jan 17, 2020. URL: https://www.who.int/publications-detail/laboratory- testing-for-2019-novel-coronavirus-in-suspected-human-cases-20200117. (Accessed April 26, 2020) 18. John W Harris, Influenza occurring in pregnant women. JAMA 1919; 72: 978–80. doi:10.1001/ jama.1919.02610140008002 19. Taubenberger JK, Morens DM. 1918 Influenza: the mother of all pandemics. Emerg Infect Dis 2006; 12: 15-22. [doi: 10.3201/eid1201.050979] [PMID: 16494711] 20. Zhang L, Jiang Y, Wei M, et al. Analysis of the pregnancy outcomes in pregnant women with COVID-19 in Hubei Province [in Chinese]. Zhonghua Fu Chan Ke Za Zhi 2020; 55: E009. [doi: 10.3760/ cma.j.cn112141-20200218-00111] [PMID: 32145714] 21. Guan WJ, Ni ZY, Hu Y, et al. Clinical Characteristics of Coronavirus Disease 2019 in China. N Engl J Med. Published online Feb 28, 2020. [doi: 10.1056/NEJMoa2002032] [PMID: 32109013] 22. Chen T, Wu D, Chen HL, et al. Clinical characteristics of 113 deceased patients with coronavirus disease 2019: retrospective study. References: 1. Report of clustering pneumonia of unknown etiology in Wuhan City. Wuhan Municipal Health Commission, 2019. URL: http://wjw.wuhan.gov.cn/front/web/ showDetail/2019123108989 (Accessed April 26, 2020) 1. Report of clustering pneumonia of unknown etiology in Wuhan City. Wuhan Municipal Health Commission, 2019. URL: http://wjw.wuhan.gov.cn/front/web/ showDetail/2019123108989 (Accessed April 26, 2020) 2. Zhu N, Zhang D, Wang W, et al. A Novel Coronavirus from Patients with Pneumonia in China 2019. N Engl J Med 2020; 382: 727-33. [doi: 10.1056/ NEJMoa2001017] [PMID: 31978945] 3. WHO. Clinical management of severe acute respiratory infection when Novel coronavirus (nCoV) infection is suspected: interim guidance. March 13, 2020. URL: https://www.who.int/publications-detail/clinical-management-of-severe-acute-respiratory- infection-when-novel-coronavirus-(ncov)-infection-is-suspected (Accessed April 19, 2020) infection when novel coronavirus (ncov) infection is suspected (Accessed April 19, 2020) 4. Drosten C, Günther S, Preiser W, et al. Identification of a novel coronavirus in patients with severe acute respiratory syndrome. N Engl J Med 2003; 348:1967-76. [doi:10.1056/NEJMoa030747] [PMID: 12690091] ( ) p ( p ) 4. Drosten C, Günther S, Preiser W, et al. Identification of a novel coronavirus in patients with severe acute respiratory syndrome. N Engl J Med 2003; 348:1967-76. [doi:10.1056/NEJMoa030747] [PMID: 12690091] 5. Zaki AM, van Boheemen S, Bestebroer TM, Osterhaus AD, Fouchier RA. Isolation of a novel coronavirus from a man with pneumonia in Saudi Arabia. N Engl J Med 2012; 367:1814-20. [doi:10.1056/NEJMoa1211721] [PMID: 23075143] 5. Zaki AM, van Boheemen S, Bestebroer TM, Osterhaus AD, Fouchier RA. Isolation of a novel coronavirus from a man with pneumonia in Saudi Arabia. N Engl J Med 2012; 367:1814-20. [doi:10.1056/NEJMoa1211721] [PMID: 23075143] 6. Zhou P, Yang XL, Wang XG, et al. A pneumonia outbreak associated with a new coronavirus of probable bat origin. Nature 2020; 579: 270-3. [doi: 10.1038/s41586-020-2012-7] [PMID: 32015507] 6. Zhou P, Yang XL, Wang XG, et al. A pneumonia outbreak associated with a new coronavirus of probable bat origin. Nature 2020; 579: 270-3. [doi: 10.1038/s41586-020-2012-7] [PMID: 32015507] 7. World Health Organization. Coronavirus (COVID-19) Last updated: 2020/4/1 tps://preprints.jmir.org/preprint/19642 [unpublished, non-peer-reviewed preprint] Wang et al JMIR Preprints URL: https://who.sprinklr.com/. (Accessed April 26, 2020) 8. Office of Health Emergency, National Health Commission of the People’s Republic of China. 2020. URL: http://www.nhc.gov.cn/yjb/s7860/202004/ 2d391a171acc4624a50a1188c8de7361.shtml. (Accessed April 26, 2020) 8. Office of Health Emergency, National Health Commission of the People’s Republic of China. 2020. URL: http://www.nhc.gov.cn/yjb/s7860/202004/ 9. Weinberger SE, Weiss ST, Cohen WR, Weiss JW, Johnson TS. Pregnancy and the lung. Am Rev Respir Dis 1980; 121: 559-81. [doi: 10.1164/arrd.1980.121.3.559] [PMID: 6998334] 10. https://preprints.jmir.org/preprint/19642 derivatives against the pathogen of severe acute respiratory syndrome in the cell cultures [in Russian]. Vopr Virusol 2008; 53: 9-13. [PMID: 18756809] p 25. Wang Z, Yang B, Li Q, et al. Clinical Features of 69 cases with coronavirus disease 2019 in Wuhan, China. Clin Infect Dis. Published online Mar 16, 2020. [doi:10.1093/cid/ciaa272] [PMID: 32176772] https://preprints.jmir.org/preprint/19642 26. Lan L, Xu D, Ye G, et al. Positive RT-PCR Test Results in Patients Recovered From COVID-19. JAMA. Published online February 27, 2020. [doi:10.1001/jama.2020.2783] [PMID: 32105304] References: BMJ 2020. [doi: 10.1136/bmj.m1295] [PMID: 32234718] 23. Kadam RU , Wilson IA. Structural basis of influenza virus fusion inhibition by the antiviral drug Arbidol. Proc Natl Acad Sci U S A 2017; 114: 206-214. [doi:10.1073/pnas.1617020114] [PMID: 28003465] 24. Khamitov RA, Loginova SIa, Shchukina VN, et al. Antiviral activity of tps://preprints.jmir.org/preprint/19642 [unpublished, non-peer-reviewed preprint] https://preprints.jmir.org/preprint/19642 Wang et al JMIR Preprints Wang et al [ ] 26. Lan L, Xu D, Ye G, et al. Positive RT-PCR Test Results in Patients Recovered From COVID-19. JAMA. Published online February 27, 2020. [doi:10.1001/jama.2020.2783] [PMID: 32105304] [ ] 26. Lan L, Xu D, Ye G, et al. Positive RT-PCR Test Results in Patients Recovered From COVID-19. JAMA. Published online February 27, 2020. [doi:10.1001/jama.2020.2783] [PMID: 32105304] 26. Lan L, Xu D, Ye G, et al. Positive RT-PCR Test Results in Patients Recovered From COVID-19. JAMA. Published online February 27, 2020. [doi:10.1001/jama.2020.2783] [PMID: 32105304] 26. Lan L, Xu D, Ye G, et al. Positive RT-PCR Test Results in Patients Recovered From COVID-19. JAMA. Published online February 27, 2020. [doi:10.1001/jama.2020.2783] [PMID: 32105304] [unpublished, non-peer-reviewed preprint] [unpublished, non-peer-reviewed preprint] https://preprints.jmir.org/preprint/19642 Wang et al JMIR Preprints https://preprints.jmir.org/preprint/19642 https://preprints.jmir.org/preprint/19642 Supplementary Files [unpublished, non-peer-reviewed preprint] [unpublished, non-peer-reviewed preprint] https://preprints.jmir.org/preprint/19642 Wang et al JMIR Preprints Figures Figures Figures [unpublished, non-peer-reviewed preprint] [unpublished, non-peer-reviewed preprint] https://preprints.jmir.org/preprint/19642 Wang et al JMIR Preprints https://preprints.jmir.org/preprint/19642 Untitled. Untitled. https://preprints.jmir.org/preprint/19642 [unpublished, non-peer-reviewed preprint]
https://openalex.org/W4387274861	https://masujournal.org/store_file/archive/67-8-8-524-527.pdf	Maltese	null	Production of Protopectinase by Fusarium solani the Incitant of Root Rot of Onion	Madras Agricultural Journal	1,980	cc-by	2	https://doi.org/10.29321/MAJ.10.A02870 https://doi.org/10.29321/MAJ.10.A02870
https://openalex.org/W4388539277	https://www.qeios.com/read/C54DYX/pdf	English	null	Review of: "Assessment of soil erosion in the Cesar watershed, an initial step toward the restoration of the Cesar River"	null	2,023	cc-by	276	Qeios, CC-BY 4.0 · Review, November 9, 2023 Qeios ID: C54DYX · https://doi.org/10.32388/C54DYX Review of: "Assessment of soil erosion in the Cesar watershed, an initial step toward the restoration of the Cesar River" Yibeltal Zewdu Kumlachew Yibeltal Zewdu Kumlachew Potential competing interests: No potential competing interests to declare. G General comment on Assessment of soil erosion in the Cesar watershed, an initial step toward the restoration of the Cesar River. General comment on Assessment of soil erosion in the Cesar watershed, an initial step toward the restoration of the Cesar River. 1. The Title of the article is not understandable and easily captured. So I recommended that is better to modify the title of the article. 2. The abstract may not include the conclusion and results. So author should be incorporating the finding in the abstract, in quantitative way. 3. In the map of the study area, the legend of the each items are not visible, so please put your legend clearly, the colors you provide for each biomass should be differ. 4. The source of the data was not mentioned, please cite the source. 5. The table format is not good, please use scientific table format. 5. The table format is not good, please use scientific table format. 6. The source of the equations should be justified kindly, if they are not developed by the author. 7. And also denoted the equation like equation1,2,3… 8. Reference: Re-check the whole reference list just to make sure you have added all the references that you cited in your manuscript. 9. Apart from this the quality of the overall paper is good. Qeios ID: C54DYX · https://doi.org/10.32388/C54DYX 1/1
https://openalex.org/W2127250722	https://ogst.ifpenergiesnouvelles.fr/articles/ogst/pdf/2006/05/ogst06051.pdf	English	null	Combining Solid State Physics Concepts and X-Ray Absorption Spectroscopy to Understand DeNOx Catalysis	Oil & gas science and technology	2,006	cc-by	10,426	Oil & Gas Science and Technology – Rev. IFP, Vol. 61 (2006), No. 5, pp. 677-689 Copyright © 2006, Institut français du pétrole DOI: 10.2516/ogst:2006006 Oil & Gas Science and Technology – Rev. IFP, Vol. 61 (2006), No. 5, pp. 677-689 Copyright © 2006, Institut français du pétrole DOI: 10.2516/ogst:2006006 Oil & Gas Science and Technology – Rev. IFP, Vol. 61 (2006), No. 5, pp. 677-689 Copyright © 2006, Institut français du pétrole DOI: 10.2516/ogst:2006006 Combining Solid State Physics Concepts and X-Ray Absorption Spectroscopy to Understand DeNOx Catalysis D. Bazin1, D. Sayers2, J. Lynch3, L. Guczi4, G. Tréglia5 and C. Mottet5 1 LPS, Bâtiment 510, Université Paris XI, 91405 Orsay Cedex - France. 2 Department of Physics, North Carolina State University, 27695, Raleigh, North Carolina - United States 3 IFP,1 et 4, avenue de Bois-Preau, 92852 Rueil-Malmaison Cedex - France. 4 Department of Surface Chemistry and Catalysis, Institute of Isotope and Surface Chemistry, CRC HAS, P.O. Box 77, 1525 Budapest - Hungary. 5 CRMCN, Campus de Luminy, Case 913, 13288 Marseille Cedex 9 - France. e-mail: bazin@lps.u-psud.fr - john.lynch@ifp.fr - guczi@sunserv.kfki.hu - treglia@crmcn.univ-mrs.fr - mottet@crmcn.univ-mrs.fr Many of the experimental projects of the authors were inspired by the work of Dale Sayers whose death on Nov 25 2004 is much regretted Many of the experimental projects of the authors were inspired by the work of Dale Sayers whose death on Nov. 25 2004 is much regretted. Résumé — Contribution des concepts de la physique des solides et de la spectroscopie d'absorption des rayons X à la compréhension du processus de catalyse DeNOx — Partant des considérations théoriques, l’évaluation de la variation des propriétés d’agrégats métalliques en fonction de leur taille et leur composition mène à une prédiction du mode d’adsorption du NO sur ces nanoentités. En nous appuyant sur les résultats expérimentaux obtenus surtout par la spectroscopie d’absorption de rayons X, nous proposons de lier le mode d’adsorption de la molécule au comportement de l’agrégat métallique suite à l’adsorption. Une hypothèse simple nous mène à envisager une interprétation de l’activité catalytique d’agrégats métalliques nanométriques lors de l’adsorption du NO. Abstract — Combining Solid State Physics Concepts and X-Ray Absorption Spectroscopy to Understand DeNOx Catalysis — Considering the NO adsorption process, starting from theoretical considerations, evaluation of the variation of the properties of metallic clusters versus their size and composition leads to a prediction of the NO adsorption mode on these nanoentities. Then, based on experimental results obtained mostly through X-ray absorption spectroscopy, we have connected the adsorption mode of the molecule to the behaviour of the metallic cluster following the adsorption process. A simple hypothesis leads us to discuss the catalytic activity of nanometer scale metallic clusters following NO adsorption. Oil & Gas Science and Technology – Rev. IFP, Vol. 61 (2006), No. 5 678 interatomic distances. Combining Solid State Physics Concepts and X-Ray Absorption Spectroscopy to Understand DeNOx Catalysis In the following sections, we show that these structural parameters can be obtained by X-ray absorp- tion spectroscopy (XAS) and can be affected by the prepara- tion procedure. Then, in order to illustrate the possibilities offered by solid state physics in describing molecular adsorp- tion processes, the adsorption of a simple molecule, NO, on nanometer-scale metallic aggregates is discussed [15-17]. More precisely, we propose that a link exists between the adsorption mode of the NO molecule and the behaviour of the metallic cluster following the NO adsorption, this behav- iour being obtained from X-ray absorption studies. For tran- sition metals which lead to dissociative adsorption, metal oxide clusters are formed. This means that the catalytic activ- ity through the formation of nitrogen and oxygen from NO decreases significantly for these metals. On the contrary, when non dissociative adsorption is observed, a high temper- ature regime leads to growth of the metallic cluster. In the later case, a decrease in the catalytic activity is not observed. Thus, combining solid state physics and X-ray absorption spectroscopy seems to offer the opportunity to predict catalytic activity. Significant progress has also been achieved describing bimetallic clusters. Thus, a same approach to the prediction of catalytic activity is proposed for bimetallic systems. 1.1 Brief Description of the Theoretical Formalism Electronic structure methods extend from ab initio calculations to semi-empirical models such as pseudo-potential theory for transition metals or the tight-binding approximation for transi- tion metals. The tight-binding method starts from isolated atoms with discrete levels, which form energy bands when the atomic wave functions overlap. It assumes that any electronic state ψ(r) delocalised throughout the solid, can be written as a linear combination of atomic orbitals (LCAO): φλ(r – n) = \| n, λ 〉, where λ denotes the orbital at site n: this approximation becomes more accurate as the overlapping among the orbitals weakens (d states of transition metals). In parallel with the size dependence of the electronic structure, one observes inhomogeneous contractions of the cluster as well as a curvature of the facets. It is possible to obtain the equilibrium atomic configuration of the cluster at 0 K by performing a Quenched Molecular Dynamics (QMD) study in the second moment of interatomic potential. This results in the relaxation profiles shown in Figure 4 for icosa- hedron Pd clusters with sizes between 13 and 4000 atoms: (1) (1) (3) 0: means contraction of p shell) An essential advantage of this formalism is to give a simple access to the local density of states (LDOS) at a given site n0 from the Green function G. (3) (> 0: means contraction of p shell) (2) (2) where R(p) is the radius of the p-th shell from the center (R0(p) is the value before relaxation). 0 1 2 3 4 5 0 2 4 6 8 10 n (E) E (eV) Pd cuboctahedron EF 13 at. 55 at. 147 at. 309 at. 923 at. bulk Figure 3 Size effect on the Pd cuboctahedron cluster average density of states calculated in the tight-binding formalism. 0 1 2 3 4 5 0 2 4 6 8 10 n (E) E (eV) Pd cuboctahedron EF 13 at. 55 at. 147 at. 309 at. 923 at. bulk Figure 3 0 1 2 3 4 0 2 4 6 8 10 n (E) E (eV) s-p-d hybridization d band only 0 1 2 3 4 5 0 2 4 6 8 10 n (E) E (eV) Pd cuboctahedron EF 13 at. 55 at. 147 at. 309 at. 923 at. bulk Figure 2 Pd bulk density of states, with and without sp-d hybridiza- tion, calculated in the tight-binding framework. INTRODUCTION Nanomaterials [1] occupy a key position in the physical sciences [2, 3] in particular due to their specific surface [4, 5] and chemical [6, 7] properties. One example is supported metal catalysts [8, 9] consisting of a porous oxide such as alumina with a large specific area (> 200 m2/g) on which metal nanoparticles are dispersed. This family of materials is used in several major industrial or environmental applica- tions [10] among which are the reduction of nitrogen oxides emitted from car exhaust systems [11, 12] and the Fischer- Tropsch synthesis [13, 14]. In order to increase the perfor- mance of these industrial processes, it is necessary to investi- gate the chemical reaction at the atomic level i.e. to understand the adsorption phenomena of small molecules on such nanoentities. The aim of this paper is to made a bridge between solid state physics and heterogeneous catalysis, the ultimate goal being to use recent theoretical methods in solid state physics to predict the catalytic activity. The starting point is given by a brief description of the theoretical formalism developed for metallic clusters containing a small number of atoms, intro- ducing different significant parameters such the size and the morphology as well as the possibility of relaxation of the Cuboctahedron 13 55 147 Icosahedron 13 55 147 Figure 1 Cuboctahedron and icosahedron clusters containing 13, 55, 147 and 309 atoms. Cuboctahedron 13 147 147 147 147 309 309 147 Icosahedron 13 55 147 55 Figure 1 Cuboctahedron and icosahedron clusters containing 13, 55, 147 and 309 atoms. Cuboctahedron and icosahedron clusters containing 13, 55, 147 and 309 atoms. D Bazin et al. / Combining Solid State Physics Concepts and X-Ray Absorption Spectroscopy to Understand DeNOx Catalysis 679 D Bazin et al. / Combining Solid State Physics Concepts and X-Ray Absorption Spectroscopy to Understand DeNOx Catalysis 679 1 MONOMETALLIC NANOPARTICLES This concept of LDOS is particularly important for clusters since many non-equivalent sites exist at the surface (vertices, edges, facets) with various co-ordinations and atomic environments, which then define different LDOS. Note that at the end of the transition metal series, the s and p valence electrons and their hybridisation with d-orbitals have to be taken into account and that one has to be careful with charge self-consistency rules. In Figure 2 this is illustrated for Pd, for which such hybridisation is required in order to obtain a density of states in good agreement with that derived from ab initio calculations. Nanometer-scale metallic clusters can first be described in terms of their geometrical properties. As underlined by several studies [18, 19], a wide range of morphologies exist. In Figure 1, clusters (icosahedron with fivefold symmetry and cuboctahedron with fcc symmetry respectively) contain- ing a few hundred atoms are presented. In this section, we introduce some basic elements regarding the theoretical for- malism associated with the calculation of the electronic struc- ture of such entities. This approach also gives clear evidence that the local den- sities are significantly modified near the Fermi level, the amplitude of the modification depending on the site. In Figure 3, a significant size effect is observed for the Pd clus- ter with average density of states characteristic of a cubocta- hedron (note that similar results have obtained regarding icosahedron clusters). 1.1 Brief Description of the Theoretical Formalism Figure 3 Size effect on the Pd cuboctahedron cluster average density of states calculated in the tight-binding formalism. 0 1 2 3 4 0 2 4 6 8 10 n (E) E (eV) s-p-d hybridization d band only Figure 2 Figure 3 Size effect on the Pd cuboctahedron cluster average density of states calculated in the tight-binding formalism. Pd bulk density of states, with and without sp-d hybridiza- tion, calculated in the tight-binding framework. Pd bulk density of states, with and without sp-d hybridiza- tion, calculated in the tight-binding framework. Oil & Gas Science and Technology – Rev. IFP, Vol. 61 (2006), No. 5 Oil & Gas Science and Technology – Rev. IFP, Vol. 61 (2006), No. 5 680 Figure 4 Radial relaxation profiles in Pd clusters, for icosahedron shapes. Here, where R(p) is the radius of the p-th shell from the centre and R0(p) is the value before relaxation. 0 2 4 6 8 10 0 5 10 15 p δRp (%) (5) (4) (2) (3) (....) 0 2 4 6 8 10 0 5 10 15 p δRp (%) (5) (4) (2) (3) (....) and confirmed by several experiments (see for example references 37 and 38). The relaxation process, however, is constrained by symmetry requirements, thus it is not homogeneous throughout the particle. On the other hand, recent results obtained on Pd nano- particles show a modification of the structure and morphol- ogy induced by the substrate [39, 40]. We have thus to emphasize that we are now in the position to take into account the interaction between the metallic cluster and the support (or absorbed molecules). Lodziana and Nørskov [41] have shown that steps on the α-Al2O3 (0001) surface are enriched in oxygen, binding to Pd atoms and small clusters much more strongly than do the terraces. Microkinetic mod- eling of the NO + CO reaction on Pd particles supported on MgO [42] clearly demonstrates that the rate limiting step is, at low temperature, NO decomposition and, at high tempera- ture, CO adsorption. Figure 4 Radial relaxation profiles in Pd clusters, for icosahedron shapes. 1.3 Advantages and Limitations of the X-Ray Absorption Spectroscopy Here, where R(p) is the radius of the p-th shell from the centre and R0(p) is the value before relaxation. Among the different synchrotron techniques, X-ray absorp- tion spectroscopy (XAS) [43] has several specific advantages for the study of nanomaterials. Surprisingly, the icosahedron presents an “accordeon- like” profile with a contraction both at the surface and in the core, leading to a large distribution of distances around the bulk value. Derived from X-ray absorption near edge spectroscopy (XANES), the electronic state of the absorbing atoms has been qualitatively linked to the density of final states to which the transitions are made. However, at least two physi- cal phenomena affect the intensity of the “white line”: the size of the cluster, which is considered as an intrinsic effect (density of states in the platinum nanoparticles is signifi- cantly different from that of bulk Pt), and a possible charge transfer between the cluster and the support, which is consid- ered as an extrinsic effect [44]. Thus, special attention has to be paid if a simulation of the XANES spectra is performed with a linear combination of the XANES spectra of well- crystallised reference compounds [45]. 1.5 NO Adsorption on Nanometer Scale Metallic Clusters Combining recent calculations of the electronic structure of nanometer-scale metallic clusters with the suggestion of Brown [73] relating melting point to the ability of metallic surfaces to dissociate NO, we have proposed a relationship between the adsorption mode (dissociative or molecular) of NO at room temperature and the behaviour of nanometer- scale metallic particles (sintering or disruption) in response to this adsorption. Considering a range of elements, a straight line (see Fig. 5) separates two possibilities: associative adsorption of NO accompanied by sintering of the particles and dissociative adsorption accompanied by particle frag- mentation. This initial model was supported by experimental data on platinum [74-76] and ruthenium [77]. More recent data indicate that the behaviour of rhodium, iridium, palla- dium as well as copper is in line with this simple model. The metal/support interaction has a strong influence on the morphology of the cluster and also on the metal-metal dis- tance. Epitaxy between particles and the support has been observed by XAS [64]. Asakura et al. [65] found that plat- inum was stabilised in the form of raft-like nanoparticles when deposited on α-Al2O3(0001), as a result of direct Pt-O-Al bonding. Recent theoretical studies have taken this interaction into account. Yamauchi et al. [66] evaluated the structural and elec- tronic characteristics of Pd3 clusters on the MgO (100) surface from quantum chemical calculations based on density func- tional theory (DFT). A disruption of the Pd3 cluster may occur depending on the nature of the cluster/support interaction. More recently, Lopez et al. [67] performed an ab-initio study of metal deposition on SiO2. The authors considered the inter- action of Cun clusters (n = 1-5) with a non-bridging oxygen, constituting a paramagnetic point defect of silica, by DFT cal- culations. They predicted that the partial charge transfer to the oxide favours the formation of electrostatic interactions between the metal cluster and the oxygen atoms. Kantorovich et al. [68] characterised Mg clusters on MgO surfaces by ab-initio calculations. The calculations suggested that nucle- ation starts at the extended defects, e.g. on surface steps. A recent publication [78] described the interaction of NO with Rh nanoparticles supported on γ-Al2O3. According to X-ray absorption spectroscopy (XAS), the immediate neighbourhood of Rh atoms in the initial state is composed of 8 Rh atoms, with Rh-Rh bond length 2.68 Å. 1.4 The Structural Properties of Metallic Clusters the particles are dramatically transformed into a spherical shape. H2 adsorption can also modify the interaction between the cluster and the support. Deutsch et al. [101] pointed out that the interatomic distance between iridium and oxygen increased after treatment of the sample in H2. Reifsnyder et al. [102] produced Pt clusters by heating in vacuum at 300°C and measured a contracted Pt-Pt distance (2.66 Å). When H2 is chemisorbed, the interatomic distance is relaxed to 2.76 Å. The properties of catalysts depend on their electronic and structural characteristics, which ultimately are a function of the preparation procedure [57], the metal/support interaction, the stability of small metal nanoparticles and, last but not least, surface restructuring during the catalytic process itself. A crucial first point is the nature of the interaction between the precursor and the support during the preparation proce- dure [58, 59]. For example, in manufacturing monolithic auto- motive catalysts, using adsorption on a washcoat of Pt nitrate, rapid adsorption occurs due to a strong interaction, leading to low dispersion, in contrast to the highly dispersed Pt obtained when H2PtCl6 precursor is used in reforming [60]. Note that an elegant method to point out the influence of the support on catalytic properties, based on variation of the organometallic precursor, has been proposed by Dal Santo et al. [61]. Finally, we have to mention the effect of the reduction tem- perature. Vaarkamp et al. [71] found by XAS that after low temperature reduction (300°C), Pt particles were 3-dimen- sional with a Pt-O interatomic distance of 2.7 Å. At higher temperature (450°C), the morphology of the Pt particles changed from 3-dimensional to rafts with a structure similar to the Pt (100) surface. Comparison of XAS results of Pt/zeo- lite samples prepared using different techniques [72] showed that the average metal particle size roughly constant (in this case 10 ± 2 Å) for reduction temperatures up to 360oC, and above this a rapid growth of metal particles is seen. A large variety of metal oxides [62] have been used to stabilise and modify the electronic and structural properties of metallic clusters. Yoshitake and Iwasawa [63] have stud- ied a range systems such as Pt/Y2O3, Pt/ZrO2, Pt/V2O5 and Pt/TiO2. They noticed that the modification of the density of the unoccupied 5d states of Pt strongly depends on the nature of the support. 1.2 Applications to Metallic Systems The specific properties of metal nanoparticles, such as their activity/selectivity in catalysing chemical reactions [20-22], have motivated several fundamental studies. A large number of metallic systems such as V [23], Cr [24], Fe [25], Co [26], Ni [27], Cu [28], Y [29], Rh [30], Ru [31], Pd [32] and Pt [33] have been investigated, leading to a generalised under- standing of the electronic structure of different metals. Modern approaches simultaneously handle atomic and elec- tronic structures in a consistent way by performing a relax- ation process using molecular dynamics based, for example, on the method of Car and Parrinello [34]. For metal nanoparticles [46] the knowledge of the different structural parameters (i.e. co-ordination numbers and inter- atomic distances) allows us to determine cluster size [47-49], morphology [50] and the degree of relaxation [51]. As an analysis tool, XAS [52, 53] is best suited to very small clusters [54]. The numbers of nearest neighbours vary rapidly with the diameter of the particle for such materials. For larger particles, coordination numbers are similar to those in the bulk metal. One major limitation is related to the possible existance of a distribution of sizes. It is easy to show that one set of coordina- tion numbers may correspond to different size distributions. Although in nanoparticles the electronic structure and morphology largely depend on the nature of the metal, some preferential geometries exist. For instance, in metallic Ag clusters [35], different structures coexist between 2 and 4 nm, while for Au clusters [36], only a truncated octahedron mor- phology has been observed. The co-existence of clusters with different geometries implies that these different species display different electronic structures (and thus different activity/selectivity properties). Nevertheless, it should be emphasised that a unique advantage of the XAS technique is to give a precise measure- ment of the electronic state of the metal and to give direct structural evidence of heterometallic bonds (in the case of multimetallic systems) under in situ conditions, i.e. while the chemical reaction occurs [55, 56]. Relaxation of interatomic distances with decreasing cluster size has been predicted from theoretical calculations D Bazin et al. / Combining Solid State Physics Concepts and X-Ray Absorption Spectroscopy to Understand DeNOx Catalysis 681 Figure 5 Diagram showing a suggested correlation [15] between the adsorption mode and the behaviour of metallic clusters. The thick straight line represents the frontier between dissociative adsorption with fragmentation (above the line) and associa- tive adsorption with sintering (below the line). Diagram showing a suggested correlation [15] between the adsorption mode and the behaviour of metallic clusters. The thick straight line represents the frontier between dissociative adsorption with fragmentation (above the line) and associa- tive adsorption with sintering (below the line). Diagram showing a suggested correlation [15] between the adsorption mode and the behaviour of metallic clusters. The thick straight line represents the frontier between dissociative adsorption with fragmentation (above the line) and associa- tive adsorption with sintering (below the line). Here we wish to propose, even if there is so far no unambiguous structural evidence of this hypothesis, two different mechanisms which can explain the link between the adsorption mode of the molecule and the structural behaviour of the nanocluster. An interesting case is given by a metal of the second transition group i.e. palladium. On a perfect Pd(111) surface, NO adsorbs molecularly [82] while molecular adsorption and dissociation strongly compete on Pd(100) [83] and Pd(110) [84] surfaces. Finally, the stepped Pd(311) surface is active for the thermal dissociation of NO [85]. Thus it seems that the position of this metal close to the straight line in our inter- pretation is coherent with the high structural sensitivity of the ability of a Pd surface to dissociate NO. For nanometer scale Pd particles, no overall morphological changes of the nanocrystals were observed during NO gas exposure [86]. Thus it seems that the straight line can be considered as a “structural stability” line. When we consider metals below the straight line in Figure 5, after the initial state given by nanometer scale metallic cluster in presence of molecular NO, a non-dissocia- tive adsorption mode occurs. If the temperature is low, then we may observe a structural situation close to the one observed by Asakura et al. [65] in the case of platinum i.e. monomer species linked to several NO molecules. Then, with increase of temperature, the mobility of the nitrosyl species over the surface of the support leads to the formation of large clusters. 1.5 NO Adsorption on Nanometer Scale Metallic Clusters After exposure to 4% NO/He at 313 K for 5 seconds, the number of Rh-Rh bonds significantly decreased (from 8 to 2). Nitrogen atoms (NRhN = 1, dRhN = 1.78 Å) and oxygen atoms (NRhO = 2, RRhO = 2.05 Å) were present in the first coordination sphere. Comparable results were obtained previously on SiO2- supported Rh clusters by Krause and Schmidt [79]. They showed that NO alone disperses Rh over the support. The behaviour of Rh seems to be independent of the nature of the support. The fact that the metallic entities are generated by reduction under hydrogen alters the contraction of the interatomic dis- tances [69] and thus the geometrical properties as well as the electronic structure of the particles. For example, Wang et al. [70] observed that only a small truncation occurs in the tem- perature range between 350°C and 450°C. Above 500°C Iridium clusters have been investigated using X-ray diffraction as well as by thermal analysis. For a series of Oil & Gas Science and Technology – Rev. IFP, Vol. 61 (2006), No. 5 682 Figure 5 Diagram showing a suggested correlation [15] between the adsorption mode and the behaviour of metallic clusters. The thick straight line represents the frontier between dissociative adsorption with fragmentation (above the line) and associa- tive adsorption with sintering (below the line). 0 500 1000 1500 2000 2500 3000 3500 4000 10 20 30 40 50 60 70 80 90 3d 4d 5d Melting point (°C) Atomic number Pt Au Ir Os Re W Ta Hf Lu Ag Pd Rh Ru Mo Nb Zr Y Zn Cu Ni Mn Cr Na + Oa Fragmentation process of the metallic cluster NOa Sintering process of the metallic cluster Ir–H–ZSM-5 systems, Wögerbauer et al. [80] observed three different processes during interaction of NO and metallic Ir: Ir–H–ZSM-5 systems, Wögerbauer et al. [80] observed three different processes during interaction of NO and metallic Ir: Ir–H–ZSM-5 systems, Wögerbauer et al. 1.5 NO Adsorption on Nanometer Scale Metallic Clusters [80] observed three different processes during interaction of NO and metallic Ir: 0 500 1000 1500 2000 2500 3000 3500 4000 10 20 30 40 50 60 70 80 90 3d 4d 5d Melting point (°C) Atomic number Pt Au Ir Os Re W Ta Hf Lu Ag Pd Rh Ru Mo Nb Zr Y Zn Cu Ni Mn Cr Na + Oa Fragmentation process of the metallic cluster NOa Sintering process of the metallic cluster – decomposition of NO with formation of N2; – decomposition of NO with formation of N2O; – adsorption of NO on the Ir surface. The dominating process depends on temperature. Moreover, through a set of experiments performed at 500°C, they showed that NO in contact with Ir0 was decomposed to N2 with concurrent oxidation of Ir0 to IrO2. Their work underlines the influence of another key parameter, the crystallite size. In the case of a recent study on nanometer scale copper clusters [81], the sample was prepared by a method different from the classical impregnation technique. A 5 Å γ-Al2O3 film was created by exposing the clean surface of a stoichio- metric NiAl{110} crystal to oxygen and heating to 1200 K, produced a sharp LEED pattern characteristic of the Al2O3 structure. Copper was then evaporated from a bead source made by melting high-purity Cu wire wrapped around a tungsten filament. The adsorption and reaction of NO on the Cu clusters was studied using infrared, molecular beam, and scanning tunnelling spectroscopies over the temperature range of 90-300 K. The adsorption mode of NO was studied by IR spectroscopy. At low coverage two main bands were observed, at 1586 and 1500 cm-1, associated respectively with two distinct monomeric NO species, adsorbed in two- fold bridge sites on (110) type facets and three-fold hollow sites on facets which resemble (111) surfaces. Figure 5 As a preliminary conclusion, while this simple model was first supported by a set of experimental data on Pt and Ru, it seems that the behaviour of Rh, Ir, Cu and Pd clusters following the NO adsorption process is also in line with this simple scheme. When we consider metals above the “stability” line, a dissociative adsorption mode occurs. According to the work of Campbell et al. [78] the metallic atoms may have bonds either to oxygen or nitrogen atoms. At high temperature, the formation of a metallic oxide can be observed due to desorp- tion of nitrogen species (as N2 molecules). We would like now to address the effect of temperature on NO adsorption and propose physicochemical mechanisms. The starting point of this discussion is the work of Asakura et al. [65] who used XAS to study the structural transformation of Pt clusters on α-Al2O3 (0001) on NO adsorption. A nondissociative adsorption of NO on Pt clusters is observed. Moreover, the initial Pt raft is destroyed to form monomer species by the interaction with NO. We would like now to discuss the validity of the “stability” line which separates the two regimes for the NO adsorption. When a diatomic molecule such as NO approaches a metal surface [3], it may encounter three kinds of potential wells corresponding respectively to a physisorbed state (far from the surface), a molecular chemisorbed state and D Bazin et al. / Combining Solid State Physics Concepts and X-Ray Absorption Spectroscopy to Understand DeNOx Catalysis 683 683 D Bazin et al. / Combining Solid State Physics Concepts and X-Ray Absorption Spectroscopy to Understand DeNOx finally to a dissociative chemisorption state. Molecular chemisorption is the most stable situation if: It is firstly important to predict the equilibrium structure of the particles, since breaking the bulk symmetry leads to both atomic and chemical rearrangements. Unfortunately, it is not easy to treat both on the same level. (4) \|Eads(N)\| + \|Eads(O)\| < Edis(NO) + \|Eads(NO)\| (4) where Eads(X) is the atomic adsorption energy (< 0) of atom X while Edis(NO) and Eads(NO) are respectively the dissocia- tion energy (> 0) of the free molecule and the molecular chemisorption energy (< 0). The dissociation energy Ediss(NO) is clearly independent of the nature of the metal. 2.1 Effect of Finite Size on Surface Segregation Qualitatively, one expects that two effects play an important role. Firstly, we have to consider the effect of a finite number of atoms of each type. This is important for dilute systems when the available quantity of the element tending to segre- gate may be lower than the number of available surface sites. In this case, even if all atoms of one component are segre- gated to the surface, full coverage is not achieved. This may be the case for small particle size leading to a high surface/bulk ratio. Secondly, there is a geometrical effect due to the coexistence of facets with different orientations. This results in mixed concentration profiles which could be antag- onistic, leading to anti-phases boundaries. 2.2 Pt-4d-Metal Clusters the energy balance involved when exchanging a core atom with one at a p-site (p = vertices, edges, facets), only requires the knowledge of two parameters. The first one, issued from the linear term, is the variation of the site energy between the two pure in which a linear term and a quadratic one appear with respects to occupation numbers P in which are equal to 1 if site n is occupied by an atom of type i, and to zero if not. Both terms are calculated from the electronic structure of the disorder state described in the CPA approximation. Due to broken bonds, the equilibrium concentration on the various non-equivalent sites has no reason to be the same as in the bulk either for a surface (phenomenon of surface segrega- tion) or for vertex, edge, facet and core sites in clusters. In that case, the segregation energy, i.e. the energy balance involved when exchanging a core atom with one at a p-site (p = vertices, edges, facets), only requires the knowledge of two parameters. The first one, issued from the linear term, is the variation of the site energy between the two pure Figure 5 As adsorption energies increase, the two atomic adsorption terms on the left hand side of Equation 4 eventually come to dominate the single molecular adsorption term on the left hand side and dissociative chemisorption becomes more sta- ble. Thus the variation of adsorption mode along a transition series may be rationalised on the basis that the adsorption energies are higher for metals which are at the middle of the transition series, metals for which we observed here a disso- ciative adsorption. 2.2 Pt-4d-Metal Clusters For Pt-4d metal binary systems, it is necessary to develop another type of approximation, which will only be suited to study the effect of chemical ordering on a rigid lattice. Compared to monometallic systems, the total energy of the alloy, for a given configuration, cannot be described by a sum of pair interactions. Nevertheless, a (small) part of the energy which depends explicitly on configuration (essential due to ordering) can be described in an Ising-like form by developing the energy with perturbation with respect to that of the disordered state: as treated within the Coherent Potential Approximation (CPA) 91. For bimetallic systems A-B possessing non-equivalent sites, such as extended surfaces or clusters, one obtains: Is it possible to extract some information regarding the activ- ity of the metallic clusters from this purely energetic approach? For a first family of metals (above the stability line), the final structural characteristic of the supported phases is a metal oxide. Assuming that the metal oxide has no catalytic activity, the rate of conversion of NO molecules would tend to be quite low. Such a conclusion is in line with the experimental results obtained by Campbell et al. [78]. On the contrary, in the case of the second family of metals (below the line) large metallic clusters are finally generated. Let’s suppose that the metallic part of the cluster consists of a mixture of very small and large clusters. We may suppose that the first participate in the growth process after the adsorption of NO and thus do not contribute to catalytic activity. With time, the small particles disappear to form large ones. Following this hypothesis, the behaviour of the activity versus time may pass throuogh a maximum as under- lined by the work of Garcia-Cortès et al. [87]. (5) (5) in which a linear term and a quadratic one appear with respects to occupation numbers P in which are equal to 1 if site n is occupied by an atom of type i, and to zero if not. Both terms are calculated from the electronic structure of the disorder state described in the CPA approximation. Due to broken bonds, the equilibrium concentration on the various non-equivalent sites has no reason to be the same as in the bulk either for a surface (phenomenon of surface segrega- tion) or for vertex, edge, facet and core sites in clusters. In that case, the segregation energy, i.e. 2 BIMETALLIC CLUSTERS As pointed out by Barbier et al. [88], from an experimental point of view, the preparation procedure of bimetallic cata- lysts influences the type of interaction between the two metallic species. For example, electrochemical methods can be used to bring the two metals into close contact [89]. Modelling of catalytic reactions for a bimetallic cluster requires knowledge both of morphology, as is the case for monometallic entities, and of local chemical ordering before interaction with adsorbed molecules can be addressed [90]. Oil & Gas Science and Technology – Rev. IFP, Vol. 61 (2006), No. 5 Oil & Gas Science and Technology – Rev. IFP, Vol. 61 (2006), No. 5 684 elements (surface energy in the case of surface segregation: τA - τB). The second one, linked to the quadratic term, is nothing but the linear combination: systems which present respectively a strong tendency to chemical ordering (M = Mo: V >> 0), a weaker one (M = Pd, Rh: V > 0) and a tendency to phase separation (M = Ru, V < 0). Recently published studies show that progress in theoretical physics now allows the distribution of metals inside bimetallic nanoparticles to be predicted, at least in the case of compounds of elements chosen at the end of transition metal series. Although we are concerned with a description at thermodynamic equilibrium, the predicted distribution of the metals between the core and the surface of the particles is in general consistent with experimental data, despite the system being affected by preparation parameters such as the nature of the precursor. which can be used to calculate the mixing or ordering ener- gies of the system under study, but not its cohesive energy. Moreover, Vnm decreases rapidly with the distance (n-m) so that its sign for first neighbour interactions (Vnm = V) indi- cates the tendency to order (V > 0) or phase-separate (V < 0). It is easy to extend the Tight Binding Ising Method (TBIM) to this finite state by replacing layer concentrations by shell concentrations in which the shell i has Ni sites. One can divide the surface shell into four sub-shells corresponding to vertices (p = 1), edges (p = 2), (100) and (111) facets (p = 3, 4). Figure 6 Competition between segregation and ordering in PtPd clus- ters (Pt atoms are white, Pd atoms are black): cuboctahedron (left) and icosahedron (right). 2 BIMETALLIC CLUSTERS To simplify, let us first assume that the concentra- tions are homogeneous and depend only on coordination: cp = cc (the concentration in the “core”) for p > 4 (sites other than surface sites). The concentration profile is then given by: As a conclusion, this model can be used as a useful start- ing point for the preparation of such objects, even though, up to now, very few studies dedicated to bimetallic clusters have undertaken on this type of theoretical basis. Cuboctahedron PtPd2 (55 atoms) (6) (6) The effect of finite matter appears on the matter conserva- tion rule, since now there is no longer an infinite reservoir. If dispersion is defined by: Cuboctahedron PtPd2 (55 atoms) Cuboctahedron PtPd2 (55 atoms) Icosahedron Pt3Pd (55 atoms) this rule allows cc and then cp to be determined. this rule allows cc and then cp to be determined. (7) (7) Moreover, there may exist a competition between ordering and surface tension effects. This is illustrated in Figure 6 showing the results of Monte Carlo simulations in the TBIM potential for Pt-Pd clusters [92] at sufficiently low tempera- ture (T = 10 K). It is known that this system presents a ten- dency to ordering and that Pd should segregate due to its lower surface tension (τPd < τPt). For small clusters consisting of 55 atoms and two concentrations (PtPd2 cuboctahedron, Pt3Pd icosahedron) Pd is first segregated to the sites with low coordination (vertices, edges, etc.). For larger clusters of 923 atoms of concentration Pt6Pd4, Pd occupies successively the vertices, the edges, then square facets, but it begins to occupy bulk sites before the triangular sites are filled, which allows the L10 ordered structure to be preserved. As a result, one finds many non-equivalent surface sites, leading to some local order specific to the size and concentration. Icosahedron Pt3Pd (55 atoms) 2.3 Other Bimetallic Systems Bimetallic clusters are increasingly the subject of experimen- tal measurement [94-96]. Renouprez et al. [97] described the catalytic activity of a series of silica-supported Pd-Ni cata- lysts in the hydrogenation of 1,3-butadiene, for which pure palladium is an order of magnitude more active than pure nickel. The surface composition was measured by low In a recent paper [93], we limited ourselves to the treatment of Pt-M bimetallic systems where M is a 4d transi- tion metal. The chosen sequence allows us to consider D Bazin et al. / Combining Solid State Physics Concepts and X-Ray Absorption Spectroscopy to Understand DeNOx Catalysis 6 685 energy ion scattering (LEIS). The results were in agreement with the theoretical predictions for bulk alloys based on ther- modynamic calculations, once these are corrected for the effect of particle size, and show the migration of palladium from the bulk to the surface of the particles. Zhu et al. [98] reported a study on the structure of bimetallic Pd0.5Cu0.5 clus- ters, using X-ray diffraction and computer simulations. The structures obtained in the simulations exhibit face-dependent surface segregation, and alternating layers rich in one of the metals. As an example regarding electronic structure, in the bimetallic system Pt-Sn/SiO2, quantum chemical calculations employing density functional theory (DFT) for Pt19 and Pt16Sn3 clusters confirmed experimental evidence that Sn donates electrons to the 6s, 6p and 5d orbitals of platinum [99, 100]. It is important to underline that information regarding the electronic transfer between a molecule and the metal is also potentially available. tering process was not observed [14]. At the start of the process, the presence of Pt-Rh bonds is clearly shown by the shape of the Fourier transform modulus (Fig. 7). Oxidation of the bimetallic particles then takes place. It is thus clear that the presence of Rh in the environment of Pt significantly modifies the behaviour of the particles during the NO adsorp- tion process. The question now arises whether it is possible to explain the behaviour of the bimetallic cluster using Figure 5. A pos- sible approach is to consider a straight line between platinum and rhodium on the diagram proposed for monometallic sys- tems. In the Pt rich region, the behaviour of PtRh bimetallic clusters will be similar to the Pt monometallic system whereas, in the Rh rich region, their behaviour will be similar to the Rh monometallic system. 2.3 Other Bimetallic Systems If we consider now the experimental result obtained on the 1% Pt-1% Rh bimetallic system, we see that this result is in line with the conclusion obtained with this simple approach. CONCLUSION Recent theoretical calculations provide important information on the electronic structure and morphology of metal nanopar- ticles. The key goals are to understand the morphology of the cluster, the distribution of the two metals inside bimetallic clusters and the electronic transfer between two metals as well as between adsorbed molecules and atoms at the cluster surface. It is thus quite clear that the experimental data obtained by X-ray absorption spectroscopy can help in an 2.4 Advantages and Limitations of XAS Regarding Bimetallic Systems Figure 7 FT magnitude of the reduced PtRh catalyst at 25°C (a), under NO+N2 at 100°C (b), 200°C (c) and 300°C (d). Thick solid line: PtO2 reference, thick dashed line: Pt metal reference. 0 1 2 3 4 5 6 7 R (Å) (a) (d) (c) (b) FT Magnitude (a. u.) Pt, LIII 0 1 2 3 4 5 6 7 R (Å) (a) (d) (c) (b) FT Magnitude (a. u.) Pt, LIII Several studies have been performed on bimetallic systems using XAS because this is one of the few techniques capable of revealing the distribution of the two metals inside a nanometer scale cluster. In the case of a homogeneous sys- tem for which the core of the cluster is composed of NA atoms of type A and the surface consists of NB atoms of type B, the total coordination (NAA + NAB for A atoms, NBA + NBB for B atoms) is equal to 12 for A atoms, and less than 12 for B atoms [45]. More specifically, in the case of bimetallic clusters, the following relationships apply: NAA + NAB = 12 > NBA + NBB, NA NAB = NB * NBA (8) (8) Nevertheless, this property is no longer valid when monometallic clusters coexist with bimetallic ones. We have previously pointed out that the number of heterobonds decreases significantly as the content of monometallic cluster increases. If most of the atoms A are involved in monometal- lic clusters, we have NAA + NAB < NBA + NBB and thus the supposed distribution of the two metals inside the cluster, derived from the different coordination values as if only bimetallic cluster existed, is simply false. This point has never been clearly addressed in various experimental studies on bimetallic clusters. Figure 7 FT magnitude of the reduced PtRh catalyst at 25°C (a), under NO+N2 at 100°C (b), 200°C (c) and 300°C (d). Thick solid line: PtO2 reference, thick dashed line: Pt metal reference. 2.5 The Catalytic Activity of Bimetallic Clusters XAS experiments performed at the Pt LIII edge have been carried out recently on two catalysts: a monometallic 1 wt% Pt and a bimetallic 1 wt% Pt-1 wt% Rh system. Under an atmosphere of 1% NO in N2, strong sintering of Pt particles was observed above 473 K, while for Pt-Rh particles this sin- Oil & Gas Science and Technology – Rev. IFP, Vol. 61 (2006), No. 5 686 9 Gates, B.C. (2000) Supported nanostructured catalysts: Metal complexes and metal clusters. J. Mol. Catal. A-Chem., 163, 55. understanding at the atomic level. Analysis conditions enable the preparation to be followed and mimicking different chemical processes will give opportunities to optimise sev- eral industrial processes. 10 Russell, A.E. and Rose, A. (2004) X-ray absorption spec- troscopy of low temperature fuel cell catalysts, Chem. Rev., 104, 4613. Regarding the NO adsorption process, starting from sur- face science considerations, the theoretical evaluation of the variation of the cohesion energy of metallic clusters versus their size allows us a knowledge of the NO adsorption mode on these nanoentities. Then, based on experimental results obtained mostly through X-ray absorption spectroscopy, we have connected the adsorption mode of the molecule to the behaviour of the metallic cluster following the adsorption process. This simple hypothesis leads us to discuss the cat- alytic activity of nanometer scale metallic cluster regarding the NO adsorption process. Research work to confirm or modify this simple model is in progress. 11 Pârvulescu, V.I., Grange, P. and Delmon, B. (1998) Catalytic removal of NO. Catal. Today, 46, 233. 12 Garin, F. (2001) Environmental catalysis, Appl. Catal. A-Gen., 222, 183. 13 Dry, M.E. (1981) The Fischer-Tropsch Synthesis in Catalysis Sciences and Technology, Anderson, J.R. and Boudart, M., (Eds.) 1, Springer Verlag, Berlin, 159. 14 Van Wechem, V.M.H. and Senden, M.M.G., (1994) Conversion of natural gas to transportation fuels via the Shell Middle Distillate Process, Stud. Surf. Sci. Catal., 81, 43. 15 Bazin, D., Mottet, C., Tréglia, G. and Lynch, J. (2000) New trends in heterogeneous catalysis processes on metallic clus- ters from synchrotron radiation and theoretical studies, Appl. Surf. Sci. 164, 140. ACKNOWLEDGEMENTS 16 Bazin, D. (2002) Solid state concepts to understand catalysis using nanoscale metallic particles, Top. Catal., 18, 79. The authors are indebted to Dr. D. Spanjaard from the Laboratoire de Physique des Solides (Paris XI University, France), Dr. M.C. Desjonqueres (from the CEA-Saclay), Dr. F. Garin, Dr. G. Maire from the LERCSI (Strasbourg, France), Dr. C. Pichon, Dr. B. Rebours, Dr. R. Revel for the IFP, Dr. A. Seigneurin, Dr. Y. Klur, Dr. A. Pourpoint from Rhodia (Aubervilliers, France), Dr. M. Capelle, Dr. F. Maire, Dr. G. Meunier, Dr. R. Noirot, from PSA Peugeot-Citroën (Centre technique de Belchamp, France) for invaluable dis- cussions on the topic of nanometer-scale clusters. 17 Bazin, D. (2003) Solid State Physics and Synchrotron Radiation Techniques to Understand Heterogeneous Catalysis in nanotechnology, Ed. G.A. Somorjai, S. Hermans, B. Zhou, Ed. kluwer. 18 Ino, A. (1969) Stability of multiply-twinned particles J. Phys. Soc. Jpn, 27, 941. 19 Wulff, G.V. (1901) Zur frage der Geschwindigkeit des Wachstums und der Auflosung der Krystalflachen, Z. Krystallogr., 34, 449. 20 Pesant, L., Matta, J., Garin, F., Ledoux, M.J., Bernhardt, P., Pham, C. and Pham-Huu C. (2004) A high-performance Pt/- SiC catalyst for catalytic combustion of model carbon parti- cles (CPs). Appl. Catal. A-Gen., 266, 21. Also, the authors are indebted to the Hungarian Science and Research Fund (T-034920) for financial support and the COST D15/0005/99 program and the Chemistry Department of the CNRS for supporting a collaborative program with LURE synchrotron facilities. 21 Nohair, B., Especel, C., Marécot, P., Montassier, C., Hoang, L.C., and Barbier, J. (2004) Selective hydrogenation of sun- flower oil over supported precious metals. C. R. Acad. Sci. II C, 7, 113. 22 Henry, C. (2000) Catalytic activity of supported nanometer- sized metal clusters, Appl. Surf. Sci., 164, 252. REFERENCES (1979) Exafs of small Cu and Ni clusters: Binding-energy and bond- length changes with cluster size, Phys. Rev. Lett., 13, 165. 55 Lynch, J. (2002), Development of structural characterisation tools for catalysts. Oil Gas Sci. Technol., 57, 281. 38 Moraweck, B., Clugnet, G. and Renouprez, A.J. (1979) Contraction and relaxation of interatomic distances in small platinum particles from extended X-ray absorption fine struc- ture (EXAFS) spectroscopy, Surf. Sci., 81, L631. 56 Grunwaldt, J.D., Caravati, M., Hannemann, S. and Baiker, A. (2004) X-ray absorption spectroscopy under reaction condi- tions: suitability of different reaction cells for combined cata- lyst characterization and time-resolved studies, Phys. Chem. Chem. Phys., 6, 3037 39 Vervish, W., Mottet, C. and Goniakowski, J. (2002) Theoretical study of the atomic structure of Pd nanoclusters deposited on a MgO(100) surface, Phys. Rev. B, 65, 245411. 57 Shih, C.C. and Chang, J.R. (2005) Genesis and growth of platinum subnano-particles on activated-carbon characterized by X-ray absorption spectroscopy: effects of preparation con- ditions, Mater. Chem. Phys., 92, 89-97. 40 Mottet, C., Goniakowski, J., Baletto, F., Ferrando, R. and Tréglia, G. (2004) Modeling free and supported metalic nan- oclusters: Structure and dynamics, Phase Transit., 77, 101. 58 Yao, N., Pinckney, C., Lim, S., Pak, C. and Haller, G.L. (2001) Synthesis and characterization of Pt/MCM-41 cata- lysts, Micropor. Mesopor. Mat., 44-45, 377. 41 Lodziana, Z. and Nørskov, J.K. (2002) Interaction of Pd with steps on Al2O3 (0001), Surf. Sci., 518, L577. 59 Chen, Y., Ciuparu, D., Lim, S., Yang, Y., Haller, G.L. and Pfefferla, L. (2004) Synthesis of uniform diameter single-wall carbon nanotubes in Co-MCM-41: effects of the catalyst prere- duction and nanotube growth temperatures, J. Catal., 225, 453. 42 Prevot, G. and Henry, C.R. (2002) Microkinetic modeling of the CO + NO reaction on Pd/MgO particle, J. Phys. Chem. B, 106, 12191. 60 Dou, D., Liu, D.J., Williamson, W.B., Kharas, K.C. and Robota, H.J. (2001) Structure and chemical properties of Pt nitrate and application in three-way automotive emission cat- alysts, Appl. Catal. B-Environ., 30, 11. 43 Sayers, D.A., Lytle, F.W. and Stern, E.A. (1970) Advances in X-ray Analysis, Ed. Plenum, New-York, 13, 1970. 44 Bazin, D., Sayers, D., Rehr, J., Mottet, C. (1997) Numerical simulation of the Pt LIII edge white line relative to nanometer scale clusters. J. Phys. Chem., 100, 5332. 61 Dal Santo, V., Dossi, C., Recchia, S., Colavita, P.E., Vlaic, G. and Psaro, R. REFERENCES 23 Alvarado, P., Dorantes-Davila, J.L. and Pastor G.M. (1998) Magnetic properties of 3d transition-metal nanostructures: Cr and V clusters embedded in bulk Fe, Phys. Rev. B, 58, 12116. 1 Somorjai, G.A. (2004) On the Move, Nature, 430, 730. 2 Friedel, J. (1978) Physics of Metals, Cambridge University Press, Cambridge. 24 Reddy, B.V., Khanna, S.N. and Jena, P. (1999) Structure and magnetic ordering in Cr8 and Cr13 clusters, Phys. Rev. B 60, 15598. 3 Desjonqueres, M.C. and Spanjaard, D. (1998) Concepts in Surface Physics, Springer-Verlag, Berlin. 4 Sinfelt, J. (2002) Role of surface science in catalysis. Surf. Sci., 500, 923. 25 Oda, T., Pasquarello, A. and Car, R. (1998) Fully uncon- strained approach to noncollinear magnetism: Application to small Fe clusters, Phys. Rev. Lett., 80, 3622. 5 Zaera, F. (2001) Probing catalytic reactions at surfaces, Prog. Surf. Sci., 69, 1. 26 Guirado-Lopez, R. (2001) Magnetic anisotropy of fcc transi- tion-metal clusters: Role of surface relaxation, Phys. Rev. B, 63, 174420. 6 Thomas, J.M. and Thomas, W.J., (1997) Principles and Practice of Heterogeneous Catalysis, Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim. 27 Calleja, M., Rey, C., Alemany, M.M.G., Gallego, L.J. Ordejon, P., Sanchez-Portal, D., Artacho, E. and Soler, J.M. (1999) Self-consistent density-functional calculations of the geometries, electronic structures, and magnetic moments of Ni-Al clusters, Phys. Rev. B, 60, 2020. 7 Iwasawa, Y. (2003) In situ characterization of supported metal catalysts and model surfaces by time-resolved and three-dimensional XAFS techniques, J. Catal., 216, 165. 8 Somorjai, G.A. and McCrea, K. (2001) Roadmap for cataly- sis science in the 21st century: a personal view of building the future on past and present accomplishments. Appl. Catal. A-Gen., 222, 3. 28 Mottet, C. Tréglia, G. and Legrand, B. (1997) New magic numbers in metallic clusters: an unexpected metal depen- dence, Surf. Sci. Lett., 383, L719. D Bazin et al. / Combining Solid State Physics Concepts and X-Ray Absorption Spectroscopy to Understand DeNOx Catalysis 687 D Bazin et al. / Combining Solid State Physics Concepts and X-Ray Absorption Spectroscopy to Understand DeNOx Catalysis 687 48 Ramallo-López, J.M., Requejo, F.G., Craievich, A.F., Wei, J., Avalos-Borja, M. and Iglesia, E. (2005) Complementary methods for cluster size distribution measurements: sup- ported platinum nanoclusters in methane reforming catalysts, J. Mol. Catal. A-Chem., 228, 299. 29 Guirado-Lopez, R., Desjonqueres, M.C. and Spanjaard, D. (1999) Electronic and magnetic structure in 4d transition metal clusters, Appl. Surf. Sci., 144, 663. REFERENCES 30 Barreteau, C., Spanjaard, D. and Desjonqueres, M.C. (1999) Electronic structure and energetics of transition metal sur- faces and clusters from a new spd tight-binding method, Surf. Sci., 433, 751. 49 Womes, M., Cholley, T., Le Peltier, F., Morin, S., Didillon, B., and Szydlowski-Schildknecht, N. (2005) Study of the reaction mechanisms between Pt(acac)2 and alumina surface sites: Application to a new refilling technique for the con- trolled variation of the particle size of Pt/Al2O3 catalysts, Appl. Catal. A-Gen., In Press. 31 Guirado-Lopez, R., Desjonqueres, M.C., Spanjaard, D. and Aguilera-Granja, F. (1998) Electronic and geometrical effects on the magnetism of small RuN clusters, J. Magn. Magn. Mat., 186, 214. 50 Jacobs, G., Ghadiali, F., Pisanu, A., Borgna, A., Alvarez, W.E. and Resasco, D.E (1999) Characterization of the morphology of Pt clusters incorporated in a KL zeolite by vapor phase and incipient wetness impregnation. Influence of Pt particle morphology on aromatization activity and deacti- vation, Appl. Catal. A-Gen., 188, 79. 32 Mottet, C., Tréglia, G. and Legrand, B. (1996) Electronic structure of Pd cluster in the tight-binding approximation: influence of spd hybridization, Surf. Sci., 352, 675. 33 Khoutami, A. (1993) PhD, University of Paris XI. 51 Frenkel, A.I., Hills, C.W., and Nuzzo, R.G. (2001) A view from the inside: Complexity in the Atomic Scale Ordering of Supported Metal Nanoparticles, J. Phys. Chem. B., 105, 12689. 34 Car, R. and Parrinello, M. (1985) Unified Approach for Molecular Dynamics and Density-Functional Theory, Phys. Rev. Lett., 55, 2471. 52 Chao, K.J., and We, A.C. (2001) Characterization of hetero- geneous catalysts by X-ray absorption spectroscopy, J. Electron Spectroscopy and Related Phenomena, 119, 175. 35 Hall, B.D., Flueli, M., Monot, R. and Borel, J.P. (1991) Multiply twinned structures in unsupported ultrafine silver particles observed by electron diffraction, Phys. Rev. B, 43, 3906. 53 Stakheev, A.Y. and Kustov, L.M. (1999) Effects of the sup- port on the morphology and electronic properties of sup- ported metal clusters: modern concepts and progress in 1990s, Appl. Catal. A-Gen., 188, 3. 36 Pinto, A., Pennisi, A.R., Faraci, G., D’agostino, G., Mobilio, S. and Boscherini, F. (1995) Evidence for truncated octahe- dral structures in supported gold clusters, Phys. Rev. B, 51, 5315. 54 Bazin, D., Sayers, D., Rehr, J. (1997) Comparison between Xas, Awaxs, Asaxs & Dafs applied to nanometer scale metal- lic clusters, J. Phys. Chem., 101, 11040. 37 Apai, G., Hamilton, J.F., Stohr, J. and Thompson, A. REFERENCES (2002) Carbon tetrachloride hydrodechlori- nation with organometallics-based Pt and Pd catalysts on MgO, J. Mol. Catal. A-Chem., 182-183, 157. 45 Bazin, D., Rehr, J.J. (2003) Limits and advantages of X-ray absorption near edge structure for nanometer scale metallic clusters, J. Phys. Chem. B 107, 12398. 62 Benfield, R.E., Grandjean, D., Dore, J.C., Esfahanian, H., Wu, Z., Kröll, M., Geerkens, M. and Schmid, G. (2004) Structure of assemblies of metal nanowires in mesoporous alumina membranes studied by EXAFS, XANES, X-ray diffraction and SAXS, Faraday Discuss., 125, 327. 46 Bazin, D., Lynch, J., Ramos-Fernandez, M. (2003) Xas and Awaxs: Two basic tools in heterogeneous catalysis, Oil Gas Sci. Technol., 58, 683. 47 Clausen, B.S., Grabaek, L., Topsoe, H., Hansen, L.B., Stoltze, P., Norskov, J.K. and Nielsen, O.H. (1993) A new procedure for particle size determination by EXAFS based on molecular dynamics simulations, J. Catal., 141, 368. 63 Yoshitake, H. and Iwasawa, Y. (1992) Electronic metal sup- port interaction in platinum catalysts under deuterium-ethene reaction conditions and the microscopic nature of the active sites, J. Phys. Chem. B, 96, 1329. Oil & Gas Science and Technology – Rev. IFP, Vol. 61 (2006), No. 5 688 80 Wögerbauer, C., Maciejewski, M. and Baiker, A. (2002) Structure Sensitivity of NO Reduction over Iridium Catalysts in HC–SCR, J. Catal., 205, 157. 64 Bazin, D., Dexpert, H., Guyot-Sionnest, N.S., Bournonville, J.P. and Lynch, J. (1989) Exafs characterization of reforming catalysts: examples of recent applications, J. Chim. Phys., 7, 86. 81 Haq, S., Carew, A. and Raval, R. (2004) Nitric oxide reduc- tion by Cu nanoclusters supported on thin Al2O3 films, J. Catal., 221, 204. 65 Asakura, K., Chun, W.J., Shirarai, M., Tomishige, K. and Iwasawa, Y. (1997) In-situ polarization-dependent total- reflection fluorescence XAFS studies on the structure Transformation of Pt Clusters on Al2O3(0001), J. Phys. Chem. B, 101, 5549. 82 Ramsier, R.D., Gao, Q., Neergaard Waltenburg, H., Lee, K.W., Nooij, O.W., Lefferts, L. and Yates, J.T. (1994) NO adsorption and thermal behavior on Pd surfaces. A detailed comparative study, Surf. Sci., 320, 209. 66 Yamauchi, R., Gunji, I., Endou, A., Yin, X., Kubo, M., Chatterjee, A. and Miyamoto, A. (1998) Electronic and struc- tural features of Pd3 cluster on MgO(100) surface cluster, Appl. Surf. Sci., 130, 572. 83 Sugai, S., Watanabe, H., Kioka, T., Miki, H. and Kawasaki, K. (1991) Chemisorption of NO on Pd(100), (111) and (110) surfaces studied by AES, UPS and XPS, Surf. REFERENCES Sci., 259, 109. 67 Lopez, N., Illas, F. and Pacchioni, G. (1999) Electronic effects in the activation of supported metal clusters: Density functional theory. study of H2 dissociation on Cu/SiO2, J. Phys. Chem. B, 103, 1712. 84 Sharpe, R.G. and Bowker, M. (1996) The adsorption and decomposition of NO on Pd(110), Surf. Sci., 360, 21. 85 Nakamura, I., Fujitani, T. and Hamada, H. (2002) Adsorption and decomposition of NO on Pd surfaces, Surf. Sci., 514, 409. 68 Kantorovich, L., Shluger, A., Günster, J., Stultz, J., Krischok, S., Goodman, D.W., Stracke, P., and Kempter, V. (1999) Mg clusters on MgO surfaces: characterization by MIES and electronic structure ab initio calculations, Nucl. Instrum. Meth. B, 157, 162. 86 Højrup Hansen, K., Sljivananin, Z., Laesgsgaard, E., Besenbacher, F. and Stensgaardet, I. (2002) Adsorption of O2 and NO on Pd nanocrystals supported on Al2O3/NiAl(110): overlayer and edge structures, Surf. Sci., 505, 25. 87 Garcia-Cortès, J.M., Pérez-Ramirez, J., Rouzaud, J.N., Vaccaro, A.R., Illàn-Gémez, M.J. and Salinas-Martinez de Lecea, C. (2003) On the structure sensitivity of deNOx HC- SCR over Pt-beta catalysts, J. Catal., 218, 111. 69 Montano, P.A., Schulzte, W., Tesche, B., Shenoy, G.K. and Morrison T.I. (1984) Exafs study of Ag particles isolated in solid argon, Phys. Rev., 30, 672. 70 Wang, Z.L., Petroski, J.L., Green, T.C. and El-Sayed, M.A. (1998) Shape transformation and surface melting of cubic and tetrahedral Platinum nanocrystals, J. Phys. Chem. B, 32, 6145. 88 Barbier, J., Chollier, M.J. and Epron, F. (1997) In “Catalysis by Metals” Renouprez, A.J. and Jobic, H. (Eds.), EDP Sciences-Springer. 89 Barbier, J. (1992) Redox reactions in the tailoring of bimetal- lic catalysts in Advances in Catalyst Preparation, Catalytica Studies Division, Mountain View, California. 71 Vaarkamp, M., Miller, J.T., Modica, F.S. and Koningsberger, D.C. (1996) On the relation between particle morphology, structure of the metal-support Interface, and catalytic proper- ties of Pt/-Al2O3, J. Catal. 163, 294. 90 Derosa, P.A., Seminario, J.M. and Balbuena, P.B. (2001) Properties of Small Bimetallic Ni-Cu Clusters, J. Phys. Chem. A, 105, 7917. 72 Pandya, K.I., Heald, S.M., Hriljac, J.A., Petrakis, L. and Fraissard, J. (1996) Characterization by EXAFS, NMR, and other techniques of Pt/NaY Zeolite at industrially relevant low concentration of Platinum, J. Phys. Chem. B, 100, 5070. 91 Tréglia, G., Legrand, B., Ducastelle, F., Saúl, A., Gallis, C., Meunier, I., Mottet, C. and Senhaji, A. (1999) Alloy surfaces: segregation, reconstruction and phase transitions, Comp. Mater. Sci., 15, 196. Copyright © 2006 Institut français du pétrole Permission to make digital or hard copies of part or all of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for components of this work owned by others than IFP must be honored. Abstracting with credit is permitted. To copy otherwise, to republish, to post on servers, or to redistribute to lists, requires prior specific permission and/or a fee: Request permission from Documentation, Institut français du pétrole, fax. +33 1 47 52 70 78, or revueogst@ifp.fr. REFERENCES 73 Brown, W. and King, D.A. (2000) NO chemisorption and reactions on metal surfaces: a new perspective, J. Phys. Chem. B. 104, 2578. 92 Khoutami, A., Legrand, B., Mottet, C. and Tréglia, G. (1994) On the influence of topology on the energy profile in metallic Pd clusters, Surf. Sci., 307-309, 735. 74 Schneider, S., Bazin, D., Garin, F., Maire, G., Dexpert, H., Meunier, G., Noirot, R. and Capelle, M. (1999) NO reaction over nanometer scale platinum clusters deposited on γ-alu- mina: an XAS study, Appl. Catal., 189, 39. 93 Bazin, D., Mottet, C. and Tréglia, G. (2000) New opportuni- ties to understand heterogeneous catalysis processes through synchrotron radiation studies and theoretical calculations of density of states: The case of nanometer scale bimetallic par- ticles, Appl. Catal. A-Gen., 200, 47. 75 Loof, P., Stenbom, B., Norden, H. and Kasemo, B. (1993) Rapid Sintering in NO of Nanometer-Sized Pt Particles on small g-Al2O3 Observed by CO Temperature-Programmed Desorption and Transmission Electron Microscopy, J. Catal., 44, 60. 94 Garcia-Gutierrez, D.I., Gutierrez-Wing, C.E., Giovanetti, L.E., Ramallo-Lopez, J.M., Requejo, F.G. and Jose- Yacaman, M. (2005) Temperature Effect on the Synthesis of Au-Pt Bimetallic Nanoparticles, J. Phys. Chem. B, 109, 3813. 76 Wang, X., Sigmon, S.M., Spivey, J.J. and Lamb, H.H. (2004) Support and particle size effects on direct NO decomposition over platinum, Catal. Today, 96, 11. 95 Nutt, M.O., Hughes, J.B. and Wong, M.S. (2005) Designing Pd on Au bimetallic nanoparticles catalysts for trichloroethene hydrodechlorination, Environ. Sci. Technol., 39, 1346-1353. 77 Hashimoto, T., Hayashi, H., Udagawa, Y. and Ueno, A. (1995) NO induecd morphology changes by Xafs study, Physica B, 208/209, 683. 96 Borgna, A., Anderson, B.G., Saib, A.M., Bluhm, H., Havecker, M., Knop-Gericke, A., Kuiper, A.E.T., Tamminga, Y. and Niemantsverdriet, J.W. (2004) Pt-Co/SiO2 Bimetallic Planar Model Catalysts for Selective Hydrogenation of Crotonaldehyde, J. Phys. Chem. B, 108, 7905. 78 Campbell, T., Dent, A.J., Diaz-Moreno, S., Evans, J., Fiddy, S.G., Newton, M.A. and Turin, S. (2002) Susceptibility of a heterogeneous catalyst, Rh/Al2O3, to rapid structural change by exposure to NO, Chem. Commum., 30, 304-305. 97 Renouprez, A., Faudon, J.F., Massardier, J., Rousset, J.L., Delichère, L. and Bergeret, G. (1997) Properties of supported PdNi catalysts prepared by coexchange and organometallic chemistry, J. Catal., 170, 181. 79 Krause, K.R., and Schmidt, L.D. (1993) Microstructural changes and volatilization of Rh and Rh/Ce on SiO2 and Al2O3 in NO + CO, J. Catal., 140, 424. D Bazin et al. Final manucript received in January 2006 REFERENCES / Combining Solid State Physics Concepts and X-Ray Absorption Spectroscopy to Understand DeNOx Catalysis 689 101 Deutsch, S.E., Miller, J.T., Tomoshige, K., Iwasawa, Y., Weber, W.A. and Gates, B.C. (1996) Supported Ir and Pt Clusters: Reactivity with oxygen investigated by extended X- ray absorption fine structure spectroscopy, J. Phys. Chem. B, 100, 13408. 98 Zhu, L., Liang, K.S., Zhang, B., Bradley, J.S. and DePristo, A.E. (1997) Bimetallic PdCu catalysts: X-ray diffraction and theoretical modeling studies, J. Catal., 167, 412. 99 Shen, J., Hill, J.H., Watve, R.M., Spiewak, B.E. and Dumesic, J.A. (1999) Microcalorimetric, infrared spectroscopic, and DFT studies of ethylene adsorption on Pt/SiO2 and Pt- Sn/SiO2 catalysts, J. Phys. Chem. B, 103, 3923. 102 Reifsnyder, S.N., Otten, M.M., Sayers, D.E. and Lamb, H. (1997) Hydrogen chemisorption on Silica-supported Pt clus- ters: In situ X-ray absorption spectroscopy, J. Phys. Chem., 101, 4972. 100 Hill, J.M., Shen, J., Watwe, R.M. and Dumesic, J.A. (2000) Microcalorimetric, infrared spectroscopic, and DFT studies of ethylene adsorption on Pd and Pd/Sn, Catal. Langmuir, 16, 2213. Final manucript received in January 2006
https://openalex.org/W2109151134	https://europepmc.org/articles/pmc3831257?pdf=render	English	null	Cardiopulmonary bypass has a modest association with cancer progression: a retrospective cohort study	BMC cancer	2,013	cc-by	7,295	RESEARCH ARTICLE Open Access Pinto et al. BMC Cancer 2013, 13:519 http://www.biomedcentral.com/1471-2407/13/519 © 2013 Pinto et al.; licensee BioMed Central Ltd. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Abstract Background: Given their frequency of occurrence in the United States, cancer and heart disease often coexist. For patients requiring open-heart surgery, this raises concern that the use of cardiopulmonary bypass (CPB) may cause a transient immunosuppression with the potential to promote the spread and growth of coexisting cancer cells. This study examined the association of cardiopulmonary bypass with cancer progression in a large population- based setting using linked data from several state-wide registries. Methods: A retrospective cohort study of cancer risk, stage, and mortality in 43,347 patients who underwent coronary artery bypass graft (CABG) surgery with and without CPB in New Jersey between 1998–2004 was conducted. A competing risk analogue of the Cox proportional hazards model with propensity score adjustment and regression on the cause-specific hazard was used to compute relative risk ratios (95% confidence intervals [CIs]) for patients undergoing CABG surgery with and without CPB. Results: An increased risk for overall cancer incidence (17%) and cancer-specific mortality (16% overall, 12% case fatality) was observed; yet these results did not reach statistical significance. Of 11 tumor-specific analyses, an increased risk of skin melanoma (1.66 [95% CI, 1.08-2.55: p=0.02]) and lung cancer (1.36 [95% CI, 1.02-1.81: p=0.03]) was observed for patients with pump versus off-pump open-heart surgery. No association was found with cancer stage. Conclusions: These results suggest that there may be a relationship between CPB and cancer progression. However, if real, the effect is likely modest at most. Further research may still be warranted with particular focus on skin melanoma and lung cancer which had the strongest association with CPB. Keywords: Cardiopulmonary bypass, Cancer progression, Population-based cohort study IL-10, a major immunoregulatory cytokine with inhibitory effects of IFN-γ, tumor necrosis factor, IL-1, IL-6, and IL-8. CPB has also been shown to increase TGF-β, a cytokine with several potent immunosuppressive and immunomod- ulatory effects that may contribute to negative feedback regulation of T cell-mediated immune response [1,2]. It is therefore possible that such biochemical changes may lead to clinically relevant changes in immune system function and cancer surveillance with the potential to promote the spread and growth of co-existing cancer cells [1,2]. Al- though changes in immunoregulatory factors caused by CPB are short-lived and not likely to induce carcinogen- esis, it is plausible that CPB may be linked to cancer pro- gression. Further research is warranted, and may provide * Correspondence: cathy.pinto@merck.com 1Department of Epidemiology, Rutgers, Piscataway, NJ, USA Full list of author information is available at the end of the article © 2013 Pinto et al.; licensee BioMed Central Ltd. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Cardiopulmonary bypass has a modest association with cancer progression: a retrospective cohort study Cathy Anne Pinto1, Stephen Marcella1, David A August2, Bart Holland3, John B Kostis4 and Kit nto1, Stephen Marcella1, David A August2, Bart Holland3, John B Kostis4 and Kitaw Demissie1,5 Background Cardiac disease and cancer occur commonly in the United States, and it is therefore not infrequent that patients who undergo open heart surgery also develop cancer. Cancer therapy generally should be performed as soon as possible after diagnosis, except in cases where surgery may take priority for patients who are at high risk of suffering a more imminent major cardiac event. Open-heart surgery with cardiopulmonary bypass (CPB) is known to cause a transient immunosuppression, as evi- denced by increases in immunoregulatory factors including * Correspondence: cathy.pinto@merck.com 1Department of Epidemiology, Rutgers, Piscataway, NJ, USA Full list of author information is available at the end of the article Pinto et al. BMC Cancer 2013, 13:519 http://www.biomedcentral.com/1471-2407/13/519 Page 2 of 10 insight into the optimal strategy for management of cancer patients with cardiovascular co-morbidities. CABG surgery with incidental valve surgery were ex- cluded from the study as these patients have very different propensities for CPB compared to those with isolated CABG surgeries. Few studies have examined the association between CPB and cancer progression. A recent posthoc analysis of 611 patients by Vieira et. al. examined the association of cardiac and non-cardiac mortality with coronary artery bypass surgery with CPB and non-surgical inter- ventions (PCI, medical treatment) [3]. Compared with the non-surgical controls, CPB surgery was associated with a lower incidence of cardiac death (76.8% and 47.1%, respectively) and higher incidence of non-cardiac mortality, with a higher tendency toward cancer related death (7.2% and 20.6%, respectively). Two other recent studies examined the effects of CPB after cancer diagno- sis on all-cause and cancer-specific mortality rates [4,5]. The results demonstrated that all-cause and cancer- specific mortality after CPB increased with shorter time intervals between diagnosis and the surgical interven- tion, especially for those patients with less than 2 years between the cancer diagnosis and subsequent cardiac procedure (p<0.0001). Endpoints h The primary endpoint was cancer-specific mortality. Cancer-specific case fatality for those with cancers diag- nosed within 1-year, 2-years, and 4-years of surgery was also examined, as well as cardiovascular- and other cause-specific mortality. The main source of vital statis- tics for the study was MIDAS, which is linked annually with mortality data obtained from the National Center for Health Statistics. An additional sensitivity analysis of cancer-specific mortality was also performed using vital statistics data from the National Center for Health Sta- tistics Multiple Cause of Death File, which lists as many as 20 contributing causes of death in addition to the re- ported underlying cause of death [11]. Further research is warranted to enable a robust assess- ment of the association between CPB and cancer progres- sion. The aim of the present study is to evaluate cancer incidence, stage, and cancer-specific mortality in a large population-based cohort who previously underwent open heart surgery with and without CPB using linked data from several state-wide registries. Background For those with isolated CABG surgery, probabilistic record linkage was used to match records in several state- wide registries including 1) the Myocardial Data Acquisi- tion System (MIDAS), which includes hospital discharge records for patients with a myocardial infarction and other invasive cardiovascular procedures who have been admit- ted to New Jersey non-federal acute care hospitals [8], 2) the New Jersey Department of Health and Senior Services Open-Heart Surgery Registry (OHSR), which includes open-heart surgery data required to be reported by NJ car- diac surgery hospitals necessary to maintain licensing [9], and 3) the New Jersey State Cancer Registry (NJSCR), which is a population-based National Cancer Institute’s Surveillance, Epidemiology and End Results (SEER) Regis- try and North American Association of Central Cancer Registry (NAACCR) and collects data on all cancers diag- nosed and/or treated in New Jersey [10]. Institutional review board approvals for the study were obtained from the Rutgers and the Department of Health and Senior Services. NJSCR data were publically available for research purposes and access to the data was granted by NJ State Cancer Epidemiology Services. Access and permission to use the linked MIDAS and NJ Open Heart Surgery Registry data were provided by the Cardiovascular Institute at Robert Wood Johnson Medical School. After data linkage was complete, data were combined to form a comprehensive database for analysis, and all personal identifying information needed for linkage of the registries were deleted from the source records. Although data suggests an association between CPB and cancer progression, these results have not been con- sistently demonstrated. A study by Platell and colleagues in 33 patients with histologically proven colorectal showed a lower 5-year colon cancer-specific survival rate for those who underwent CPB surgery than for non-surgical con- trols: 34% and. 71%, respectively, p<0.05; HR=2.9 (95% CI: 1.5-4.4). However, a sensitivity analysis that excluded pa- tients with Stage IV cancer (1 CPB patient, 1 control) showed no significant difference in the cancer-specific sur- vival rate (p=0.1) [6]. Additionally, in a recent multicenter study of CPB and cancer progression performed by Suzuki et. al. in 74 patients with metastatic cancer who under- went open-heart surgery with and without CPB, no signifi- cant difference was observed in cancer-specific mortality (26.7% and 24.1%, respectively, p=0.8) [7]. Statistical methods the analysis due to missing perfusion data or death on the date of open-heart surgery. The majority of patients (>99.7%) included in the final analysis had a single open- heart surgery and the mean perfusion time for surgeries performed on pump was 86 minutes. Additional baseline characteristics are summarized in Table 1. Given the observational nature of the study, propensity scores were used as an adjustment in the model to help minimize bias related to differences in baseline risk fac- tors, as those with and without CPB may differ in import- ant prognostic factors related to outcome [12]. The propensity score model was developed using a stepwise lo- gistic regression with CPB as the outcome variable in the model. The model was developed with baseline character- istics reported at the time of the first open-heart surgery. Univariate modeling was performed to identify potential confounders and covariates with a significant association with the outcome of mortality [13,14]. The discriminatory power of the model was assessed using the area under the receiver operating characteristic curve (AUC), or C statis- tic; however, this model diagnostic was not used to guide variable selection into the propensity score model. Asym- metric restriction of the propensity score distribution (‘trimming’) was applied to improve overlap of the pro- pensity score distribution and improve baseline covariate balance between the groups. The group with pump procedures included a higher percentage of men, and those with a higher prevalence of coronary artery disease, prior CV intervention, number of diseased vessels, hypertension, diabetes, cardiogenic shock, and left main disease; whereas, patients with off- pump procedures were slightly older, included a greater proportion of Blacks and Hispanics, patients with Medi- care/Medicaid insurance and those with urgent CABG sur- gery, and patients with a higher prevalence of peripheral vascular disease, renal failure, and moderate/severe lung disease (Table 1). The majority of patients in each group were taking aspirin and/or beta blockers preoperatively. The final propensity score model included age, race, myocardial infarction, number of diseased vessels, prior intervention, congestive heart failure, hypertension, smoking, preoperative status (ie, urgency of the proced- ure), lung disease, peripheral vascular disease, cerebro- vascular disease, renal failure, ejection fraction, left main disease, endocarditis, and cardiogenic shock, each of which was significant in the model (p<0.01, AUC=0.679 [95% CI: 0.673, 0.685], p<.0001). Population and data source The study population included patients who underwent at least one open-heart coronary artery bypass surgery (CABG), with no incidental valve surgery, who had a hospital discharge date between January 1, 1998 and December 31, 2004 in the state of New Jersey. Patients had no prior cancer diagnosis with cancer records dating back to 1979. Patients with open-heart valve surgery or Other key secondary endpoints included cancer inci- dence and stage. Overall cancer incidence was examined and incidence for a subset of 11 commonly occurring can- cers, including those thought to be more susceptible to chronic immune modulation (i.e. skin melanoma, non- Hodgkin lymphomas, kidney cancer). Pinto et al. BMC Cancer 2013, 13:519 http://www.biomedcentral.com/1471-2407/13/519 Page 3 of 10 Pinto et al. BMC Cancer 2013, 13:519 http://www.biomedcentral.com/1471-2407/13/519 Pinto et al. BMC Cancer 2013, 13:519 http://www.biomedcentral.com/1471-2407/13/519 Cancer incidence and stage Of h 3 3 i i Of the 43,347 patients included in the final analysis, a total of 2,960 (6.8%) patients were diagnosed with a total of 3,182 primary cancers (Table 2). The majority of pa- tients diagnosed with cancer were male (79.2%), white (87.3%), and the median age at the time of cancer diag- nosis was 74 yrs. Of those diagnosed with cancer, the most commonly reported cancers types included prostate (25.2%), lung and bronchus (15.0%), colorectal (14.1%), and cancer of the urinary bladder (8.2%). Logistic and linear regression models, diagnostic test- ing, and sensitivity analyses were also employed as ap- propriate. Given the exploratory nature of this research, no multiplicity adjustments were performed. All analyses were performed in SAS V9.2 (SAS Institute Inc, Cary, NC, USA). The risk of developing cancer was proportionally greater in patients who underwent pump procedures compared with off-pump procedures (7.2% versus 5.8%, respectively), but the difference did not achieve statis- tical significance with or without adjustment for baseline risk factors (crude RR=1.06 [95% CI: 0.86-1.30, p=0.59], adjusted RR=1.17 [95% CI: 0.93-1.47, p=0.19]) (Table 2). These findings correspond to a cancer rate of 3.8 per 100,000 person-years of risk in the group with pump Statistical methods There was fairly good overlap in the resulting propensity score dis- tributions, except for the tail ends (Figure 1). The final dataset, after asymmetric trimming to improve overlap of the propensity score distributions and balance among baseline covariates, included a total of 43,347 or 90% of the original dataset. For cancer-specific mortality and cancer incidence, the primary analysis was performed using a competing risk analogue of the Cox Proportional Hazard models [15]. To account for the correlation among patients within a hos- pital or surgeon cluster, a robust covariance matrix was used to compute hazard ratios and 95% confidence inter- vals as measures of relative risk [16]. The zero time for the analysis was the date of the first surgery, and patients were followed until the date of the event (death, cancer inci- dence) or December 31, 2006 (date of censorship), which- ever came first. The proportional hazards assumption was tested using Schoenfeld residuals [17]. Each model was ad- justed, as applicable, for potential confounders including age, gender, race, cancer type, stage, time interval between surgery and cancer diagnosis, type of cancer treatment, use of blood products during surgery, year of surgery, and propensity score quantiles. For each analysis, crude and adjusted hazard ratios are presented. Differences in cancer stage was examined using a Cochran Armitage trend test. Sample and baseline covariates BMC Cancer 2013, 13:519 http://www.biomedcentral.com/1471-2407/13/519 Page 5 of 10 Table 1 Baseline characteristics for patients with open-heart surgery in New Jersey between 1998–2004, by CPB status (Continued) 30-39% 5,002 (14.0) 1,422 (11.7) 40-49% 8,703 (24.3) 2,641 (21.6) ≥50% 18,577 (51.9) 6,963 (56.8) aincludes patients with ≥1 isolated CABG surgeries; excludes patients with incidental valve surgery . bincludes patients with ≥1 on-pump procedure prior to incident cancer. Less than 0.2% of patients also had an off-pump procedure. celective: procedure deferred without increased risk of compromised cardiac outcome; urgent: not elective/emergent, required during same hospitalization to minimize chance of further clinical deterioration, worsening/sudden chest pain, congestive heart failure, acute myocardial infarction (AMI), anatomy, IABP, unstable angina with IV nitroglycerin or rest angina; emergent: ischemic dysfunction (ongoing ischemia including rest angina despite maximal medical therapy (medical and/or IABP), MI within 24 hours before surgery, or pulmonary edema requiring intubation), mechanical dysfunction (shock with or without circulatory support); emergent salvage: CPR enrout to the operating room or prior to anesthesia induction. Baseline characteristics were assessed at the time of the first open-heart surgery. %= n/N patients where n=frequency count in each category and N=total number of subjects. Due to rounding, sum of all percentages may equal 100%. CABG= coronary artery bypass graft, CPB=cardiopulmonary bypass (or “pump” procedure). Table 1 Baseline characteristics for patients with open-heart surgery in New Jersey between 1998–2004, by CPB status (Continued) 30 39% 5 002 (14 0) 1 422 (11 7) Table 1 Baseline characteristics for patients with open-heart surgery in New Jersey between 19 (Continued) 1 Baseline characteristics for patients with open-heart surgery in New Jersey between 1998–2004, nued) acteristics for patients with open-heart surgery in New Jersey between 1998–2004, by CPB status procedures and 3.5 per 100,000 person-years of risk in the group with off-pump procedures. y Of the 43,347 patients with isolated CABG surgery, the adjusted relative risk of cancer-specific mortality was 1.16 (95% CI: 0.92-1.46, p=0.20) (Table 3). Cardiovascu- lar and other-cause specific mortality was also increased for those who underwent pump versus off-pump surgery (1.15 (95% CI: 0.86-1.55, p=0.34) and 1.20 (95% CI: 0.65- 2.19, p=0.56), respectively. Of those patients with a cancer diagnosis after surgery, the adjusted risk of cancer-specific mortality was 1.12 (95% CI: 0.89-1.41, p=0.33) (Table 4). A Kaplan Meier plot illustrating the risk of cancer-specific mortality for those patients with cancers diagnosed after surgery is presented in Figure 2. Sample and baseline covariates A total of 48,009 patients who underwent isolated open- heart CABG surgery with (35,795) and without (12,214) CPB with extracorporeal circulation (or “pump” proce- dures) were included in the final analysis. Less than 2% of patients in the linked database were excluded from Pinto et al. BMC Cancer 2013, 13:519 http://www.biomedcentral.com/1471-2407/13/519 Page 4 of 10 Page 4 of 10 Table 1 Baseline characteristics for patients with open-heart surgery in New Jersey between 1998–2004, by CPB statu All CABG patientsa (n=48,009) Pumpb (n=35,795) Off pump (n=12,214) Length of hospital stay, mean (range) 11 (0–303) 10 (0–175) Males, n (%) 25,868 (72.7) 8,547 (70.0) Age, n (%) <50 3,231 (9.0) 1,096 (9.0) 50-59 7,439 (20.8) 2,351 (19.3) 60-69 11,304 (31.6) 3,637 (29.8) 70-79 10,999 (30.7) 3,743 (30.7) ≥80 2,822 (7.9) 1,387 (11.4) Race/ethnicity, n (%) White 30,030 (83.9) 9,744 (79.8) Black 1,927 (5.4) 931 (7.6) Hispanic 1,511 (4.2) 757 (6.2) Other/unknown 2,327 (6.5) 782 (6.4) Primary insurance, n (%) Medicare/medicaid 16,254 (45.4) 6,450 (52.8) Blue cross/commercial 6,021 (17.3) 2,070 (17.0) HMO 7,801 (21.8) 2,351 (19.3) Uninsured/indigent 924 (2.6) 366 (3.0) Self pay 676 (1.9) 205 (1.7) Other 3,939 (11.0) 772 (6.3) Preoperative statusc, n (%) Elective 13,428 (37.5) 4,297 (35.2) Urgent 20,369 (56.9) 7,412 (60.7) Emergent 1,793 (5.0) 403 (3.3) Salvage 78 (0.2) 13 (0.1) Myocardial infarction, n (%) 15,934 (44.5) 5,073 (41.5) Prior CV intervention, n (%) 8,987 (25.1) 2,823 (23.1) Diabetes, n (%) 13,069 (36.5) 4,009 (32.8) Hypertension, n (%) 27,153 (75.9) 8,779 (71.9) Congestive heart failure, n (%) 5,981 (16.7) 2,034 (16.7) Cardiogenic shock, n (%) 1,497 (4.2) 253 (2.1) Endocarditis, n (%) 77 (0.2) 19 (0.2) Cerebrovascular disease, n (%) 4,042 (11.3) 1,417 (11.6) Smoke, ever, n (%) 18,724 (52.3) 6,454 (52.8) Renal failure, n (%) 1,660 (4.6) 691 (5.7) Lung disease, n (%) Mild 3,427 (9.6) 1,078 (8.8) Moderate 753 (2.1) 294 (2.4) Severe 343 (1.0) 181 (1.5) Peripheral vascular disease, n (%) 5,359 (15.0) 2,019 (16.5) Disease vessels, n (%) One 958 (2.7) 759 (6.2) Two 5,721 (16.0) 2,010 (16.5) Three 23,407 (65.4) 7,699 (63.0) LM disease (>50% Occlusion), n (%) 10,055 (28.1) 3,208 (26.3) Ejection fraction (%), n (%) <20% 881 (2.5) 490 (4.0) 20-29% 2,494 (7.0) 725 (6.0) Table 1 Baseline characteristics for patients with open-heart surgery in New Jersey between 1998–2004, by CPB status All CABG patientsa (n=48,009) Pinto et al. Sample and baseline covariates No trend was observed for cancers diagnosed within shorter timeframes after sur- gery (Table 4). Of the 11 tumor-specific analyses performed, there was an increased risk of skin melanoma [RR=1.66 (95% CI, 1.08-2.55: p=0.02)], lung cancer [RR=1.36 (95% CI, 1.02-1.81: p=0.03)]), and a borderline in- crease in the relative risk of colorectal cancer (1.22 [95% CI: 0.98-1.53], p=0.08) after adjustment for base- line risk factors (Table 2). Similar findings were ob- served using standard Kaplan Meier methods (data not shown). For those cancers with a reported stage, there was no significant difference in the stage of cancers between the groups with and without pump exposure (p=0.65). Approximately 20% of reported cancers were metastatic at the time of diagnosis. Figure 1 Boxplot of propensity scores by exposure status and quantiles of propensity score distribution. Legend. Final analytical dataset including a total of 43,347 coronary artery bypass patients. 1 and 0 designates patients with pump and off-pump exposure status, respectively. Figure 1 Boxplot of propensity scores by exposure status and quantiles of propensity score distribution. Legend. Fin ncluding a total of 43,347 coronary artery bypass patients. 1 and 0 designates patients with pump and off-pump exposure s Figure 1 Boxplot of propensity scores by exposure status and quantiles of propensity score distribution. Legend. Final analytical dataset including a total of 43,347 coronary artery bypass patients. 1 and 0 designates patients with pump and off-pump exposure status, respectively. t of propensity scores by exposure status and quantiles of propensity score distribution. Legend. Final analytical dat of 43,347 coronary artery bypass patients. 1 and 0 designates patients with pump and off-pump exposure status, respectivel Pinto et al. Sample and baseline covariates BMC Cancer 2013, 13:519 http://www.biomedcentral.com/1471-2407/13/519 Page 6 of 10 Table 2 Relative risk of cancer, including any cancer and tumor-specific cancers CABG patientsa (n=43,347) Pumpb (n=33,357) Off Pump (n=9,900) Unadjusted Adjustedc n %d n %d RRe (95% CI) p-value RRe (95% CI) p-value Patients ≥1 primary cancerf 2,388 7.2% 572 5.8% Patients w/tumor-specific cancersg • Lung and bronchus 369 1.1% 75 0.8% 1.24 (0.94-1.62) 0.125 1.36 (1.02- 1.81) 0.034 • Prostate 609 1.8% 137 1.4% 1.13 (0.85-1.51) 0.401 1.25 (0.92-1.68) 0.149 • Pancreas 63 0.2% 18 0.2% 0.86 (0.57-1.32) 0.495 0.95 (0.62-1.47) 0.820 • Stomach 65 0.2% 23 0.2% 0.73 (0.43-1.22) 0.226 0.80 (0.48-1.35) 0.404 • Breast 100 0.3% 30 0.3% 0.84 (0.52-1.35) 0.467 0.92 (0.57-1.51) 0.752 • Colon/rectum 338 1.0% 79 0.8% 1.11 (0.90-1.38) 0.343 1.22 (0.98-1.53) 0.083 • Kidney/renal pelvis 93 0.3% 25 0.3% 0.95 (0.75-1.21) 0.676 1.05 (0.83- 1.32) 0.698 • Urinary bladder 197 0.6% 45 0.5% 1.07 (0.79-1.45) 0.672 1.18 (0.84-1.66) 0.339 • Corpus uterus 31 0.1% 3 0.0% * ***** * ******* ** • Non-hodgkin lymphoma 90 0.3% 23 0.2% 0.94 (0.65-1.35) 0.721 1.03 (0.71- 1.50) 0.874 • Skin melanoma 116 0.3% 19 0.2% 1.50 (0.99-2.27) 0.053 1.66 (1.08- 2.55) 0.022 afinal dataset after asymmetric trimming of the propensity score model to improve overlap of the propensity score distributions for patients with and without CPB CABG surgery. The final dataset includes 43,347 patients or 90% of the original dataset. Table 2 Relative risk of cancer, including any cancer and tumor-specific cancers CABG surgery. The final dataset includes 43,347 patients or 90% of the original dataset. p p p p p cmodel adjusted for age at time of surgery, gender, race, year of surgery, use of blood products, and propensity score. d d%=number of patients with tumor-specific cancer/ total number in group100. % number of patients with tumor specific cancer/ total number in group 100. erisk ratio for pump/off-pump modeled using Cox proportional hazards model with a robust covariance matrix that accounted for survival times for individuals within a hospital. For patients with multiple primaries, other cancers are censored at the time of occurrence. Competing risk model with regression of exposure on cause-specific hazard. p p g p erisk ratio for pump/off-pump modeled using Cox proportional hazards model with a robust covariance matrix that accounted for survival times for individuals within a hospital. Sample and baseline covariates BMC Cancer 2013, 13:519 http://www.biomedcentral.com/1471-2407/13/519 Page 7 of 10 Page 7 of 10 Table 4 Cancer-specific mortality for patients with diagnosis during follow-up, stratified by duration between surgery and diagnosis Cancer patientsa (n=2,960) Pumpb Off pump Kaplan Meierc Competing risk modelc (n=2,388) (n=572) n %d n %d HRe (95% CI) p-value HRe (95% CI) p-value Cancers diagnosed within 1 yr of surgery Number of patients with cancer 455 126 Patients who died from cancer 155 34.1% 41 32.5% • Unadjusted model 1.02 (0.72-1.43) 0.932 1.02 (0.72-1.43) 0.932 • Adjusted modelf 1.01 (0.72-1.40) 0.969 1.07 (0.82-.1.41) 0.615 • Sensitivity analysis-MCDg 173 38.0% 43 34.1% o Unadjusted model 1.08 (0.77-1.52) 0.665 1.08 (0.77-1.52) 0.665 o Adjusted modelf 1.11 (0.83-1.45) 0.514 1.19 (0.94-1.51) 0.148 Cancers diagnosed within 2 yrs of surgery Number of patients with cancer 868 250 Patients who died from cancer 270 31.3% 75 30.0% • Unadjusted model 1.00 (0.78-1.28) 0.987 1.00 (0.78-1.28) 0.987 • Adjusted modelf 1.05 (0.76-1.47) 0.762 107 (0.79-1.45) 0.656 • Sensitivity analysis-MCDg 299 34.4% 78 31.2% o Unadjusted model 1.07 (0.83-1.36) 0.611 1.07 (0.83-1.36) 0.612 o Adjusted modelf 1.10 (0.80-1.50) 0.558 1.15 (0.87-1.54) 0.327 Cancers diagnosed within 4 yrs of surgery Number of patients with cancer 1,635 448 Patients who died from cancer 484 29.6% 123 27.5% • Unadjusted model 1.00 (0.79-1.28) 0.980 1.00 (0.80-1.28) 0.980 • Adjusted modelf 1.02 (0.79-1.32) 0.868 109 (0.85-1.39) 0.491 • Sensitivity analysis-MCDg 529 32.4% 132 29.5% o Unadjusted model 1.02 (0.80-1.29) 0.874 1.02 (0.89-1.29) 0.874 o Adjusted modelf 1.05 (0.80-1.38) 0.678 1.10 (0.87-1.40) 0.417 Cancers diagnosed at any time during follow-up • Unadjusted model 668 28.0% 141 24.7% 0.96 (0.78-1.17) 0.665 0.96 (0.78-1.17) 0.665 • Adjusted modelf 1.06 (0.85-1.32) 0.595 1.12 (0.89-1.41) 0.330 • Sensitivity analysis-MCDg o Unadjusted model 730 30.6% 154 26.9% 0.96 (0.78-1.16) 0.648 0.96 (0.78-1.16) 0.648 o Adjusted modelf 1.04 (0.98-1.09) 0.211 1.11 (0.89-1.39) 0.366 afinal dataset after asymmetric trimming of the propensity score model to improve overlap of the propensity score distributions for patients with and without CPB CABG surgery. The final dataset includes 43,347 patients or 90% of the original dataset. bincludes patients with ≥1 on-pump procedure prior to incident cancer. cKaplan-Meier estimate treats failures from competing causes as censored observations; competing risk model with regression of exposure on cause-specific hazard. d%=number of patients who died / total number in each group100. Sample and baseline covariates For patients with multiple primaries, other cancers are censored at the time of occurrence. Competing risk model with regression of exposure on cause specific hazard fmultiple primary tumor-specific cancers (e.g. prostate, colorectal) may have been reported for any given patient. For the relative risk reported for patients with multiple primaries was used in the time to event analysis. gfor tumor-specific cancers, other cancers were censored at the time of diagnosis. results for cancers with cell count ≤5 were suppressed to as a way to ensure statistical reliability and protect patient confidentiality. CABG=coronary artery bypass graf, CPB=cardiopulmonary bypass (or “pump” procedure). Table 3 Cause-specific mortality for isolated open-heart surgery patients with no prior cancer diagnosis, by CPB status CABG patientsa (n=43,347) Pumpb Off pump Kaplan Meierc Competing risk modelc (n=33,357) (n=9,990) n %d n %d HRe (95% CI) p-value HRe (95% CI) p-value Cancer-specific mortality • Unadjusted model 923 2.8% 218 2.2% 1.02 (0.76-1.37) 0.898 1.02 (0.76-1.37) 0.898 • Adjusted modelf 1.13 (0.90-1.42) 0.306 1.16 (0.92-1.46) 0.203 Cardiovascular-specific mortality • Unadjusted model 2,979 8.9% 754 7.5% 1.01 (0.78-1.31) 0.938 1.01 (0.78-1.31) 0.938 • Adjusted modelf 1.26 (0.89-1.79) 0.186 1.15 (0.86-1.55) 0.344 Other cause-specific mortality • Unadjusted model 1,862 5.6% 491 4.9% 0.94 (0.73-1.22) 0.658 0.94 (0.73-1.22) 0.658 • Adjusted modelf 1.24 (0.67-2.29) 0.495 1.20 (0.65-2.19) 0.561 afinal dataset after asymmetric trimming of the propensity score model to improve overlap of the propensity score distributions for patients with and without CPB CABG surgery. The final dataset includes 43,347 patients or 90% of the original dataset. bincludes patients with ≥1 on-pump procedure prior to incident cancer. cKaplan-Meier estimate treats failures from competing causes as censored observations; competing risk model with regression of exposure on cause-specific hazard. d%=number of patients who died / total number in each group100. ehazard ratio for pump/off-pump modeled using Cox proportional hazards model with a robust covariance matrix that accounted for survival times for individuals within a hospital. Zero time for analysis was time of the open-heart surgery. fmodel adjusted for age at time of surgery, gender, race, year of surgery, use of blood products, and propensity score. CABG=coronary artery bypass graft, CPB=cardiopulmonary bypass (or “pump” procedure). 3 Cause-specific mortality for isolated open-heart surgery patients with no prior cancer diagnosis, CABG patientsa (n=43,347) mortality for isolated open-heart surgery patients with no prior cancer diagnosis, by CPB status CABG patientsa (n=43,347) Pinto et al. Sample and baseline covariates ehazard ratio for pump/off-pump modeled using Cox proportional hazards model with a robust covariance matrix that accounted for survival times for individuals within a hospital. Zero time for analysis was time of the open-heart surgery. fmodel adjusted age at time of initial cancer diagnosis, gender, race, cancer type, cancer stage, cancer treatment (e.g. chemotherapy, surgery), duration between surgery and cancer diagnosis, use of blood products, year of surgery, and propensity score. gthe analysis used cancer-specific mortality reported as primary cause of death or underlying cause of death using data from the NCHS multiple cause of death file. CABG=coronary artery bypass graft, “pump” procedure= CPB=cardiopulmonary bypass. Table 4 Cancer-specific mortality for patients with diagnosis during follow-up, stratified by duration between surgery and diagnosis y for patients with diagnosis during follow-up, stratified by duration between surgery Cancers diagnosed at any time during follow-up Pinto et al. BMC Cancer 2013, 13:519 http://www.biomedcentral.com/1471-2407/13/519 Pinto et al. BMC Cancer 2013, 13:519 http://www.biomedcentral.com/1471-24 Pinto et al. BMC Cancer 2013, 13:519 Page 8 of 10 http://www.biomedcentral.com/1471-2407/13/519 Figure 2 Kaplan Meier survival curve for cancer-specific mortality for patients with open-heart surgery and cancer diagnosed during follow-up, by CPB status Legend: Includes 2,960 cancer patients in study cohort. Figure 2 Kaplan Meier survival curve for cancer-specific mortality for patients with open-heart surgery and cancer diagnosed during follow-up, by CPB status Legend: Includes 2,960 cancer patients in study cohort. The lack of temporal association between surgery and cancer diagnosis was also observed in a separate study of CPB in patients with a prior cancer diagnosis (data unpublished). In this separate study, the relative risk for patients with pump versus off-pump procedures for cancers diagnosed within a year prior to surgery was 1.05 (95% CI, 0.85-1.30: p=0.67) compared to HR=1.10 (95% CI, 0.87-1.42: p=0.41) and HR=1.07 (95% CI, 0.84- 1.36: p=0.57) for those patients with cancers diagnosed 2 years and 4 years prior to surgery, respectively. The largest increase in cancer incidence, was observed for skin melanomas. This is consistent with the underlying hy- pothesis that immunosuppression and decreased immuno- surveillance are risk factors for melanoma. Prior studies have shown an increased risk of skin melanoma with chronic immune suppression [18]. The possible etiology of the observed association of CPB with lung cancer develop- ment is less clear. Sample and baseline covariates Although prior studies have shown an in- creased risk of lung cancer in patients with HIV and heart transplant recipients, it is unclear if these prior findings are related to immune suppression, medical surveillance bias, or an increase in behavioral risk factors such as smoking for which there may be residual confounding [19,20]. In the present study, there was no notable difference in the history of smoking between the pump versus off-pump groups. Although the results of each analysis presented herein demonstrates an increased mortality risk for those with CPB exposure, none of the results reached statistical signifi- cance. Similar findings were observed using standard Kaplan Meier methods and using the NJSCR multiple cause of death file (Tables 3, 4). For each analysis of cancer- specific mortality, the proportional hazards assumption was satisfied. Compared to open-heart surgery patients undergoing off- pump procedures, the results also show a non-statistically significant increase in the risk of cancer-specific mortality for patients who underwent cardiopulmonary bypass sur- gery prior to any diagnosis of cancer 1.16 (95% CI, 0.92- 1.46: p=0.20) as well as a non-statistically significant increase in the case-fatality rate for those cancer patients who underwent surgery with cardiopulmonary bypass prior to the cancer diagnosis (1.12: 95% CI, 0.89-1.41: p=0.33). No difference was observed in cancer stage at the time of diagnosis (p=0.65). The time between surgery and cancer diagnosis, which would serve as the most compelling evi- dence given its temporal association, did not show any im- pact on the relative risk of mortality. Even so, there were too few events to draw any definitive conclusions. Authors’ contributions d Authors contributions CAP, SM, DA, BH, and KD have each made substantial contributions to the study design, analysis, interpretation of the data, and drafting of the manuscript. JK assured the creation and maintenance of the longitudinal information of MIDAS, and made contributions to the study design. All authors have reviewed the manuscript critically for important intellectual content and given final approval of the version to be published. Authors’ informations l d d l CAP is a lead epidemiologist for a major pharmaceutical company, specializing in cardiovascular disease, diabetes, and urology, and is a member of the international society of pharmacoepidemiology (ISPE) and international society of pharmacoeconomics and outcomes research (ISPOR). SM is a clinical epidemiologist, a general internist and pediatrician with an interest in the metabolic syndrome and cancer. He currently is a consultant in global health outcomes research. DAA is the Chief of Surgical Oncology at The Rutgers Cancer Institute of New Jersey and Robert Wood Johnson Medical School. His clinical practice as a cancer surgeon led him to become interested in the role of immunosuppression in cancer pathogenesis. BH is a biostatistician with many years’ experience conducting epidemiological studies. He has a special interest in survival models and hazards regression. JK is the Director of the Cardiovascular Institute, Associate Dean for Cardiovascular Research at Rutgers-Robert Wood Johnson Medical, and Prin- cipal Investigator for MIDAS. KD is the Chair of Epidemiology at the Rutgers School of Public Health and a member of the Cancer Institute of New Jersey. He is a cancer epidemiologist with a research focus in understanding cancer progression. Although this research entailed one of the most robust assessments of cancer incidence and cancer-specific mor- tality in CPB patients to date in a population-based set- ting, there are design limitations. First, although robust probabilistic record linkage methods were used to link re- cords in the different state-wide databases, there is still the potential for selection bias and information bias result- ing from this process. As well, the generalizability of this study is limited given asymmetric trimming of ~10% of the original database to improve overlap in propensity score distributions. Additionally, even with the use of pro- pensity score adjustment for important risk factors, there is still the chance of imbalances between the groups in important unmeasured confounders and residual con- founding due to lack of detailed information collected for other important risk factors included in the propen- sity score modeling. Lastly, with no control for false dis- covery rates and multiple testing, spurious results are possible and, as such, the results should be interpreted with caution. Author details 1 1Department of Epidemiology, Rutgers, Piscataway, NJ, USA. 2Division of Surgery Oncology, Department of Surgery, The Rutgers Cancer Institute of New Jersey and Robert Wood Johnson Medical School, New Brunswick, NJ, USA. 3Department of Preventive Medicine, Rutgers New Jersey Medical School, Newark, NJ, USA. 4Cardiovascular Institute, Rutgers-Robert Wood Johnson Medical School for the MIDAS Study Group, New Brunswick, NJ, USA. 5The Rutgers Cancer Institute of New Jersey and Robert Wood Johnson Medical School, New Brunswick, NJ, USA. Received: 22 March 2013 Accepted: 9 October 2013 Published: 3 November 2013 Acknowledgments The authors would like to acknowledge the significant efforts of Pam Agovino, MPH from the Cancer Surveillance Unit of the New Jersey Department of Health & Senior Services for her assistance with probabilistic record linkage of MIDAS and the New Jersey State Cancer Registry. Additionally, the authors would like to acknowledge the efforts of those listed below for their general guidance on use of the state-wide databases in this study. Kevin Henry, PhD, Yingzi Deng, MD. Discussion This is the first population-based multicenter cohort study in patients with underlying cardiovascular disease and no pre-existing cancer diagnosis which used robust propensity score modeling to adjust for baseline imbalances and which accounted for competing risks. Results of this research show a statistically significant increase in the relative risk of skin melanoma (RR=1.66: 95% CI, 1.08-2.55: p=0.02), can- cer of the lung and bronchus (RR=1.36: 95% CI, 1.02-1.81: p=0.03), and an increase in overall cancer incidence in pa- tients who underwent isolated CABG surgery with cardio- pulmonary bypass compared to those patients undergoing off-pump surgery (RR=1.17: 95% CI, 0.93-1.47: p=0.19). The general strengths of this study includes its population-based cohort design, the use of medical Pinto et al. BMC Cancer 2013, 13:519 http://www.biomedcentral.com/1471-2407/13/519 Pinto et al. BMC Cancer 2013, 13:519 http://www.biomedcentral.com/1471-2407/13/519 Page 9 of 10 information recorded by clinicians rather than self- reports by patients or by proxy, the virtually complete ac- counting of open-heart surgeries given the need for hospitalization of these patients, and the mandatory reporting requirements for cancers in the State of New Jersey. The number of patients with procedures or reloca- tion outside of New Jersey is assumed to be minimal and non-differential for patients with and without cardiopul- monary bypass surgery. Other strengths of this study in- clude the completeness of vital status records captured using a variety of methods including, but not limited to, a review of state and national death files, state taxation files, hospital discharge files, Medicare and Medicaid files, and motor vehicle registration records, and the ability to per- form sensitivity analyses using the National Center for Health Statistics (NCHS) Multiple Cause of Death file that captures a cancer diagnosis reported as the primary or underlying cause of mortality on the death certificate. Additionally, robust propensity score modeling was used to help minimize potential confounding, and competing risk methods were used given that competing risks are of notable concern due to the age and prevalence of comor- bidities in this particular patient population. the spread and growth of pre-existing cancer cells with particular focus on skin melanoma and lung cancer which had the largest association in this study. Competing interests The authors declare that they have no competing interests. References S bl The data suggest there may be some degree of associ- ation between CPB and cancer progression. However, if real, the effect is likely to be modest at best. Although clinical practice guidelines will not likely change based on these findings, the results may assure clinicians that the choice of cardiopulmonary bypass should be deter- mined by other clinical considerations. Further research may still be warranted to assess whether the transient immunosuppression associated with CPB can promote 1. Sablotzki A, Welters I, Lehmann N, Menges T, Gorlach G, et al: Plasma levels of immunoinhibitory cytokine interleukin-10 and transforming growth factor-β in patients undergoing coronary artery bypass grafting. Eur J Cardio Thorac Surg 1997, 11(4):763–768. 2. Markewitz A, Lante W, Franke A, Marohl K, Kuhlmann WD, Weinhold C: Alterations of cell mediated immunity following cardiac operations: clinical implications and open questions. Shock 2001, 16(Suppl 1):10–15. 3. Viera RD, Pereira AC, Lima EG, Garzillo CL, Rezende PC, et al: Cancer-related deaths among different treatment options in chronic coronary artery disease: results of a 6-year follow-up of the MASS II study. Coronary Artery Diesase 2012, 23:79–84. 1. Sablotzki A, Welters I, Lehmann N, Menges T, Gorlach G, et al: Plasma levels of immunoinhibitory cytokine interleukin-10 and transforming growth factor-β in patients undergoing coronary artery bypass grafting. Eur J Cardio Thorac Surg 1997, 11(4):763–768. 2. Markewitz A, Lante W, Franke A, Marohl K, Kuhlmann WD, Weinhold C: Alterations of cell mediated immunity following cardiac operations: clinical implications and open questions. Shock 2001, 16(Suppl 1):10–15. 3. Viera RD, Pereira AC, Lima EG, Garzillo CL, Rezende PC, et al: Cancer-related deaths among different treatment options in chronic coronary artery disease: results of a 6-year follow-up of the MASS II study. Coronary Artery Diesase 2012, 23:79–84. Page 10 of 10 Page 10 of 10 Pinto et al. BMC Cancer 2013, 13:519 http://www.biomedcentral.com/1471-2407/13/519 Pinto et al. BMC Cancer 2013, 13:519 http://www.biomedcentral.com/1471-2407/13/519 4. Mistiaen W, Van Cauwelaert P, Muylaert P, Wuyts F, Harrisson F, Bortier H: Effect of prior malignancy on survival after cardiac surgery. Ann Thorac Surg 2004, 77:1593–1597. 5. Carrascal Y, Gualis J, Arevalo A, Fulquet E, Florez S, Rey J: Cardiac surgery with extracorporeal circulation in cancer patients: influence on surgical morbidity and mortality on survival. Rev Esp Cardiol 2008, 61(4):369–375. 6. Platell C: Influence of cardiopulmonary bypass surgery on cancer-specific survival rate of patients with colorectal cancer. Dis Colon Rectum 1998, 41:1371–1375. 7. Suzuki S, Usui A, Yoshida K, Matsuura A, Ichihara T: Effect of cardiopulmonary bypass on cancer prognosis. Asian Cardiovasc Ann 2010, 18:536–540. 7. Suzuki S, Usui A, Yoshida K, Matsuura A, Ichihara T: Effect of cardiopulmonary bypass on cancer prognosis. Asian Cardiovasc Thorac Ann 2010, 18:536–540. 7. Suzuki S, Usui A, Yoshida K, Matsuura A, Ichihara T: Effect of 8. The Cardiovascular Institute of New Jersey MIDAS - Myocardial Infarction Data Acquisition System (MIDAS). http://rwjms.umdnj.edu/cvinj/Research/MIDAS_1. html Accessed June 4, 2011. 9. State of New Jersey Department of Health and Senior Services Cardiac Surgery Registry. http://www.state.nj.us/health/healthcarequality/cardiacsurgery. shtml. Accessed June 4, 2011. 10. State of New Jersey Department of Health and Senior Services New Jersey State Cancer Registry (NJSCR); 2011. http://www.nj.gov/health/ces/. Accessed June 4, 2011. 11. Redelings MD, Wise M, Sorvillo FC: Using Multiple Cause-of-Death Data to Investigate Associations and Causality Between Conditions Listed on the Death Certificate. Am J Epidemiol 2007, 166:104–108. 12. Rosenbaum PR, Rubin DB: The central role of the propensity, score in observational studies for causal effects. Biometrika 1983, 79:516–524. 13. Brookhart MA, Schneeweiss S, Rothman KJ, Glynn RJ, Avorn J, et al: Variable selection for propensity score models. Am J Epidemiol 2006, 163:1149–1156. selection for propensity score models. Am J Epidemiol 2006, 163:1149 14. Westreich D, Cole SR, Funk MJ, Brookhart A, Stuermer T: The role of the c- statistic in the variable select for propensity score models. Pharmacoepidemiol Drug Saf 2011, 20:317–320. 15. Putter H, Fiocco M, Geskus RB: Tutorial in biostatistics: competing risks and multistate models. Stat Med 2007, 26:2389–2430. and multistate models. Stat Med 2007, 26:2389–2430. 16. Lin DY, Wei LJ: The robust inference for the proportional hazards model. Journal of the American Statistical Assocation 1989, 84:1074–1078. 16. Lin DY, Wei LJ: The robust inference for the proportion 16. Pinto et al. BMC Cancer 2013, 13:519 http://www.biomedcentral.com/1471-2407/13/519 Lin DY, Wei LJ: The robust inference for the proportional hazards model Journal of the American Statistical Assocation 1989, 84:1074–1078. 17. Harrell F, Lee K: Verifying assumptions of the proportional hazards model. Proceedings of the eleventh annual SAS users’group international; 1986:823–828. Journal of the American Statistical Assocation 1989, 84:1074–107 17. Harrell F, Lee K: Verifying assumptions of the proportional hazards model. Proceedings of the eleventh annual SAS users’group international; 1986:823–828. 18. Gerlini G, Romagnoli P, Pimpinelli N: Skin cancer and immunosuppression. Crit Rev Oncol Hematol 2005, 56:127–136. Elsevier Ireland Ltd. Available at: http://ac.els-cdn.com/S1040842805000910/1-s2.0-S1040842805000910-main. pdf?_tid=bc91744a-31d7-11e3-a216-00000aacb362&acdnat=1381429020_ 442946be3f049ec6081886577f16d347. Accessed on October 10, 2013. 19. Crespo-Leiro, Villa-Arranz A, Manito-Lorie N, Paniagua-Martin MJ, Rabago G, et al: Lung Cancer After Heart Transplantation. Results from a Large Mul- ticenter Registry. Am J Transplant 2011, 11:1035–1040. 20. Kirk GD, Merlo C, Driscoll PO, Mehta SH, Galai N, et al: HIV Infection is Associated with an Increased Risk for Lung Cancer, Independent of Smoking. Clin Infect Dis 2007, 45(1):103–110. doi:10.1186/1471-2407-13-519 Cite this article as: Pinto et al.: Cardiopulmonary bypass has a modest association with cancer progression: a retrospective cohort study. BMC Cancer 2013 13:519. Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color ﬁgure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color ﬁgure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color ﬁgure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit Submit your next manuscript to BioMed Central and take full advantage of: Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review
W4213034329.txt	https://zenodo.org/records/4898951/files/%D0%9F%D0%BE%D0%BD%D0%BE%D0%BC%D0%B0%D1%80%D0%B5%D0%BD%D0%BA%D0%BE.pdf	ru		КОНЦЕПТ «КАЧЕСТВО», КАК КЛЮЧЕВОЙ, В ДЕЯТЕЛЬНОСТИ ПРЕДПРИЯТИЙ ПРОДОВОЛЬСТВЕННОГО КОМПЛЕКСА: НАУЧНО-МЕТОДИЧЕСКИЙ ПОДХОД	Zenodo (CERN European Organization for Nuclear Research)	2,021	cc-by	1,911	УДК 330.342 DOI КОНЦЕПТ «КАЧЕСТВО», КАК КЛЮЧЕВОЙ, В ДЕЯТЕЛЬНОСТИ ПРЕДПРИЯТИЙ ПРОДОВОЛЬСТВЕННОГО КОМПЛЕКСА: НАУЧНО-МЕТОДИЧЕСКИЙ ПОДХОД Пономаренко Н. Ш. канд. экон. наук, доцент ГОУ ВПО «Донецкий национальный университет», Донецк, ДНР В статье представлена параметрическая модель обеспечения качества покупки товара или оказания услуги на основе потребительской ценности с использованием методического инструментария. Для четкости распределения предприятий продовольственного комплекса по качеству товаров или услуг применен матричный метод, который позволит обосновать позиционирование предприятий в зависимости от качества предоставляемых потребителю товаров или услуг. Ключевые слова: продовольственный комплекс, качество, товар, услуга, имидж, система жизнеобеспечения. THE CONCEPT OF "QUALITY" AS KEY IN THE ACTIVITIES OF FOOD COMPLEX ENTERPRISES: SCIENTIFIC AND METHODICAL APPROACH Ponomarenko N. Sh. Candidate of Economic Sciences, Associate Professor SEE HPE « Donetsk National University», Donetsk, DPR The article presents a parametric model for ensuring the quality of the purchase of a product or the provision of a service based on customer value using methodological tools. For the clarity of the distribution of food complex enterprises by the quality of goods or services, a matrix method has been applied, which will justify the positioning of enterprises depending on the quality of goods or services provided to the consumer. Keywords: food complex, quality, product, service, image, life support system. Постановка проблемы. Стабильность экономики и всего народного хозяйства государства во многом зависит от показателя качества. Особое значение концепт «качество» имеет в продовольственном комплексе, как приоритетной составляющей системы жизнеобеспечения. 200 Анализ исследований и публикаций. Проблемы обеспечения качества продукции и услуг, влияющих на качество жизни, содержательно изложены в работах Л.Е. Басовского, Б.И. Герасимова, М.И. Гельвановского, B.Ю. Огвоздина. Развитию теории управления качеством продукции уделено внимание в работах И.И. Антоновой, В.Я. Белобрагина, Д.С. Демиденко, А.В. Гличева, Н.В. Злобиной, В.А. Лапидуса, В.Д. Малыгиной, В.А. Качалова, С.Н. Кузьминой, Т.А. Салимовой, И.Г. Ушачева, Р.А. Фатхутдинова. Концептуальные основы категории «качество» были обоснованы и заложены также в работах зарубежных ученых Дж. Джурана, Ф. Кросби, Ж.-Ж. Ламбена, A. Фейгенбаума, К. Исикавы, Г. Тагути. Важным в потребительской составляющей качества при покупке товара или предоставленной услуге является имидж предприятия продовольственного комплекса – целостное его восприятие различными группами общественности, формирующееся на основе информации о различных сторонах деятельности предприятия, которая хранится в памяти людей [3]. Принимая во внимание отсутствие в научной литературе четкого элементного состава имиджа предприятия торговли и сферы услуг, можно предположить, что формирующими его ключевыми компонентами являются: визуальный имидж и бизнес-имидж предприятия, причем, в данной ситуации определенную лепту имеет информационное пространство и технологии [2]. Визуальный имидж предприятия формируется с учетом визуального восприятия зала, в котором оказывается услуга, стиль и оформление интерьера, атмосферы, внешнего вида персонала и фирменной символики. В свою очередь, бизнес-имидж основывается на восприятии предприятия торговли или сферы услуг как субъекта предпринимательской деятельности [1]. Одной из важных составляющих исследования концепта «качество» является анализ инвариантного аспекта. Цель статьи. Используя методический инструментарий оценки качества товара или услуги на основе потребительской ценности, обосновать положение концепта «качество», как ключевого, в деятельности предприятий продовольственного комплекса. Изложение основного материала. На примере качества предоставляемой конечному потребителю товара или услуги разработана параметрическая модель (рис. 1). 201 Потребительская составляющая Ценовая составляющая качество товара и услуги на основе потребительской ЦЕННОСТИ Потребительские ожидания информативность; социальное назначение; точность и своевременность исполнения; соответствие целевому назначению; безопасность и экологичность; степень соответствия качества товара и услуг запросам потребителей; стабильность; инновационность и др. Прейскурант средний чек; развитость системы ценовых скидок; уровень цен по прейскурантам Персонал профессионализм и мастерство; соблюдение норм и правил, соответсвующих торговле, определенной сфере услуг; время ожидания покупки, начала обслуживания; коммуникабельность; аккуратность; вежливость;. внимательность при обслуживании потребителя и др. Имидж *составлено автором визуальное восприятие зала, в котором приобретается товар, оказывается услуга; стиль и оформление интерьера; атмосфера; внешний вид персонала; фирменная символика и др. Рисунок 1 - Параметрическая модель обеспечения качества покупки товара или оказания услуги на основе потребительской ценности В основу предложенной параметрической модели обеспечения качества покупки или оказания услуги положены потребительская и ценовая составляющие. Следует подчеркнуть, что при выборе того или иного, например торгового предприятия продовольственного комплекса, потребитель особое внимание обращает на его имидж. Имидж способствует и обеспечивает уверенность потребителя в удовлетворении соответствующих потребностей на соответствующем уровне [1]. 202 Вследствие этого в представленную модель обеспечения качества покупки или оказания услуги включен такой компонент как категория «имидж», где учтено и визуальное восприятие зала, в котором приобретается товар или оказывается услуга; и стиль и оформление интерьера, и фирменная символика, и атмосфера, и даже внешний вид персонала. Говоря об имидже предприятия, участвующего в реализации товара или оказывающего услугу, нельзя не отметить, что при этом учитываются не только удовлетворение физиологических потребностей, но и потребности безопасности, полезности и принадлежности к соответствующей социальной среде. Этим объясняется стремление потребителей при наличии выбора удовлетворять потребности в определенном предприятии торговли или предприятии сферы услуг. Любая модель обеспечения качества покупки или оказания услуги, базирующаяся на основе потребительской ценности, предполагает потребительские ожидания, которые как минимум требований о покупке или услуге включают такие, как: информативность; социальное назначение; точность и своевременность исполнения; соответствие целевому назначению; безопасность и экологичность; степень соответствия качества товара и услуг запросам потребителей; стабильность; инновационность. Определение ценности в данном случае оправданно, поскольку цель оценки качества приобретаемого товара или полученной при этом услуги – именно потребительская характеристика ценности, которую предлагает предприятие своему потребителю, в частности, торговое предприятие, входящее в продовольственный комплекс. Концепт «качество» и качественные характеристики объекта (товара или услуги) требуют их количественного измерения. Для этого, как правило, переходят от качественной оценки к количественной путем использования механизма балльной оценки и обобщения информации, и на основе постепенного свертывания единичных и групповых показателей - к единому обобщающему показателю, отражающему уровень качества продукции или услуги по шкале: высокий, средний, низкий. Последовательность методического инструментария оценки качества товара или услуги на основе потребительской ценности предусматривает четыре этапа. Первый этап позволяет по системе единичных показателей осуществить оценку характеристик услуги, персонала, имиджа предприятия сферы услуг, а также определяется уровень цен на товары или услуги и осуществляется сравнение прейскурантов на соответствующий элемент с предприятиями-конкурентами. В данном 203 случае предлагается использование соответствующей карты наблюдений. В перечень показателей включены параметры как формальные (соответствие требованиям санитарных, технических норм и правил, действующих ГОСТов, ТУ и др.), так и неформальные (психофизиологические), в совокупности позволяющие определить уровень определенного элемента ценности, предлагаемой предприятием продовольственного комплекса. Данная оценка осуществляется путем анкетирования потребителей. Для осуществления количественной оценки предлагается соблюдение зависимости «лучшее качество товара или услуги, ценовое предложение  высший балл». Обработку первичных данных, полученных на основе наблюдения, предусмотрено осуществлять по соответствующей шкале оценки, представленной следующим образом: а) высокий уровень характеристики – 5 баллов; б) низкий уровень характеристики  1 балл; в) промежуточные характеристики – 2, 3, 4 балла. На втором этапе осуществляется определение индикаторов ценности по каждому из элементов оценки: «потребительской» и «ценовой» по конкретному предприятию. Обобщение полученных результатов осуществляется на основе подсчета среднеарифметической количества баллов по каждому из элементов оценки (Vzеj) с определением в конце среднего количества баллов в целом по потребительской (Vzp) и ценовой (Vzq) составляющими ценности товара или услуги также на основании среднеарифметической. n Vez j  B z i і 1 , n (1) где Vzеj – показатель ценности услуги j–го элемента z–го предприятия сферы услуг; Вi – количество баллов по единичной характеристике товара или услуги «і» z–го предприятия; n– количество единичных характеристик товара или услуги. Третий этап предполагает определение показателей ценности товара или услуги по «потребительской» и «ценовой» составляющими. m Vpz(q)  V j 1 m z ej , (2) Vzp(q) – показатель ценности услуги z-го торгового предприятия или сферы услуг по «потребительской» и «ценовой» характеристикам; m – количество квалиметрических параметров, рассматриваемых по определенному направлению оценки (по потребительской 204 составляющей  потребительские ожидания, персонал, имидж; по ценовой – прейскурант цен). На основании предельных значений обобщающих показателей «потребительской» и «ценовой» составляющих ценности товара или услуги (четвёртый этап), определяется уровень качества товара или услуги предприятия с учетом определения предельных значений по потребительской ( V ) и ценовой ( V ) характеристикам её ценности (табл. 1). гран p гран q гран p V V min р  Vqгран  Vqmin  Vрmax  Vрmin (3) 2 max Vq  Vqmin (4) 2 Таблица 1 Идентификация уровня качества товара или услуги Варианты № 1 комбинаций z гран  Vp  Vр  z гран  Vq  Vq Пространственное размещение в матрице Вывод І квадрант низкий уровень качества товара или услуги 2 z гран  Vp  Vр  z гран  Vq  Vq ІІ квадрант средний уровень качества товара или услуги 3 z гран  Vp  Vр  z гран  Vq  Vq ІІІ квадрант средний уровень качества товара или услуги 4 z гран  Vp  Vр  z гран  Vq  Vq IVквадрант высокий уровень качества товара или услуги На рис. 2 схематически представлена алгоритмическая модель оценки качества товара или услуги на основе потребительской ценности в оперативном режиме. Для четкости распределения предприятий продовольственного комплекса по качеству товаров или услуг применен матричный метод (рис. 3). В основе позиционирования – предельные значения качественных и ценовых характеристик товара или услуг предприятий продовольственного комплекса. Распределение характеристик на две группы по признакам «выше граничного значения» – «ниже граничного значения» позволяет выделить предприятия с высоким уровнем качества производимых и реализуемых товаров или предоставляемых услуг. 205 Начало Низкий уровень качества товара или услуги Высокий уровень качества услуги Высокий уровень качества услуги Конец Рисунок 2 – Алгоритмическая модель оценки качества товара или услуги на основе потребительской ценности в оперативном режиме Согласно матрице позиционирования предприятий в зависимости от качества предоставляемых потребителю товаров или услуг, для предприятий с высоким уровнем качества товаров или услуг характерны наиболее высокие значения квалиметрических и ценовых характеристик товара или услуги. К предприятиям с низким уровнем качества предоставляемых покупателю товаров или услуг – относятся, 206 по ценовой характеристике Показатель ценности товара или услуги соответственно, с низкими значениями качественных и ценовых характеристик. ІІ квадрант IV квадрант І квадрант ІІІ квадрант І квадрант Показатель ценности товара или услуги по потребительской характеристике Рисунок 3 – Макет матрицы позиционирования предприятий в зависимости от качества предоставляемых потребителю товаров или услуг Выводы по данному исследованию и направления дальнейших разработок по данной проблеме. Предложенный научно-методический подход к акцентированию внимания на роли качества товара или услуги предприятия продовольственного комплекса на основе их потребительской ценности, предусматривает расчет интегрального показателя качества предоставляемого потребителю товара или услуги по потребительским и ценовым составляющим. Показатель сформирован на основе комплексных оценок потребительских ожиданий, персонала и имиджа предприятия продовольственного комплекса, а также прейскуранта цен, что позволяет сконцентрировать усилия отдельных составляющих потребительской ценности товара или услуги для реализации основных функций предприятия продовольственного комплекса. Данный подход позволит также сформировать в конкурентной среде стратегические перспективы развития системы жизнеобеспечения с приоритетом качества товара или услуги в рамках продовольственного комплекса. 207 Обоснование концепта «качество», как ключевого в деятельности предприятий продовольственного комплекса, может стать основой для разработки Концепции константности качества продовольственного комплекса как приоритетной составляющей системы жизнеобеспечения, как фактора обеспечения качества жизни населения. Список использованных источников 1. Беляева, М.А, Самкова, В.А. Азы имиджелогии: имидж личности, организации, территории [Текст]: учебное пособие для вузов / М.А. Беляева, В.А. Самкова; Урал. гос. пед. ун-т. – Екатеринбург, 2016. – 184 с. 2. Тимиргалеева, Р.Р. Формирование корпоративного имиджа предприятия /Р.Р. Тимиргалеева //Научно-методический электронный журнал «Концепт». – Текст электронный.  2017. – №1. С.2-7. – URL: http://e-koncept.ru/2017/170004.htm (дата обращения: 10.02.2021). 3. Томилова, М.В. Модель имиджа организации / М.В. Томилова.  Текст: электронный // BizEducation online: [сайт]. – URL: http://www.bizeducation.ru (дата обращения: 16.02.2021). УДК 334.012.61.02(477.62) DOI СТРАТЕГИЧЕСКОЕ РАЗВИТИЕ МАЛОГО БИЗНЕСА В ДНР Рославцева Е. А. канд. экон. наук, доцент ГО ВПО «Донецкий национальный университет экономики и торговли имени Михаила Туган-Барановского», Донецк, ДНР Саенко В. Б. канд. гос. упр., доцент ГОУ ВПО «Донецкая академия управления и государственной службы при Главе Донецкой Народной Республики», Донецк, ДНР В статье разработана модель стратегического развития малого бизнеса в ДНР, сформулированы направления развития, выделены основные элемента стратегического развития и даны рекомендации для разработки долгосрочной стратегии развития малого предпринимательства в ДНР до 2030 года. Ключевые слова: стратегия; малый бизнес; государственные структуры, модель стратегического развития. 208
https://openalex.org/W2897006601	https://inria.hal.science/hal-02156461/file/474841_1_En_1_Chapter.pdf	English	null	Formalising BPMN Service Interaction Patterns	Lecture notes in business information processing	2,018	cc-by	10,253	To cite this version: Chiara Muzi, Luise Pufahl, Lorenzo Rossi, Mathias Weske, Francesco Tiezzi. Formalising BPMN Service Interaction Patterns. 11th IFIP Working Conference on The Practice of Enterprise Modeling (PoEM), Oct 2018, Vienna, Austria. pp.3-20, 10.1007/978-3-030-02302-7_1. hal-02156461 Formalising BPMN Service Interaction Patterns Chiara Muzi, Luise Pufahl, Lorenzo Rossi, Mathias Weske, Francesco Tiezzi To cite this version: Chiara Muzi, Luise Pufahl, Lorenzo Rossi, Mathias Weske, Francesco Tiezzi. Formalising BPMN Service Interaction Patterns. 11th IFIP Working Conference on The Practice of Enterprise Modeling (PoEM), Oct 2018, Vienna, Austria. pp.3-20, 10.1007/978-3-030-02302-7_1. hal-02156461 Formalising BPMN Service Interaction Patterns Chiara Muzi, Luise Pufahl, Lorenzo Rossi, Mathias Weske, Francesco Tiezzi Distributed under a Creative Commons Attribution 4.0 International License Formalising BPMN Service Interaction Patterns Chiara Muzi1, Luise Pufahl2, Lorenzo Rossi1, Mathias Weske2, and Francesco Tiezzi1 Chiara Muzi1, Luise Pufahl2, Lorenzo Rossi1, Mathias Weske2, and Francesco Tie 1 School of Science and Technology, University of Camerino, Italy {chiara.muzi,lorenzo.rossi,francesco.tiezzi}@unicam.it 2 Hasso-Plattner-Institute, University of Potsdam, Potsdam, Germany {luise.pufahl,mathias.weske}@hpi.de 1 School of Science and Technology, University of Camerino, Italy {chiara.muzi,lorenzo.rossi,francesco.tiezzi}@unicam.it 2 Hasso-Plattner-Institute, University of Potsdam, Potsdam, Germany {luise.pufahl,mathias.weske}@hpi.de Abstract. Business process management is especially challenging when cross- ing organisational boundaries. Inter-organisational business relationships are con- sidered as a ﬁrst-class citizen in BPMN collaboration diagrams, where multiple participants interact via messages. Nevertheless, proper carrying out of such in- teractions may be difﬁcult due to BPMN lack of formal semantics. In particular, no formal studies have been speciﬁcally done to cope with complex BPMN in- teraction scenarios uniﬁed under the name of Service Interaction Patterns. In this work the depiction of the service interaction patterns in BPMN collabora- tion diagrams is revisited and fully formalised via a direct semantics for BPMN multi-instance collaborations, thus leaving no room for ambiguity and validating the BPMN semantics. To make the formalisation more accessible, a visualisation of the patterns execution by means of a BPMN model animation tool is provided. Keywords: BPMN · Collaboration · Service Interaction Patterns · Formalisation HAL Id: hal-02156461 https://inria.hal.science/hal-02156461v1 Submitted on 14 Jun 2019 L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés. HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. Distributed under a Creative Commons Attribution 4.0 International License 1 Introduction The effective and efﬁcient handling of business processes is a primary goal of organ- isations. Business Process Management (BPM) provides methods and techniques to support these endeavors [17]. Thereby, the main artefacts are business process models which help to document, analyse, improve, and automate organisation processes [13]. To this aim, nowadays BPMN (Business Process Model and Notation) [16] is the mod- elling notation most widely applied in industry and academia. g y pp y For conducting a successful business, an organisation does not act alone, but it is usually involved in collaborations with other organisations. The importance of interac- tions has been underlined by many authors [1,10,2] and a lot of effort has been done to identify the most common interaction scenarios from a business perspective, which have been called Service Interaction Patterns [3]. Interactions are considered as a ﬁrst- class citizen in BPMN collaboration diagrams, where multiple participants cooperate by exchanging messages and sharing thereby data. This motivated the use of BPMN to model service interaction patterns [17], initially deﬁned only in terms of textual de- scriptions. This effort provided a graphical, more intuitive, description of the patterns and allowed to assess the suitability of BPMN to express common interaction scenar- ios. However, a severe issue in this study is that the precise behaviour of the BPMN models corresponding to some patterns may result unclear, in particular when multiple instances of the interacting participants are involved. This problem is mainly due to the fact that the BPMN standard comes without a formal semantics, which is needed in presence of tricky features, like multiple process instantiation. In this paper, we then aim at formalising the execution semantics of service interac- tion patterns speciﬁed in BPMN. This is a particularly important challenge in the BPM domain, as a precise semantics of the message exchanges as well as their dependencies is a prerequisite to ensure the appropriate carrying out, in practice, of such interactions. To achieve this goal, we resort to a formal semantics for BPMN collaborations includ- ing multiple instances introduced in [6]. The operational semantics is directly deﬁned on BPMN elements in terms of Labelled Transition Systems (LTSs), rather than as an encoding into other formalisms. Speciﬁcally, for each service interaction pattern we re- port the related BPMN collaboration model and provide its formalisation in terms of transitions of the corresponding LTS. 1 Introduction A direct formal characterisation is crucial, as it does not leave any room for am- biguity, and increases the potential for formal reasoning. This is especially important when dealing with multiple instances, whose static and compact BPMN representation hides their complex semantics. Moreover, the BPMN formalisation in [6] enables the use of the MIDA animation tool [7] that provides a visualisation, faithfully follow- ing the semantics, of patterns execution. This makes the behaviour of patterns easily understandable also to an audience non-familiar with formal methods. Finally, since the service interaction patterns have been used to evaluate different choreography lan- guages [5], their formalisation allows to validate the BPMN semantics itself, both in terms of the considered BPMN elements and of the expected semantic behaviour. In the remainder of this paper, we start with the introduction of a motivational ex- ample in Sec. 2, followed by an overview of the formalisation of BPMN collaboration diagrams in Sec. 3. In Sec. 4 we provide the representation of the service interaction patterns in BPMN and their formalisation. We present the patterns animation in Sec. 5. Finally, related work is discussed in Sec. 6 and the paper is concluded in Sec. 7. 2 Motivating Scenario Thus, the Retailer rather stops waiting as soon as enough responses have arrived or a given timeout is expired. This latter behaviour is rendered by the timer event attached to the sub-process, whose activation produces the execution of task “Reduce Invoice by Missing Items”. The Re- tailer then packs the items and passes the needed information to the Logistic Provider, who is in charge to send the package to the Customer. When the Customer receives the ordered items, via a receiving task, its process completes. ther to “Send Item” or not, by following one of the outgoing edges of the XOR split gateway, according to the information in the “Stock” data object. In case the item is not available, the Provider does not respond back to the Retailer. Thus, the Retailer rather stops waiting as soon as enough responses have arrived or a given timeout is expired. This latter behaviour is rendered by the timer event attached to the sub-process, whose activation produces the execution of task “Reduce Invoice by Missing Items”. The Re- tailer then packs the items and passes the needed information to the Logistic Provider, who is in charge to send the package to the Customer. When the Customer receives the ordered items, via a receiving task, its process completes. In this scenario, already various interaction patterns can be observed, such as One- to-many send/receive between the Retailer and the Item Provider(s), or the Request with Referral between the Customer, Retailer and Logistic Provider [3]. The service interac- tions represent different types of dynamic behaviour, ranging from simple message ex- changes, to scenarios involving multiple participants and multiple message exchanges, as well as routing behaviour, where information is routed to a new collaboration partner during an interaction (e.g. from the Retailer to the Logistic Provider). As the service interaction patterns are textually provided, a visualisation as well as a formalisation is crucial to precisely render the message exchanges between participants, especially be- cause multiple instances are involved. In particular, the interactions between the Retailer sub-process and the Item Provider may result quite intricate, as both generate multiple instances and, in addition, the sub-process is constrained by a timer event. Without a clear understanding of the interplay between these features, formally provided by the operational semantics, different interpretations may easily arise. 2 Motivating Scenario In this section we introduce an order fulﬁlment scenario to illustrate BPMN 2.0 col- laboration diagrams and the depiction of service interaction patterns. The considered scenario shows an interaction among a Customer, a Retailer, multiple Item Providers and a Logistic Provider (Fig. 1). The processes of the different interacting partners are represented inside rectangles, called pools, and their interaction is given by message edges (dashed connectors) visualising communication ﬂows. The order fulﬁlment process is started by the Customer who sends an “Order Re- quest” to the Retailer, via a send task. The arrival of this message starts, via a message start event, a process instance of the Retailer pool. This latter creates a list of needed items and stores this information in the “Item” data object collection. For each item, the Retailer sends out a request to an Item Provider and waits for the response. This interac- tion is rendered by a multi-instance sub-process communicating with the multi-instance pool of the Item Provider: each message of the Retailer creates a new process instance of the Item Provider pool. Each Provider checks for items availability and decides ei- 2 Customer Order Management Send Order Receive Pack Order Process Completed Retailer Request Received Create Item List Request Management Request Item Receive Item Request Completed Reduce Invoce by Missing Items Pack Items Give Order to Logistic Provider Request Process Completed Item Provider Item Requested Check Item Availability Send Item Item Request Completed Item Request Uncompleted Logistic Provider Logistic Order Received Send Pack Sending Process Completed Order Order Request Item Item Request Sent Items Package Logistic Order Stock Is item in stock? yes no Fig. 1: BPMN collaboration diagram of the Order Fulﬁlment Scenario. ther to “Send Item” or not, by following one of the outgoing edges of the XOR split gateway, according to the information in the “Stock” data object. In case the item is not available, the Provider does not respond back to the Retailer. Thus, the Retailer rather stops waiting as soon as enough responses have arrived or a given timeout is expired. This latter behaviour is rendered by the timer event attached to the sub-process, whose activation produces the execution of task “Reduce Invoice by Missing Items”. The Re- tailer then packs the items and passes the needed information to the Logistic Provider, who is in charge to send the package to the Customer. 2 Motivating Scenario When the Customer receives the ordered items, via a receiving task, its process completes. In this scenario, already various interaction patterns can be observed, such as One- Customer Order Management Send Order Receive Pack Order Process Completed Retailer Request Received Create Item List Request Management Request Item Receive Item Request Completed Reduce Invoce by Missing Items Pack Items Give Order to Logistic Provider Request Process Completed Item Provider Item Requested Check Item Availability Send Item Item Request Completed Item Request Uncompleted Logistic Provider Logistic Order Received Send Pack Sending Process Completed Order Order Request Item Item Request Sent Items Package Logistic Order Stock Is item in stock? yes no Fig. 1: BPMN collaboration diagram of the Order Fulﬁlment Scenario. Customer Order Management Send Order Receive Pack Order Process Completed Retailer Request Received Create Item List Request Management Request Item Receive Item Request Completed Reduce Invoce by Missing Items Pack Items Give Order to Logistic Provider Request Process Completed Item Provider Item Requested Check Item Availability Send Item Item Request Completed Item Request Uncompleted Logistic Provider Logistic Order Received Send Pack Sending Process Completed Order Order Request Item Item Request Sent Items Package Logistic Order Stock Is item in stock? yes no Fig. 1: BPMN collaboration diagram of the Order Fulﬁlment Scenario. Fig. 1: BPMN collaboration diagram of the Order Fulﬁlment Scenario. g g ther to “Send Item” or not, by following one of the outgoing edges of the XOR split gateway, according to the information in the “Stock” data object. In case the item is not available, the Provider does not respond back to the Retailer. Thus, the Retailer rather stops waiting as soon as enough responses have arrived or a given timeout is expired. This latter behaviour is rendered by the timer event attached to the sub-process, whose activation produces the execution of task “Reduce Invoice by Missing Items”. The Re- tailer then packs the items and passes the needed information to the Logistic Provider, who is in charge to send the package to the Customer. When the Customer receives the ordered items, via a receiving task, its process completes. ther to “Send Item” or not, by following one of the outgoing edges of the XOR split gateway, according to the information in the “Stock” data object. In case the item is not available, the Provider does not respond back to the Retailer. 3 Background Notions on the BPMN Formalisation In this section we provide an overview of the formal semantics of BPMN multi-instance collaborations given in [6]. The formalisation relies on a textual representation of the structure of BPMN collaboration models, deﬁned by the Backus-Naur Form (BNF) grammar in Fig. 2. In the proposed grammar, the non-terminal symbols C, P and A 3 C ::“ poolpp, Pq \| miPoolpp, Pq \| C1 ∥C2 P ::“ startpeenb, eoq \| startRcvpm:˜t, eoq \| endpeiq \| endSndpei, m: ˜ expq \| terminatepeiq \| andSplitpei, Eoq \| xorSplitpei, Gq \| andJoinpEi, eoq \| xorJoinpEi, eoq \| eventBasedpei, pm1 :˜t1, eo1q, . . . , pmh :˜th, eohqq \| taskpei, exp, A, eoq \| taskRcvpei, exp, A, m:˜t, eoq \| taskSndpei, exp, A, m: ˜ exp, eoq \| interRcvpei, m:˜t, eoq \| interSndpei, m: ˜ exp, eoq \| P1 ∥P2 A ::“ ϵ \| d.f ::“ exp, A Fig. 2: BNF syntax of BPMN collaboration structures. Fig. 2: BNF syntax of BPMN collaboration structures. represent Collaboration Structures, Process Structures and Data Assignments, respec- tively. The ﬁrst two syntactic categories directly refer to the corresponding notions in BPMN, while the latter refers to list of assignments used to specify updating of data objects. The terminal symbols, denoted by the sans serif font, are the typical elements of a BPMN model, i.e. pools, events, tasks and gateways. Intuitively, a BPMN collaboration model is rendered in this syntax as a collection of pools, each one specifying a process. Formally, a collaboration C is a composition, by means of the ∥operator, of pools. A pool is either of the form poolpp, Pq (for single- instance pools), or miPoolpp, Pq (for multi-instance pools) where p is the name that uniquely identiﬁes the pool, and P is the enclosed process. At process level, e P E uniquely denotes a sequence edge, while E P 2E is a set of edges. For the convenience of the reader, ei refers to the edge incoming in an element, while eo to the outgoing edge, and eenb to the (spurious) edge denoting the enabled status of a start event. To describe the semantics of collaboration models, we enrich the structural infor- mation with a notion of execution state, deﬁned by the state of each process instance and the store of the exchanged messages. We call process conﬁgurations and collabo- ration conﬁgurations these stateful descriptions. 3 Background Notions on the BPMN Formalisation 4 xtaskSndpei, exp1, A, m: ˜ exp, eoq, σ, αy !m :˜v ÝÝÝÑ xincpdecpσ, eiq, eoq, α1y σpeiq ą 0, evalpexp1, α, trueq, updpα, A, α1q, evalp ˜ exp, α, ˜vq pP-TaskSndq xinterRcvpei, m:˜t, eoq, σ, αy ?m : ˜et,ϵ ÝÝÝÝÝÑ xincpdecpσ, eiq, eoq, αy σpeiq ą 0, evalp˜t , α, ˜et q pP-InterRcvq xxorSplitpei, tpe, expquYGq, σ, αy ϵÝÑ xincpdecpσ, eiq, eq, αy σpeiq ą 0, evalpexp, α, trueq pP-XorSplit1q xP1, σ, αy ℓÝÑ xσ1, α1y xP1 ∥P2, σ, αy ℓÝÑ xσ1, α1y ℓ‰ kill pP-Int1q ιppq “ txσ, αyu xP, σ, αy !m :˜v ÝÝÝÑ xσ1, α1y pC-Deliverq xpoolpp, Pq, ι, δy !m :˜v ÝÝÝÑ xupdI pι, p, txσ1, α1yuq, addpδ, m, ˜vqy ιppq “ txσ, αyu xP, σ, αy ?m : ˜et,A ÝÝÝÝÝÑ xσ1, α1y ˜v P δpmq matchp ˜et, ˜vq “ A1 pC-Receiveq xpoolpp, Pq, ι, δy ?m :˜v ÝÝÝÑ xupdI pι, p, txσ1, updpα1, pA1, Aqqyuq, rmpδ, m, ˜vqy Fig. 3: An excerpt of BPMN semantic rules. Fig. 3: An excerpt of BPMN semantic rules. We refer the interested reader to [6] for a full account of the deﬁnition of these relations, while we report in Fig. 3, by way of example, some operational rules. Rule P-TaskSnd is used for the execution of send tasks possibly equipped with data objects. These latter are associated to a task by means of a conditional expression, exp1, and a list of assignments A, each of which assigns the value of an expression to a data ﬁeld (the ﬁeld f of the data object named d is accessed via d.f). Sending tasks also have as argument a pair of the form m : ˜ exp, where m is a message name and ˜ exp is a tuple of expressions. The task is activated only when there is a token in the incoming edge of the task (σpeiq ą 0) and the task’s guard exp1 is satisﬁed (evalpexp1, α, trueq). The effects of the task execution are as follows: the marking σ of the process instance is updated with the movement of one token from ei to eo, by means of functions dec and inc, and the message action !m : ˜v is produced, where the message content ˜v results from the evaluation of the expression tuple ˜ exp (evalp ˜ exp, α, ˜vq). The produced label is used to deliver the message at the collaboration layer (see rule C-Deliver). 3 Background Notions on the BPMN Formalisation Formally, a process conﬁguration has the form xP, σ, αy, where: P is a process structure; σ : E Ñ N is a sequence edge state function specifying, for each sequence edge, the current number of tokens marking it (N is the set of natural numbers); and α : F Ñ V is the data state function assigning values (possibly null) to data object ﬁelds (F is the set of data ﬁelds and V the set of values). A collaboration conﬁguration has the form xC, ι, δy, where: C is a collaboration struc- ture, ι : P Ñ 2SσˆSα is the instance state function mapping each pool name (P is the set of pool names) to a multiset of instance states (ranged over by I and containing pairs of the form xσ, αy), with Sσ and Sα the sets of edges and data states, and δ : M Ñ 2Vn is a message state function specifying for each message name m P M a multiset of value tuples representing the messages received along the message edge with the label m. The operational semantics is deﬁned by means of a labelled transition system (LTS), whose deﬁnition relies on an auxiliary LTS on the behaviour of processes. The latter is a triple xP, L, Ñy where: P ranged over by xP, σ, αy is a set of process conﬁgurations, L ranged over by ℓ, is a set of labels, and ÑĎ P ˆ L ˆ P is a transition relation. We will write xP, σ, αy ℓÝÑ xP, σ1, α1y to indicate that pxP, σ, αy, ℓ, xP, σ1, α1yq PÑ. Now, the labelled transition relation on collaboration conﬁgurations formalises the message exchange and the data update according to the process evolution. The LTS is a triple xC, Lc, Ñcy where: C, ranged over by xC, ι, δy, is a set of collaboration conﬁgurations; Lc, ranged over by l, is a set of labels; and ÑcĎ C ˆ Lc ˆ C is a transition relation. 3 Background Notions on the BPMN Formalisation Rule P-InterRcv is similar, but it produces a label corresponding to the reception of a message, which is actually consumed by rule C-Receive. Rule P-XorSplit1 is applied when a token is available in the incoming edge of a XOR split gateway and a conditional expression of one of its outgoing edges is evaluated to true; the rule decrements the token in the incoming edge and increments the token in the selected outgoing edge. Finally, rule P-Int1 deals with interleaving in a standard way for process elements. More details on these rules are given in the next section, from time to time when they are applied. 4.1 Send Pattern Informal Description. A party sends a message to another one. This pattern can be p Task 1 m e1 e2 q Fig. 4: Send pattern. modelled as the BPMN collaboration fragment in Fig. 4. Notably, this is only a way to model it: the send task could be replaced by an intermediate send event or by a message end event. However, up to some technicalities, all cases behave in the same way, thus we report here only one of them. Informal Description. A party sends a message to another one. This pattern can be p Task 1 m e1 e2 q Fig. 4: Send pattern. modelled as the BPMN collaboration fragment in Fig. 4. Notably, this is only a way to model it: the send task could be replaced by an intermediate send event or by a message end event. However, up to some technicalities, all cases behave in the same way, thus we report here only one of them. Fig. 4: Send pattern. q e1 e2 Textual Speciﬁcation. The collaboration fragment in Fig. 4 is repre- sented in the textual notation as C “ poolpp, Pq ∥poolpq, Qq with P “ taskSndpe1, exp1, ϵ, m: ˜ exp2, e2q ∥P 1, where p is the sender and q a generic re- ceiver (represented by a black-box pool in the graphical notation, whose process Q is left unspeciﬁed in the textual one). m p Formal Semantics. According to the form of process P, and the current state xσ, αy of pool p’s instance, the collaboration can evolve as follows: – Process P moves by executing Task 1. This execution step takes place by ap- plying rule P-TaskSnd, which requires the incoming edge e1 of the task be marked by at least one token (σpe1q ą 0), and the task’s guard exp1 be satisﬁed (evalpexp1, α, trueq). The effects of the task execution are as follows: the mes- sage action !m : ˜v is produced, where the message content ˜v results from the evaluation of the expression tuple ˜ exp2 (evalp ˜ exp2, α, ˜vq), and the marking σ of the process instance is updated with the movement of one token from e1 to e2, that is σ1 “ incpdecpσ, e1q, e2q. 4 Patterns Formalisation In this section we present and formalise the Service Interaction Patterns [3] supported by BPMN. Since BPMN is not speciﬁcally tailored to the needs of service interaction 5 patterns, the notation cannot completely support all their features. For instance, while the informal and general description of these patterns leaves it open if in an interaction the counter-party is known at design-time or not, in BPMN it is expected to have a priori knowledge of the interacting partners, i.e. the target pool of a message edge cannot be dynamically selected. On the other hand, in case a message is directed to a multi-instance pool, BPMN supports a form of runtime binding of the message with the correct process instance by means of the correlation mechanism [16, Sec. 8.3.2]. Moreover, it is also possible to dynamically specify other model features, such as the number of involved participants and exchanged messages. Each pattern is presented according to the following structure: Informal Description consists of a natural language description, and a graphical rep- resentation in terms of a BPMN collaboration fragment. Textual Speciﬁcation provides the textual notation of the BPMN collaboration model. extual Speciﬁcation provides the textual notation of the BPMN collaboration model. ormal Semantics describes the operational rules applied to perform each execution step, and shows the results in terms of the execution state functions evolution. p Formal Semantics describes the operational rules applied to perform each execution step, and shows the results in terms of the execution state functions evolution. p In the following we present ﬁrst those patterns concerning single transmissions, both bilateral (Sec. 4.1-4.3) and multilateral (Sec. 4.4-4.7), and then the routing patterns involving multiple transmissions (Sec. 4.8-4.9). 4.2 Receive Pattern Informal Description. A party receives a message from another party. This q e1 e2 m p Fig. 5: Receive pattern. pattern can be modelled as the BPMN collabora- tion fragment shown in Fig. 5. Also here the in- termediate receive event could be replaced, in this case by a receive task or by a receiving start event. Textual Speciﬁcation. The textual representation of the collaboration fragment in Fig. 5 has again the form C of the previous pattern, with Q “ interRcvpe1, m:˜t, e2q ∥Q1. Informal Description. A party receives a message from another party. This q e1 e2 m p Fig. 5: Receive pattern. pattern can be modelled as the BPMN collabora- tion fragment shown in Fig. 5. Also here the in- termediate receive event could be replaced, in this case by a receive task or by a receiving start event. Textual Speciﬁcation. The textual representation of the collaboration fragment in Fig. 5 has again the form C of the previous pattern, with Q “ interRcvpe1, m:˜t, e2q ∥Q1. Informal Description. A party receives a message from another party. This q e1 e2 m pattern can be modelled as the BPMN collabora- tion fragment shown in Fig. 5. Also here the in- termediate receive event could be replaced, in this case by a receive task or by a receiving start event. p Fig. 5: Receive p Textual Speciﬁcation. The textual representation of the collaboration fragment in Fig. 5 has again the form C of the previous pattern, with Q “ interRcvpe1, m:˜t, e2q ∥Q1. Formal Semantics. Assuming that the intermediate receive event is enabled by a token in e1, the process Q can perform a receiving action, that is the transition xinterRcvpe1, m:˜t, e2q, σ, αy ?m : ˜et,ϵ ÝÝÝÝÝÑ xincpdecpσ, e1q, e2q, αy is produced by apply- ing rule P-InterRcv. Then, the process Q evolves by means of the interleaving rule P-Int1. The produced label ?m: ˜et, ϵ indicates the willingness of process Q to con- sume a message of type m matching the template ˜et. If present, the message is actually consumed by rule C-Receive at collaboration level. 4.1 Send Pattern Therefore, the application of rule P-TaskSnd produces the transition xtaskSndpe1, exp1, ϵ, m: ˜ exp2, e2q, σ, αy !m :˜v ÝÝÝÑ xσ1, αy, where the data state α remains unchanged because no data object is connected 6 to Task 1. Hence, the overall process P can evolve according to the interleav- ing rule P-Int1, that is xP, σ, αy !m :˜v ÝÝÝÑ xσ1, αy. Similarly, by applying the rule C-Deliver, and then the interleaving rule at collaboration level, the execution step of the overall collaboration C is represented by the transition xC, ι, δy !m :˜v ÝÝÝÑ xupdI pι, p, txσ1, αyuq, addpδ, m, ˜vqy, with ιppq “ txσ, αyu. Its effects are: updating the marking in the p’s instance (updI pι, p, txσ1, αyuq), and updating the message state function (addpδ, m, ˜vq) by adding a value tuple ˜v to the m’s message list, in order to be subsequently consumed by the receiving participant q. to Task 1. Hence, the overall process P can evolve according to the interleav- ing rule P-Int1, that is xP, σ, αy !m :˜v ÝÝÝÑ xσ1, αy. Similarly, by applying the rule C-Deliver, and then the interleaving rule at collaboration level, the execution step of the overall collaboration C is represented by the transition xC, ι, δy !m :˜v ÝÝÝÑ xupdI pι, p, txσ1, αyuq, addpδ, m, ˜vqy, with ιppq “ txσ, αyu. Its effects are: updating the marking in the p’s instance (updI pι, p, txσ1, αyuq), and updating the message state function (addpδ, m, ˜vq) by adding a value tuple ˜v to the m’s message list, in order to be subsequently consumed by the receiving participant q. step of the overall collaboration C is represented by the transition xC, ι, δy !m : v ÝÝÝÑ xupdI pι, p, txσ1, αyuq, addpδ, m, ˜vqy, with ιppq “ txσ, αyu. Its effects are: updating the marking in the p’s instance (updI pι, p, txσ1, αyuq), and updating the message state function (addpδ, m, ˜vq) by adding a value tuple ˜v to the m’s message list, in order to be subsequently consumed by the receiving participant q. q y y g p p – Process P moves by executing an (unspeciﬁed) activity of P 1. 4.2 Receive Pattern Indeed, this rule requires that there is a message in the m’s message queue (˜v P δpmq) that matches the template ˜et of the receiving event (matchp ˜et, ˜vq “ A); the assignments A produced by this matching are then applied to the data state α of the q’s instance, and the message is removed from the queue (rmpδ, m, ˜vq). 4.1 Send Pattern Thus, we have a transition xP 1, σ, αy ℓÝÑ xσ1, α1y, from which P can evolve by means of the sym- metric rule of P-Int1, and the overall collaboration can then evolve accordingly. This execution step, anyway, is not relevant for the pattern semantics, and hence is not discussed in more detail. – Process Q moves by executing an (unspeciﬁed) activity. Again this execution step is not relevant for the pattern semantics. p Task 1 e1 e2 Relying on asynchronous communication, we are able to formalise an unreliable and non-guaranteed delivery. The sending action in fact just updates the message state function by adding a message, without requiring this to be received. m q 4.3 Send/Receive Pattern Informal Description. Two parties, p and q, engage in two causally related interactions. In the ﬁrst interaction, p sends a message (the request) to q, while in the second one p receives a message (the response) from q. 7 p Task 1 m Task 1 q Task 3 m1 m2 e1 e2 e2 e1 e3 e4 e5 e6 1 2 q Fig. 6: Send/Receive pattern. This pattern can be modelled by combining the Send and the Receive patterns, as shown in Fig. 6. Textual Speciﬁcation. The textual representation of the collaboration fragment in Fig. 6 has again the form C of the previous patterns, with P “ taskSndpe1, exp1, ϵ, m1: ˜ exp2, e2q ∥interRcvpe2, m2: ˜t1, e3q ∥P 1 Q “ interRcvpe4, m1: ˜t2, e5q ∥taskSndpe5, exp3, ϵ, m2: ˜ exp4, e6q ∥Q1 Task 1 e1 e2 This pattern can be modelled by combining the Send and the Receive patterns, as shown in Fig. 6. m q Textual Speciﬁcation. The textual representation of the collaboration fragment in Fig. 6 has again the form C of the previous patterns, with P “ taskSndpe1, exp1, ϵ, m1: ˜ exp2, e2q ∥interRcvpe2, m2: ˜t1, e3q ∥P 1 Q “ interRcvpe4, m1: ˜t2, e5q ∥taskSndpe5, exp3, ϵ, m2: ˜ exp4, e6q ∥Q1 m p Formal Semantics. The execution steps of this pattern are realised by combining the semantic rules for the Send and Receive patterns. In detail: let us suppose that there is a token in the incoming edge of Task 1 (σpe1q ą 0) and the other pre- conditions of rule P-TaskSnd are satisﬁed; by applying this rule we have that xtaskSndpe1, exp1, ϵ, m1: ˜ exp2, e2q, σ, αy !m1 :˜v ÝÝÝÑ xincpdecpσ, e1q, e2q, αy. Then, P evolves by performing a sending action, by means of the interleaving rule P-Int1, that is xP, σ, αy !m1 :˜v ÝÝÝÑ xσ1, αy. At the collaboration layer, by applying rule C-Deliver, the message m1 is delivered to q. Now, on the receiving party, assuming that there is a to- ken on e4 and that the template ˜t2 evaluates to ˜ et2, by applying rules P-InterRcv and P-Int1, we have xQ, σ2, α2y ?m1 : ˜ et2,ϵ ÝÝÝÝÝÝÑ xincpdecpσ2, e4q, e5q, α2y. The observed label indicates the willingness to receive a message of type m1. 4.3 Send/Receive Pattern Thus, at collaboration level, rule C-Receive can be applied to allow process Q to actually consume the sent request message. Now, Task 3 is enabled and, by proceeding in a specular way, Q can send the response message m2 and P can consume it. m p Formal Semantics. The execution steps of this pattern are realised by combining the semantic rules for the Send and Receive patterns. In detail: let us suppose that there is a token in the incoming edge of Task 1 (σpe1q ą 0) and the other pre- conditions of rule P-TaskSnd are satisﬁed; by applying this rule we have that xtaskSndpe1, exp1, ϵ, m1: ˜ exp2, e2q, σ, αy !m1 :˜v ÝÝÝÑ xincpdecpσ, e1q, e2q, αy. Then, P evolves by performing a sending action, by means of the interleaving rule P-Int1, that is xP, σ, αy !m1 :˜v ÝÝÝÑ xσ1, αy. At the collaboration layer, by applying rule C-Deliver, the message m1 is delivered to q. Now, on the receiving party, assuming that there is a to- ken on e4 and that the template ˜t2 evaluates to ˜ et2, by applying rules P-InterRcv and 4.4 Racing Incoming Messages Pattern Informal Description. A party expects to receive one among a set of messages. These messages may be structurally different (i.e. different types) and may come from differ- ent categories of partners. The way a message is processed depends on its type and/or the category of partner from which it comes. This pattern can be modelled in BPMN by using in the receiving participant an event-based gateway connected to receiving events. Messages can be expected from one participant (Fig. 7) or they can arrive from different participants (Fig. 8). Textual Speciﬁcation. Let us ﬁrst consider the case in which messages arrive from one participant (Fig. 7). In the textual notation the diagram is rendered as the collaboration of the usual form C, with P “ xorSplitpe1, tpe2, exp1q, pe3, exp2quq ∥taskSndpe2, exp3, ϵ, m1: ˜ exp4, e4q ∥ taskSndpe3, exp5, ϵ, m2: ˜ exp6, e5q ∥P 1 Q B dp 6 p 1 ˜ 7q p 2 ˜ 8qq ∥Q1 Q “ eventBasedpe6, pm1:˜t1, e7q, pm2:˜t2, e8qq ∥Q1 Q “ eventBasedpe6, pm1:˜t1, e7q, pm2:˜t2, e8qq ∥Q1 Q “ eventBasedpe6, pm1:˜t1, e7q, pm2:˜t2, e8qq ∥Q1 The case in which messages arrive from two different participants (Fig. 8) is ren- dered in the textual notation as C “ poolpp, P 2q ∥poolpr, Rq ∥poolpq, Qq, where 8 P P p Task 1 Task 2 q e1 e2 e3 e4 e5 e6 e7 e8 m1 m2 A m e1 e2 P R Task A B e1 e2 e3 e4 A N e2 e1 e(n-1) m1 Q Fig. 7: Racing incoming messages (a). q e6 e7 e8 m1 m2 p r Fig. 8: Racing incoming messages (b). process Q is as the above one, while P 2 and R are left unspeciﬁed (because they are included in black-box pools). p Task 1 Task 2 q e1 e2 e3 e4 e5 e6 e7 e8 m1 m2 Fig. 7: Racing incoming messages (a). P A m e1 e2 Q q e6 e7 e8 m1 m2 p r Fig. 8: Racing incoming messages (b). Fig. 8: Racing incoming messages (b). Fig. 7: Racing incoming messages (a). P P R T k A B e1 e2 e3 e4 A N e2 e1 e(n-1) m1 process Q is as the above one, while P 2 and R are left unspeciﬁed (because they are included in black-box pools). Qk Q1 Q e5 e6 e7 m1 m2 e(k-1) ek m Formal Semantics. 4.4 Racing Incoming Messages Pattern Let us start with the case in which messages arrive from a single participant, and assume that a token is available in the incoming edge of the XOR split gateway of P (σpe1q ą 0) and the conditional expression exp1 is evaluated to true (evalpexp1, α, trueq). Thus, rule P-XorSplit1 can be applied and the token is moved to the edge e2, hence enabling Task 1. Formally, this step corresponds to the transition xP, σ, αy ϵÝÑ xincpdecpσ, e1q, e2q, αy, where label ϵ denotes the movement of the token internally to the process. The next step corresponds to the execution of Task 1, which is as in the case of the Send pattern. Once the message m1 has been sent (hence, there exists a ˜v such that ˜v P δpm1q), and assuming that there is a token in e6 (σpe6q ą 0), the event-based gateway can evolve by applying the corresponding rule. This corre- sponds to the transition xeventBasedpe6, pm1 :˜t1, e7q, pm2:˜t2, e8qq, σ1, α1y ?m1 : ˜ et1,ϵ ÝÝÝÝÝÝÑ xincpdecpσ1, e6q, e7q, α1y, with template ˜et1 matching the message ˜v. The rule moves the token from the incoming edge to the outgoing edge corresponding to the received message. The produced label enables the application of rule C-Receive at collaboration level, which takes care of consuming the message ˜v of type m1 in δ. The case where message m2 is selected to be sent is similar. In the scenario shown in Fig. 8, even if the transitions produced by the collaboration have the same labels, the pattern semantics is quite different. In fact, in the previous case the organisation p internally decides which message will be sent and only one message will be delivered and consumed, while in this case the organisations p and r act independently from each other and it may occur that both m1 and m2 are sent to q. In such a case, one of the two messages will be consumed, depending on their arrival time, and the other message will be pending forever. 4.5 One-To-Many Send Pattern Informal Description. A party sends messages to several parties. All messages have the same type (although their contents may be different). The number of parties to whom the message is sent may or may not be known at design time. In BPMN, this pattern can be modelled as the collaboration fragment in Fig. 9, where each party is represented as an instance of a multi-instance pool and a message is sent to each process instance via a sequential multi-instance send task. From now on, when a message is sent/received to/by several parties, we will model these parties as a multiple instance pool. This is the most interesting 9 among various interpretations which are not considered in this work (e.g., repre- senting multiple receiving parties as different single-instance pools). We can have p A B e4 e5 Task 1 m e1 e2 that the number of sent messages is either known at design time (by setting the LoopCardinality at- tribute of the send task) or it is read from a data object during the process execution. p Task 1 m e1 e2 q Fig. 9: One-To-Many-Send. e7 e8 m1 m2 e2 e1 e(n 1) q Fig. 9: One-To-Many-Send. Textual Speciﬁcation. Here, to keep the pattern formalisation more manageable, the sequential multi-instance task is rendered as a macro. The macro encloses the task in a FOR- loop expressed by means of a pair of XOR join and split gateways, and an additional data object c1 for the loop counter. In the textual notation we have C “ poolpp, Pq ∥ miPoolpq, Qq, where process Q is left unspeciﬁed and in P the attribute LoopCardinal- ity is set to n: P A e1 e2 Fig. 9: One-To-Many-Send. Fig. 9: One-To-Many-Send. P m1 r P “ xorJoinpte1, e13u, e11q ∥taskSndpe11, c1.c ‰ null, c1.c :“ c1.c ` 1, m: ˜ exp1, e12q ∥xorSplitpe12, tpe13, c1.c ď nq, pe2, defaultquq ∥P 1 e7 m Qk Q1 e3 e3 m1 m2 m Formal semantics. The execution steps are realised as in the previous cases, by repeat- edly applying the semantic rules of the XOR gateway and the send task. It is worth noticing that at each application of rule P-TaskSnd the ﬁeld c of the data object c1 is updated with the assignment c1.c :“ c1.c ` 1. At the end of the pattern execution, the message list δpmq contains n sent messages. 4.5 One-To-Many Send Pattern Q e3 e4 e5 4.6 One-From-Many Receive Pattern P Informal Description. A party receives several logically related messages arising from autonomous events occurring at different parties. The arrival of messages must be timely so that they can be correlated as a single logical request. The interaction may complete successfully or not depending on the messages gathered. In this pattern the receiver does not know the number of messages that will arrive, and stops waiting as soon as a certain number of messages have arrived or a timeout occurs. Q C m1 m2 e4 e5 g This pattern can be modelled as the collaboration fragment shown in Fig 10. q Task 1 e1 e3 e2 e4 m p Fig. 10: One-From-Many-Receive. q Task 1 e1 e3 e2 e4 m p Fig. 10: One-From-Many-Receive. Textual Speciﬁcation. Also in this case, to sim- plify the formal treatment, we rely on a macro for the multi-instance receive task with a timer. In particular, the multi-instance behaviour is rep- resented by enclosing the receive task in a FOR- loop (as for the sequential multi-instance task). The timer attached to the receive task is instead abstracted via a non-deterministic choice, by resorting to a race condition. In detail, the receiving party q will get, via an event-based gateway, either a message from a sending party (i.e., an instance of p) or a time-out message from a speciﬁc pool t representing the timer. In the textual notation we have C “ miPoolpp, Pq ∥poolpq, Qq ∥poolpt, Tq, where Fig. 10: One-From-Many-Receive. 10 Q “ taskSndpe1, exp1, ϵ, mstartTimer : ˜ exp2, e1q ∥xorJoinpte1, evu, e2q ∥ eventBasedpe2, pm:˜t1, e3q, pmtimeout :˜t2, e3qq ∥ taskpe3, c1.c ‰ null, c1.c :“ c1.c ` 1, eivq ∥ xorSplitpeiv, tpev, c1.c ď nq, pe2, defaultquq ∥xorJoinpte2, e3u, e4q ∥Q1 T “ startRcvpmstartTimer : ˜t3, e5q ∥taskSndpe5, exp3, ϵ, mtimeout : ˜ exp4, e6q ∥endpe6q Formal Semantics. Once a token arrives at e1 in the process Q, a mstartTimer message is sent to the pool t by means of the send task, in order to activate an instance of the timer process T. This instance will perform a send task, delivering a message mtimeout, to signal that the timeout is expired, and then it terminates. As effect of the execution of the send task in Q, a token is moved in e1, which enables the looping behaviour regulated by the XOR gateways. 4.6 One-From-Many Receive Pattern P At each iteration, the event-based gateway consumes either a message m or mtimeout; in the former case the non-communicating task increments the loop counter and the execution of another interaction is evaluated (by means of the XOR split conditions), while in the latter case the edge e3 is followed and the pattern execution completes. 4.7 One-To-Many Send/Receive Pattern Informal Description. A party sends a request to several other parties. Responses are expected within a given time-frame. However, some responses may not arrive within the time-frame and some parties may even not respond at all. This pattern can be rendered as the collaboration fragment in Fig. 11. A practical use of this pattern is shown in the scenario in Fig. 1. q p Task 1 Task 2 Task 3 m1 m2 e2 e3 e4 e8 e9 e1 e5 e7 e6 Fig. 11: One-To-Many Send/Receive. Textual Speciﬁcation. This pat- tern relies on a multi-instance sub- process with a speciﬁc form, i.e. it is characterised by a sequence of a send task and a receive task, pro- ceeded and followed by a start and an end event, respectively. As usual, to simplify the formal treatment we resort to a macro. In this case, it con- sists of a sequential send task followed by a multi-instance receive task with a timer. We have C “ poolpp, Pq ∥miPoolpq, Qq, where process P is rendered in terms of macros as already shown in the previous patterns (hence, for the sake of presentation, its speciﬁcation is omitted), while process Q is as follows: q p Task 1 Task 2 Task 3 m1 m2 e2 e3 e4 e8 e9 e1 e5 e7 e6 Fig. 11: One-To-Many Send/Receive. Textual Speciﬁcation. This pat- tern relies on a multi-instance sub- process with a speciﬁc form, i.e. it is characterised by a sequence of a send task and a receive task, pro- ceeded and followed by a start and an end event, respectively. As usual, to simplify the formal treatment we resort to a macro. In this case, it con- sists of a sequential send task followed by a multi-instance receive task with a timer. We have C “ poolpp, Pq ∥miPoolpq, Qq, where process P is rendered in terms of macros as already shown in the previous patterns (hence, for the sake of presentation, its speciﬁcation is omitted), while process Q is as follows: q p Task 1 Task 2 Task 3 m1 m2 e2 e3 e4 e8 e9 e1 e5 e7 e6 Fig. 11: One-To-Many Send/Receive. Fig. 11: One-To-Many Send/Receive. Fig. 11: One-To-Many Send/Receive. resort to a macro. In this case, it con- sists of a sequential send task followed by a multi-instance receive task with a timer. 4.7 One-To-Many Send/Receive Pattern Now, the correlation takes place according to the data states, which we assume to be as follows: α1pd.fq “ α2pd.fq “ “foo”, α1pd.idq “ 7, and α2pd.idq “ 5. Therefore, the ﬁrst message is delivered to the second instance, updating α2 with the assignment d.code “9876, while the second message is delivered to the ﬁrst instance, updating α1 with the assignment d.code “1234. 4.8 Request with Referral Pattern In the textual speciﬁcation we have C “ poolpp, Pq ∥ poolpq, Qq ∥poolpr, Rq, where: P “ startpeenb, e1q ∥taskSndpe1, exp1, A1, m1: ˜ exp2, e2q ∥interRcvpe2, m2: ˜t1, e3q ∥ taskSndpe3, exp3, A2, m3: ˜ exp4, e4q ∥interRcvpe4, m4: ˜t2, e5q ∥endpe5q and Q and R are deﬁned in a similar way. and Q and R are deﬁned in a similar way. Formal Semantics. The execution steps and their results are simply realised by applying the semantic rules for the different BPMN elements, as already shown for the previous patterns. It is up to the message sent by pool p to pool r to specify in its content the reference to pool q, whose process waits for the routed message. 4.7 One-To-Many Send/Receive Pattern We have C “ poolpp, Pq ∥miPoolpq, Qq, where process P is rendered in terms of macros as already shown in the previous patterns (hence, for the sake of presentation, its speciﬁcation is omitted), while process Q is as follows: Q “ interRcvpe7, m1: ˜t1, e8q ∥taskSndpe8, exp, A, m2: ˜ exp, e9q ∥Q1 Q “ interRcvpe7, m1: ˜t1, e8q ∥taskSndpe8, exp, A, m2: ˜ exp, e9q ∥Q1 Formal Semantics. In this pattern we have that process P sends out, by means of rule P-TaskSnd, several messages of type m1 that need to be properly correlated with the correct process instance of Q. The content of the messages themselves provides the correlation information. For example, let us assume that two messages of type m1 are sent to q, and that consist of three ﬁelds, say x“foo”, 5, 1234y and x“foo”, 7, 9876y. Also, let us consider the case where there are two receiving instances, i.e. ιpqq “ Formal Semantics. In this pattern we have that process P sends out, by means of rule P-TaskSnd, several messages of type m1 that need to be properly correlated with the correct process instance of Q. The content of the messages themselves provides the correlation information. For example, let us assume that two messages of type m1 are sent to q, and that consist of three ﬁelds, say x“foo”, 5, 1234y and x“foo”, 7, 9876y. Also, let us consider the case where there are two receiving instances, i.e. ιpqq “ 11 txσ1, α1y, xσ2, α2yu, and that template ˜t1 of the intermediate receiving event is deﬁned as xd.f, d.id, ?d.codey, meaning that the ﬁelds f and id of the data object d identify cor- relation data while code is a formal ﬁeld. Now, the correlation takes place according to the data states, which we assume to be as follows: α1pd.fq “ α2pd.fq “ “foo”, α1pd.idq “ 7, and α2pd.idq “ 5. Therefore, the ﬁrst message is delivered to the second instance, updating α2 with the assignment d.code “9876, while the second message is delivered to the ﬁrst instance, updating α1 with the assignment d.code “1234. txσ1, α1y, xσ2, α2yu, and that template ˜t1 of the intermediate receiving event is deﬁned as xd.f, d.id, ?d.codey, meaning that the ﬁelds f and id of the data object d identify cor- relation data while code is a formal ﬁeld. 4.8 Request with Referral Pattern r q p Task 1 Task 3 Task 4 Task 5 e1 e2 e3 e4 e5 e6 e7 e8 e9 e10 e11 m1 m2 m3 m4 m5 Task 2 Fig. 12: Request with Referral pattern. 1 r q p Task 1 Task 3 Task 4 Task 5 e1 e2 e3 e4 e5 e6 e7 e8 e9 e10 e11 m1 m2 m3 m4 m5 Task 2 Fig. 12: Request with Referral pattern. Informal Description. A party p sends a request to another party q indicating that any follow-up should be sent to another party r. An example of a BPMN col- laboration involving the request with referral pattern is shown in Fig. 12, and also in the com- munication between the Retailer and the Logistic Provider in Fig. 1. r q p Task 1 Task 3 Task 4 Task 5 e1 e2 e3 e4 e5 e6 e7 e8 e9 e10 e11 m1 m2 m3 m4 m5 Task 2 Fig. 12: Request with Referral pattern. Informal Description. A party p sends a request to another party q indicating that any follow-up should be sent to another party r. An example of a BPMN col- laboration involving the request with referral pattern is shown in Fig. 12, and also in the com- munication between the Retailer and the Logistic Provider in Fig. 1. Textual Speciﬁcation. In the textual speciﬁcation we have C “ poolpp, Pq ∥ poolpq, Qq ∥poolpr, Rq, where: P “ startpeenb, e1q ∥taskSndpe1, exp1, A1, m1: ˜ exp2, e2q ∥interRcvpe2, m2: ˜t1, e3q ∥ taskSndpe3, exp3, A2, m3: ˜ exp4, e4q ∥interRcvpe4, m4: ˜t2, e5q ∥endpe5q r q p Task 1 Task 3 Task 4 Task 5 e1 e2 e3 e4 e5 e6 e7 e8 e9 e10 e11 m1 m2 m3 m4 m5 Task 2 Fig. 12: Request with Referral pattern. 1 Informal Description. A party p sends a request to another party q indicating that any follow-up should be sent to another party r. An example of a BPMN col- laboration involving the request with referral pattern is shown in Fig. 12, and also in the com- munication between the Retailer and the Logistic Provider in Fig. 1. Fig. 12: Request with Referral pattern. Textual Speciﬁcation. In the textual speciﬁcation we have C “ poolpp, Pq ∥ poolpq, Qq ∥poolpr, Rq, where: Textual Speciﬁcation. while P and Q are deﬁned in a similar way. while P and Q are deﬁned in a similar way. Formal Semantics. Similar to the previous pattern, the formal semantics of this pattern is determined by the application of rules for sending and receiving message already described, except for the AND split gateway that simply consumes a token in e7 and, simultaneously, produces one token in e8 and one token in e9. 3 http://pros.unicam.it/mida 5 Patterns Animation via MIDA The MIDA (Multiple Instances and Data Animator) tool3 is a web application writ- ten in JavaScript, based on the Camunda bpmn.io modeller. MIDA, whose graphical interface is shown in Fig. 14, is an animator of collaboration models that can involve multiple instances and data objects. MIDA animates process models by means of the visualisation of tokens ﬂow and data evolution. To correctly enact the collaboration be- haviour, the implementation of the tool relies on the formalisation presented in Sec. 3. The core feature of MIDA is the model animation. It re- sults helpful both in educational contexts, for explaining the be- haviour of BPMN elements, and in practical modelling activities, for debugging errors. In fact, de- signers can achieve a precise un- derstanding of the behaviour of processes and collaborations by means of the visualisation of the Fig. 14: MIDA tool interface. The core feature of MIDA is the model animation. It re- sults helpful both in educational contexts, for explaining the be- haviour of BPMN elements, and in practical modelling activities, for debugging errors. In fact, de- signers can achieve a precise un- derstanding of the behaviour of processes and collaborations by means of the visualisation of the model execution. Fig. 14: MIDA tool interface. The core feature of MIDA is the model animation. It re- sults helpful both in educational contexts, for explaining the be- haviour of BPMN elements, and in practical modelling activities, for debugging errors. In fact, de- signers can achieve a precise un- derstanding of the behaviour of processes and collaborations by means of the visualisation of the model execution. Fig. 14: MIDA tool interface. Fig. 14: MIDA tool interface. We have exploited MIDA to model and animate the interaction patterns presented in Sec. 4, providing an intuitive knowledge of their behaviour. We have also used MIDA to animate the motivating example in Fig. 1, thus showing how the tool sup- ports the study of more intricate scenarios resulting from the combination of var- ious patterns. These animations are available from http://pros.unicam.it/ service-interaction-patterns/. 4.9 Relayed Request Pattern r q p Task 1 Task 2 Task 3 Task 4 e1 e2 e3 e4 e5 e6 e8 e7 e9 e10 e11 m1 m2 m3 m4 Fig. 13: An example of Relayed Request pattern. Informal Description. A party p makes a request to party q, which delegates the request pro- cessing to another party r. This latter party interacts with party p while party q observes a view of the interactions. This pattern can be rendered as the collaboration fragment in Fig. 13. Fig. 13: An example of Relayed Request pattern. 12 Textual Speciﬁcation. In the textual notation the pattern is rendered as follows: C “ poolpp, Pq ∥poolpq, Qq ∥poolpr, Rq, where: Textual Speciﬁcation. In the textual notation the pattern is rendered as follows: C “ poolpp, Pq ∥poolpq, Qq ∥poolpr, Rq, where: R “ startRcvpm2: ˜t1, e7q ∥andSplitpe7, te8, e9uq ∥ taskSndpe8, exp1, ϵ, m4: ˜ exp2, e10q ∥taskSndpe9, exp3, ϵ, m3: ˜ exp4, e11q ∥R1 6 Related Work The most common interaction scenarios from a business perspective, named Service Interaction Patterns, have been described in [3,17]. However, they lack of visualisation as well as formal semantics. Since then, effort has been devoted to visualise [5,1,9] and 13 formalise these patterns [10,2,15], as shown in Table 1. This provides a comparison among the state-of-the-art approaches dealing with service interaction patterns with respect to: (i) the language used for patterns speciﬁcation, (ii) the main contribution of the work and (iii) its limitations. In the following, these works are compared to the contribution of this paper. Paper Year Language Contribution Limitations [2] 2005 ASM Formalisation and extension No models analysis [10] 2006 π-calculus Formalisation Ambiguities [15] 2007 CPN Formalisation and extension No correlation [9] 2008 BPMN 2.0 iBPMN: extension for interaction modelling No formalisation [1] 2009 Open Nets Overview of services domain challenges No formalisation [5] 2014 BPMN 2.0 Extension of BPMN supported patterns No formalisation Table 1: Review on the Service Interaction Patterns literature. Considering the patterns speciﬁed in BPMN, relevant works are [5,9]. Campagna et al. [5] discuss BPMN 2.0 support for the service interaction patterns and propose a set of enhancements to broaden it. However, they do not formalise these patterns and thus, they do not provide formal validation of the proposed solutions. Decker and Barros [9] introduce iBPMN, a set of extensions to the BPMN standard for interaction modelling. They show that most service interaction patterns can be expressed using iBPMN and present an algorithm for deriving interface behaviour models from simple interaction models. However, they do not aim at providing a formal characterisation of the proposed extensions. Both the above works are more interested in overcoming BPMN lacks for supporting interaction patterns rather than clarifying the semantics of the supported patterns, which is instead a major challenge when using BPMN collaborations [8]. Abstract State Machines (ASM) [4], π-calculus [14] and Petri Nets [12] have been proposed as a solid ground to formalise service interaction patterns. The ﬁrst formal- ization of the patterns was given by Barros and Börger [2] proposing a compositional framework and interaction ﬂows. They provide ASM for eight service interaction sce- narios and illustrate how, by combinations and reﬁnements of them, one can deﬁne arbitrarily complex interaction patterns. The ASMs offer an implementation draft of the patterns, but are less suited for the analysis of collaborations. 6 Related Work Decker and Puhlmann [10] provide a formalisation of service interactions via π-calculus as a ﬁrst step to anal- yse collaborations. However, their work shows still some ambiguities. For instance, the Racing Incoming Messages pattern allows to receive multiple messages at once, but the work does not clarify how one among the competing messages is chosen for consump- tion. Moreover, the authors refer to a synchronous communication model, not compliant with the BPMN standard. Mulyar et al. [15] formalise the semantics of the interaction patterns by means of Coloured Petri Nets (CPN). Moreover, they extend the scope of the original service interaction patterns by describing various pattern variants. How- ever, even if Petri Nets provide support for multiple instance patterns, process instances are characterised by their identities, rather than by the values of their data, which are necessary for correlation [11]. Finally, van der Aalst et al. [1] provide an overview of the challenges in the domain of service interaction patterns and they propose to use open nets as a formal framework for addressing these challenges. However, they do not 14 aim at formalising the interaction patterns, of which they only provide a brief descrip- tion. Differently from the mentioned works, we focus on BPMN by directly deﬁning the semantics of the supported patterns, thus avoiding the mapping to other formalisms equipped with their own semantics. 7 Concluding Remarks In this work we focus on service interaction patterns, visualising them in BPMN col- laborations and providing a comprehensive formalisation by means of a direct formal semantics for BPMN collaboration diagrams. This allows to validate the semantics in [6], as we show it is suitable to cover the interaction patterns expressed in BPMN. Fur- ther, the animation tool MIDA has been exploited to model and animate the BPMN collaborations, allowing an intuitive understanding of the patterns execution. As a future work, we plan to investigate the formalisation of new BPMN interaction patterns where data objects play a more central role. References 1. van der Aalst, W.M., Mooij, A.J., Stahl, C., Wolf, K.: Service interaction: Patterns, formal- ization, and analysis. In: Formal Methods for Web Services. pp. 42–88. Springer (2009) y pp p g ( ) 2. Barros, A., Börger, E.: A compositional framework for service interaction patterns and inter- action ﬂows. In: Formal Engineering Methods. pp. 5–35. Springer (2005) 3. Barros, A., Dumas, M., Ter Hofstede, A.H.: Service interaction patterns. In: Business Process Management, pp. 302–318. Springer (2005) 4. Börger, E., Thalheim, B.: Modeling Workﬂows, Interaction Patterns, Web Services and Busi- ness Processes: The ASM-Based Approach. In: ASM, B and Z, pp. 24–38. Springer (2008) 5. Campagna, D., Kavka, C., Onesti, L.: Enhancing BPMN 2.0 support for service interaction patterns. In: Software Engineering and Applications (ICSOFT-EA) (2014) 6. Corradini, F., Muzi, C., Re, B., Tiezzi, F., Rossi, L.: Animating multiple instances in bpmn collaborations: from formal semantics to tool support. In: BPM 2018. LNCS, Springer (2018), to appear 7. Corradini, F., Muzi, C., Re, B., Tiezzi, F., Rossi, L.: Mida: Multiple instances and data ani- mator. In: BPM 2018 (Demo). LNCS, Springer (2018), to appear 8. Cortes-Cornax, M., Dupuy-Chessa, S., Rieu, D.: Choreographies in BPMN 2.0: new chal- lenges and open questions. In: ZEUS. vol. 847, pp. 50–57. CEUR-WS.org (2012) 9. Decker, G., Barros, A.: Interaction modeling using BPMN. LNCS 4928, 208–219 (2008) , , , g g , ( ) 10. Decker, G., Puhlmann, F., Weske, M.: Formalizing service interactions. In: International Conference on Business Process Management. pp. 414–419. Springer (2006) 11. Decker, G., Weske, M.: Instance isolation analysis for service-oriented architectures. In: SCC. vol. 1, pp. 249–256. IEEE (2008) pp 12. Dijkman, R.M., Dumas, M., Ouyang, C.: Semantics and analysis of business process models in BPMN. Information and Software Technology 50(12), 1281–1294 (2008) 13. Dumas, M., La Rosa, M., Mendling, J., Reijers, H.A.: Fundamentals of business process management. Springer (2013) 14. Milner, R.: A Calculus of Communicating Systems. Springer, Secaucus, NJ, USA (1980) 15. Mulyar, N., van der Aalst, W.M., Aldred, L., Russell, N.: Service interaction patterns: A conﬁgurable framework. management 34, 29 (2007) g g 16. OMG: Business Process Model and Notation (BPMN V 2.0) (2011) 17. Weske, M.: Business Process Management. Springer (2012) 15
https://openalex.org/W2466084116	https://publicatio.bibl.u-szeged.hu/9329/1/3087549_Kurgyis_ijms.pdf	English	null	Melanoma-Derived BRAFV600E Mutation in Peritumoral Stromal Cells: Implications for in Vivo Cell Fusion	International journal of molecular sciences	2,016	cc-by	5,914	Melanoma-Derived BRAFV600E Mutation in Peritumoral Stromal Cells: Implications for in Vivo Cell Fusion Zsuzsanna Kurgyis 1,†, Lajos V. Kemény 1,†, Tünde Buknicz 1, Gergely Groma 2, Judit Oláh 1, Ádám Jakab 1, Hilda Polyánka 2, Kurt Zänker 3, Thomas Dittmar 3, Lajos Kemény 1,2,,‡ and István B. Németh 1,,‡ 1 Department of Dermatology and Allergology, University of Szeged, Szeged 6720, Hungary; kurgyis.zsuzsanna@med.u-szeged.hu (Z.K.); kemeny.lajos@gmail.com (L.V.K.); buknicz.tunde@med.u-szeged.hu (T.B.); lazarne.olah.judit@med.u-szeged.hu (J.O.); yakoo@freemail.hu (A.J.) 1 Department of Dermatology and Allergology, University of Szeged, Szeged 6720, Hungary; kurgyis.zsuzsanna@med.u-szeged.hu (Z.K.); kemeny.lajos@gmail.com (L.V.K.); buknicz.tunde@med.u-szeged.hu (T.B.); lazarne.olah.judit@med.u-szeged.hu (J.O.); yakoo@freemail.hu (A.J.) y 2 MTA-SZTE Dermatological Research Group, Szeged 6720, Hungary; groma.gergo@gmail.com (G.G.); hilda@mail.derma.szote.u-szeged.hu (H.P.) 2 MTA-SZTE Dermatological Research Group, Szeged 6720, Hungary; groma.gergo@gmail.com (G.G.); hilda@mail.derma.szote.u-szeged.hu (H.P.) g 3 Institute of Immunology & Experimental Oncology, Witten/Herdecke University, Witten 58453, Germa k t k @ i h d (K Z ) th ditt @ i h d (T D ) 3 Institute of Immunology & Experimental Oncology, Witten/Herdecke University, Witten 58453, Germany; kurt.zaenker@uni-wh.de (K.Z.); thomas.dittmar@uni-wh.de (T.D.) * Correspondence: kemeny.lajos@med.u-szeged.hu (L.K.); nemeth.istvan.balazs@med.u-szeged.hu (I.B.N Tel.: +36-62-545-277 (L.K. & I.B.N.) † These authors contributed equally to this work. ‡ These authors contributed equally to this work. Academic Editor: Terrence Piva Received: 31 March 2016; Accepted: 13 June 2016; Published: 21 June 2016 Abstract: Melanoma often recurs in patients after the removal of the primary tumor, suggesting the presence of recurrent tumor-initiating cells that are undetectable using standard diagnostic methods. presence of recurrent tumor initiating cells that are undetectable using standard diagnostic methods. As cell fusion has been implicated to facilitate the alteration of a cell’s phenotype, we hypothesized that cells in the peritumoral stroma having a stromal phenotype that initiate recurrent tumors might originate from the fusion of tumor and stromal cells. Here, we show that in patients with BRAFV600E melanoma, melanoma antigen recognized by T-cells (MART1)-negative peritumoral stromal cells express BRAFV600E protein. To conﬁrm the presence of the oncogene at the genetic level, peritumoral stromal cells were microdissected and screened for the presence of BRAFV600E with a mutation-speciﬁc polymerase chain reaction. Interestingly, cells carrying the BRAFV600E mutation were not only found among cells surrounding the primary tumor but were also present in the stroma of melanoma metastases as well as in a histologically tumor-free re-excision sample from a patient who subsequently developed a local recurrence. We did not detect any BRAFV600E mutation or protein in the peritumoral stroma of BRAFWT melanoma. Melanoma-Derived BRAFV600E Mutation in Peritumoral Stromal Cells: Implications for in Vivo Cell Fusion Therefore, our results suggest that peritumoral stromal cells contain melanoma-derived oncogenic information, potentially as a result of cell fusion. These hybrid cells display the phenotype of stromal cells and are therefore undetectable using routine histological assessments. Our results highlight the importance of genetic analyses and the application of mutation-speciﬁc antibodies in the identiﬁcation of potentially recurrent-tumor-initiating cells, which may help better predict patient survival and disease outcome. Keywords: cell fusion; melanoma; mutation detection; macrophage; ﬁbroblast; BRAFV600E International Journal of Molecular Sciences International Journal of Molecular Sciences 1. Introduction Melanoma is a highly aggressive malignancy, which often gives rise to metastases and local recurrences following tumor-free staging [1,2]. The histopathological examination of excisional Int. J. Mol. Sci. 2016, 17, 980; doi:10.3390/ijms17060980 www.mdpi.com/journal/ijms www.mdpi.com/journal/ijms Int. J. Mol. Sci. 2016, 17, 980 2 of 9 and re-excisional samples and sentinel lymph node biopsies constitutes a fundamental part of the prognostic evaluation of this disease. Thus, the ability to detect recurrent-tumor-initiating cells in the tissue sample is highly important both for prognostic purposes and for managing the disease. As a result of tumor heterogeneity, tumor cells often differentially express phenotypic markers [3,4], rendering their detection during routine histological assessments difﬁcult. Epithelial-mesenchymal transition, dedifferentiation, and acquisition of stem-cell characteristics have been shown to make tumor cells more malignant [3,5,6]. Although the mechanisms by which these features are acquired have not yet been fully elucidated, epigenomic and genomic alterations are thought to be the key drivers of this malignant heterogeneity [7,8]. Cell fusion has been implicated in playing a major role in tumor progression in numerous tumor types [9–11]. Several studies have found that hybrid cells of tumor-stromal cell fusion are more malignant than the parental tumor cells and contribute to tumor heterogeneity [12]. Moreover, it has been shown that cell fusion might alter the phenotype of the parental cells [13–15] and that it can be an alternative way of epithelial mesenchymal transition [16–18]. Therefore, we hypothesized that tumor cells could similarly adopt a stromal phenotype by fusing with peritumoral stromal cells, while maintaining their oncogenic properties. Such fusion may lead to a cell type with mixed features that can evade immune recognition and clinical detection and therefore promote tumor recurrence. Thus, the detection of such hybrid cells in the intra- and peritumoral stroma could be signiﬁcant for the accurate histopathological diagnosis and subsequent therapy. 2.1. The BRAFV600E Protein Is Expressed in Subpopulations of Peritumoral Stromal Cells To identify potential recurrent tumor-initiating cells, we screened for peritumoral cells that display a stromal phenotype but carry tumor-derived oncogenic information. For this we performed dual immunohistochemical staining on n = 11 patient-derived tissue samples of BRAFV600E melanoma with MART1 (melanoma antigen recognized by T-cells) and BRAFV600E. BRAFV600E is an oncogenic somatic mutation present in approximately 50%–60% of malignant melanomas. The staining revealed that, in addition to melanoma cells, BRAFV600E is also expressed in some subpopulations of MART1´ peritumoral stromal cells with ﬁbroblast and macrophage morphology (Figure 1a,b). To investigate whether the mutant protein originated from the tumor cells, we tested the stroma of n = 5 BRAFWT melanomas for the presence of the BRAFV600E protein. However, we could not detect the mutant protein in BRAFWT melanoma tissue samples (Supplementary Materials Figure S1), suggesting that the presence of the mutation in the stromal cells could not occur without the neighboring melanoma cells carrying the mutation. Nevertheless, though the antibody has been reported to be highly speciﬁc, and we did not see any speciﬁc signals in BRAFWT melanoma tissue samples, we cannot rule out nonspeciﬁc staining especially in case of macrophages. 2.2. Some Peritumoral Fibroblasts and Macrophages Carry the BRAFV600E Mutation in Primary Melanoma, Melanoma Metastasis and a Tumor-Free Re-Excision Sample 2.2. Some Peritumoral Fibroblasts and Macrophages Carry the BRAFV600E Mutation in Primary Melanoma, Melanoma Metastasis and a Tumor-Free Re-Excision Sample (a,b) Tissue samples of a patient with BRAFV600E melanoma stained for the melanoma marker melanoma antigen recognized by T-cells (MART1) (red), the BRAFV600E mutant protein (brown) and hematoxylin (blue). Brown arrows indicate peritumoral MART1−/BRAFV600E+ cells displaying fibroblast (a) or macrophage (b) morphology. Blue arrows indicate MART1−/BRAFV600E− stromal cells (light blue from hematoxylin). Red lines indicate MART1+ melanoma cells. Black bars indicate 50 µm. Figure 1. Peritumoral stromal cells express the melanoma-derived oncogenic BRAFV600E protein. (a,b) Tissue samples of a patient with BRAFV600E melanoma stained for the melanoma marker melanoma antigen recognized by T-cells (MART1) (red), the BRAFV600E mutant protein (brown) and hematoxylin (blue). Brown arrows indicate peritumoral MART1´/BRAFV600E+ cells displaying ﬁbroblast (a) or macrophage (b) morphology. Blue arrows indicate MART1´/BRAFV600E´ stromal cells (light blue from hematoxylin). Red lines indicate MART1+ melanoma cells. Black bars indicate 50 µm. Figure 1. Peritumoral stromal cells express the melanoma-derived oncogenic BRAFV600E protein. (a,b) Tissue samples of a patient with BRAFV600E melanoma stained for the melanoma marker melanoma antigen recognized by T-cells (MART1) (red), the BRAFV600E mutant protein (brown) and hematoxylin (blue). Brown arrows indicate peritumoral MART1−/BRAFV600E+ cells displaying fibroblast (a) or macrophage (b) morphology. Blue arrows indicate MART1−/BRAFV600E− stromal cells (light blue from hematoxylin). Red lines indicate MART1+ melanoma cells. Black bars indicate 50 µm. Figure 1. Peritumoral stromal cells express the melanoma-derived oncogenic BRAFV600E protein. (a,b) Tissue samples of a patient with BRAFV600E melanoma stained for the melanoma marker melanoma antigen recognized by T-cells (MART1) (red), the BRAFV600E mutant protein (brown) and hematoxylin (blue). Brown arrows indicate peritumoral MART1´/BRAFV600E+ cells displaying ﬁbroblast (a) or macrophage (b) morphology. Blue arrows indicate MART1´/BRAFV600E´ stromal cells (light blue from hematoxylin). Red lines indicate MART1+ melanoma cells. Black bars indicate 50 µm. p g y BRAFV600E mutation at the genetic level. 2.2. Some Peritumoral Fibroblasts and Macrophages Carry the BRAFV600E Mutation in Primary Melanoma, Melanoma Metastasis and a Tumor-Free Re-Excision Sample We wanted to conﬁrm that the staining we observed in the BRAFV600E samples was speciﬁc, so we examined whether this oncogenic information is also present in the peritumoral cells at the genetic level. Therefore, we dual-stained BRAFV600E primary melanoma tissue samples with MART1 and either the ﬁbroblast marker smooth muscle actin (SMA) or the monocyte-macrophage marker CD68, and isolated cell compartments consisting of 20–50, clearly MART1´ but either SMA+ or CD68+ cells with ﬁbroblast and macrophage morphology, respectively, using laser-capture microdissection (Figure 2a,b). Subsequent allele speciﬁc mutation detection PCR analyses performed on the dissected stromal cells revealed that the BRAFV600E mutant allele was present in peritumoral MART1´/SMA+ ﬁbroblasts and MART1´/CD68+ macrophages. In addition to primary melanoma tissue samples, such peritumoral stromal cells carrying the melanoma-derived mutation were detected in lymph node 3 of 9 Int. J. Mol. Sci. 2016, 17, 980 and cutaneous melanoma metastases (Figure 2c). Surprisingly, BRAFV600E was also detected when analyzing MART1´ macrophages dissected from a histologically tumor-free re-excision sample from a patient who subsequently developed a local recurrence. Having dissected peritumoral stromal cells from BRAFWT melanoma tissue samples, we only detected the wild-type allele, providing further evidence that BRAFV600E is only present in peritumoral cells that are adjacent to a mutant tumor. In total, we examined 86 dissected samples of 19 patients, and out of 12 patients with BRAFV600E melanoma, we found BRAFV600E-containing peritumoral cells in the tissue samples of ﬁve patients. The proportion of mutant alleles detected in these cell populations varied between 0.5% and 30% (Table 1 and Supplementary Materials Table S1). Int. J. Mol. Sci. 2016, 17, 980 3 of 9 analyzing MART1− macrophages dissected from a histologically tumor-free re-excision sample from a patient who subsequently developed a local recurrence. Having dissected peritumoral stromal cells from BRAFWT melanoma tissue samples, we only detected the wild-type allele, providing further evidence that BRAFV600E is only present in peritumoral cells that are adjacent to a mutant tumor. In total, we examined 86 dissected samples of 19 patients, and out of 12 patients with BRAFV600E melanoma, we found BRAFV600E-containing peritumoral cells in the tissue samples of five patients. The proportion of mutant alleles detected in these cell populations varied between 0.5% and 30% (Table 1 and Supplementary Materials Table S1). Figure 1. Peritumoral stromal cells express the melanoma-derived oncogenic BRAFV600E protein. 2.2. Some Peritumoral Fibroblasts and Macrophages Carry the BRAFV600E Mutation in Primary Melanoma, Melanoma Metastasis and a Tumor-Free Re-Excision Sample Tissue 1 Npos/Nex 2,3 Fibroblasts 4 Dissected Macrophages 5 Dissected BRAFV600E primary melanoma 2/2 1/1 BRAFV600E melanoma metastasis 2/4 2/4 Histologically tumor-free tissue (from patients with BRAFV600E melanoma) 0/4 1/3 BRAFWT primary melanoma 0/4 0/1 1 Tissue samples were dual-stained either with melanoma antigen recognized by T-cells (MART1) & smooth muscle actin (SMA) or with MART1 & CD68 and examined by a certified pathologist (IBN); 2 Npos: number of patients where BRAFV600E-positive peritumoral stromal cells were found; Nex: the number of patients examined; 3 Dissected cell samples were considered positive for BRAFV600E if the prevalence of the mutant allele was higher than 0.39% (ΔCt < 8); 4 MART1−/SMA+ cells were considered fibroblasts; 5 MART1−/CD68+ cells were considered macrophages. Table 1. Peritumoral fibroblasts and macrophages in BRAFV600E melanoma carry the melanoma-derived BRAFV600E mutation at the genetic level. Tissue 1 Npos/Nex 2,3 Fibroblasts 4 Dissected Macrophages 5 Dissected BRAFV600E primary melanoma 2/2 1/1 BRAFV600E melanoma metastasis 2/4 2/4 Histologically tumor-free tissue (from patients with BRAFV600E melanoma) 0/4 1/3 BRAFWT primary melanoma 0/4 0/1 1 Tissue samples were dual-stained either with melanoma antigen recognized by T-cells (MART1) & smooth muscle actin (SMA) or with MART1 & CD68 and examined by a certiﬁed pathologist (IBN); 2 Npos: number of patients where BRAFV600E-positive peritumoral stromal cells were found; Nex: the number of patients examined; 3 Dissected cell samples were considered positive for BRAFV600E if the prevalence of the mutant allele was higher than 0.39% (∆Ct < 8); 4 MART1´/SMA+ cells were considered ﬁbroblasts; 5 MART1´/CD68+ cells were considered macrophages. p g y BRAFV600E mutation at the genetic level. Ti 1 Npos/Nex 2,3 Table 1. Peritumoral fibroblasts and macrophages in BRAFV600E melanoma carry the melanoma-derived BRAFV600E mutation at the genetic level. 4 of 9 Int. J. Mol. Sci. 2016, 17, 980 Figure 2. Peritumoral stromal cells contain the melanoma-derived oncogenic BRAFV600E at the genomic level. (a) BRAFV600E melanoma tissue sample stained for MART1 (red), smooth muscle actin (SMA) (brown), and hematoxylin (blue) before (left panel) and after (right panel) laser-capture microdissection of MART1−/SMA+ fibroblasts (white line); (b,c) Tissue samples of BRAFV600E (b) primary melanoma and (c) melanoma metastasis stained for MART1 (red), CD68 (brown), and hematoxylin (blue) before (left panels) and after (right panels) laser-capture microdissection of MART1−/CD68+ macrophages (black line). Black bars indicate 150 µm. Figure 2. Peritumoral stromal cells contain the melanoma-derived oncogenic BRAFV600E at the genomic level. 2.2. Some Peritumoral Fibroblasts and Macrophages Carry the BRAFV600E Mutation in Primary Melanoma, Melanoma Metastasis and a Tumor-Free Re-Excision Sample (a) BRAFV600E melanoma tissue sample stained for MART1 (red), smooth muscle actin (SMA) (brown), and hematoxylin (blue) before (left panel) and after (right panel) laser-capture microdissection of MART1´/SMA+ ﬁbroblasts (white line); (b,c) Tissue samples of BRAFV600E (b) primary melanoma and (c) melanoma metastasis stained for MART1 (red), CD68 (brown), and hematoxylin (blue) before (left panels) and after (right panels) laser-capture microdissection of MART1´/CD68+ macrophages (black line). Black bars indicate 150 µm. Figure 2. Peritumoral stromal cells contain the melanoma-derived oncogenic BRAFV600E at the genomic level. (a) BRAFV600E melanoma tissue sample stained for MART1 (red), smooth muscle actin (SMA) (brown), and hematoxylin (blue) before (left panel) and after (right panel) laser-capture microdissection of MART1−/SMA+ fibroblasts (white line); (b,c) Tissue samples of BRAFV600E (b) primary melanoma and (c) melanoma metastasis stained for MART1 (red), CD68 (brown), and hematoxylin (blue) before (left panels) and after (right panels) laser-capture microdissection of MART1−/CD68+ macrophages (black line). Black bars indicate 150 µm. Figure 2. Peritumoral stromal cells contain the melanoma-derived oncogenic BRAFV600E at the genomic level. (a) BRAFV600E melanoma tissue sample stained for MART1 (red), smooth muscle actin (SMA) (brown), and hematoxylin (blue) before (left panel) and after (right panel) laser-capture microdissection of MART1´/SMA+ ﬁbroblasts (white line); (b,c) Tissue samples of BRAFV600E (b) primary melanoma and (c) melanoma metastasis stained for MART1 (red), CD68 (brown), and hematoxylin (blue) before (left panels) and after (right panels) laser-capture microdissection of MART1´/CD68+ macrophages (black line). Black bars indicate 150 µm. These results indicate that some peritumoral stromal cells contain a melanoma derived oncogene at both the DNA and protein levels. The possible recurrent tumor-initiating potential of these cells is supported by our clinical observations of the local recurrence rate in our melanoma patients diagnosed before 1998. Patients who had primary melanoma removed completely, based on a clinical diagnosis of a non-melanoma skin lesion and with histologically tumor-free excision margins developed local recurrence significantly more often (p = 0.7 × 10−5, 12 out of 228 patients) than patients with melanomas excised with 5-cm margins (8 out of 935 patients), implying that recurrent- tumor-initiating cells could indeed be present in the peritumoral stroma of malignant melanoma. These results indicate that some peritumoral stromal cells contain a melanoma-derived oncogene at both the DNA and protein levels. The possible recurrent tumor-initiating potential of these cells is supported by our clinical observations of the local recurrence rate in our melanoma patients diagnosed before 1998. 2.2. Some Peritumoral Fibroblasts and Macrophages Carry the BRAFV600E Mutation in Primary Melanoma, Melanoma Metastasis and a Tumor-Free Re-Excision Sample Patients who had primary melanoma removed completely, based on a clinical diagnosis of a non-melanoma skin lesion and with histologically tumor-free excision margins developed local recurrence signiﬁcantly more often (p = 0.7 ˆ 10´5, 12 out of 228 patients) than patients with melanomas excised with 5-cm margins (8 out of 935 patients), implying that recurrent-tumor-initiating cells could indeed be present in the peritumoral stroma of malignant melanoma. 3. Discussion Aft th 3. Discussion After the complete resection of primary cutaneous melanoma, patients occasionally develop local recurrence, which is considered an independent prognostic factor [1]. This can be explained by hidden tumor-initiating cells containing tumor-derived genetic information as local minimal residual disease in the peritumoral area. In addition, such tumor-derived cells can also spread to distant areas such as lymph nodes. Tu o ell a e o o ly ide tified du i outi e hi tolo i al a e e t u i a ke that After the complete resection of primary cutaneous melanoma, patients occasionally develop local recurrence, which is considered an independent prognostic factor [1]. This can be explained by hidden tumor-initiating cells containing tumor-derived genetic information as local minimal residual disease in the peritumoral area. In addition, such tumor-derived cells can also spread to distant areas such as lymph nodes. Tumor cells are commonly identified during routine histological assessments using markers that are often insufficiently sensitive to distinguish tumor cells from stromal cells, which is supported by additional genetic analyses of sentinel lymph nodes [19]. The possible role of isolated tumor cells in the sentinel lymph node is debated, and the predictive value of additional diagnostic approaches, such as melanoma-marker detection at the mRNA level, to date, is inconclusive [20]. Nevertheless, some studies using enhanced pathological assessments have found a link between the presence of Tumor cells are commonly identiﬁed during routine histological assessments using markers that are often insufﬁciently sensitive to distinguish tumor cells from stromal cells, which is supported by additional genetic analyses of sentinel lymph nodes [19]. The possible role of isolated tumor cells in the sentinel lymph node is debated, and the predictive value of additional diagnostic approaches, such as melanoma-marker detection at the mRNA level, to date, is inconclusive [20]. Nevertheless, 5 of 9 Int. J. Mol. Sci. 2016, 17, 980 some studies using enhanced pathological assessments have found a link between the presence of single tumor cells in the sentinel lymph node and worse disease outcome compared with completely tumor-free sentinel lymph nodes [21–23], and other improved tumor detection methods, such as quantitative immunocytochemistry combined with single-cell comparative genomic hybridization, have been found to better predict patient survival and disease outcome [24]. 3. Discussion Aft th 3. Discussion The incomplete removal of tumor cells often results in a higher recurrence rate: small excision margins lead to a more frequent tumor recurrence even if the excision margins are assessed to be tumor-free [2,25–27]. These data imply that recurrent-tumor-initiating cells with an altered phenotype could indeed be present in the peritumoral stroma of malignant melanoma, allowing them to remain undetected during routine diagnostic procedures. In this study, we examined patient-derived melanoma tissue samples and detected peritumoral cells displaying stromal cell phenotype but carrying the oncogenic BRAFV600E mutation characteristic of the adjacent melanoma cells. First, we detected the mutated protein with a mutation-speciﬁc antibody. Even though nonspeciﬁc staining of the antibody in peritumoral mononuclear cells have been reported [28,29], the fact that peritumoral ﬁbroblasts were also positive argues for a speciﬁc signal. Moreover, we only observed positivity in peritumoral cells adjacent to BRAFV600E melanoma. Nevertheless, to conﬁrm the presence of the oncogene in these cells, we also performed genomic analyses and showed that the mutation is also present in cells with a stromal cell phenotype adjacent to BRAFV600E melanoma cells at the genetic level. We believe the mutation originates from tumor cells, as we did not detect the mutation in the stroma of BRAFWT melanoma samples. Possible mechanisms for the transfer of tumor-derived information to stromal cells include tumor-stromal cell fusion, exosomal protein transfer and the phagocytosis of tumor debris. However, subpopulations of peritumoral stromal cells carried the BRAFV600E mutation not only at protein but also at genomic level, suggesting cell fusion as the underlying mechanism. It is important to mention that, since melanoma cells can lose MART1 expression, especially on the periphery of the tumor, cell morphology and the expression patterns of CD68 or SMA were all taken into account during the process of peritumoral cell identiﬁcation. Cell fusion studies based on in vitro observations and mouse models suggest that cell fusion can confer an evolutionary beneﬁt, such as increased metastatic potential [30] or drug resistance [31] to certain tumor–stromal cell hybrid clones. However, the majority of in vitro spontaneously formed hybrid cells undergo apoptosis [32], most likely as a result of mitotic stress. Therefore, it is possible that tumor stromal cell fusion is rather an antitumor mechanism, eliminating most of the hybrid tumor cells but occasionally giving rise to highly malignant tumor cell clones. 3. Discussion Aft th 3. Discussion Studies in mice [33] imply that cell fusion may also take place in human tumors in vivo. However, it is very difﬁcult to detect tumor–stromal cell fusion on a genetic level in humans. To address this difﬁculty, tumors developing in patients who have received bone marrow transplants have been investigated: genetic material originating from the donor was detected in the patients’ tumor cells [34,35], strongly suggesting a fusion event between recipient tumor cells and donor hemopoietic cells. However, inﬂammation resulting from treatment (e.g., whole body irradiation or chemotherapeutics) preceding bone-marrow transplantation has been reported to promote cell fusion [36–38]; therefore, further studies investigating fusion between tumor and stromal cells are clearly required. Nevertheless, human in vivo studies demonstrating the relevance of tumor–stromal cell fusion in the clinical diagnosis or treatment of cancer have been lacking. In conclusion, our results suggest that peritumoral stromal cells contain melanoma-speciﬁc oncogenic properties such as the BRAFV600E mutation derived from the neighboring tumor cells, potentially as a result of cell fusion. Based on the literature data, these cells, following a reversal to a tumorous phenotype [12], can possibly contribute to tumor recurrence. These cells are phenotypically indistinguishable from peritumoral stromal cells and are therefore not accurately detectable by routine histological assessments. Our results highlight the importance of genetic analysis and mutation-speciﬁc antibodies, especially in the case of histologically tumor-free tissue samples, 6 of 9 Int. J. Mol. Sci. 2016, 17, 980 for which improved methods for the detection of tumor cells have already been shown to better predict survival and outcome. Although yet to be conﬁrmed in clinical trials, the use of genetic analyses and mutation-speciﬁc antibodies could have important prognostic and therapeutic consequences and could enable the detection of recurrent-tumor-initiating cells. 4.1. Tissue Samples and Determination of BRAF Mutational Status Tissue samples from n = 11 patients with BRAFV600E and n = 5 patients with BRAFWT melanoma were examined for BRAFV600E protein expression. The corresponding patient numbers for genetic analyses were n = 11 in the case of BRAFV600E melanoma and n = 8 in the case of BRAFWT melanoma. The BRAF mutational status of melanomas was determined with cobas® 4800 BRAFV600 Mutation Test (Roche Molecular Diagnostics, Pleasanton, CA, USA). 4.2. Immunohistochemistry Formalin-ﬁxed, parafﬁn-embedded tissue sections of patients with BRAFV600E malignant melanoma were stained with rabbit monoclonal MART1 (clone A19-P, DB Biotech, Kosice, Slovakia), mouse monoclonal SMA (clone 1A4, DAKO, Glostrup, Denmark), CD68 (clone PG-M1, DAKO), and BRAFV600E (Clone VE1, Springbio, Pleasanton, CA, USA) primary antibodies. The Bond Polymer Reﬁne Detection Kit and the ChromoPlex 1 Dual Detection Kit (both from Leica, Wetzlar, Germany) were used for the visualization of single and dual histochemical staining, respectively, according to the manufacturer’s instructions. All immunohistochemical staining was scanned with a Pannoramic MIDI Slide Scanner (3DHISTECH Ltd., Budapest, Hungary) and analyzed with Pannoramic Viewer software (3DHISTECH Ltd.). 4.3. Laser-Capture Microdissection and Detection of the BRAFV600E Allele 30–100 cells were dissected from tissue sections stained either with MART1 and SMA or with MART1 and CD68 using a Palm Microbeam (Carl Zeiss Microscopy, Jena, Germany). Samples were digested with proteinase K (QIAGEN, Venlo, The Netherlands). Mutant allele (BRAF_476_mu) and gene reference (Braf_rf) TaqMan® Mutation Detection Assays (Life Technologies, Carlsbad, CA, USA) were used to detect the BRAFV600E mutant allele in the dissected samples. Samples with ∆Ct (Ctmut ´ Ctref) < 8 and Ctmut < 40 were considered carrying the mutant allele. MART1+ tumor cells dissected from BRAFWT and BRAFV600E melanoma were used as negative and positive controls, respectively. Abbreviations MART1 melanoma antigen recognized by T-cells SMA smooth muscle actin MART1 melanoma antigen recognized by T-cells SMA smooth muscle actin Local recurrence rate in melanoma patients were compared with the Fisher’s exact test. The level of signiﬁcance was set to 0.05. Local recurrence rate in melanoma patients were compared with the Fisher’s exact test. The level of signiﬁcance was set to 0.05. Supplementary Materials: Supplementary materials can be found at http://www.mdpi.com/1422-0067/ 17/6/980/s1. Acknowledgments: The authors would like to thank their colleagues at the Dermatohistopathological Laboratory, University of Szeged, Hungary: Erika Függ and Mónika Kohajda for technical support and assistance, Márta Széll, Kornélia Szabó, Attila Bebes, Anikó Göblös, and Gábor Tax for our useful discussions and their comments, and Zoltán Tóbiás for preparing the illustrations. We are also very grateful to Ferhan Ayaydin (Biological Research Center, Hungarian Academy of Science, Szeged, Hungary) for providing the Palm Microbeam microscope and for the technical assistance. This work was supported by TÁMOP-4.2.1/B-09/1/KONV-2010-0005 and by TÁMOP-4.2.2/B grants. István B. Németh was supported by the Bolyai Scholarship 2015/17 of the Hungarian Academy of Sciences. Thomas Dittmar was supported by the Fritz-Bender-Foundation, Munich, Germany. Author Contributions: István B. Németh conceived the study; Lajos Kemény and István B. Németh supervised the entire project; Zsuzsanna Kurgyis, Lajos V. Kemény and István B. Németh designed the experiments, analyzed the data, and generated the ﬁgures with the help of Gergely Groma, Thomas Dittmar, Lajos Kemény; István B. Németh, 7 of 9 Int. J. Mol. Sci. 2016, 17, 980 Zsuzsanna Kurgyis and Lajos V. Kemény performed the experiments supported by Tünde Buknicz, Ádám Jakab, Hilda Polyánka; István B. Németh, Gergely Groma, Thomas Dittmar and Kurt Zänker provided critical analysis support; Judit Oláh provided the clinical data of the melanoma patients; Zsuzsanna Kurgyis and Lajos V. Kemény wrote the manuscript. Conﬂicts of Interest: The authors declare no conﬂict of interest. References 1. Hocevar, M.; Dragonja, Z.; Pilko, G.; Gazic, B.; Zgajnar, J. Residual melanoma after an excisional biopsy is an independent prognostic factor for local recurrence and overall survival. Eur. J. Surg. Oncol. 2014, 40, 1271–1275. [CrossRef] [PubMed] 1. Hocevar, M.; Dragonja, Z.; Pilko, G.; Gazic, B.; Zgajnar, J. Residual melanoma after an excisional biopsy is an independent prognostic factor for local recurrence and overall survival. Eur. J. Surg. Oncol. 2014, 40, 1271–1275. [CrossRef] [PubMed] 2. Thomas, J.M.; Newton-Bishop, J.; A’Hern, R.; Coombes, G.; Timmons, M.; Evans, J.; Cook, M.; Theaker, J.; Fallowﬁeld, M.; O’Neill, T.; et al. Excision margins in high-risk malignant melanoma. N. Engl. J. Med. 2004, 350, 757–766. [CrossRef] [PubMed] 2. Thomas, J.M.; Newton-Bishop, J.; A’Hern, R.; Coombes, G.; Timmons, M.; Evans, J.; Cook, M.; Theaker, J.; Fallowﬁeld, M.; O’Neill, T.; et al. Excision margins in high-risk malignant melanoma. N. Engl. J. Med. 2004, 350, 757–766. [CrossRef] [PubMed] 3. Polyak, K.; Weinberg, R.A. Transitions between epithelial and mesenchymal states: Acquisition of malignant and stem cell traits. Nat. Rev. Cancer 2009, 9, 265–273. [CrossRef] [PubMed] 3. Polyak, K.; Weinberg, R.A. Transitions between epithelial and mesenchymal states: Acquisition of malignant and stem cell traits. Nat. Rev. Cancer 2009, 9, 265–273. [CrossRef] [PubMed] 4. Landsberg, J.; Kohlmeyer, J.; Renn, M.; Bald, T.; Rogava, M.; Cron, M.; Fatho, M.; Lennerz, V.; Wölfel, T.; Hölzel, M.; et al. Melanomas resist T-cell therapy through inﬂammation-induced reversible dedifferentiation. Nature 2012, 490, 412–416. [CrossRef] [PubMed] 4. Landsberg, J.; Kohlmeyer, J.; Renn, M.; Bald, T.; Rogava, M.; Cron, M.; Fatho, M.; Lennerz, V.; Wölfel, T.; Hölzel, M.; et al. Melanomas resist T-cell therapy through inﬂammation-induced reversible dedifferentiation. Nature 2012, 490, 412–416. [CrossRef] [PubMed] 5. Giancotti, F.G. Mechanisms governing metastatic dormancy and reactivation. Cell 2013, 155, 750–764. [CrossRef] [PubMed] 5. Giancotti, F.G. Mechanisms governing metastatic dormancy and reactivation. Cell 2013, 155, 750–764. [CrossRef] [PubMed] 6. Thiery, J.P.; Acloque, H.; Huang, R.Y.J.; Nieto, M.A. Epithelial-mesenchymal transitions in development and disease. Cell 2009, 139, 871–890. [CrossRef] [PubMed] 6. Thiery, J.P.; Acloque, H.; Huang, R.Y.J.; Nieto, M.A. Epithelial-mesenchymal transitions in development and disease. Cell 2009, 139, 871–890. [CrossRef] [PubMed] 7. Easwaran, H.; Tsai, H.-C.; Baylin, S.B. Cancer epigenetics: Tumor heterogeneity, plasticity of stem-like states, and drug resistance. Mol. Cell 2014, 54, 716–727. [CrossRef] [PubMed] 7. Easwaran, H.; Tsai, H.-C.; Baylin, S.B. Cancer epigenetics: Tumor heterogeneity, plasticity of stem-like states, and drug resistance. Mol. Cell 2014, 54, 716–727. [CrossRef] [PubMed] 8. References Itakura, E.; Huang, R.-R.; Wen, D.-R.; Cochran, A.J. “Stealth” melanoma cells in histology-negative sentinel lymph nodes. Am. J. Surg. Pathol. 2011, 35, 1657–1665. [CrossRef] [PubMed] 21. Satzger, I.; Völker, B.; Meier, A.; Schenck, F.; Kapp, A.; Gutzmer, R. Prognostic signiﬁcance of isolated HMB45 or Melan A positive cells in Melanoma sentinel lymph nodes. Am. J. Surg. Pathol. 2007, 31, 1175–1180. [CrossRef] [PubMed] 22. Murali, R.; DeSilva, C.; McCarthy, S.W.; Thompson, J.F.; Scolyer, R.A. Sentinel lymph nodes containing very small (<0.1 mm) deposits of metastatic melanoma cannot be safely regarded as tumor-negative. Ann. Surg. Oncol. 2012, 19, 1089–1099. [PubMed] 23. Murali, R.; Thompson, J.F.; Shaw, H.M.; Scolyer, R.A. The prognostic signiﬁcance of isolated immunohistochemically positive cells in sentinel lymph nodes of melanoma patients. Am. J. Surg. Pathol. 2008, 32, 1106–1108. [CrossRef] [PubMed] 24. Ulmer, A.; Dietz, K.; Hodak, I.; Polzer, B.; Scheitler, S.; Yildiz, M.; Czyz, Z.; Lehnert, P.; Fehm, T.; Hafner, C.; et al. Quantitative measurement of melanoma spread in sentinel lymph nodes and survival. PLoS Med. 2014, 11, e1001604. [CrossRef] [PubMed] 25. Ng, A.K.; Jones, W.O.; Shaw, J.H. Analysis of local recurrence and optimizing excision margins for cutaneous melanoma. Br. J. Surg. 2001, 88, 137–142. [CrossRef] [PubMed] 26. Hudson, L.E.; Maithel, S.K.; Carlson, G.W.; Rizzo, M.; Murray, D.R.; Hestley, A.C.; Delman, K.A. 1 or 2 cm margins of excision for T2 melanomas: Do they impact recurrence or survival? Ann. Surg. Oncol. 2013, 20, 346–351. [CrossRef] [PubMed] 27. Pasquali, S.; Haydu, L.E.; Scolyer, R.A.; Winstanley, J.B.; Spillane, A.J.; Quinn, M.J.; Saw, R.P.M.; Shannon, K.F.; Stretch, J.R.; Thompson, J.F. The importance of adequate primary tumor excision margins and sentinel node biopsy in achieving optimal locoregional control for patients with thick primary melanomas. Ann. Surg. 2013, 258, 152–157. [CrossRef] [PubMed] 28. Fisher, K.E.; Cohen, C.; Siddiqui, M.T.; Palma, J.F.; Lipford, E.H.; Longshore, J.W. Accurate detection of BRAF p.V600E mutations in challenging melanoma specimens requires stringent immunohistochemistry scoring criteria or sensitive molecular assays. Hum. Pathol. 2014, 45, 2281–2293. [CrossRef] [PubMed] 29. Long, G.V.; Wilmott, J.S.; Capper, D.; Preusser, M.; Zhang, Y.E.; Thompson, J.F.; Kefford, R.F.; von Deimling, A.; Scolyer, R.A. Immunohistochemistry is highly sensitive and speciﬁc for the detection of V600E BRAF mutation in melanoma. Am. J. Surg. Pathol. 2013, 37, 61–65. [CrossRef] [PubMed] 30. Rachkovsky, M.; Sodi, S.; Chakraborty, A.; Avissar, Y.; Bolognia, J.; McNiff, J.M.; Platt, J.; Bermudes, D.; Pawelek, J. Melanoma ˆ macrophage hybrids with enhanced metastatic potential. Clin. Exp. References Stoecklein, N.H.; Klein, C.A. Genetic disparity between primary tumours, disseminated tumour cells, and manifest metastasis. Int. J. Cancer 2010, 126, 589–598. [CrossRef] [PubMed] 9. Harkness, T.; Weaver, B.A.; Alexander, C.M.; Ogle, B.M. Cell fusion in tumor development: Accelerated genetic evolution. Crit. Rev. Oncog. 2013, 18, 19–42. [CrossRef] [PubMed] 10. Clawson, G.A. Cancer. Fusion for moving. Science 2013, 342, 699–700. [CrossRef] [PubMed] 11. Duelli, D.; Lazebnik, Y. Cell fusion: A hidden enemy? Cancer Cell 2003, 3, 445–448. [CrossRef] 12. Rappa, G.; Mercapide, J.; Lorico, A. Spontaneous formation of tumorigenic hybrids between breast cancer and multipotent stromal cells is a source of tumor heterogeneity. Am. J. Pathol. 2012, 180, 2504–2515. [CrossRef] [PubMed] 13. Shabo, I.; Midtbö, K.; Andersson, H.; Åkerlund, E.; Olsson, H.; Wegman, P.; Gunnarsson, C.; Lindström, A. Macrophage traits in cancer cells are induced by macrophage-cancer cell fusion and cannot be explained by cellular interaction. BMC Cancer 2015, 15. [CrossRef] [PubMed] 14. Terada, N.; Hamazaki, T.; Oka, M.; Hoki, M.; Mastalerz, D.M.; Nakano, Y.; Meyer, E.M.; Morel, L.; Petersen, B.E.; Scott, E.W. Bone marrow cells adopt the phenotype of other cells by spontaneous cell fusion. Nature 2002, 416, 542–545. [CrossRef] [PubMed] 15. Kemény, L.V.; Kurgyis, Z.; Buknicz, T.; Groma, G.; Jakab, Á.; Zänker, K.; Dittmar, T.; Kemény, L.; Németh, I.B. Melanoma cells can adopt the phenotype of stromal ﬁbroblasts and macrophages by spontaneous cell fusion in vitro. Int. J. Mol. Sci. 2016, 17. [CrossRef] [PubMed] 16. Clawson, G.A.; Matters, G.L.; Xin, P.; Imamura-Kawasawa, Y.; Du, Z.; Thiboutot, D.M.; Helm, K.F.; Neves, R.I.; Abraham, T. Macrophage-tumor cell fusions from peripheral blood of melanoma patients. PLoS ONE 2015, 10, e0134320. [CrossRef] [PubMed] 17. Xue, J.; Zhu, Y.; Sun, Z.; Ji, R.; Zhang, X.; Xu, W.; Yuan, X.; Zhang, B.; Yan, Y.; Yin, L.; et al. Tumorigenic hybrids between mesenchymal stem cells and gastric cancer cells enhanced cancer proliferation, migration and stemness. BMC Cancer 2015, 15. [CrossRef] [PubMed] 8 of 9 Int. J. Mol. Sci. 2016, 17, 980 18. He, X.; Li, B.; Shao, Y.; Zhao, N.; Hsu, Y.; Zhang, Z.; Zhu, L. Cell fusion between gastric epithelial cells and mesenchymal stem cells results in epithelial-to-mesenchymal transition and malignant transformation. BMC Cancer 2015, 15. [CrossRef] [PubMed] 19. Mocellin, S.; Hoon, D.S.B.; Pilati, P.; Rossi, C.R.; Nitti, D. Sentinel lymph node molecular ultrastaging in patients with melanoma: A systematic review and meta-analysis of prognosis. J. Clin. Oncol. 2007, 25, 1588–1595. [CrossRef] [PubMed] 20. © 2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC-BY) license (http://creativecommons.org/licenses/by/4.0/). Espejel, S.; Romero, R.; Alvarez-Buylla, A. Radiation damage increases Purkinje neuron heterokaryons in neonatal cerebellum. Ann. Neurol. 2009, 66, 100–109. [CrossRef] [PubMed] References Metastasis 1998, 16, 299–312. [CrossRef] [PubMed] 31. Nagler, C.; Hardt, C.; Zänker, K.S.; Dittmar, T. Co-cultivation of murine BMDCs with 67NR mouse mammary carcinoma cells give rise to highly drug resistant cells. Cancer Cell Int. 2011, 11. [CrossRef] [PubMed] 32. Wang, R.; Sun, X.; Wang, C.Y.; Hu, P.; Chu, C.-Y.; Liu, S.; Zhau, H.E.; Chung, L.W.K. Spontaneous cancer-stromal cell fusion as a mechanism of prostate cancer androgen-independent progression. PLoS ONE 2012, 7, e42653. [CrossRef] [PubMed] 33. Chakraborty, A.K.; Sodi, S.; Rachkovsky, M.; Kolesnikova, N.; Platt, J.T.; Bolognia, J.L.; Pawelek, J.M. A Spontaneous murine melanoma lung metastasis comprised of host ˆ tumor hybrids. Cancer Res. 2000, 60, 2512–2519. [PubMed] 34. Lazova, R.; LaBerge, G.S.; Duvall, E.; Spoelstra, N.; Klump, V.; Sznol, M.; Cooper, D.; Spritz, R.A.; Chang, J.T.; Pawelek, J.M. A Melanoma brain metastasis with a donor-patient hybrid genome following bone marrow transplantation: First evidence for fusion in human cancer. PLoS ONE 2013, 8, e66731. [CrossRef] [PubMed] 35. Yilmaz, Y.; Lazova, R.; Qumsiyeh, M.; Cooper, D.; Pawelek, J. Donor Y chromosome in renal carcinoma cells of a female BMT recipient: visualization of putative BMT-tumor hybrids by FISH. Bone Marrow Transplant. 2005, 35, 1021–1024. [CrossRef] [PubMed] 9 of 9 Int. J. Mol. Sci. 2016, 17, 980 36. Nygren, J.M.; Liuba, K.; Breitbach, M.; Stott, S.; Thorén, L.; Roell, W.; Geisen, C.; Sasse, P.; Kirik, D.; Björklund, A.; et al. Myeloid and lymphoid contribution to non-haematopoietic lineages through irradiation-induced heterotypic cell fusion. Nat. Cell Biol. 2008, 10, 584–592. [CrossRef] [PubMed] 7. Espejel, S.; Romero, R.; Alvarez-Buylla, A. Radiation damage increases Purkinje neuron heterokaryon neonatal cerebellum. Ann. Neurol. 2009, 66, 100–109. [CrossRef] [PubMed] 38. Johansson, C.B.; Youssef, S.; Koleckar, K.; Holbrook, C.; Doyonnas, R.; Corbel, S.Y.; Steinman, L.; Rossi, F.M.V.; Blau, H.M. Extensive fusion of haematopoietic cells with Purkinje neurons in response to chronic inﬂammation. Nat. Cell Biol. 2008, 10, 575–583. [CrossRef] [PubMed] © 2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC-BY) license (http://creativecommons.org/licenses/by/4.0/).
https://openalex.org/W4210736544	https://ieeexplore.ieee.org/ielx7/6287639/9668973/09694614.pdf	English	null	Data-Driven Extraction of Uniformly Stable and Passive Parameterized Macromodels	IEEE access	2,022	cc-by	14,913	IEEE ELECTRONICS PACKAGING SOCIETY SECTION Received December 23, 2021, accepted January 13, 2022, date of publication January 27, 2022, date of current version February 11, 2022. eceived December 23, 2021, accepted January 13, 2022, date of Digital Object Identifier 10.1109/ACCESS.2022.3147034 The associate editor coordinating the review of this manuscript and approving it for publication was Ravi Mahajan. Data-Driven Extraction of Uniformly Stable and Passive Parameterized Macromodels TOMMASO BRADDE , (Member, IEEE), STEFANO GRIVET-TALOCIA , (Fellow, IEEE), ALESSANDRO ZANCO , (Member, IEEE), AND GIUSEPPE C. CALAFIORE , (Fellow, IEEE) Department of Electronics and Telecommunications, Politecnico di Torino, 10129 Torino, Italy Corresponding author: Tommaso Bradde (tommaso.bradde@polito.it) TOMMASO BRADDE , (Member, IEEE), STEFANO GRIVET-TALOCIA , (Fellow, IEEE), ALESSANDRO ZANCO , (Member, IEEE), AND GIUSEPPE C. CALAFIORE , (Fellow, IEEE) Department of Electronics and Telecommunications, Politecnico di Torino, 10129 Torino, Italy Corresponding author: Tommaso Bradde (tommaso.bradde@polito.it) TOMMASO BRADDE , (Member, IEEE), STEFANO GRIVET-TALOCIA , (Fellow, IEEE), ALESSANDRO ZANCO , (Member, IEEE), AND GIUSEPPE C. CALAFIORE , (Fellow, IEEE) Department of Electronics and Telecommunications, Politecnico di Torino, 10129 Torino, Italy Corresponding author: Tommaso Bradde (tommaso.bradde@polito.it) ABSTRACT A robust algorithm for the extraction of reduced-order behavioral models from sampled frequency responses is proposed. The system under investigation can be any Linear and Time Invariant structure, although the main emphasis is on devices that are relevant for Signal and Power Integrity and RF design, such as electrical interconnects and integrated passive components. We assume that the device under modeling is parameterized by one or more design variables, which can be related to geometry or materials. Therefore, we seek for multivariate macromodels that reproduce the dynamic behavior over a predeﬁned frequency band, with an explicit embedded dependence of the model equations on these external parameters. Such parameterized macromodels may be used to construct component libraries and prove very useful in fast system-level numerical simulations in time or frequency domain, including optimization, what- if, and sensitivity analysis. The main novel contribution is the formulation of a ﬁnite set of convex constraints that are applied during model identiﬁcation, which provide sufﬁcient conditions for uniform model stability and passivity throughout the parameter space. Such constraints are characterized by an explicit control allowing for a trade-off between model accuracy and runtime, thanks to some special properties of Bernstein polynomials. In summary, we propose a method to systematically address the longstanding problem of multivariate stability and passivity enforcement in data-driven model order reduction, which insofar has been tackled only via either over-conservative or heuristic and possibly unreliable methods. INDEX TERMS Passive macromodeling, reduced-order modeling, parameterized modeling, data-driven model order reduction, Bernstein polynomials, linear matrix inequalities. under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/ VOLUME 10, 202 IEEE ELECTRONICS PACKAGING SOCIETY SECTION This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0 I. INTRODUCTION Two main problems affect this strategy: ﬁrst, identiﬁcation of passivity violations in a multivariate setting requires sampling, so that it is possible that some passivity violations are not identiﬁed and thus not removed; second, removal of a passivity violation requires the solution of a nonlinear optimization problem, whose linearization during iterations may lead to further loss of accuracy and possibly lack of convergence. ports are deﬁned, a ﬁnite-bandwidth characterization of the system is obtained in terms of samples associated with a speciﬁed network function, typically in the scattering representation. A rational ﬁtting process [7]–[10] is then performed over this available data in order to obtain a closed form expression which best explains the associated input-output behavior. The resulting rational model of the network function is then synthesized into an equivalent and low-complexity SPICE netlist, that can be exploited to perform fast numerical simulations at the system level. The above procedure can be extended by requiring that the macromodel mimics the behavior of the structure for different conﬁgurations of a set of design or physical parameters, whose value is not ﬁxed a priori but is known to belong to a prescribed set. In such a case, the network function data are sampled in correspondence of a ﬁnite number of parameters conﬁgurations, and a multivariate modeling strategy is pursued to obtain a parameterized macromodel that can replace the original structure for all of the parameters conﬁgurations of interest [11]–[15]. This work proposes a novel constrained multivariate ratio- nal ﬁtting framework, that overcomes all above limitations and drawbacks. The approach can be summarized as follows. 1) The model structure is deﬁned as a ratio between a matrix numerator and a scalar denominator, which are both expanded into a partial fraction basis (with stable basis poles) along frequency and multivariate Bernstein polynomials in the parameter space. When the macromodel is to be used within system-level simulations, it is crucial that the associated equivalent circuit preserves some fundamental structural properties of the true device. In particular, passive components and interconnects, which are unable to generate energy on their own, must be represented by certiﬁed passive macromodels [6], [16]. Otherwise, the models can be the root cause of spurious insta- bilities, thus impairing the entire modeling and simulation workﬂow. 2) Stability conditions along frequency are expressed as a (continuously) parameterized Kalman-Yakubovick- Popov (KYP) linear matrix inequality, which depends only on the model denominator. I. INTRODUCTION 3) A ﬁnite number of convex constraints providing a sufﬁcient condition for uniform stability is derived by expanding all terms of the above KYP condition in terms of Bernstein polynomials, and by exploiting some unique properties of such polynomials. While the generation of passive univariate (non- parameterized) macromodels can mostly be considered as a solved problem [17]–[19], this is not true for the more complex parameterized case, and the topic is still subject of active research. Many different approaches have been proposed in the literature to tackle the passive parameterized macromodeling problem, but an ultimate fast, efﬁcient and robust methodology is still not available. 4) Passivity conditions along frequency are expressed as a parameterized KYP (similarly to point 2 above), which depends only on the model numerator. 5) A ﬁnite number of convex constraints providing a sufﬁcient condition for uniform passivity is derived as in point 3, through a Bernstein polynomial expansion of the above KYP formulation. Among the available strategies, some [11], [12], [20], [21] rely on passivity-preserving interpolation schemes. These approaches aim at building multivariate macromodels by interpolating a set of ‘‘root’’ univariate macromodels constructed at several discrete parameter instances. In case the root models are passive, the use of passivity preserving interpolation schemes guarantees the uniform passivity of the resulting multivariate model. The simplicity of these approaches comes with some serious drawbacks. For instance, the resulting multivariate model may show nonphysical augmented complexity and/or incorrect model behavior for parameter combinations that are different from the training samples. 6) The conservativity introduced by the discretization in points 3 and 5 is reduced by exploiting a special degree elevation property of Bernstein polynomials. In the above list, points 1, 2 and 4 are reformulations of known results. Preliminary ideas based on point 5 have been recently published in [26] for parameterized macromodels including a single external parameter; expanding on such preliminary ideas, we present a full treatment of the theoret- ical derivations that legitimate the validity of points 3 and 5 when generating macromodels that include an arbitrary number of external parameters; the practical effectiveness of the proposed approach is further enhanced by introducing the conservativity reduction strategy of point 6. Other techniques waive these structural passivity pre- serving properties in favor of more compact parameterized macromodels, which are identiﬁed through a well-established multivariate rational ﬁtting procedure [14]. I. INTRODUCTION the behavior of a given system with minimal computational efforts, by accurately reproducing those physical quantities that are of interest within a speciﬁc simulation, i.e. the required system outputs [1]–[3]. The intrinsic complexity of the ﬁrst-principle physical laws (e.g. Maxwell’s equations) is reduced to a small set of explanatory instrumental variables that are sufﬁcient to predict the input-output relationship of interest. In particular, in the ﬁeld of electronics manufacturing, the enormous complexity of state-of-the-art devices is such that behavioral models ﬁnd major room for practical exploitation, in particular when dealing with passive electromagnetic devices [4]–[6] such as electrical interconnects and integrated components. Mathematical modeling is a cornerstone for modern tech- nological advancement and industrial manufacturing. The possibility of accurately predicting the behavior of a given design allows engineers to perform preliminary testing and veriﬁcation stages without relying on the construction of physical prototypes, which is highly consuming in terms of strategical assets. In this view, the industrial interest for mathematical models is moved not only by their effectiveness in predicting physical phenomena, but also by their potential for saving resources in terms of manpower and time-to- market. For this reason, reduced order models or ‘‘macromodels’’ gained an increasing importance in the ﬁeld of Computer Aided Design. The rationale behind such models is to predict The generation of a macromodel is usually performed by following a well established workﬂow. A physical model for the device is ﬁrst instantiated within a CAD environment able to provide a highly detailed description of the structure based on the Maxwell’s equations. Once the interface electrical 15786 VOLUME 10, 2022 VOLUME 10, 2022 T. Bradde et al.: Data-Driven Extraction of Uniformly Stable and Passive Parameterized Macromodels sampling in the parameter space [22], [23], or by imposing some sign properties in the model coefﬁcients [24]. The former approach may miss small stability violations due to the ﬁnite number of constraints that can be constructed, while the latter is known to be over-conservative and may lead to a model with reduced accuracy. In this framework, passivity enforcement is usually performed by post-processing, so that the model is iteratively perturbed until all the passivity violations are removed [25]. I. INTRODUCTION Uniform stability can be enforced by embedding some constraints in the ﬁtting process. These constraints can be either based on (adaptive) We remark that the effectiveness of the constraints dis- cretization strategy based on Bernstein polynomials involved 15787 15787 VOLUME 10, 2022 T. Bradde et al.: Data-Driven Extraction of Uniformly Stable and Passive Parameterized Macromodels in points 3 and 5 has been also recently discussed in [27] in a more general setting. sum operation j + k between two multi-indices returns a multi-index i = (j1 + k1, . . . , jd + kd). The max(j, k) [resp. min(j, k)] function returns the component-wise maximum (resp. minimum) for each entry of its arguments. In this context, we also deﬁne the multi-index binomial coefﬁcient From the computational standpoint, the proposed method consists of • an iterative low-complexity least squares identiﬁcation of the denominator of the model, where special Linear Matrix Inequality (LMI) constraints are used to enforce uniform stability; j k = dY i=1 ji ki . (1) (1) • a single higher-complexity LMI-constrained least squares problem for the identiﬁcation of the numer- ator (matrix) coefﬁcients, which guarantees uniform model passivity. Let p(ϑ) : Rd →R be a generic multivariate polynomial in d variables ϑ = (ϑ1 . . . , ϑd). In particular, we make extensive use of Bernstein polynomials. For d = 1, Let p(ϑ) : Rd →R be a generic multivariate polynomial in d variables ϑ = (ϑ1 . . . , ϑd). In particular, we make extensive use of Bernstein polynomials. For d = 1, b ¯ℓ ℓ(ϑ) = ¯ℓ ℓ ϑℓ(1 −ϑ) ¯ℓ−ℓ, ℓ= 0, . . . , ¯ℓ (2) The resulting algorithm is thus fully deterministic and robust, since it does not rely on sampling and is based on a convex formulation which is solved in ﬁnite time using standard optimization software. (2) deﬁnes the ℓ-th Bernstein polynomial of degree ¯ℓin the scalar variable ϑ. For d > 1, the ℓ-th multivariate Bernstein polynomial of multi-degree ℓis deﬁned as The proposed approach has a single main limitation, in terms of the overall complexity of the models that can be processed. The scalability analysis and the numerical examples that follow show that only small and medium-scale models are tractable, thus providing an applicability limit of proposed framework. I. INTRODUCTION This limit is in fact common to all applications that are based on LMI constraints, not only in model order reduction but also in the more general ﬁeld of modern control system engineering and numerical linear algebra. bℓ ℓ(ϑ) = b ¯ℓ1 ℓ1(ϑ1) × · · · × b ¯ℓd ℓd (ϑd) (3) (3) where ℓ= ( ¯ℓ1, . . . , ¯ℓd) (4) (4) is the multi-index collecting the degrees of the polynomials in each individual variable ϑk. For a given set of maximum degrees ℓ, we deﬁne the associated set of admissible indices as This paper is organized as follows. Section II introduces some notation and general facts, considered as preliminaries and background. Section III states the main considered problem, deﬁnes the adopted model structure, and recalls the existing model identiﬁcation methods; also this section is to be regarded as background material. Sections IV and V provide a complete derivation of proposed uniform stability and passivity conditions, respectively. Together with Sec. VI dedicated to the reduction of conservativity, they form the key novel material of this work. Section VII presents a set of numerical results together with discussion on performance and assessment of applicability limits. Conclusions are ﬁnally drawn in section VIII. as Iℓ= {ℓ∈Nd : ℓ≤ℓ}. (5) (5) For any multivariate polynomial matrix function in the Bernstein basis Bernstein basis F(ϑ) : Rd →Rm×n = X ℓ∈Iℓ Fℓbℓ ℓ(ϑ) (6) (6) we denote as control points the elements of the set {Fℓ: ℓ∈ Iℓ}, in short {Fℓ}. We will exploit some notable properties of multivariate Bernstein polynomials, which we report here following [28]. First, we recall that such polynomials are non-negative and provide a partition of unity for any generic dimension d and maximum degree ℓ II. PRELIMINARIES AND NOTATION C. MODEL IDENTIFICATION The model coefﬁcients are found by minimizing the model-data error according to ﬁtting condition (12), which is solved through a sequence of linear least squares problems based on the linearized approximation ˜Hk,m = ˜H(jωk, ϑm), k = 1, . . . ¯k, m = 1, . . . , ¯m. (11) (11) Nµ(jωk, ϑm) −Dµ(jωk, ϑm) ˜Hk,m Dµ−1(jωk, ϑm) ≈0, k = 1, . . . ¯k, m = 1, . . . , ¯m (15) Most commonly, such data samples are available as the scattering matrix of the reference device. When the structure under modeling is known to be passive, it is assumed that the data samples are compliant with the appropriate passivity conditions. It is also assumed that the available samples are sufﬁcient to characterize the variations of the frequency responses over the target frequency band: indeed, any information that is not embedded in the input data cannot be reproduced by any model constructed using such data. (15) where µ = 1, 2, . . . is the iteration index and Dµ−1 is known at each iteration µ since based on estimates of the denomina- tor coefﬁcients ri,ℓat the previous iteration µ −1. The ﬁrst iteration is initialized with D0(jω, ϑ) = 1. Condition (15) is equivalent to (12) whenever Dµ(jω, ϑ) = Dµ−1(jω, ϑ), which represents a convergence condition. The goal of parameterized macromodeling is to synthesize a reduced-order rational model with a transfer function H(s, ϑ) that matches the set of input training data p g All the conditions (15) can be collected i All the conditions (15) can be collected in a compact form 9µ x 9µ y xµ yµ ≈0 (16) (16) H(jωk, ϑm) ≈˜Hk,m, k = 1, . . . ¯k, m = 1, . . . , ¯m. (12) (12) where vectors xµ, yµ collect the numerator and denominator coefﬁcients Ri,ℓand ri,ℓ, respectively, and 9µ x and 9µ y are constant (iteration-dependent) regressor matrix blocks. System (16) is solved in least squares sense, suitably complemented by a non-triviality constraint to rule out the all-zero solution [30]. Due to the particular structure of the regression matrix entering problem (16), each iteration can The rational structure of the model allows for a straightfor- ward conversion of the transfer function into a parameterized equivalent circuit of reduced order that can be exploited within off-the-shelf SPICE environments (not discussed here, see e.g. A. GENERAL SETTING We consider a generic P-port Linear and Time-Invariant (LTI) system, whose behavior depends on d real-valued physical or design parameters. Without loss of generality, the parameter vector ϑ = (ϑ1, . . . , ϑd) is assumed to belong to a normalized d-dimensional hypercube 2 = [0, 1]1 × . . . × [0, 1]d, called the design space. Denoting with ˜H(s, ϑ) the P×P transfer function of the system, we assume that a highly detailed ﬁrst-principle model is available, which can be used to evaluate the system response through its frequency-domain samples over a ﬁnite bandwidth of interest and for any given combination of the design variables II. PRELIMINARIES AND NOTATION In the following, we denote with N, R, and C, the ﬁelds of natural, real and complex numbers, respectively. The symbol s is reserved for the Laplace variable, and j = √ −1 is the imaginary unit. Scalars are denoted with a plain lowercase font x, while uppercase fonts denote matrices X, whose size is speciﬁed if not clear from the context. Matrix transpose and Hermitian transpose are denoted with X⊤ and X⋆, respectively. The set of symmetric matrices of size n is denoted as Sn; accordingly, S− n denotes the cone of negative semi-deﬁnite matrices of size n. A given transfer function is denoted as H(s), and IP is reserved for the identity matrix of size P. bℓ ℓ(ϑ) ≥0 ∀ℓ∈Iℓ, X ℓ∈Iℓ bℓ ℓ(ϑ) = 1, ∀ϑ ∈2. (7) (7) These two properties imply that all the values attained by functions (6) are obtained via a convex combination of the control points {Fℓ}. A multivariate Bernstein polynomial of degree ℓcan be equivalently rewritten in terms of another Bernstein polynomial of higher degree, exploiting the so-called degree elevation property. Consider a Bernstein polynomial p(ϑ) of degree ℓand a degree increment e = (e1, . . . , e2) ≥0. Then, p(ϑ) admits the two following equivalent representations We deﬁne a multi-index as a d-dimensional collection of indices i = (i1, . . . , id) ∈Nd. Given two multi-indices j and k, we write j ≤k meaning j1 ≤k1, . . . , jd ≤kd. The p(ϑ) = X ℓ∈Iℓ pℓbℓ ℓ(ϑ) = X γ ∈Iγ gγ bγ γ (ϑ), γ = ℓ+ e (8) (8) 15788 VOLUME 10, 2022 VOLUME 10, 2022 T. Bradde et al.: Data-Driven Extraction of Uniformly Stable and Passive Parameterized Macromodels where the set of coefﬁcients gγ are obtained as convex combinations of the original coefﬁcients pℓ B. MODEL STRUCTURE Consistently with most of the existing literature on this subject [14], [29], we assume the following Parameterized- Sanathanan-Koerner (PSK) model structure gγ = X s∈S ℓ s e γ −s ℓ+e γ ps, γ ∈Iγ , (9) (9) H(s, ϑ) = N(s, ϑ) D(s, ϑ) = P¯n i=0 P ℓ∈IℓRi,ℓbℓ ℓ(ϑ) ϕi(s) P¯n i=0 P ℓ∈Iℓri,ℓbℓ ℓ(ϑ) ϕi(s) , (13) (13) where the sum is performed over the set of multi-indices S = {s : s = max(0, γ −e), . . . , min(ℓ, γ )}. Finally, given two Bernstein polynomials p(ϑ), g(ϑ) having total degree ℓ and γ respectively, their product h(ϑ) can be expressed as a polynomial of total degree β = ℓ+ γ , whose coefﬁcients hβ are where the basis functions ϕi(s) are constructed from a set of predeﬁned poles {q1, . . . , q¯n} with ℜ{qi} < 0 ∀i as where the basis functions ϕi(s) are constructed from a set of predeﬁned poles {q1, . . . , q¯n} with ℜ{qi} < 0 ∀i as    ϕi(s) = (s −qi)−1, qi ∈R ϕi(s) = [(s −qi)−1 + (s −q⋆ i )−1] qi ∈C ϕi+1(s) = j[(s −qi)−1 −(s −q⋆ i )−1] qi+1 = q⋆ i ∈C are hβ = X s∈S ℓ s γ β−s ℓ+γ β psgβ−s, β ∈Iβ (10) ϕ + j q qi q + qi (14) (14) (10) with ϕ0(s) = 1. Therefore, both N(s, ϑ) and D(s, ϑ) are stable rational functions of the Laplace variable s, sharing the same set of poles. The Bernstein bases are exploited to parameterize N(s, ϑ) and D(s, ϑ) via the unknown model coefﬁcients ri,ℓ∈R and Ri,ℓ∈RP×P. As these two transfer functions share the same set of common poles, the zeros and the poles of H(s, ϑ) coincide with the zeros of N(s, ϑ) and D(s, ϑ) respectively. The PSK model structure (13) thus provides a parameterization of both zeros and poles of each individual model response. Note that the basis poles qi cancel out in (13) and are not poles of the model. They are only instrumental for the deﬁnition of the barycentric basis functions ϕi(s) upon which the model structure is constructed. with ϕ0(s) = 1. Therefore, both N(s, ϑ) and D(s, ϑ) are stable rational functions of the Laplace variable s, sharing the same set of poles. B. MODEL STRUCTURE The Bernstein bases are exploited to parameterize N(s, ϑ) and D(s, ϑ) via the unknown model coefﬁcients ri,ℓ∈R and Ri,ℓ∈RP×P. As these two transfer functions share the same set of common poles, the zeros and the poles of H(s, ϑ) coincide with the zeros of N(s, ϑ) and D(s, ϑ) respectively. The PSK model structure (13) thus provides a parameterization of both zeros and poles of each individual model response. Note that the basis poles qi cancel out in (13) and are not poles of the model. They are only instrumental for the deﬁnition of the barycentric basis functions ϕi(s) upon which the model structure is constructed. where S = {s : s = max(0, β −γ ), . . . , min(ℓ, β)}. E. STATE-SPACE REALIZATIONS is solved to ﬁnd the numerator unknowns xµ. See [31] for further details about the algorithmic aspects of this decoupling strategy. A state-space realization for the denominator transfer func- tion can be constructed as follows [14] Iterations are stopped when the denominator estimate stabilizes, i.e., when the following condition is met D(s, ϑ) ↔6D = A1 B1 C1(ϑ) D1(ϑ) , (21) (21) δµ = yµ −yµ−1 2 ∥yµ∥2 ≤ϵ (19) where the constant matrices A1, B1 are where the constant matrices A1, B1 are (19) A1 = blkdiag{A1,i} ∈R¯n×¯n (22) B1 = [. . . , B1,i, . . . ]⊤∈R¯n, (23) (22) (23) being ϵ a given threshold. Alternatively, the iteration is stopped if a maximum prescribed iteration number is reached. Notice that convergence criterion (19) does not involve the numerator unknowns xµ; therefore, the solution of (18) can be deferred to the last iteration, once (19) is met. being ϵ a given threshold. Alternatively, the iteration is stopped if a maximum prescribed iteration number is reached. Notice that convergence criterion (19) does not involve the numerator unknowns xµ; therefore, the solution of (18) can be deferred to the last iteration, once (19) is met. (23) with A1,i =    qi, qi ∈R " σi ωi −ωi σi # , qi = σi ± jωi ∈C (24) B1,i = ( 1, qi ∈R h 2 0 i , qi = σi ± jωi ∈C (25) (24) (17) (17) obtained by elimination of xµ in (16) through a QR decomposition. Once yµ is available (e.g. as the singular vector of 0µ y associated to its least singular value), the following system The solution of these two problems requires a set of state-space realizations for both model numerator and denominator, seen as individual transfer functions, which are introduced next. 9µ x xµ ≈−9µ y yµ (18) (18) D. PROBLEM STATEMENT The objective of this work is to guarantee that the parameter- ized model (13) is uniformly stable and possibly uniformly passive throughout the design space 2. We will achieve this goal by modifying the model identiﬁcation steps by adding a set of semi-deﬁnite constraints providing provable sufﬁcient conditions for uniform stability and passivity. We anticipate that uniform stability is achieved by constraining only the denominator estimate (17) and provides a necessary prerequisite to uniform passivity. The latter is controlled by constraining all model coefﬁcients. (25) Note that, by construction, the pair (A1, B1) is controllable and A1 is Hurwitz, as ℜ{qi} < 0 ∀i. The parameterized output matrices are available as Bernstein polynomials and read C1(ϑ) = X ℓ∈Iℓ Cℓ 1 bℓ ℓ(ϑ), Cℓ 1 = [r1,ℓ, . . . , r¯n,ℓ] ∈R1×¯n (26) (27) (26) (26) D1(ϑ) = X ℓ∈Iℓ Dℓ 1 bℓ ℓ(ϑ), Dℓ 1 = r0,ℓ∈R. (27) Let us recall the general conditions for the uniform passivity of a generic parameter-dependent scattering or immittance LTI system in terms of its transfer matrix (27) y 1) H(s, ϑ) regular for ℜ{s} > 0 ∀ϑ ∈2 2) H∗(s, ϑ) = H(s∗, ϑ) ∀s ∈C, ∀ϑ ∈2 3) 8(s, ϑ) ⪰0 for ℜ{s} > 0, ∀ϑ ∈2 where ∗denotes the complex conjugate, and y 1) H(s, ϑ) regular for ℜ{s} > 0 ∀ϑ ∈2 Using a similar construction, we can realize the numerator transfer function as follows 3) 8(s, ϑ) ⪰0 for ℜ{s} > 0, ∀ϑ ∈2 where ∗denotes the complex conjugate, and 3) 8(s, ϑ) ⪰0 for ℜ{s} > 0, ∀ϑ ∈2 where ∗denotes the complex conjugate, and 3) 8(s, ϑ) ⪰0 for ℜ{s} > 0, ∀ϑ ∈2 where ∗denotes the complex conjugate, and N(s, ϑ) ↔6N = A B C2(ϑ) D2(ϑ) (28) (28) 8(s, ϑ) = ( IP −H⋆(s, ϑ)H(s, ϑ) scattering, H⋆(s, ϑ) + H(s, ϑ) immittance. (20) (20) where A = IP ⊗A1 and B = IP ⊗B1 with ⊗denoting the matrix Kronecker product, and where Condition 1 is related to uniform stability, whereas the realness condition 2 is enforced by construction by adopted model structure (13). Condition 3 deﬁnes uniform dissipa- tivity in terms of Bounded Realness (in the scattering case) and Positive Realness (in the immittance case). Wihtout loss of generality, we will only consider the Bounded Realness conditions in the following, since Positive Realness can be achieved with a straightforward adaptation. C. MODEL IDENTIFICATION [6], [23]). 15789 VOLUME 10, 2022 VOLUME 10, 2022 T. Bradde et al.: Data-Driven Extraction of Uniformly Stable and Passive Parameterized Macromodels be split in two steps, which seek for xµ and yµ separately, in order to improve the algorithm efﬁciency. The ﬁrst step ﬁnds yµ by solving a reduced least-squares system of the form Problem 1: Derive a numerically viable approach to estimate the model coefﬁcients ri,ℓ, so that Condition 1 is fulﬁlled (uniform stability) Problem 2: Assuming uniform stability, derive a numeri- cally viable approach to estimate the model coefﬁcients Ri,ℓ, so that Condition 3 is fulﬁlled (uniform passivity). 0µ y yµ ≈0, (17) IV. UNIFORM STABILITY CONDITIONS This section presents a solution for Problem 1 and derives a set of algebraic and convex constraints providing a guaranteed uniform stability of the model. (P, Q, R) = P⊤R + RP RQ Q⊤R 0 (35) (35) Given the adopted model structure (13), stability is attained by constraining all the zeros of the denominator D(s, ϑ) to have a negative real part, since these zeros coincide with the parameter-dependent model poles. The denominator D(s, ϑ) satisﬁes by construction Conditions 1, 2 of Sec. III-D being a real rational and strictly stable function. If we are able to additionally enforce the dissipativity Condition 3, then D(s, ϑ) becomes a certiﬁed uniformly Positive Real function. Since any Positive Real function is also minimum phase [32] with stable zeros, we conclude that enforcing (31) guarantees indirectly the uniform stability of the model H(s, ϑ). Note that, under the working assumptions, Condition 3 can be replaced by the simpler condition for any triplet of matrices P, Q, R = R⊤with compat- ible size, and where L(ϑ) plays the role of an energy storage function (Lyapunov matrix). Notice that (34) is a parameterized version of the well-established Positive Real Lemma [35], in which the additional requirement L(ϑ) ⪰0 is automatically entailed by the fact that A1 is Hurwitz. Condition (34) is a robust Linear Matrix Inequality (LMI) condition on both the instrumental matrix L(ϑ) and the denominator coefﬁcients ri,ℓ, which enter the various matrix blocks according to (26)-(27). Although such conditions are convex, solving (17) for the denominator coefﬁcients while enforcing (34) for all ϑ ∈2 is still a computationally intractable task. We have removed dependence on frequency, but a continuous dependence of the constraints in the parameters ϑ remains. (31) D⋆(jω, ϑ) + D(jω, ϑ) ≥0 ∀ϑ ∈2, ∀ω, (31) This problem is addressed by restricting the class of the storage functions L(ϑ) to a ﬁnite-dimensional space. In particular, we adopt the following structure using an expansion into Bernstein polynomials since D(s, ϑ) is strictly stable and bounded for s →∞, without poles on the imaginary axis s = jω. This consideration has been extensively exploited in the literature to generate stable parameterized macromodels. From the numerical standpoint, (31) cannot be enforced directly, as it embeds an inﬁnite number of constraints that must be veriﬁed over the entire continuous frequency- parameter space. where we deﬁne the auxiliary block matrix where we deﬁne the auxiliary block matrix IV. UNIFORM STABILITY CONDITIONS IV. UNIFORM STABILITY CONDITIONS Its enforcement has been addressed either by discretization into a ﬁnite set via sampling-based strategies [23], or by deriving over-conservative sufﬁcient conditions on the sign of individual terms in the denominator expansion [24]. These two strategies are complementary from the point of view of the modeling performances: while the former guarantees high level of accuracy but does not provide a theoretical guarantee for uniform stability, the latter leads to provable stability and is very efﬁcient but can result in accuracy degradation due to the approximate nature of the applied constraints. In what follows, we propose a scheme that retains the advantages of both approaches. L(ϑ) = X ℓ∈Iℓ Lℓbℓ ℓ(ϑ), Lℓ∈S¯n ∀ℓ∈Iℓ (36) (36) based on a set of unknown symmetric matrix coefﬁcients {Lℓ}. Using (36), (34) becomes ∀ϑ ∈2, ∃Lℓ∈S¯n, ℓ∈Iℓ: S(ϑ) = X ℓ∈Iℓ Sℓbℓ ℓ(ϑ) ⪯0, (37) (37) with Sℓ= (A1, B1, Lℓ) − 0 Cℓ⊤ 1 Cℓ 1 2Dℓ 1 ∈S¯n+1. (38) (38) As all Bernstein polynomials are nonnegative, bℓ ℓ(ϑ) ≥ 0 ∀ℓ∈Iℓ, we see that (34) is implied by the following sufﬁcient conditions Let us consider the state-space realization (21) and deﬁne Z1(jω) = (jωIP −A1)−1B1. Condition (31) can be rewritten as Sℓ⪯0 ∀ℓ∈Iℓ (39) as (39) Z1(jω)⋆C1(ϑ)⊤+ C1(ϑ)Z1(jω) + 2D1(ϑ) ≥0, (32) that must hold ∀ϑ ∈2, ∀ω, or in the more compact matrix form Z1(jω)⋆C1(ϑ)⊤+ C1(ϑ)Z1(jω) + 2D1(ϑ) ≥0, (32) Z1(jω)⋆C1(ϑ)⊤+ C1(ϑ)Z1(jω) + 2D1(ϑ) ≥0, (32) that must hold ∀ϑ ∈2, ∀ω, or in the more compact matrix form (32) which can be exploited as a set of standard non-parameterized LMI constraints to be enforced during the estimation of the model denominator coefﬁcients. More precisely, our proposed solution for uniform stability enforcement can be cast as the following semi-deﬁnite program Z1(jω) IP ⋆ 0 −C1(ϑ)⊤ −C1(ϑ) −2D1(ϑ) Z1(jω) IP ≤0. (33) min yµ 0µyµ 2 s.t. Sℓ⪯0 ∀ℓ∈Iℓ (40) (40) Since the pair (A1, B1) is controllable, we can apply the Yakubovich lemma [33], [34] to cast this frequency domain inequality as the equivalent algebraic inequality replacing the unconstrained least squares problem (17). The optimization problem (40) is a semi-deﬁnite program that can be solved through off-the-shelf convex optimization solvers. ∀ϑ ∈2, ∃L(ϑ) ∈S¯n : (A1, B1, L(ϑ)) − 0 C1(ϑ)⊤ C1(ϑ) 2D1(ϑ) ⪯0, (34) Some remarks about the computational cost are in order. Let Card(Iℓ) = V. D. PROBLEM STATEMENT Condition 1 is related to uniform stability, whereas the realness condition 2 is enforced by construction by adopted model structure (13). Condition 3 deﬁnes uniform dissipa- tivity in terms of Bounded Realness (in the scattering case) and Positive Realness (in the immittance case). Wihtout loss of generality, we will only consider the Bounded Realness conditions in the following, since Positive Realness can be achieved with a straightforward adaptation. C2(ϑ) = X ℓ∈Iℓ Cℓ 2 bℓ ℓ(ϑ) Cℓ 2 ∈RP×¯nP, (29) D2(ϑ) = X ℓ∈Iℓ Dℓ 2 bℓ ℓ(ϑ) Dℓ 2 = R0,ℓ∈RP×P. (30) (29) (30) For ﬁxed ℓ, matrix Cℓ 2 collects the elements of the model coefﬁcients Ri,ℓ, i > 0 with a compatible ordering. The pair (A, B) inherits the controllability property from (A1, B1). All of the eigenvalues of A are stable. In summary, we will propose a solution to the following two problems: 15790 VOLUME 10, 2022 T. Bradde et al.: Data-Driven Extraction of Uniformly Stable and Passive Parameterized Macromodels V. UNIFORM PASSIVITY CONDITIONS Condition (47) is similar to (34), with the additional complication that the numerator unknowns xµ that are embedded in the parameterization of C2(ϑ), D2(ϑ) appear as quadratic terms in X2(ϑ). It is nonetheless possible to reformulate (47) as a LMI. First, we write X⊗(ϑ) as a sum of Bernstein polynomials of total degree m = 2ℓ This section proposes a solution to the Problem 2 of Sec. IV, which aims at enforcing uniform dissipativity of the model. In the assumed scattering representation, this is equivalent to enforcing the model response H(s, ϑ) to be Bounded Real throughout the design space 2. As Bounded Realness requires as a prerequisite the uniform stability, we assume that the coefﬁcients yµ of the model denominator have already been identiﬁed by solving the convex program (40). Therefore, this section will focus on the identiﬁcation of the coefﬁcients xµ of the model numerator, assuming as frozen the denominator coefﬁcients. X⊗(ϑ) = X m∈Im Xm bm m(ϑ) (48) (48) with symmetric matrix coefﬁcients Xm. Since the denomina- tor coefﬁcients ri,ℓare available, each Xm can be computed in closed form as the product of polynomials in Bernstein basis, using formula (10). Note that this expansion is exact. Second, we apply the inverse Schur complement to (47), obtaining the equivalent LMI condition Let us consider the non-expansivity Condition 3 of Sec. III-D, that we can equivalently restrict to the imaginary axis similarly to (31) as H⋆(jω, ϑ)H(jω, ϑ) ⪯IP ∀ω ∈R, ∀ϑ ∈2. (41) (41)   (A, B, P(ϑ)) −X⊗(ϑ) C2(ϑ) D2(ϑ)⊤ C2(ϑ) D2(ϑ) −IP  ⪯0 (49) Exploiting the model structure (13) provides the equivalent form Exploiting the model structure (13) provides the equivalent form (49) C2(ϑ) D2(ϑ) N⋆(jω, ϑ)N(jω, ϑ) −IP D⋆(jω, ϑ)D(jω, ϑ) ⪯0, ∀ω ∈R, ∀ϑ ∈2. (42) which must be veriﬁed ∀ϑ ∈ 2. Third, we apply the degree elevation property of Bernstein polynomials (8) to the off-diagonal blocks of (49). More precisely, we rewrite C2(ϑ) and D2(ϑ) as a sum of Bernstein polynomials of total degree m starting from their original degree-ℓexpansions (29)-(30), and we cast the result in compact form as (42) ∀ω ∈R, ∀ϑ ∈2. (42) We now replace numerator and denominator with their state- space realization, as in Sec. IV. IV. UNIFORM STABILITY CONDITIONS The solution of (40) involves (34) VOLUME 10, 2022 15791 T. Bradde et al.: Data-Driven Extraction of Uniformly Stable and Passive Parameterized Macromodels (¯n + 1)V unknown model coefﬁcients and V((¯n + 1)2 + ¯n + 1)/2 instrumental variables. The size of the regressor matrix is 0µ ∈RP2V ¯n×V ¯n, as explained in [31]. The size of the symmetric matrices involved in the constraints is ¯n + 1, and the number of matrix constraints is V. As one could expect, the proposed approach suffers from a curse of dimensionality when the dimension of the design space d increases, as both the number of denominator unknowns and the cardinality V of the admissible indices grows exponentially with d. However, as experimentally demonstrated in Section VII, the numerical solution of (40) requires affordable (desktop-level) computing resources when making use of state-of-the-art convex optimization solvers, at least for moderate dimension d (few units). the state matrices A and B are the same as for the numerator realization (28). Setting now Z(jω) = (jωI¯nP −A)−1B and using (43), (28), allows to cast condition (42) as (¯n + 1)V unknown model coefﬁcients and V((¯n + 1)2 + ¯n + 1)/2 instrumental variables. The size of the regressor matrix is 0µ ∈RP2V ¯n×V ¯n, as explained in [31]. The size of the symmetric matrices involved in the constraints is ¯n + 1, and the number of matrix constraints is V. As one could expect, the proposed approach suffers from a curse of dimensionality when the dimension of the design space d increases, as both the number of denominator unknowns and the cardinality V of the admissible indices grows exponentially with d. However, as experimentally demonstrated in Section VII, the numerical solution of (40) requires affordable (desktop-level) computing resources when making use of state-of-the-art convex optimization solvers, at least for moderate dimension d (few units). Z(jω) IP ⋆ X2(ϑ) −X⊗(ϑ) Z(jω) IP ⪯0, (45) (45) which must hold ∀ϑ ∈2 and ∀ω, and where we have deﬁned the auxiliary matrices Xν(ϑ) = C⊤ ν (ϑ) D⊤ ν (ϑ) Cν(ϑ) Dν(ϑ) , (46) (46) being the symbol ν a place-holder for the subscripts {2, ⊗}. IV. UNIFORM STABILITY CONDITIONS As the pair (A, B) is controllable, the application of Yakubovich lemma [34] translates (45) into the following equivalent parameterized algebraic condition We conclude this section by noting that both (36) and (39) do introduce some amount of conservativity in the formulation with respect to the continuously parameterized form (34). We will show in Sec. VI that this amount can be effectively controlled and reduced thanks to the degree elevation property of the Bernstein polynomials. ∀ϑ ∈2, ∃P(ϑ) ∈S¯nP : X2(ϑ) −X⊗(ϑ) + (A, B, P(ϑ)) ⪯0, (47) which can be interpreted as a reformulation of the classical Bounded Real lemma [35] for stable transfer functions which are parameterized according to the proposed model structure (13). V. UNIFORM PASSIVITY CONDITIONS Finally, since Bernstein polynomials are non-negative, we see that (49) is implied by the following set of semi-deﬁnite constraints Fm ⪯0, ∀m ∈Im. (54) (54) The above conditions can be easily incorporated within the numerator estimation procedure expressed by (18). This unconstrained least squares system in the unknown variables xµ is thus replaced by the following LMI-constrained convex program We consider the replacement of (52) with the discretized set (54). Let F be the set spanned by all matrices F(ϑ) as ϑ spans the design space 2. By construction, this set is embedded in the convex hull generated by all matrix coefﬁcients Fm min xµ 9µ x xµ + 9µ y yµ 2 s.t. Fm ⪯0, ∀m ∈Im. (55) (55) F = {x : x = F(ϑ), ϑ ∈2} ⊆Conv({Fm}). (56) (56) The numerical solution (55) provides a set of numerator coefﬁcients xµ that guarantee uniform passivity of the model. Due to the convex formulation with a ﬁnite number of constraints, this solution is attained in polynomial time with standard convex optimization solvers. The data matrices 9µ x,y entering (55) have a row size depending on the amount of training data samples. If the passive structure under modeling is reciprocal, then the problem involves M(¯n + 1)(P2/2 + P/2) unknown model coefﬁcients and M(P2¯n2/2 + P¯n/2) instrumental variables, where M = Card(Im). From these expressions we see that the complexity of the problem depends not only on the total number M of elements in the multivariate Bernstein basis used to represent the dissipativity constraints, but also on the number of interface ports P of the model. This implies that the proposed approach is only applicable to small-medium scale systems. This limitation is common to all convex formulations of passivity constraints based on the Positive and Bounded Real Lemmas. Enforcing (54) guarantees uniform passivity by constraining all individual matrix coefﬁcients Fm to be negative semi- deﬁnite, so that Fm ∈S− g ⇒Conv({Fm}) ⊆S− g ⇒F ⊆S− g . (57) (57) The degree of conservativity depends on the distance between F and the boundary of Conv({Fm}). The larger this distance, the larger the degree of conservativity in the passivity (stability) enforcement. Therefore, reduction of this distance will lead to a reduction of conservativity and to an improved model accuracy. V. UNIFORM PASSIVITY CONDITIONS Before proceeding, we need to construct a Multi-Input Multi-Output (MIMO) realization of the auxiliary system IPD(s, ϑ) appearing in (42), which replicates the (scalar) denominator D(s, ϑ) along the diagonal of a P × P matrix. This realization can be written as Y(ϑ) = X m∈Im Y m bm m(ϑ) = X ℓ∈Iℓ Cℓ⊤ 2 Dℓ⊤ 2 bℓ ℓ(ϑ). (50) (50) IPD(s, ϑ) ↔ IP ⊗A1 IP ⊗B1 IP ⊗C1(ϑ) IP ⊗D1(ϑ) (43) = A B C⊗(ϑ) D⊗(ϑ) , (44) (43) The matrix coefﬁcients Y m depend linearly on the numerator coefﬁcients Ri,ℓ. Fourth, we enforce the following structure for the instrumental matrix P(ϑ) (44) P(ϑ) = X m∈Im Pm bm m(ϑ), Pm ∈S¯nP ∀m ∈Im. (51) where all state-space matrices are known since the denom- inator coefﬁcients have already been determined. Note that (51) 15792 15792 VOLUME 10, 2022 VOLUME 10, 2022 T. Bradde et al.: Data-Driven Extraction of Uniformly Stable and Passive Parameterized Macromodels Thanks to this assumption, all terms in (49) are Bernstein polynomials of total degree m, so that (49) can be rewritten in compact form using (50) and (51) as leading to the LMI constraints (39) and (54) are based on a chain of necessary and sufﬁcient conditions, exception made for two main steps. First, the imposition of a particular polynomial structure for the instrumental matrices L(ϑ) and P(ϑ) may restrict the class of storage functions that may provide a stability or passivity certiﬁcate for the model. This in turn may restrict the class of models that can be obtained. In other words, these assumptions introduce some degree of conservativity in the identiﬁcation process. F(ϑ) = X m∈Im Fm bm m(ϑ) ⪯0 ∀ϑ ∈2 (52) (52) with symmetric matrix coefﬁcients Fm = (A, B, Pm) −Xm Y m Y m⊤ −IP ∈Sg (53) (53) A second source of conservativity arises from the dis- cretization of (37) into (39) and of (52) into (54). Considering the latter, the employed discretization is over-conservative because F(ϑ) may be uniformly negative semi-deﬁnite even in case some of the Fm are not. We now analyze this limitation in detail, and we propose an effective strategy to reduce the amount of conservativity, thereby improving the overall model accuracy. of size g = P¯n + P. V. UNIFORM PASSIVITY CONDITIONS Fortunately, the properties of Bernstein polynomials come at hand for this task, since it is well known that, for any matrix function in form (6), the set of control points (the matrix coefﬁcients in the Bernstein expansion) converge uniformly to the value of the expanded function under repeated application of the degree elevation property [36]–[38]. This is graphically illustrated in Fig 1. As a ﬁnal remark, we note that in our derivations we assumed the output matrices associated with the transfer functions N(s, ϑ) and IPD(s, ϑ) to be expressed as Bernstein polynomials sharing the same total degree ℓ. If one drops this assumption, the proposed derivations are still valid, as the elements of the matrix function (49) can always be represented as Bernstein polynomial series of equal degree, by exploiting the degree elevation property. p p y g p y g Let us apply this property to the present passivity (stability) enforcement case. We deﬁne e = m+(e, . . . , e). Then for any e we can always write F(ϑ) = X m∈Im Fm bm m(ϑ) = X e∈Ie Fe be e(ϑ), (58) (58) where the new control points {Fe} are obtained as convex combinations of {Fm} according to (9). We have VI. DEGREE ELEVATION AND CONSERVATIVITY REDUCTION F ⊆Conv({Fe}) ⊆Conv({Fm}), ∀e ∈N. (59) (59) The derivations of Sec. IV-V led to the pair of convex optimization problems (40) and (55) which, when solved in sequence, provide a guarantee of uniform stability and passivity of the parameterized model. All the derivations For increasing e, we have the uniform convergence prop- erty [36] For increasing e, we have the uniform convergence prop- erty [36] lim e→∞{Fe} = F(ϑ). (60) (60) VOLUME 10, 2022 T. Bradde et al.: Data-Driven Extraction of Uniformly Stable and Passive Parameterized Macromodels FIGURE 1. Graphical demonstration of the degree elevation effects. The red line represents the set F of the values attained by a function defined according to (6) for d = 1, ¯ℓ= 4 and F ℓ∈R3; the purple volume is the convex hull of F; the light blue polyhedra are the convex hulls of the control points {F ℓ} for different levels of degree elevation. As the degree of the representation increases, the polyhedron approaches the underlying set F, thus providing better and better outer approximations. FIGURE 1. Graphical demonstration of the degree elevation effects. The red line represents the set F of the values attained by a function defined according to (6) for d = 1, ¯ℓ= 4 and F ℓ∈R3; the purple volume is the convex hull of F; the light blue polyhedra are the convex hulls of the control points {F ℓ} for different levels of degree elevation. As the degree of the representation increases, the polyhedron approaches the underlying set F, thus providing better and better outer approximations. with a convergence rate 1/e, see Fig. 1. For any given e, we can therefore replace (55) with a less conservative optimization problem the imposed polynomial structure on P(ϑ), as far as e is sufﬁciently large. In our experiments, we observed that the degree elevation process is very effective in reducing the conservativity of the passivity constraints (54). Conversely, we did not observe relevant advantages in applying the same strategy to improve the stability constraints (40). min xµ 9µ x xµ + 9µ y yµ 2 s.t. Fe ⪯0, ∀e ∈Ie (61) (61) where the constraint Fe ⪯0 becomes practically equivalent to (52) for sufﬁciently large e. Switching to (61) does not modify the number of decision variables in the optimization. VI. DEGREE ELEVATION AND CONSERVATIVITY REDUCTION However, the number of LMI constraints increases reaching E = Card(Ie), implying that conservativity reduction comes with an increase in computational cost. As a beneﬁcial side effect, the degree elevation property may also lead to a relaxation of the structure imposed on the instrumental matrix P(ϑ), thereby addressing the ﬁrst source of conservativity discussed at the beginning of this Section. If applying the degree elevation after imposing a given structure of the storage function (51), this structure will not change even if expressed as a higher degree polynomial, and the dimension of the space spanned by the allowed storage functions will remain the same. Conversely, if a new degree-elevated structure A. AN INTEGRATED INDUCTOR reference data ˜Hi,j, we also deﬁne the error index We consider a 2-port, 1.5 turns integrated inductor, parame- terized by its side-length ϑ ∈[1.02, 1.52] mm. The structure, depicted in Fig. 2, is characterized in terms of its scattering parameters in the bandwidth [0.1, 12] GHz. A total of ¯m = 11 parameter conﬁgurations are available as training data from a ﬁeld solver sweep, with each dataset including ¯k = 477 logarithmically distributed frequency samples. ϵi,j = max m=1,..., ¯m v u u u t1 ¯k ¯k X k=1 Hi,j(jωk, ϑm) −˜Hi,j k,m ˜Hi,j k,m 2 , (63) (63) which is representative of the worst case relative error of the model against the data over the design space. This training data is used to generate a passive parame- terized macromodel of dynamic order ¯n = 7, using degree ¯ℓ= 2 polynomials to represent the dependence of the model responses on the inductor side-length. With these settings, the solution of the semi-deﬁnite program (40) required 0.22 s on average for the 10 performed iterations. Figure 3 shows the evolution of the convergence index (19) as the iteration number increases. In the considered datasets, the frequency-parameter spaces are sampled over logarithmically or linearly spaced grids. In the considered datasets, the frequency-parameter spaces are sampled over logarithmically or linearly spaced grids. However, the proposed technique can be applied also in case the data are obtained according to some adaptive sam- pling strategy which leads to unstructured data distribution. Additionally, in some application scenarios, some a priori knowledge of the transfer function properties (e.g. degree of smoothness or resonance and anti-resonance locations) may be exploited to reduce the number ¯k of frequency samples retrieved for each parameter conﬁguration. This is not restrictive for the applicability of the method, provided that the data samples are sufﬁcient to fully characterize the structure behavior. However, the proposed technique can be applied also in case the data are obtained according to some adaptive sam- pling strategy which leads to unstructured data distribution. Additionally, in some application scenarios, some a priori knowledge of the transfer function properties (e.g. degree of smoothness or resonance and anti-resonance locations) may be exploited to reduce the number ¯k of frequency samples retrieved for each parameter conﬁguration. This is not restrictive for the applicability of the method, provided that the data samples are sufﬁcient to fully characterize the structure behavior. VII. EXPERIMENTS We now report the results of various numerical tests of increasing complexity, in order to investigate the performance and the applicability limits of proposed approach. All exper- iments have been performed using a workstation equipped with 32 GB of memory and a 3.3 GHz Intel i9-X7900 CPU using a prototypal MATLAB implementation. g p yp p All tests are based on the following settings. The denominator coefﬁcients yµ are always computed by solving problem (40), while the numerator coefﬁcients xµ are estimated at the last iteration by solving problem (61) with a given level of degree elevation e. When solving this problem, we always deﬁne the structure of the matrix P(ϑ) as in (62); this implies that the number of variables involved in the problem is proportional to the number of considered constraints, i.e. Card(Ie). The above mentioned semi-deﬁnite programs are handled via the YALMIP toolbox [39], exploiting the MOSEK interior point method for conic problems [40]. P(ϑ) = X e∈Ie Pe be e(ϑ), Pe ∈S¯nP ∀e ∈Ie (62) (62) of total degree e is used, all the corresponding control points Pe will provide independent degrees of freedom in a degree-e expansion, therefore increasing the space of allowed storage functions enabling certiﬁcation of model passivity (stability). Since polynomials converge to any arbitrary smooth multivariate function on a compact domain, this second strategy practically removes the limitations of Iterations are stopped when the convergence index δµ ≤ 10−3; the evolution of δµ over iterations is displayed below for each test case, in order to monitor convergence based on the stabilization of denominator coefﬁcients. With reference to a given transfer function element Hi,j and the associated 15794 VOLUME 10, 2022 VOLUME 10, 2022 T. Bradde et al.: Data-Driven Extraction of Uniformly Stable and Passive Parameterized Macromodels FIGURE 2. A 1.5-turn integrated inductor parameterized by the side-length ϑ ∈[1.02, 1.52] mm. Drawing is not to scale. FIGURE 3. Integrated inductor. Normalized deviation of the denominator coefficients estimates, as a function of the iteration index µ. FIGURE 4. Integrated inductor. Residual norm of the constrained numerator coefficients estimation problem, as a function of the degree elevation level e. The experimental results are compared to a reference asymptotic 1/e trend, which is expected based on the theory [36]. FIGURE 5. Integrated inductor. VII. EXPERIMENTS Time required to solve problem (61) as a function of the number of decision variables; the latter is directly proportional to the degree elevation e FIGURE 4. Integrated inductor. Residual norm of the constrained numerator coefficients estimation problem, as a function of the degree elevation level e. The experimental results are compared to a reference asymptotic 1/e trend, which is expected based on the theory [36]. FIGURE 2. A 1.5-turn integrated inductor parameterized by the side-length ϑ ∈[1.02, 1.52] mm. Drawing is not to scale. FIGURE 2. A 1.5-turn integrated inductor parameterized by the side-length ϑ ∈[1.02, 1.52] mm. Drawing is not to scale. FIGURE 3. Integrated inductor. Normalized deviation of the denominator coefficients estimates, as a function of the iteration index µ. FIGURE 5. Integrated inductor. Time required to solve problem (61) as a function of the number of decision variables; the latter is directly proportional to the degree elevation e. FIGURE 3. Integrated inductor. Normalized deviation of the denominator coefficients estimates, as a function of the iteration index µ. A. AN INTEGRATED INDUCTOR At the last iteration µ = 7, we solve problem (61) for different values of the degree elevation level e = 1, . . . , 50, in order to show the effect of the proposed conservativity reduction. Fig. 4 reports the trend of the optimal cost function resulting from the solution of (61), as a function of e. This ﬁgure conﬁrms the effectiveness of degree elevation in the reduction of the conservativity of passivity constraints, since the residual norm of the cost function is reduced by almost one order of magnitude. The corresponding CPU time Finally, we remark that the automated selection of the model hyper-parameters ¯ℓand ¯n is still an open problem. In the following examples, this selection was performed in a preprocessing stage with a basic trial and error strategy. 15795 VOLUME 10, 2022 T. Bradde et al.: Data-Driven Extraction of Uniformly Stable and Passive Parameterized Macromodels FIGURE 6. Integrated inductor. Comparison between parameterized model responses and training data for a degree elevation level e = 50; all the ¯m = 11 parameter configurations are shown. FIGURE 6. Integrated inductor. Comparison between parameterized model responses and training data for a degree elevation level e = 50; all the ¯m = 11 parameter configurations are shown. The quality of the resulting model (for the case e = 50) is conﬁrmed by comparing the model responses to the training data in Fig. 6. Finally, Fig. 7 reports the relative error index ϵ2,1 as a function of the degree elevation e. The ﬁgure reports also the error that would be obtained by generating a model without enforcing any passivity constraint. We see that starting from e ≈6 the proposed approach is able to achieve a model accuracy that is not distinguishable from the unconstrained case. We conclude that proposed framework is able to guarantee uniform model passivity by construction, with no accuracy degradation, and with limited overhead in computing time, at least for this small-scale example. requirements are depicted in Fig. 5, as a function of the total number of variables involved in the optimization, which in turn depends on the degree elevation order e. For this small- scale example, the computational time is modest even in the case e = 50, which is associated to a total of 5847 unknowns. requirements are depicted in Fig. A. AN INTEGRATED INDUCTOR Evolution of the convergence index δµ through iterations. A. AN INTEGRATED INDUCTOR Coupled transmission line. Optimal values of the cost function in (61) for various degree elevation levels. FIGURE 10. Coupled transmission line. Optimal values of the cost function in (61) for various degree elevation levels. FIGURE 7. Integrated inductor. Relative error ϵ2,1 for different degree elevation levels e. The blue line reports the corresponding error for a model generated without enforcing any passivity constraint. FIGURE 7. Integrated inductor. Relative error ϵ2,1 for different degree elevation levels e. The blue line reports the corresponding error for a model generated without enforcing any passivity constraint. FIGURE 8. A partially-coupled multiconductor transmission line system. The parameter ϑ represents the length of the coupling. The drawing is not to scale FIGURE 11. Coupled transmission line. Time required to solve problem (61) as a function of the number of decision variables corresponding to the various degree elevation levels reported in Fig. 10. The increase in the number of variables is mostly due to the increased order of the instrumental polynomial matrix P(ϑ). FIGURE 8. A partially-coupled multiconductor transmission line system. The parameter ϑ represents the length of the coupling. The drawing is not to scale. FIGURE 9. Coupled transmission line. Evolution of the convergence index δµ through iterations. form a square with adjacent center-to-center distance equal to 1.61 mm. The total length of the interconnect is L = 10 cm, but the coupling between the two pairs in the corresponding per-unit-length matrices is considered only over a portion of the length Lc = ϑ ∈2 = [20, 40] mm, which is the independent parameter considered for this study. The lines are considered as uncoupled for the remaining length L −Lc. Figure 8 provides a graphical description for the structure. This example is selected to illustrate the shifting of the resonances as ϑ changes, and the capability of the model to track such resonances (the parameterized model poles) thanks to the adopted model structure. The design space is sampled with ¯m = 11 linearly spaced values. For each parameter conﬁguration a total of ¯k = 499 logarithmically spaced frequency samples of the 4 × 4 scattering matrix are extracted in the bandwidth [0.01, 5] GHz. These samples are used to generate a model of dynamic order ¯n = 28, whereas numerator N(s, ϑ) and denominator D(s, ϑ) are parameterized by Bernstein polynomials of order 4 and 2, respectively. FIGURE 9. Coupled transmission line. A. AN INTEGRATED INDUCTOR 5, as a function of the total number of variables involved in the optimization, which in turn depends on the degree elevation order e. For this small- scale example, the computational time is modest even in the case e = 50, which is associated to a total of 5847 unknowns. Considering as an example the model obtained for e = 50, we veriﬁed a-posteriori the non-expansivity condition 3 of the model. Therefore, we computed the model singular values over a ﬁnely sampled frequency-parameter grid, using 3000 log-spaced frequency values in the bandwidth [0, 1011] Hz, and 3000 linearly spaced samples in the design space. The maximum observed singular value resulted less than one, with a passivity margin 1 −σmax = 3 × 10−10. requirements are depicted in Fig. 5, as a function of the total number of variables involved in the optimization, which in turn depends on the degree elevation order e. For this small- scale example, the computational time is modest even in the case e = 50, which is associated to a total of 5847 unknowns. Considering as an example the model obtained for e = 50, we veriﬁed a-posteriori the non-expansivity condition 3 of the model. Therefore, we computed the model singular values over a ﬁnely sampled frequency-parameter grid, using 3000 log-spaced frequency values in the bandwidth [0, 1011] Hz, and 3000 linearly spaced samples in the design space. The maximum observed singular value resulted less than one, with a passivity margin 1 −σmax = 3 × 10−10. Considering as an example the model obtained for e = 50, we veriﬁed a-posteriori the non-expansivity condition 3 of the model. Therefore, we computed the model singular values over a ﬁnely sampled frequency-parameter grid, using 3000 log-spaced frequency values in the bandwidth [0, 1011] Hz, and 3000 linearly spaced samples in the design space. The maximum observed singular value resulted less than one, with a passivity margin 1 −σmax = 3 × 10−10. 15796 VOLUME 10, 2022 T. Bradde et al.: Data-Driven Extraction of Uniformly Stable and Passive Parameterized Macromodels FIGURE 7. Integrated inductor. Relative error ϵ2,1 for different degree elevation levels e. The blue line reports the corresponding error for a model generated without enforcing any passivity constraint. FIGURE 8. A partially-coupled multiconductor transmission line system. The parameter ϑ represents the length of the coupling. The drawing is not to scale. FIGURE 9. Coupled transmission line. A. AN INTEGRATED INDUCTOR Evolution of the convergence index δµ through iterations. B. MULTICONDUCTOR TRANSMISSION LINE WITH FIGURE 10. Coupled transmission line. Optimal values of the cost function in (61) for various degree elevation levels. FIGURE 11. Coupled transmission line. Time required to solve problem (61) as a function of the number of decision variables corresponding to the various degree elevation levels reported in Fig. 10. The increase in the number of variables is mostly due to the increased order of the instrumental polynomial matrix P(ϑ). form a square with adjacent center-to-center distance equal to 1.61 mm. The total length of the interconnect is L = 10 cm, but the coupling between the two pairs in the corresponding per-unit-length matrices is considered only over a portion of the length Lc = ϑ ∈2 = [20, 40] mm, which is the independent parameter considered for this study. The lines are considered as uncoupled for the remaining length L −Lc. Figure 8 provides a graphical description for the structure. This example is selected to illustrate the shifting of the resonances as ϑ changes, and the capability of the model to track such resonances (the parameterized model poles) thanks to the adopted model structure. The design space is sampled with ¯m = 11 linearly spaced values. For each parameter conﬁguration a total of ¯k = 499 logarithmically spaced frequency samples of the 4 × 4 scattering matrix are extracted in the bandwidth FIGURE 10. Coupled transmission line. Optimal values of the cost function in (61) for various degree elevation levels. FIGURE 11. Coupled transmission line. Time required to solve problem (61) as a function of the number of decision variables corresponding to the various degree elevation levels reported in Fig. 10. The increase in the number of variables is mostly due to the increased order of the instrumental polynomial matrix P(ϑ). FIGURE 10. Coupled transmission line. Optimal values of the cost FIGURE 7. Integrated inductor. Relative error ϵ2,1 for different degree elevation levels e. The blue line reports the corresponding error for a model generated without enforcing any passivity constraint. FIGURE 8. A partially-coupled multiconductor transmission line system. The parameter ϑ represents the length of the coupling. The drawing is not to scale. FIGURE 9. Coupled transmission line. Evolution of the convergence index δµ through iterations. FIGURE 10. Coupled transmission line. Optimal values of the cost function in (61) for various degree elevation levels. FIGURE 10. B. MULTICONDUCTOR TRANSMISSION LINE WITH VARIABLE COUPLING LENGTH Relative error ϵ2,1 for the passive models based on different degree elevations, compared to the error of the model obtained without enforcing any passivity constraint. FIGURE 14. Parameterized poles trajectories of the coupled transmission line model. Left panel: in-band poles. Right panel: enlarged view on the low-frequency region. FIGURE 16. High-speed link. Convergence of denominator coefficients estimates through iterations. FIGURE 17. High-speed link. Time required to enforce the model passivity, as a function of the number of variables required by different degree elevation levels. FIGURE 16. High-speed link. Convergence of denominator coefficients estimates through iterations. FIGURE 17. High-speed link. Time required to enforce the model passivity, as a function of the number of variables required by different degree elevation levels. FIGURE 16. High-speed link. Convergence of denominator coefficients estimates through iterations. FIGURE 16. High-speed link. Convergence of denominator coefficients estimates through iterations. FIGURE 13. Coupled transmission line. Relative error ϵ2,1 for the passive models based on different degree elevations, compared to the error of the model obtained without enforcing any passivity constraint. FIGURE 14. Parameterized poles trajectories of the coupled transmission line model. Left panel: in-band poles. Right panel: enlarged view on the low-frequency region. FIGURE 17. High-speed link. Time required to enforce the model passivity, as a function of the number of variables required by different degree elevation levels. FIGURE 17. High-speed link. Time required to enforce the model passivity, as a function of the number of variables required by different degree elevation levels. FIGURE 14. Parameterized poles trajectories of the coupled transmission line model. Left panel: in-band poles. Right panel: enlarged view on the low-frequency region. FIGURE 18. High-speed link. Cost function reduction for increasing degree elevation during passivity enforcement. FIGURE 15. High-speed PCB stripline interconnect parameterized by via pad and antipad radii. Drawing for illustration only and not to scale. FIGURE 18. High-speed link. Cost function reduction for increasing degree elevation during passivity enforcement. FIGURE 15. High-speed PCB stripline interconnect parameterized by via pad and antipad radii. Drawing for illustration only and not to scale. for this case the error approaches the limit corresponding to the unconstrained (hence not guaranteed passive) model, computed using the same training dataset. A graphical representation of the model parameterized poles trajectories is given in Fig 14, computed over a very ﬁne sweep of the free parameter ϑ ∈2. B. MULTICONDUCTOR TRANSMISSION LINE WITH VARIABLE COUPLING LENGTH This second test case provides an academic example with a distributed coupling parameter. We consider a multiconductor transmission line with two differential pairs, each made of two equal parallel wires (radius of copper core rw = 0.5 mm and dielectric coating re = 0.8 mm). The two differential pairs are placed next to each other, so that the wire centers The convergence of the identiﬁcation algorithm is demon- strated in Fig. 9, where the value of δµ for µ ≥ 1 is reported. For this example, we built 10 different models, 15797 VOLUME 10, 2022 T. Bradde et al.: Data-Driven Extraction of Uniformly Stable and Passive Parameterized Macromodels FIGURE 12. Coupled transmission line. Fitting results for the first-column of the transfer matrix. The elements that are not shown here exhibit similar trends and a comparable model accuracy. solving each time problem (61) with different levels of degree elevation e = 1, 2, . . . , 10. Figure 10 reports the optimal cost function value from the solution of problem (61) for different degree elevation levels e. The average time required to solve FIGURE 12. Coupled transmission line. Fitting results for the first-column of the transfer matrix. The elements that trends and a comparable model accuracy. FIGURE 12. Coupled transmission line. Fitting results for the first-column of the transfer matrix. The elements that are not shown here exhibit similar trends and a comparable model accuracy. solving each time problem (61) with different levels of degree elevation e = 1, 2, . . . , 10. Figure 10 reports the optimal cost solving each time problem (61) with different levels of degree elevation e = 1, 2, . . . , 10. Figure 10 reports the optimal cost function value from the solution of problem (61) for different degree elevation levels e. The average time required to solve 15798 VOLUME 10, 2022 T. Bradde et al.: Data-Driven Extraction of Uniformly Stable and Passive Parameterized Macromodels FIGURE 16. High-speed link. Convergence of denominator coefficients estimates through iterations. FIGURE 17. High-speed link. Time required to enforce the model passivity, as a function of the number of variables required by different degree elevation levels. FIGURE 18. High-speed link. Cost function reduction for increasing degree elevation during passivity enforcement. FIGURE 13. Coupled transmission line. B. MULTICONDUCTOR TRANSMISSION LINE WITH VARIABLE COUPLING LENGTH As expected, all the poles are stable with a negative real part, uniformly in the parameter space. The presence of bifurcations further conﬁrms the effectiveness of the proposed approach in modeling non-smooth poles problem (40) is 0.72 s, while the time required to solve (61) depends on the degree elevation level. The actual runtimes for this test case are reported in Fig. 11. Figure 12 reports the modeling results obtained with e = 10 for the entire ﬁrst column of the scattering matrix, while Fig. 13 depicts the model relative error ϵ2,1 as a function of the degree elevation e. These results conﬁrm that also 15799 VOLUME 10, 2022 T. Bradde et al.: Data-Driven Extraction of Uniformly Stable and Passive Parameterized Macromodels FIGURE 19. High-speed link. Comparison of model responses (e = 25) with the corresponding raw data over a random subset of 14 out of the total 81 available frequency responses. FIGURE 19. High-speed link. Comparison of model responses (e = 25) with the corresponding raw data over a random subset of 14 out of the total 81 available frequency responses. behaviors, thanks to the implicit parameterization provided by the adopted model structure. the interval [0.02, 5] GHz, and over a 9 × 9 uniform grid in the parameter space. These data are used to generate a parameterized macromodel with ¯n = 25 poles and polynomial order of numerator and denominator ℓ= (3, 2). C. A TWO-PARAMETER HIGH-SPEED PCB LINK The simulation is performed by considering eight different antipad radius configurations in the interval [400, 600]µm, while keeping fixed the pad radius to 300µm. Two periods of the output signal are shown. FIGURE 21. Parameterized transient analysis of the high-speed link equivalent circuit. The simulation is performed by considering eight different antipad radius configurations in the interval [400, 600]µm, while keeping fixed the pad radius to 300µm. Two periods of the output signal are shown. FIGURE 20. High-speed link. Evolution of the relative error ϵ1,1 for the passive models with different degree elevations, compared to the unconstrained model error. The value of the optimal cost function value of the semi- deﬁnite program (55) for different values of e is reported in Fig. 18, and conﬁrms the same decreasing trend that has been observed in single-parameter test cases. FIGURE 22. The LNA circuit schematic. For the case e = 25, a visual comparison between the parameterized model frequency response and the reference data is provided in Fig. 19, considering a subset of 14 random parameter conﬁgurations out of the available 81. Also in this case, the accuracy of the model is remarkable throughout the considered frequency band, with no visual difference between model and data on this scale. Finally, Fig. 20 reports the relative error ϵ1,1 for different degree elevations. Also for this case the error stabilizes to the same error of the unconstrained (non-passive) model; this occurs at about e = 15. In order to assess the inﬂuence of the design parameters on the time-domain responses, and to demonstrate the efﬁciency of the parameterized models in a typical use case scenario, we performed a transient simulation of the equivalent parameterized SPICE circuit synthesized from the model. The simulation setup includes a 50 voltage driver launching a pulse sequence with amplitude 1 V, period T = 3 ns, rise and fall time 200 ps, and bit duration 0.8 ns. The receiver side is instead terminated by a RC parallel load, with R = 1 k and C = 2 pF. We considered a ﬁxed pad radius ϑ1 = 300 µm, and we let the antipad radius vary, by considering a linear sweep of 8 conﬁgurations within its allowed range. The results of the simulation are depicted in Fig. 21, where the voltage signals at the receiver are shown. C. A TWO-PARAMETER HIGH-SPEED PCB LINK The time requirements of each simulation amounted to 0.5 s using a freeware SPICE solver. FIGURE 22. The LNA circuit schematic. The purpose of this test case is two-fold. First, we show that even in case of high-dimensional design spaces, the generation of uniformly stable parameterized macromodels can be efﬁciently tackled by solving problem (40). Second, we show how the proposed approach is less conservative than the current state-of-the-art method [24] providing a formal guarantee of uniform stability given model structure (13). Of course, uniform passivity is not applicable since this is an active device. We consider a ﬁxed operating point VSUP = 4.5 V and we construct a small-signal linearized model. A total of ¯m = 1400 parameter conﬁgurations are considered according to a latin-hypercube distribution in the design space. For each ﬁxed conﬁguration, the reﬂection coefﬁcient at the ampliﬁer input port is sampled at ¯k = 701 logarithmically spaced frequency points in the interval [1, 10] GHz. Only 595 parameter conﬁgurations are exploited to generate a parameterized macromodel with ¯n = 10 and ℓ = (1, 1, 1, 1, 1, 1, 2), while the remaining samples are left for model validation. C. A TWO-PARAMETER HIGH-SPEED PCB LINK The convergence of the denominator coefﬁcient estimation is illustrated by plotting δµ in Fig. 16. With the considered model structure, the time required to solve (40), averaged over the 8 PSK iterations amounts to 1.5 s. For this example, we considered a number of possible degree elevations levels e ranging from 1 to 25. The time required to build each of the 25 models is depicted in Fig. 17. We observe that, although the time requirements for this example are larger, the solver scales favourably with the increase in the number of instrumental variables induced by the higher degree elevations (almost linearly, at least up to 105 variables). This test case considers a 2-parameter structure, namely a high-speed stripline link running through two PCBs attached by a connector and the corresponding via ﬁelds, ﬁrst presented in [41]. A schematic layout of the structure is depicted in Fig. 15. The PCB substrate has permittivity ϵr = 3 and tanδ = 0.002. The vertical vias are parameterized by the pad radius ϑ1 ∈[100, 300] µm and the associated antipad radius ϑ2 ∈[400, 600] µm. See [41] for full details. The scattering parameters of the structure are available from a ﬁeld solver (courtesy of Prof. Schuster, TUHH, Germany) at ¯k = 250 frequency points linearly spaced in VOLUME 10, 2022 15800 VOLUME 10, 2022 T. Bradde et al.: Data-Driven Extraction of Uniformly Stable and Passive Parameterized Macromodels FIGURE 20. High-speed link. Evolution of the relative error ϵ1,1 for the passive models with different degree elevations, compared to the unconstrained model error. FIGURE 21. Parameterized transient analysis of the high-speed link equivalent circuit. The simulation is performed by considering eight different antipad radius configurations in the interval [400, 600]µm, while keeping fixed the pad radius to 300µm Two periods of the output signal FIGURE 20. High-speed link. Evolution of the relative error ϵ1,1 for the passive models with different degree elevations, compared to the unconstrained model error. FIGURE 21. Parameterized transient analysis of the high-speed link equivalent circuit. The simulation is performed by considering eight different antipad radius configurations in the interval [400, 600]µm, while keeping fixed the pad radius to 300µm. Two periods of the output signal are shown. FIGURE 21. Parameterized transient analysis of the high-speed link equivalent circuit. VIII. CONCLUSION This work presented a passive macromodeling strategy that can be successfully used to generate surrogates of small- to-medium size passive multiport structures characterized by a limited number of degrees of freedom. The approach combines the desirable model compactness feature, typical of approaches based on multivariate rational ﬁtting, with the theoretical warranty of uniform model passivity throughout the design space. As a particular case, removing passivity conditions enables parameterized (linearized small-signal) macromodeling of active devices with uniform stability constraints. The proposed stability and passivity constraints are conservative since based on a discretization of continuous positive and/or bounded realness conditions in a multidi- mensional parameter space. However, the amount of the conservativity in the stability and passivity constraints can be effectively controlled in terms of the Bernstein polynomial degree elevation, which provides an algorithm control knob. Therefore, the proposed method naturally allows users to select the most appropriate trade-off between computational time requirements for the model extraction and model accuracy. FIGURE 23. LNA example. Comparison between model responses and validation data over 19 different frequency responses, randomly selected in the design space. TABLE 1. Free parameters considered for the modeling of the LNA test case. First six parameters: parasitic inductances and capacitances of the transistor. Parameter h is the substrate thickness for lines TL1, TL2, TL3. TABLE 1. Free parameters considered for the modeling of the LNA test case. First six parameters: parasitic inductances and capacitances of the transistor. Parameter h is the substrate thickness for lines TL1, TL2, TL3. TABLE 1. Free parameters considered for the modeling of the LNA test case. First six parameters: parasitic inductances and capacitances of the transistor. Parameter h is the substrate thickness for lines TL1, TL2, TL3. Various numerical examples show that stability enforce- ment is attained in seconds for typical small-medium scale problems, whereas passivity enforcement requires a larger runtime, which depends on the cumulative number of decision variables. The latter depends on the number of model poles, the degree of polynomials providing model parameterization, as well as the number of instrumental variables that are required to cast the proposed constraints in a convex form. Future research directions will be devoted to the reduction of the computational burden required by the proposed strategy, with the objective of handling larger and more complex electrical, electronic or electromagnetic structures, and/or the concurrent dependence on more free parameters. [1] T. Zhang, K. Chakrabarty, and R. B. Fair, ‘‘Behavioral modeling and performance evaluation of microelectroﬂuidics-based PCR systems using SystemC,’’ IEEE Trans. Comput.-Aided Design Integr., vol. 23, no. 6, pp. 843–858, Jun. 2004. D. AN ACTIVE DEVICE In this last example we generate a reduced-order small-signal model of the Low Noise Ampliﬁer (LNA) depicted in Fig. 22, which includes both lumped elements and lossy transmission lines. The circuit depends on d = 7 design parameters, which are listed in Table 1. The device was ﬁrst presented in [42]; additionally, it was considered as a test-bench for the generation of uniformly stable parameterized macromodels in [24]. With this conﬁguration, the modeling algorithm reaches the stopping threshold δµ = 10−3 in only 3 iterations. The 15801 VOLUME 10, 2022 T. Bradde et al.: Data-Driven Extraction of Uniformly Stable and Passive Parameterized Macromodels FIGURE 23. LNA example. Comparison between model responses and validation data over 19 different frequency responses, randomly selected in the design space. The stop criterion δµ = 10−3 was met after 4 iterations, with an average computing time required to solve (40) equal to 1.2 s. We computed the relative error index of the resulting model, obtaining ϵ = 6.36 × 10−5; the same index for a model based on [24] was ϵ = 1.94 × 10−2. Thus, the proposed technique provides a decrease of the worst case relative error of about 3 orders of magnitude, while guaranteeing the uniform model stability by construction. This improvement is attained in approximately the same runtime. [2] G. Casinovi and A. Sangiovanni-Vincentelli, ‘‘A macromodeling algorithm for analog circuits,’’ IEEE Trans. Comput.-Aided Design Integr. Circuits Syst., vol. 10, no. 2, pp. 150–160, Feb. 1991. VIII. CONCLUSION average time to solve (40) is 6 s, and the relative error index results ϵ = 1.42 × 10−3, conﬁrming that the model is highly accurate also in correspondence of the validation samples. In Fig. 23, we provide a visual comparison between the model and the data, for 19 different randomly-selected validation responses. In order to show the low degree of conservativity of the proposed stability constraints, we repeated the same experiment performed in [24], where the uniform stability is enforced by imposing a sign inequality directly on the denominator coefﬁcients rn,ℓduring the model generation. For this purpose, we considered the same LNA device and restricted the dimension of the design space to d = 5, by taking into account only the ﬁrst ﬁve parameters listed in Table 1. We built a model by setting ¯n = 16 and ℓ= (1, 1, 1, 1, 1), as in the referenced article. REFERENCES REFERENCES 8, pp. 863–886, Oct. 2009. [17] S. Grivet-Talocia, ‘‘Passivity enforcement via perturbation of Hamiltonian matrices,’’ IEEE Trans. Circuits Syst. I, Reg. Papers, vol. 51, no. 9, pp. 1755–1769, Sep. 2004. [42] T. Buss, ‘‘2GHz low noise ampliﬁer with the BFG425W,’’ Philips Semi- cond., Nijmegen, The Netherlands, Tech. Rep. RNR-T45-96-B-773, 1996. [Online]. Available: http://application-notes.digchip.com/004/4-7999.pdf [18] D. Deschrijver and T. Dhaene, ‘‘Fast passivity enforcement of S-parameter macromodels by pole perturbation,’’ IEEE Trans. Microw. Theory Techn., vol. 57, no. 3, pp. 620–626, Mar. 2009. [19] S. Grivet-Talocia and A. Ubolli, ‘‘On the generation of large passive macromodels for complex interconnect structures,’’ IEEE Trans. Adv. Packag., vol. 29, no. 1, pp. 39–54, Feb. 2006. [20] F. Ferranti, L. Knockaert, T. Dhaene, and G. Antonini, ‘‘Passivity- preserving parametric macromodeling for highly dynamic tabulated data based on lur’e equations,’’ IEEE Trans. Microw. Theory Techn., vol. 58, no. 12, pp. 3688–3696, Dec. 2010. TOMMASO BRADDE (Member, IEEE) received the bachelor’s degree in electronic engineering from Roma Tre University, Rome, Italy, in 2015, and the master’s degree in mechatronic engineer- ing, Politecnico di Torino, Turin, Italy, in 2018, where he is currently pursuing the Ph.D. degree in electrical, electronics and communication engi- neering. His research interests include data-driven parameterized macromodeling and its applications to system level signal and power integrity assess- ments, with the inclusion of active devices. He was a co-recipient of the 2018 Best Paper Award of the IEEE International Symposium on Electromagnetic Compatibility, the 2020 Best Paper Award of the IEEE Conference on Electrical Performance of Electronic Packaging and Systems (EPEPS), the Best Student Paper Award of the 23rd IEEE Workshop on Signal and Power Integrity, and the Best Student Paper Award of the 2021 EPEPS Symposium. [21] F. Ferranti, L. Knockaert, and T. Dhaene, ‘‘Passivity-preserving parametric macromodeling by means of scaled and shifted state-space systems,’’ IEEE Trans. Microw. Theory Techn., vol. 59, no. 10, pp. 2394–2403, Oct. 2011. [22] S. Grivet-Talocia, ‘‘A perturbation scheme for passivity veriﬁcation and enforcement of parameterized macromodels,’’ IEEE Trans. Compon., Packag., Manuf. Technol., vol. 7, no. 11, pp. 1869–1881, Nov. 2017. [23] S. Grivet-Talocia and R. Trinchero, ‘‘Behavioral, parameterized, and broadband modeling of wired interconnects with internal discontinuities,’’ IEEE Trans. Electromagn. Compat., vol. 60, no. 1, pp. 77–85, Feb. 2018. [24] A. Zanco, S. Grivet-Talocia, T. Bradde, and M. De Stefano, ‘‘Uniformly stable parameterized macromodeling through positive deﬁnite basis functions,’’ IEEE Trans. Compon., Packag., Manuf. Technol., vol. 10, no. REFERENCES 15802 15802 VOLUME 10, 2022 T. Bradde et al.: Data-Driven Extraction of Uniformly Stable and Passive Parameterized Macromodels [3] I. S. Stievano, I. A. Maio, and F. G. Canavero, ‘‘Parametric macromodels of digital I/O ports,’’ IEEE Trans. Adv. Packag., vol. 25, no. 2, pp. 255–264, May 2002. [26] T. Bradde, A. Zanco, and S. Grivet-Talocia, ‘‘Bivariate macromodeling with passivity constraints,’’ in Proc. IEEE 30th Conf. Electr. Perform. Electron. Packag. Syst. (EPEPS), Oct. 2021, pp. 1–3. [27] A. Kojima, ‘‘A characterization of parameter-dependent LMIs on Bernstein polynomial basis,’’ in Proc. IEEE Conf. Decis. Control (CDC), Dec. 2018, pp. 4687–4694. [4] G. Antonini, D. Deschrijver, and T. Dhaene, ‘‘Broadband rational macromodeling based on the adaptive frequency sampling algorithm and the partial element equivalent circuit method,’’ IEEE Trans. Electromagn. Compat., vol. 50, no. 1, pp. 128–137, Feb. 2008. [28] J. Berchtold and A. Bowyer, ‘‘Robust arithmetic for multivariate bernstein- form polynomials,’’ Comput.-Aided Des., vol. 32, no. 11, pp. 681–689, Sep. 2000. [5] R. Araneo, ‘‘Extraction of broad-band passive lumped equivalent circuits of microwave discontinuities,’’ IEEE Trans. Microw. Theory Techn., vol. 54, no. 1, pp. 393–401, Jan. 2006. [29] A. Zanco and S. Grivet-Talocia, ‘‘Toward fully automated high- dimensional parameterized macromodeling,’’ IEEE Trans. Compon., Packag., Manuf. Technol., vol. 11, no. 9, pp. 1402–1416, Sep. 2021. [6] S. Grivet-Talocia and B. Gustavsen, Passive Macromodeling: Theory and Applications, vol. 239. Hoboken, NJ, USA: Wiley, 2015. [30] B. Gustavsen, ‘‘Improving the pole relocating properties of vector ﬁtting,’ in Proc. IEEE Power Eng. Soc. Gen. Meeting, Jun. 2006, p. 1. [7] B. Gustavsen and A. Semlyen, ‘‘Rational approximation of frequency domain responses by vector ﬁtting,’’ IEEE Trans. Power Del., vol. 14, no. 3, pp. 1052–1061, Jul. 1999. [31] T. Bradde, S. Grivet-Talocia, M. De Stefano, and A. Zanco, ‘‘A scalable reduced-order modeling algorithm for the construction of parameterized interconnect macromodels from scattering responses,’’ in Proc. IEEE Symp. Electromagn. Compat., Signal Integrity Power Integrity (EMC, SI PI), Jul. 2018, pp. 650–655. [8] Y. Nakatsukasa, O. Sète, and L. N. Trefethen, ‘‘The AAA algorithm for rational approximation,’’ SIAM J. Sci. Comput., vol. 40, no. 3, pp. A1494–A1522, Jan. 2018. [9] M. Berljafa and S. Güttel, ‘‘Generalized rational Krylov decompositions with an application to rational approximation,’’ SIAM J. Matrix Anal. Appl., vol. 36, no. 2, pp. 894–916, 2015. [32] M. R. Wohlers, Lumped and Distributed Passive Networks. New York, NY, USA: Academic, 1969. [33] S. V. Gusev and A. L. REFERENCES Likhtarnikov, ‘‘Kalman-Popov-Yakubovich lemma and the S-procedure: A historical essay,’’ Autom. Remote Control, vol. 67, no. 11, pp. 1768–1810, Nov. 2006. [10] S. Lefteriu and A. C. Antoulas, ‘‘A new approach to modeling multiport systems from frequency-domain data,’’ IEEE Trans. Comput.-Aided Design Integr. Circuits Syst., vol. 29, no. 1, pp. 14–27, Jan. 2010. [34] L. Vandenberghe, V. R. Balakrishnan, R. Wallin, A. Hansson, and T. Roh, ‘‘Interior-point algorithms for semideﬁnite programming problems derived from the KYP lemma,’’ in Positive Polynomials Control. Berlin, Germany: Springer, 2005, pp. 195–238. [11] E. R. Samuel, L. Knockaert, F. Ferranti, and T. Dhaene, ‘‘Guaranteed passive parameterized macromodeling by using Sylvester state-space realizations,’’ IEEE Trans. Microw. Theory Techn., vol. 61, no. 4, pp. 1444–1454, Apr. 2013. [35] B. D. Anderson and S. Vongpanitlerd, Network Analysis and Synthesis: A Modern Systems Theory Approach. New York, NY, USA: Dover, 2013. [12] F. Ferranti, L. Knockaert, and T. Dhaene, ‘‘Parameterized S-parameter based macromodeling with guaranteed passivity,’’ IEEE Microw. Wireless Compon. Lett., vol. 19, no. 10, pp. 608–610, Oct. 2009. [36] F. J. Li, ‘‘Interpolation and convergence of Bernstein–Bézier c Acta Math. Sinica, English Ser., vol. 27, no. 9, p. 1769, 2011. [13] A. C. Rodriguez and S. Gugercin, ‘‘The p-AAA algorithm for data driven modeling of parametric dynamical systems,’’ 2020, arXiv:2003.06536. [37] G. Farin, Curves and Surfaces for Computer-Aided Geometric Design: A Practical Guide. Amsterdam, The Netherlands: Elsevier, 2014. [38] H. Prautzsch and L. Kobbelt, ‘‘Convergence of subdivision and degree elevation,’’ Adv. Comput. Math., vol. 2, no. 1, pp. 143–154, Jan. 1994. [14] P. Triverio, S. Grivet-Talocia, and M. S. Nakhla, ‘‘A parameterized macromodeling strategy with uniform stability test,’’ IEEE Trans. Adv. Packag., vol. 32, no. 1, pp. 205–215, Feb. 2009. [39] J. Lofberg, ‘‘YALMIP : A toolbox for modeling and optimization in MATLAB,’’ in Proc. IEEE Int. Conf. Robot. Autom., Sep. 2004, pp. 284–289. [15] Y. Q. Xiao, S. Grivet-Talocia, P. Manfredi, and R. Khazaka, ‘‘A novel framework for parametric Loewner matrix interpolation,’’ IEEE Trans. Compon., Packag., Manuf. Technol., vol. 9, no. 12, pp. 2404–2417, Dec. 2019. pp [40] M. ApS. (2021). MOSEK Optimizer API for C. [Online]. Available: https://docs.mosek.com/9.3/capi.pdf [41] J. Preibisch, T. Reuschel, K. Scharff, J. Balachandran, B. Sen, and C. Schuster, ‘‘Exploring efﬁcient variability-aware analysis method for high-speed digital link design using PCE,’’ in Proc. DesignCon, 2017, pp. 1–19. [16] S. Grivet-Talocia, ‘‘On driving non-passive macromodels to instability,’’ Int. J. Circuit Theory Appl., vol. 37, no. REFERENCES 11, pp. 1782–1794, Nov. 2020. [25] A. Zanco, S. Grivet-Talocia, T. Bradde, and M. De Stefano, ‘‘Enforcing passivity of parameterized LTI macromodels via Hamiltonian-driven multivariate adaptive sampling,’’ IEEE Trans. Comput.-Aided Design Integr. Circuits Syst., vol. 39, no. 1, pp. 225–238, Jan. 2020. 15803 VOLUME 10, 2022 VOLUME 10, 2022 T. Bradde et al.: Data-Driven Extraction of Uniformly Stable and Passive Parameterized Macromodels STEFANO GRIVET-TALOCIA (Fellow, IEEE) received the Laurea and Ph.D. degrees in elec- tronic engineering from the Politecnico di Torino, Turin, Italy. From 1994 to 1996, he was with the NASA/Goddard Space Flight Center, Greenbelt, MD, USA. He is currently a Full Professor of electrical engineering with the Politecnico di Torino. He co-founded the academic spinoff company IdemWorks, Turin, in 2007, where he served as the President until its acquisition by CST, in 2016. He has authored about 200 journals and conference papers. His current research interests include passive macromodeling of lumped and distributed interconnect structures, model-order reduction, modeling and simulation of ﬁelds, circuits, and their interaction, wavelets, time-frequency transforms, and their applications. He was a co-recipient of the 2007 Best Paper Award of the IEEE TRANSACTIONS ON ADVANCED PACKAGING. He received the IBM Shared University Research Award in 2007, 2008, and 2009. He was the General Chair of the 20th and 21st IEEE Workshops on Signal and Power Integrity (SPI2016 and SPI2017). He is currently the Chair of the Technical Committee on Electrical Design, Modeling and Simulation, IEEE Electronic Packaging Society. He was an Associate Editor of the IEEE TRANSACTIONS ON ELECTROMAGNETIC COMPATIBILITY, from 1999 to 2001. He is currently serving as an Associate Editor for the IEEE TRANSACTIONS ON COMPONENTS, PACKAGING AND MANUFACTURING TECHNOLOGY. GIUSEPPE C. CALAFIORE (Fellow, IEEE) received the master’s degree in ﬁnancial data sci- ence from the University of California, Berkeley. He held visiting positions at the Information Sys- tems Laboratory (ISL), Stanford University, Cal- ifornia, in 1995; the Ecole Nationale Supérieure de Techniques Avancées (ENSTA), Paris, in 1998; and the University of California, in 1999, 2003, 2007, 2017, 2018, and 2019. He was a Senior Fellow with the Institute of Pure and Applied Mathematics (IPAM), University of California, Los Angeles, in 2010. He is currently a Full Professor with DET, Politecnico di Torino, where he coordinates the Control Systems and Data Science Group and an Associate Fellow of IEIIT-CNR. He has over 20 years of teaching experience in master-level and Ph.D. REFERENCES courses in the areas of systems and control theory, convex optimization and machine learning. He is the author of about 200 journals and conference proceedings papers and eight books. His research interests include convex optimization, identiﬁcation and control of uncertain systems, with applications ranging from ﬁnance and economic systems to robust control, machine learning, and data science. He received the IEEE Control System Society ‘‘George S. Axelby’’ Outstanding Paper Award, in 2008. ALESSANDRO ZANCO (Member, IEEE) received the bachelor’s degree in electrical engineering and the master’s degree in mechathronic engineering from the Politecnico di Torino, Torino, Italy, in 2015 and 2018, respectively, where he is currently pursuing the Ph.D. degree in electrical, electronics and communication engineering. His research interest includes high-dimensional parameterized black-box modeling for EMC. 15804 VOLUME 10, 2022
https://openalex.org/W2151371357	http://www.scirp.org/journal/PaperDownload.aspx?paperID=59010	English	null	Drug Prescribing Trends among Consultants and General Practitioners in Sharjah-UAE	Pharmacology & pharmacy	2,015	cc-by	3,476	Drug Prescribing Trends among Consultants and General Practitioners in Sharjah-UAE Copyright © 2015 by authors and Scientific Research Publishing Inc. This work is licensed under the Creative Commons Attribution International License (CC BY). http://creativecommons.org/licenses/by/4.0/ Copyright © 2015 by authors and Scientific Research Publishing Inc. This work is licensed under the Creative Commons Attribution International License (CC BY). http://creativecommons.org/licenses/by/4.0/ How to cite this paper: Sharif, S.I., Fazli, H., Tajrobehkar, Y., Namvar, Z. and Bugaighis, L.M.T. (2015) Drug Prescribing Trends among Consultants and General Practitioners in Sharjah-UAE. Pharmacology & Pharmacy, 6, 374-379. http://dx.doi.org/10.4236/pp.2015.68038 Corresponding author. harmacology & Pharmacy, 2015, 6, 374-379 ublished Online August 2015 in SciRes. http://www.scirp.org/journal/pp ttp://dx.doi.org/10.4236/pp.2015.68038 Drug Prescribing Trends among Consultants and General Practitioners in Sharjah-UAE Suleiman I. Sharif, Hoda Fazli, Yasamin Tajrobehkar, Zeinab Namvar, Laila M. T. Bugaighis Department of Pharmacy Practice & Pharmacotherapeutics, College of Pharmacy, University of Sharjah, Sharjah, United Arab Emirates Email: *sharifsi@sharjah.ac.ae Received 28 July 2015; accepted 21 August 2015; published 24 August 2015 Copyright © 2015 by authors and Scientific Research Publishing Inc. This work is licensed under the Creative Commons Attribution International License (CC BY). http://creativecommons.org/licenses/by/4.0/ Abstract Pharmacology & Pharmacy, 2015, 6, 374-379 Published Online August 2015 in SciRes. http://www.scirp.org/journal/pp http://dx.doi.org/10.4236/pp.2015.68038 Pharmacology & Pharmacy, 2015, 6, 374-379 Published Online August 2015 in SciRes. http://www.scirp.org/journal/pp http://dx.doi.org/10.4236/pp.2015.68038 Pharmacology & Pharmacy, 2015, 6, 374-379 Published Online August 2015 in SciRes. http://www.scirp.org/journal/pp http://dx.doi.org/10.4236/pp.2015.68038 Prescription Analysis, Trends, Consultants, General Practitioners Prescription Analysis, Trends, Consultants, General Practitioners Prescription Analysis, Trends, Consultants, General Practitioners 1. Introduction Inappropriate use of medications in healthcare facilities is a common problem not only in developing but also in developed countries [1]-[4]. However, the situation is worse in developing countries as in addition to the delete- rious effects on health care outcomes; it also exhausts the limited health budget. In the latter countries, the ex- cessive and indiscriminate use of drugs significantly and advertently influences health care delivery as it in- creases the incidence and seriousness of adverse drug reactions and interactions and bacterial resistance [1]. The physician has to prescribe medicines that meet the patient clinical need at doses adequate to cover their individ- ual requirements for adequate period of time and at affordable cost [5]. Dispensing errors can be dangerous [6]-[8]. This entails the dispenser to provide the patient with the appropriate strengths of the prescribed drugs and clearly indicating to the patient the appropriate dose and how to measure, administer and frequently use the medicine. Changing a medication to ensure affordability requires a consultation of the prescriber. The dispenser, the pharmacist, must also provide some instruction on proper storage of medicines. The role of the consumer requires strict adherence to the instruction given by both the prescriber and the dispenser to avoid over dosage, under treatment, drug interactions and degradation of active constituents of medicines due to bad storage condi- tions. Among the easiest ways to examine an aspect of drug misuse is prescription analysis. Such investigation examines the patterns of drug prescribing. Moreover, prescription analysis whether of private practice or that of health facilities can determine areas for improvement towards rational drug use. Moreover, feeding back results of prescription analysis to the actual prescribers and health authorities is a useful method of intervention [1]. In the present study, prescribing behavior, dispensing and use of medicines will be approached to pinpoint various factors that may influence rational drug use. 2. Methods For studying the prescribing trends among consultants in a general hospital, and general practitioners, a total 1239 and 980 prescriptions were collected during the month of April, 2011 from a general hospital out-patients pharmacy and a community pharmacy in Emirate of Sharjah respectively. The prescriptions from the hospital were issued by consultants while those of the pharmacy represent general practitioners (GPs). Prescriptions were subjected to analysis using the World Health Organization (WHO) suggested indicators [9]. These include pre- scriber’s information such as name, registration number and signature; patient’s information including the name, age, sex, and address of patient plus brief diagnosis and history of allergy. All prescriptions were examined for eligibility. In addition information on the medicine were determined including dosage regimen, number of en- counters per prescription, % generic drugs prescribed, most common therapeutic classes and the most common drug of each class. Prescriptions were also examined for medications prescribed as injections. In this study, names of patients were concealed in consideration of patients’ privacy protection. Data were collected and ana- lyzed using Microsoft Excel® and expressed in terms of both counts and percentages. Abstract Background: Inappropriate prescribing can lead to errors in dispensing medications and serious problems for patients. Objectives: Prescription analysis can identify such drawbacks of prescrib- ing, increase awareness of prescribers of rational prescribing and consequently lead to proper de- livery of pharmaceutical care and enhance therapeutic outcomes. Methods: In the present study, prescriptions issued by consultants from a hospital and by general practitioners from private practice in Sharjah-United Arab Emirates were analyzed using indicators suggested by World Health Organizations. These include information with regard to prescriber, patient and the medi- cation prescribed. We also determined the average number of drugs/encounter and % of pre- scriptions with antibiotics and those with injections. Data were collected and analyzed using Mi- crosoft Excel® and expressed in terms of both counts and percentages. Results: Almost all pre- scriptions were handwritten with easily readable ones being 65% for consultants and 46% for general practitioners. Average number of drugs/encounter was 2.1 and 2.8 for consultants and general practitioners, respectively. Antibiotics were prescribed in 27% and 44%; generic pre- scribing was 5% and 10% by consultants and general practitioners respectively and 8% of pre- scriptions by consultants contained injections. Variable results were obtained on information re- garding the patient but consultants seem to be better in documenting patient’s age and gender. Consultants and general practitioners tend to prescribe 3 drugs and more in 35% and 25% re- spectively. The most commonly prescribed therapeutic classes for both groups of prescribers were NSAIDs and antibiotics with ibuprofen and amoxicillin-clavulanic acid combination being the most commonly prescribed drugs of each class. Conclusion: To improve prescription writing, in- terventions must include, among others, incorporation of topics on prescription writing in medi- cal curriculum and programs of continuing medical education. S. I. Sharif et al. Keywords 3. Results In the present study, most prescriptions were hand written with 35% - 50% of them being difficult to read mak- ing dispensing at increased risk of errors. All prescriptions issued by consultants or general practitioners (GPs) were deficient in important information regarding the prescriber, the patient and the dosage regimen (Table 1 and Table 2). The average number of drugs per prescription was 2.1 for consultants and 2.8 for general practi- tioners. Prescribing generic drugs was higher in the case of GPs (10%) as compared to consultants (5%). The trend for poly-pharmacy was more evident in prescriptions by consultants where prescriptions with 4 or more drugs totaled to 16% (Table 3). Prescriptions by consultants contained injections (8%) as compared to none of those issued by GPs. As shown in Table 1, prescribing two drugs of the same therapeutic class was more in case of GPs as it counted to 16% of all prescriptions while it was evident in only 3% of prescriptions by consultants. 375 S. I. Sharif et al. Table 1. Comparison of % presence of prescriber’s information, eligibility of prescription, average number of drugs/encounter, injections, antibiotics and medication information in prescriptions by consultants and general practitioners. Information Number (%) of prescriptions Consultants (n = 1239) General practitioners (n = 980) Prescriber’s Name 1090 (88%) 813 (83%) Signature 1115 (90%) 941 (96%) Specialty 991 (80%) 725 (74%) License’s number 558 (45%) 333 (34%) Prescriptions Typed 0 (0%) 59 (6%) Hand written 1090 (100%) 921 (94%) Easily readable 805 (65%) 451 (46%) Medication Average number of drugs/encounter 2.1 2.8 Dose 620 (50%) 892 (91%) Route of administration 372 (30%) 186 (19%) Duration of treatment 942 (76%) 608 (62%) Generic drugs 62 (5%) 98 (10%) Injections 37 (3%) 0 (0%) Antibiotics 335 (27%) 431 (44%) Two drugs of the same class 37 (3%) 431 (44%) Possible drug-drug interaction 50 (4%) 44 (4.5%) T bl 2 C i f ti t’ i f ti t i i ti b lt t d l titi Table 1. Comparison of % presence of prescriber’s information, eligibility of prescription, average number of drugs/encounter, injections, antibiotics and medication information in prescriptions by consultants and general practitioners. Information Number (%) of prescriptions Consultants (n = 1239) General practitioners (n = 980) Table 1. 4. Discussion The present study aimed at comparing prescribing behavior of consultants in a hospital and GPs in private clin- ics. Prescription analysis can indicate areas of irrational drug use practice. Name of patient was deficient in very small % of prescriptions by both consultants and GPs. These results are similar to those reported for Saudi Ara- bia [2] [10]. Age of patient was not mentioned in 70% of prescriptions by GPs as compared to 10% of encoun- ters by consultants. In the present study, lack of sex and age from prescriptions by consultants was within the ranges mentioned in the above Saudi studies. Gender of patients was not mentioned in 64% of prescriptions by GPs as compared to 12% of those by con- sultants. Such information are essential for pharmacists dispensing these prescriptions to check on whether doses prescribed were appropriate or not and also in case of female patients to take precautions in cases of pregnancy and lactation. Similar to previous observations in Saudi Arabia [2], patient’s address was present in only 14% of prescriptions by GPs and not mentioned at all in those by consultants. This is alarming since both the prescribing physician and the dispensing pharmacist need to immediately contact the patient in case of either prescription or dispensing error. In the present study diagnosis and history of allergy were completely omitted from all studied prescriptions. This is in contrast to results reported by others in Saudi Arabia [2] [10] [11] but similar to our previous observations in Sharjah [12]. In the present study, the name, signature, specialty and license number of the prescriber were not mentioned in small % of prescriptions. The percentage lack of such information was rather similar with both groups of pre- scribers. Most prescriptions analyzed were hand written with readability being of 65% and 46% for consultants and GPs respectively. Our results are less than those reported for Saudi Arabia [2] but far in excess of those re- ported in other studies [3] [10]. Poor legibility of handwriting can lead to misinterpretation by the pharmacist and result in errors in drug dispensing and administration [13] with consequent risks to the patient. The average number of drugs per prescription for consultants was only slightly higher than that recommended by WHO [9]. However, in prescriptions by GPs the number of drugs/encounter was 2.8 which is similar to that described in India [14] [15]. 3. Results Comparison of % presence of prescriber’s information, eligibility of prescription, average number of drugs/encounter, injections, antibiotics and medication information in prescriptions by consultants and general practitioners. Number (%) of prescriptions Table 2. Comparison of patient’s information present in prescriptions by consultants and general practitioners. Table 2. Comparison of patient’s information present in prescriptions by consultants and general practitioner Table 2. Comparison of patient’s information present in prescriptions by consultants and general practitioners. Patient’s information Number (%) of prescriptions Consultants (n = 1239) General practitioners (n = 980) Name 1202 (97%) 941 (96%) Age 1115 (90%) 294 (30%) Gender 1090 (88%) 353 (36%) Address 0 (0%) 137 (14%) Brief diagnosis 0 (0%) 7 (0.7%) History of allergy 0 (0%) 0 (0%) Table 3. Number and % of prescriptions with various numbers of medications issued by consultants and general practitioners. Number of drugs Number (%) of prescriptions Consultants (n = 1239) General practitioners (n = 980) No drugs 25 (2%) 0 (0%) One drugs 434 (35%) 470 (48%) Two drugs 319 (28%) 265 (27%) Three drugs 235 (19%) 157 (16%) Four drugs 99 (8%) 59 (6%) More than 4 drugs 99 (8%) 29 (3%) 376 S. I. Sharif et al. Drug interactions were observed in 4% and 4.5% of encounters by consultants and GPs respectively (Table 1) counted to The most commonly prescribed class of drugs was the NSAIDs constituting 30% and 51% for con- sultants and general practitioners respectively (Table 4). In both cases ibuprofen was the most commonly pre- scribed of this class as more than 50% of prescriptions contained this drug. Antibiotics come second on the list. These drugs were prescribed in 27% and 44% of prescriptions by consultants and GPs respectively. The combi- nation of amoxicillin and clavulanic acid was the favorite drug. The third class of drugs prescribed by consul- tants was gastrointestinal drugs (22%) and vitamins (36%) for general practitioners (Table 4). 4. Discussion In the present study, the % of encounters with more than 4 drugs for both consul- tants (8%) and GPS (6%) was markedly less than that observed in encounters by consultants (25%) in our earlier study in Sharjah [12]. Surprisingly, the dose of the drug was not mentioned in only 9% of GPs prescriptions as compared to 50% of prescriptions by consultants. These results on GPs patterns of prescribing seem, to some extent, better than those Table 4. Most commonly prescribed therapeutic classes and drug of each class in prescriptions by consultants and general practitioners. The most commonly prescribed therapeutic class and drug of each class Number (%) of prescriptions Consultants (n = 1239) General practitioners (n = 980) NSAIDs 372 (30%) NSAIDs 500 (51%) Ibuprofen 644 (52%) Ibuprofen 666 (68%) Antibiotics 335 (27%) Antibiotics 431 (44%) Amoxicillin + clavulanic acid 818 (66%) Amoxicillin + clavulanic acid 843 (86%) GI drugs 273 (22%) Vitamins 353 (36%) Hyoscine-N-butylbromide 966 (78%) Multi-vitamins 82 (84%) 377 S. I. Sharif et al. of consultants. Reasons for such negligence could be attributed to overconfidence of consultants, ignorance, short consultation time or stressful conditions of work. None of these can be an accepted justification for such behavior. Route of administrations was lacking in about 70% and 80% of prescriptions by consultants and GPs respectively. On the other hand, duration of treatment was lacking in 24% and 38% of prescriptions by consul- tants and GPs respectively. Such prescribing trends of both consultants and GPs although it is not appropriate; it strongly emphasizes the role of pharmacists in providing rational pharmaceutical care through complementing the prescriber’s deficiencies, if any, in dispensed prescriptions. Generic prescribing was low in prescriptions by both groups of prescribers and even lower than that reported in our earlier study in Sharjah [12] and in some Western countries [3]-[5]. This is may be attributable to a more influential role of pharmaceutical promotional activities and medical insurance strategies in UAE. Inability of patients to purchase costly medications significantly contributes to the problem of noncompliance and prescrib- ing generic drugs should be advocated among prescribers unless there is an issue of bioavailability. None of the prescriptions by GPS included injections whereas 8% of those issued by consultants contained injections. This may increase the possibility of non-compliance in case of expensive injectable formulations that cannot be afforded by the patient. 5. Limitations of the Study The time constraint and difficulty in obtaining prescriptions did not allow us to analyze samples of prescriptions issued by consultants and GPs in other Emirates, therefore the results cannot be generalized to other Emirates than Dubai. Future studies will focus on prescriptions collected from various Emirates and hospitals. 4. Discussion Analysis of the number of drugs per encounter demonstrated that the inci- dence of poly-pharmacy is also higher in prescriptions issued by consultants than those by GPS. Such trend in- creases the risk of adverse effects and drug interactions. The later was evident in about 4% of encounters by both groups of prescribers. In general, one would expect a better prescribing pattern by consultants. Unfortu- nately, in the present study, this was not always the case. Focus in programs of medical continuing education on prescription writing may help improving the patterns of prescribing of both groups of prescribers. The most commonly prescribed drugs were NSAIDs followed by antibiotics. The latter were reasonably pre- scribed by consultants (27%). This is similar to our earlier findings in a hospital in Dubai [1]. On the other hand, GPs prescribed antibiotics in more than 40% of prescriptions. This is still far less than that prescribed in India [15]. The third common class was gastrointestinal drugs for consultants and multivitamins for GPs. No other medications were included in the prescriptions with multivitamins and the extensive use of multivitamins may be explained by the insistence of patients on receiving drugs to feel better. Despite the fact that in some aspects of prescribing a more appropriate trend was shown by GPs than by con- sultants, prescribing by both groups still not ideal. Periodical monitoring of prescriptions can always be benefi- cial as it can pinpoint areas for improvement towards rational drug prescribing. 6. Conclusion A prescription should be wisely and appropriately written to include all the information that would be of help to the pharmacist who dispenses the medications and the patient who uses them. Periodical monitoring of prescrip- tions is useful as it can identify areas for improvement in prescription writing. Moreover, we suggest that inter- ventions to improve prescribing behavior may include among others, feedback of results to the prescribers through personal interviews and incorporation of principles of rational prescription writing in medical curricu- lum and continuing medical education programs for both GPs and consultants. [1] Sharif, S.I., Al-Shaqra, M., Hajjar, H., Shamout, A. and Wess, L. (2008) Patterns of Drug Prescribing in a Hospital in Dubai, United Arab Emirates. Libyan Journal of Medicine, 3, 10-12. http://dx.doi.org/10.4176/070928 [2] Irshaid, Y.M., Al Homrany, M., Hamdi, A.A., Adjepon-Yamoah, K.K. and Mahfouz, A.A. (2005) Compliance with Good Practice in Prescription Writing at Outpatient Clinic in Saudi Arabia. Eastern Mediterranean Health Journal, 11, 922-928. [3] Meyer, T.A. (2000) Improving the Quality of the Order-Writing Process for Inpatient Orders and Outpatient Prescrip- tions. American Journal of Health-System Pharmacy, 57, S18-S12. [4] Montastruc, F., Gardette, V., Cantet, C., Piau, A., Lapeyre-Mestre, M., Vellas, B., Montastruc, J.-L. and Andrieu, S. 4] Montastruc, F., Gardette, V., Cantet, C., Piau, A., Lapeyre-Mestre, M., Vellas, B., Montastruc, J.-L. and [3] Meyer, T.A. (2000) Improving the Quality of the Order-Writing Process for Inpatient Orders and Outpatient Prescrip- tions. American Journal of Health-System Pharmacy, 57, S18-S12. [2] Irshaid, Y.M., Al Homrany, M., Hamdi, A.A., Adjepon-Yamoah, K.K. and Mahfouz, A.A. (2005) Compliance with Good Practice in Prescription Writing at Outpatient Clinic in Saudi Arabia. Eastern Mediterranean Health Journal, 11, 922-928. [1] Sharif, S.I., Al-Shaqra, M., Hajjar, H., Shamout, A. and Wess, L. (2008) Patterns of Drug Prescribing in a Hospital in Dubai, United Arab Emirates. Libyan Journal of Medicine, 3, 10-12. http://dx.doi.org/10.4176/070928 References [2] Irshaid, Y.M., Al Homrany, M., Hamdi, A.A., Adjepon-Yamoah, K.K. and Mahfouz, A.A. (2005) Compliance with Good Practice in Prescription Writing at Outpatient Clinic in Saudi Arabia. Eastern Mediterranean Health Journal, 11, 922-928. [3] Meyer, T.A. (2000) Improving the Quality of the Order-Writing Process for Inpatient Orders and Outpatient Prescrip- tions. American Journal of Health-System Pharmacy, 57, S18-S12. 378 S. I. Sharif et al. (2013) Potentially Inappropriate Medication Use among Patients with Alzheimer Disease in the REAL.FR Cohort: Be Aware of Atropinic and Benzodiazepine Drugs! European Journal of Clinical Pharmacology, 69, 1589-1597. http://dx.doi.org/10.1007/s00228-013-1506-8 [5] DeVries, T.P., et al., Eds. (1995) Guide to Good Prescribing: A Practical Manual. World Health Organization, Geneva, 51-5 (WHO/DAP/94.11). [6] (1983) The Pharmaceutical Professions and Institutions. UAE Federal Law No: 4. [7] Brahams, D. (1984) Legal Liability and the Negligent Prescription. Practitioner, 228, 444-445. [8] Yousif, E., Ahmed, A.M., Abdalla, M.E. and Abdelgadir, M.A. (2006) Deficiencies in Medical Prescriptions in a Su- danese Hospital. Eastern Mediterranean Health Journal, 12, 915-918. [9] WHO (1995) How to Investigate Drug Use in Health Facilities: Selected Drug Use Indicators. World Health Organiza- tion, Geneva, WHO/DAP/93.1.1995. [10] Balbaid, O.M. and Al-Dawood, K.M. (1988) Assessment of Physician’s Prescribing Practices at Ministry of Health Hospitals in Jeddah City, Saudi Arabia. Saudi Medical Journal, 19, 28-35. [11] Bawazir, S.A. (1992) Prescribing Patterns at Community Pharmacies in Saudi Arabia. International Pharmacy Journal, 6, 222-224. [12] Sharif, S.I., Alabdouli, A.H. and Sharif, R.S. (2013) Drug Prescribing Trends in a General Hospital in Sharjah-United Arab Emirates. American Journal of Pharmacological Sciences, 1, 6-9. http://dx.doi.org/10.12691/ajps-1-1-2 [13] Velo, G.P. and Minuz, P. (2009) Medication Errors: Prescribing Faults and Prescription Errors. British Journal of Clinical Pharmacology, 67, 624-628. http://dx.doi.org/10.1111/j.1365-2125.2009.03425.x [14] Bapna, J.S., Tekur, U., Gitanjali, B., Shashindran, C.H., Pradhan, S.C., Thulasimani, M., et al. (1992) Drug Utilization at Primary Health Care Level in Southern India. European Journal of Clinical Pharmacology, 43, 413-415. http://dx.doi.org/10.1007/BF02220618 [15] Kshirsagar, M.J., Langade, D., Patil, S. and Patki, P.S. (1998) Prescribing Patterns among Medical Practitioners in Pune. Bulletin World Health Organization, India. 379
https://openalex.org/W2898033497	https://europepmc.org/articles/pmc6180453?pdf=render	English	null	The meaning of blood pressure	Critical care	2,018	cc-by	9,746	Elastic energy Three types of energy produce arterial pressure: elastic, kinetic, and gravitational. By far the most significant is elastic energy. The volume inside vascular structures stretches their elastic walls and produces a recoil force, which, based on the elastic properties of the structure, cre- ates a pressure. The materials making up vascular struc- tures are not homogenous so that the volume-to-pressure relationship of arterial vessels is not linear and has a con- vex curvilinearity [2–4] (Fig. 1). Resistance to stretch of a substance is called elastance and the inverse, the ease of stretch, compliance. Normal blood flow is pulsatile be- cause of the cyclic nature of cardiac emptying and filling. The consequent cyclic changes in the volume of the aorta © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. The meaning of blood pressure S. Magder in a single direction, whereas pressure can be used over the curved surfaces of vessels and has the units of force per cross-sectional area. Force is the product of mass and acceleration, and the standard unit is a Pascal, which is a newton per square meter. However, vascular pres- sures still most often are measured in length-based units of millimeters of mercury or centimeters of water. This has historical origins. Before the availability of electronic transducers, pressures were measured with columns of water or mercury. The mass of the column is the prod- uct of the volume and density. The density of water is 1 and that of mercury is 13.6 times that of water. The height of a column of fluid is proportional to the volume over the cross-sectional area of the column, and thus has units of length. Pressure, therefore, is proportional to the product of the density of the fluid and the height of the fluid and gives force per cross-sectional area. The force on the column of water, or mercury, is the acceler- ation of the column by gravity. These “length” measure- ments of force are relative, for they depend upon the position on the earth relative to the center of the earth. However, the acceleration due to gravity is similar over all the earth’s earth. It is even only about 0.2% lower at the top of Mt. Everest. Thus, length-based units are still useful for biological measurements. Units of millimeters of mercury (mmHg) are converted to kilopascals by multiplying by 0.13. Background Blood pressure is one of the most commonly measured clinical parameters and blood pressure values are major determinants of therapeutic decisions. However, interpret- ation of the physiological meaning of blood pressure in an individual patient is not always an easy task. This paper reviews the physical basis and physiological determinants of arterial pressure, and the relationship of arterial pres- sure to tissue perfusion. Some of the issues have been cov- ered in a previous review on blood pressure [1]. The objective of this paper is to provide guidance when con- sidering therapeutic options but it is not possible to give a definitive algorithm with current knowledge. Open Access Open Access Abstract Measurement of arterial pressure is one of the most basic elements of patient management. Arterial pressure is determined by the volume ejected by the heart into the arteries, the elastance of the walls of the arteries, and the rate at which the blood flows out of the arteries. This review will discuss the three forces that determine the pressure in a vessel: elastic, kinetic, and gravitational energy. Emphasis will be placed on the importance of the distribution of arterial resistances, the elastance of the walls of the large vessels, and critical closing pressures in small arteries and arterioles. Regulation of arterial pressure occurs through changes in cardiac output and changes in vascular resistance, but these two controlled variables can sometimes be in conflict. Keywords: Arterial resistance, Conductance, Critical closing pressure, Cardiac output, Blood flow, Perfusion, Gravitational energy, Elastic energy, Kinetic energy Magder Critical Care (2018) 22:257 https://doi.org/10.1186/s13054-018-2171-1 Magder Critical Care (2018) 22:257 https://doi.org/10.1186/s13054-018-2171-1 Correspondence: sheldon.magder@mcgill.ca Department of Critical Care, McGill University Health Centre, 1001 Decarie Blvd., Montreal, Quebec H4A 3J1, Canada Physical basis of vascular pressures Pressure is a force distributed over a surface area and, as such, it has the same units as tension. The term pressure is used instead of tension because tension is determined Correspondence: sheldon.magder@mcgill.ca Department of Critical Care, McGill University Health Centre, 1001 Decarie Blvd., Montreal, Quebec H4A 3J1, Canada Fig. 1 Effect of age and initial volume on thoracic aortic elastance. The slopes of the lines are elastance. The right upper insert shows the increase in circumferential tension versus increases in aortic circumference in percent for age < 18 to > 80 years [33]. The lower left shows a schematic pressure–volume relationship for the aorta. The boxes represent stroke volumes. The same stroke volume A starting from the same initial volume produces increasing pulse pressures depending upon the shape and position of the start of the stroke volume. The stroke volume B is the same size as in A but starts at a higher initial volume and produces a much larger pulse pressure Fig. 1 Effect of age and initial volume on thoracic aortic elastance. The slopes of the lines are elastance. The right upper insert shows the increase in circumferential tension versus increases in aortic circumference in percent for age < 18 to > 80 years [33]. The lower left shows a schematic pressure–volume relationship for the aorta. The boxes represent stroke volumes. The same stroke volume A starting from the same initial volume produces increasing pulse pressures depending upon the shape and position of the start of the stroke volume. The stroke volume B is the same size as in A but starts at a higher initial volume and produces a much larger pulse pressure Fig. 1 Effect of age and initial volume on thoracic aortic elastance. The slopes of the lines are elastance. The right upper insert shows the increase in circumferential tension versus increases in aortic circumference in percent for age < 18 to > 80 years [33]. The lower left shows a schematic pressure–volume relationship for the aorta. The boxes represent stroke volumes. The same stroke volume A starting from the same initial volume produces increasing pulse pressures depending upon the shape and position of the start of the stroke volume. The stroke volume B is the same size as in A but starts at a higher initial volume and produces a much larger pulse pressure produce the cyclic changes in arterial pressure. Although the elastance of the wall of arterial vessels varies with vol- ume, over short periods of time the actual curvilinear rela- tionship of volume to pressure is constant because it is determined by the composition of the vascular wall [2, 4]. Changes in this curvilinear relationship of aortic elastance require changes in the matrix of the wall which do not occur acutely, but rather occur over time with chronic processes such as long-standing hypertension and aging. very thin relative to the radius of the structure, as is the case of a soap bubble [5]. Thus, although commonly used, this simplification is not valid for vascular struc- tures and the full equation for the assessment of wall tension must be used with the pressure values relative to absolute zero pressure [6, 7]. When tension is calculated this way, the tension across the wall is a negative value in most vessels, which means that vessels are tending to explode rather than collapse, and wall tensions are actu- ally more or less negative values. g g y g g Pressure must be measured relative to a reference value which is defined as zero. Most often the reference for zero pressure is atmospheric pressure because this is the pres- sure that surrounds the body. The pressure inside a vessel relative to the pressure outside a vessel gives the pressure that distends the wall of the vessel and is called transmural pressure. For example, if the outside atmospheric pressure is called zero and the pressure inside the aorta is 120/ 80 mmHg, the transmural pressure simply is 120/80 mmHg. However, if atmospheric pressure is 760 mmHg, the real pressure across the arterial wall relative to absolute zero pressure is 880/840 mmHg, but if this absolute value were to be used to determine if transmural pressure changed, one would have to first determine if atmospheric pressure had changed! Page 2 of 10 Magder Critical Care (2018) 22:257 Fig. 1 Effect of age and initial volume on thoracic aortic elastance. The slopes of the lines are elastance. The right upper insert shows the increase in circumferential tension versus increases in aortic circumference in percent for age < 18 to > 80 years [33]. The lower left shows a schematic pressure–volume relationship for the aorta. The boxes represent stroke volumes. The same stroke volume A starting from the same initial volume produces increasing pulse pressures depending upon the shape and position of the start of the stroke volume. The stroke volume B is the same size as in A but starts at a higher initial volume and produces a much larger pulse pressure Kinetic energy h d f The second force determining arterial pressure is kinetic energy, which is due to the velocity of the flowing blood [8]. Kinetic energy is equal to the product of one half the mass (m) of the blood, which is the product of the volume and density of the blood, and the square of blood velocity (v): Kinetic energy ¼ 1=2 m v2 : Velocity of flowing blood is in units of distance over time. The product of the velocity of blood and the cross-sectional area of a vessel gives flow of blood in units of volume per time. Kinetic energy contributes only about 3% of the total force at the peak of normal systolic pressure, but kinetic energy makes up a greater proportion of the pressure in large veins and pulmonary Tension across vessel walls often is calculated with use of the Laplace relationship and the value of pressure in- side the vessel relative to atmospheric pressure. How- ever, the Laplace relationship assumes that the wall is Magder Critical Care (2018) 22:257 Page 3 of 10 Page 3 of 10 is flowing from an area of lower pressure to an area of higher pressure when a catheter facing the flow is used. vessels because the velocity of blood is similar to that in the large arteries, whereas the elastic energy is much smaller. Kinetic energy likely has a greater role in septic pa- tients with high cardiac outputs, for the higher flow means that there is a greater kinetic component and, at the same time, elastic energy is decreased by the vaso- dilation. This will produce a difference between pressure measured with intravascular catheters facing the flow and the pressure measured with a non-invasive device which only measures the lateral elastic component of energy. Furthermore, the decreased lateral elastic force could alter myogenic responses whereas the increased velocity in small vessels will alter shear stress and the two could alter proper matching of flow to the tissue’s metabolic needs. Kinetic energy can produce some confusing results, in- cluding blood seeming to flow from a lower to a higher pressure and the appearance that blood is flowing uphill! This occurs because flow is based on the total energy difference across a system, and not just the difference in elastic energy. Examples of this occur when sections of vessels either widen or narrow. Gravitational energy Th i f The importance of the gravitational component of the energy for blood flow is important when pressure is measured with a fluid filled system. This is because the position of the transducer and the choice of the refer- ence level have a large impact on the measured value and it is essential that the reference level is standardized. The gravitational effect on the body is very significant in the upright position. For example (Fig. 3), in a person who is 182 cm tall, and who has a systolic pressure of 110/70 mmHg and mean pressure of 83 mmHg mea- sured at the level of the heart, the pressure measured with a transducer placed at the top of the head is only Fig. 2 Pressure measurements in a vessel with an aneurysmal region. Pressures measured with fluid filled tubes facing the flow [1, 3, 5] measure elastic and kinetic energy, whereas tubes with the opening perpendicular to the flow just measure lateral pressure [2, 4, 6]. An assumption is that the energy loss due to resistance (dashed line) is minimal. In A, the tube facing the flow [1] shows a higher pressure than the tube measuring lateral pressure [2] because it includes kinetic energy. In B the vessel diameter is larger and velocity of flow is slower. The kinetic energy converts into elastic energy and the difference between tubes 3 and 4 is much smaller than between 1 and 2. In C, the tube narrows again so that kinetic energy increases and lateral energy decreases, which again increases the difference between 5 and 6 Fig. 2 Pressure measurements in a vessel with an aneurysmal region. Pressures measured with fluid filled tubes facing the flow [1, 3, 5] measure elastic and kinetic energy, whereas tubes with the opening perpendicular to the flow just measure lateral pressure [2, 4, 6]. An assumption is that the energy loss due to resistance (dashed line) is minimal. In A, the tube facing the flow [1] shows a higher pressure than the tube measuring lateral pressure [2] because it includes kinetic energy. In B the vessel diameter is larger and velocity of flow is slower. The kinetic energy converts into elastic energy and the difference between tubes 3 and 4 is much smaller than between 1 and 2. Kinetic energy h d f Figure 2 shows an ex- ample of an aneurysmal dilatation of a vessel. Pressure is measured with a fluid filled catheter with the opening fa- cing the oncoming flow, as is the practice with most ar- terial catheters, and another catheter that has an opening perpendicular to the flow (side pressure). Flow in L/min must be the same in each section of the vessel because what goes in must go out to maintain conserva- tion of mass. However, in the region of the aneurysmal dilatation where the diameter is much larger, velocity is much slower because the cross-sectional area changes with the square of the radius. Energy cannot be created nor destroyed so the decrease in kinetic energy is con- verted into elastic energy. This increases the pressure on the wall of the dilated area and leads to further dilatation and a further increase in pressure on the already weak- ened wall until the wall stretches to a critical value and ruptures. In the vessel segment distal to the aneurysm, the velocity is again higher and elastic energy is con- verted back into kinetic energy so that it looks like blood Resistances Resistance to flow in a tube is given by Poiseuille’s law, which says that, in a tube with laminar flow, the resistance, which is the frictional loss of energy, is determined by the length of the tube, the viscosity of the blood, and the in- verse of the radius of the tube raised to the fourth power [4]. Vessel radius is thus the dominant determinant of re- sistance and the only factor that can significantly change rapidly. The total resistance of tubes in series is determined by summing up all the individual resistances in the series; in contrast, the sum of parallel resistances is determined by: 1/Rtotal = 1/R1 + 1/R2 + 1/R3…1/Rn Fig. 3 Gravitation effect on arterial pressures (adapted from [9]). The numbers on the right in mmHg refer to the gravitational potential energy related to the difference between the base of the measuring device relative to the mid-point of the right atrium (dashed line) assuming a 182-cm male. The loss of pressure due to resistance is assumed to be 5 mmHg. When the transducer is set at the level of the right atrium, the mean pressure is 83 mmHg. At the top of the head the pressure would only be 39 mmHg. If the transducer was at the level of the foot, the pressure would be a mean of 171 mmHg Fig. 3 Gravitation effect on arterial pressures (adapted from [9]). The numbers on the right in mmHg refer to the gravitational potential energy related to the difference between the base of the measuring device relative to the mid-point of the right atrium (dashed line) assuming a 182-cm male. The loss of pressure due to resistance is assumed to be 5 mmHg. When the transducer is set at the level of the right atrium, the mean pressure is 83 mmHg. At the top of the head the pressure would only be 39 mmHg. If the transducer was at the level of the foot, the pressure would be a mean of 171 mmHg p 1/Rtotal = 1/R1 + 1/R2 + 1/R3…1/Rn This is because the greater the number of parallel channels, the greater the overall cross-sectional area, and the greater the overall effective radius. Resistances vary among different vascular beds. Factors include the size of the vascular bed and the density of vessels. Be- cause of their sizes, the splanchnic and muscle beds have overall low vascular resistances. Gravitational energy Th i f In C, the tube narrows again so that kinetic energy increases and lateral energy decreases, which again increases the difference between 5 and 6 Fig. 2 Pressure measurements in a vessel with an aneurysmal region. Pressures measured with fluid filled tubes facing the flow [1, 3, 5] measure elastic and kinetic energy, whereas tubes with the opening perpendicular to the flow just measure lateral pressure [2, 4, 6]. An assumption is that the energy loss due to resistance (dashed line) is minimal. In A, the tube facing the flow [1] shows a higher pressure than the tube measuring lateral pressure [2] because it includes kinetic energy. In B the vessel diameter is larger and velocity of flow is slower. The kinetic energy converts into elastic energy and the difference between tubes 3 and 4 is much smaller than between 1 and 2. In C, the tube narrows again so that kinetic energy increases and lateral energy decreases, which again increases the difference between 5 and 6 Magder Critical Care (2018) 22:257 Page 4 of 10 Page 4 of 10 Fig. 3 Gravitation effect on arterial pressures (adapted from [9]). The numbers on the right in mmHg refer to the gravitational potential energy related to the difference between the base of the measuring device relative to the mid-point of the right atrium (dashed line) assuming a 182-cm male. The loss of pressure due to resistance is assumed to be 5 mmHg. When the transducer is set at the level of the right atrium, the mean pressure is 83 mmHg. At the top of the head the pressure would only be 39 mmHg. If the transducer was at the level of the foot, the pressure would be a mean of 171 mmHg outflow and falls after the peak of ejection because out- flow exceeds inflow. The outflow is dependent upon the resistance emptying the arterial tree and the elastance of the vessel walls. The product of the inverse of elastance (compliance) and the downstream resistance gives the time constant of emptying of the arterial vessels. The time constant is the time it takes to get to 63% of a new steady state after a step change in flow or pressure. Gravitational energy Th i f Time constants are important in pulsatile systems because they set the amount of filling and emptying of aortic vol- ume that can occur based on the cardiac frequency, the proportions of contraction and relaxation times during systole, and the diastole time. Resistances However, when flows are related to the mass of tissue, muscle tissue has a high baseline resistance because the flow per mass is low. The importance of this is that the change in flow in dif- ferent vascular beds with a fall in arterial pressure de- pends upon the slope of the pressure–flow line in that region [1]. The steeper the slope of the relationship, the greater the fall in flow for a given decrease in pressure. The kidney starts with a very steep pressure–flow rela- tionship, whether assessed by mass or as a proportion of total body cardiac output, and it has a small capacity to dilate further [10] (Fig. 4). about 66/26 mmHg with a mean of 39 mmHg. On the other hand, if the transducer is placed at the level of the foot, the pressure would be 198/158 mmHg and a mean of 171 mmHg. It is worth noting that normal pressures for brain perfusion in the upright posture are much below clinically recommended targets but we do not need norepinephrine to walk around! Gravitational energy is not a large factor for assess- ment of arterial pressure in the supine position, but it still is a significant factor for venous return because ven- ous pressures are low, and the pressure difference be- tween the region of systemic venous compliance and the right atrium is in the range of only 4 to 8 mmHg, which is about 6 to 11 cm of height. This can produce differ- ences in venous return in the supine and prone positions and consequently in cardiac output [9]. The important factor in assessing the reserves of flow in a vascular region is the maximum slope of the re- gional pressure–flow line because this indicates the physical limit to flow at a given pressure (Fig. 4). Coron- ary blood flow can increase fivefold above the flow at a resting heart rate of 70 beats per minute. Thus, at low heart rates, the heart has very large blood flow reserves, which allow the heart to tolerate large decreases in arter- ial pressure. However, this is not true when there is a fixed coronary obstruction that limits the decrease in Determinants of arterial pressure The main determinant of arterial pressure is the stretch of the walls of the arteries by the volume they contain. This volume increases in systole because inflow exceeds Magder Critical Care (2018) 22:257 Page 5 of 10 Fig. 4 Flow vs pressure for kidney (left) and heart (right) based on data from hemorrhaged dogs [11]. The dotted line indicates baseline state and the solid line indicates maximal vasodilation with nitroprusside. The initial flow vs pressure line for the kidney is steep and is only a little steeper with vasodilation. The heart starts with a much flatter flow vs pressure line but can increase fivefold in the pressure range of 70–80 mmHg. Note that the peak conductance of flow to the heart is only mildly higher than the baseline conductance to the kidney Fig. 4 Flow vs pressure for kidney (left) and heart (right) based on data from hemorrhaged dogs [11]. The dotted line indicates baseline state and the solid line indicates maximal vasodilation with nitroprusside. The initial flow vs pressure line for the kidney is steep and is only a little steeper with vasodilation. The heart starts with a much flatter flow vs pressure line but can increase fivefold in the pressure range of 70–80 mmHg. Note that the peak conductance of flow to the heart is only mildly higher than the baseline conductance to the kidney Fig. 4 Flow vs pressure for kidney (left) and heart (right) based on data from hemorrhaged dogs [11]. The dotted line indicates baseline state and the solid line indicates maximal vasodilation with nitroprusside. The initial flow vs pressure line for the kidney is steep and is only a little steeper with vasodilation. The heart starts with a much flatter flow vs pressure line but can increase fivefold in the pressure range of 70–80 mmHg. Note that the peak conductance of flow to the heart is only mildly higher than the baseline conductance to the kidney 25 mmHg under baseline conditions [15]. Unfortunately, mean arterial critical closing pressure currently cannot be assessed in an intact person either for the whole body or in local regions. coronary resistance. On the other hand, the capacity to increase the slope of the pressure–flow relationship in the kidney is limited, which makes the kidney very sensi- tive to decreases in blood pressure. Determinants of arterial pressure g When a critical closing pressure is present, use of the right atrial or central venous pressure as the value of downstream pressure for the vasculature produces an important error in the common assessment of vascular resistance. This is because the slope of the true flow ver- sus pressure relationship, i.e., the inverse of resistance, is much steeper than that obtained with this standard cal- culation. Even worse, the error gets larger the lower the pressure or flow because the pressure below the critical closing pressure does not affect flow yet it takes up an increasingly greater proportion of the total pressure used for the calculation. This error makes it look like there is an increase in vascular resistance when flow decreases, which would make sense physiologically as being a defense against a fall in arterial pressure, but it occurs from the measurement error even if there is no actual vasoconstriction. This error makes it difficult to know if a drug such as milrinone improved cardiac output by its inotropic action or because it dilated vessels and reduced afterload. To truly know what happened, it is necessary to have two points on a pressure–flow line, but this can- not be readily obtained in human subjects, and for the matter, it is not easy to obtain in most animal studies. A useful point is that if the cardiac output rises with a rise or no change in arterial pressure, there was a true in- crease in cardiac function. The message is that resistance numbers are of little use and noting the relative change in blood flow and blood pressure is much more useful. Critical closing pressure Resistance to flow through a tube is calculated as the dif- ference between the upstream and downstream pressures, divided by the flow between the two pressures. Accord- ingly, systemic vascular resistance typically is calculated as the difference between aortic mean pressure and the right atrial pressure, or central venous pressure, which usually are the same. This calculation assumes that the vascular system functions as a continuous tube, but this is not true. Most tissues have critical closing pressures at the level of the arterioles. These are also called vascular waterfalls or Starling resistors [11]. The presence of a critical closing pressure creates the same phenomena that exist in veins when the pressure inside a vessel is less than the pressure outside, but in arterioles flow limitation likely is created by the flow characteristics in small vessels without true collapse. When waterfall-like properties exist, the down- stream pressure no longer effects flow, and arterial resist- ance should be calculated from mean arterial pressure to the critical closing pressure, and not to the right atrial pressure. Animal studies suggest that the average critical closing pressure for the whole circulation is around 30 mmHg [12] but the critical closing pressure differs among vascular beds [13]. For example, in resting skeletal muscle the critical closing pressure was estimated to be over 60 mmHg [14]. In the coronary circulation the crit- ical closing pressure likely is in the range of 15 to Magder Critical Care (2018) 22:257 Page 6 of 10 Page 6 of 10 The arteriolar critical closing pressure is increased by a decrease in the carotid sinus pressure and alpha-adrenergic agonists [16, 17]. It is decreased by increased arterial pressure through the myogenic response [18] and by cal- cium channel blockers [19]. It also decreases with reactive hyperemia and exercise-induced hyperemia [14, 20], indi- cating that it also responds to local metabolic activity. elastance) of the aortic wall determine the time constant of arterial emptying and the volume left in the aorta at the end of each cycle. An increase in true aortic elastance (i.e., the shape and position of the whole curve; Fig. 1) is im- portant because it is a determinant of the diastolic pres- sure at which the aortic valve opens, the shape of the pulse pressure, and the speed of the forward and back- ward pressure waves in the aorta [23, 24]. Dynamic elastance l Dynamic elastance has recently become popular. It is ar- gued that it can be a useful measure for assessing the coupling of the heart and circulation [25–28]. It is de- rived from concepts introduced by Sunagawa and co-workers [29, 30], who attempted to derive an equa- tion that relates stroke volume to the mechanical prop- erties of the ventricle and vascular system. Their equations predicted stroke volume based on derived aor- tic and ventricular elastances. Unappreciated require- ments were that ventricular diastolic pressure was considered to be on the flatter part of the ventricular diastolic filling curve, and that heart rate was constant, neither of which can be assured in the intact circulation. When these assumptions are true, the prediction of stroke volume from the formula essentially represented the ascending portion of a cardiac function curve with a constant heart rate, contractility, and afterload. Fig. 5 Pressure versus volume relationship of the left ventricle. The series of lines with increasing slopes indicate the time-varying elastance of the aorta as described by Sagawa and colleagues [22]. Note that aortic valve opening occurs much before peak aortic elastance, peak left ventricular pressure, and accordingly peak aortic pressure The term dynamic elastance currently used by investi- gators [26–28, 31] is based on the ratio of respiratory variation in pulse pressure that occurs with each positive pressure breath as a percentage of the mean pressure di- vided by the corresponding change in stroke volume as a percentage of the mean change during the breath. This makes for a very complex measure. True elastance can only be assessed in a static state by increasing or de- creasing volume in an elastic structure by a known amount with no flow and then observing the change in pressure. Elastance is also different in the thoracic and abdominal aorta and in the different large vessels [32]. The total elastance is determined by the sum of the elas- tances of all the arterial vascular segments. When flow is present, especially pulsatile flow, there are also resistance and kinetic components to this dy- namic measure. A further problem is the curvilinear shape of the aortic volume–pressure relationship. Be- cause of this shape, the change in pressure with a change in volume is greater at higher initial volumes because Fig. 5 Pressure versus volume relationship of the left ventricle. Critical closing pressure Ultimately, the final value of arterial pressure is set by the strong regula- tory mechanisms that ensure that cardiac output and the return of blood to the heart match metabolic needs and as adjustments in vascular resistance and regional critical closing pressures to maintain a constant arterial pressure. This means that arterial pressure should not be consid- ered in isolation. Cardiac-aortic coupling The main determinant of the stroke volume by the eject- ing heart is the pressure at which the aortic valve opens, because this is the pressure at which heart muscle begins to shorten with a quasi-isotonic contraction (Fig. 5). When the aortic valve opens, the left ventricle is not yet at peak systolic elastance, and ejection continues until maximal left ventricular elastance is reached [21, 22]. Maximum ventricular elastance, i.e., the slope of the end-systolic pressure–volume line, is only a property of the heart and it is not a function of the load on the heart. The slope of this relationship is the same whether the heart contracts isometrically or isotonically [21]. The diastolic pressure at which the aortic valve opens is a function of the volume that is still in the aorta at the end of diastole. That volume is determined by a composite of factors: the amount of volume that was put into the aorta during the previous systole, the time allowed for the vol- ume to empty, which is dependent upon the length of dia- stole, the downstream arterial resistance, the critical closing pressures in small arteries or arterioles, and aortic elastance. The resistance and compliance (inverse of Dynamic elastance l The series of lines with increasing slopes indicate the time-varying elastance of the aorta as described by Sagawa and colleagues [22]. Note that aortic valve opening occurs much before peak aortic elastance, peak left ventricular pressure, and accordingly peak aortic pressure Page 7 of 10 Page 7 of 10 Magder Critical Care (2018) 22:257 a mean pressure of only 10 to 15 mmHg. Pulmonary ar- terial pressure remains low at peak exercise even with a fivefold increase in cardiac output. There are two main advantages for our high systemic arterial pressures. First, by keeping arterial pressure relatively constant, regional flows can change by altering regional arterial resistances according to regional needs for flow, without a change in aortic pressure. This works much like opening taps in your house, which allows a common pressure head to allow water to flow into the sink, bathtub, or toilet. The alternative way to increase flow to a region, such as the working muscle, would be to increase blood pressure by increasing the resistance in every vascular region except for the one that needs more flow. This obviously is a much more complex process than simply dilating one region and would have major consequences for regions that did not need more flow. It also would mean that the initial low arterial pressure would have to increase when there is a regional need for more flow, and this would increase the strain on the heart. This brings up a second advantage for having a high baseline arterial pressure. Because mean aortic pressure changes little with changes in regional flows or cardiac output, the load on the heart is relatively constant. This is important because the heart tolerates pressure loads much less well than volume loads (i.e., volume ejected) so that by hav- ing a relatively constant pressure, the load on the heart is relatively constant. the volume is moving up the steeper part of the relation- ship, but the actual shape of the relationship itself is constant over short periods of time. It can become stiffer over time with increases in age and hypertension (Fig. 1). The clinically obtained “dynamic” elastance is not a static measurement and is dominated by changes in ar- terial resistance [29], the critical closing pressure, and, to some extent, the position on the arterial volume–pres- sure curve. Dynamic elastance l Since the dynamic elastance term uses cyclic respiratory challenges to produce changes in pulse pres- sure and stroke volume, the changes likely are occurring mainly through the changes in return of blood to the right heart and to some extent by changes in loading of the right heart with lung inflation. This means that this measurement is affected by changes in blood volume, the size of the change in pleural pressure, and the change in transpulmonary pressure. Heart rate too is a factor because the length of diastole is a determinant of the volume that remains in the aorta at the end of dia- stole and thus a determinant of where the arterial vol- ume is on the elastance curve [33]. It also is affected by the emptying of the pulmonary venous reserves during the respiratory cycle [34]. The respiratory rate and length of inspiration and expiration add other factors. It is thus not surprising that dynamic elastance does not always act as expected [31] and at best may reflect gen- eral patterns. It is likely preferable to just examine the change in stroke volume, cardiac output, and blood pressure that were used to derive the measurement to interpret the response to a therapy. Autoregulation I f i It often is argued that ideal targets for perfusion pressure should be in the range of normal autoregulated blood flow [13]. This is defined as the pressure range in which flow does not change with increases or deceases in pressure [36]. Maintenance of flow with changes in arterial pressure can occur through four general mechanisms that regulate vascular resistance and regional critical closing pressures: neuro-humeral, myogenic, metabolic, and flow-mediated processes. However, the range of autoregulated flow, the mechanisms that regulate autoregulation in a tissue, the effects of drugs, the effects of disease, and the effects of central nervous activity differ among vascular beds and cannot be generalized. Regulation of arterial pressure Arterial pressures in all mammals from mice to humans is in the same range and blood pressure is one of the body’s most tightly regulated variables. The feedback control is remarkable. A young male exercising to near maximum aerobic capacity can increase cardiac output fivefold or more but mean arterial pressure does not change or even slightly decreases. For this to occur, ar- terial resistance must decrease by more than 80%. This tight regulation of arterial pressure occurs primarily through baroreceptor regulation, but regional myogenic mechanisms and metabolic activity also contribute to local autoregulation. Flow-mediated adjustments also occur, by which increased flow produces a decrease in downstream vascular resistance by the release of nitric oxide from the endothelium [35]. Let us start with what are the dominant controlled variables in the whole circulation. As already stated, we are pressure-regulated beings, meaning that mainten- ance of a constant blood pressure is a priority for the body. Blood pressure is approximated by flow (cardiac output) and the systemic vascular resistance. Cardiac out- put is determined by metabolic needs of tissues, which can be quantified by their consumption of oxygen. Con- sidering that control of arterial pressure is a priority for the body, and that cardiac output is strongly related to metabolic needs, it can be appreciated that changes in The tight control of blood pressure raises the physio- logical question as to why this evolved and why is arter- ial pressure much higher in mammals and birds than in all other species? The high arterial pressure is not neces- sary for baseline flow because the full cardiac output goes from the right to left heart through the lungs with Page 8 of 10 Page 8 of 10 Magder Critical Care (2018) 22:257 Baroreceptor-induced vasoconstriction is greater in peripheral vascular beds, which are primarily muscle tis- sue, than in the splanchnic bed [38, 39]. This shifts the distribution of blood flow to the splanchnic bed. By itself this would result in a decrease in cardiac output and a further decrease in blood pressure because volume accu- mulates in the very compliant splanchnic vasculature [40, 41]. However, the effect of this redistribution is compensated for by a decrease in the capacitance in the splanchnic bed (recruitment of unstressed into stressed volume) and a decrease in splanchnic venous resistance at the same time as the arterial resistance to the splanch- nic bed increases [39]. Regulation of arterial pressure A fall in arterial pressure with a normal cardiac output requires an in- crease in systemic vascular resistance to restore arterial pressure, but the rise in arterial resistance increases the load on the left ventricle, which could lead to a decrease in cardiac output. The hypotension would be fixed, but tissue perfusion would not. If the increase in vasoconstric- tion also increases venous resistance, cardiac output would fall even more [37]. If the fall in arterial pressure occurs because of a decrease in cardiac output, an increase in arterial resistance in all vascular beds will restore blood pressure, but not regional organ blood flows. The hope when a pure vasoconstrictor drug is used is that local metabolic activity will override the constricting effect of the drug in critical vascular beds such as the brain and heart so that these regions will receive a greater propor- tion of the available flow. How much this occurs likely de- pends upon the ability of these regions to modify the generalized vasoconstriction through their local signals, and likely also is affected by the receptor density for the vasoconstricting drug. Very high doses may just constrict all regions non-discriminately. The clinically important point is that if tissue perfusion is low, a treatment must increase cardiac output without a change in arterial pressure and not overwhelm regional mechanisms that match flow to tissue needs. This type of strategy re- quires some measure of blood flow or indirect mea- sures of tissue perfusion such as lactate and central venous saturation as well as clinical indicators such as wakefulness, skin temperature and color, and urine out- put when the kidneys are working. The fourth factor regulating local blood flow is flow-mediated dilatation. This provides a feed-forward process and decreases the downstream resistance when flow increases [44]. It primarily is mediated by release of nitric oxide (NO) through the effect of shear stress on vascular endothelial cells [45]. This mechanism would spiral out of control if something else does not happen because the decrease in downstream resistance would result in more flow, more release of NO, greater flow, and so on. It is typical of nature to drive with her foot on the gas and brake at the same time; consider simul- taneous parasympathetic and sympathetic nerve activ- ities and the cardio-inhibitory and cardio-stimulatory centers in the brain. Regulation of arterial pressure It is likely that infused vasocon- strictors also affect the peripheral vasculature more than the splanchnic bed, but at higher doses the difference may no longer be active and these drugs may then alter the normal distribution of resistances. If the vasocon- strictor cannot recruit more unstressed volume because there are insufficient reserves, and constricts venous re- sistance, cardiac output and tissue perfusion will fall. Tissues need flow and not pressure unless they can se- lectively dilate. This is what happens in most cases when phenylephrine is given; arterial pressure rises but cardiac output falls [37, 42]. In contrast, norepinephrine in mod- erate doses does not increase venous resistance and also produces a moderate increase in cardiac function [43]. Again, monitoring perfusion or cardiac output can be helpful to know what is happening. systemic vascular resistance dominate the normal regula- tion of arterial pressure. Regulation of systemic vascular resistance first occurs through neural mechanisms that provide afferent feedback to the medullary cardio-in- hibitory and cardio-stimulatory regions that regulate vascular tone by sending efferent signals through para- sympathetic and sympathetic pathways, and to some extent through humoral signals, to maintain the cen- trally set pressure. systemic vascular resistance dominate the normal regula- tion of arterial pressure. Regulation of systemic vascular resistance first occurs through neural mechanisms that provide afferent feedback to the medullary cardio-in- hibitory and cardio-stimulatory regions that regulate vascular tone by sending efferent signals through para- sympathetic and sympathetic pathways, and to some extent through humoral signals, to maintain the cen- trally set pressure. The relationship of blood flow to metabolic need for the whole body is dominated by tissues that can greatly increase their oxygen needs, which are skeletal and car- diac muscle. In these tissues, as is the case for the whole body, there is a linear relationship between blood flow and oxygen consumption, indicating that the primary regulator of blood flow is metabolic activity. This strong metabolic coupling can override neural-mediated vaso- constriction. Metabolic activity likely plays an important role in the brain, too; however, the limited space in the skull means that increased volume and pressure must be controlled, likely by the myogenic process. y y y g p The two dominantly controlled cardiovascular variables, systemic arterial blood pressure and cardiac output rela- tive to metabolic need, can be in conflict. Regulation of arterial pressure Dynamic arterial elastance as a predictor of arterial pressure response to fluid administration: a validation study. Crit Care. 2014;18(6):626. 28. Pinsky MR. Defining the boundaries of bedside pulse contour analysis: dynamic arterial elastance. Crit Care. 2011;15(1):120. 29. Sunagawa K, Maughan WL, Burkhoff D, Sagawa K. Left ventricular interaction with arterial load studied in isolated canine ventricle. Am J Phys. 1983;245(5 Pt 1):H773–80. 30. Sunagawa K, Maughan WL, Sagawa K. Optimal arterial resistance for the maximal stroke work studied in isolated canine left ventricle. Circ Res. 1985; 56(4):586–95. 31. Monge Garcia MI, Guijo Gonzalez P, Gracia Romero M, Gil Cano A, Rhodes A, Grounds RM, et al. Effects of arterial load variations on dynamic arterial elastance: an experimental study. Br J Anaesth. 2017;118(6):938–46. 4. Burton AC. The vascular bed. Physiology and biophysics of the circulation. Chicago: Year Book Medical Publishers Inc.; 1965. p. 61–92. Chicago: Year Book Medical Publishers Inc.; 1965. p. 61–92. 5. Burton AC. On the physical equilibrium of small blood vessels. Am J Physiol. 1951;164:319–29. 6. Azuma T, Oka S. Mechanical equilibrium of blood vessel walls. Am J Physiol. 1971;221:1210–318. 7. Oka S, Azuma T. Physical theory of tension in thick-walled blood vessels in equilibrium. Biorheology. 1970;7:109–17. 8. Burton AC. Total fluid energy, gravitational potential energy, effects of posture. physiology and biophysics of the circulation: an introductory text. Chicago: Year Book Medical Publishers Inc.; 1965. p. 95–111. Publisher’s Note Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. 28. Pinsky MR. Defining the boundaries of bedside pulse contour analysis: dynamic arterial elastance. Crit Care. 2011;15(1):120. 29. Sunagawa K, Maughan WL, Burkhoff D, Sagawa K. Left ventricular interaction with arterial load studied in isolated canine ventricle. Am J Phys. 1983;245(5 Pt 1):H773–80. Received: 29 January 2018 Accepted: 27 August 2018 30. Sunagawa K, Maughan WL, Sagawa K. Optimal arterial resistance for the maximal stroke work studied in isolated canine left ventricle. Circ Res. 1985; 56(4):586–95. Competing interests 27. Garcia MI, Romero MG, Cano AG, Aya HD, Rhodes A, Grounds RM, et al. Dynamic arterial elastance as a predictor of arterial pressure response to fluid administration: a validation study. Crit Care. 2014;18(6):626. Conclusions l 9. Magder S. Is all on the level? Hemodynamics during supine versus prone ventilation. Am J Respir Crit Care Med. 2013;188(12):1390–1. 9. Magder S. Is all on the level? Hemodynamics during supine versus prone ventilation. Am J Respir Crit Care Med. 2013;188(12):1390–1. Mean arterial pressure is determined by cardiac output, systemic vascular resistance, and a critical closing pres- sure at the level of the arterioles. Each of these factors is controlled by mechanisms that work at the level of the whole organism, but also interact with important local regulatory mechanisms. Arterial pulse pressure brings in another set of variables which are related to the elas- tance of the aortic wall, the volume of blood in the aorta, the cardiac frequency, and the proportion of time in systole and diastole. Because of the complexities of these interactions, it is not possible to make simple pre- dictions of the response to vasopressor therapies. This becomes even more complicated when pathologies are added that alter the potential of vessels to respond, or because there are fixed obstructions to flow. Only em- piric studies can determine the best approach for the management of hypotension and hypoperfusion. Finally, it must be remembered that what counts for tissues is blood flow and not the arterial pressure and, even more importantly, the matching of flow to metabolic needs. The body does this masterfully through multiple coun- teracting control mechanisms. It is very unlikely that a single therapeutic agent can match the naturally occur- ring well-orchestrated control mechanisms. 10. Magder SA. Pressure-flow relations of diaphragm and vital organs with nitroprusside-induced vasodilation. J Appl Physiol. 1986;61:409–16. 11. Permutt S, Riley S. Hemodynamics of collapsible vessels with tone: the vascular waterfall. J Appl Physiol. 1963;18(5):924–32. 11. Permutt S, Riley S. Hemodynamics of collapsible vessels with tone: the vascular waterfall. J Appl Physiol. 1963;18(5):924–32. 12. Sylvester JL, Traystman RJ, Permutt S. Effects of hypoxia on the closing pressure of the canine systemic arterial circulation. Circ Res. 1981;49:980–7. 12. Sylvester JL, Traystman RJ, Permutt S. Effects of hypoxia on the closing pressure of the canine systemic arterial circulation. Circ Res. 1981;49:980–7. 13. Kato R, Pinsky MR. Personalizing blood pressure management in septic shock. Ann Intensive Care. 2015;5(1):41. 13. Kato R, Pinsky MR. Personalizing blood pressure management in septic shock. Ann Intensive Care. 2015;5(1):41. 14. Magder S. Starling resistor versus compliance. Which explains the zero-flow pressure of a dynamic arterial pressure-flow relation? Conclusions l Circ Res. 1990;67:209–20. 14. Magder S. Starling resistor versus compliance. Which explains the zero-flow pressure of a dynamic arterial pressure-flow relation? Circ Res. 1990;67:209–20. y 15. Bellamy RF. Diastolic coronary artery pressure-flow relations in the dog. Circ Res. 1978;43(1):92–101. 15. Bellamy RF. Diastolic coronary artery pressure-flow relations in the dog. Circ Res. 1978;43(1):92–101. 16. Shrier I, Hussain SNA, Magder S. Effect of carotid sinus stimulation on resistance and critical closing pressure of the canine hindlimb. Am J Physiol. 1993;264:H1560–H6. 17. Shrier I, Magder S. NG-nitro-L-arginine and phenylephrine have similar effects on the vascular waterfall in the canine hindlimb. Am J Physiol. 1995; 78(2):478–82. 18. Shrier I, Magder S. Response of arterial resistance and critical closing pressure to change in perfusion pressure in canine hindlimb. Am J Physiol. 1993;265:H1939–H45. 19. Shrier I, Magder S. The effects of nifedipine on the vascular waterfall and arterial resistance in the canine hindlimb. Am J Physiol. 1995;268:H372–H6. y 20. Shrier I, Magder S. Effects of adenosine on the pressure-flow relationships in an in vitro model of compartment syndrome. J Appl Physiol. 1997;82(3):755–9. 21. Sagawa K. The ventricular pressure-volume diagram revisited. Circ Res. 1978; 43:677–87. 22. Suga H, Saeki Y, Sagawa K. End-systolic force-length relationship of nonexcised canine papillary muscle. Am J Phys. 1977;233(6):H711–H7. 23. Guarracino F, Baldassarri R, Pinsky MR. Ventriculo-arterial decoupling in acutely altered hemodynamic states. Crit Care. 2013;17(2):213. 22. Suga H, Saeki Y, Sagawa K. End-systolic force-length relationship of nonexcised canine papillary muscle. Am J Phys. 1977;233(6):H711–H7. 23. Guarracino F, Baldassarri R, Pinsky MR. Ventriculo-arterial decoupling in acutely altered hemodynamic states. Crit Care. 2013;17(2):213. Abbreviations Abbreviations cm: Centimeter; L: Liter; m: Mass; mmHg: Millimeter of mercury; Rn: Segment resistance; RTotal: Total resistance; v: Velocity (L/sec) 24. Wang JJ, O'Brien AB, Shrive NG, Parker KH, Tyberg JV. Time-domain representation of ventricular-arterial coupling as a windkessel and wave system. Am J Physiol Heart Circ Physiol. 2003;284(4):H1358–68. 25. Monge Garcia MI, Gil Cano A, Gracia Romero M. Dynamic arterial elastance to predict arterial pressure response to volume loading in preload- dependent patients. Crit Care. 2011;15(1):R15. Author’s contributions The author read and approved the final manuscript. Author’s contributions The author read and approved the final manuscript. 26. Cecconi M, Monge Garcia MI, Gracia Romero M, Mellinghoff J, Caliandro F, Grounds RM, et al. The use of pulse pressure variation and stroke volume variation in spontaneously breathing patients to assess dynamic arterial elastance and to predict arterial pressure response to fluid administration. Anesth Analg. 2015;120(1):76–84. Consent for publication Not applicable. Competing interests The author declares that he has no competing interests. Competing interests The author declares that he has no competing interests. Regulation of arterial pressure In this case the brakes are the local myogenic and central neuro-humeral mechanisms, as well as local metabolic needs. The advantage to such a process is that it allows rapid adaptation to increased needs for flow and fine tuning of the matching of flow to local metabolic activity. Flow-mediated dilatation is lost when the endothelium is damaged in vascular dis- ease and contributes to further vascular damage. Returning to the question of the usefulness of target- ing the autoregulatory range, I would argue that what really counts is avoiding the lower autoregulatory range in which flow falls when arterial pressure falls and, even more so, when this is combined with a decrease in Magder Critical Care (2018) 22:257 Page 9 of 10 Page 9 of 10 oxygen consumption, because dilatation and oxygen ex- traction are maximal. When this limit is reached, the only treatments that can help tissue perfusion are an in- crease in cardiac output or constriction of some other region, but constriction of these other regions would compromise their function. This means that organs can- not be considered in isolation and the reserves of the whole system need to be considered. 3. Roach MR, Burton AC. The reason for the shape of the distensibility curves of arteries. Can J Biochem Physiol. 1957;35(8):681–90. 4. Burton AC. The vascular bed. Physiology and biophysics of the circulation. Chicago: Year Book Medical Publishers Inc.; 1965. p. 61–92. 3. Roach MR, Burton AC. The reason for the shape of the distensibility curves of arteries. Can J Biochem Physiol. 1957;35(8):681–90. 3. Roach MR, Burton AC. The reason for the shape of the distensibility curves of arteries. Can J Biochem Physiol. 1957;35(8):681–90. 4. Burton AC. The vascular bed. Physiology and biophysics of the circulation. Chicago: Year Book Medical Publishers Inc.; 1965. p. 61–92. 5. Burton AC. On the physical equilibrium of small blood vessels. Am J Physiol. 1951;164:319–29. 6. Azuma T, Oka S. Mechanical equilibrium of blood vessel walls. Am J Physiol. 1971;221:1210–318. 7. Oka S, Azuma T. Physical theory of tension in thick-walled blood vessels in equilibrium. Biorheology. 1970;7:109–17. 8. Burton AC. Total fluid energy, gravitational potential energy, effects of posture. physiology and biophysics of the circulation: an introductory text. Chicago: Year Book Medical Publishers Inc.; 1965. p. 95–111. 9. Magder S. Is all on the level? Hemodynamics during supine versus prone ventilation. Am J Respir Crit Care Med. 2013;188(12):1390–1. 10. Magder SA. Regulation of arterial pressure Pressure-flow relations of diaphragm and vital organs with nitroprusside-induced vasodilation. J Appl Physiol. 1986;61:409–16. 11. Permutt S, Riley S. Hemodynamics of collapsible vessels with tone: the vascular waterfall. J Appl Physiol. 1963;18(5):924–32. 12. Sylvester JL, Traystman RJ, Permutt S. Effects of hypoxia on the closing pressure of the canine systemic arterial circulation. Circ Res. 1981;49:980–7. 13. Kato R, Pinsky MR. Personalizing blood pressure management in septic shock. Ann Intensive Care. 2015;5(1):41. 14. Magder S. Starling resistor versus compliance. Which explains the zero-flow pressure of a dynamic arterial pressure-flow relation? Circ Res. 1990;67:209–20. 15. Bellamy RF. Diastolic coronary artery pressure-flow relations in the dog. Circ Res. 1978;43(1):92–101. 16. Shrier I, Hussain SNA, Magder S. Effect of carotid sinus stimulation on resistance and critical closing pressure of the canine hindlimb. Am J Physiol. 1993;264:H1560–H6. 17. Shrier I, Magder S. NG-nitro-L-arginine and phenylephrine have similar effects on the vascular waterfall in the canine hindlimb. Am J Physiol. 1995; 78(2):478–82. 18. Shrier I, Magder S. Response of arterial resistance and critical closing pressure to change in perfusion pressure in canine hindlimb. Am J Physiol. 1993;265:H1939–H45. 19. Shrier I, Magder S. The effects of nifedipine on the vascular waterfall and arterial resistance in the canine hindlimb. Am J Physiol. 1995;268:H372–H6. 20. Shrier I, Magder S. Effects of adenosine on the pressure-flow relationships in an in vitro model of compartment syndrome. J Appl Physiol. 1997;82(3):755–9. 21. Sagawa K. The ventricular pressure-volume diagram revisited. Circ Res. 1978; 43:677–87. 22. Suga H, Saeki Y, Sagawa K. End-systolic force-length relationship of nonexcised canine papillary muscle. Am J Phys. 1977;233(6):H711–H7. 23. Guarracino F, Baldassarri R, Pinsky MR. Ventriculo-arterial decoupling in acutely altered hemodynamic states. Crit Care. 2013;17(2):213. 24. Wang JJ, O'Brien AB, Shrive NG, Parker KH, Tyberg JV. Time-domain representation of ventricular-arterial coupling as a windkessel and wave system. Am J Physiol Heart Circ Physiol. 2003;284(4):H1358–68. 25. Monge Garcia MI, Gil Cano A, Gracia Romero M. Dynamic arterial elastance to predict arterial pressure response to volume loading in preload- dependent patients. Crit Care. 2011;15(1):R15. 26. Cecconi M, Monge Garcia MI, Gracia Romero M, Mellinghoff J, Caliandro F, Grounds RM, et al. The use of pulse pressure variation and stroke volume variation in spontaneously breathing patients to assess dynamic arterial elastance and to predict arterial pressure response to fluid administration. Anesth Analg. 2015;120(1):76–84. 27. Garcia MI, Romero MG, Cano AG, Aya HD, Rhodes A, Grounds RM, et al. References 31. Monge Garcia MI, Guijo Gonzalez P, Gracia Romero M, Gil Cano A, Rhodes A, Grounds RM, et al. Effects of arterial load variations on dynamic arterial elastance: an experimental study. Br J Anaesth. 2017;118(6):938–46. 1. Magder SA. The highs and lows of blood pressure: toward meaningful clinical targets in patients with shock. Crit Care Med. 2014;42(5):1241–51. 2. Dobrin PB. Mechanical properties of arteries. Physiol Rev. 1978;58(2):397–460. 2. Dobrin PB. Mechanical properties of arteries. Physiol Rev. 1978;58(2):397–460. Page 10 of 10 Magder Critical Care (2018) 22:257 32. Nakashima T, Tanikawa J. A study of human aortic distensibility with relation to atherosclerosis and aging. Angiology. 1971;22(8):477–90. 33. Maughan WL, Sunagawa K, Burkhoff D, Graves WL Jr, Hunter WC, Sagawa K. Effect of heart rate on the canine end-systolic pressure-volume relationship. Circulation. 1985;72(3):654–9. 34. Magder S, Guerard B. Heart-lung interactions and pulmonary buffering: lessons from a computational modeling study. RespirPhysiol Neurobiol. 2012;182(2–3):60–70. 35. Moncada S, Higgs A. The L-arginine-nitric oxide pathway. N Engl J Med. 1993;329(27):2002–12. 35. Moncada S, Higgs A. The L-arginine-nitric oxide pathway. N Engl J Med. 1993;329(27):2002–12. . Johnson PC. Autoregulation of blood flow. Circ Res. 1986;59:483–9 37. Magder S. Phenylephrine and tangible bias. Anesthesia Analgesia. 2011; 113(2):211–3. 38. Hainsworth R, Karim F, McGregor KH, Rankin AJ. Effects of stimulation of aortic chemoreceptors on abdominal vascular resistance and capacitance in anaesthetized dogs. J Physiol. 1983;334:421–31. 39. Deschamps A, Magder S. Baroreflex control of regional capacitance and blood flow distribution with or without alpha adrenergic blockade. J Appl Physiol. 1992;263:H1755–H63. 40. Magder S. Volume and its relationship to cardiac output and venous return. Crit Care. 2016;20:271. 41. Krogh A. The regulation of the supply of blood to the right heart. Skand Arch Physiol. 1912;27:227–48. 42. Thiele RH, Nemergut EC, Lynch C III. The physiologic implications of isolated alpha 1 adrenergic stimulation. Anesth Analg. 2011;113(2):284–96. 43. Datta P, Magder S. Hemodynamic response to norepinephrine with and without inhibition of nitric oxide synthase in porcine endotoxemia. Am J Resp Crit Care Med. 1999;160(6):1987–93. 44. Furchgott RF, Zawadzki JV. The obligatory role of endothelial cells in the relaxation of arterial smooth muscle by acetylcholine. Nature. 1980;288:373–6. 45. Moncada S. The L-arginine:nitric oxide pathway. Acta Physiol Scand. 1992; 145:201–27.
https://openalex.org/W1995952892	https://europepmc.org/articles/pmc3590211?pdf=render	English	null	In Vitro and In Vivo Evaluation of Zinc-Modified Ca–Si-Based Ceramic Coating for Bone Implants	PloS one	2,013	cc-by	8,257	Abstract Competing Interests: The authors have declared that no competing interests exist. * E-mail: yespine@163.com (XJY); 061022049@fudan.edu.cn (MYW) * E-mail: yespine@163.com (XJY); 061022049@fudan.edu.cn (MYW) . These authors contributed equally to this work. ming Yu1., Kai Li2., Xuebin Zheng2, Dannong He3, Xiaojian Ye1, Meiyan Wang3 1 Department of Orthopaedics, Changzheng Hospital, Second Military Medical University, Shanghai, People’s Republic of China, 2 Key Laboratory of Inorganic Coating Materials, Shanghai Institute of Ceramics, Chinese Academy of Sciences, Shanghai, People’s Republic of China, 3 National Engineering Research Center for Nanotechnology, Shanghai, People’s Republic of China Abstract The host response to calcium silicate ceramic coatings is not always favorable because of their high dissolution rates, leading to high pH within the surrounding physiological environment. Recently, a zinc-incorporated calcium silicate-based ceramic Ca2ZnSi2O7 coating, developed on a Ti-6Al-4V substrate using plasma-spray technology, was found to exhibit improved chemical stability and biocompatibility. This study aimed to investigate and compare the in vitro response of osteoblastic MC3T3-E1 cells cultured on Ca2ZnSi2O7 coating, CaSiO3 coating, and uncoated Ti-6Al-4V titanium control at cellular and molecular level. Our results showed Ca2ZnSi2O7 coating enhanced MC3T3-E1 cell attachment, proliferation, and differentiation compared to CaSiO3 coating and control. In addition, Ca2ZnSi2O7 coating increased mRNA levels of osteoblast-related genes (alkaline phosphatase, procollagen a1(I), osteocalcin), insulin-like growth factor-I (IGF-I), and transforming growth factor-b1 (TGF-b1). The in vivo osteoconductive properties of Ca2ZnSi2O7 coating, compared to CaSiO3 coating and control, was investigated using a rabbit femur defect model. Histological and histomorphometrical analysis demonstrated new bone formation in direct contact with the Ca2ZnSi2O7 coating surface in absence of fibrous tissue and higher bone-implant contact rate (BIC) in the Ca2ZnSi2O7 coating group, indicating better biocompatibility and faster osseointegration than CaSiO3 coated and control implants. These results indicate Ca2ZnSi2O7 coated implants have applications in bone tissue regeneration, since they are biocompatible and able to osseointegrate with host bone. Citation: Yu J, Li K, Zheng X, He D, Ye X, et al. (2013) In Vitro and In Vivo Evaluation of Zinc-Modified Ca–Si-Based Ceramic Coating for Bone Implants. PLoS ONE 8(3): e57564. doi:10.1371/journal.pone.0057564 Editor: Wei-Chun Chin, University of California, Merced, United States of America Yu et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestr d reproduction in any medium, provided the original author and source are credited. Copyright: 2013 Yu et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: This work was supported by the National Natural Science Foundation of China (Grant No. 81071455) and the Fund for Key Science and Technology Program of Shanghai Science and Technology Committee (Grant No. 09441900106). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. PLOS ONE \| www.plosone.org Introduction However, March 2013 \| Volume 8 \| Issue 3 \| e57564 1 PLOS ONE \| www.plosone.org Zinc-Modified Ceramic Coating for Bone Implants technology. Ca2ZnSi2O7 coating exhibited chemically stable (low dissolution) and good bioactivity in comparison with CaSiO3 coating [22]. In the present study, we investigate how Ca2ZnSi2O7 coating affects adhesion, morphology, orientation, proliferation and osteoblast differentiation of MC3T3-E1 osteoprogenitor cells (Fig. 1a). The in vivo osseointegration potential of Ca2ZnSi2O7 coating was also assessed using a rabbit femur defect implantation model (Fig. 1b). Materials and Methods Specimen Preparation and Characterization Introduction major limitations of calcium silicate coatings include deleterious biological effects due to their high dissolution rate and induction of high pH within the surrounding tissues, which limits further their biomedical application [5]. In recent years, calcium silicate-based ceramics have become more promising as potential implant biomaterials for bone tissue engineering due to their bioactive and biocompatible properties [1–3]. Despite the beneficial influence on bone response, well- known drawbacks associated with this material exist, such as poor mechanical properties, which limit the scope of its clinical application [4–6]. However, these limitations can be overcome by surface-modification techniques [7,8]. The bioactive ceramic coating would confer adequate bioactivity to the surface of the implant, preventing direct contact between the substrate and surrounding bone tissue, thus reducing release of problematic ions from the metallic substrate. Recently, ion-modification of Ca–Si-based ceramics has been developed to improve their chemical stability and biomedical properties and includes divalent (Mg [11], Zn [12,13] and Sr [14]) and tetravalent (Ti [5] and Zr [15]) modification. Moreover, Ti and Zr-incorporated Ca–Si-based ceramics have been used as stable coatings that improve biomedical properties compared with calcium silicate coatings [16,17]. However, Zn-modified calcium silicate (Ca2ZnSi2O7) ceramic coatings have not been fully investigated to date. It is well established that zinc, an essential trace element, plays an important role in various physiological processes [18]. Zn has been shown to have a stimulatory effect on bone formation and an inhibitory or biphasic effect on osteoclastic bone resorption [19]. Zinc deficiency results in the arrest of bone growth, bone development, and the overall maintenance of bone health [20,21]. This indicates Ca2ZnSi2O7 is an interesting biomaterial coating that warrants further investigation. In our previous work, Zn was selected for incorporation into a Ca–Si system to form hardystonite (Ca2ZnSi2O7), which was used as a feedstock coating on the Ti-6Al-4V substrate using plasma-spray Several coating techniques have been developed and include flame spraying, sputtering, electrophoretic coating, hot isostatic pressing and solution coating [9]. Each approach has its advantages and disadvantages; however, plasma-spraying has shown the most promise as a coating method. In fact, plasma- spraying is reported as the method of choice for coating CaSiO3 and Ca2SiO4 onto metal substrates, both of which have been used clinically to enhance bioactivity and bonding strength with titanium alloy in comparison to HA coating [10]. In vitro Testing Cell culture. The osteoblast-like cell line, MC3T3-E1, was obtained from the Chinese Academy of Sciences Cell Bank. Cells were cultured in culture plates containing alpha-minimal essential medium (a-MEM, Gibco BRL, Invitrogen, Life Technologies, USA) supplemented with 10% (v/v) fetal bovine serum (FBS) and 1% (v/v) of a 100 U/ml penicillin and 100 mg/ml streptomycin solution (Gibco BRL, respectively). Cells were incubated at 37uC in the presence of 5% CO2. On reaching 80% confluence, cells were detached every 2–3 days using 0.25% trypsin/EDTA solution (Gibco BRL). Quantitative measurement of OC secretion. MC3T3-E1 cells were cultured on control, CaSiO3 and Ca2ZnSi2O7 coated substrates in differentiation medium for 1, 7, 14 and 21 days. Quantitative levels of OC secreted into the culture medium were determined using an enzyme-linked immunoassay (ELISA) kit (R&D) following the manufacturer’s instructions. Cell attachment and morphology. MC3T3-E1 cells were seeded onto uncoated Ti-6Al-4V titanium (control), CaSiO3 and Ca2ZnSi2O7 coated substrates within individual wells of 24-well culture plates at a density of 16104 cells/cm2. Cells were incubated in a-MEM culture medium supplemented with 10% FBS at 37uC in 5% CO2 for 24 h. Samples were rinsed with PBS (pH 7.4) to remove non-adherent cells and fixed in 4% para- formaldehyde for 1 h. Cells were subsequently rinsed with PBS buffer containing 4% (w/v) sucrose and post-fixed in 1% osmium tetroxide in PBS, followed by sequential dehydration in an ethanol series of 30%, 50%, 70%, 90%, 95% and 100% for 5 minutes in each concentration. Specimens were sputter-coated with gold and the morphological characteristics of the attached cells were determined using scanning electron microscope (SEM, S-4800, Hitachi, Japan). ( ) g Quantitative reverse transcription-PCR (qRT- PCR). Total RNA was isolated from MC3T3-E1 cells on each coated substrate using Trizol reagent (Fermentas, Maryland, NY, USA). First-strand cDNA synthesis was carried out using Superscript III reverse transcriptase (Invitrogen, Carlsbad, CA, USA) for 60 min at 50uC. Quantitative RT-PCR for MC3T3-E1 cell osteoblast-related genes of alkaline phosphatase (ALP), procollagen a1(I) (Col a1(I)), osteocalcin (OC)and growth factors gene of insulin-like growth factor-I (IGF-I) and transforming growth factor-b1 (TGF-b1), was performed using the ABI PRISMs 7000 Sequence Detection System (Applied Biosystems, Foster City,CA, USA). Specimen Preparation and Characterization substrates with dimensions of 1061062 mm and ø1061 mm, for in vitro cell culture and in vivo study respectively. The atmosphere plasma spray (APS) system (Sulzer Metco, Switzerland) was used to fabricate all coatings. The thickness of the coating was approximately 170 mm. Before plasma-spraying, substrates were ultrasonically cleaned in acetone and grit-blasted using corundum sand of F60 grade. CaSiO3 coatings were prepared using the same conditions as the control. Samples were ultrasonically cleaned and sterilized in acetone, ethanol and distilled water for 10 min. Quantitative measurement of ALP activity. MC3T3-E1 cells were cultured on control, CaSiO3 and Ca2ZnSi2O7 coated substrates in differentiation medium for 1, 4, 7 and 14 days. ALP activity was measured using stable p-nitrophenol phosphate substrate. At each time point, culture medium was removed by decantation and cells were washed with PBS and harvested in 1 ml universal ALP buffer (100 mM citric acid, 100 mM KH2PO4, 100 mM sodium tetraborate.10 H2O, 100 mM Tris, 100 mM KCl; pH 11). Cells were sonicated twice for 20 sec and centrifuged at 3000 rpm for 5 min at 4uC. ALP activity in the supernatants was determined following addition of p-nitrophenyl phosphate substrate and the reaction was stopped using 100 ml of 0.1 N NaOH. The optical density was measured at 405 nm using a microplate reader (SPECTRA MAX PLUS 384 MK3, Thermo). The ALP activity was calculated from a standard curve after normalization to total protein content, which was measured using the Bradford protein assay kit (Pierce, Rockford, IL, USA). ALP experiments were repeated twice with n = 3 for each substrate. The phase composition, surface morphology, surface roughness (Ra) and bonding strength of the Ca2ZnSi2O7 coating was characterized in our previous work [22]. The phase composition, surface morphology, surface roughness (Ra) and bonding strength of the Ca2ZnSi2O7 coating was characterized in our previous work [22]. Specimen Preparation and Characterization Ca2ZnSi2O7 powders and coatings were prepared according to methods described previously [22]. Briefly, Ca2ZnSi2O7 powders were synthesized by the sol-gel method using zinc nitrate hexahydrate (Zn(NO3)2.6H2O; Sinopharm Chemical Reagent Co., Ltd (SCRC), China), calcium nitrate tetrahydrate (Ca(- NO3)2.4H2O; SCRC, China) and tetraethyl orthosilicate (TEOS, (C2H5O)4Si); SCRC, China). The obtained Ca2ZnSi2O7 powders were sieved through 80 meshes and sprayed onto Ti-6Al-4V Figure 1. (a) Schematic representation of MC3T3-E1 cell proliferation on uncoated Ti-6Al-4V titanium (control), CaSiO3 and Ca2ZnSi2O7 coated substrates. (b) Schematic illustration of in vivo analysis. White arrows indicated new bone was formed at the Ca2ZnSi2O7 coating periphery with direct contact to the implant surface in absence of a connective tissue layer in group 3. In contrast, a wide band of fibrous tissue was clearly seen around the surface of the CaSiO3 coating and control in group 2 and 1. Group 1, control; Group 2, CaSiO3 coating; Group 3, Ca2ZnSi2O7 coating. doi:10.1371/journal.pone.0057564.g001 PLOS ONE \| www.plosone.org 2 March 2013 \| Volume 8 \| Issue 3 \| e57564 Figure 1. (a) Schematic representation of MC3T3-E1 cell proliferation on uncoated Ti-6Al-4V titanium (control), CaSiO3 and Ca2ZnSi2O7 coated substrates. (b) Schematic illustration of in vivo analysis. White arrows indicated new bone was formed at the Ca2ZnSi2O7 coating periphery with direct contact to the implant surface in absence of a connective tissue layer in group 3. In contrast, a wide band of fibrous tissue was clearly seen around the surface of the CaSiO3 coating and control in group 2 and 1. Group 1, control; Group 2, CaSiO3 coating; Group 3, Ca2ZnSi2O7 coating. doi:10.1371/journal.pone.0057564.g001 Figure 1. (a) Schematic representation of MC3T3-E1 cell proliferation on uncoated Ti-6Al-4V titanium (control), CaSiO3 and Ca2ZnSi2O7 coated substrates. (b) Schematic illustration of in vivo analysis. White arrows indicated new bone was formed at the Ca2ZnSi2O7 coating periphery with direct contact to the implant surface in absence of a connective tissue layer in group 3. In contrast, a wide band of fibrous tissue was clearly seen around the surface of the CaSiO3 coating and control in group 2 and 1. Group 1, control; Group 2, CaSiO3 coating; Group 3, Ca2ZnSi2O7 coating. doi:10.1371/journal.pone.0057564.g001 2 2 7 g doi:10.1371/journal.pone.0057564.g001 March 2013 \| Volume 8 \| Issue 3 \| e57564 PLOS ONE \| www.plosone.org 2 Zinc-Modified Ceramic Coating for Bone Implants using ELISA kits (R&D Systems, Minneapolis, MN, USA) as described below. using ELISA kits (R&D Systems, Minneapolis, MN, USA) as described below. In vitro Testing The amount of target gene transcript in each sample was determined using SYBRH Premix Ex TaqTM II (Takara, Kyoto, Japan) in a final volume of 20 ml containing the same amount of RT product (50 ng cDNA), 10 ml of 26SYBRH Premix Ex TaqTM II and 0.8 ml of 10 mM forward and reverse primers. Primers for the selected genes are listed in Table 1. Amplification conditions were as follows: Stage 1:95uC for 30 s; Stage 2:30 repetitions of 95uC for 5 s, 60uC for 30 s. Each RT- PCR quantification experiment was performed in triplicate for individual samples. Final results were reported as the relative expression normalized with transcript level of the housekeeping gene, Glyceraldehyde 3-phosphate dehydrogenase (GAPDH). Cell proliferation. Cell proliferation was measured using a standard methyl thiazolyl tetrazolium (MTT) assay (Sigma- Aldrich, Saint Louis, MO, n = 6 for each sample). MC3T3-E1 cells were cultured on control, CaSiO3 and Ca2ZnSi2O7 coated substrates placed individually in a 24-well culture plate at a density of 16104 cells/cm2 and allowed to grow for 1, 4, 7 and 14 days. At the specified time-points, the substrates were washed in PBS and transferred to another new 24-well plate. MTT stock solution (10% of total volume) was added to 24-well plates containing the coatings and incubated for 4 h at 37uC and 5% CO2. Medium was withdrawn and dimethyl sulphoxide (Sigma–Aldrich) was added to each well to dissolve the formazan dye crystals. 100 ul of the reacted reagent from each well was transferred to 96-well plates and the absorbance measured using a microplate reader (SPECTRA MAX PLUS 384 MK3, Thermo Fisher Scientific, Waltham, MA, USA) at a wavelength of 490 nm. In vivo Analysis Surgical procedure. Total nine White New Zealand rabbits (obtained from Shanghai Jiao Tong University, Laboratory Animal Center, male, 2–2.5 kg body weight) were randomly divided into 3 groups. The use of animals and the experimental protocols was approved by the Institutional Animal Welfare Committee of Shanghai Jiao Tong University. A total of six cylindrical implants of approximately 1 mm in diameter and 1 cm in length were implanted into the femur of each rabbit. Three implants were inserted into the femur of the left hind leg and another three into the femur of the right hind leg. A schedule was prepared to ensure equal placement of the three implant types within each of the six femoral sites and one of each implant types placed into each femur. Rabbits were anesthetized by injecting 3% Surgical procedure. Total nine White New Zealand rabbits (obtained from Shanghai Jiao Tong University, Laboratory Animal Center, male, 2–2.5 kg body weight) were randomly divided into 3 groups. The use of animals and the experimental protocols was approved by the Institutional Animal Welfare Committee of Shanghai Jiao Tong University. A total of six cylindrical implants of approximately 1 mm in diameter and 1 cm in length were implanted into the femur of each rabbit. Three implants were inserted into the femur of the left hind leg and another three into the femur of the right hind leg. A schedule was prepared to ensure equal placement of the three implant types within each of the six femoral sites and one of each implant types placed into each femur. Rabbits were anesthetized by injecting 3% Cell differentiation and mineralization. MC3T3-E1 cell osteoblast differentiation and mineralization were determined using cells (16104 cells/cm2) cultured in triplicate (n = 3) on each coating group in differentiation medium comprising a-MEM medium supplemented with 50 mg/ml ascorbic acid (Sigma- Aldrich) and 10 mM b-glycerophosphate (Sigma-Aldrich). Cell differentiation and mineralization were characterized from alka- line phosphatase (ALP) activity and Osteocalcin (OC) secretion. Concentrations of ALP and OC were individually determined March 2013 \| Volume 8 \| Issue 3 \| e57564 PLOS ONE \| www.plosone.org 3 Zinc-Modified Ceramic Coating for Bone Implants Table 1. Primers used for qRT-PCR. In vivo Analysis Gene Sequence(59–39) Melting Temperature(uC) product length (bp) GAPDH F: TCCACTCACGGCAAATTCAACG 60 145 R: TAGACTCCACGACATACTCAGC ALP F: GCTGATCATTCCCAGGTTTT 60 204 R: CTGGGCCTGGTAGTTGTTGT Col a1(I) F: TTCTCCTGGTAAAGATGGTGC 60 255 R: GGACCAGCATCACCTTTAACA OC F: CCTCAGTCCCCAGCCCAGATC 60 220 R: CAGGGCAGAGAGAGAGGACAG IGF-I F: CACTCATCCACAATGCCTGTCT 60 118 R: CTGAGCTGGTGGATGCTCTTC TGF-b1 F: CCCTATATTTGGAGCCTGGA 60 141 R: CTTGCGACCCACGTAGTAGA Shown are the details of the primers used for qRT-PCR, including melting temperatures, forward (F) and reverse (R) sequences and product length. GAPDH, glyceraldehyde-3-phosphate dehydrogenase; ALP, alkaline phosphatase; Col a1(I), procollagen a1(I); OC, osteocalcin; IGF-I, Insulin-like growth factor-I; TGF-b1, transforming growth factor-b1. doi:10.1371/journal.pone.0057564.t001 Shown are the details of the primers used for qRT-PCR, including melting temperatures, forward (F) and reverse (R) sequences and product length. GAPDH, glyceraldehyde-3-phosphate dehydrogenase; ALP, alkaline phosphatase; Col a1(I), procollagen a1(I); OC, osteocalcin; IGF-I, Insulin-like growth factor-I; TGF-b1, transforming growth factor-b1. doi 10 1371/journal pone 0057564 t001 Shown are the details of the primers used for qRT-PCR, including melting temperatures, forward (F) and reverse (R) sequences and product length. GAPDH, glyceraldehyde-3-phosphate dehydrogenase; ALP, alkaline phosphatase; Col a1(I), procollagen a1(I); OC, osteocalcin; IGF-I, Insulin-like growth factor-I; TGF-b1, transforming growth factor-b1. doi:10.1371/journal.pone.0057564.t001 Nembutal (30 mg/kg) via the ear vein and a longitudinal incision was made by scalpel in the rabbit femur under rigorous aseptic conditions. Circular holes, 1 mm diameter by 1.0 cm deep, were drilled using a surgical electronic drill and thoroughly rinsed with physiological saline to remove shards of bone. Implants of Ti–6Al– 4V (control), CaSiO3-coated Ti–6Al–4V and Ca2ZnSi2O7-coated Ti–6Al–4V were used in this study. Before insertion of implants, blood was removed using sterile cotton balls and the holes were manually filled with implants as tightly as possible. The wound was sutured with nylon thread. Rabbits were sacrificed 1.5 month after implantation. Cell Attachment and Morphology MC3T3-E1 cell attachment and morphology on control, CaSiO3 and Ca2ZnSi2O7 coated substrates were examined using SEM (Fig. 2). After 24 h of culture, MC3T3-E1 cells were found to attach to all tested samples surface. The cells displayed a triangle- like morphology on the surface of the titanium control (Fig. 2a, b) but appeared spindle-like on the surface of both the Ca2ZnSi2O7 (Fig. 2e, f) and CaSiO3 (Fig. 2c, d) coated substrates. A greater number of cells had cytoplasmic extensions, which were longer in protrusion, on the Ca2ZnSi2O7 coated surface in comparison to CaSiO3 and control. Specimen preparation. After sacrifice, Excised specimens were fixed in 4% paraformaldehyde for 3 days, dehydrated in a series of ethanol solutions (70, 80, 90, 95 and 100%) and embedded in polymethylmethacrylate resin. Undecalcified sec- tions with a thickness of 100 mm were cut using a saw microtome (Leica ST1600) and stained with trinitrophenol- fuchsin prior to histological analysis. A minimum of 6 sections were made from each implant and used for descriptive and morphometric analyses. Specimen preparation. After sacrifice, Excised specimens were fixed in 4% paraformaldehyde for 3 days, dehydrated in a series of ethanol solutions (70, 80, 90, 95 and 100%) and embedded in polymethylmethacrylate resin. Undecalcified sec- tions with a thickness of 100 mm were cut using a saw microtome (Leica ST1600) and stained with trinitrophenol- fuchsin prior to histological analysis. A minimum of 6 sections were made from each implant and used for descriptive and morphometric analyses. Cell Proliferation MC3T3-E1 cell proliferation on control, CaSiO3 and Ca2Zn- Si2O7 coated substrates was determined using an MTT assay. A statistically significant increase in cell proliferation was measured on all samples with increasing culture time (Fig. 3). Cell proliferation was increased on days 1, 4, 7 and 14 on Ca2ZnSi2O7 coated substrate compared to either CaSiO3 coated substrate or control (p,0.05). Histology and histomorphometry. Histological observa- tion and histomorphometrical analysis of the sections were performed under a biologic fluorescence microscope (Olympus BX-60). The microscopic images of the sections were digitized and analyzed using an image-analyzing software (KS400). The bone-implant contact rate (BIC) was used as the index of osseointegration, equal to the surface length contact of the implant with the bone/the overall surface length of implant inserted into the bone. Histology and histomorphometry. Histological observa- tion and histomorphometrical analysis of the sections were performed under a biologic fluorescence microscope (Olympus BX-60). The microscopic images of the sections were digitized and analyzed using an image-analyzing software (KS400). The bone-implant contact rate (BIC) was used as the index of osseointegration, equal to the surface length contact of the implant with the bone/the overall surface length of implant inserted into the bone. Cell Differentiation and Mineralization Cell differentiation from pre-osteoblasts to osteoblasts revealed osteogenic functioning of MC3T3-E1 cells on all substrates. ALP assays were performed on MC3T3-E1 cells cultured on control, CaSiO3 and Ca2ZnSi2O7 coated substrates at various time points. ALP activity presented a similar profile for all samples, with increased activity throughout the culture duration (Fig. 4). For all conditions tested, there was no significant difference in ALP activity between days 1 and 4, after which levels rapidly increased up to day 14. ALP activity within cells incubated on Ca2ZnSi2O7 coating increased above those on the CaSiO3 coating and control (p,0.05 for CaSiO3 coating and control at day 7, 14). Results therefore indicate that Ca2ZnSi2O7 coating can promote cell maturation as reflected by increased ALP activity. Statistical Analysis For statistical analysis, first Levene’s test was performed to determine the homogeneity of variance for all the data, and then one way analysis of variance followed by Tukey or Tamhane’s T2 post-hoc test for multiple comparison was performed for the comparisons between different groups. SPSS 17.0 program was employed for all statistical analysis and differences were considered significant if p,0.05. All data were expressed as a mean 6 standard deviation (SD). Levels of the differentiation marker, OC, were investigated at days 1, 7, 14 and 21 of culture. As shown in Fig. 5, no significant difference in OC levels were found in MC3T3-E1 cells cultured on March 2013 \| Volume 8 \| Issue 3 \| e57564 PLOS ONE \| www.plosone.org March 2013 \| Volume 8 \| Issue 3 \| e57564 4 Zinc-Modified Ceramic Coating for Bone Implants Figure 2. SEM images of MC3T3-E1 cell morphology on the different substrates 24 h after seeding:(a, b) control, (c, d) CaSiO coating and (e, f) Ca2ZnSi2O7 coating. The arrows indicated cells cultured on the different substrates. doi:10.1371/journal.pone.0057564.g002 Figure 2. SEM images of MC3T3-E1 cell morphology on the different substrates 24 h after seeding:(a, b) control, (c, d) CaSiO3 coating and (e, f) Ca2ZnSi2O7 coating. The arrows indicated cells cultured on the different substrates. doi:10.1371/journal.pone.0057564.g002 Ca2ZnSi2O7, CaSiO3 and control substrates at day 1. However, increases in OC were observed at days 7, 14 and 21 (p,0.05) in cells grown on the Ca2ZnSi2O7 coating compared to CaSiO3 coated substrates and control. control. These results are consistent with the previous ALP activity findings. procollagen a1(I) mRNA expression (Fig. 6b) decreased with time for all samples, but levels were significantly higher in cells on Ca2ZnSi2O7 coating in comparison to those on CaSiO3 coating and control (p,0.05 for CaSiO3 coating and control at week 1; p,0.05 for control alone at week 2). Osteocalcin mRNA expression (Fig. 6c) increased slightly over two weeks for all samples, after which levels rapidly increased up to week 3. In comparison to CaSiO3 coating and control, osteocalcin mRNA levels in cells grown on Ca2ZnSi2O7 coating were significantly increased (p,0.05 for CaSiO3 coating and control). IGF-I mRNA expression (Fig. 6d) increased with time for all samples although Osteoblast-related Gene Expression MC3T3-E1 cells were cultured for 1, 2 and 3 weeks on control, CaSiO3 and Ca2ZnSi2O7 coatings and expression of typical bone- related genes was examined using qRT-PCR (Fig. 6). ALP mRNA expression levels (Fig. 6a) in MC3T3 cells cultured on Ca2ZnSi2O7 coating steadily increased with time and were significantly higher (p,0.05) compared to cells cultured on CaSiO3 coating and March 2013 \| Volume 8 \| Issue 3 \| e57564 PLOS ONE \| www.plosone.org PLOS ONE \| www.plosone.org 5 Zinc-Modified Ceramic Coating for Bone Implants Figure 3. MC3T3-E1 cell proliferation on control, CaSiO3 coated and Ca2ZnSi2O7 coated substrates after 1, 4, 7 and 14 days in culture. Cell proliferation of MC3T3-E1 cells cultured on the different substrates was measured by MTT assay. Ca2ZnSi2O7 coating enhanced the proliferation of preosteoblasts. All experiments were repeated in triplicate Ca2ZnSi2O7 coating compared with CaSiO3 coating or control, p,0.05. Data presented as mean 6 SD (n = 6). doi:10.1371/journal.pone.0057564.g003 response of osteoblast-like MC3T3-E1 cells cultured on uncoated Ti-6Al-4V titanium (control), CaSiO3 coated and Ca2ZnSi2O7 Figure 3. MC3T3-E1 cell proliferation on control, CaSiO3 coated and Ca2ZnSi2O7 coated substrates after 1, 4, 7 and 14 days in culture. Cell proliferation of MC3T3-E1 cells cultured on the different substrates was measured by MTT assay. Ca2ZnSi2O7 coating enhanced the proliferation of preosteoblasts. All experiments were repeated in triplicate Ca2ZnSi2O7 coating compared with CaSiO3 coating or control, p,0.05. Data presented as mean 6 SD (n = 6). doi:10.1371/journal.pone.0057564.g003 Figure 4. ALP activity in MC3T3-E1 cells cultured on control, CaSiO3 and Ca2ZnSi2O7 coated substratesfor 1, 4, 7 and 14 days. All experiments were repeated twice with n = 3 for each substrate. Ca2ZnSi2O7 coating compared with CaSiO3 coating or control, p,0.05. Results are presented as mean 6 SD. doi:10.1371/journal.pone.0057564.g004 Zinc-Modified Ceramic Coating for Bone Implants Figure 4. ALP activity in MC3T3-E1 cells cultured on control, CaSiO3 and Ca2ZnSi2O7 coated substratesfor 1, 4, 7 and 14 days. All experiments were repeated twice with n = 3 for each substrate. Ca2ZnSi2O7 coating compared with CaSiO3 coating or control, p,0.05. Results are presented as mean 6 SD. doi:10.1371/journal.pone.0057564.g004 Figure 4. ALP activity in MC3T3-E1 cells cultured on control, CaSiO3 and Ca2ZnSi2O7 coated substratesfor 1, 4, 7 and 14 days. All experiments were repeated twice with n = 3 for each substrate. Ca2ZnSi2O7 coating compared with CaSiO3 coating or control, p,0.05. Results are presented as mean 6 SD. doi:10.1371/journal.pone.0057564.g004 Figure 3. In vivo Analysis Histology. The excellent in vitro properties of Ca2ZnSi2O7 coating led to analysis of its in vivo biocompatible properties following implantation into a rabbit femur defect model. No inflammation or other implant-associated complications were observed macroscopically or microscopically in any of the histopathology sections for the implanted Ca2ZnSi2O7 coatings throughout the experimental periods. After 1.5 month of implantation, new bone was primarily observed at the Ca2Zn- Si2O7 coating periphery with direct contact to the implant surface in absence of a connective tissue layer (Fig. 7g–i). In contrast, fibrous tissue was clearly seen around the surface of the CaSiO3 coating (Fig. 7d–f) and control (Fig. 7a–c), resulting in markedly reduced bone-implant contact. New bone formation was sparse around these two implants. Ideally bioactive biomaterials need to interact actively with cells and stimulate cell growth [27]. MTT assay showed MC3T3-E1 cells could proliferate on all three samples. Furthermore, the proliferation ratio on Ca2ZnSi2O7 ceramic coating was higher than that on CaSiO3 coating or control at all time points, suggesting Ca2ZnSi2O7 coating modulates cell proliferation Figure 5. OC level in MC3T3-E1 cells cultured on control, CaSiO3 and Ca2ZnSi2O7 coated substrates for 1, 7, 14 and 21 days determined using ELISA assay. All experiments were repeated twice with n = 3 for each substrate.Ca2ZnSi2O7 coating compared with CaSiO3 coating or control, p,0.05. Results presented are the mean 6 SD. doi:10.1371/journal.pone.0057564.g005 Histomorphometry. The results of the histomorphometric measures are shown in Fig. 8. The bone-implant contact rate (BIC) was 70.6168.42% in the Ca2ZnSi2O7 coating group, which was significantly (p,0.05) higher than the values of 25.2962.43% observed in the CaSiO3 coating group and 16.4761.39% in the control group. Osteoblast-related Gene Expression MC3T3-E1 cell proliferation on control, CaSiO3 coated and Ca2ZnSi2O7 coated substrates after 1, 4, 7 and 14 days in culture. Cell proliferation of MC3T3-E1 cells cultured on the different substrates was measured by MTT assay. Ca2ZnSi2O7 coating enhanced the proliferation of preosteoblasts. All experiments were repeated in triplicate Ca2ZnSi2O7 coating compared with CaSiO3 coating or control, p,0.05. Data presented as mean 6 SD (n = 6). doi:10.1371/journal.pone.0057564.g003 response of osteoblast-like MC3T3-E1 cells cultured on uncoated Ti-6Al-4V titanium (control), CaSiO3 coated and Ca2ZnSi2O7 coated substrates and investigate their capacity to conduct osteoregeneration after implantation into a rabbit femur defect model. levels were significantly higher in cells grown on Ca2ZnSi2O7 coating compared to CaSiO3 coating and control (p,0.05 for CaSiO3 coating and control). TGF-b1 mRNA expression (Fig. 6e) increased with time for all samples and was significantly higher in cells grown on Ca2ZnSi2O7 coating compared to CaSiO3 coating and control (p,0.05 for CaSiO3 coating and control). Biomaterial surface specifications can influence adsorption of biological molecules and, in the second stage, adsorption of cells [26]. Hence, adequate adhesion and spreading of cells are prerequisite for the interaction of cells with their substrate, which will determine subsequent cellular activities including proliferation and differentiation. Investigation into cell attachment found that MC3T3-E1 cells adhere to Ca2ZnSi2O7 coating with greater spread and number of filopodia projections anchored to the surface than those adhering to CaSiO3 coating or control, which indicates that Ca2ZnSi2O7 coating has a good bioactive surface favorable for cell adhesion and growth with good biocompatibility. Discussion It is well-known that biomaterial physicochemical character- istics are important in tissue engineering because of their direct effect on cellular response and ultimately tissue regeneration [23]. Surface modifications have been applied to metallic biomaterials in order to improve their wear properties, corrosion resistance, and biocompatibility [24,25]. Our previous studies have shown the Zn-modified calcium silicate (Ca2ZnSi2O7) ceramic coating exhibits significantly improved stability in physiological solution and good bioactivity, as well as high bonding strength with the titanium alloy substrate [22]. We therefore conducted a more detailed study to examine and compare the in vitro and in vivo Figure 5. OC level in MC3T3-E1 cells cultured on control, CaSiO3 and Ca2ZnSi2O7 coated substrates for 1, 7, 14 and 21 days determined using ELISA assay. All experiments were repeated twice with n = 3 for each substrate.Ca2ZnSi2O7 coating compared with CaSiO3 coating or control, p,0.05. Results presented are the mean 6 SD. doi:10.1371/journal.pone.0057564.g005 March 2013 \| Volume 8 \| Issue 3 \| e57564 PLOS ONE \| www.plosone.org 6 Figure 6. qRT-PCR analysis of the expression of different osteoblast MC3T3-E1 cell markers (a) ALP, (b) Col a1(I), (c) OC, (d) IGF-I and (e) TGF-b1 on control, CaSiO3 and Ca2ZnSi2O7 coated substrates at weeks 1, 2 and 3. Ca2ZnSi2O7 coating compared with CaSiO3 coating and control,p,0.05. Results were normalized using GAPDH as a house keeping gene. Results presented are the mean 6 SD (n = 3) of three independent experiments. doi:10.1371/journal.pone.0057564.g006 Zinc-Modified Ceramic Coating for Bone Implants Zinc-Modified Ceramic Coating for Bone Implants Figure 6. qRT-PCR analysis of the expression of different osteoblast MC3T3-E1 cell markers (a) ALP, (b) Col a1(I), (c) OC, (d) IGF Figure 6. qRT-PCR analysis of the expression of different osteoblast MC3T3-E1 cell markers (a) ALP, (b) Col a1(I), (c) OC, (d) IGF-I and (e) TGF-b1 on control, CaSiO3 and Ca2ZnSi2O7 coated substrates at weeks 1, 2 and 3. Ca2ZnSi2O7 coating compared with CaSiO3 coating and control,p,0.05. Results were normalized using GAPDH as a house keeping gene. Results presented are the mean 6 SD (n = 3) of three independent experiments. doi:10.1371/journal.pone.0057564.g006 activity. Besides proliferation, the ability of preosteoblasts to differentiate on the biomaterials is an important stage that occurs before bone mineralization. ALP was used as a marker for early and mid-stages of osteoblast maturation and bone matrix pro- duction [28–30]. March 2013 \| Volume 8 \| Issue 3 \| e57564 Discussion In our study, MC3T3-E1 cell ALP activity on Ca2ZnSi2O7 coating exhibited higher levels of expression than on CaSiO3 coating or control from day 7. This is in accordance with the results of ALP gene regulation, indicating Ca2ZnSi2O7 coating might possess a greater ability to stimulate preosteoblast differen- tiation compared to CaSiO3 coating and control. As expected, Zn ions appear to be the relevant inducers of ALP activity, since ALP activity. Besides proliferation, the ability of preosteoblasts to differentiate on the biomaterials is an important stage that occurs before bone mineralization. ALP was used as a marker for early and mid-stages of osteoblast maturation and bone matrix pro- duction [28–30]. In our study, MC3T3-E1 cell ALP activity on Ca2ZnSi2O7 coating exhibited higher levels of expression than on CaSiO3 coating or control from day 7. This is in accordance with the results of ALP gene regulation, indicating Ca2ZnSi2O7 coating might possess a greater ability to stimulate preosteoblast differen- tiation compared to CaSiO3 coating and control. As expected, Zn ions appear to be the relevant inducers of ALP activity, since ALP is a Zn-dependent enzyme [31]. OC is secreted solely by osteoblasts and thought to play a role in cell mineralization and calcium ion homeostasis [32]. Our results show that Ca2ZnSi2O7 coating indeed markedly raises levels of OC over control levels at later periods in culture, which indicates their higher maturation state at this period. Further investigation into OC expression found mRNA levels followed a similar trend to OC levels, indicating that Zn-modified calcium silicate (Ca2ZnSi2O7) ceramic coating possesses a greater ability to promote osteoblast mineral- ization of MC3T3-E1 cells compared to CaSiO3 coating or control. Our findings are consistent with data reported by other March 2013 \| Volume 8 \| Issue 3 \| e57564 7 PLOS ONE \| www.plosone.org PLOS ONE \| www.plosone.org Zinc-Modified Ceramic Coating for Bone Implants Figure 7. Histological morphology of the interface between implant and bone tissue after 1.5 month implantation in rabbit femurs. (a–c) control, (b) magnification of the left white square in panel (a), (c) magnification of the right white square in panel (a); (d–f) CaSiO3 coating, (e) magnification of the bottom white rectangle in panel (d), (f) magnification of the top white rectangle in panel (d), (g–i) Ca2ZnSi2O7 coating, (h) magnification of the left white rectangle in panel (g), (i) magnification of the right white square in panel (g). Discussion FT, fibrous tissue; Im, implant; NB, new bone. doi:10 1371/journal pone 0057564 g007 Figure 7. Histological morphology of the interface between implant and bone tissue after 1.5 month implantation in rabbit femurs. (a–c) control, (b) magnification of the left white square in panel (a), (c) magnification of the right white square in panel (a); (d–f) CaSiO3 coating, (e) magnification of the bottom white rectangle in panel (d), (f) magnification of the top white rectangle in panel (d), (g–i) Ca2ZnSi2O7 coating, (h) magnification of the left white rectangle in panel (g), (i) magnification of the right white square in panel (g). FT, fibrous tissue; Im, implant; NB, new bone. doi:10.1371/journal.pone.0057564.g007 associated genes: ALP, Col a1(I) and OC by qRT-PCR. These genes are major phenotypic markers for preosteoblast differenti- ation during bone formation [39]. Col-I is the most abundant protein synthesized by active osteoblasts and is essential to mineral deposition, so its expression represents the start of osteoblast differentiation [40]. Col a1(I) mRNA expression (early marker, major extracellular collagenous protein for organic matrix formation) was upregulated in cells on Ca2ZnSi2O7 coating compared to CaSiO3 coating and control. The increase was pronounced during the early stage time points, which represent the beginning of osteoblast differentiation. Similar results were observed by Ehara et al, who found that calcium phosphate compounds cultured with MC3T3-E1 cells accelerated osteoblast differentiation in the early phase and promoted matrix production [41]. The results of this study demonstrate that at molecular level, osteoblast bone marker gene expression (ALP, Col a1(I), OC) on Ca2ZnSi2O7 coating is up-regulated, indicating Ca2ZnSi2O7 as research groups. Chesters [33] and Cousins [34] report that zinc is an essential trace element involved in diverse metabolic and cellular signaling pathways and is involved in the modulation of gene expression for proteins involved in bone formation. Ito et al also report that incorporation of zinc into calcium phosphate cement significantly promotes preosteoblast proliferation and differentiation in vitro [35–37]. More recently, Kwun et al. suggest zinc is involved in regulating the transcription of preosteoblast differentiation genes, such as Col-I, ALP, osteopontin and OC [38]. Hence, Zn is considered to be one promising agent for enhancing the bone-forming ability of implant materials, which can be achieved by controlling the release of Zn ions. Our studies suggest that Zn ions released form Ca2ZnSi2O7 coatings might play an important role in MC3T3-E1 cell proliferation and differentiation. Discussion In the present study, MC3T3-E1 cells were used to investigate the influence of Ca2ZnSi2O7 coating on regulation of osteoblast- March 2013 \| Volume 8 \| Issue 3 \| e57564 PLOS ONE \| www.plosone.org March 2013 \| Volume 8 \| Issue 3 \| e57564 8 Zinc-Modified Ceramic Coating for Bone Implants Figure 8. Bone-implant contact rate (BIC) of each group after 1.5 month insertion. It showed the mean BIC over the total implant length. For the Ca2ZnSi2O7 group, the BIC showed significant differences compared with CaSiO3 coating and control (p,0.05). Results presented are the mean 6 SD (n = 6) of two independent experiments. doi:10.1371/journal.pone.0057564.g008 osteoblast cells. Yamaguchi et al. also report that culture with zinc stimulates protein production of IGF-I, TGF-b1 or osteocalcin in osteoblastic MC3T3-E1 cells [50]. All the results obtained in vitro indicate that Ca2ZnSi2O7 coating is more favorable to the cellular activities of MC3T3-E1 cells, indicating that this coating might be suitable for bone regeneration and tissue engineering compared to CaSiO3 coating and control. The behavior of the bone tissue around Ca2ZnSi2O7 coating following implantation in rabbit femurs was also investigated. We found new bone was formed and in direct contact with the implants after 1.5 month of Ca2ZnSi2O7 coating implantation, in absence of fibrous tissue infiltration around the implant. In contrast, a wide band of fibrous tissue was clearly seen around the surface of the CaSiO3 coating and control. These observations were in line with the histomorphometrical data. Our results showed the direct bone contact with the implant (BIC) of Ca2ZnSi2O7 coating group was significantly greater than that for CaSiO3 coating and control after 1.5 month implantation. So, we concluded that Ca2ZnSi2O7 coatings have better bone integration properties after implantation, compared to CaSiO3 coating and control. This is consistent with the results obtained by Zreiqat et al., who observed that the incorporation of Sr and Zn into their Ca–Si system induced in vivo osteoconductivity at 3 and 6 weeks following implantation in tibial bone defects in rats [51]. Figure 8. Bone-implant contact rate (BIC) of each group after 1.5 month insertion. It showed the mean BIC over the total implant length. For the Ca2ZnSi2O7 group, the BIC showed significant differences compared with CaSiO3 coating and control (p,0.05). Results presented are the mean 6 SD (n = 6) of two independent experiments. Conclusions Plasma-sprayed Ca2ZnSi2O7 coating enhanced MC3T3-E1 cell attachment, proliferation, differentiation and up-regulated bone marker gene expression of ALP, Col a1(I) and OC and growth factors genes of IGF-I and TGF-b1, compared to CaSiO3 coating and control. The in vivo study demonstrates further efficacy of Ca2ZnSi2O7 coating in modulating bone formation around the implant and enhancement of osseointegration in absence of fibrous tissue response. In conclusion, the biological properties observed for the investigated Ca2ZnSi2O7 coating strongly suggest it is a good candidate for orthopedic and dental implant fixation. Discussion doi:10.1371/journal.pone.0057564.g008 doi:10.1371/journal.pone.0057564.g008 a more favorable substrate for MC3T3-E1 cells with the potential for applications in bone tissue regeneration. Growth factors have become an important component for tissue engineering and regenerative medicine [42–44]. In particular, IGF-I and TGF-b1 have great significance in bone tissue engineering [45]. IGF-I has been implicated as a regulator of osteoblast abundance to maintain bone matrix by enhancing osteoblast matrix production [46] and regulating osteoclast bone resorption [47]. TGF-b1 is considered another important regula- tor of osteoblast and osteoclast activity [48]. It is activated by resorbing osteoclasts which in turn attenuates further bone resorption by impairing osteoclastogenesis and promoting bone formation through chemotactic attraction of osteoblasts, enhance- ment of osteoblast proliferation and the early stages of differen- tiation with production of ECM proteins (including Col-I, osteopontin, OC) [49]. However, the effect of biomaterials on gene expressions of these factors in osteoblast cells has not been fully clarified. This study investigated mRNA expressions of IGF-I and TGF-b1 on the control, CaSiO3 and Ca2ZnSi2O7 coated substrates using qRT-PCR. We found that culture on Ca2ZnSi2O7 coating caused a significant increase in IGF-I and TGF-b1 mRNA expression in MC3T3-E1 cells, both of which are involved in the stimulation of bone formation and cell proliferation within Author Contributions Conceived and designed the experiments: MYW XJY. Performed the experiments: JMY KL. Analyzed the data: XBZH DNH MYW XJY. Contributed reagents/materials/analysis tools: XBZH DNH. Wrote the paper: MYW XJY. Conceived and designed the experiments: MYW XJY. Performed the experiments: JMY KL. Analyzed the data: XBZH DNH MYW XJY. Contributed reagents/materials/analysis tools: XBZH DNH. Wrote the paper: MYW XJY. Conceived and designed the experiments: MYW XJY. Performed the experiments: JMY KL. Analyzed the data: XBZH DNH MYW XJY. Contributed reagents/materials/analysis tools: XBZH DNH. Wrote the paper: MYW XJY. References 9. Sun LM, Berndt CC, Gross KA, Kucuk A (2001) Material fundamentals and clinical performance of plasma-sprayed hydroxyapatite coatings: A review. J Biomed Mater Res 58: 570–592. 1. Siriphannon P, Kameshima Y, Yasumori A, Okada K, Hayashi S (2000) Influence of preparation conditions on the microstructure and bioactivity of a- CaSiO3 ceramics: Formation of hydroxyapatite in simulated body fluid. J Biomed Mater Res 52: 30–39. 10. Liu XY, Tao SY, Ding CX (2002) Bioactivity of plasma sprayed dicalcium silicate coatings. Biomaterials 23: 963–968. 2. Ni SY, Chang J, Chou L, Zhai WY (2007) Comparison of osteoblast-like cell responses to calcium silicate and tricalcium phosphate ceramics in vitro. J Biomed Mater Res Part B 80B: 174–183. 11. Wu CT, Chang JA, Wang JY, Ni SY, Zhai WY (2005) Preparation and characteristics of a calcium magnesium silicate (bredigite) bioactive ceramic. Biomaterials 26: 2925–2931. 3. Xue W, Liu X, Zheng X, Ding C (2005) In vivo evaluation of plasma-sprayed wollastonite coating. Biomaterials 26: 3455–3460. 12. Ramaswamy Y, Wu CT, Zhou H, Zreiqat H (2008) Biological response of human bone cells to zinc-modified Ca-Si-based ceramics. Acta Biomater 4: 1487–1497. 4. Siriphannon P, Kameshima Y, Yasumori A, Okada K, Hayashi S (2002) Formation of hydroxyapatite on CaSiO3 powders in simulated body fluid. J European Ceram Soc 22: 511–520. 13. Lu HX, Kawazoe N, Tateishi T, Chen GP, Jin XG, et al. (2010) In vitro Proliferation and osteogenic differentiation of human bone marrow-derived mesenchymal stem cells cultured with hardystonite (Ca2ZnSi2O7) and b-TCP Ceramics. J Biomater Appl 25: 39–56. 5. Wu C, Ramaswamy Y, Soeparto A, Zreiqat H (2008) Incorporation of titanium into calcium silicate improved their chemical stability and biological properties. J Biomed Mater Res Part A. 86A: 402–410. 14. Wu CT, Ramaswamy Y, Kwik D, Zreiqat H (2007) The effect of strontium incorporation into CaSiO3 ceramics on their physical and biological properties. Biomaterials 28: 3171–3181. 6. Liu XY, Chu PK, Ding CX (2004) Surface modification of titanium, titanium alloys, and related materials for biomedical applications. Mater Sci Eng R-Rep 47: 49–121. 15. Ramaswamy Y, Wu C, Van Hummel A, Combes V, Grau G, et al. (2008) The responses of osteoblasts, osteoclasts and endothelial cells to zirconium modified calcium-silicate-based ceramic. Biomaterials 29: 4392–4402. 7. Liu XY, Ding CX, Wang ZY (2001) Apatite formed on the surface of plasma- sprayed wollastonite coating immersed in simulated body fluid. Biomaterials 22: 2007–2012. calcium-silicate-based ceramic. Biomaterials 29: 4392–4402. 8. 1. Siriphannon P, Kameshima Y, Yasumori A, Okada K, Hayashi S (2000) Influence of preparation conditions on the microstructure and bioactivity of a- CaSiO3 ceramics: Formation of hydroxyapatite in simulated body fluid. J Biomed Mater Res 52: 30–39. Zinc-Modified Ceramic Coating for Bone Implants Tissue Eng 13: 1135–1150. 25. Ma PX (2008) Biomimetic materials for tissue engineering. Adv Drug Deliv Rev 60: 184–198. 26. Zeng HT, Chittur KK, Lacefield WR (1999) Analysis of bovine serum albumin adsorption on calcium phosphate and titanium surfaces. Biomaterials 20: 377– 384. g 43. Levenberg S, Huang NF, Lavik E, Rogers AB, Itskovitz Eldor J, et al. (2003) Differentiation of human embryonic stem cells on three-dimensional polymer scaffolds. Proc Natl Acad Sci USA 100: 12741–12746. 27. Sader MS, LeGeros RZ, Soares GA (2009) Human osteoblasts adhesion and proliferation on magnesium-substituted tricalcium phosphate dense tablets. J Mater Sci: Mater Med 20: 521–527. 44. Richardson TP, Peters MC, Ennett AB, Mooney DJ (2001) Polymeric system for dual growth factor delivery. Nat Biotechnol 19: 1029–1034. 45. Takahashi T, Ogasawara T, Kishimoto J, Liu GY, Asato H, et al. (2005) Synergistic effects of FGF-2 with insulin or IGF-I on the proliferation of human auricular chondrocytes. Cell Transplant 14: 683–693. J 28. Benoit DSW, Schwartz MP, Durney AR, Anseth KS (2008) Small functional groups for controlled differentiation of hydrogel-encapsulated human mesen- chymal stem cells. Nat Mater 7: 816–823. 46. Canalis E, Rydziel S, Delany AM, Varghese S, Jeffrey JJ (1995) Insulin-like growth-factors inhibit interstitial collagenase synthesis in bone cell-cultures. Endocrinology 136: 1348–1354. 29. Malaval L, Modrowski D, Gupta AK, Aubin JE (1994) Cellular expression of bone-related proteins during in vitro osteogenesis in rat bone-marrow stromal cell-cultures. J Cell Physiol 158: 555–572. 47. Hill PA, Reynolds JJ, Meikle MC (1995) Osteoblasts mediate insulin-like growth- factor-I and growth-factor-II stimulation of osteoclast formation and function. Endocrinology 136: 124–131. y 30. Maddi A, Hai H, Ong ST, Sharp L, Harris M, et al. (2006) Long wave ultrasound may enhance bone regeneration by altering OPG/RANKL ratio in human osteoblast-like cells. Bone 39: 283–288. 48. Fox SW, Lovibond AC (2005) Current insights into the role of transforming growth factor-b in bone resorption. Mol Cell Endocrinol 243: 19–26. 31. Habibovic P, Barralet JE (2011) Bioinorganics and biomaterials: Bone repair. Acta Biomater 7: 3013–3026. 49. Janssens K, ten Dijke P, Janssens S, Van Hul W (2005) Transforming growth factor-b1 to the bone. Endocr Rev 26: 743–774. 32. Lee NK, Sowa H, Hinoi E, Ferron M, Ahn JD, et al. (2007) Endocrine regulation of energy metabolism by the skeleton. Cell 130: 456–469. 50. Zinc-Modified Ceramic Coating for Bone Implants Zinc-Modified Ceramic Coating for Bone Implants 36. Kawamura H, Ito A, Miyakawa S, Layrolle P, Ojima K, et al. (2000) Stimulatory effect of zinc-releasing calcium phosphate implant on bone formation in rabbit femora. J Biomed Mater Res 50: 184–190. 17. Liang Y, Xie YT, Ji H, Huang LP, Zheng XB (2010) Excellent stability of plasma-sprayed bioactive Ca3ZrSi2O9 ceramic coating on Ti-6Al-4V. Appl Surf Sci 256: 4677–4681. 18. Tapiero H, Tew KD (2003) Trace elements in human physiology and pathology: zinc and metallothioneins. Biomed Pharmacother 57: 399–411. 37. Ito A, Otsuka M, Kawamura H, Ikeuchi M, Ohgushi H, et al. (2005) Zinc- containing tricalcium phosphate and related materials for promoting bone formation. Curr Appl Phys 5: 402–406. 19. Yamaguchi M, Oishi H, Suketa Y (1987) Stimulatory effect of zinc on bone formation in tissue culture. Biochemical pharmacology 36: 4007–4012. g y formation in tissue culture. Biochemical pharmacology 36: 4007–4 38. Kwun IS, Cho YE, Lomeda RAR, Shin HI, Choi JY, et al. (2010) Zinc deficiency suppresses matrix mineralization and retards osteogenesis transiently with catch-up possibly through Runx 2 modulation. Bone 46: 732–741. 20. Hsieh HS, Navia JM (1980) Zinc deficiency and bone formation in guinea pig alveolar implants. J Nutr 110: 1581–1588. p J 21. Oner G, Bhaumick B, Bala RM (1984) Effect of zinc deficiency on serum somatomedin levels and skeletal growth in young rats. Endocrinology 114: 1860–1863. 39. Stein GS, Lian JB, Owen TA (1990) Relationship of cell growth to the regulation of tissue-specific gene expression during osteoblast differentiation. FASEB J 4: 3111–3123. 22. Li K, Yu J, Xie Y, Huang L, Ye X, et al. (2011) Chemical stability and antimicrobial activity of plasma sprayed bioactive Ca2ZnSi 2O7)coating. J Mater sci: Mater Med 22: 2781–2789. 40. Franceschi RT, Iyer BS (1992) Relationship between collagen synthesis and expression of the osteoblast phenotype in MC3T3-E1 cells. J Bone Miner Res 7: 235–246. 23. Anselme K (2000) Osteoblast adhesion on biomaterials. Biomaterials 21: 667– 681. 41. Ehara A, Ogata K, Imazato S, Ebisu S, Nakano T, et al. (2003) Effects of a-TCP and TetCP on MC3T3-E1 proliferation, differentiation and mineralization. Biomaterials 24: 831–836. 24. Guleryuz H, Cimenoglu H (2004) Effect of thermal oxidation on corrosion and corrosion-wear behaviour of a Ti-6A1–4V alloy. Biomaterials 25: 3325–3333. 42. Kimelman N, Pelled G, Helm GA, Huard J, Schwarz EM, et al. (2007) Review: Gene and stem cell based therapeutics for bone regeneration and repair. References Liu XY, Ding CX, Chu PK (2004) Mechanism of apatite formation on wollastonite coatings in simulated body fluids. Biomaterials 25: 1755–1761. 16. Liu XY, Ding CX (2002) Plasma sprayed wollastonite/TiO2 composite coatings on titanium alloys. Biomaterials 23: 4065–4077. 8. Liu XY, Ding CX, Chu PK (2004) Mechanism of apatite formation on wollastonite coatings in simulated body fluids. Biomaterials 25: 1755–1761. 9 March 2013 \| Volume 8 \| Issue 3 \| e57564 March 2013 \| Volume 8 \| Issue 3 \| e57564 PLOS ONE \| www.plosone.org Zinc-Modified Ceramic Coating for Bone Implants Zinc-Modified Ceramic Coating for Bone Implants Yamaguchi M, Hashizume M (1994) Effect of b-alanyl-L-histidinato zinc on protein-components in osteoblastic MC3T3-El cells increase in osteocalcin, insulin-like growth-factor-I and transforming growth-factor-b. Mol Cell Biochem 136: 163–169. 33. Chesters JK (1998) Molecular biological approaches to nutrient-gene interac- tions. Proc Nutr Soc 57: 249–256. 34. Cousins RJ (1998) A role of zinc in the regulation of gene expression. Proc Nutr Soc 57: 307–311. 51. Zreiqat H, Ramaswamy Y, Wu CT, Paschalidis A, Lu ZF, et al. (2010) The incorporation of strontium and zinc into a calcium-silicon ceramic for bone tissue engineering. Biomaterials 31: 3175–3184. 35. Ito A, Kawamura H, Otsuka M, Ikeuchi M, Ohgushi H, et al. (2002) Zinc- releasing calcium phosphate for stimulating bone formation. Mater Sci Eng C- Biomimetic Supramol Syst 22: 21–25. March 2013 \| Volume 8 \| Issue 3 \| e57564 PLOS ONE \| www.plosone.org PLOS ONE \| www.plosone.org 10
https://openalex.org/W2482597709	http://publications.iodp.org/proceedings/353/EXP_REPT/CHAPTERS/353_101.PDF	English	null	Expedition 353 summary	Proceedings of the International Ocean Discovery Program. Expedition reports	2,016	cc-by	31,429	1 Clemens, S.C., Kuhnt, W., LeVay, L.J., Anand, P., Ando, T., Bartol, M., Bolton, C.T., Ding, X., Gariboldi, K., Giosan, L., Hathorne, E.C., Huang, Y., Jaiswal, P., Kim, S., Kirkpatrick, J.B., Littler, K., Marino, G., Martinez, P., Naik, D., Peketi, A., Phillips, S.C., Robinson, M.M., Romero, O.E., Sagar, N., Taladay, K.B., Taylor, S.N., Thirumalai, K., Uramoto, G., Usui, Y., Wang, J., Yamamoto, M., and Zhou, L., 2016. Expedition 353 summary. In Clemens, S.C., Kuhnt, W., LeVay, L.J., and the Expedition 353 Scientists, Indian Monsoon Rainfall. Proceedings of the International Ocean Discovery Program, 353: College Station, TX (International Ocean Discovery Program). http://dx.doi.org/10.14379/iodp.proc.353.101.2016 2 Expedition 353 Scientists’ addresses. MS 353-101: Published 29 July 2016 doi:10.14379/iodp.proc.353.101.2016 Contents 1 Abstract 1 Introduction 2 Objectives 2 Background 10 Site summaries 27 Expedition 353 synthesis 27 Expedition 353 science assessment 28 References Expedition 353 summary1 S.C. Clemens, W. Kuhnt, L.J. LeVay, P. Anand, T. Ando, M. Bartol, C.T. Bolton, X. Ding, K. Gariboldi, L. Giosan, E.C. Hathorne, Y. Huang, P. Jaiswal, S. Kim, J.B. Kirkpatrick, K. Littler, G. Marino, P. Martinez, D. Naik, A. Peketi, S.C. Phillips, M.M. Robinson, O.E. Romero, N. Sagar, K.B. Taladay, S.N. Taylor, K. Thirumalai, G. Uramoto, Y. Usui, J. Wang, M. Yamamoto, and L. Zhou2 Keywords: International Ocean Discovery Program, IODP, Expedition 353, JOIDES Resolution, Site U1443, Site U1444, Site U1445, Site U1446, Site U1447, Site U1448, Indian monsoon, monsoon, Bay of Bengal, Andaman Sea, paleoclimate, paleoceanography, Miocene, Pliocene, Pleistocene, Holocene, Cretaceous, Paleogene, Ninetyeast Ridge, Indian Ocean, salinity, Bengal Fan, orbital, millennial, centennial, abrupt climate change Abstract ween river basins. IODP Site U1443 (Ninetyeast Ridge) is an open- ocean site with modern surface water salinity very near to the global mean but is documented to have recorded changes in monsoonal circulation over orbital to tectonic timescales. This site serves as an anchor for establishing the extent to which the north to south (19°N to 5°N) salinity gradient changes over time. International Ocean Discovery Program (IODP) Expedition 353 (29 November 2014–29 January 2015) drilled six sites in the Bay of Bengal, recovering 4280 m of sediments during 32.9 days of on-site drilling. Recovery averaged 97%, including coring with the ad- vanced piston corer, half-length advanced piston corer, and ex- tended core barrel systems. The primary objective of Expedition 353 is to reconstruct changes in Indian monsoon circulation since the Miocene at tectonic to centennial timescales. Analysis of the sediment sections recovered will improve our understanding of how monsoonal climates respond to changes in forcing external to the Earth’s climate system (i.e., insolation) and changes in forcing inter- nal to the Earth’s climate system, including changes in continental ice volume, greenhouse gas concentrations, sea level, and the ocean-atmosphere exchange of energy and moisture. All of these mechanisms play critical roles in current and future climate change in monsoonal regions. Clemens, S.C., Kuhnt, W., LeVay, L.J., and the Expedition 353 Scientists Proceedings of the International Ocean Discovery Program Volume 353 publications.iodp.org Clemens, S.C., Kuhnt, W., LeVay, L.J., and the Expedition 353 Scientists Proceedings of the International Ocean Discovery Program Volume 353 publications.iodp.org Clemens, S.C., Kuhnt, W., LeVay, L.J., and the Expedition 353 Scientists Proceedings of the International Ocean Discovery Program Volume 353 publications.iodp.org 1 Clemens, S.C., Kuhnt, W., LeVay, L.J., Anand, P., Ando, T., Bartol, M., Bolton, C.T., Ding, X., Gariboldi, K., Giosan, L., Hathorne, E.C., Huang, Y., Jaiswal, P., Kim, S., Kirkpatrick, J.B., Littler, K., Marino, G., Martinez, P., Naik, D., Peketi, A., Phillips, S.C., Robinson, M.M., Romero, O.E., Sagar, N., Taladay, K.B., Taylor, S.N., Thirumalai, K., Uramoto, G., Usui, Y., Wang, J., Yamamoto, M., and Zhou, L., 2016. Expedition 353 summary. In Clemens, S.C., Kuhnt, W., LeVay, L.J., and the Expedition 353 Scientists, Indian Monsoon Rainfall. Proceedings of the International Ocean Discovery Program, 353: College Station, TX (International Ocean Discovery Program). http://dx.doi.org/10.14379/iodp.proc.353.101.2016 2 Expedition 353 Scientists’ addresses Campanian–Miocene • Understand the timing and conditions under which monsoonal circulation initiated and reconstruct the variability of the Indian monsoon at orbital timescales; and Figure F2. Location of Expedition 353 sites and seafloor depths. Map was generated using GeoMapApp (http://www.geomapapp.org), using topog- raphy and bathymetry from the Global Multi-Resolution Topography syn- thesis (Ryan et al., 2009). • Understand the relationship between Indian monsoon variabil- ity and major past global climatic events such as Oligocene/Mio- cene cooling (Zachos et al., 1997), the onset of the mid-Miocene climatic optimum (Holbourn et al., 2007, 2015; Zachos et al., 2001), and mid-Miocene cooling and Antarctic cryosphere ex- pansion (Holbourn et al., 2013, 2014). • Understand the relationship between Indian monsoon variabil- ity and major past global climatic events such as Oligocene/Mio- cene cooling (Zachos et al., 1997), the onset of the mid-Miocene climatic optimum (Holbourn et al., 2007, 2015; Zachos et al., 2001), and mid-Miocene cooling and Antarctic cryosphere ex- pansion (Holbourn et al., 2013, 2014). Site U1446 (1430 mbsl) Site U1447 (1391 mbsl) Site U1444 (3132 mbsl) Site U1445 (2502 mbsl) Site U1443 (2925 mbsl) Site U1448 (1097 mbsl) 22° N 78°E 80° 82° 84° 86° 88° 90° 92° 94° 96° 98° 20° 18° 16° 14° 12° 10° 8° 6° Bengal that span 5°N to 20°N (Figure F2). Analysis of these sedi- ments will help to address the scientific objectives outlined below. Scientific drilling last occurred in the Bay of Bengal in 1972 (Deep Sea Drilling Project Leg 22) when the D/V Glomar Challenger drilled Sites 217 and 218 in the southernmost Bay of Bengal (8°N to 9°N). Scientific studies performed on India’s National Gas Hydrate Program Expedition 1 cores collected aboard the JOIDES Resolution in 2006 (Collett et al., 2008; Ramana et al., 2014), which provided the foundation for selecting most of the sites for Expedition 353, in- dicate the excellent potential for paleoclimatic and paleoceano- graphic reconstructions in the region (i.e., Ponton et al., 2012; Cawthern et al., 2014; Flores et al., 2014; Johnson et al., 2014; Phil- lips et al., 2014a, 2014b; Ali et al., 2015). Site U1446 (1430 mbsl) Site U1447 (1391 mbsl) Site U1444 (3132 mbsl) Site U1445 (2502 mbsl) Site U1443 (2925 mbsl) Site U1448 (1097 mbsl) 22° N 78°E 80° 82° 84° 86° 88° 90° 92° 94° 96° 98° 20° 18° 16° 14° 12° 10° 8° 6° Pliocene–Pleistocene ODP Leg 117 IODP Exp 355 IODP Exp 359 IODP Exp 353 IODP Exp 346 IODP Exp 363 ODP Leg 184 Figure F2. Location of Expedition 353 sites and seafloor depths. Map was generated using GeoMapApp (http://www.geomapapp.org), using topog- raphy and bathymetry from the Global Multi-Resolution Topography syn- thesis (Ryan et al., 2009). • Establish the sensitivity and timing of changes in monsoon cir- culation relative to insolation forcing, latent heat export from the Southern Hemisphere, global ice volume extent, and green- house gas concentrations; • Determine the extent to which Indian and East Asian monsoon winds and precipitation are coupled and at what temporal and geographic scales; • Better separate the effects of climate change and tectonics on erosion and runoff; • Provide verification targets for climate models, including the rapidly evolving water isotope–enabled, time-dependent mod- els; and • Understand the relationship between Indian monsoon variabil- ity and the Pliocene–Pleistocene intensification of Northern Hemisphere glaciation. Oligocene–present • Establish a complete Oligocene–present astronomically tuned timescale based on high-resolution benthic and planktonic iso- tope reference curves for the Indian Ocean, and • Establish a complete Oligocene–present astronomically tuned timescale based on high-resolution benthic and planktonic iso- tope reference curves for the Indian Ocean, and • Integrate high-resolution distribution studies of well-preserved Oligocene–recent calcareous and siliceous microfossils from the Indian Ocean into global compilation studies of paleoclimatic and biotic evolution. Background Motivation for drilling in the Bay of Bengal Pliocene–Pleistocene Introduction The R/V JOIDES Resolution has conducted scientific ocean drilling in many of the marginal basins surrounding monsoon-influ- enced regions of India and Asia including the Arabian Sea (Ocean Drilling Program [ODP] Leg 117), the South China Sea (ODP Leg 184), the East China Sea (Integrated Ocean Drilling Program Expe- dition 346), and the marginal sea bordered by the Eurasian conti- nent, the Korean Peninsula, and the Japanese Islands (Expedition 346). Sediments recovered have been used to reconstruct changes in summer monsoon upwelling and eolian transport (Leg 117), summer and winter monsoon surface and intermediate water dy- namics in the northern and southern regions of the South China Sea (Leg 184), the influence of the westerlies on monsoonal circulation in the marginal sea bordered by the Eurasian continent, the Korean Peninsula, and the Japanese Islands (Expedition 346), and the influ- ence of Yangtze River runoff on the surface waters of the East China Sea (Expedition 346). These records will be complemented by fu- ture drilling during scheduled International Ocean Discovery Pro- gram (IODP) expeditions in the Eastern Arabian Sea (Expedition 355), the Maldives (Expedition 359), and the Timor Sea (Expedition 363) (Figure F1). Prior to Expedition 353, however, no scientific drilling has occurred in the core convective region of the Indo- Asian monsoon system, the northern Bay of Bengal. Expedition 353 fills this scientific/geographic gap by drilling targets in the Bay of The primary signal targeted is the exceptionally low salinity sur- face waters that result, in roughly equal measure, from both direct summer monsoon precipitation above the Bay of Bengal and runoff from the numerous large river basins that drain into the Bay of Ben- gal. Changes in rainfall and surface ocean salinity are captured and preserved in a number of chemical, physical, isotopic, and biological components of sediments deposited in the Bay of Bengal. Expedi- tion 353 sites are strategically located in key regions where these signals are the strongest and best preserved. Salinity changes at IODP Sites U1445 and U1446 (northeast Indian margin) result from direct precipitation as well as runoff from the Ganges-Brahmaputra river complex and the many river basins of peninsular India. Salinity changes at IODP Sites U1447 and U1448 (Andaman Sea) result from direct precipitation and runoff from the Irrawaddy and Sal- Expedition 353 summary S.C. Clemens et al. Motivation for drilling in the Bay of Bengal Pliocene–Pleistocene vation for drilling in the Bay of Bengal Objectives Figure F1. ODP and IODP monsoon-related expeditions. Satellite image from Stöckli et al. (2005). Figure F1. ODP and IODP monsoon-related expeditions. Satellite image from Stöckli et al. (2005). Pliocene–Pleistocene The three regions chosen for drilling during Expedition 353 (Mahanadi basin, Andaman Islands, and northern Ninetyeast Ridge) delineate a north–south transect designed to capture past changes in the exceptionally strong salinity gradient observed in the modern Bay of Bengal (19°N to 5°N). A threefold motivation exists for targeting this precipitation/salinity signal. Bengal that span 5°N to 20°N (Figure F2). Analysis of these sedi- ments will help to address the scientific objectives outlined below. Scientific drilling last occurred in the Bay of Bengal in 1972 (Deep Sea Drilling Project Leg 22) when the D/V Glomar Challenger drilled Sites 217 and 218 in the southernmost Bay of Bengal (8°N to 9°N). Scientific studies performed on India’s National Gas Hydrate Program Expedition 1 cores collected aboard the JOIDES Resolution in 2006 (Collett et al., 2008; Ramana et al., 2014), which provided the foundation for selecting most of the sites for Expedition 353, in- dicate the excellent potential for paleoclimatic and paleoceano- graphic reconstructions in the region (i.e., Ponton et al., 2012; Cawthern et al., 2014; Flores et al., 2014; Johnson et al., 2014; Phil- lips et al., 2014a, 2014b; Ali et al., 2015). First, the Bay of Bengal/Andaman Sea and surrounding catch- ments are within the Earth’s strongest hydrological regime (Figure F3), impacting billions of people; a solid understanding of the phys- ics behind monsoonal climate change is of significant societal rele- vance (Nicholls et al., 2007). The net annual surface water exchange (precipitation plus runoff minus evaporation) within the Bay of Bengal and Andaman Sea during the summer monsoon is 184 × 1010 m3, dominating the winter signal of −32 × 1010 m3 for an annual average of 152 × 1010 m3 (Varkey et al., 1996). The effects of this Volume 353 IODP Proceedings 2 S.C. Clemens et al. Expedition 353 summary Figure F3. Climatological (A) July–August and (B) January–February mean precipitation rates (shading in mm/day) and 925 hPa wind vectors (arrows). Precipitation and wind climatology are derived from CMAP (Xie and Arkin, 1997) (1979–2000) and NCEP/NCAR reanalysis (1951–2000), respectively. coupling high- and low-latitude climate change. However, the phys- ical mechanisms behind these links are not fully understood; this was a primary goal of Expedition 346. Progress is also being made in understanding winter monsoon and summer monsoon linkages at the millennial timescale. Pliocene–Pleistocene For example, recent work offshore Goa, western India, shows synchronous breakdown in summer and win- ter monsoon airflow over the Arabian Sea during Heinrich events (Singh et al., 2011), which is in contrast to the East Asian Monsoon system that shows an asynchronous relationship between summer and winter monsoon strength at the millennial scale (Yancheva et al., 2007). A July-August mean 5 6 7 8 9 11 12 13 14 15 16 17 18 4 3 2 20 m/s Consensus does not yet exist on the extent of the coupling or the ultimate forcing of monsoon winds and precipitation at orbital and longer timescales (An et al., 2011; Caley et al., 2011a, 2011b, 2011c; Cheng et al., 2009; Clemens and Prell, 2007; Clemens et al., 1996, 1991, 2008; Clift et al., 2008; Ruddiman, 2006; Wang et al., 2008; Ziegler et al., 2010). Some argue for close coupling between changes in Indian and East Asian summer monsoon winds and precipitation across the entire region spanning the Arabian Sea (Leg 117), the South China Sea (Leg 184), and terrestrial records from eastern China. In this case, changes in the strength of summer monsoon circulation across these regions are thought to be sensitive to Northern Hemisphere sensible heating (insolation), the timing of energy release from the Southern Hemisphere Indian Ocean, and the timing of global ice volume minima (An et al., 2011; Caley et al., 2011b, 2011c; Clemens and Prell, 2003, 2007; Clemens et al., 1996, 2008). In contrast, others interpret the timing of summer monsoon circulation on the basis of speleothem records from southeast China, as forced directly by external insolation with little or no in- fluence from internal boundary conditions such as ice volume or Southern Hemisphere ocean–atmosphere latent heat exchange (Cheng et al., 2009; Ruddiman, 2006; Wang et al., 2008). Caballero- Gill et al. (2012) demonstrate that these contrasting interpretations are not attributable to differences in terrestrial and marine chronol- ogies. Therefore, this lack of consensus points either to a strong deficit in our understanding of monsoon sensitivity to changes in the fundamental boundary conditions including insolation, ocean/atmosphere energy exchange, ice volume, and atmospheric greenhouse gas concentrations or to the confounding influence of seasonality on interpretation of proxy records. Pliocene–Pleistocene The latter is illus- trated in Figure F5 where rain-bearing air masses span both winter and summer seasons (eastern China cave region, lower panel), whereas rain-bearing air masses are strongly biased toward the summer season for the Mahanadi basin (upper panel). 5 6 7 8 9 11 12 13 14 15 16 17 18 4 3 2 B January-February mean 5 6 7 8 10 11 12 13 14 15 16 2 1 4 3 9 20 m/s (mm/day) B 5 6 7 8 10 11 12 13 14 15 16 2 1 4 3 9 (mm/day) budget are clearly evident in the surface salinity climatology (Figure F4) (Antonov et al., 2010), indicating a well-defined, strong signal that can be used to monitor changes in monsoonal precipitation via chemical, physical, and isotopic indicators for changes in precipita- tion, salinity, and terrestrial erosion/runoff. The extent of the low- salinity surface waters is sufficient to mute (via stratification) what would otherwise yield strong summer season productivity in re- sponse to wind-driven upwelling along the eastern Indian margin, similar to that seen in the Arabian Sea (Guptha et al., 1997; Kumar et al., 2002; Phillips et al., 2014b). Hence, the Bay of Bengal is opti- mal for isolating and recording the summer monsoon precipitation signal in that the vast majority of precipitation, sourced from the Southern Hemisphere Indian Ocean, occurs during boreal summer (Figure F5). Lack of consensus also extends to the tectonic scale, where the timing of monsoon intensification to modern strength is debated. Some proxy records suggest intensification occurred at ~7–8 Ma (e.g., Kroon et al., 1991; Prell et al., 1992), whereas others suggest a considerably earlier intensification, perhaps as early as early Mio- cene (~22 Ma) (Clift et al., 2008; Guo et al., 2002; Sun and Wang, 2005) or even late Eocene (~45 Ma) (Licht et al., 2014). Emergence and expansion of arid-adapted C4 flora in South Asia argues for re- duced precipitation since ~8 Ma (e.g., Cerling et al., 1997; Huang et al., 2007; Quade and Cerling, 1995), whereas proxies dedicated to reconstructing seasonality instead suggest little variability in the monsoon over the last 10 My (Dettman et al., 2001). Clift and Plumb (2008), Molnar et al. (2010), and the report from the Detailed Plan- ning Group “Asian Monsoon and Cenozoic Tectonic History” (http://www.iodp.org/doc_download/2336-mmdpgreport) pro- vide comprehensive overviews of these issues. More recently, Rodri- guez et al. IODP Proceedings Pliocene–Pleistocene 28 32 28 30 32 31 33 31 33 33 33 3231 33 32 31 34 34 34 33 32 33 32 30 33 32 34 31 33 32 33 32 34 31 33 32 33 32 34 31 33 32 30 33 32 34 31 24 20 20 24 24 28 31 28 36 36 36 36 33 32 30 33 32 34 31 24 20 24 28 31 28 24 33 1 33 32 34 36 20 28 31 28 24 36 29 30 31 31 20 32 24 33 34 36 33 28 33 32 3130 30 33 33 32 34 30 28 31 29 24 36 33 30 32 24 28 32 33 33 32 34 30 28 34 29 24 36 33 30 32 1 33 31 30 32 3231 29 29 33 35 34 35 34 32 28 35 35 32 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec U1445/U1446 U1445/U1446 U1445/U1446 U1445/U1446 U1445/U1446 U1445/U1446 U1445/U1446 U1445/U1446 U1445/U1446 U1445/U1446 U1445/U1446 U1445/U1446 U1447/U1448 U1443 U1447/U1448 U1443 U1447/U1448 U1443 U1447/U1448 U1443 U1447/U1448 U1443 U1447/U1448 U1443 U1447/U1448 U1443 U1447/U1448 U1443 U1447/U1448 U1443 U1447/U1448 U1443 U1447/U1448 U1443 U1447/U1448 U1443 33 1 33 32 34 36 20 28 31 28 24 36 29 30 31 32 Nov U1445/U1446 U1447/U1448 U1443 33 1 33 32 34 36 20 28 31 28 24 36 29 30 31 32 Nov U1445/U1446 U1447/U1448 U1443 31 20 32 24 33 34 36 33 28 33 32 3130 30 Dec U1445/U1446 U1447/U1448 U1443 on the basis of increased Globigerina bulloides concentrations may be an artifact related to uplift of the Owen Ridge at this time, result- ing in enhanced carbonate preservation. This, however, cannot ac- count for the similar signal observed in the radiolarian assemblage data (Prell et al., 1992). on the basis of increased Globigerina bulloides concentrations may be an artifact related to uplift of the Owen Ridge at this time, result- ing in enhanced carbonate preservation. This, however, cannot ac- count for the similar signal observed in the radiolarian assemblage data (Prell et al., 1992). plication in the interpretation of the OMZ signal as a direct response to atmospheric circulation in the core region of summer monsoon winds (i.e., oxygen drawdown in response to decay of up- welling-produced organic carbon). Pliocene–Pleistocene (2014) suggested that the ~8 Ma intensification inferred Second, recent studies have called into question the extent to which basin-scale monsoon winds and continental precipitation are coupled over a range of timescales and spatial scales (Clemens et al., 2010; Clemens and Prell, 2007; Liu et al., 2006; Molnar, 2005; Ruddi- man, 2006; Wang et al., 2008; Ziegler et al., 2010). Nearly all proxy records indicate strong coupling between summer monsoon winds and precipitation across the Indo-Asian monsoon subsystems at the millennial scale (Altabet et al., 2002; Cai et al., 2006; Clemens, 2005; Schulz et al., 1998; Sun et al., 2011; Wang et al., 2001); this tight cou- pling is likely attributed to the strong role of the winter westerlies in 3 Volume 353 S.C. Clemens et al. Expedition 353 summary Figure F4. World Ocean Atlas Monthly Mean Salinity for 1955–2006 (Antonov et al., 2010). Areas drilled in the Andaman Sea (Sites U1447 and U1448), Mahanadi basin, northern Bay of Bengal (Sites U1445 and U1446), and Ninetyeast Ridge (Site U1443) are shown. sis of increased Globigerina bulloides concentrations may fact related to uplift of the Owen Ridge at this time, result- hanced carbonate preservation. This, however, cannot ac- the similar signal observed in the radiolarian assemblage l et al., 1992). , recent work suggests that interpretation of the oxygen m zone (OMZ) signal in the northern Arabian Sea (Leg 117) omplicated by changing oxygen content of southern-source ( d l l l h plication in the interpretation of the OMZ signal response to atmospheric circulation in the core region monsoon winds (i.e., oxygen drawdown in response to welling-produced organic carbon). Precipitation, salini off indicators are not influenced by the chemistry o sourced intermediate and deep water masses, offering t to disentangle the influences of these factors in interp soon proxy records. h d d ll ff h hern Bay of Bengal (Sites U1445 and U1446), and Ninetyeast Ridge (Site U1443) are shown. Pliocene–Pleistocene This is consistent with World Meteorological Organization rainfall records indi- cating that June, July, and August rainfall only accounts for 50% of the total annual rainfall in eastern China. The Bay of Bengal location isolates summer season dynamics. Deep-sea benthic foraminiferal diversity in the Indian Ocean further decreased between 8 and 6 Ma and is associated with a neg- ative δ13C shift at 3.2–2.3 and 1.6–0.9 Ma, coinciding with an in- creased abundance of species indicative of increased organic carbon flux (Singh and Gupta, 2005). Since ~2.8 Ma, roughly coeval with the onset of Northern Hemisphere glaciation, benthic foraminiferal species that are well adapted to seasonally strong fluctuating carbon flux dominate the assemblages. This has been related to increased duration and strength of the northeast (winter) monsoon, which is accompanied by relatively low primary production in the eastern equatorial Indian Ocean (Gupta and Thomas, 2003). Expedition 353 will provide new Neogene intermediate water benthic foramin- iferal assemblage, δ18O, and δ13C records along a meridional transect to better understand the relative contribution of monsoon- induced changes in carbon flux and changes in intermediate water circulation linked to high-latitude climatic events such as fluctua- tions in the extension of the East Antarctic ice sheet. Analysis of Bay of Bengal surface sediments indicates that the foraminiferal lyso- cline, the depth delimiting well-preserved from noticeably dissolved assemblages, shoals significantly from south to north. The foramin- iferal lysocline rises from 3800 to 3300 m between 0°N and 7°N (about Site U1443) then systematically shoals to ~2000 m at ~20°N (Indian margin sites) (Cullen and Prell, 1984), indicating enhanced carbon flux to the deep sea in the northern part of the Bay of Ben- gal. The Expedition 353 meridional transect provides a unique op- portunity to investigate the nature and timing of variations in deep- water radiogenic isotope composition in response to restriction of the deepwater connection between the Pacific and Indian Oceans through the Indonesian Gateway since the mid-Miocene. Com- pared to the fan sediments of IODP Expedition 354, Expedition 353 sediments will help to disentangle climatic variability from direct tectonic influences using non-Himalayan sedimentary sequences on the Indian continental margin, and to evaluate the influence of enhanced Himalayan erosion since the late Oligocene at sites af- fected by a Himalayan signal in Andaman Sea and Ninetyeast Ridge sediments. outstanding issues. Pliocene–Pleistocene Resolving these questions using the geological archive is important for providing verification targets for climate models, especially given that the majority of current Atmosphere- Ocean General Circulation Models used in the Intergovernmental Panel on Climate Change reports do not accurately simulate the spatial or intraseasonal variability of monsoon precipitation (Ran- dall et al., 2007; Sharmila et al., 2015). The broad Oligocene passage between the Indian and Pacific Oceans must have enabled significant surface and intermediate wa- ter exchange and the possibility of deepwater flow from the Indian Ocean to the Pacific Ocean (Thomas et al., 2003). The progressive closure of the Indonesian Gateway caused by the northward move- ment of Australia (Hall, 2002; Hall et al., 2011; Kuhnt et al., 2004) induced changes in deepwater and intermediate water circulation through the Indonesian Gateway, which may have resulted in a sig- nificant change in the Eastern Indian Ocean neodymium (Nd) iso- tope composition during the middle Miocene (Frank et al., 2006; Gourlan et al., 2008). A second shift in Eastern Indian Ocean Nd Pliocene–Pleistocene Precipitation, salinity, and run- off indicators are not influenced by the chemistry of externally sourced intermediate and deep water masses, offering the potential to disentangle the influences of these factors in interpreting mon- soon proxy records. Third, recent work suggests that interpretation of the oxygen minimum zone (OMZ) signal in the northern Arabian Sea (Leg 117) may be complicated by changing oxygen content of southern-source intermediate waters (Anand et al., 2008; Caley et al., 2011c; Schmit- tner et al., 2007; Ziegler et al., 2010). This presents a potential com- The Expedition 353 drilling effort, targeting the core geographic region of summer monsoon precipitation, directly addresses these IODP Proceedings Volume 353 4 S.C. Clemens et al. Expedition 353 summary creases in productivity started at 14 Ma (Gupta et al., 2013) or 10 Ma (Gupta et al., 2004) and reached levels associated with present Indian monsoon conditions around 2.3 Ma. This late Miocene to Pliocene “biogenic bloom” (Farrell et al., 1995) implies important changes in nutrient cycling in the Indian Ocean and probably on a global scale, which in particular affected the silica and phosphorus cycles (Dickens and Owen, 1999). The productivity increase be- tween 10 and 8 Ma in the eastern equatorial Indian Ocean (onset of the biogenic bloom) may have been linked either to global cooling and the expansion of Antarctic ice sheets leading to a major change in deep ocean circulation and nutrient cycling, to the initiation of the Indian monsoons, or to a combination of both (Gupta et al., 2004). Figure F5. Ten-day National Oceanic and Atmospheric Administration HYSPLIT backtracks (2007–2011). Green indicates rain-bearing air trajecto- ries. Rain-bearing trajectories over India are dominated by summer season (southwest) winds, whereas both winter (northerly) and summer (southerly) air masses carry precipitation to the monsoonal region of eastern China. This is consistent with World Meteorological Organization rainfall records indi- cating that June, July, and August rainfall only accounts for 50% of the total annual rainfall in eastern China. The Bay of Bengal location isolates summer season dynamics. Figure F5. Ten-day National Oceanic and Atmospheric Administration HYSPLIT backtracks (2007–2011). Green indicates rain-bearing air trajecto- ries. Rain-bearing trajectories over India are dominated by summer season (southwest) winds, whereas both winter (northerly) and summer (southerly) air masses carry precipitation to the monsoonal region of eastern China. IODP Proceedings Geological setting Terrigenous sediment supply to the An- daman Sea originates dominantly from the Irrawaddy and Salween Rivers (Colin et al., 1998; Ali et al., 2015). Analysis of Andaman Sea surface sediments indicates that foraminifers are abundant and well preserved shallower than ~1800 mbsl (>100,000 individuals/gram) and decrease to <100 individuals/gram deeper than 3000 mbsl (Frerichs, 1971). Modern atmospheric and oceanographic circulation The Indian summer monsoon is characterized by low atmo- spheric pressure over the Indo-Asian continent (Indo-Asian Low) relative to high atmospheric pressure over the southern subtropical Indian Ocean (Mascarene High). The resulting pressure gradient Krishna Godavari Mahanadi Ganges Damodar Irrawaddi Salween Brahmani-Baitarani Subernarekha Penner Kaveri Brahmaputra Site U1446 Site U1447 Site U1448 Site U1444 Site U1445 Site U1443 30° N 25° 20° 15° 10° 5° 65°E 70° 75° 80° 85° 90° 95° 100° 105° Figure F7. Summary sketch after Singh et al. (2013) showing the subduction zone through the back-arc basin to ~40 km depth. EMF = Eastern Margin fault, DF = Diligent fault, NAB = Nicobar Andaman Basin, ANF = Andaman-Nicobar fault. Krishna Godavari Mahanadi Ganges Damodar Irrawaddi Salween Brahmani-Baitarani Subernarekha Penner Kaveri Brahmaputra Site U1446 Site U1447 Site U1448 Site U1444 Site U1445 Site U1443 30° N 25° 20° 15° 10° 5° 65°E 70° 75° 80° 85° 90° 95° 100° 105° Deep time Benthic foraminiferal diversity and assemblage composition, in conjunction with geochemical proxies of carbon flux, indicate a stepwise increase in primary production and carbon flux and an ex- pansion of the intermediate water OMZ in the northeastern Indian Ocean since the late Oligocene (Gupta et al., 2013). The main in- 5 IODP Proceedings Volume 353 S.C. Clemens et al. Expedition 353 summary tinuation in Prydz Bay, Antarctica, known as Lambert Graben (Federov et al., 1982). isotopes may have been related to a shift in the source area of the Indonesian Throughflow toward the North Pacific around 3.5 Ma (Cane and Molnar, 2001; Gourlan et al., 2008). The Ninetyeast Ridge (NER) is an aseismic volcanic ridge span- ning ~31°S to ~10°N, where it is buried beneath Bengal Fan sedi- ments. The NER is thought to have formed by age-progressive hotspot volcanism from plume sources currently beneath the Ker- guelen Plateau (Royer et al., 1991; Sager et al., 2010). The ridge top rises to a height of ~3.5 km above the surrounding abyssal plain with depths as shallow as ~2000 meters below sea level (mbsl). Site U1443 is located at ~5°N at 2925 mbsl. This location provides good preservation of carbonate microfossils given that the foraminiferal lysocline in this region is close to 3300 mbsl (Cullen and Prell, 1984). Geological setting The eastern continental margin of India is the result of the sepa- ration of India and the Australia/Antarctica portion of Gondwana- land during the Early Cretaceous at ~130 Ma (Powell et al., 1988; Scotese et al., 1988). The 2000 m isobaths of the northeast Indian continental margin and the East Antarctica margin (Prydz Bay) are closely matched, supporting the inferred alignment of India and Antarctica prior to rifting (Subrahmanyam et al., 2008). The major Indian peninsular rivers draining into the Bay of Bengal (Figure F6) are thought to be associated with graben features resulting from the rifting of India from Antarctica as well as subsequent Indian plate motion. The Mahanadi Graben, for example, appears to have a con- The Andaman Sea is situated between the Andaman Islands and the Malaya Peninsula (Figure F2). The Andaman-Sumatra island arc system results from oblique subduction of the Indo-Australian plate beneath the Eurasian plate (Singh et al., 2013). Stretching and rifting of the overriding plate in the early Miocene (~25 Ma) has resulted in two distinct plates (Sunda and Burma) separated by an active spreading center (Curray, 1991) located in the deepest portion of the Andaman Sea. An accretionary wedge complex, scraped off the subducting slab, lies west of the spreading center, forming a series of shallower basins associated with backthrust faulting within the ac- creted sediments (Figure F7). The Andaman Sea drilling sites are within the western part of the Andaman forearc basin (west of the Nicobar-Andaman Basin), bounded on either side by the Diligent and Eastern Margin faults. Terrigenous sediment supply to the An- daman Sea originates dominantly from the Irrawaddy and Salween Rivers (Colin et al., 1998; Ali et al., 2015). Analysis of Andaman Sea surface sediments indicates that foraminifers are abundant and well preserved shallower than ~1800 mbsl (>100,000 individuals/gram) and decrease to <100 individuals/gram deeper than 3000 mbsl (Frerichs, 1971). Figure F6. Main river basins draining into the Bay of Bengal and Andaman Sea. Map was generated using GeoMapApp (http://www.geoma- papp.org), using topography and bathymetry from the Global Multi-Resolu- tion Topography synthesis (Ryan et al., 2009). shallower basins associated with backthrust faulting within the ac- creted sediments (Figure F7). The Andaman Sea drilling sites are within the western part of the Andaman forearc basin (west of the Nicobar-Andaman Basin), bounded on either side by the Diligent and Eastern Margin faults. Modern atmospheric and oceanographic circulation The Indian summer monsoon is characterized by low atmo- spheric pressure over the Indo-Asian continent (Indo-Asian Low) relative to high atmospheric pressure over the southern subtropical Indian Ocean (Mascarene High). The resulting pressure gradient Figure F7. Summary sketch after Singh et al. (2013) showing the subduction zone through the back-arc basin to ~40 km depth. EMF = Eastern Margin fault, DF = Diligent fault, NAB = Nicobar Andaman Basin, ANF = Andaman-Nicobar fault. IODP Proceedings 6 Volume 353 g Accretionary prism Fore-arc high Fore-arc basin Volcanic arc Spreading center Continental crust Oceanic crust Oceanic crust Oceanic mantle Continental mantle ? Oceanic mantle Accreted sediments Basin sediments 0 20 25 15 10 5 35 30 40 Depth (km) ANF Spreading center Oceanic plate 100 200 300 400 DF NAB West East U1447, U1448 EMF Accretionary prism Fore-arc high Fore-arc basin Volcanic arc Spreading center Continental crust Oceanic crust Oceanic crust Oceanic mantle Continental mantle ? Oceanic mantle Accreted sediments Basin sediments 0 20 25 15 10 5 35 30 40 Depth (km) ANF Spreading center Oceanic plate 100 200 300 400 DF NAB West East U1447, U1448 EMF Spreading center East Fore-arc high Spreading center Oceanic crust 6 IODP Proceedings IODP Proceedings Volume 353 Volume 353 S.C. Clemens et al. Expedition 353 summary leads to large-scale displacement of the Intertropical Convergence Zone and cross-equatorial flow of low-level winds carrying mois- ture that is ultimately released over South Asia, the Bay of Bengal, and southeast China (Hastenrath and Greischar, 1993; Liu et al., 1994; Loschnigg and Webster, 2000; Webster, 1987a, 1987b, 1994; Webster et al., 1998). Modern meteorological observations and moisture transport budgets (Figure F8) quantitatively show that the Southern Hemisphere Indian Ocean is the dominant source of moisture (latent heat) to the Indian and East Asian summer mon- soons during June, July, and August (JJA) (Bosilovich and Schubert, 2002; Ding and Chan, 2005; Ding et al., 2004; Emile-Geay et al., 2003; Liu and Tang, 2004, 2005; Park et al., 2007; Simmonds et al., 1999; Wajsowicz and Schopf, 2001; Xie and Arkin, 1997; Zhu and Newell, 1998). The Arabian Sea is a very minor moisture source (evaporation > precipitation), whereas the Bay of Bengal/Andaman Sea, India, the South China Sea, and southeast China are all mois- ture sinks (precipitation > evaporation) (Figure F8). Modern atmospheric and oceanographic circulation that carries high-salinity waters eastward around the tip of India and Sri Lanka into the southern Bay of Bengal at a rate of 8.4 Sver- drup (Sv = 106 m3/s). This influx of high-salinity water is reflected in the southern Bay of Bengal July, August, and September salinity pat- terns (Figure F4) and is successfully modeled as a passive tracer in mixed-layer ocean models (Jensen, 2001, 2003). Southwest summer monsoon winds in the Bay of Bengal also drive the northward-flow- ing East Indian Coastal Current (EICC). During the winter mon- soon, northeast winds drive all these surface currents in the opposite directions, transporting 11 Sv of water toward the eastern Arabian Sea. Salinity on the Indian margin, northwest Bay of Bengal, reaches a minimum of ~22 in September (Figure F4); this is a lagged re- sponse to JJA rainfall over the Bay of Bengal and the surrounding drainage basins as well as seasonal surface current patterns. Salinity at this location reaches a maximum of 34 during the spring months. The Andaman Sea sites, situated between the modern 32 and 33 an- nual average isohalines, monitor drainage from the Irrawaddy and Salween Rivers. Salinity in the southern Bay of Bengal (Site U1443) is closely pinned to the 34 isohaline year round, anchoring the southern end of the modern salinity gradient at near open-ocean values. Although this site does not currently experience significant seasonal salinity variability, it does record large-scale changes in precipitation and runoff at the millennial, orbital, glacial–intergla- cial, and tectonic scales as discussed below. The full meridional transect (spanning the Indian margin, Andaman Sea, and northern NER) has a modern salinity range of 12. g A total of 12 major rivers feed the Bay of Bengal/Andaman Sea (Ganges, Brahmaputra, Meghna, Damodar, Mahanadi, Godavari, Krishna, Irrawaddy, Salween, Penner, Kavery, and Mahaweli Rivers) (Figure F6), discharging in total 943 × 109 m3 of water during the summer monsoon months (JJA) (Varkey et al., 1996), accounting for ~50% of the total freshwater input, the remaining being from direct precipitation (Akhil et al., 2014). Annual rainfall within and sur- rounding the Bay of Bengal is dominated by precipitation during JJA with the exception of the Chennai Basin in southernmost peninsu- lar India, where rainfall peaks in November (Figure F9). Modern atmospheric and oceanographic circulation Cherrapunji precipitation is scaled separately. Data from World Meteorological Organization (http://cli- mexp.knmi.nl). Figure F10. World Ocean Atlas Monthly Mean Temperature (°C) for 1955– 2006 (Locarnini et al., 2010). Summer and winter monsoon current systems follow Schott and McCreary (2001). Site locations in the Andaman Sea (Sites U1447 and U1448), Mahanadi basin, Bay of Bengal (Sites U1445 and U1446), and Site U1443 (redrill ODP Site 758) are shown. WMC = Winter Monsoon Current. Two water masses occupy the thermocline of the Indian Ocean, Indian Central Water (ICW) and Indonesian Throughflow Water (ITW) or Australasian Mediterranean Water (Figure F14). ICW originates from downwelling in the subtropical convergence south of 30°S, and ITW is derived from North Pacific Intermediate Water, strongly modified during its passage through the Indonesian archi- pelago. There is no formation of thermocline water in the Bay of follow Schott and McCreary (2001). Site locations in the Andaman Sea (Sites U1447 and U1448), Mahanadi basin, Bay of Bengal (Sites U1445 and U1446), and Site U1443 (redrill ODP Site 758) are shown. WMC = Winter Monsoon Current. 70°E 75° 80° 85° 90° 95° 100° 25° N 20° 15° 10° 5° 0° 70°E 75° 80° 85° 90° 95° 100° 25° N 20° 15° 10° 5° 0° Jan U1447/U1448 U1446/U1446 25 U1443 WMC WICC EICC Jul SMC WICC EICC U1446/U1446 U1443 U1447/U1448 ing, and transport at timescales from millennial to tectonic (Clift and Plumb, 2008; Wang et al., 2005). Changes in monsoon strength are well documented at ~23, 15, 8–7, and 2.75 Ma (Clift and Plumb, 2008). The Expedition 353 drilling targets will allow measurement of the consequent impact on weathering rates and transport of par- ticulate materials to the ocean basins in a variety of settings both proximal and distal relative to river inputs. mexp.knmi.nl). 0 20 40 60 80 100 0 50 100 150 200 250 300 Purushottampur Chennai Kolkata Dhaka Yangon Bhubaneswar Berhampur Chittagong Salem Bhadrak Port Blair Cherrapunji Precipitation (cm/month) Cherrapunji precipitation (cm/month) Month 2 4 6 8 10 12 0 20 40 60 80 100 0 50 100 150 200 250 300 Precipitation (cm/month) Cherrapunji precipitation (cm/month) Month 2 4 6 8 10 12 70°E 75° 80° 85° 90° 95° 100° 25° N 20° 15° 10° 5° 0° Jul SMC WICC EICC U1446/U1446 U1443 U1447/U1448 Cherrapunji precipitation (cm/mont Precipitation (cm/month) ing, and transport at timescales from millennial to tectonic (Clift and Plumb, 2008; Wang et al., 2005). Modern atmospheric and oceanographic circulation Changes in monsoon strength are well documented at ~23, 15, 8–7, and 2.75 Ma (Clift and Plumb, 2008). The Expedition 353 drilling targets will allow measurement of the consequent impact on weathering rates and transport of par- ticulate materials to the ocean basins in a variety of settings both proximal and distal relative to river inputs. Two water masses occupy the thermocline of the Indian Ocean, Indian Central Water (ICW) and Indonesian Throughflow Water (ITW) or Australasian Mediterranean Water (Figure F14). ICW originates from downwelling in the subtropical convergence south of 30°S, and ITW is derived from North Pacific Intermediate Water, strongly modified during its passage through the Indonesian archi- pelago. There is no formation of thermocline water in the Bay of 70°E 75° 80° 85° 90° 95° 100° 25° N 20° 15° 10° 5° 0° Jan U1447/U1448 U1446/U1446 25 U1443 WMC WICC EICC 70°E 75° 80° 85° 90° 95° 100° 25° N 20° 15° 10° 5° 0° Jan U1447/U1448 U1446/U1446 25 U1443 WMC WICC EICC Modern atmospheric and oceanographic circulation The domi- nance of the JJA precipitation signal is reflected in the Bay of Bengal surface salinity patterns (Figure F4), spanning salinities of 20–34 over both seasonal (summer–winter) and spatial (north–south) scales, with lowest values in August and September. Estimates for expected δ18OG.ruber values can be calculated by combining climatologies for monthly temperature and salinity with regional δ18Oseawater-salinity equations (Figure F11). Northern (20°N) δ18OG.ruber estimates range between −1.9‰ for January and −3.3‰ for October with strong sensitivity to salinity seasonality. Es- timates for 5°N range from −2.5‰ (August–December) to −2.9‰ (April) with strong sensitivity to temperature seasonality. Primary surface ocean currents (Schott and McCreary, 2001; Schott et al., 2009) reflect the seasonal wind forcing in both the eastern Arabian Sea and the Bay of Bengal (Figure F10). The West Indian Coastal Current (WICC) flows south during the summer monsoon, connecting with the Southwest Monsoon Current (SMC) Terrestrial runoff products are also of significant utility in as- sessing linkages between monsoon circulation, chemical weather- Figure F8. Summer monsoon moisture budget and transport path (inset) after Ding et al. (2004). Moisture budget averaged for 1990–1999 (June, July, and August; units are 106 kg/s). The southern Indian Ocean is the dominant moisture (latent heat) source. The Bay of Bengal, Indochina, the South China Sea, and China are all moisture sinks. No significant Pacific moisture source is indicated. Summer monsoon moisture transport patterns (inset) averaged for 1990–1999 (the 5th pentad of May through the 2nd pentad of July; units are kg/m/s). IODP Proceedings 7 Volume 353 40 60 80 100 120 0 20 40 268.5 69.2 6.8 30.8 40.6 259.7 73 46.4 186.3 77.8 58.6 159.4 72.1 148.3 34.1 595.5 357.7 316.8 625.8 88.7 68.5 115.2 157.5 100.3 (18) (-28) (-45.4) (-114.5) (4) (-41.9) (-295) 240.1 0.3 8.6 40 60 80 100 120 0 20 40 268.5 69.2 6.8 30.8 40.6 259.7 73 46.4 186.3 77.8 58.6 159.4 72.1 148.3 34.1 595.5 357.7 316.8 625.8 88.7 68.5 115.2 157.5 100.3 (18) (-28) (-45.4) (-114.5) (4) (-41.9) (-295) 240.1 0.3 8.6 Volume 353 Volume 353 7 IODP Proceedings S.C. Clemens et al. Expedition 353 summary Figure F9. Monthly precipitation data for 1948–2011 (as available) for loca- tions with drainage into the Bay of Bengal. Cherrapunji precipitation is scaled separately. Data from World Meteorological Organization (http://cli- mexp.knmi.nl). Figure F9. Monthly precipitation data for 1948–2011 (as available) for loca- tions with drainage into the Bay of Bengal. IODP Proceedings Water masses and deep circulation Expedition 353 summary Figure F11. Estimated monthly G. ruber δ18O derived from monthly salinity and temperature climatologies at (A) 5°N, 90°E and (B) 20°N, 86°E, Bay of Bengal. Monthly mean ±1σ over 1958–2013. Sea-surface temperature (SST) from HadISST data set (Rayner et al., 2003). Sea-surface salinity (SSS) from ORA-S4 data set (Balmaseda et al., 2012). Pseudo-δ18O of G. ruber (white) constructed from SST and SSS data set using low-light equation of Bemis et al. (1998) and open-ocean Bay of Bengal δ18Oseawater-salinity relationship from Singh et al. (2010). S.C. Clemens et al. Expedition 353 summary Figure F11. Estimated monthly G. ruber δ18O derived from monthly salinity and temperature climatologies at (A) 5°N, 90°E and (B) 20°N, 86°E, Bay of Bengal. Monthly mean ±1σ over 1958–2013. Sea-surface temperature (SST) from HadISST data set (Rayner et al., 2003). Sea-surface salinity (SSS) from ORA-S4 data set (Balmaseda et al., 2012). Pseudo-δ18O of G. ruber (white) constructed from SST and SSS data set using low-light equation of Bemis et al. (1998) and open-ocean Bay of Bengal δ18Oseawater-salinity relationship from Singh et al. (2010). Figure F11. Estimated monthly G. ruber δ18O derived from monthly salinity and temperature climatologies at (A) 5°N, 90°E and (B) 20°N, 86°E, Bay of Bengal. Monthly mean ±1σ over 1958–2013. Sea-surface temperature (SST) from HadISST data set (Rayner et al., 2003). Sea-surface salinity (SSS) from ORA-S4 data set (Balmaseda et al., 2012). Pseudo-δ18O of G. ruber (white) constructed from SST and SSS data set using low-light equation of Bemis et al. (1998) and open-ocean Bay of Bengal δ18Oseawater-salinity relationship from Singh et al. (2010). Figure F11. Estimated monthly G. ruber δ18O derived from monthly salinity and temperature climatologies at (A) 5°N, 90°E and (B) 20°N, 86°E, Bay of Bengal. Monthly mean ±1σ over 1958–2013. Sea-surface temperature (SST) from HadISST data set (Rayner et al., 2003). Sea-surface salinity (SSS) from ORA-S4 data set (Balmaseda et al., 2012). Pseudo-δ18O of G. ruber (white) constructed from SST and SSS data set using low-light equation of Bemis et al. (1998) and open-ocean Bay of Bengal δ18Oseawater-salinity relationship from Singh et al. (2010). its thermocline water masses to 1500 m water depth are m ICW and ITW. Transfer of ICW to the northern In- is accompanied by a rapid decrease in oxygen content, ging along the path. The lowest oxygen values occur in engal, which contains the oldest ICW. IODP Proceedings Water masses and deep circulation Comprehensive descriptions of eastern Indian Ocean regional oceanography are provided in Wyrtki (1971), Mantyla and Reid (1995), Tomczak and Godfrey (2003), and Schott et al. (2009), from which we briefly summarize descriptions of water masses and circu- lation patterns relevant for Bay of Bengal drilling in the depth range between 1100 and 3000 m targeted for Expedition 353. Indian Deep Water (IDW) occupies the depth range from 3800 to ~1500 m within the equatorial and northern Indian Ocean (Fig- ure F12). IDW in the eastern Indian Ocean is characterized by high salinities, reaching maxima of 34.8 in the southwestern Indian Ocean and 34.75 in the southeastern Indian Ocean, where the IDW upper limit rises to 500 m (Tomczak and Godfrey, 2003). IDW tem- perature, salinity, and oxygen properties in the high-salinity core are virtually identical with those of North Atlantic Deep Water (NADW) in the Atlantic sector of the Southern Ocean, indicating that IDW is mainly of NADW origin and not originally formed in the Southern Ocean, as is the Antarctic Bottom Water (AABW) that occupies the Indian Ocean deeper than 3800 m (Tomczak and Godfrey, 2003). The flow of IDW is northward and concentrated along the African margin and the NER as indicated by the World Ocean Circulation Experiment (WOCE) 95°E oxygen, silicate, and temperature profiles (Figure F13). IDW further penetrates north- ward into the Northern Hemisphere, is modified by mixing with thermocline water from above and upwelling of AABW from below, and spreads into the Arabian Sea and the Bay of Bengal. Two water masses occupy the thermocline of the Indian Ocean, Indian Central Water (ICW) and Indonesian Throughflow Water (ITW) or Australasian Mediterranean Water (Figure F14). ICW originates from downwelling in the subtropical convergence south of 30°S, and ITW is derived from North Pacific Intermediate Water, strongly modified during its passage through the Indonesian archi- pelago. There is no formation of thermocline water in the Bay of Two water masses occupy the thermocline of the Indian Ocean, Indian Central Water (ICW) and Indonesian Throughflow Water (ITW) or Australasian Mediterranean Water (Figure F14). ICW originates from downwelling in the subtropical convergence south of 30°S, and ITW is derived from North Pacific Intermediate Water, strongly modified during its passage through the Indonesian archi- pelago. There is no formation of thermocline water in the Bay of 8 Volume 353 IODP Proceedings Volume 353 S.C. Clemens et al. Site U1443 Ocean data view Water masses and deep circulation As a result of northward movement, Site U1443 moved from tem- 20° N 10° 0° 10° 20° 30° 40° 30°E 40° 70° 80° 90° 100° 110° 120° 130° 50° 60° Figure F13. Salinity, silicate, and oxygen from eastern Indian Ocean WOCE meridional Transect I09, eastern Indian Ocean (from eWOCE Gallery, http://www.ewoce.org/gallery [Schlitzer, 2000]). Salinity (psu) Silicate (µmol/kg) Oxygen (µmol/kg) 0 1000 4000 2000 3000 5000 Depth (m) 0 1000 4000 2000 3000 5000 Depth (m) 0 1000 4000 2000 3000 5000 Depth (m) 20°S EQ 20°N Ocean data view Ocean data view Ocean data view 36 35.5 34 35 34.5 125 100 25 75 50 0 250 200 50 150 100 0 350 300 (black) and Indonesian Throughflow Water (red). From Xu et al. (2006). A B 20° N 10° 0° 10° 20° 30° 40° 30°E 40° 70° 80° 90° 100° 110° 120° 130° 50° 60° into the Bay of Bengal. Further freshening is observed in the Bay of Bengal near 90°E, resulting from ITW advection directly from its outflow area into the tropical eastern Indian Ocean. Variability and evolution of thermocline circulation in the Bay of Bengal are strongly dependent on monsoonal forcing; however, the present ex- tremely low oxygen levels indicate a very low renewal rate for the thermocline waters of the Bay of Bengal. The uppermost 100 m of the eastern Indian Ocean in the Bay of Bengal consists of a low-salinity water mass derived from river run- off from India and Indochina, the Bay of Bengal Water (BBW), with surface salinity strongly fluctuating with seasons but remaining be- low 33 throughout the year. The lower boundary to the ICW is characterized by a strong halocline. The southward extension of the BBW surface water mass is highest during October–December when it reaches the area along the western Indian coast and is low- est during April–June before the summer monsoon leads to a new expansion of the BBW. Site summaries Site U1443 B k d d bj ti A B A B A B B Figure F13. Salinity, silicate, and oxygen from eastern Indian Ocean WOCE meridional Transect I09, eastern Indian Ocean (from eWOCE Gallery, http://www.ewoce.org/gallery [Schlitzer, 2000]). Water masses and deep circulation The strong oxy- gen decrease in the northern Indian Ocean can be explained duction of the transfer of ICW during the southwest m season, resulting in a small annual net transfer rate. ITW al tributes to the renewal of thermocline water in the northern Ocean, resulting in significant freshening of the ICW along i t al., 2012). Pseudo δ O of G. ruber (white) constructed from SST and SSS data set using low light equation of Bemis et al. (1998) and ope δ18Oseawater-salinity relationship from Singh et al. (2010). Site U1443 (5°N, 90°E) SST Site U1443 (5°N, 90°E) Pseudo-δ18O G. ruber Site U1443 (5°N, 90°E) SSS A B Site U1446 (20°N, 86°E) SST Site U1446 (20°N, 86°E) SSS Site U1446 (20°N, 86°E) Pseudo-δ18O G. ruber Month of the year Month of the year Month of the year Month of the year Month of the year Month of the year δ18Opseudo (‰) δ18Opseudo (‰) Temperature (°C) Salinity (PSU) Temperature (°C) Salinity (PSU) Site U1443 (5°N, 90°E) SST A Month of the year Temperature (°C) Site U1443 (5°N, 90°E) Pseudo-δ18O G. ruber Month of the year δ18Opseudo (‰) Site U1443 (5°N, 90°E) SST Site U1443 (5°N, 90°E) SSS A Month of the year Month of the year Temperature (°C) Salinity (PSU) A Month of the year δ18Opseudo (‰) Month of the year B Site U1446 (20°N, 86°E) SST Month of the year Temperature (°C) B Site U1446 (20°N, 86°E) SST Site U1446 (20°N, 86°E) SSS Month of the year Month of the year Temperature (°C) Salinity (PSU) Site U1446 (20°N, 86°E) Pseudo-δ18O G. ruber Month of the year δ18Opseudo (‰) B Site U1446 (20°N, 86°E) Pseudo-δ18O G Bengal, and its thermocline water masses to 1500 m water depth are derived from ICW and ITW. Transfer of ICW to the northern In- dian Ocean is accompanied by a rapid decrease in oxygen content, indicating aging along the path. The lowest oxygen values occur in the Bay of Bengal, which contains the oldest ICW. The strong oxy- gen decrease in the northern Indian Ocean can be explained by re- duction of the transfer of ICW during the southwest monsoon season, resulting in a small annual net transfer rate. ITW also con- tributes to the renewal of thermocline water in the northern Indian Ocean, resulting in significant freshening of the ICW along its path IODP Proceedings Volume 353 Volume 353 9 Volume 353 S.C. Water masses and deep circulation Salinity (psu) Silicate (µmol/kg) Oxygen (µmol/kg) 0 1000 4000 2000 3000 5000 Depth (m) 0 1000 4000 2000 3000 5000 Depth (m) 0 1000 4000 2000 3000 5000 Depth (m) 20°S EQ 20°N Ocean data view Ocean data view Ocean data view 36 35.5 34 35 34.5 125 100 25 75 50 0 250 200 50 150 100 0 350 300 Salinity (psu) 0 1000 4000 2000 3000 5000 Depth (m) Ocean data view Ocean data view into the Bay of Bengal. Further freshening is observed in the Bay of Bengal near 90°E, resulting from ITW advection directly from its outflow area into the tropical eastern Indian Ocean. Variability and evolution of thermocline circulation in the Bay of Bengal are strongly dependent on monsoonal forcing; however, the present ex- tremely low oxygen levels indicate a very low renewal rate for the thermocline waters of the Bay of Bengal. into the Bay of Bengal. Further freshening is observed in the Bay of Bengal near 90°E, resulting from ITW advection directly from its outflow area into the tropical eastern Indian Ocean. Variability and evolution of thermocline circulation in the Bay of Bengal are strongly dependent on monsoonal forcing; however, the present ex- tremely low oxygen levels indicate a very low renewal rate for the thermocline waters of the Bay of Bengal. Silicate (µmol/kg) 0 1000 4000 2000 3000 5000 Depth (m) Ocean data view The uppermost 100 m of the eastern Indian Ocean in the Bay of Bengal consists of a low-salinity water mass derived from river run- off from India and Indochina, the Bay of Bengal Water (BBW), with surface salinity strongly fluctuating with seasons but remaining be- low 33 throughout the year. The lower boundary to the ICW is characterized by a strong halocline. The southward extension of the BBW surface water mass is highest during October–December when it reaches the area along the western Indian coast and is low- est during April–June before the summer monsoon leads to a new expansion of the BBW. Ocean data view Oxygen (µmol/kg) 0 1000 4000 2000 3000 5000 Depth (m) 20°S EQ 20°N Ocean data view Depth (m) Water masses and deep circulation Clemens et al. Expedition 353 summary Figure F12. Indian Ocean deep and bottom water circulation (from Frank et al., 2006). Blue arrows represent the deep and bottom water flow patterns in the Indian Ocean. After Mantyla and Reid (1995). Figure F14. Mean (A) temperature and (B) salinity in the upper thermocline on isopycnal Surface 25.7, located in the depth range 150–200 m (after You and Tomczak, 1993). Arrows indicate movement of Indian Central Water (black) and Indonesian Throughflow Water (red). From Xu et al. (2006). Figure F14. Mean (A) temperature and (B) salinity in the upper thermocline on isopycnal Surface 25.7, located in the depth range 150–200 m (after You and Tomczak, 1993). Arrows indicate movement of Indian Central Water (black) and Indonesian Throughflow Water (red). From Xu et al. (2006). 20° N 10° 0° 10° 20° 30° 40° 30°E 40° 70° 80° 90° 100° 110° 120° 130° 50° 60° Figure F13. Salinity, silicate, and oxygen from eastern Indian Ocean WOCE meridional Transect I09, eastern Indian Ocean (from eWOCE Gallery, http://www.ewoce.org/gallery [Schlitzer, 2000]). into the Bay of Bengal. Further freshening is observed in the Bay of Bengal near 90°E, resulting from ITW advection directly from its outflow area into the tropical eastern Indian Ocean. Variability and evolution of thermocline circulation in the Bay of Bengal are strongly dependent on monsoonal forcing; however, the present ex- tremely low oxygen levels indicate a very low renewal rate for the thermocline waters of the Bay of Bengal. The uppermost 100 m of the eastern Indian Ocean in the Bay of Bengal consists of a low-salinity water mass derived from river run- off from India and Indochina, the Bay of Bengal Water (BBW), with surface salinity strongly fluctuating with seasons but remaining be- low 33 throughout the year. The lower boundary to the ICW is characterized by a strong halocline. The southward extension of the BBW surface water mass is highest during October–December when it reaches the area along the western Indian coast and is low- est during April–June before the summer monsoon leads to a new expansion of the BBW. Site summaries Site U1443 Background and objectives The NER represents the trace of the Kerguelen/Ninetyeast hotspot prior to middle Eocene rifting (Peirce, Weissel, et al., 1989). Background and objectives Re- cent research using Site 758 sediments include high-resolution (suborbital) Late Pleistocene reconstruction of changes in upper water column structure based on multispecies planktonic foramin- ifer records (Bolton et al., 2013), reconstruction of the Li isotope composition of seawater over the past 70 My (Misra and Froelich, 2012), Late Cretaceous to early Eocene reconstruction of seawater neodymium (Le Houedec et al., 2012), and glacial–interglacial scale reconstruction of Os isotopic composition of seawater (Burton et al., 2010). Drilling Site U1443 had the following main objectives: • Extend the Pliocene–Pleistocene stable isotope record for Site 758, which is the only “high-resolution” record across the initia- tion of Northern Hemisphere glaciation in the Indian Ocean (Hoogakker et al., 2006; Mudelsee and Raymo, 2005), into the Miocene and Oligocene. • Use the Nd isotope composition of NER sediments to extend re- constructions of the relative contribution of suspended sedi- ment from the Ganges-Brahmaputra, Irrawaddy, Arakan, and Indian peninsular rivers as indicators of glacial–interglacial vari- ability in monsoon strength into the Miocene (Burton and Vance, 2000; Stoll et al., 2007; Gourlan et al., 2008, 2010). A complete Neogene sediment archive at Site U1443 opens the possibility to obtaining proxy records for river discharge in suffi- • Establish a composite section for the entire Miocene–Oligocene sedimentary succession at the NER, which would be the basis for establishing the first orbitally tuned Indian Ocean isotope strati- graphy for this time interval. Table T1. Expedition 353 hole summary. Download table in .csv format. Background and objectives The NER represents the trace of the Kerguelen/Ninetyeast hotspot prior to middle Eocene rifting (Peirce, Weissel, et al., 1989). As a result of northward movement, Site U1443 moved from tem- EQ 10 IODP Proceedings Volume 353 Volume 353 Volume 353 S.C. Clemens et al. Expedition 353 summary • Recover a more complete record of the Cretaceous/Paleogene boundary interval, which was only incompletely retrieved at the base of the core catcher of Core 121-758A-31X. perate southern latitudes during the Campanian to ~5°S near the Oligocene/Miocene boundary and to its present 5°N location in the southernmost Bay of Bengal (Figure F2). The site has been within 10° of the Equator for the past 35 My (Shipboard Scientific Party, 1989). The ridge-top location (2925 mbsl) has prevented the deposi- tion of sedimentary sequences typically associated with fan trans- port processes and is thus a useful location for recovery of open- ocean pelagic sediments. • Correlate the onset of increased terrigenous clay component and sedimentation rates in the late Miocene to orbitally tuned isotope and magnetic reversal stratigraphy. • Precisely determine the timing of intensifications of the Indian monsoon, as evident from increased freshwater input to the Bay of Bengal and northern end of the NER using salinity proxies based on Mg/Ca-temperature estimates and δ18O of surface- dwelling planktonic foraminifers. g Site U1443 is located ~100 m southeast of ODP Site 758 on the crest of the NER (Figure F2; Table T1). Previous drilling at this site included overlapping holes (758A and 758B) only for the uppermost 92 m of the total 527 m of sediments drilled. This limited the result- ing composite record to the past 7.3 My (Farrell and Janecek, 1991). • Investigate variability and possible influence of orbital forcing on fluxes of terrigenous material to the northern NER since the middle Miocene. In particular, relate terrigenous pulses at ~7.0– 5.6 and ~3.9–2.0 Ma, which were interpreted to represent varia- tions in the fluvial flux resulting from the uplift and erosion of the Himalaya (Hovan and Rea, 1992), to the variability of the In- dian monsoon. In spite of these limitations, the IODP sample database shows that more than 16,000 sediment samples have been taken from Site 758 cores in the last 10 y. Parts of the uppermost sections are almost entirely depleted, and sampling has moved to the archive half. Lithostratigraphy Sediments recovered from Site U1443 reveal a range of pelagic and hemipelagic sediments of Late Pleistocene to Campanian age, comprising four distinct lithostratigraphic units (I–IV): The age-depth relationship for Hole U1443A is based on the biostratigraphy of the three fossil groups studied (calcareous nan- nofossils, planktonic foraminifers, and diatoms) and paleomagnetic stratigraphy. Age datums of the fossil groups and paleomagnetism show good agreement for the Pleistocene. Calcareous nannofossils and planktonic foraminifers show consistent age-depth relation- ships throughout the Cenozoic and Late Cretaceous, with no major outliers. The combined biostratigraphic and paleomagnetic age model indicates a mean sedimentation rate of 1.20 cm/ky in the up- per part of lithostratigraphic Unit I (0–80 m CSF-A, Pleistocene to late Miocene). The mean sedimentation rate decreases to around 0.41 cm/ky between 100 and 130 m CSF-A (the upper part of litho- stratigraphic Subunit IIa, late Miocene to middle Miocene). Sedi- mentation rates between 135 and 200 m CSF-A (the lower half of lithostratigraphic Subunit IIa) average 0.81 cm/ky (early Miocene to Oligocene). Finally, following a hiatus that spans the late Oligocene to late Paleocene, the mean sedimentation rate in the late Paleocene and Late Cretaceous (lithostratigraphic Units III and IV) is 0.36 cm/ky. These sedimentation rate and age estimates broadly agree with those published for Site 758 (Shipboard Scientific Party, 1989). • Unit I (0–107.80 m core depth below seafloor, Method A [CSF‑A]) is composed of Late Pleistocene–late Miocene light– dark gray nannofossil oozes with varying proportions of fora- minifers, clay, and volcanic ash. y • Unit II (107.80–242.36 m CSF-A) is composed of pale yellow to white nannofossil ooze and chalk with variable contents of au- thigenic carbonate and foraminifers that are late Miocene to early Oligocene in age. • Unit III (242.36–308.68 m CSF-A) is late Paleocene to late Cam- panian in age and comprises a 66 m thick package of pale yellow and brown nannofossil chalks with varying proportions of authi- genic carbonate and occasional chert and porcellanite as nod- ules and thin beds. • In Unit IV, a total of 31 m (308.68–341.35 m CSF-A), compris- ing a succession of greenish gray marlstones with glauconite of late Campanian age, was recovered before Hole U1443A was terminated at 344 m DSF. Lithostratigraphic units are defined by changes in lithology (as identified by visual core description and smear slide observations), physical properties, and color reflectance (L, a, and b). Background and objectives Site Hole Latitude Longitude Seafloor depth (mbrf) Cores (N) Cored (m) Recovered (m) Recovery (%) Drilled (m) Total penetration (m) Total depth (mbrf) Time on hole (h) Time on site (days) U1443 A 5°23.0098′N 90°21.7100′E 2940.2 48 344.00 326.80 95 0.0 344.0 3284.2 77.50 U1443 B 5°23.0190′N 90°21.7091′E 2935.5 40 326.40 308.51 95 0.0 326.4 3261.9 44.75 U1443 C 5°23.0078′N 90°21.6984′E 2935.4 28 207.90 182.87 88 1.5 209.4 3144.8 38.00 U1443 D 5°22.9991′N 90°21.6992′E 2935.3 2 8.20 7.48 91 0.0 8.2 2943.5 12.25 Site U1443 totals: 118 886.50 825.66 93 1.5 888.0 7.12 U1444 A 14°00.0057′N 84°49.7405′E 3143.4 37 330.60 226.05 68 0.0 330.6 3474.0 70.75 U1444 B 13°59.9940′N 84°49.7412′E 3142.5 9 81.10 74.16 91 47.5 128.6 3271.1 24.75 Site U1444 totals: 46 411.70 300.21 73 47.5 459.2 4.00 U1445 A 17°44.7217′N 84°47.2518′E 2513.1 77 672.60 666.40 99 0.0 672.6 3185.7 155.50 U1445 B 17°44.7098′N 84°47.2498′E 2514.5 4 33.00 33.36 101 0.0 33.0 2547.5 6.00 U1445 C 17°44.7095′N 84°47.2387′E 2513.5 36 305.20 305.60 100 0.0 305.2 2818.7 64.50 Site U1445 totals: 117 1010.80 1005.36 99 0.0 1010.8 9.42 U1446 A 19°5.0090′N 85°44.0894′E 1441.2 21 180.00 186.63 104 0.0 180.0 1621.2 28.75 U1446 B 19°5.0085′N 85°44.0786′E 1440.5 3 27.10 27.20 100 0.0 27.1 1467.6 2.50 U1446 C 19°5.0215′N 85°44.0780′E 1441.1 23 182.00 180.55 99 0.0 182.0 1623.1 24.50 Site U1446 totals: 47 389.10 394.38 101 0.0 389.1 2.32 U1447 A 10°47.4061′N 92°59.9999′E 1402.2 88 738.00 732.96 99 0.0 738.0 2140.2 122.50 U1447 B 10°47.3945′N 93°00.0028′E 1403.1 3 24.40 24.26 99 0.0 24.4 1427.5 5.25 U1447 C 10°47.3952′N 93°00.0114′E 1403.6 17 158.90 158.32 100 2.0 160.9 1564.5 19.25 Site U1447 totals: 108 921.30 915.54 99 2.0 923.3 6.13 U1448 A 10°38.0315′N 93°00.0036′E 1109.7 60 421.00 434.82 103 0.0 421.0 1530.7 47.75 U1448 B 10°38.0202′N 93°00.0032′E 1107.9 57 357.10 369.38 103 1.5 358.6 1466.5 36.00 U1448 C 10°38.0830′N 93°00.0233′E 1108.2 4 34.30 34.77 101 0.0 34.3 1142.5 9.25 Site U1448 totals: 121 812.40 838.97 103 1.5 813.9 3.90 Expedition 353 totals: 557 4431.8 4280.12 97 52.5 4484.3 789.75 32.90 11 Volume 353 Volume 353 IODP Proceedings Volume 353 S.C. Clemens et al. Expedition 353 summary cient time resolution to document the relation between global climate and the Indian monsoon throughout the Neogene. (SEM) analyses revealed that Emiliania huxleyi (<0.29 Ma) is pres- ent in intervals 353-U1443A-1H-1, 0–50 cm, and 353-U1443B-1H- 1, 0–10 cm. Cretaceous sediments contain abundant moderately to poorly preserved calcareous nannofossils. Lithostratigraphy The ob- served lithologic differences between the units are primarily the re- sult of varying abundances of nannofossils, clay, and authigenic carbonate, with glauconite influencing the color and magnetic sus- ceptibility (MS) properties in Unit IV. Lithologic descriptions are based primarily on sediments recovered from Hole U1443A, sup- plemented with observations from Holes U1443B–U1443D. Background and objectives The Cretaceous/Paleo- gene (K/Pg) boundary was identified in Core 353-U1443A-39X but is highly bioturbated and appears to be incomplete. Geochemistry The composition of the interstitial water (IW) and bulk sedi- ment samples reflects the variation in sediment composition and reactions that occurred since deposition. Overall, the sediments have high carbonate content (>80 wt% throughout most of the sec- tion) and low total organic carbon (<0.3 wt%). The depositional en- vironment changed as the site migrated from the Southern Hemisphere to the current location, as the collision of India with Asia delivered more and more terrigenous material to the location. In the upper sediments, ash deposition from the nearby Indonesian arc plays an important role. In general, the data from Site U1443 agree well with data from Site 758, but the new data provide a better resolution in the sections recovered by APC and HLAPC coring compared to the same depth intervals recovered by XCB coring at Site 758. Coring summary g y At Site U1443, Holes U1443A–U1443D were drilled to total depths of 344.0, 326.4, 209.4, and 8.2 m drilling depth below sea- floor (DSF), respectively (Table T1). In Holes U1443A and U1443B, the full-length (9.7 m long) advanced piston corer (APC), half- length advanced piston corer (HLAPC) (4.7 m long), and extended core barrel (XCB) systems were used. The XCB system was de- ployed after reaching APC refusal depth. The APC and HLAPC sys- tems were used for Hole U1443C, and only the HLAPC system was used for Hole U1443D. A total of 118 cores were recovered at Site U1443. A total of 444.06 m of core over a 471.7 m interval was re- covered using the APC system (94% recovery). The HLAPC system was used to core a 156.4 m interval and 161.25 m of core was recov- ered (103%). The cored interval with the XCB system was 258.4 m with a recovery of 220.35 m of core (85%). The overall recovery per- centage for Site U1443 was 93%. Planktonic foraminifers are dominant to abundant in Samples 353-U1443A-1H-CC through 43X-CC, which are Pleistocene to Late Cretaceous in age. Abundance decreases from common to few in Samples 44X-CC through 48X-CC. Preservation is good to mod- erate throughout the Cenozoic with a few exceptions in the late Miocene (Sample 13H-CC) and the Oligocene to Paleocene (Sam- ples 33X-CC and 34X-CC), where preservation is poor. The latest Maastrichtian is identified in Sample 39X-CC by the presence of Abathomphalus mayaroensis. A low-diversity assemblage charac- terizes sediments in older samples. Preservation in the Late Creta- ceous (Samples 39X-CC through 48X-CC) is moderate to poor. At Site U1443, diatoms are most abundant from 0 to 28 m CSF‑A; their occurrence then becomes sporadic downcore until 192 m CSF-A. The diatom assemblage in the uppermost 28 m of Site U1443 is diverse and mainly consists of Pleistocene to Holocene species, mostly typical of warm to temperate low-latitude ocean wa- ters, and includes species indicating transport of coastal waters to Site U1443. Biostratigraphy Calcareous nannofossils are abundant throughout the section in Hole U1443A, which is Late Pleistocene (<0.29 Ma) to Campanian (>72.1 Ma) in age, with a large unconformity that spans most of the Eocene and the latest Paleocene. Nannofossil assemblages are typi- cally tropical to subtropical and are well preserved in the Pleisto- cene to late Miocene sections (~0–110 m CSF-A, lithostratigraphic Unit I). Below this interval, evidence of diagenetic overgrowth of nannofossils becomes apparent. Scanning electron microscope 12 Volume 353 IODP Proceedings S.C. Clemens et al. S.C. Clemens et al. Expedition 353 summary Paleomagnetism low seafloor (CCSF-A) because several tephra layers coincided with positive MS peaks. NGR and reflectance spectroscopy b data were used throughout the rest of the records. A continuous splice was created spanning the upper 180 m, approximately tripling the length of time over which a continuous record was achieved for Site 758. This is mostly due to the use of the HLAPC, which greatly ex- tended the length of the piston-cored section. Paleomagnetic measurements were conducted on all archive- half sections and 127 discrete samples taken from Holes U1443A– U1443C. Some of the discrete samples were also subjected to rock magnetic analyses. Magnetic polarity patterns were recovered for most of the APC and HLAPC cores but not for the XCB cores. The resulting magnetostratigraphy was constructed for two time peri- ods: 0–6 Ma (Holes U1443A and U1443C) and 18–25 Ma (Hole U1443A). The age-depth model from paleomagnetic measurements agrees well with the biostratigraphic ages. Between these age inter- vals, patterns based on pass-through magnetic measurements are inconclusive because of a decrease in natural remanent magnetiza- tion (NRM) intensity. An increase in carbonate content and a de- crease in the concentration of fine-grained ferrimagnetic minerals explain the poor signal. Preliminary rock magnetic experiments suggest that changes in bulk magnetic properties are primarily con- trolled by variations in the concentration of fine-grained magne- tite/maghemite. Physical properties 0 2 4 a* 3 4 5 0 400 800 1200 1600 2000 2400 2800 3200 LR04 Age (ka) -2 2 6 b* 50 100 RGB blue 50 100 150 RGB green 50 100 150 RGB red 10 30 50 Magnetic susceptibility 24 40 56 NGR 8 40 50 60 L* 0 2 4 a* 3 4 5 0 400 800 1200 1600 2000 2400 2800 3200 LR04 Age (ka) -2 2 6 b* 50 100 RGB blue 50 100 150 RGB green 50 100 150 RGB red 10 30 50 Magnetic susceptibility 24 40 56 NGR 8 40 50 60 L* Pliocene–Pleistocene tephrochronology The Pliocene to Holocene tephra layers of the northern NER provide a unique record of explosive volcanism derived from the In- donesian island of Sumatra, the northernmost part of the Sunda arc, and the nearest volcanically active region (Dehn et al., 1991). Recov- ery of a complete succession of tephra layers within a complete splice with a high-resolution orbitally tuned isotope stratigraphy and the possibility of correlation of individual tephra layers between three holes make Site U1443 a unique observatory of volcanic activ- ity in this region over the last 5 My (Figures F16, F17). The volcanic tephra layers are excellent tools for lithostratigraphic correlation of the marine sediment successions within the four different holes on the northern NER. Detailed studies of the tephra layers will contrib- ute to the understanding of the magmatic evolution in the source region and allow determination and precise dating of eruptive cy- cles. In particular, the eruption history of the Toba Caldera in IODP Proceedings Physical properties Downhole variations in physical property measurements at Site U1443 reflect changes in lithology, condensed sections associated with depositional hiatuses, and diagenetic processes. The overall suite of physical property data from Hole U1443A, which is the deepest hole at this site, is divided into three broad physical prop- erty (PP) units (1–3) based on significant transitions within the data. The general porosity trend decreases with depth with a few ex- cursions that tend to align with anomalies in P-wave velocity. Unit 1 is characterized by high variability in MS, natural gamma radiation (NGR), P-wave velocity, and color reflectance. This unit is inter- preted to reflect numerous ash layers intercalated within clayey nannofossil ooze. MS peaks are likely the result of relative increases in iron-bearing minerals within the ash layers, and the color reflec- tance data display changes between darker ash and lighter nanno- fossil ooze as well as changes in the clay content of the nannofossil ooze. Unit 2 is characterized by a rapid transition to an overall lighter color (high reflectance), with relatively uniform low NGR and low P-wave velocity trends. The sedimentologists identify this unit to be dominantly composed of chalk. There are two anomalous intervals identified within Unit 2 that are based on an increase in density, P-wave velocity, and NGR but no corresponding change in color reflectance. Variations in physical properties within Unit 2 are visually undetectable. The physical property anomalies were used to assist in directing sampling for smear slide and SEM analysis, which together provided evidence that the anomalous intervals could be due to high authentic overgrowth and carbonate cementation cor- relating with strong, positive seismic reflections and biostratigraph- ically defined depositional hiatuses. Unit 3 is characterized by a rapid transition to darker color (low reflectance) along with higher NGR, MS, and P-wave values than in Unit 2, suggesting a transition to a stronger, more lithified material with a likely increase in mag- netic minerals and clays that are likely the result of increased terrig- enous input. All three units could also be identified in Hole U1443B. The first two units could also be identified in Hole U1443C. Figure F15. Site U1443 splice RGB, magnetic susceptibility (loop and spot), L, a, b, and NGR data tuned to LR04 benthic isotope stack for the last 3.4 My. Stratigraphic correlation northern Sumatra is well documented on the northern NER (Ninkovich, 1979; Ninkovich et al., 1978; Dehn et al., 1991), and in- dividual tephra layers at Site U1443 can be tentatively correlated to the main eruptions of the Toba Caldera complex at 75, ~450, ~840, and ~1200 ka (Chesner and Rose, 1991; Dehn et al., 1991; Farrell and Janecek, 1991). The possibility of 40Ar/39Ar dating of these ash layers and correlation to a robust paleomagnetic reversal scheme at Site U1443 in combination with high-resolution stable isotope re- cords will further contribute to intercalibrating Pliocene–Pleisto- cene geomagnetic reversal and δ18O chronostratigraphy (Hall and Farrell, 1993, 1995). Stratigraphic correlation A composite section and splice to establish a continuous sedi- ment sequence were created using Holes U1443A–U1443D. Cor- relation and splicing were based on MS, NGR, and reflectance spectroscopy b data (Figure F15). The MS data proved particularly useful for correlation in the upper ~63 m core composite depth be- 13 Volume 353 IODP Proceedings Volume 353 Volume 353 S.C. Clemens et al. Expedition 353 summary Figure F16. Toba Tephra layers D, d, and E in (A) Holes U1443A and (B) U1443B. Figure F16. Toba Tephra layers D, d, and E in (A) Holes U1443A and (B) U1443B. Figure F17. Correlation of mudline cores in Holes U1443A, U1443B, U1443C, and U1443D. Figure F17. Correlation of mudline cores in Holes U1443A, U1443B, U1443C, and U1443D. northern Sumatra is well documented on the northern NER (Ninkovich, 1979; Ninkovich et al., 1978; Dehn et al., 1991), and in- dividual tephra layers at Site U1443 can be tentatively correlated to the main eruptions of the Toba Caldera complex at 75, ~450, ~840, and ~1200 ka (Chesner and Rose, 1991; Dehn et al., 1991; Farrell and Janecek, 1991). The possibility of 40Ar/39Ar dating of these ash layers and correlation to a robust paleomagnetic reversal scheme at Site U1443 in combination with high-resolution stable isotope re- cords will further contribute to intercalibrating Pliocene–Pleisto- cene geomagnetic reversal and δ18O chronostratigraphy (Hall and Farrell, 1993, 1995). Transitional Cretaceous/Paleogene boundary The sedimentary record of the Cretaceous/Paleogene (K/Pg) boundary transition was difficult to assess at Site 758 because this interval was only recovered in Hole 758A. The transition occurs at the base of the core catcher of Core 121-758A-31X (295.6 meters below seafloor [mbsf]) and is potentially incomplete because of the nearby coring gap between Cores 121-758A-31X and 32X. The Site U1443 K/Pg boundary (66.04 Ma) was identified between Samples 353-U1443A-39X-4W, 40 cm, and 39X-5W, 60 cm (Figure F18). Sediments in this interval are heavily bioturbated (containing bur- rows, blebs, patches, color banding, and mottles), and most samples studied within this interval contain a mixture of Cretaceous and Pa- leocene nannoplankton species. In particular, Zoophycos spreite burrows are common in the K/Pg boundary succession and are a prominent feature in upper Maastrichtian sediments of Holes U1443A and U1443B. The interval below Section 353-U1443A- 39X-5W contains moderately to poorly preserved Cretaceous nan- nofossil assemblages. No latest Maastrichtian marker events were found, possibly a result of a minor hiatus. Figure F16. Stratigraphic correlation Toba Tephra layers D, d, and E in (A) Holes U1443A and (B) U1443B. D d E D d E A Toba tephra layer D d E D d E B Toba tephra layer Figure F17. Correlation of mudline cores in Holes U1443A, U1443B, U1443C, and U1443D. U1443A-1H U1443B-1H U1443C-1H U1443D-1F Toba ash 100 cm 150 cm 50 cm 0 cm duplication? U1443C-2H Toba ash Toba ash U1443A-1H U1443B-1H U1443C-1H U1443D-1F Toba ash duplication? Toba ash Toba ash D d E D d E Toba tephra layer U1443A-1H U1443B-1H U1443C-1H U1443D-1F Toba ash duplication? Toba ash Toba ash D laye D d E D d E d U1443B-1H duplication? E 100 cm 150 cm 50 cm 0 cm U1443C-2H B D d D d E Toba tephra layer d E interval was only recovered in Hole 758A. The transition occurs at the base of the core catcher of Core 121-758A-31X (295.6 meters below seafloor [mbsf]) and is potentially incomplete because of the nearby coring gap between Cores 121-758A-31X and 32X. The Site U1443 K/Pg boundary (66.04 Ma) was identified between Samples 353-U1443A-39X-4W, 40 cm, and 39X-5W, 60 cm (Figure F18). Sediments in this interval are heavily bioturbated (containing bur- rows, blebs, patches, color banding, and mottles), and most samples studied within this interval contain a mixture of Cretaceous and Pa- leocene nannoplankton species. In particular, Zoophycos spreite burrows are common in the K/Pg boundary succession and are a prominent feature in upper Maastrichtian sediments of Holes U1443A and U1443B. The interval below Section 353-U1443A- 39X-5W contains moderately to poorly preserved Cretaceous nan- nofossil assemblages. No latest Maastrichtian marker events were found, possibly a result of a minor hiatus. northern Sumatra is well documented on the northern NER (Ninkovich, 1979; Ninkovich et al., 1978; Dehn et al., 1991), and in- dividual tephra layers at Site U1443 can be tentatively correlated to the main eruptions of the Toba Caldera complex at 75, ~450, ~840, and ~1200 ka (Chesner and Rose, 1991; Dehn et al., 1991; Farrell and Janecek, 1991). The possibility of 40Ar/39Ar dating of these ash layers and correlation to a robust paleomagnetic reversal scheme at Site U1443 in combination with high-resolution stable isotope re- cords will further contribute to intercalibrating Pliocene–Pleisto- cene geomagnetic reversal and δ18O chronostratigraphy (Hall and Farrell, 1993, 1995). IODP Proceedings A note on the Hole U1443A CSF-A depth scale Stratigraphic and visual correlation indicates that ~5.4 m of sed- iment was lost from the bottom of Core 353-U1443A-1H (Figure F20). This is not surprising in light of the fact that four APC core barrels were bent or broken during attempts to establish the various holes at Site U1443. In this case the addition of ~5.4 m to the CSF-A depth of all cores below Core 1H is necessary to make the CSF-A depths more consistent with those in Holes U1443B–U1443D. The Campanian to Maastrichtian part of the succession at Site U1443 is characterized by moderate to high recovery of indurated yellowish brown nannoplankton chalk, which grades downcore into greenish glauconite-rich chalk with intervals of abundant Inocera- mus shell fragments (Figure F19). The lower part of this interval in- cludes several centimeter-thick indurated chert nodules and nodular layers, which caused a significant decrease in drilling prog- ress and recovery in the lowermost two cores of Hole U1443A (Cores 353-U1443A-47X and 48X). The recovery of Upper Creta- ceous sediments above these Campanian siliceous layers was good in both Holes U1443A and U1443B, resulting in a complete strati- graphic record of this interval, which was characterized by only Late Oligocene to middle Miocene paleomagnetic record g p g The high-quality paleomagnetic record of Site 758 was limited to the uppermost 100 m, which was retrieved by APC coring. Below this level, biscuiting of the cored sediment by XCB coring prohib- ited high-quality magnetic measurements. At Site U1443, it was possible to deploy the APC and HLAPC core barrels in sediment deeper than 200 m CSF-A, which opened up a new window for ob- taining paleomagnetic records for the lower part of the Miocene and upper Oligocene. In particular, the record between 130 and 200 m CSF-A is characterized by high-quality paleomagnetic data that allow establishment of a complete paleomagnetic reversal scale that covers the entire lower Miocene and upper Oligocene, including the Oligocene/Miocene stage boundary. fragments recovered at Site 758. The low degree of diagenesis and the complete recovery may allow construction of a spliced bulk car- bonate carbon isotope record that covers the entire Maastrichtian to the middle/late Campanian including the Campanian/Maastrich- tian boundary. IODP Proceedings Transitional Cretaceous/Paleogene boundary The low degree of diagenesis and the complete recovery may allow construction of a spliced bulk car- bonate carbon isotope record that covers the entire Maastrichtian to the middle/late Campanian including the Campanian/Maastrich- tian boundary. A note on the Hole U1443A CSF-A depth scale Stratigraphic and visual correlation indicates that ~5.4 m of sed- iment was lost from the bottom of Core 353-U1443A-1H (Figure F20). This is not surprising in light of the fact that four APC core b l b t b k d i tt t t t bli h th i Figure F18. Site U1443 Cretaceous/Paleocene transitional boundary. Pg K Figure F19. A. Transition from greenish glauconitic chalk to brownish-red- dish nannofossil marls, Hole U1443A (42X-1). B. Glauconitic marl with ino- ceramid shell material (46X-1). A B Figure F18. Site U1443 Cretaceous/Paleocene transitional boundary. Figure F18. Site U1443 Cretaceous/Paleocene transitional boundary. Figure F18. Site U1443 Cretaceous/Paleocene transitional boundary. Figure F19. A. Transition from greenish glauconitic chalk to brownish-red- dish nannofossil marls, Hole U1443A (42X-1). B. Glauconitic marl with ino- ceramid shell material (46X-1). Figure F19. A. Transition from greenish glauconitic chalk to brownish-red- dish nannofossil marls, Hole U1443A (42X-1). B. Glauconitic marl with ino- ceramid shell material (46X-1). First complete spliced record of Neogene deepwater sediments in the Indian Ocean Pg K A A First complete spliced record of Neogene deepwater sediments in the Indian Ocean B B B fragments recovered at Site 758. The low degree of diagenesis and the complete recovery may allow construction of a spliced bulk car- bonate carbon isotope record that covers the entire Maastrichtian to the middle/late Campanian including the Campanian/Maastrich- tian boundary. B The continuous pelagic sediment record of Site U1443 with well-preserved benthic and planktonic foraminifers will allow for the extension of existing Pliocene–Pleistocene isotope records from Site 758 to the base of the Neogene and tracking of deepwater and surface water isotopic signals in the eastern equatorial Indian Ocean over the last 25 My. Sedimentation rates >0.41 cm/ky will still allow analyses of stable isotopes and geochemical records in a resolution sufficient to resolve orbital-scale climate variability and to correlate these records to existing Atlantic and Pacific orbitally tuned isotope curves. Transitional Cretaceous/Paleogene boundary The sedimentary record of the Cretaceous/Paleogene (K/Pg) boundary transition was difficult to assess at Site 758 because this 14 Volume 353 IODP Proceedings Expedition 353 summary fragments recovered at Site 758. The low degree of diagenesis and the complete recovery may allow construction of a spliced bulk car- bonate carbon isotope record that covers the entire Maastrichtian to the middle/late Campanian including the Campanian/Maastrich- tian boundary. A note on the Hole U1443A CSF-A depth scale Figure F19. A. Transition from greenish glauconitic chalk to brownish-red- dish nannofossil marls, Hole U1443A (42X-1). B. Glauconitic marl with ino- ceramid shell material (46X-1). A B S.C. Clemens et al. Expedition 353 summary First complete spliced record of Neogene deepwater sediments in the Indian Ocean The continuous pelagic sediment record of Site U1443 with well-preserved benthic and planktonic foraminifers will allow for the extension of existing Pliocene–Pleistocene isotope records from Site 758 to the base of the Neogene and tracking of deepwater and surface water isotopic signals in the eastern equatorial Indian Ocean over the last 25 My. Sedimentation rates >0.41 cm/ky will still allow analyses of stable isotopes and geochemical records in a resolution sufficient to resolve orbital-scale climate variability and to correlate these records to existing Atlantic and Pacific orbitally tuned isotope curves. Late Oligocene to middle Miocene paleomagnetic record The high-quality paleomagnetic record of Site 758 was limited to the uppermost 100 m, which was retrieved by APC coring. Below this level, biscuiting of the cored sediment by XCB coring prohib- ited high-quality magnetic measurements. At Site U1443, it was possible to deploy the APC and HLAPC core barrels in sediment deeper than 200 m CSF-A, which opened up a new window for ob- taining paleomagnetic records for the lower part of the Miocene and upper Oligocene. In particular, the record between 130 and 200 m CSF-A is characterized by high-quality paleomagnetic data that allow establishment of a complete paleomagnetic reversal scale that covers the entire lower Miocene and upper Oligocene, including the Oligocene/Miocene stage boundary. Recovery of expanded Campanian–Maastrichtian sediment succession The Campanian to Maastrichtian part of the succession at Site U1443 is characterized by moderate to high recovery of indurated yellowish brown nannoplankton chalk, which grades downcore into fragments recovered at Site 758. Background and objectives The main scientific objective of Expedition 353 is to reconstruct changes in precipitation and runoff on suborbital to orbital time- 15 Volume 353 Volume 353 S.C. Clemens et al. Expedition 353 summary Expedition 353 summary Figure F20. Holes U1443A/U1443B correlation of Cores 353-U1443A-1H through 2H and 353-U1443B-1H and 2H, indicating a 5.4 m gap at the base of Core 353-U1443A-1H. including the Mahanadi River and the Ganges-Brahmaputra river complex, and the Andaman Sea, including the Irrawaddy and Sal- ween river systems. scales. To achieve this objective, site locations were selected accord ing to the likelihood of continuous sedimentation and proximity t the main sources of freshwater feeding the Northern Bay of Benga of Core 353-U1443A-1H. U1443A U1443B U1443B-1H U1443B-2H U1443B-2H U1443B-1H Tephra A (75 ka) Tephra D (731-750 ka) Tephra d (756-760 ka) Tephra E (774-780 ka) One month prior to the start of the expedition, an alternate drilling plan was requested, one that included only sites in interna- tional waters. To accommodate this, we utilized existing site survey data acquired by colleagues at the University of Bremen in 1997 and 2006 to locate drilling targets in the central Bengal Fan, following a latitudinal transect approach. This alternate plan included sets of sites at three different latitudes in international waters at 11°N, 14°N, and 17° to 18°N. U1443A U1443B U1443B-1H Tephra A (75 ka) Sediments at these alternative sites do not meet the paleocli- matic time-series oriented scientific objectives of Expedition 353 because the sedimentary archive of the Bengal Fan includes abun- dant coarse-grained turbidite sequences that do not have the strati- graphic integrity required for high-resolution reconstruction of climate change. However, the alternate sites on the Bengal Fan are complementary to the objectives of Expedition 354 (France-Lanord et al., 2014). One of the crucial issues for Expedition 354 is the con- tinuity of the terrigenous flux from the Himalayan source to the ~8°N drilling targets. Channel migration may, at times, decouple the middle fan from the supply. The addition of sites north of the 8°N Expedition 354 transect sites may help to evaluate these pro- cesses and provide additional material for the understanding of ter- rigenous flux into the Bengal Fan, one of the main objectives of the Expedition 354 drilling proposal. Background and objectives p g p p Site U1444 (14°N, 84°49.74′E; 3132 mbsl) is located at common midpoint (CMP) 1302 on seismic Line GeoB97-041 (Schwenk and Spieß, 2009) (Figure F21A; Table T1). The site is located in the western part of the lower Bengal Fan, close to the westernmost abandoned channel that fed the western part of the lower fan (Em- mel and Curray, 1984). Seismics (Figure F21B–F21C) suggest that these sediments are composed of a series of buried channel-levee sequences incised into hemipelagic sediments atop the underlying 85°E Ridge. Schwenk and Spieß (2009) identified two prominent seismic reflectors (Unconformities Uc and Ud) bounding a more transparent hemipelagic unit between 4.32 and 4.20 s two-way trav- eltime (TWT), corresponding to ~105 to 185 mbsf. These uncon- formities can be traced for several hundred kilometers to the east along Line GeoB97-041 (Schwenk and Spieß, 2009). Similar re- gional-scale unconformities in the lower fan have been correlated to Site 218 and dated as earliest Pliocene (~4.8 Ma) and middle Pleisto- cene (~0.65 Ma) in age (von der Borch, Sclater, et al., 1974; Schwenk and Spieß, 2009). Unconformities Uc and Ud were inferred to be Pleistocene and Pliocene in age as well, but prior to drilling Site U1444, they had not been dated. Thus, a primary objective of this site was to determine the lithologic changes associated with Uncon- formities Uc and Ud and to date these reflectors in an effort to as- sess the degree to which turbidite and intercalated hemipelagic sequences are continuous from the upper to lower fan regions. U1443B-2H IODP Proceedings Coring summary At Site U1444, Holes U1444A and U1444B were drilled to total depths of 330.6 and 128.6 m DSF, respectively (Table T1). In Holes U1444A and U1444B, the full-length (9.7 m long) APC system and the XCB system were primarily used. The HLAPC (4.7 m long) was used only for Core 353-U1444A-24F. For Holes U1444A and U1444B, the APC system was used to refusal followed by the XCB. A total of 46 cores were recovered at Site U1444. The APC cored interval was 160.0 m with a recovery of 156.21 m (98% recovery). The HLAPC cored interval was 4.8 m with a recovery of 3.94 m of core (82%). The XCB cored interval was 246.9 m with a recovery of scales. To achieve this objective, site locations were selected accord- ing to the likelihood of continuous sedimentation and proximity to the main sources of freshwater feeding the Northern Bay of Bengal, 16 Volume 353 IODP Proceedings S.C. Clemens et al. Expedition 353 summary Figure F21. (A) Location of Site U1444 (after Emmel and Curray, 1984), (B) seismic line, and (C) line-drawing interpretation (after Schwenk and Spieß, 2009). In the line drawing, gray indicates buried channel-levee systems and bold black lines indicate seismic unconformities (Uc and Ud). (Continued on next page.) e (57%). Hole U1444B contained a drilled interval 0 m DSF. The overall recovery for Site U1444 was tion, smear slide analysis, and physical property measurem recovered lithologies are siliciclastic and are composed of with intercalated hemipelagic intervals ray indicates buried channel-levee systems and bold black lines indicate seismic unconformities (Uc and Ud). (Continued on nex DSDP Site 218 IODP Site U1444 India Sri Lanka ODP Sites 717, 718, 719 GeoB97-041 GeoB97-028 GeoB97-059/069 GeoB97-020/027 A E N Abandoned major channel with levees Erosive channel or mass wasting Major channel with levees Constructive distributary channel with levees locally Small channel belonging to depositional lobe? Small unleveed channel crossing (channel pattern unknown) Abandoned Fan divisions Gradient in m/m Channel dimensions: height versus width in m Seamounts or topographic highs Dominant lithology Gravel Sand Sandy/Silty mud Mud Abandoned lobe Slump Debris flow A Constructive distributary channel with levees locally tion, smear slide analysis, and physical property measurements. The recovered lithologies are siliciclastic and are composed of turbidites with intercalated hemipelagic intervals. tion, smear slide analysis, and physical property measurements. The recovered lithologies are siliciclastic and are composed of turbidites with intercalated hemipelagic intervals. • Unit I (0–95.01 m CSF-A) is composed of silty sand and silty clay with numerous turbidites. Coring summary 140.06 m of core (57%). Hole U1444B contained a drilled interval from 47.5 to 95.0 m DSF. The overall recovery for Site U1444 was 73%. • Unit I (0–95.01 m CSF-A) is composed of silty sand and silty clay with numerous turbidites. • Unit I (0–95.01 m CSF-A) is composed of silty sand and silty clay with numerous turbidites. IODP Proceedings Lithostratigraphy Sediments recovered from Holes U1444A and U1444B can be divided into four lithostratigraphic units based on visual descrip- 17 Volume 353 Expedition 353 summary S.C. Clemens et al. Figure F21 (continued). ~1 km Uc Ud 85°E Ridge B U1444 4.1 4.2 4.3 4.4 4.5 4.6 4.7 4.8 4.9 5.0 Two-way traveltime (s) ~1 km Uc Ud 85°E Ridge B U1444 4.1 4.2 4.3 4.4 4.5 4.6 4.7 4.8 4.9 5.0 Two-way traveltime (s) Uc Ud 85°E Ridge C U1444 Two-way traveltime (s) 4.3 4.5 4.7 4.9 Uc Ud 85°E Ridge C U1444 Two-way traveltime (s) 4.3 4.5 4.7 4.9 C B Two-way traveltime (s) • Unit II (95.01–168.91 m CSF-A) is primarily nannofossil-rich clay with silt and foraminifers. Turbidites in Unit II are less abundant, thinner, and finer than in Unit I. • Unit II (95.01–168.91 m CSF-A) is primarily nannofossil-rich clay with silt and foraminifers. Turbidites in Unit II are less abundant, thinner, and finer than in Unit I. microfossils were present, preservation was good to moderate and rarely poor. Foraminifers ranged in preservation from poor to good in the samples in which they occur. • Unit III (168.91–255.60 m CSF-A) is characterized by very poor recovery and is dominated by silty fine to medium sand. Calcareous nannofossil assemblages were of Pleistocene to late Miocene age and tropical/subtropical in character. Most Pleisto- cene, Pliocene, and late Miocene calcareous nannofossil bioevents used to define the biostratigraphic zones were present. This enabled us to assign all samples from hemipelagic sediment sequences and many samples from turbidite-rich sequences to a single biozone. Emiliania huxleyi is present in Sample 353-U1444A-9H-CC near the bottom of lithostratigraphic Unit I, suggesting that this entire unit (95 m) was deposited during the Late Pleistocene and Holo- cene. The top of lithostratigraphic Unit II (95–177 m CSF-A) was placed within nannofossil Zone NN20. Sample 353-U1444A-20X- CC (168.89 m CSF‑A), near the base of Unit II, falls within nanno- fossil Zone NN16 (late Pliocene). The lithostratigraphic Unit IV se- quence spans Zone NN13 (middle Pliocene) to Zone NN11 (late Miocene). Sample 353-U1444A-37X-CC, 18 cm, collected 12 cm above the bottom of Hole U1444A, contained common Discoaster loeblichii, which suggests that the oldest sediments in Hole U1444A are younger than 7.53 Ma (late Miocene). • Unit IV (255.60–323.39 m CSF-A) is composed primarily of nannofossil-rich clay and clayey silt interbedded with sand and silt turbidites. Lithostratigraphy Turbidites at this site show typical erosional bases and normal (fining-upward) grading but rarely show structures typical of classic Bouma or Stow sequences (e.g., parallel, wavy, or lenticular lami- nae). Structureless turbidites suggest very rapid sedimentation and/or disturbance of the water-saturated sands during drilling and recovery. Turbidites were classified visually as either silt or sand as the maximum grain size. At Site U1444, turbidites can be classified compositionally from visual and smear slide descriptions as mica rich, well sorted quartz dominant, organic debris rich, or glauconite rich. The overall lithologic differences between units and variation in turbidite grain size and thickness are consistent with fluctuations in the proximity of active turbidity current channels on the Bengal Fan. Planktonic foraminifer assemblages in Hole U1444A are tropical to subtropical throughout the Neogene and include species indica- tive of upwelling. The last occurrence (LO) of Globigerinoides ruber (pink) in Sample 353-U1444A-4H-CC and the first occurrence (FO) of Globorotalia flexuosa in Sample 14X-CC indicate that sediments between 35.83 and 119.42 m CSF-A are of Pleistocene age. Pliocene datums were found in Samples 16X-CC through 19X-CC (137.41– 166.53 m CSF-A) and in Cores 32X and 33X (283.02–290.50 m CSF‑A). Eleven core catcher samples from 168.89 to 270.34 m CSF‑A were either barren or contained no index species. Miocene • Unit II (95.01–168.91 m CSF-A) is primarily nannofossil-rich clay with silt and foraminifers. Turbidites in Unit II are less abundant, thinner, and finer than in Unit I. Paleomagnetism Paleomagnetic measurements were conducted on archive-half sections for both Holes U1444A and U1444B. Sections dominated by sand were not measured because of their unstable texture and the risk of contamination to the magnetometer. All sections from XCB cores were severely affected by drilling-related overprint and were difficult to interpret. A selection of discrete samples (N = 119) taken from working-half sections was also analyzed. Based on APC section measurements and discrete sample data, we propose tenta- tive magnetozones throughout Hole U1444A. Magnetostrati- graphic ages are given as 0.781 Ma at ~112–118 m CSF-A and 2.581 Ma at ~148–151 m CSF-A. Other magnetozones were difficult to correlate with the geomagnetic polarity timescale because of poor recovery in some intervals and drilling-related overprint rema- nence. Most of the discrete data downhole to ~100 m CSF-A are in- fluenced by gyroremanent magnetization, likely due to the presence of greigite (a ferrimagnetic iron sulfide). A rock magnetic evaluation of greigite abundance revealed a depth distribution in broad agree- ment with interstitial water geochemistry. Regional Unconformity Uc (Schwenk and Spieß, 2009) is clearly recognizable in the sedimentary succession of Site U1444. It is de- fined by a lithologic change from hemipelagic clays and silty clays (with foraminifer and nannofossil ages of ~200 ka) to a turbidite- dominated sequence composed of silty sands and silty clays that are largely barren of microfossils. This 95 m long sequence represents the last 200 ky of sedimentary deposition at Site U1444. The deeper regional unconformity (Ud in Schwenk and Spieß, 2009) is charac- terized by a change from very poorly recovered turbiditic silty sands and clayey silts to hemipelagic clays and silty clays in the middle Pliocene (between ~3.6 and 3.8 Ma). Within the error of the dating, which includes sedimentation rate–based extrapolations over tens of meters at Site 218, the Late Pleistocene regional Unconformity Uc could be the same age as the unconformity at Site 218. The deeper Pliocene Unconformity Ud at Site U1444 appears to be on the order of 1 My younger relative to the unconformity at Site 218. The onset of channel-levee systems in the latest Miocene is indi- cated by downhole data from ODP Leg 116 and seismic data along an east–west transect at 8°N (Schwenk and Spieß, 2009). Geochemistry Site U1444 varies geochemically, reflecting the depositional his- tory of this section of the Bengal Fan. Significant changes in sedi- ment and pore water chemistry reflect the major changes in lithology. Lithostratigraphic Unit I features a complete drawdown of sulfate and increases in methane, alkalinity, and nutrients. Deeper units are less reducing with opposing trends in many elements, in- cluding pronounced secondary peaks in sulfate, silicate, B, Li, Mn, and K. Organic C is low but variable, ranging between 0 and 2 wt%, whereas carbonate content tracks lithostratigraphic units: Units I and III are low, whereas Units II and IV are variable but generally higher (10–30 wt%). Physical properties Downhole variations in physical property measurements at Site U1444 reflect changes in lithology, condensed sections associated with depositional hiatuses, and diagenetic processes. The overall suite of physical property data is divided into four broad physical property units in Hole U1444A (PP Units A1–A4) and three broad physical property units in Hole U1444B (PP Units B1–B3). The tur- biditic lithostratigraphy is the primary reason why units in Holes U1444A and U1444B do not correspond well to each other. How- ever, some general observations remain valid for both holes. First, MS values are higher in coarse-grained sediments than in clay-rich sections. Second, core-length MS trends show high values at the bottom of cores and low values at the top, possibly indicating me- chanical sorting at core-length scales. These trends are mainly pres- ent in the APC cores that are almost entirely composed of coarse and watery sands, suggesting that settling of the coarse-grained fac- Biostratigraphy The biostratigraphic age model for Site U1444 was established by combining calcareous nannofossil and planktonic foraminifer datums from Hole U1444A. Age-depth relationships for the two fossil groups show good agreement. All samples were nearly or completely barren of diatoms. Calcareous microfossils (nannofos- sils, planktonic foraminifers, and benthic foraminifers) at Site U1444 were rare or absent in turbidite-rich sequences and common to abundant in hemipelagic sediment sequences. Where calcareous 18 IODP Proceedings Volume 353 S.C. Clemens et al. Expedition 353 summary sediments are identified in Samples 35X-CC and 36X-CC. The oc- currence of Candeina nitida in Sample 36X-CC indicates that it can be no older than 8.43 Ma. tion, which is generally associated with magnetic material, could have occurred from severe disturbance during coring and recovery. Third, throughout the hole the physical property data are highly variable, which makes it challenging to reliably interpret them for subtle changes in sediment type and character. Site U1444 is characterized by large variations in sedimentation rate that occur because of high-frequency large-scale episodes of sediment redeposition. Mean sedimentation rates reach >20 cm/ky in lithostratigraphic Unit I (0–95 m CSF-A) and >30 cm/ky in Unit III (168.91–255.60 m CSF-A). Lithostratigraphic Units II and IV are mostly hemipelagic with sedimentation rates an order of magnitude lower than in Units I and III, around 1.39 cm/ky for Unit II and 1.60 cm/ky for Unit IV. Highlights: dating seismic Unconformities Uc and Ud in Hole U1444A The middle Bengal Fan is characterized by channel-levee sys- tems that are erosionally incised into underlying hemipelagic sedi- ments (Schwenk and Spieß, 2009). Seismic profiles indicate two major regional unconformities, which were dated at Site 218 as ear- liest Pliocene (~4.8 Ma) and Middle Pleistocene (0.65 Ma) (von der Borch, Sclater, et al., 1974; Schwenk and Spieß, 2009). The extent of the regional unconformities, the onset of levee systems, and faults terminating within Pleistocene sediments suggest that tectonic events, in addition to changes in sediment supply and transport, ex- erted major controls on the sedimentation patterns of the Bengal Fan (Schwenk and Spieß, 2009). Stratigraphic correlation No composite depth scale or splice was constructed for Site U1444. Coring disturbance and mechanical sorting of turbidite sands led to a restructuring of the physical properties on a core-by- core basis in the upper part of the site (see Physical properties), and a lack of signal in the XCB cores in the lower part of the site prevented establishment of reliable tie points between Holes U1443A and U1444B. Paleomagnetism These changes were possibly related to changes in the erosion and weath- ering regime in the drainage area of the Ganges River (Schwenk and Spieß, 2009). At Site U1444, the oldest sediments of Hole U1444A (3.8 to ~6 Ma) consist of fine-grained mudstones with intercalated thin-bedded levee sediments followed by a strong increase in coarse-grained silty–sandy levee and turbidite sediments represent- ing a fan lobe deposit in the middle Pliocene around 3.8 Ma. Lithostratigraphy Sediments recovered from Holes U1445A–U1445C are princi- pally composed of hemipelagic clays with a significant biogenic component and occasional thin turbidites of Late Pleistocene (Ho- locene?) to late Miocene age. Because of the consistent fine-grained nature of the sediments, only one stratigraphic unit is recognized and divided into two subunits (Ia and Ib), primarily based on nan- nofossil and biosilica content: The percentage of planktonic foraminifers is high (mean = 86.4%) in the uppermost 158.97 m of Hole U1445A but lower (mean = 59.9%) throughout the remaining Hole U1445A core catchers. The total number of foraminifers per 10 cm3 raw sediment is highly variable, ranging from 0 to >24,000, with a marked decrease in numbers deeper than 180 m CSF-A. The number of benthic fora- minifers per 10 cm3 raw sediment follows the same pattern, averag- ing 1667 between 0 and 180 m CSF-A and 129 in deeper sediments. Reworking is apparent in many late Miocene through Pleistocene samples. Pleistocene planktonic foraminifer assemblages were re- covered from 6.9 to 261.51 m CSF-A. Pleistocene planktonic assem- blages are dominated by tropical to warm subtropical species and by some temperate species. Species commonly associated with upwell- ing zones are common to abundant throughout Samples 353- U1445A-1H-CC through 20H-CC, indicating coastal upwelling or vertical mixing of water masses. Species of markedly different last appearance datums are found within a zone encompassing Samples 9H-CC through 18H-CC, which implies reworking. Pliocene plank- tonic foraminifer assemblages were recovered from Samples 29X- CC through 68X-CC. Miocene sediments span Samples 69X-CC through 77X-CC. The occurrence of Globigerinoides conglobatus in Sample 76X-CC suggests a maximum age for this sample of 6.20 Ma. • Subunit Ia (0–165.28 m CSF-A) is a sequence of Late Pleistocene (Holocene?) to Middle Pleistocene olive-gray to dark greenish gray clay with biosilica and clay with nannofossils, with occa- sional beds of biosilica-rich clay. • Subunit Ib (165.28–667.56 m CSF-A) comprises the interval be- tween the first downcore occurrence of biosilica-rich clays to the bottom of Hole U1445A, and spans the Middle Pleistocene to late Miocene. Foraminifers are a persistent but variable component of the Subunit Ia clays, whereas calcareous nannofossils become less abun- dant downhole. Subunit Ib is dominated by very dark greenish gray biosilica-rich clay with glauconite, with significant intervals of in- creased diatom content between ~160 and 330 m CSF-A. Background and objectives Site U1445 (2503 mbsl) is located near the southern end of the Mahanadi basin, on the eastern margin of India (Figure F2). This location, ~94 km offshore, offers the opportunity to drill sediments underlying the low-salinity waters of the Indian margin, a result of 19 IODP Proceedings Volume 353 S.C. Clemens et al. Expedition 353 summary Biostratigraphy summer monsoon rainfall and runoff from the peninsular rivers of India, including the Ganges-Brahmaputra river complex and the Mahanadi River. The location of this site, seaward of the base of the slope, offers the potential to reach Miocene sediments but is not protected from turbidite deposition. Objectives at this site were to recover late Miocene to Holocene sediment sections in order to re- construct changes in the Indian summer monsoon at orbital to sub- orbital timescales. Calcareous and siliceous microfossils are present in Hole U1445A with variable downcore abundance trends. Calcareous nannofossil preservation is good to moderate, and abundances vary between rare and abundant. Pleistocene to late Miocene nannofossil assemblages are typical of tropical/subtropical paleoenvironments. Foraminifers are dominant to abundant in the uppermost 188 m in Hole U1445A. Abundance decreases rapidly deeper than 188 m CSF-A and falls to few or rare deeper than 290 m CSF-A. Foramin- ifer preservation is good to moderate, with the exception of three samples that show poor preservation. Diatom preservation ranges from good to poor and tends to be better when diatom abundance is higher. Coring summary At Site U1445, Holes U1445A, U1445B, and U1445C were drilled to total depths of 672.6, 33.0, and 305.2 m DSF, respectively (Table T1). In Holes U1445A and U1445C, the APC and XCB sys- tems were primarily used; only the APC system was used in Hole U1445B. For Holes U1445A and U1445C, the APC system was used to refusal. Following refusal of the APC, the XCB was deployed to total depth. Overall, 117 cores were recorded for the site. A total of 487.34 m of core over a 476.3 m cored interval was recovered using the APC (102% recovery). The cored interval with the XCB was 534.5 m with a core recovery of 518.02 m (97%). The overall recov- ery percentage for Site U1445 was 99%. The total time spent on Site U1445 was 9.3 days. All Pleistocene nannofossil marker species were found, with the exception of Reticulofenestra asanoi. The Pliocene/Pleistocene boundary (2.59 Ma), located between Cores 353-U1445A-31X and 38X, is bracketed by a number of Discoaster LOs that are dated be- tween 2.39 and 2.8 Ma. The Miocene/Pliocene boundary (5.33 Ma) is well constrained between 603.22 and 606.19 m CSF-A and is based on the LO of Triquetrorhabdulus rugosus (5.28 Ma) and the FO of Ceratolithus acutus (5.35 Ma) in this interval. The oldest cal- careous nannofossil sample studied (Sample 353-U1445-A-77X- CC; 667.46 m CSF-A) contained Discoaster quinqueramus and Dis- coaster berggrenii, suggesting an age between 5.59 and 7.53 Ma. Be- yond the age model, nannofossil biostratigraphy was of great use in dating rip-up clasts found in several horizons in Holes U1445A and U1445C. The matrix surrounding the clasts was of the same age as the sediment above and below the horizons containing the clasts, whereas the rip-up clasts themselves were of different ages, ranging from Late Pleistocene to late Eocene. Geochemistry Wireline logging in Hole U1445A was initiated following coring operations. The hole was swept with a heavy-weight water-based mud that included barite as a weighting additive to improve hole stability. The presence of barite attenuates gamma rays, lowers the NGR measurements, and affects (increasing) the photoelectric fac- tor. Therefore, the effects of barite in the mud on the log data needs to be taken into consideration when interpreting the data. Two log- ging passes were made, but a bridge was encountered at ~440 m, which prevented the tools from descending any further. The upper 80 m of the hole was not logged as it was occupied by the drill string. The first logging run was made with the triple combo and the second pass was made with the Formation MicroScanner tool. The wireline logging suite is divided into three units based on hole con- ditions inferred from the caliper data. The caliper data for Units 1 (100–237 m CSF-A) and 3 (380–440 m CSF-A) show highly irregu- lar hole conditions with large washout zones and reduce the confi- dence in the data quality for these units. The caliper data for Unit 2 (237–380 m CSF-A) show that the hole is close to bit size. NGR counts were found to agree in magnitude with shipboard Natural Gamma Radiation Logger logs, and the wireline density data are in excellent agreement with moisture and density measurements, sug- gesting that the data in this unit are reliable. However, the density logs from the Whole-Round Multisensor Logger and Special Task Multisensor Logger appear to consistently underestimate the true density of the formation. Neutron porosity data suggest higher ap- parent porosity than density porosity as a result of water bound in the structure of the clays that dominate the downhole lithology. There are no resistivity spikes suggestive of concentrated gas hy- drate occurrence, and thus the gas hydrate at this site, if present, could be dispersed. The geochemistry of Site U1445 strongly reflects the processes of sulfate reduction and methanogenesis associated with microbial degradation of organic matter. High methane concentrations are found in headspace and void space gas samples. A high meth- ane/ethane ratio suggests that the methane is mostly of biogenic or- igin. The organic C content is as high as 4 wt%, and carbonate is associated with intervals of more abundant calcite microfossils. Lithostratigraphy Foramini- fers and nannofossils are both much less common in Subunit Ib than in Subunit Ia but are present in small numbers throughout, particularly between ~570 and 670 m CSF-A where an increase in calcareous nannofossils is observed. Thin (~2–20 cm) turbidites are present in sediments from both Subunits Ia and Ib, varying in com- position from silt-sized quartz-rich silt/sands to foraminifer-rich sands with occasional bioclastic-rich sands. Soupy and mousselike intervals, characteristic of gas hydrate dissociation, were identified in sediments recovered using the APC system in Holes U1445A– U1445C from both Subunits Ia and Ib. Overall, cores from Hole U1445C were considerably less disturbed (drilling disturbance and gas expansion) than those recovered from Hole U1445A at equiva- lent depths, possibly due to reduced heave during Hole U1445C op- erations. Diatoms are useful for age estimation throughout the entire sed- imentary column of Hole U1445A. Several diatom events were rec- ognized between the LO of Nitzschia reinholdii (0.90–1.0 Ma) and the LO of Nitzschia miocenica (5.7 Ma). Valve preservation is mostly good to moderate. Strong variations in abundance and shifts in the species composition of the diatom assemblage will help to re- IODP Proceedings 20 Volume 353 S.C. Clemens et al. Expedition 353 summary construct paleoceanographic changes in the eastern Bay of Bengal between the late Miocene and the Late Pleistocene. The highly di- verse diatom community mainly consists of species typical of warm to temperate low-latitude ocean waters. High-productivity species, including Thalassionema nitzschioides var. nitzschioides and resting spores of Chaetoceros, tend to dominate whenever total diatom abundance is higher than “few.” A certain degree of freshwater/ter- rigenous input is revealed by the recurrent presence of numerous phytoliths and freshwater diatoms. slide analysis suggests that the prevalence of diatom-rich clay has enhanced the porosities while maintaining a low MS, density, and NGR. It also appears that the change in color reflectance from Units 1 to 2 could be a result of transition to XCB coring from APC cor- ing. Unit 3 is characterized by high NGR peaks, which may be re- lated to drilling disturbance or coarse-grained fractions that are likely to occur in the few turbiditic layers that were identified. Paleomagnetism Paleomagnetic measurements were conducted on archive-half sections for all three holes at Site U1445, with alternating field (AF) demagnetization up to 10 mT. Discrete samples (N = 219) taken from working-half sections were also analyzed, with AF demagneti- zation typically up to 30 mT. Characteristic remanent magnetiza- tions (ChRMs) of discrete samples were calculated using the principal component analysis (PCA) technique. Remanence intensities, in addition to bulk magnetic parameters, drop significantly in the uppermost 10–20 m because of diagenetic reduction. Nonetheless, a fairly straightforward magnetic polarity stratigraphy was constructed, except for the lower part of Hole U1445A (deeper than ~4 Ma or ~470 m CSF-A). Rock magnetic measurements on discrete samples show promising variations in the bulk magnetic properties for the last 6 My. The significance of these trends has yet to be evaluated; however, a periodicity, possibly on the order of 40–60 ky, may be revealing astronomical-scale cy- cles. Stratigraphic correlation A composite scale (CCSF-A) and a splice (CCSF-D) were con- structed for Site U1445 using MS, NGR, and red, green, and blue (RGB) data from Holes U1445A and U1445C. Splicing among these holes enabled us to construct a continuous stratigraphic sequence to ~252 m CCSF-D. Because of shipboard data quality issues, cor- relation should be viewed with caution deeper than ~50 m CCSF-A and especially deeper than ~236 m CCSF-A. The age model for Site U1445 was established by combining cal- careous nannofossil, planktonic foraminifer, and diatom datums with paleomagnetic reversal datums. Age-depth relationships for Hole U1445A of the three fossil groups show good agreement and match the magnetochron boundary datums well. The combined biostratigraphic/magnetostratigraphic age model suggests a mean sedimentation rate of 11.4 cm/ky, assuming a linear fit of all data. Highlights Sediments at this site were remarkably conducive to XCB cor- ing. The deepest hole at Site U1445 was cored to 673 mbsf, 447 m of which was cored using the XCB, with an overall recovery of 99%. Within intervals where core expansion was strong because of mi- crobial gas, an 8 m advance of the 9.5 m XCB was employed to pro- vide expansion room within the liner, thereby significantly reducing the loss of sediment out of the top of the corer. Geochemistry The pore water chemistry of Site U1445 reflects reducing conditions with sulfate depleted at around 18 m CSF-A. Other examples are alkalinity peaking at around 30 mM near 50 m CSF-A and dissolved Ba and silicate increasing downhole. The influence of seawater con- tamination and oxygenation during activities associated with XCB coring is noticeable in the profiles of some ions, most notably sul- fate, phosphate, and Fe. Lithostratigraphy Sediments recovered from Holes U1446A–U1446C are Holo- cene to Middle Pleistocene in age and are a hemipelagic mixture of a dominant lithogenic fraction diluting a minor biogenic fraction. They are primarily composed of dark gray to gray clay with nanno- fossils, nannofossil-rich clay, clay with foraminifers, clay, clay with biosilica, and biosilica-rich clay. Because of the consistent clayey na- ture of the sediments, only one lithostratigraphic unit (Unit I) is rec- ognized at this site (0–180.11 m CSF-A) with no subunit divisions. Visual core description and smear slide observations were used with supporting information from physical properties (primarily NGR) and geochemical parameters (weight percent CaCO3) to evaluate the varying abundances of the siliciclastic fraction (clays, silts, and occasionally sands) versus the biogenic fraction (calcareous and sili- ceous). Turbidites are rare at Site U1446, with only occasional thin quartz-rich or shallow-water carbonate-rich intervals indicative of transported sediments. Overall, drilling disturbance is generally slight to moderate, as all cores were retrieved using the APC system, characterized by voids and gas expansion cracks. Physical properties Downhole variations in physical property measurements at Site U1445 reflect changes in lithology, condensed sections associated with depositional hiatuses, and diagenetic processes. Based on sig- nificant transitions within the data, the overall suite of physical property data is divided into three broad PP units (1–3) in Hole U1445A. Due to a shorter drilling depth, only the first two of the three physical property units could be identified in Hole U1445C. Unit 2 is characterized by anomalous porosity preservation. Smear Discrete intervals in the late Miocene, middle Pliocene, and Middle Pleistocene are characterized by abundant diatoms, includ- ing species characteristic of high-productivity environments and those found in coastal and freshwaters. Similarly, species of plank- 21 IODP Proceedings Volume 353 Volume 353 S.C. Clemens et al. Expedition 353 summary Expedition 353 summary Background and objectives Site U1446 (1430 mbsl) is located within the Mahanadi basin on the eastern margin of India (Figure F2). This location, ~70 km off- shore, offers the opportunity to drill sediments underlying the low- salinity waters of the Indian margin, a result of summer monsoon rainfall and runoff from the peninsular rivers of India, including the Ganges-Brahmaputra river complex and the Mahanadi River. Lo- cated atop a small rise near the base of the slope, this site should be protected from turbidite deposition. Objectives at this site were to recover Late Pleistocene to Holocene sediment sections in order to reconstruct changes in the Indian summer monsoon at orbital to suborbital timescales. The age model for Site U1446 was established by combining cal- careous nannofossil, planktonic foraminifer, and diatom datums with paleomagnetic reversal datums. The combined biostrati- graphic/magnetostratigraphic age model indicates a mean sedimen- tation rate of ~16 cm/ky from 0 Ma to just over 1 Ma. The oldest planktonic foraminifer datum encountered is the LO of Globorota- lia tosaensis (0.61 Ma) in Sample 353-U1446A-12H-CC, whereas the combination of nannofossils and diatoms datums constrains the basal age of Hole U1446A to between 0.90 and 1.26 Ma. Coring summary At Site U1446, Holes U1446A, U1446B, and U1446C were drilled to total depths of 180.0, 27.1, and 182.0 m DSF, respectively (Table T1). In Holes U1446A and U1446C, the APC and HLAPC systems were deployed. In Hole U1446B, only the APC system was used. Overall, 47 cores were recorded for the site. A total of 344.95 m of core over a 342.3 m cored interval was recovered using the APC system (101% recovery). The HLAPC cored interval was 46.8 m with a recovery of 49.43 m of core (106%). The overall recovery percentage for Site U1446 was 101%. The total time spent on Site U1446 was 2.3 days. Geochemistry The organic C content of Site U1446 ranges from 0.8 to 1.6 wt% (average = 1.2 wt%), and the geochemistry of the site mainly reflects the anaerobic processes of sulfate reduction and methanogenesis associated with microbial degradation of organic matter. Sulfate de- clines rapidly from 28 mM at the sediment/water interface to near zero values at approximately 20 m CSF-A. Alkalinity peaks at 20 m depth, consistent with the production of bicarbonate during sulfate reduction. Sulfate reduction appears to continue throughout the core, as suggested by the gradual increase in dissolved Ba concen- tration with depth, as this Ba is likely derived from barite. Changes in the concentration of other cations and anions (Fe, Mg, Ca, am- monium, and Sr) in pore waters can be readily explained by the mi- crobial-induced chemical reactions and their effects on pH, alkalinity, and mineral dissolution and precipitation. Headspace methane concentrations are generally low shallower than 30 m CSF‑A but rise between 30 and 70 m CSF-A, peaking at ~50 m CSF‑A, with a moderate concentration of 1000 ppm. High methane to ethane ratios suggest that the methane is mostly of biogenic ori- gin (methanogenesis). Carbonate content varies significantly be- tween 2 and 20 wt% within the scale of a few meters, with the low content intervals corresponding to high NGR (terrestrial clay-rich materials). Biostratigraphy tonic foraminifers characteristic of higher productivity environ- ments increase in abundance in the middle Pliocene and Pleistocene, possibly suggesting nutrient input related to changes in surface water stratification or runoff. Calcareous microfossils are continuously present in the sedi- ments of Hole U1446A, whereas siliceous microfossils are sporadi- cally present. Calcareous nannofossils show abundances ranging from few to abundant in the smear slides studied, and their preser- vation is generally very good to good and occasionally moderate. Foraminifers are dominant to abundant in Hole U1446A in 19 of 21 core catcher samples and common or few in the remaining two samples. Preservation is good to moderate in all foraminifer sam- ples. Diatoms are present in the lower and upper part of the Hole U1446A record. Valve preservation ranges from good to poor and tends to be better when abundance is higher. All cores were infrared scanned on the catwalk prior to section- ing. A number of centimeter-scale cold spots were detected, having up to 7°C difference relative to background core temperatures. Rhi- zon pore water sampling indicated chloride values well below that of seawater, confirming these anomalies as hydrates. Splicing among overlapping holes (0 to ~305 m) at Site U1445 will yield the first continuous section recovered from the Indian margin that spans the Pleistocene. Over this interval, fine-scale tur- bidites averaged ~1 per core (<5 cm thick). Excellent XCB recovery (99%) over the remaining ~360 m will allow for high-resolution monsoon reconstruction into the late Miocene. All Late Pleistocene calcareous nannofossil events were identi- fied. Late to Middle Pleistocene assemblages are typical of tropi- cal/subtropical paleoenvironments. All core catcher samples from Hole U1446A contain Pleistocene planktonic foraminifer assem- blages. Planktonic assemblages are dominated by tropical to warm subtropical species. The diatom community is highly diverse and resembles that of Site U1445. The diatom assemblage consists of species typical of warm to temperate, low- to mid-latitude ocean waters. Paleomagnetism Paleomagnetic measurements were conducted on archive-half sections for all three holes at Site U1446, with AF demagnetization up to 10 mT. Discrete samples taken from working-half sections of Holes U1446A (N = 53) and U1446C (N = 9) were also analyzed, with AF demagnetization up to 40–80 mT. ChRMs of these discrete 22 IODP Proceedings Volume 353 S.C. Clemens et al. Expedition 353 summary the Irrawaddy and Salween Rivers as well as nearby Little Andaman Island at tectonic to suborbital timescales. samples were calculated using the PCA technique. The paleomag- netic signal is generally good downhole to ~120 m CSF-A, but data are highly scattered deeper. A magnetostratigraphy is constructed from 0 to 1.173 Ma (to ~170 m CSF-A) in Hole U1446A, albeit with a certain degree of ambiguity for some of the chrons. In addition, anhysteretic remanent magnetization (ARM) was acquired and measured on a selection of discrete samples from Hole U1446A. A significant decrease in ARM as a consequence of diagenetic reduc- tion at 30 m CSF-A was observed. ARM intensity increases slightly from 40 to 60 m CSF-A and decreases from 100 to 130 m CSF-A. These variations could relate to multiple factors, such as paleopro- ductivity, postdepositional alteration, or unseen lithologic varia- tions, but unraveling these factors requires further exploration. Coring summary At Site U1447, Holes U1447A, U1447B, and U1447C were drilled to total depths of 738.0, 24.4, and 160.9 m DSF, respectively (Table T1). In Hole U1447A the APC, HLAPC, and XCB systems were deployed. In Holes U1447B and U1447C only the APC system was used. Overall, 108 cores were recorded for the site. A total of 451.27 m of core over a 444.3 m cored interval was recovered using the APC system (102% recovery). The HLAPC cored interval was 67.8 m with a core recovery of 69.12 m (102%). The XCB cored in- terval was 409.2 m with a core recovery of 395.15 m (97%). The overall recovery percentage for Site U1447 was 99%. The total time spent on Site U1447 was 6.13 days. Highlights Site U1446 recovered ~180 m of section from two holes, ranging from Middle Pleistocene to Holocene, with a mean sedimentation rate of ~16 cm/ky. Hole U1446C recovered an excellent mudline, with common fragile epifaunal tubular agglutinated foraminifers. The entire interval was APC cored with excellent recovery. Toba ash (likely) was recovered at ~15 m CSF-A. Although the site is within the hydrate stability zone and in the same general vicinity as Site U1445, no indications of hydrates were found, possibly due to the finer grained nature of the sediments. The spliced section at this site will yield an excellent continuous Late Pleistocene section. The high sedimentation rate offers excellent opportunities to reconstruct monsoon climates at suborbital to orbital timescales over the entire interval. y g g • Unit IV (489.80–740.46 m CSF-A) is composed of late Miocene greenish gray–gray clayey nannofossil ooze with glauconite or biosilica, biosilica-rich clay with varying proportions of glauco- nite and nannofossils, and nannofossil-rich clay with biosilica. The observed lithologic differences between the units are pri- marily the result of varying abundances of biosilica (principally dia- toms and sponge spicules), turbidites, and nannofossils. Lithostratigraphy Sediments recovered from Holes U1447A, U1447B, and U1447C are principally composed of Late Pleistocene to late Mio- cene hemipelagic clays with a significant biogenic component and numerous turbidites (primarily calcitic turbidites), composing four distinct lithostratigraphic units (I–IV): • Unit I (0–126.00 m CSF-A) is composed of Late Pleistocene greenish gray clayey nannofossil oozes with foraminifers and foraminifer-rich nannofossil oozes with clay. • Unit II (126.00–329.12 m CSF-A) is composed of Late Pleisto- cene–late Pliocene greenish gray clayey nannofossil oozes with foraminifers, foraminifer-rich nannofossil oozes with clay, and clayey calcareous oozes with varying proportions of foramini- fers. This unit is also characterized by the presence of numerous thin to thick light gray calcitic turbidites, described as foramin- ifer-rich and bioclastic-rich layers with authigenic carbonate and foraminifers dominated by the sand or silt size fraction. Physical properties Downhole variations in physical property measurements at Site U1446 likely reflect changes in lithology, sedimentation rate, and diagenetic processes. Variation in physical properties as a function of depth at Site U1446 were not as significant as seen at other sites in this expedition. Only subtle changes in density and porosity were observed and formed the basis for dividing Hole U1446A into three PP subunits (1a–1c). Physical property data in Hole U1446C was very similar to Hole U1446A, and all the subunits could be identi- fied. In both holes, cyclic variability in NGR values are interpreted to be reflecting changes in relative lithogenic and biogenic sedimen- tation. Stratigraphic correlation A composite scale (CCSF-A) and a splice were constructed for Site U1446 using MS, NGR, and RGB data from Holes U1446A and U1446C. Splicing among these holes enabled us to construct a con- tinuous stratigraphic sequence to ~108 m CCSF-D. Because of data quality and time availability issues, correlation should be viewed with caution deeper than ~78 m CCSF-A. • Unit III (329.12–489.80 m CSF-A) is composed of early Pliocene to late Miocene greenish gray clayey nannofossil oozes to calcar- eous oozes with varying amounts of glauconite. Biostratigraphy Calcareous nannofossils are abundant or common throughout Hole U1447A. Their preservation is generally very good to moder- ate, except in two samples that contain poorly preserved, overgrown nannofossils. Foraminifers are dominant to abundant and very well preserved throughout the Pliocene–Pleistocene, with a few excep- tions in the turbiditic interval of lithostratigraphic Unit II where preservation decreases to moderate or poor and/or abundance de- creases to common. Foraminifers are dominant to abundant in most Miocene sediments as well, though abundance decreases to com- mon or few in samples where diatom abundance is high. Preserva- tion in the Miocene continues to be good with only two exceptions where moderate preservation accompanies low foraminifer abun- dances. Diatoms are sporadically present in the uppermost 566 m of Hole U1447A. Their abundance varies between abundant and few Background and objectives Site U1447 is located in the Andaman Sea, ~45 km offshore Lit- tle Andaman Island at 1392 mbsl (Figure F2). The site lies within a basin on the eastern flank of a rise that separates two north-south– oriented basins structurally related to the Eastern Margin and Dili- gent faults penetrating upward from the underlying accretionary wedge complex. The location yields access to older sediments but is not protected from potential turbidite deposition. The objectives at this site were to recover Miocene to recent sediments in order to reconstruct changes in summer monsoon rainfall and runoff from 23 Volume 353 IODP Proceedings S.C. Clemens et al. Expedition 353 summary measurements and PCA directions. In addition, ARM was acquired and measured on a selection of Hole U1447A discrete samples for preliminary insight into depth variations in sediment bulk magnetic properties. downcore. Valve preservation ranges from good to poor, and valves are better preserved in samples with abundances that are higher than common. The frequency and distribution of nannofossil and foraminifer bioevents enabled the construction of a high-resolution age model for Site U1447. Diatoms were also useful for age estimation in sam- ples deeper than 556.11 m CSF-A. Derived paleomagnetic reversal datums fit well into the biostratigraphic age model. The base of the Late Pleistocene was delineated by the LO of G. ruber (pink) at 32.25 m CSF-A. The Pliocene/Pleistocene boundary (2.59 Ma) is placed just below the LO of Discoaster pentaradiatus (2.39 Ma), which oc- curs between 299.78 and 301.89 m CSF-A. The Miocene/Pliocene boundary (5.33 Ma) occurs between 468.89 and 470.19 m CSF-A based on the LO of T. rugosus (5.28 Ma) and the FO of C. acutus (5.35 Ma) in this interval. Physical properties Downhole variations in physical property measurements at Site U1447 reflect changes in lithology, condensed sections associated with depositional hiatuses, and diagenetic processes. The overall suite of physical property data in Hole U1447A, which is the deep- est hole, is divided into four broad PP units (1–4) based on signifi- cant transitions within the data. The depth of Holes U1447B and U1447C were not enough to use them for understanding hole-to- hole physical property correlation. Overall, a long-term cyclic vari- ability in NGR can be seen in Hole U1447A, which may be reflective of periodic changes in relative lithogenic and biogenic sedimenta- tion. The oldest calcareous nannofossil sample studied contained Discoaster hamatus, suggesting an age older than 9.53 Ma. The old- est planktonic foraminifer datum encountered is the FO of Neoglo- boquadrina acostaensis (9.83 Ma), and combined with the presence of Globorotalia limbata in Sample 87X-CC, suggests that the basal age of Hole U1447A is between 9.83 and 10.66 Ma. The oldest dia- tom datum is the FO of Thalassiosira burckliana (between 719.87 and 721.31 m CSF-A), suggesting that the bottom of Hole U1447A is slightly older than 9.1 Ma. Based on linear fits, including all bio- stratigraphic and magnetostratigraphic data, sedimentation rates in Hole U1447A are highest in the Pleistocene (average = 12 cm/ky lin- ear sedimentation rate [LSR]), moderate in the Pliocene and most of the Miocene (average = 6.5 cm/ky LSR), and possibly lower in the Miocene deeper than 700 m CSF-A. Stratigraphic correlation A composite scale (CCSF-A) and a splice were constructed for Site U1447 using MS, NGR, and RGB data from Holes U1447A and U1447C. Splicing among these holes enabled us to construct a con- tinuous stratigraphic sequence to ~147 m CCSF-D. Because of data quality and time availability issues, correlation should be viewed with caution deeper than ~112 m CCSF-A. Geochemistry The geochemistry of Site U1447 mainly reflects the anaerobic processes of sulfate reduction and methanogenesis associated with microbial degradation of organic matter. The organic C content ranges from 0.1 to 2.0 wt% (average = 0.8 wt%). Sulfate decreases rapidly from 28 mM at the sediment/water interface to near zero values at approximately 20 m CSF-A. Alkalinity peaks at 20 m CSF‑A, consistent with the production of bicarbonate during sulfate reduction. A gradual increase in dissolved Ba concentration with depth suggests ongoing barite dissolution. Changes in the concen- tration of other elements and ions (Fe, Mn, Ca, B, ammonium, and Sr) in pore water can be readily explained by microbially mediated chemical reactions and their effects on pH, alkalinity, and mineral dissolution and precipitation. Headspace methane concentrations peak at 125 m CSF-A, abruptly decrease below the peak, and are consistently low to the bottom of the hole. High methane to ethane ratios suggest that the methane is mostly of biogenic origin (me- thanogenesis). Carbonate content varies significantly between 9 and 55 wt%. Unconformities with possible short-term stratigraphic hiatuses in the upper part of the sedimentary succession include the follow- ing: • A massive carbonate turbidite bed in the uppermost (Holocene) part of the succession at ~1.40–2.4 m CSF-A. • A hiatus at ~290 m CSF-A (~2.17 s TWT on Line AN-01-25A) is indicated by the co-occurrence of several biostratigraphic events within the same stratigraphic interval. This unconformity may be associated with a prominent turbidite bed at interval 353- U1447A-36F-1, 40 cm, close to the Pliocene–Pleistocene transi- tion. • A second possible hiatus is indicated by a cluster of biostrati- graphic events at 340 m CSF-A (~2.23 s TWT on Line AN-01- 25A) within the early–late Pliocene transition. The unconfor- mity would be located within Core 46X, which had no recovery, where drilling progress indicated a noticeable change in lithol- ogy. • A third possible hiatus is located at 470 m CSF-A (~2.38 s TWT on Line AN-01-25A), within the Miocene–Pliocene transition (several nannoplankton events clustering around 5.35 Ma da- tum) with the LO of Globoquadrina dehiscens (5.92 Ma) in Core 61X-CC just below this interval. The sedimentary succession in this interval shows a distinct unconformity at the base of a dis- tinct black layer at interval 60X-4, 70–73 cm. Highlights Site U1447 exhibits a nearly complete hemipelagic succession from the recent (mudline with common fragile epifaunal tubular ag- glutinated foraminifers of the genera Rhabdammina and Sac- corhiza) to middle Miocene (~9.8 Ma), with all nannofossil biostratigraphic markers present. S.C. Clemens et al. S.C. Clemens et al. Expedition 353 summary nannofossils and foraminifers, and characterized by a low biosil- ica content (<1%). The expanded stratigraphic section in Hole U1447A provides a first high-resolution record of late Miocene Indian monsoon vari- ability in the Bay of Bengal. Shipboard investigations provided the first evidence for a link between monsoonal intensity and biosili- ceous paleoproductivity patterns in the Andaman Sea over the last 9.8 My. All Pleistocene and most of the Pliocene sediments of Site U1447 are virtually barren of diatoms, which exhibit last common occurrences below the Miocene/Pliocene boundary. This temporal distribution pattern is distinctly different from the late Miocene– early Pliocene “biogenic bloom” observed throughout the open tropical Indo-Pacific and Atlantic Oceans (e.g., Farrell et al., 1995; Dickens and Owen, 1999; Hermoyian and Owen, 2001). The maxi- mum in diatom abundance in the Andaman area (roughly between 9.8 and 6.5 Ma at Site U1447) together with an increase in authi- genic pyrite and carbonate started earlier and was of shorter dura- tion compared to the open-ocean biogenic bloom. These peculiar productivity changes within the Bay of Bengal could either reflect the availability of nutrients brought into the Bay of Bengal by rivers under a variable monsoonal intensity since the middle Miocene (Clift, 2006) or be related to the presence or absence of a seasonal freshwater lid, which would prevent upwelling of nutrient-rich in- termediate waters during the monsoon season. • Unit III (338.60–379.11 m CSF-A) is composed of late Miocene greenish dark to light gray nannofossil-rich clay and clay with nannofossils, and characterized by the increased abundances of glauconite (up to ~7%) and siliceous sponge spicules (up to g p p g p p ~4%). A hiatus representing ~8 My occurs at the base of Unit III at 379.11 m CSF-A, where there is an abrupt change to litho- stratigraphic Unit IV. • Unit IV (379.11–420.60 m CSF-A) is composed of middle–early Miocene greenish gray–light greenish gray biosiliceous ooze with varying proportions of clay and nannofossils. The observed lithologic differences between the units are pri- marily the result of varying abundances of biogenic components (nannofossils, foraminifers, diatoms, and sponge spicules), clay, and glauconite. Background and objectives Site U1448 (1097 mbsl) is located in the Andaman Sea, ~44 km offshore Little Andaman Island, 17.7 km south of Site U1447 (Figure F2). The site lies atop a rise separating north-south–oriented basins associated with the Eastern Margin and Diligent faults, both related to the underlying accretionary wedge complex. The ridge-top loca- tion should have a lower sedimentation rate and be better protected from turbidite deposition. The objectives at this site are to recover Pliocene to recent sediments in order to reconstruct changes in summer monsoon rainfall and runoff from the Irrawaddy and Sal- ween Rivers as well as nearby Little Andaman Island at tectonic to suborbital timescales. p g The age model for Site U1448 was established by combining cal- careous nannofossil, planktonic foraminifer, and diatom datums (the latter below the hiatus only). The oldest calcareous nannofossil sample studied above the hiatus suggests an age between 5.94 and 6.91 Ma. Planktonic foraminifer datums constrain the age of the oldest core catcher sample above the hiatus to between 5.92 and 8.58 Ma. The three different microfossil groups suggest slightly dif- ferent ages for the sediments below the hiatus. The interval between the hiatus and the bottom of Hole U1448A was constrained to be- tween 14.91 and 17.71 Ma based on nannofossils. The foraminifer assemblage in the youngest core catcher sample below the hiatus is older than 14.53 Ma, whereas the oldest planktonic foraminifer da- tum encountered is the LO of Praeorbulina glomerosa (14.78 Ma). The foraminifer P. glomerosa is found in the deepest sample (353- U1448A-60X-CC), defining the basal age of Hole U1448A as 14.78– 16.27 Ma. Combining the ages of biostratigraphic events of nanno- fossil and planktonic foraminifers provides an estimate for the dura- tion of the Miocene hiatus of approximately 8 My between 5.94– 6.91 and 14.91–16.38 Ma. The co-occurrence of the diatom species Rhaphidodiscus marylandicus and Annellus californicus in the low- ermost part of Hole U1448A suggests the bottom of Hole U1448A is older than 16.7 Ma (LO of R. marylandicus) and younger than 17.3 Ma (FO of A. californicus). Coring summary At Site U1448, Holes U1448A, U1448B, and U1448C were drilled to total depths of 421.0, 358.6, and 34.3 m DSF, respectively (Table T1). In Hole U1448A the APC, HLAPC, and XCB systems were deployed. In Hole U1448B the APC and HLAPC systems were used, and only the APC system was deployed for Hole U1448C. Overall, 121 cores were recorded for the site. A total of 427.52 m of core over a 416.2 m cored interval was recovered using the APC sys- tem (103% recovery). The HLAPC cored interval was 318.6 m with a core recovery of 333.24 m (105%). The XCB cored interval was 77.6 m with a core recovery of 78.21 m (101%). The overall recovery percentage for Site U1448 was 103%. The total time spent on Site U1448 was 3.9 days. Biostratigraphy Calcareous nannofossils are abundant throughout Hole U1448A, and their preservation is very good in the Pleistocene, Pliocene, and uppermost Miocene sediments (0–379.1 m CSF-A), with few reworked species. A hiatus was identified at 379.11 m CSF‑A, separating late Miocene and middle Miocene sediments. Below the hiatus, nannofossils are abundant and moderately pre- served. Nannofossil assemblages are typical of tropical/subtropical paleoenvironments. Foraminifers are well-preserved in all core catcher samples from Hole U1448A; they are dominant in the upper 340 m, whereas their abundance decreases to common just above the hiatus and drops to few to common below the hiatus. Diatoms are rarely present in the uppermost 379.04 m of Hole U1448A. Their abundance abruptly increases just below the hiatus, and their preservation varies from moderate to good. Paleomagnetism Paleomagnetic measurements were conducted on archive-half sections for all three holes at Site U1447, with AF demagnetization up to 10 mT. Discrete samples taken from working-half sections of Hole U1447A were also analyzed, with stepwise AF demagnetiza- tion up to 40–80 mT. ChRMs of these discrete samples were calcu- lated using the PCA technique. Both types of data became noisy deeper than ~90 m CSF-A, with a predominant drilling-related overprint. Two tentative magnetostratigraphies were proposed to at least 1.778 Ma at ~220 m CSF-A for Hole U1447A. These two classes use sets of data with different reliability from pass-through These unconformities may be related to regional tectonic events; further investigation of their stratigraphic and regional ex- tent will provide new insight into the Miocene–recent tectonic evo- lution of the Andaman accretionary wedge complex. 24 Volume 353 IODP Proceedings Physical properties Downhole variations in physical property measurements at Site U1448 reflect changes in lithology, condensed sections associated with depositional hiatuses, and diagenetic processes. The overall suite of physical property data from Hole U1448A, which is the deepest hole, is divided into six broad PP units (1–6) based on sig- nificant transitions within the data. The top five units in Hole U1448A were based on subtle changes in the physical property trends rather than rapid transitions. The NGR data best show the long-term cyclicity at this site and are possibly related to variations in terrigenous input. Unit 1 shows high variability in almost all physical property measurements, likely from unconsolidated sedi- ment. Unit 2 shows minor MS variability, increasing density, de- creasing porosity, and relatively steady NGR counts. Unit 3 is characterized by an increase in MS, increase in density, increase in NGR counts, decrease in porosity, and lower b* values, suggesting higher terrigenous input and clay relative to carbonate or siliceous Lithostratigraphy Sediments recovered from Site U1448 are principally composed of hemipelagic clays with a significant biogenic component, com- posing four distinct lithostratigraphic units (I–IV) of Late Pleisto- cene to middle–early Miocene age: Age-depth relationships for Hole U1448A based on biostratigra- phy for the three fossil groups studied generally show good agree- ment. Sedimentation rates above the hiatus are between 5 and 6 cm/ky. Sedimentation rates below the hiatus are difficult to quantify given the wide range of estimated ages. • Unit I (0–183.11 m CSF-A) is composed of Late to early Pleisto- cene greenish gray clay with varying proportions of nannofossils and foraminifers, and clayey nannofossil ooze. • Unit II (183.11–338.60 m CSF-A) is composed of early Pleisto- cene to late Miocene greenish clay with varying proportions of IODP Proceedings 25 Volume 353 S.C. Clemens et al. Expedition 353 summary Stratigraphic correlation A composite scale (CCSF-A) and a splice were constructed for Site U1448 using MS, NGR, and RGB data from Holes U1448A, U1448B, and U1448C. Splicing among these holes enabled us to construct a continuous stratigraphic sequence to ~203 m CCSF-D. Following a gap estimated to be 2 m, a floating composite scale was constructed between ~205 and ~260 m CCSF-A using Holes U1448A and U1448B. Because of data quality and time availability issues, correlation should be viewed with caution deeper than ~115 m CCSF‑A. Paleomagnetism Paleomagnetic measurements were conducted on archive-half sections for all three holes at Site U1448. To accommodate the core flow, only APC and HLAPC cores from Hole U1448A were AF de- magnetized up to 10 mT. XCB cores from Hole U1448A and APC cores from Holes U1448B and U1448C were measured predomi- nantly for NRM, with a few exceptions. HLAPC cores from Hole U1448B have not been measured. Declination values from section measurements suggest the uppermost 200 m CSF-A is largely di- vided into three magnetozones, corresponding to the Brunhes (C1n), Matuyama (C1r), and Gauss (C2An) Chrons. Discrete sam- ples taken from working-half sections of Hole U1448A (N = 90) were also analyzed, with stepwise AF demagnetization typically up to 80 mT, but generally the determination of ChRM was difficult likely due to drilling-related overprint. Stability of NRM appears to be high just below the hiatus (379.11 m CSF-A), with higher resolu- tion measurements possibly inferring polarity transitions. ARM was acquired and measured on a selection of Hole U1448A discrete samples for preliminary insight into downcore variation in sedi- ment bulk magnetic properties. The high-resolution sedimentary succession of Site U1448 also allowed recognition of a suite of volcanic tephra layers that appear at least partly correlative to the tephra succession of Site U1443. In particular, the tephra of the Late Pleistocene Toba eruption (0.0738 Ma) could be recognized in all three holes of Site U1448 and pro- vides an excellent marker bed for stratigraphic correlation. The main highlight at Site U1448 was the recovery of pelagic middle Miocene sediment below the distinct seismic reflector at 1.865 s TWT on seismic Line AN01-26A (see Figure F3 in the Site U1448 chapter [Clemens et al., 2016]). The reflector represents a major unconformity between late Miocene (oldest biostratigraphic datum just above the unconformity is 6.91 Ma, LO of Reticulofenes- tra rotaria) and middle Miocene (14.53 Ma, LO of Praeorbulina si- cana just below the unconformity). The middle Miocene pelagic succession below the unconformity comprises a carbonate-rich bio- siliceous sedimentary succession that was fully recovered in very high quality XCB cores and will provide the first insight in mon- soonal influenced biogenic sedimentation during the middle Mio- cene climatic optimum. Highlights Hole U1448C recovered an excellent mudline, with common fragile epifaunal tubular agglutinated foraminifers. In contrast to the sedimentation at nearby Site U1447, the lower Pleistocene and middle–upper Pliocene at Site U1448 are not affected by gravity de- position (turbidites or fine-grained grain flows/debris flows). The lower Pleistocene and Pliocene of Site U1448 exhibit a complete hemipelagic succession with all calcareous biostratigraphic markers represented. The intermediate to high sedimentation rates of this succession (5–6 cm/ky) will allow high-resolution paleoclimatic re- constructions using calcareous and organic proxy indicators of pa- leotemperature and salinity in a continuous succession reaching into the late Miocene (6.91 Ma). Geochemistry fractions. Unit 4 corresponds to lithostratigraphic Unit II. It has generally high MS values, high density, a continued decrease in po- rosity, variable NGR counts, and the lowest b* values. In Unit 5, MS, NGR, and density decrease, whereas porosity and b* increase. The physical properties of Unit 5 likely indicate an increase in abun- dance of more biosilica-rich clays. All physical property measure- ments change rapidly at 379.1 m CSF-A, which corresponds to an approximately 8 My hiatus. The top five units in Hole U1448A could be identified in Hole U1448B. The geochemistry of Site U1448 mainly reflects the anaerobic processes of sulfate reduction and methanogenesis associated with microbial degradation of organic matter. The organic C content ranges from 0.2 to 1.4 wt% (average = 0.6 wt%). Sulfate decreases from 28 mM at the sediment/water interface to near zero values at approximately 40 m CSF-A. Alkalinity has a broad peak, signifi- cantly lower than at Site U1447, starting from 40 m to 150 m CSF-A before gradually decreasing, consistent with the production of bi- carbonate during sulfate reduction. Headspace methane concentra- tions increase immediately below the sulfate interface and continue downcore with the highest values between 200 and 250 m CSF-A. The overall concentrations (20–30 ppmv), however, are quite low compared to Site U1447. High methane to ethane ratios suggest that the methane is mostly of biogenic origin (methanogenesis). A grad- ual increase in dissolved Ba concentration with depth suggests on- going barite dissolution. Changes in the concentration of other elements and ions (Fe, Mn, Ca, B, ammonium, and Sr) in pore wa- ters can be readily explained by microbially mediated chemical re- actions and their effects on pH, alkalinity, and mineral dissolution and precipitation. Carbonate content varies significantly between 12 and 37 wt%. Critical aspects Physical property data quality and stratigraphic correlation Critical aspects Physical property data quality and stratigraphic correlation Operations at Site U1445, as planned, included two holes to 680 mbsf with logging. This was abbreviated to two holes spanning 672 and 305 m and logging. This site, located on the upper continental rise in the southern Mahanadi basin, although susceptible to turbi- dite and other gravity flow deposits, allowed recovery of lower Pleistocene to upper Miocene sediments that are not available at the other Mahanadi site (U1446). Recovery was a remarkable 99% at this site, including 447 m of XCB coring. Loss of sediment because of gas expansion was minimized by using 8 m XCB advances, allow- ing 1.5 m of accommodation space within the core barrel. This site was the only site at which centimeter-scale cold spots were detected by infrared scanning of the core on the catwalk prior to sectioning. These intervals were documented to be related to gas hydrates by Rhizon pore water sampling, which revealed reduced chlorinity driven by gas hydrate disassociation. Relatively poor data quality of MS (Section Half Multisensor Logger [SHMSL]), NGR, and color reflectance measurements (SHMSL), as well as muted physical properties signals made it diffi- cult to create reliable splices at sea for crucial parts of the records recovered during Expedition 353. In particular, sections that were cored using the HLAPC, which were of outstanding quality, could not be correlated with sufficient precision because of the lack of de- tail in the physical property data. This stemmed from two factors, the primary being the extremely noisy SHMSL data associated with the inability of the instrument to land the MS and spectral sensors flush on the core surface; it is possible that these measurements may simply have to be done by hand in order to attain reliable data. The other contributing factor is specific to the sediments recovered. In many cases, the MS signal of the sediment was low because diagen- etic factors and color variation within sediments was muted. Even so, the only data set that proved of reliable quality and sufficient res- olution was the line scan RGB data. These data, however, cannot stand alone as the only data available for creating reliable splices at sea. A possible solution to this dilemma in the future would be to install a shipboard X-ray fluorescence (XRF) scanning facility. Expedition 353 synthesis Six sites were drilled during Expedition 353 in the Bay of Bengal, recovering a total of 4280 m of sediment section during 32.9 days of on-site operations. One hole was logged. This is substantially less than the initially planned campaign targeting 6256 m of section, in- cluding downhole logging at three sites (Clemens et al., 2014). This 26 IODP Proceedings Volume 353 S.C. Clemens et al. Expedition 353 summary difference amounts to approximately 15 days of lost operational time. Weather accounted for 1.6 days of lost operations. Mechanical breakdowns and broken coring equipment accounted for 1.98 days of lost operations. The remainder of the lost operational time was due to delays in attaining the necessary permissions to operate in Indian exclusive economic zone (EEZ) waters and to scheduling and conducting the necessary vessel inspections required to operate in Indian EEZ waters. ~7 to ~14 Ma. Operations at Site U1447 initially targeted two deep holes to 738 m with downhole logging at one site. Restrictions on operational time reduced this site to one deep hole (738 m) and one shallow hole (160 m) with no logging. Similarly, operations at Site U1448 called for three holes to 422 m. This plan was truncated to two holes, one to target depth and the other to 356 m. These Anda- man Sea sites allow for the first time reconstruction of monsoonal climates from late Miocene to present in this region. Operations at Site U1443 (NER) were scheduled to be three holes to 350 mbsf with logging. This was largely accomplished with the exception of logging. Recovery was excellent, and triple coring produced the first complete spliced record of Neogene deepwater sediments in the Indian Ocean and a paleomagnetic reversal record spanning the lower Miocene and upper Oligocene. This will provide the opportunity for high-resolution climate reconstructions with solid chronostratigraphic control. Critical aspects Physical property data quality and stratigraphic correlation This would reliably compensate for low MS and color variability and greatly enhance the lithologic description available at sea. Site U1446 is located on the continental slope within the Maha- nadi basin on a northwest-southeast–trending ridge, effectively iso- lated from turbidite deposition. Operations here called for three holes to 184 mbsf. Loss of operational time resulted in an abbrevi- ated plan, including two holes to the target depth. Both holes were accomplished with APC coring with more than 100% recovery and excellent core quality. Recovered sediments spanned the past 1.2 My, including the initiation of high-amplitude Northern Hemi- sphere glaciation. Combined with sediments from Site U1445, re- construction of monsoon climate change from the latest Miocene through recent is now possible for the first time in this important region. To overcome significant problems with high-resolution sam- pling plans for Expedition 353, several core sections are being sent to the Woods Hole Oceanographic Institute and to the Bremen Core Repository for XRF core scanning. The core scanning results will be used to extend the reliable stratigraphic splices created on the ship prior to the sampling party at the Kochi Core Center. Andaman Sea Sites U1447 and U1448 are ~10 nmi apart but within different depositional settings. Site U1447 is located within a basin on the flank of a rise, susceptible to turbidite deposition but reaching back to the middle Miocene, whereas Site U1448 is on the top of the ridge, sheltered from turbidite deposition but reaching only into the late Miocene, where a hiatus transitions abruptly from Expedition 353 science assessment Drilling operations—core handling Drilling operations and core handling were performed abso- lutely professionally and with outstanding commitment. Results were accordingly extraordinary (i.e., full-recovery APC coring to more than 350 mbsf in two holes and complete APC records to the early Oligocene at Site U1443, where previous APC coring termi- nated within the Pleistocene). In the cases of technical issues or sed- iment issues such as degassing within the core barrel, appropriate drilling and core handling adjustments were implemented rapidly, preserving the quality of the cored sections. Overall, the outstand- ing skills of the drillers and core technicians were the main reason why, despite significant scheduling setbacks, the scientific goals of Expedition 353 were able to be accomplished. Bengal Fan Site U1444 was not targeted in the original opera- tions plan of for Expedition 353. Drilling this site was necessary be- cause it was in international waters and occupied a portion of the time during which negotiations over fees, duties, and inspection matters were ongoing. Nevertheless, this site will promote the sci- entific goals of Expedition 354. We successfully dated the Unconfor- mities Uc and Ud defined in Schwenk and Spieß (2009). Preliminary results indicate that Unconformity Uc, within the error of dating, is the same age as found at Site 218, whereas Unconformity Ud may be considerably younger than the Site 218 unconformity initially hy- pothesized to be of the same age. S.C. Clemens et al. S.C. Clemens et al. Expedition 353 summary As initially planned, the expedition targeted 6256 m of recovery and three logged holes (Clemens et al., 2014). The lost logging and nearly 2000 m of unrecovered sediment section are accounted for by close to 15 days of lost operational time. Weather accounted for 1.6 days of lost operations. Mechanical breakdowns, four broken core barrels, rusted winch wire, and parting of the winch wire ac- counted for 1.98 days of lost operations. This lost time can be largely attributed to the hazards of working in difficult environ- ments. Bolton, C.T., Chang, L., Clemens, S.C., Kodama, K., Ikehara, M., Medina-Eliz- alde, M., Paterson, G.A., Roberts, A.P., Rohling, E.J., Yamamoto, Y., and Zhao, X., 2013. A 500,000 year record of Indian summer monsoon dynamics recorded by eastern equatorial Indian Ocean upper water-col- umn structure. Quaternary Science Reviews, 77:167–180. http://dx.doi.org/10.1016/j.quascirev.2013.07.031 http://dx.doi.org/10.1016/j.quascirev.2013.07.031 Bosilovich, M.G., and Schubert, S.D., 2002. Water vapor tracers as diagnostics of the regional hydrologic cycle. Journal of Hydrometeorology, 3(2):149– 165. http://dx.doi.org/10.1175/1525- 7541(2002)003<0149:WVTADO>2.0.CO;2 The large majority of the lost operational time was due to delays in attaining the necessary permissions to operate in Indian EEZ wa- ters, scheduling and conducting the vessel inspections required to operate in Indian EEZ waters, and protracted negotiations over con- cern that the JOIDES Resolution would incur excessive fees/duties upon entering port for inspection. These issues are typically ad- dressed and resolved prior to an expedition and would not normally impact operational time. Expedition 353 had to compensate by cut- ting two of the three planned logging operations, canceling one pri- mary site in the Mahanadi basin entirely, and canceling or shortening additional holes required to ensure complete spliced re- cords to full target depths in the Mahanadi basin and Andaman Sea. Fortunately, formation characteristics and sea state worked strongly in our favor to produce excellent (near 100%) recovery, even in XCB cored intervals. This will serve to reduce the negative impact of lost operational time. Burton, K.W., Gannoun, A., and Parkinson, I.J., 2010. Climate driven glacial– interglacial variations in the osmium isotope composition of seawater recorded by planktic foraminifera. Earth and Planetary Science Letters, 295(1–2):58–68. http://dx.doi.org/10.1016/j.epsl.2010.03.026 Burton, K.W., and Vance, D., 2000. Glacial–interglacial variations in the neo- dymium isotope composition of seawater in the Bay of Bengal recorded by planktonic foraminifera. Earth and Planetary Science Letters, 176(3– 4):425–441. http://dx.doi.org/10.1016/S0012-821X(00)00011-X Caballero-Gill, R.P., Clemens, S.C., and Prell, W.L., 2012. http://dx.doi.org/10.1029/2011PA002268 Cai, Y., An, Z., Cheng, H., Edwards, R.L., Kelly, M.J., Liu, W., Wang, X., and Shen, C.-C., 2006. High-resolution absolute-dated Indian Monsoon record between 53 and 36 ka from Xiaobailong Cave, southwestern China. Geology, 34(8):621–624. http://dx.doi.org/10.1130/G22567.1 Caley, T., Malaizé, B., Bassinot, F., Clemens, S.C., Caillon, N., Linda, R., Char- lier, K., and Rebaubier, H., 2011a. The monsoon imprint during the “atyp- ical” MIS 13 as seen through north and equatorial Indian Ocean records. Quaternary Research, 76(2):285–293. References Akhil, V.P., Durand, F., Lengaigne, M., Vialard, J., Keerthi, M.G., Gopal- akrishna, V.V., Deltel, C., Papa, F., and de Boyer Montégut, C., 2014. A modeling study of the processes of surface salinity seasonal cycle in the Bay of Bengal. Journal of Geophysical Research: Oceans, 119(6):3926– 3947. http://dx.doi.org/10.1002/2013JC009632 http://dx.doi.org/10.1016/j.yqres.2011.07.001 Caley, T., Malaizé, B., Revel, M., Ducassou, E., Wainer, K., Ibrahim, M., Shoeaib, D., Migeon, S., and Marieu, V., 2011b. Orbital timing of the Indian, East Asian and African boreal monsoons and the concept of a ‘global monsoon’. Quaternary Science Reviews, 30(25–26):3705–3715. http://dx.doi.org/10.1016/j.quascirev.2011.09.015 Ali, S., Hathorne, E.C., Frank, M., Gebregiorgis, D., Stattegger, K., Stumpf, R., Kutterolf, S., Johnson, J.E., and Giosan, L., 2015. South Asian monsoon history over the past 60 kyr recorded by radiogenic isotopes and clay min- eral assemblages in the Andaman Sea. Geochemistry, Geophysics, Geosys- tems, 16(2):505–521. http://dx.doi.org/10.1002/2014GC005586 Caley, T., Malaizé, B., Zaragosi, S., Rossignol, L., Bourget, J., Eynaud, F., Marti- nez, P., Giraudeau, J., Charlier, K., and Ellouz-Zimmermann, N., 2011c. New Arabian Sea records help decipher orbital timing of Indo-Asian monsoon. Earth and Planetary Science Letters, 308(3–4):433–444. http://dx.doi.org/10.1016/j.epsl.2011.06.019 Altabet, M.A., Higginson, M.J., and Murray, D.W., 2002. The effect of millen- nial-scale changes in Arabian Sea denitrification on atmospheric CO2. Nature, 415(6868):159–162. http://dx.doi.org/10.1038/415159a Cane, M.A., and Molnar, P., 2001. Closing of the Indonesian Seaway as a pre- cursor to East African aridification around 3–4 million years ago. Nature, 411(6834):157–162. http://dx.doi.org/10.1038/35075500 An, Z., Clemens, S.C., Shen, J., Qiang, X., Jin, Z., Sun, Y., Prell, W.L., Luo, J., Wang, S., Xu, H., Cai, Y., Zhou, W., Liu, X., Liu, W., Shi, Z., Yan, L., Xiao, X., Chang, H., Wu, F., Ai, L., and Lu, F., 2011. Glacial–interglacial Indian summer monsoon dynamics. Science, 333(6043):719–723. http://dx.doi.org/10.1126/science.1203752 Cawthern, T., Johnson, J.E., Giosan, L., Flores, J.A., Rose, K., and Solomon, E., 2014. A late Miocene–early Pliocene biogenic silica crash in the Andaman Sea and Bay of Bengal. Marine and Petroleum Geology, 58(Part A):490– 501. http://dx.doi.org/10.1016/j.marpetgeo.2014.07.026 Anand, P., Kroon, D., Singh, A.D., Ganeshram, R.S., Ganssen, G., and Elder- field, H., 2008. Coupled sea surface temperature–seawater δ18O recon- structions in the Arabian Sea at the millennial scale for the last 35 ka. Paleoceanography, 23(4):PA4207. http://dx.doi.org/10.1029/2007PA001564 Cerling, T.E., Harris, J.M., MacFadden, B.J., Leakey, M.G., Quade, J., Eisen- mann, V., and Ehleringer, J.R., 1997. Global vegetation change through the Miocene/Pliocene boundary. Nature, 389(6647):153–158. http://dx.doi.org/10.1038/38229 http://dx.doi.org/10.1038/38229 S.C. Clemens et al. Direct correlation of Chinese speleothem δ18O and South China Sea planktonic δ18O: transfer- ring a speleothem chronology to the benthic marine chronology. Pale- oceanography, 27(2):PA2203. Administrative preparation for operations within Indian EEZ waters During Expedition 353, 4280 m of sediment section was recov- ered during 32.9 days of on-site operations. One hole was logged. IODP Proceedings 27 Volume 353 http://dx.doi.org/10.1029/2007PA001564 Cheng, H., Edwards, R.L., Broecker, W.S., Denton, G.H., Kong, X., Wang, Y., Zhang, R., and Wang, X., 2009. Ice age terminations. Science, 326(5950):248–252. http://dx.doi.org/10.1126/science.1177840 Antonov, J.I., Seidov, D., Boyer, T.P., Locarnini, R.A., Mishonov, A.V., Garcia, H.E., Baranova, O.K., Zweng, M.M., and Johnson, D.R., 2010. World Ocean Atlas 2009 (Volume 2): Salinity. In Levitus, S. (Ed.), NOAA Atlas NESDIS 69: Washington D.C. (U.S. Government Printing Office). ftp://ftp.nodc.noaa.gov/pub/WOA09/DOC/woa09_vol2_text_fig- ures.pdf Chesner, C.A., and Rose, W.I., 1991. Stratigraphy of the Toba tuffs and the evolution of the Toba caldera complex, Sumatra, Indonesia. Bulletin of Volcanology, 53(5):343–356. http://dx.doi.org/10.1007/BF00280226 Clemens, S.C., 2005. Millennial-band climate spectrum resolved and linked to centennial-scale solar cycles. Quaternary Science Reviews, 24(5–6):521– 531. http://dx.doi.org/10.1016/j.quascirev.2004.10.015 Balmaseda, M.A., Mogensen, K., and Weaver, A.T., 2012. Evaluation of the ECMWF ocean reanalysis system ORAS4. Quarterly Journal of the Royal Meteorological Society, 139(674):1132–1161. http://dx.doi.org/10.1002/qj.2063 g Clemens, S.C., Kuhnt, W., and LeVay, L.J., 2014. Expedition 353 Scientific Pro- spectus: iMonsoon. International Ocean Discovery Program. Clemens, S.C., Kuhnt, W., and LeVay, L.J., 2014. Expedition 353 Scientific Pro- spectus: iMonsoon. International Ocean Discovery Program. http://dx.doi.org/10.14379/iodp.sp.353.2014 Bemis, B.E., Spero, H.J., Bijma, J., and Lea, D.W., 1998. Reevaluation of the oxygen isotopic composition of planktonic foraminifera: experimental results and revised paleotemperature equations. Paleoceanography, 13(2):150–160. http://dx.doi.org/10.1029/98PA00070 http://dx.doi.org/10.14379/iodp.sp.353.2014 Clemens, S.C., Kuhnt, W., LeVay, L.J., Anand, P., Ando, T., Bartol, M., Bolton, C.T., Ding, X., Gariboldi, K., Giosan, L., Hathorne, E.C., Huang, Y., Jaiswal, P., Kim, S., Kirkpatrick, J.B., Littler, K., Marino, G., Martinez, P., 28 Volume 353 IODP Proceedings S.C. Clemens et al. Expedition 353 summary Naik, D., Peketi, A., Phillips, S.C., Robinson, M.M., Romero, O.E., Sagar, N., Taladay, K.B., Taylor, S.N., Thirumalai, K., Uramoto, G., Usui, Y., Wang, J., Yamamoto, M., and Zhou, L., 2016. Site U1448. In Clemens, S.C., Kuhnt, W., LeVay, L.J., and the Expedition 353 Scientists, Indian Monsoon Rainfall. Proceedings of the International Ocean Discovery Pro- gram, 353: College Station, TX (International Ocean Discovery Program). http://dx.doi.org/10.14379/iodp.proc.353.108.2016 Naik, D., Peketi, A., Phillips, S.C., Robinson, M.M., Romero, O.E., Sagar, N., Taladay, K.B., Taylor, S.N., Thirumalai, K., Uramoto, G., Usui, Y., Dickens, G.R., and Owen, R.M., 1999. The latest Miocene–early Pliocene bio- genic bloom: a revised Indian Ocean perspective. Marine Geology, 161(1):75–91. http://dx.doi.org/10.1016/S0025-3227(99)00057-2 Ding, Y., and Chan, J.C.L., 2005. The East Asian summer monsoon: an over- view. Meteorology and Atmospheric Physics, 89(1–4):117–142. http://dx.doi.org/10.1007/s00703-005-0125-z http://dx.doi.org/10.14379/iodp.proc.353.108.2016 Ding, Y., Li, C., and Liu, Y., 2004. Overview of the South China Sea monsoon experiment. http://dx.doi.org/10.1029/2008PA001638 Clift, P.D., 2006. Controls on the erosion of Cenozoic Asia and the flux of clas- tic sediment to the ocean. Earth and Planetary Science Letters, 241(3– 4):571–580. http://dx.doi.org/10.1016/j.epsl.2005.11.028 http://dx.doi.org/10.2973/odp.proc.sr.138.143.1995 Federov, L.V., Ravich, M.G., and Hofmann, J., 1982. Geologic comparison of southeastern peninsular India and Sri Lanka with a part of East Antarc- tica (Enderby Land, MacRobertson Land, and Princess Elizabeth Land). In Craddock, C. (Ed.), Antarctic Geoscience: Madison, WI (University of Wisconsin Press), 73–78. Clift, P.D., Hodges, K.V., Heslop, D., Hannigan, R., Long, H.V., and Calves, G., 2008. Correlation of Himalayan exhumation rates and Asian monsoon intensity. Nature Geoscience, 1(12):875–880. Flores, J.A., Johnson, J.E., Mejía-Molina, A.E., Álverez, M.C., Sierro, F.J., Singh, S.D., Mahanti, S., and Giosan, L., 2014. Sedimentation rates from calcareous nannofossil and planktonic foraminifera biostratigraphy in the Andaman Sea, northern Bay of Bengal, and eastern Arabian Sea. Marine and Petroleum Geology, 58(Part A):425–437. http://dx.doi.org/10.1016/j.marpetgeo.2014.08.011 http://dx.doi.org/10.1038/ngeo351 Clift, P.D., and Plumb, R.A., 2008. The Asian Monsoon: Causes, History and Effects: Cambridge, United Kingdom (Cambridge University Press). http://dx.doi.org/10.1017/CBO9780511535833 Colin, C., Kissel, C., Blamart, D., and Turpin, L., 1998. Magnetic properties of sediments in the Bay of Bengal and the Andaman Sea: impact of rapid North Atlantic Ocean climatic events on the strength of the Indian mon- soon. Earth and Planetary Science Letters, 160(3–4):623–635. http://dx.doi.org/10.1016/S0012-821X(98)00116-2 http://dx.doi.org/10.1016/j.marpetgeo.2014.08.011 France-Lanord, C., Schwenk, T., and Klaus, A., 2014. Expedition 354 Scientific Prospectus: Bengal Fan. International Ocean Discovery Program. http://dx.doi.org/10.14379/iodp.sp.354.2014 http://dx.doi.org/10.14379/iodp.sp.354.2014 Frank, M., Whiteley, N., van de Flierdt, T., Reynolds, B.C., and O’Nions, K., 2006. Nd and Pb isotope evolution of deep water masses in the eastern Indian Ocean during the past 33 Myr. Chemical Geology, 226(3–4):264– 279. http://dx.doi.org/10.1016/j.chemgeo.2005.09.024 Collett, T.S., Riedel, M., Cochran, J., Boswell, R., Presley, J., Kumar, P., Sathe, A.V., Sethi, A., Lall, M., Sibal, V., and the NGHP Expedition 01 Scientists, 2008. Indian National Gas Hydrate Program (NGHP) Expedition 01, Ini- tial Report: Noida, India (DGH, Ministry of Petroleum and Natural Gas). Frerichs, W.E., 1971. Planktonic foraminifera in the sediments of the Anda- man Sea. Journal of Foraminiferal Research, 1(1):1–14. http://dx.doi.org/10.2113/gsjfr.1.1.1 Cullen, J.L., and Prell, W.L., 1984. Planktonic foraminifera of the northern Indian Ocean: distribution and preservation in surface sediments. Marine Micropaleontology, 9(1):1–52. http://dx.doi.org/10.1016/0377-8398(84)90022-7 http://dx.doi.org/10.1029/2010PA001926 Clemens, S.C., Prell, W.L., Sun, Y., Liu, Z., and Chen, G., 2008. Southern Hemisphere forcing of Pliocene δ18O and the evolution of Indo-Asian monsoons. Paleoceanography, 23(4):PA4210. http://dx.doi.org/10.1029/2008PA001638 http://dx.doi.org/10.1016/0377-8398(84)90022-7 Gourlan, A.T., Meynadier, L., and Allègre, C.J., 2008. Tectonically driven changes in the Indian Ocean circulation over the last 25 Ma: neodymium isotope evidence. Earth and Planetary Science Letters, 267(1–2):353–364. http://dx.doi.org/10.1016/j.epsl.2007.11.054 Curray, J.R., 1991. Possible greenschist metamorphism at the base of a 22-km sedimentary section, Bay of Bengal. Geology, 19(11):1097–1100. http://dx.doi.org/10.1130/0091- 7613(1991)019<1097:PGMATB>2.3.CO;2 http://dx.doi.org/10.1029/2007PA001564 Advances in Atmospheric Sciences, 21(3):343–360. http://dx.doi.org/10.1007/BF02915563 Clemens, S.C., Murray, D.W., and Prell, W.L., 1996. Nonstationary phase of the Plio-Pleistocene Asian monsoon. Science, 274(5289):943–948. http://dx.doi.org/10.1126/science.274.5289.943 Emile-Geay, J., Cane, M.A., Naik, N., Seager, R., Clement, A.C., and van Geen, A., 2003. Warren revisited: atmospheric freshwater fluxes and “why is no deep water formed in the North Pacific?” Journal of Geophysical Research: Oceans, 108(C6):3178. http://dx.doi.org/10.1029/2001JC001058 Clemens, S.C., and Prell, W.L., 2003. A 350,000 year summer-monsoon multi- proxy stack from the Owen Ridge, northern Arabian Sea. Marine Geology, 201(1–3):35–51. http://dx.doi.org/10.1016/S0025-3227(03)00207-X ceans, 108(C6):3178. http://dx.doi.org/10.1029/2001JC001 Clemens, S.C., and Prell, W.L., 2007. The timing of orbital-scale Indian mon- soon changes. Quaternary Science Reviews, 26(3–4):275–278. http://dx.doi.org/10.1016/j.quascirev.2006.11.010 Emmel, F.J., and Curray, J.R., 1984. Bengal Fan, Indian Ocean. In Bouma, A.H., Normark, W.P., and Barnes, N.E. (Eds.), Submarine Fans and Related Sys- tems: New York (Springer-Verlag), 248–275. http://dx.doi.org/10.1016/j.quascirev.2006.11.010 Farrell, J.W., and Janecek, T.R., 1991. Late Neogene paleoceanography and paleoclimatology of the northeast Indian Ocean (Site 758). In Weissel, J., Peirce, J., Taylor, E., Alt, J., et al., Proceedings of the Ocean Drilling Pro- gram, Scientific Results, 121: College Station, TX (Ocean Drilling Pro- gram), 297–355. http://dx.doi.org/10.2973/odp.proc.sr.121.124.1991 Clemens, S.C., Prell, W., Murray, D., Shimmield, G., and Weedon, G., 1991. Forcing mechanisms of the Indian Ocean monsoon. Nature, 353(6346):720–725. http://dx.doi.org/10.1038/353720a0 Clemens, S.C., Prell, W.L., and Sun, Y., 2010. Orbital-scale timing and mecha- nisms driving Late Pleistocene Indo-Asian summer monsoons: reinter- preting cave speleothem δ18O. Paleoceanography, 25(4):PA4207. http://dx.doi.org/10.1029/2010PA001926 Farrell, J.W., Raffi, I., Janecek, T.R., Murray, D.W., Levitan, M., Dadey, K.A., Emeis, K.-C., Lyle, M., Flores, J.-A., and Hovan, S., 1995. Late Neogene sedimentation patterns in the eastern equatorial Pacific Ocean. In Pisias, N.G., Mayer, L.A., Janecek, T.R., Palmer-Julson, A., and van Andel, T.H. (Eds.), Proceedings of the Ocean Drilling Program, Scientific Results, 138: College Station, TX (Ocean Drilling Program), 717–756. http://dx.doi.org/10.2973/odp.proc.sr.138.143.1995 7613(1991)019<1097:PGMATB>2.3.CO;2 Gourlan, A.T., Meynadier, L., Allègre, C.J., Tapponnier, P., Birck, J.-L., and Joron, J.-L., 2010. Northern Hemisphere climate control of the Bengali rivers discharge during the past 4 Ma. Quaternary Science Reviews, 29(19–20):2484–2498. http://dx.doi.org/10.1016/j.quasci- Dehn, J., Farrell, J.W., and Schmincke, H.-U., 1991. Neogene tephrochronol- ogy from Site 758 on northern Ninetyeast Ridge: Indonesian arc volca- nism of the past 5 Ma. In Weissel, J., Peirce, J., Taylor, E., Alt, J., et al., Proceedings of the Ocean Drilling Program, Scientific Results, 121: College Station, TX (Ocean Drilling Program), 273–295. http://dx.doi.org/10.2973/odp.proc.sr.121.123.1991 Guo, Z.T., Ruddiman, W.F., Hao, Q.Z., Wu, H.B., Qiao, Y.S., Zhu, R.X., Peng, S.Z., Wei, J.J., Yuan, B.Y., and Liu, T.S., 2002. Onset of Asian desertifica- tion by 22 Myr ago inferred from loess deposits in China. Nature, 416(6877):159–163. http://dx.doi.org/10.1038/416159a Dettman, D.L., Kohn, M.J., Quade, J., Ryerson, F.J., Ojha, T.P., and Hamidullah, S., 2001. Seasonal stable isotope evidence for a strong Asian monsoon throughout the past 10.7 m.y. Geology, 29(1):31–34. http://dx.doi.org/10.1130/0091- 7613(2001)029<0031:SSIEFA>2.0.CO;2 Gupta, A.K., Singh, R.K., Joseph, S., and Thomas, E., 2004. Indian Ocean high-productivity event (10–8 Ma): linked to global cooling or to the initi- ation of the Indian monsoons? Geology, 32(9):753–756. http://dx.doi.org/10.1130/G20662.1 IODP Proceedings 29 Volume 353 S.C. Clemens et al. Expedition 353 summary Gupta, A.K., Singh, R.K., and Verma, S., 2013. Deep-sea palaeoceanographic evolution of the eastern Indian Ocean during the late Oligocene–Pleisto- cene: species diversity trends in benthic foraminifera. Current Science, 104(7):904–910. htt // t i i /V l /104/07/0904 df Jensen, T.G., 2001. Arabian Sea and Bay of Bengal exchange of salt and tracers in an ocean model. Geophysical Research Letters, 28(20):3967–3970. http://dx.doi.org/10.1029/2001GL013422 Jensen, T.G., 2003. Cross-equatorial pathways of salt and tracers from the northern Indian Ocean: modeling results. Deep Sea Research, Part II, 50(12–13):2111–2127. http://dx.doi.org/10.1016/S0967-0645(03)00048-1 http://www.currentscience.ac.in/Volumes/104/07/0904.pdf Gupta, A.K., and Thomas, E., 2003. Initiation of Northern Hemisphere glacia- tion and strengthening of the northeast Indian monsoon: Ocean Drilling Program Site 758, eastern equatorial Indian Ocean. Geology, 31(1):47–50. http://dx.doi.org/10.1130/0091-7613(2003)031<0047:ION- HGA>2.0.CO;2 http://dx.doi.org/10.1016/S0967-0645(03)00048-1 Johnson, J.E., Phillips, S.C., Torres, M.E., Piñero, E., Rose, K.K., and Giosan, L., 2014. Influence of total organic carbon deposition on the inventory of gas hydrate in the Indian continental margins. Marine and Petroleum Geol- ogy, 58(Part A):406–424. Guptha, M.V.S., Curry, W.B., Ittekkot, V., and Muralinath, A.S., 1997. Seasonal variation in the flux of planktonic foraminifera: sediment trap results from the Bay of Bengal, northern Indian Ocean. Journal of Foraminiferal Research, 27(1):5–19. http://dx.doi.org/10.1130/G34890.1 Liu, X., Liu, Z., Kutzbach, J.E., Clemens, S.C., and Prell, W.L., 2006. Hemi- spheric insolation forcing of the Indian Ocean and Asian monsoon: local versus remote impacts. Journal of Climate, 19(23):6195–6208. http://dx.doi.org/10.1175/JCLI3965.1 Holbourn, A., Kuhnt, W., Schulz, M., Flores, J.-A., and Andersen, N., 2007. Orbitally-paced climate evolution during the middle Miocene “Monte- rey” carbon-isotope excursion. Earth and Planetary Science Letters, 261(3–4):534–550. http://dx.doi.org/10.1016/j.epsl.2007.07.026 Hoogakker, B.A.A., Rohling, E.J., Palmer, M.R., Tyrrell, T., and Rothwell, R.G., 2006. Underlying causes for long-term global ocean δ13C fluctuations over the last 1.20 Myr. Earth and Planetary Science Letters, 248(1–2):15–29. http://dx.doi.org/10.1016/j.epsl.2006.05.007 Locarnini, R.A., Mishonov, A.V., Antonov, J.I., Boyer, T.P., Garcia, H.E., Baranova, O.K., Zweng, M.M., and Johnson, D.R., 2010. World Ocean Atlas 2009 (Volume 1): Temperature. In Levitus, S. (Ed.), NOAA Atlas NESDIS 68: Washington, DC (U.S. Government Printing Office). ftp://ftp.nodc.noaa.gov/pub/WOA09/DOC/woa09_vol1_text_ Hovan, S., and Rea, D.K., 1992. The Cenozoic record of continental mineral deposition on Broken and Ninetyeast Ridges, Indian Ocean: southern African aridity and sediment delivery from the Himalayas. Paleoceanog- raphy, 7(6):833–860. http://dx.doi.org/10.1029/92PA02176 http://dx.doi.org/10.1016/S1367-9120(01)00069-4 Hall, R., Cottam, M.A., and Wilson, M.E.J., 2011. The SE Asian Gateway: His- tory and Tectonics of the Australia–Asia Collision. Special Publication - Geological Society of London, 355. http://dx.doi.org/10.1144/SP355.1 Le Houedec, S., Meynadier, L., Cogné, J.-P., Allègre, C.J., and Gourlan, A.T., 2012. Oceanwide imprint of large tectonic and oceanic events on sea- water Nd isotope composition in the Indian Ocean from 90 to 40 Ma. Geochemistry, Geophysics, Geosystems, 13(6):Q06008. htt //d d i /10 1029/2011GC003963 Hastenrath, S., and Greischar, L., 1993. The monsoonal heat budget of the hydrosphere atmosphere system in the Indian Ocean sector. Journal of Geophysical Research: Oceans, 98(C4):6869–6881. http://dx.doi.org/10.1029/92JC02956 http://dx.doi.org/10.1029/2011GC003963 Licht, A., van Cappelle, M., Abels, H.A., Ladant, J.-B., Trabucho-Alexandre, J., France-Lanord, C., Donnadieu, Y., Vandenberghe, J., Rigaudier, T., Lécuyer, C., Terry, D., Jr., Adriaens, R., Boura, A., Guo, Z., Soe, A.N., Quade, J., Dupont-Nivet, G., and Jaeger, J.-J., 2014. Asian monsoons in a late Eocene greenhouse world. Nature, 513(7519):501–506. http://dx.doi.org/10.1038/nature13704 Hermoyian, C.S., and Owen, R.M., 2001. Late Miocene–early Pliocene bio- genic bloom: evidence from low-productivity regions of the Indian and Atlantic Oceans. Paleoceanography, 16(1):95–100. http://dx.doi.org/10.1029/2000PA000501 Holbourn, A., Kuhnt, W., Clemens, S., Prell, W., and Andersen, N., 2013. Mid- dle to late Miocene stepwise climate cooling: evidence from a high-reso- lution deep water isotope curve spanning 8 million years. Paleoceanography, 28(4):688–699. http://dx.doi.org/10.1002/2013PA002538 Liu, W.T., and Tang, W., 2004. Oceanic influence on the precipitation in India and China as observed by TRMM and QuikSCAT [paper presented at 2nd TRMM International Science Conference, Nara, Japan, 6–10 Sep- tember 2004]. http://www.eorc.jaxa.jp/TRMM/museum/event/2nd- TISC/HP/Extended Abstract/4.10_LIU_Timothy.pdf http://dx.doi.org/10.1002/2013PA002538 Holbourn, A., Kuhnt, W., Kochhann, K.G.D., Andersen, N., and Meier, S., 2015. Global perturbation of the carbon cycle at the onset of the Miocene Climatic Optimum. Geology, 43(2):123–126. http://dx.doi.org/10.1130/G36317.1 Liu, W.T., and Tang, W., 2005. Estimating moisture transport over oceans using space-based observations. Journal of Geophysical Research: Atmo- spheres, 110(D10):D10101. http://dx.doi.org/10.1029/2004JD005300 http://dx.doi.org/10.1130/G36317.1 Liu, W.T., Zhang, A., and Bishop, J.K.B., 1994. Evaporation and solar irradi- ance as regulators of sea surface temperature in annual and interannual changes. Journal of Geophysical Research: Oceans, 99(C6):12623–12637. http://dx.doi.org/10.1029/94JC00604 Holbourn, A., Kuhnt, W., Lyle, M., Schneider, L., Romero, O., and Anderson, N., 2013. Middle Miocene climate cooling linked to intensification of eastern equatorial Pacific upwelling. Geology, 42(1):19–22. http://dx.doi.org/10.1130/G34890.1 7613(1991)019<1097:PGMATB>2.3.CO;2 http://dx.doi.org/10.2113/gsjfr.27.1.5 http://dx.doi.org/10.1016/j.marpetgeo.2014.08.021 Kroon, D., Steens, T., and Troelstra, S.R., 1991. Onset of monsoonal related upwelling in the western Arabian Sea as revealed by planktonic foramini- fers. In Prell, W.L., Niitsuma, N., et al., Proceedings of the Ocean Drilling Program, Scientific Results, 117: College Station, TX (Ocean Drilling Pro- gram), 257–263. http://dx.doi.org/10.2973/odp.proc.sr.117.126.1991 Hall, C.M., and Farrell, J.W., 1993. Laser 40Ar/39Ar age from Ash D of ODP Site 758: dating the Brunhes-Matuyama reversal and oxygen isotope Stage 19.1. Eos, Transactions of the American Geophysical Union, 74(Supple- ment):110. Kuhnt, W., Holbourn, A., Hall, R., Zuvela, M., and Käse, R., 2004. Neogene history of the Indonesian Throughflow. In Clift, P., Wang, P., Kuhnt, W., and Hayes, D. (Eds.), Continent-Ocean Interactions within East Asian Marginal Seas. Geophysical Monograph, 149:299–320. http://dx.doi.org/10.1029/149GM16 Hall, C.M., and Farrell, J.W., 1995. Laser 40Ar/39Ar ages of tephra from Indian Ocean deep-sea sediments: tie points for the astronomical and geomag- netic polarity time scales. Earth and Planetary Science Letters, 133(3– 4):327–338. http://dx.doi.org/10.1016/0012-821X(95)00087-S p y y 4):327–338. http://dx.doi.org/10.1016/0012-821X(95)00087-S Hall, R., 2002. Cenozoic geological and plate tectonic evolution of SE Asia and the SW Pacific: computer-based reconstructions, model and animations. Journal of Asian Earth Sciences, 20(4):353–431. http://dx.doi.org/10.1016/S1367-9120(01)00069-4 Kumar, S.P., Muraleedharan, P.M., Prasad, T.G., Gauns, M., Ramaiah, N., de Souza, S.N., Sardesai, S., and Madhupratap, M., 2002. Why is the Bay of Bengal less productive during summer monsoon compared to the Ara- bian Sea? Geophysical Research Letters, 29(24):2235. http://dx.doi.org/10.1029/2002GL016013 figures.pdf Loschnigg, J., and Webster, P.J., 2000. A coupled ocean–atmosphere system of SST modulation for the Indian Ocean. Journal of Climate, 13(19):3342– 3360. http://dx.doi.org/10.1175/1520- 0442(2000)013<3342:ACOASO>2.0.CO;2 Huang, Y., Clemens, S.C., Liu, W., Wang, Y., and Prell, W.L., 2007. Large-scale hydrological change drove the late Miocene C4 plant expansion in the Himalayan foreland and Arabian Peninsula. Geology, 35(6):531–534. http://dx.doi.org/10.1130/G23666A.1 Mantyla, A.W., and Reid, J.L., 1995. On the origins of deep and bottom waters of the Indian Ocean. Journal of Geophysical Research: Oceans, 100(C2):2417–2439. http://dx.doi.org/10.1029/94JC02564 IODP Proceedings 30 Volume 353 Volume 353 S.C. Clemens et al. Expedition 353 summary Misra, S., and Froelich, P.N., 2012. Lithium isotope history of Cenozoic sea- water: changes in silicate weathering and reverse weathering. Science, 335(6070):818–823. http://dx.doi.org/10.1126/science.1214697 Ramana, M.V., Ramprasad, T., Collett, T.S., Kumar, P., Boswell, R.M., Riedel, M., Sathe, A.V., Lall, M., Vishwanath, K., Muzumdar, A., Sain, K., and Cochran, J. (Eds.), 2014. Geologic Implications of Gas Hydrates in the Off- shore of India: Results of the National Gas Hydrate Program Expedition 01, Marine and Petroleum Geology, 58(Part A). http://www.science direct.com/science/journal/02648172/58/supp/PA Molnar, P., 2005. Mio–Pliocene growth of the Tibetan Plateau and evolution of East Asian climate. Palaeontologia Electronica, 8(1):1–23. http://palaeo-electronica.org/2005_1/molnar2/molnar2.pdf Molnar, P., Boos, W.R., and Battisti, D.S., 2010. Orographic controls on cli- mate and paleoclimate of Asia: thermal and mechanical roles for the Tibetan Plateau. Annual Review of Earth and Planetary Sciences, 38(1):77–102. http://dx.doi.org/10.1146/annurev-earth-040809-152456 Randall, D.A., Wood, R.A., Bony, S., Colman, R., Fichefet, T., Fyfe, J., Kattsov, V., Pitman, A., Shukla, J., Srinivasan, J., Stouffer, R.J., Sumi, A., and Taylor, K.E., 2007. Climate models and their evaluation. In Solomon, S., Qin, D., Manning, M., Chen, Z., Marquis, M., Averyt, K.B., Tingor, M., and Miller, H.L. (Eds.), Climate Change 2007: The Physical Science Basis. Contribu- tion of Working Group I to the Fourth Assessment Report of the Intergov- ernmental Panel on Climate Change: Cambridge, UK (Cambridge University Press), 589–662. http://www.ipcc.ch/pdf/assessment- report/ar4/wg1/ar4-wg1-chapter8.pdf Mudelsee, M., and Raymo, M.E., 2005. Slow dynamics of the Northern Hemi- sphere glaciation. Paleoceanography, 20(4):PA4022. http://dx.doi.org/10.1029/2005PA001153 Nicholls, R.J., Wong, P.P., Burkett, V., Codignotto, J., Hay, J., McLean, R., Ragoonaden, S., and Woodroffe, C.D., 2007. Coastal systems and low- lying areas. In Parry, M.L., Canziani, O.F., Palutikof, J.P., van der Linden, P.J., and Hanson, C.E. (Eds.), Climate Change 2007: Impacts, Adaptation and Vulnerability. http://dx.doi.org/10.1029/2006PA001384 Schott, F.A., and McCreary, J.P., Jr., 2001. The monsoon circulation of the Indian Ocean. Progress in Oceanography, 51(1):1–123. Schott, F.A., and McCreary, J.P., Jr., 2001. The monsoon circulation of the Indian Ocean. Progress in Oceanography, 51(1):1–123. http://dx.doi.org/10.1016/S0079-6611(01)00083-0 Powell, C.McA., Roots, S.R., and Veevers, J.J., 1988. Pre-breakup continental extension in East Gondwanaland and the early opening of the eastern Indian Ocean. Tectonophysics, 155(1–4):261–283. http://dx.doi.org/10.1016/0040-1951(88)90269-7 http://dx.doi.org/10.1016/S0079-6611(01)00083-0 Schott, F.A., Xie, S.-P., and McCreary, J.P.,Jr., 2009. Indian Ocean circulation and climate variability. Reviews of Geophysics, 47(1):RG1002. http://dx.doi.org/10.1029/2007RG000245 figures.pdf Contribution of Working Group II to the Fourth Assess- ment Report of the Intergovernmental Panel on Climate Change: Cam- bridge, United Kingdom (Cambridge University Press), 315–356. https://www.ipcc.ch/pdf/assessment-report/ar4/wg2/ar4-wg2- chapter6.pdf report/ar4/wg1/ar4-wg1-chapter8.pdf Rayner, N.A., Parker, D.E., Horton, E.B., Folland, C.K., Alexander, L.V., Rowell, D.P., Kent, E.C., and Kaplan, A., 2003. Global analyses of sea surface tem- perature, sea ice, and night marine air temperature since the late nine- teenth century. Journal of Geophysical Research: Atmospheres, 108(D14):4407. http://dx.doi.org/10.1029/2002JD002670 Rodriguez, M., Chamot-Rooke, N., Huchon, P., Fournier, M., and Delescluse, M., 2014. The Owen Ridge uplift in the Arabian Sea: implications for the sedimentary record of Indian monsoon in late Miocene. Earth and Plane- tary Science Letters, 394:1–12. http://dx.doi.org/10.1016/j.epsl.2014.03.011 Ninkovich, D., 1979. Distribution, age and chemical composition of tephra layers in deep-sea sediments off western Indonesia. Journal of Volcanol- ogy and Geothermal Research, 5(1–2):67–86. http://dx.doi.org/10.1016/0377-0273(79)90033-7 Royer, J.-Y., Peirce, J.W., and Weissel, J.K., 1991. Tectonic constraints on the hotspot formation of the Ninetyeast Ridge. In Weissel, J., Peirce, J., Taylor, E., Alt, J., et al., Proceedings of the Ocean Drilling Program, Scientific Results, 121: College Station, TX (Ocean Drilling Program), 763–776. http://dx.doi.org/10.2973/odp.proc.sr.121.122.1991 http://dx.doi.org/10.1016/0377-0273(79)90033-7 Ninkovich, D., Shackelton, N.J., Abdel-Monem, A.A., Obradovich, J.D., and Izett, G., 1978. K-Ar age of the Late Pleistocene eruption of Toba, north Sumatra. Nature, 276(5688):574–577. http://dx.doi.org/0.1038/276574a0 Ruddiman, W.F., 2006. What is the timing of orbital-scale monsoon changes? Quaternary Science Reviews, 25(7–8):657–658. http://dx.doi.org/10.1016/j.quascirev.2006.02.004 http://dx.doi.org/10.1029/2008GC002332 Phillips, S.C., Johnson, J.E., Giosan, L., and Rose, K., 2014a. Monsoon-influ- enced variation in productivity and lithogenic sediment flux since 110 ka in the offshore Mahanadi Basin, northern Bay of Bengal. Marine and Petroleum Geology, 58(Part A):502–525. http://dx.doi.org/10.1016/j.marpetgeo.2014.05.007 Sager, W.W., Paul, C.F., Krishna, K.S., Pringle, M., Eisin, A.E., Frey, F.A., Gopala Rao, D., and Levchenko, O., 2010. Large fault fabric of the Ninet- yeast Ridge implies near-spreading ridge formation. Geophysical Research Letters, 37(17):L17304. http://dx.doi.org/10.1029/2010GL044347 http://dx.doi.org/10.1016/j.marpetgeo.2014.05.007 Schlitzer, R., 2000. Electronic atlas of WOCE hydrographic and tracer data now available. Eos, Transactions American Geophysical Union, 81(5):45. http://dx.doi.org/10.1029/00EO00028 Phillips, S.C., Johnson, J.E., Underwood, M.B., Guo, J., Giosan, L., and Rose, K., 2014b. Long-timescale variation in bulk and clay mineral composition of Indian continental margin sediments in the Bay of Bengal, Arabian Sea, and Andaman Sea. Marine and Petroleum Geology, 58(Part A):117–138. http://dx.doi.org/10.1016/j.marpetgeo.2014.06.018 Schmittner, A., Galbraith, E.D., Hostetler, S.W., Pedersen, T.F., and Zhang, R., 2007. Large fluctuations of dissolved oxygen in the Indian and Pacific Oceans during Dansgaard–Oeschger oscillations caused by variations of North Atlantic Deep Water subduction. Paleoceanography, 22(3):PA3207. http://dx.doi.org/10.1029/2006PA001384 Ponton, C., Giosan, L., Eglinton, T.I., Fuller, D.Q., Johnson, J.E., Kumar, P., and Collett, T.S., 2012. Holocene aridification of India. Geophysical Research Letters, 39(3):L3407. http://dx.doi.org/10.1029/2011GL050722 http://dx.doi.org/10.1016/0040-1951(88)90269-7 Prell, W.L., Murray, D.W., Clemens, S.C., and Anderson, D.M., 1992. Evolu- tion and variability of the Indian Ocean summer monsoon: evidence from the western Arabian Sea drilling program. In Duncan, R.A., Rea, D.K., Kidd, R.B., von Rad, U., and Weissel, J.K. (Eds.), Synthesis of Results from Scientific Drilling in the Indian Ocean. Geophysical Monograph, 70:447– 469. http://dx.doi.org/10.1029/GM070p0447 http://dx.doi.org/0.1038/276574a0 Park, S.-C., Sohn, B.-J., and Wang, B., 2007. Satellite assessment of divergent water vapor transport from NCEP, ERA40, and JRA25 reanalyses over the Asian summer monsoon region. Journal of the Meteorological Society of Japan, 85(5):615–632. http://dx.doi.org/10.2151/jmsj.85.615 Ryan, W.B.F., Carbotte, S.M., Coplan, J.O., O’Hara, S., Melkonian, A., Arko, R., Weissel, R.A., Ferrini, V., Goodwillie, A., Nitsche, F., Bonczkowski, J., and Zemsky, R., 2009. Global multi-resolution topography synthesis. Geo- chemistry, Geophysics, Geosystems, 10(3):Q03014. http://dx.doi.org/10.1029/2008GC002332 Peirce, J., Weissel, J., et al., 1989. Proceedings of the Ocean Drilling Program, Initial Reports, 121: College Station, TX (Ocean Drilling Program). http://dx.doi.org/10.2973/odp.proc.ir.121.1989 http://dx.doi.org/10.1126/science.1064618 Singh, A., Jani, R.A., and Ramesh, R., 2010. Spatiotemporal variations of the δ18O–salinity relation in the northern Indian Ocean. Deep Sea Research, Part I, 57(11):1422–1431. http://dx.doi.org/10.1016/j.dsr.2010.08.002 Webster, P.J., 1987a. The elementary monsoon. In Fein, J.S., and Stephens, P.L. (Eds.), Monsoons: New York (Wiley), 3–32. Webster, P.J., 1987b. The variable and interactive monsoon. In Fein, J.S., and Stephens, P.L. (Eds.), Monsoons: New York (Wiley), 269–330. Singh, A.D., Jung, S.J.A., Darling, K., Ganeshram, R., Ivanochko, T., and Kroon, D., 2011. Productivity collapses in the Arabian Sea during glacial cold phases. Paleoceanography, 26(3):PA3210. http://dx.doi.org/10.1029/2009PA001923 Webster, P.J., 1994. The role of hydrological processes in ocean-atmosphere interactions. Reviews of Geophysics, 32(4):427–476. Webster, P.J., 1994. The role of hydrological processes in ocean-atmosphere interactions. Reviews of Geophysics, 32(4):427–476. http://dx.doi.org/10.1029/94RG01873 http://dx.doi.org/10.1029/94RG01873 Singh, R.K., and Gupta, A.K., 2005. Systematic decline in benthic foramin- iferal species diversity linked to productivity increases over the last 26 Ma in the Indian Ocean. Journal of Foraminiferal Research, 35(3):219–227. http://dx.doi.org/10.2113/35.3.219 Webster, P.J., Magaña, V.O., Palmer, T.N., Shukla, J., Tomas, R.A., Yanai, M., and Yasunari, T., 1998. Monsoons: processes, predictability, and the pros- pects for prediction. Journal of Geophysical Research: Oceans, ( ) h d d 103(C7):14451–14510. http://dx.doi.org/10.1029/97JC02719 Singh, S.C., Moeremans, R., McArdle, J., and Johansen, K., 2013. Seismic images of the sliver strike-slip fault and back thrust in the Andaman- Nicobar region. Journal of Geophysical Research: Solid Earth, 118(10):5208–5224. http://dx.doi.org/10.1002/jgrb.50378 Wyrtki, K., 1971. Oceanographic Atlas of the International Ocean Expedition. National Science Foundation. Xie, P., and Arkin, P.A., 1997. Global precipitation: a 17-year monthly analysis based on gauge observations, satellite estimates, and numerical model outputs. Bulletin of the American Meteorological Society, 78(11):2539– 2558. http://dx.doi.org/10.1175/1520- 0477(1997)078<2539:GPAYMA>2.0.CO;2 Stöckli, R., Vermote, E., Saleous, N., Simmon, R., and Herring, D., 2005. The Blue Marble Next Generation—A true color Earth dataset including sea- sonal dynamics from MODIS. NASA Earth Observatory. http://earthob- servatory.nasa.gov/Features/BlueMarble/bmng.pdf servatory.nasa.gov/Features/BlueMarble/bmng.pdf Stoll, H.M., Vance, D., and Arevalos, A., 2007. Records of the Nd isotope com- position of seawater from the Bay of Bengal: implications for the impact of Northern Hemisphere cooling on ITCZ movement. Earth and Plane- tary Science Letters, 255(1–2):213–228. http://dx.doi.org/10.1016/j.epsl.2006.12.016 Xu, J., Kuhnt, W., Holbourn, A., Andersen, N., and Bartoli, G., 2006. Changes in the vertical profile of the Indonesian Throughflow during Termination II: evidence from the Timor Sea. Paleoceanography, 21(4):PA4202. http://dx.doi.org/10.1016/j.palaeo.2005.03.005 Zachos, J., Pagani, M., Sloan, L., Thomas, E., and Billups, K., 2001. Trends, rhythms, and aberrations in global climate 65 Ma to present. Science, 292(5517):686–693. http://dx.doi.org/10.1126/science.1059412 Sun, Y., Clemens, S.C., Morrill, C., Lin, X., Wang, X., and An, Z., 2011. Influ- ence of Atlantic meridional overturning circulation on the East Asian winter monsoon. Nature Geoscience, 5(1):46–49. http://dx.doi.org/10.1038/ngeo1326 Zachos, J.C., Flower, B.P., and Paul, H., 1997. Orbitally paced climate oscilla- tions across the Oligocene/Miocene boundary. Nature, 388(6642):567– 570. http://dx.doi.org/10.1038/41528 http://dx.doi.org/10.1126/science.1064618 http://dx.doi.org/10.1029/2006PA001278 http://dx.doi.org/10.1029/2006PA001278 Yancheva, G., Nowaczyk, N.R., Mingram, J., Dulski, P., Schettler, G., Negend- ank, J.F.W., Liu, J., Sigman, D.M., Peterson, L.C., and Haug, G.H., 2007. Influence of the intertropical convergence zone on the East Asian mon- soon. Nature, 445(7123):74–77. http://dx.doi.org/10.1038/nature05431 Subrahmanyam, V., Subrahmanyam, A.S., Murty, G.P.S., and Murthy, K.S.R., 2008. Morphology and tectonics of Mahanadi Basin, northeastern conti- nental margin of India from geophysical studies. Marine Geology, 253(1– 2):63–72. http://dx.doi.org/10.1016/j.margeo.2008.04.007 –72. http://dx.doi.org/10.1016/j.margeo.2008.04.007 You, Y., and Tomczak, M., 1993. Thermocline circulation and ventilation in the Indian Ocean derived from water mass analysis. Deep Sea Research, Part I, 40(1):13–56. http://dx.doi.org/10.1016/0967-0637(93)90052-5 Sun, X., and Wang, P., 2005. How old is the Asian monsoon system?—Palaeo- botanical records from China. Palaeogeography, Palaeoclimatology, Palaeoecology, 222(3–4):181–222. http://dx.doi.org/10.1016/j.palaeo.2005.03.005 http://dx.doi.org/10.1016/j.palaeo.2005.03.005 http://dx.doi.org/10.1029/2007RG000245 Schulz, H., von Rad, U., and Erlenkeuser, H., 1998. Correlation between Ara- bian Sea and Greenland climate oscillations of the past 110,000 years. Nature, 393(6680):54–57. http://dx.doi.org/10.1038/31750 Schwenk, T., and Spieß, V., 2009. Architecture and stratigraphy of the Bengal Fan as response to tectonic and climate revealed from high-resolution seismic data. In Kneller, B.C., Martinsen, O.J., and McCaffrey, B. (Eds.), External Controls on Deep-Water Depositional Systems. Special Publica- tion - SEPM (Society of Sedimentary Geologists), 92:107–131. Quade, J., and Cerling, T.E., 1995. Expansion of C4 grasses in the late Miocene of northern Pakistan: evidence from stable isotopes in paleosols. Palaeo- geography, Palaeoclimatology, Palaeoecology, 115(1–4):91–116. http://dx.doi.org/10.1016/0031-0182(94)00108-K 31 IODP Proceedings Volume 353 S.C. Clemens et al. Expedition 353 summary Scotese, C.R., Gahagan, L.M., and Larson, R.L., 1988. Plate tectonic recon- structions of the Cretaceous and Cenozoic ocean basins. Tectonophysics, 155(1–4):27–48. http://dx.doi.org/10.1016/0040-1951(88)90259-4 Wajsowicz, R.C., and Schopf, P.S., 2001. Oceanic influences on the seasonal cycle in evaporation over the Indian Ocean. Journal of Climate, 14(6):1199–1226. http://dx.doi.org/10.1175/1520- 0442(2001)014<1199:OIOTSC>2.0.CO;2 Sharmila, S., Joseph, S., Sahai, A.K., Abhilash, S., and Chattopadhyay, R., 2015. Future projection of Indian summer monsoon variability under climate change scenario: an assessment from CMIP5 climate models. Global and Planetary Change, 124:62–78. http://dx.doi.org/10.1016/j.gloplacha.2014.11.004 Wang, P., Clemens, S., Beaufort, L., Braconnot, P., Ganssen, G., Jian, Z., Ker- shaw, P., and Sarnthein, M., 2005. Evolution and variability of the Asian monsoon system: state of the art and outstanding issues. Quaternary Sci- ence Review, 24(5–6):595–629. htt //d d i /10 1016/j i 2004 10 002 http://dx.doi.org/10.1016/j.gloplacha.2014.11.004 Shipboard Scientific Party, 1989. Site 758. In Peirce, J., Weissel, J., et al., Pro- ceedings of the Ocean Drilling Program, Initial Reports, 121: College Sta- tion, TX (Ocean Drilling Program), 359–453. http://dx.doi.org/10.2973/odp.proc.ir.121.112.1989 http://dx.doi.org/10.1016/j.quascirev.2004.10.002 Wang, Y., Cheng, H., Edwards, R.L., Kong, X., Shao, X., Chen, S., Wu, J., Jiang, X., Wang, X., and An, Z., 2008. Millennial- and orbital-scale changes in the East Asian monsoon over the past 224,000 years. Nature, 451(7182):1090–1093. http://dx.doi.org/10.1038/nature06692 http://dx.doi.org/10.2973/odp.proc.ir.121.112.1989 Simmonds, I., Bi, D., and Hope, P., 1999. Atmospheric water vapor flux and its association with rainfall over China in summer. Journal of Climate, 12(5):1353–1367. http://dx.doi.org/10.1175/1520- 0442(1999)012<1353:AWVFAI>2.0.CO;2 1(7182):1090–1093. http://dx.doi.org/10.1038/nature0669 Wang, Y.J., Cheng, H., Edwards, R.L., An, Z.S., Wu, J.Y., Shen, C.-C., and Dorale, J.A., 2001. A high-resolution absolute-dated late Pleistocene monsoon record from Hulu Cave, China. Science, 294(5580):2345–2348. http://dx.doi.org/10.1126/science.1064618 http://dx.doi.org/10.1038/ngeo1326 Thomas, D.J., Bralower, T.J., and Jones, C.E., 2003. Neodymium isotopic reconstruction of late Paleocene–early Eocene thermohaline circulation. Earth and Planetary Science Letters, 209(3–4):309–322. http://dx.doi.org/10.1016/S0012-821X(03)00096-7 Zhu, Y., and Newell, R.E., 1998. A proposed algorithm for moisture fluxes from atmospheric rivers. Monthly Weather Review, 126(3):725–735. http://dx.doi.org/10.1175/1520- 0493(1998)126<0725:APAFMF>2.0.CO;2 Zhu, Y., and Newell, R.E., 1998. A proposed algorithm for moisture fluxes from atmospheric rivers. Monthly Weather Review, 126(3):725–735. http://dx.doi.org/10.1175/1520- 0493(1998)126<0725:APAFMF>2.0.CO;2 Tomczak, M., and Godfrey, J.S., 2003. Regional Oceanography: An Introduc- tion (2nd ed.): Delhi (Daya Publishing House). Ziegler, M., Lourens, L.J., Tuenter, E., Hilgen, F., Reichart, G.-J., and Weber, N., 2010. Precession phasing offset between Indian summer monsoon and Arabian Sea productivity linked to changes in Atlantic overturning circu- lation. Paleoceanography, 25(3):PA3213. Ziegler, M., Lourens, L.J., Tuenter, E., Hilgen, F., Reichart, G.-J., and Weber, N., 2010. Precession phasing offset between Indian summer monsoon and Arabian Sea productivity linked to changes in Atlantic overturning circu- lation. Paleoceanography, 25(3):PA3213. http://dx.doi.org/10.1029/2009PA001884 Varkey, M.J., Murty, V.S.N., and Suryanarayana, A., 1996. Physical oceanogra- phy of the Bay of Bengal and Andaman Sea. Oceanography and Marine Biology, 34:1–70. http://drs.nio.org/drs/handle/2264/2276 von der Borch, C.C., Sclater, J.G., et al., 1974. Initial Reports of the Deep Sea Drilling Project, 22: Washington, DC (U.S. Government Printing Office). http://dx.doi.org/10.2973/dsdp.proc.22.1974 32 Volume 353 IODP Proceedings
https://openalex.org/W1899724582	https://eprints.kingston.ac.uk/id/eprint/35259/1/James_K_35259_VoR.pdf	English	null	Reducing conflict and containment rates on acute psychiatric wards: The Safewards cluster randomised controlled trial	International journal of nursing studies	2,015	cc-by	9,492	Len Bowers a,, Karen James a, Alan Quirk b, Alan Simpson c, SUGARc,1, Duncan Stewart a, John Hodsoll a Len Bowers a,, Karen James a, Alan Quirk b, Alan Simpson c, SUGARc,1, Duncan Stewart a, John Hodsoll a a Institute of Psychiatry, King’s College London, De Crespigny Park, London SE5 8AF, United Kingdom b Royal College of Psychiatrists, 21 Prescot Street, London E1 8BB, United Kingdom c City University London, Northampton Square, London EC1V 0HB, United Kingdom A R T I C L E I N F O Article history: Received 24 February 2015 Received in revised form 30 April 2015 Accepted 3 May 2015 Keywords: Absconding Inpatient Psychiatry Rapid tranquillisation Restraint Seclusion Self harm Special observation Violence Article history: Received 24 February 2015 Received in revised form 30 April 2015 Accepted 3 May 2015 Background: Acute psychiatric wards manage patients whose actions may threaten safety (conﬂict). Staff act to avert or minimise harm (containment). The Safewards model enabled the identiﬁcation of ten interventions to reduce the frequency of both. Objective: To test the efﬁcacy of these interventions. Objective: To test the efﬁcacy of these interventions. Design: A pragmatic cluster randomised controlled trial with psychiatric hospitals and wards as the units of randomisation. The main outcomes were rates of conﬂict and containment. Participants: Staff and patients in 31 randomly chosen wards at 15 randomly chosen hospitals. Results: For shifts with conﬂict or containment incidents, the experimental condition reduced the rate of conﬂict events by 15% (95% CI 5.6–23.7%) relative to the control intervention. The rate of containment events for the experimental intervention was reduced by 26.4% (95% CI 9.9–34.3%). Conclusions: Simple interventions aiming to improve staff relationships with patients can reduce the frequency of conﬂict and containment. Trial registration: IRSCTN38001825 Conclusions: Simple interventions aiming to improve staff relationships with patients can reduce the frequency of conﬂict and containment. Trial registration: IRSCTN38001825. reduce the frequency of conﬂict and containment. Trial registration: IRSCTN38001825. ublished by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). Crown Copyright 2015 Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). * Corresponding author at: PO Box 30, Section of Mental Health Nursing Institute of Psychiatry, De Crespigny Park, London SE5 8AF, United Kingdom. Tel.: +44 020 7848 5323; fax: +44 020 7848 0458. 1 Service User and carer Group for Research, led by Professor Simpson, hosted by City University, London. E-mail address: len.bowers@kcl.ac.uk (L. Bowers). Contents lists available at ScienceDirect Contents lists available at ScienceDirect International Journal of Nursing Studies 52 (2015) 1412–1422 International Journal of Nursing Studies 52 (2015) 1412–1422 http://dx.doi.org/10.1016/j.ijnurstu.2015.05.001 0020-7489/Crown Copyright 2015 Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/ b /4 0/) 015.05.001 Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/ http://dx.doi.org/10.1016/j.ijnurstu.2015.05.001 0020-7489/Crown Copyright 2015 Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/ by/4.0/). What this paper adds events, nor do they aim to reduce conﬂict and containment as a whole. All ﬁgures for the numbers of studies refer to a cross topic review of the whole conﬂict and containment literature, ultimate consisting of 1177 papers and con- ducted by the authors’ research group. Both conﬂict and containment overall can be reduced, making wards safer and less coercive environments for patients and staff. Both conﬂict and containment overall can be reduced, making wards safer and less coercive environments for patients and staff. The underlying Safewards Model is supported and should now be subjected to further tests. The underlying Safewards Model is supported and should now be subjected to further tests. Previous research has shown highly variable rates of conﬂict and containment on different wards, not explicable solely in terms of the patients admitted (Bowers, 2009). The cross topic literature review referred to above (Bowers et al., 2010; Dack et al., 2013; James et al., 2012; Owiti and Bowers, 2011; Papadopoulos et al., 2012a,b; Stewart and Bowers, 2011; Stewart et al., 2009; Van Der Merwe et al., 2013) led to the development of the Safewards Model (Bowers, 2014). This model explains variable rates of conﬂict and contain- ment and identiﬁes a large number of ‘staff modiﬁers’: aspects of staff actions that can impact on the likelihood of conﬂict or containment incidents. The model enabled the creation of a list of interventions that could enhance the staff modiﬁers and thereby reduce conﬂict and containment rates. As the focus of potential interventions in the Safewards trial was the nursing team as a whole and generalised aspects of their attitudes and behaviour towards patients, wards had to be the unit of randomisation. The list of potential interventions was scored by the research team for feasibility and impact, resulting in a short list of 30 that were taken to consultations with panels of expert nurses, service users and carers (Simpson et al., 2014). The top 16 interventions went forward to a pilot study on four wards, and were subsequently reduced, consolidated and improved into a package of ten interventions for use in a full scale cluster randomised controlled trial. The ten easy Safewards interventions should be imple- mented in practice. Acute psychiatric wards provide limited duration care to people in acute states of disturbance and distress. 1.1. Objective We aimed to evaluate the efﬁcacy of a complex intervention (Safewards), targeted at nursing staff, to reduce conﬂict and containment rates at the level of acute psychiatric wards. What this paper adds Once admitted patients may exhibit a number of different difﬁcult and risky behaviours, including verbal aggression, attempts to abscond, self-harm, refusal to eat or drink, aggression to objects or people. A range of different methods are used by nursing staff either to prevent these behaviours from occurring, or ameliorate their outcomes, including the use of extra tranquillising medication, special observation by staff, manual restraint and seclu- sion. We refer to the behaviours posing a risk to patients or those around them as ‘conﬂict’, and the actions of staff to manage them as ‘containment’. Treating these behaviours collectively is justiﬁable because the different conﬂict behaviours (aggression, self-harm, substance/alcohol use etc.) correlate strongly within patients (Bowers et al., 2005) and within wards (Bowers, 2009). Patients who engage in high rates of one type of conﬂict behaviour are more likely to engage in others, and wards with high rates of one type of conﬂict behaviour are more likely to have high rates of others. The same is the case for different containment items, both for patients and wards. What is already known about the topic? alone or seclusion alone) very few RCTs have been undertaken. There are no previous RCTs of interventions to reduce conﬂict or containment as a whole. Even for individual conﬂict behaviours or containment events (e.g. violence There are substantially more before and after studies of interventions in practice, which are mostly local and without controls, and a large quantity of observational, longitudinal and descriptive studies. Narrative reviews are available, particularly for violence, and for seclusion and mechanical restraint. They suggest that conﬂict and containment rates can be inﬂuenced by staff behaviour, but the evidence is generally weak and of poor quality. What evidence exists has been assembled into the Safewards Model, which underlies the interventions used in this trial. 1 Service User and carer Group for Research, led by Professor Simpson, hosted by City University, London. p // g/ /j j 0020-7489/Crown Copyright 2015 Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/ by/4 0/) L. Bowers et al. / International Journal of Nursing Studies 52 (2015) 1412–1422 1413 1. Materials and methods Despite the link between different conﬂict and con- tainment events, and the connection between conﬂict and containment themselves, most current methods for making psychiatric wards safer places focus on just one or two types of these events. Training courses for staff in the prevention and management of violence are the most commonly used intervention. These contain basic de- escalation skills and manual restraint training (Lee et al., 2001). The ‘six core strategies’ are another approach, in this case targeted at reducing the use of seclusion and mechanical restraint (Huckshorn, 2005) and include: senior management commitment to change, using audit to inform practice, workforce training, use of assessment tools, patient involvement, and debrieﬁng techniques. A third current method aims to reduce violence through short term (shift by shift or day by day) risk assessment based on statistically veriﬁed indicators (Abderhalden et al., 2004). The majority of evaluations of these and other methods have been by natural experiment with ofﬁcial statistics as the outcome measure (n = 103 studies). Results have been variable (Stewart et al., 2010), the use of any control group rare (n = 20 studies), the possibility of substitution of one form of containment by another seldom assessed, the issue of missing data ignored, and the number of previous randomised trials very small (n = 5). Publication bias in relation to the high numbers of natural experiments is likely. None of these methods for making psychiatric wards safer provide a comprehensive model explaining the causes of conﬂict and containment 1.2. Participants The study comprised 31 psychiatric wards at 15 hospi- tals within 100 km of central London and in 9 NHS Trusts. Inclusion criteria were acute psychiatric wards for adults of any gender. Wards were excluded if they had a specialist function, had planned major changes, or where two or more of the following criteria were met: no permanent ward manager in post, a locum consultant solely respon- sible for inpatient care, >30% nursing staff vacancy rate. Willing nurses and healthcare assistants working on the selected wards were included, with 564 staff (88% of the possible total) giving their consent. Non-consenting staff were free not to submit outcome data, questionnaires, or participate in the interventions. Signed consent on behalf of patients was given by Trust CEOs, although no data was collected from patients and all research interventions were with staff. Directors of Nursing and Medical Directors also approved the study prior to access being granted. National Health Service ethical approval was secured (11/LO/0798). L. Bowers et al. / International Journal of Nursing Studies 52 (2015) 1412–1422 1414 abscond and 1 self-harm event, but does not record which patients were responsible for those events, or whether one patient was responsible for them all. A total conﬂict score is obtained by summing the number of conﬂict incidents during the shift, and a total containment score by summing the number of containment events. The tool has been demonstrated to be reliable (Bowers et al., 2006) and valid (Bowers et al., 2005) and is accompanied by a handbook, carefully devised operational deﬁnitions, and brief struc- tured training. Its associations with ward features, stafﬁng provision, patient characteristics, physical environment, routines, surrounding local community service provision, and change over time in relation to local policy changes and other events, have been thoroughly explored in some of the largest observational studies conducted into acute psychiatry to date (Bowers et al., 2007a,b). Secondary outcomes were the Attitude to Personality Disorder Questionnaire (Bowers and Allan, 2006), the Self-harm Antipathy scale (Patterson et al., 2007), the Ward Atmo- sphere Scale (programme clarity, and order and organisa- tion subscales) (Moos, 1974); and the SF-36v2, a short form health survey (Ware et al., 2002). These additional scales were those best representative of the types of changes in staff predicted by the Safewards Model to be associated with changes in rates of conﬂict and containment. 1.2. Participants In addition the SF-36v2 was included to assess the impact of the control interventions. Fidelity was measured by a simple checklist completed by Research Assistants on every visit to the wards and by a participant end of study questionnaire. 1.3. Interventions Wards in the experimental condition implemented a package of ten ‘Safewards’ interventions: (1) mutually agreed and publicised standards of behaviour by and for patients and staff; (2) short advisory statements (called ‘soft words’) on handling ﬂashpoints, hung in the nursing ofﬁce and changed every few days; (3) a de-escalation model used by the best de-escalator on the staff (as elected by the ward concerned) to expand the skills of the remaining ward staff; (4) a requirement to say something good about each patient at nursing shift handover; (5) scanning for the potential bad news a patient might receive from friends, relatives or staff, and intervening promptly to talk it through; (6) structured, shared, innocuous, personal information between staff and patients (e.g. music preferences, favourite ﬁlms and sports, etc.) via a ‘know each other’ folder kept in the patients day room; (7) a regular patient meeting to bolster, formalise and intensify inter-patient support; (8) a crate of distraction and sensory modulation tools to use with agitated patients (stress toys, mp3 players with soothing music, light displays, textured blankets, etc.); (9) reassuring explanations to all patients following potentially frightening incidents; and (10) a display of positive messages about the ward from discharged patients. Interventions therefore occurred at the cluster level, as they were collective endeavours of the nursing team, or visible to everyone. Full descriptions of these interventions coupled with training videos are freely available online (www.safewards.net). Wards in the control condition implemented a package of interventions directed at improving staff physical health: a desk exercises poster in ward ofﬁce; pedometer based competi- tions; supplies of healthy snacks; diet assessment and individualised feedback; health and exercise magazines supplied regularly to the staff ofﬁce; health promotion literature; linkages to local sports and exercise facilities. Improvement in physical health was predicted by the Safewards Model to have no impact on conﬂict and containment. This arm of the study therefore controlled for both researcher attention and participant expectancy. All wards and their staff in both arms were primed to expect reductions in conﬂict and containment rates. Staff on the control wards were told that improvements in their own physical health would lead to them delivering nursing care more effectively, and thereby reduce conﬂict and contain- ment 1.5. Sample size The required sample size was based on the data from the City-128 study (Bowers, 2009), extrapolated to a full trial scenario through simulation, powered for two primary outcomes by Bonferroni adjustment. Conﬂict and containment have a complex and partial relationship, therefore these two types of events need to be assessed independently. The model for event counts (conﬂict or containment) was Poisson based with offset for number of beds per ward and random effects of ward nested within hospital to account for clustering due to paired randomi- sation of wards within hospital and repeated measures within ward. A conservative ﬁgure of 13 beds per ward was assumed, with 2 wards per hospital and 10 hospitals participating. For periods of 30 days (three shifts per day, 90 nursing shifts in total) in each phase the model predicted 97.9% power for conﬂict events and 93.7% power for containment events. The calculations allowed for a modest 20% decreases in target events commensurate with those obtained in our previous before and after trials (Bowers et al., 2003, 2006). In order to allow for potential ward drop outs from the trial, a target of 15 hospitals (30 wards) was set for the study. 1.4. Outcomes The primary outcomes were rates of total conﬂict and rates of total containment as measured by the Patient-staff Conﬂict Checklist (PCC) (Bowers et al., 2005). This single sheet paper form was completed by the nurse in charge at the end of every nursing shift, and logs the frequency of 22 conﬂict events (verbal aggression, suicide attempts, alcohol use, attempted absconding, etc.) and 8 uses of containment (coerced medication, seclusion, restraint, special observation, etc.). These events are recorded at the level of the shift, not individual patient, for example a PCC may record 3 verbal abuse events, 1 attempted 1.7. Procedure at each hospital, (iii) allocation to experimental or control (Fig. 1). In each case simple randomisation was used for the selection process by the designated staff member at King’s College Clinical Trials Unit. At one hospital site with three eligible wards there was uncertainty as one ward of two was potentially going to close. All three wards were therefore recruited and the two wards under threat of closure were randomised to the same experimental condition. No ward closed during the study, resulting in a total sample size of 31 wards. No wards dropped out from the study. All randomisation was independent of the researchers and trial statistician. Throughout the study, wards and their staff were blind as to which package of interventions were the experimental or control condition, each of which were given neutral titles and were described as likely to reduce conﬂict and containment rates. Additionally neither the staff nor the research assistants working with them knew which intervention package would be applied on which ward until two weeks before its introduction, so that baseline data could not be biased. at each hospital, (iii) allocation to experimental or control (Fig. 1). In each case simple randomisation was used for the selection process by the designated staff member at King’s College Clinical Trials Unit. At one hospital site with three eligible wards there was uncertainty as one ward of two was potentially going to close. All three wards were therefore recruited and the two wards under threat of closure were randomised to the same experimental condition. No ward closed during the study, resulting in a total sample size of 31 wards. No wards dropped out from the study. All randomisation was independent of the researchers and trial statistician. Throughout the study, wards and their staff were blind as to which package of interventions were the experimental or control condition, each of which were given neutral titles and were described as likely to reduce conﬂict and containment rates. Additionally neither the staff nor the research assistants working with them knew which intervention package would be applied on which ward until two weeks before its introduction, so that baseline data could not be biased. Once recruited, ward staff were trained in the use of the PCC and collection of data on the primary outcome continued throughout the study (see Fig. 1). 1.6. Randomisation and masking All hospital sites within 100 km of central London and with at least two eligible wards were identiﬁed. Three random selections were made: (i) hospitals, (ii) two wards L. Bowers et al. / International Journal of Nursing Studies 52 (2015) 1412–1422 1415 1.8. Statistical analysis 1.8. Statistical analysis served both male and female patients (n = 16) with ten serving men and ﬁve women only. The modal age group of the participating staff was 40–49 years (33.7%) a minority were white British (28.4%) and most were female (59.4%), all being typical of nursing staff working in acute psychiatry in the south east of England (Bowers et al., 2008a,b). There was no signiﬁcant difference in ward type, gender served, staff age gender or ethnicity, between the experimental and control groups. The statistical analysis plan was agreed with the Trial Steering Committee before data collection began. Data were analysed using the MCMCglmm package in R 2.15 (Hadﬁeld, 2010). The primary outcome was counts of conﬂict and containment events by ward shift (am, pm and night) collected over the course of the study phases; baseline, implementation and outcome. As the distribution of events in the data had an excess of zero counts (shifts with no events) we chose to model the data with a Poisson hurdle mixed model. Count data is generally modelled with Poisson distribution; however, it is often the case that there is greater heterogeneity in the data than expected by the Poisson model which can be manifest in numbers of zero events, as here. A hurdle model is a two-part model consisting of a binary response and count regression model, which has previously been useful in research on addictions (Atkins et al., 2013) and elsewhere. The ﬁrst part of the model, the hurdle is based on the binomial distribution and describes the risk of a conﬂict or containment incident occurring (and therefore accounts for the zero events). If an event does occur, i.e. once the hurdle has been crossed, a zero-truncated Poisson distribution models the number of events occurring. The dependent variable was the count of incidents (conﬂict or containment) per shift. Covariates were study phase (baseline, implementation, and outcome), treatment (Safe- wards intervention or control), time of shift (am, pm or night), day in study phase (centred) and the log of the number of beds per shift to be used as an offset. An interaction between outcome phase and treatment condi- tion gave the primary treatment difference at for the outcome phase. As the number of beds differed per ward the offset allows predicted rates of events per patient bed. 2.2. Missing data and sensitivity analysis Table 3 shows the number and percentage of missing observations by experimental phase and experimental condition. Baseline rates of missing data were relatively high at 36% for the control and 40% for the experimental condition. Although the increase in missingness was greater in the control than the Safewards condition, this difference was only present as a weak non-signiﬁcant effect (OR – 0.87, 95% CI 0.74–1.03). There were no other predictors of missingness than already included in the primary analysis model. 1.9. Role of the funding source Excluding wards from the analysis with high rates of missing data in the outcome phase had little impact on the treatment effect (for both the rate ratio and hurdle). Similarly, excluding hospitals with wards that did not comply with the protocol or that had operational difﬁcul- ties did not effect the results. In all cases the direction and magnitude of the effect was approximately the same (Table 4, SI). The sponsor of the trial had no role in trial design, data collection, data analysis, interpretation, or writing of the report. The corresponding author had full access to all the data in the trial and had ﬁnal responsibility for the decision to submit for publication. L. Bowers et al. / International Journal of Nursing Studies 52 (2015) 1412–1422 1416 1.8. Statistical analysis Rate ratio estimates of the treatment effect are presented together with the 95% Bayesian credible interval and associated p statistic. There were high rates of missing data in the study, and consequently we took a range of pragmatic approaches to assess the robustness of our results, under both missing at random (MAR) and missing not at random (MNAR) assumptions (White et al., 2011). Further details of data analyses are in the supplementary information (SI). No wards dropped out of the study once recruited. However, the response rate for the primary outcome was less than expected, with less than 50% of PCC forms returned in the outcome phase. There was a range of return rates with some wards providing very high return rates and others with very low rates. However, the rates of missing data were approximately the same in the experimental and control conditions (see Table 3 and below) and we present a comprehensive investigation into possible biases due to missing data. 2.1. Primary outcome Table 1 shows the baseline measures and Table 2 shows the rate ratios for the mean effect of treatment on the primary outcomes for the probability and rate of events. Relative to the control intervention, when conﬂict events occurred the Safewards intervention reduced the rate of conﬂict events by 15.0% (95% CI 5.6–23.7%). Similarly, when containment events occurred the rate of contain- ment events for ward shifts with events was reduced by 26.4% (95% CI 9.9–34.3%). There were no signiﬁcant differences in the rates of zero event shifts for conﬂict or containment. 1.7. Procedure Baseline data were collected for eight weeks, and wards then had a further eight weeks to implement their allocated package of interventions. They then continued using the inter- ventions for a further eight weeks. Secondary outcome questionnaires were collected during the baseline period and repeated during the outcome period, and were distributed to consenting staff via internal mail or in person. All wards were visited 2–3 times a week throughout the study by researchers, who picked up and delivered questionnaires, encouraged participation, liased with the team to plan introduction of the interventions, and answered any questions. The trial is registered, number IRSCTN38001825, and an indepen- dently chaired Trial Steering Committee had oversight of the project. Sample fr ame of eligible hosp ita ls (n=42) Random sample of hosp itals (n=20) Excluded (n= 7): War d closures (n= 3) Service r econﬁguraon to coinci de with trial (n= 4 ) Substute hosp ita ls (n=2) Randomised ( 31 war ds at 15 h osp itals) Allocated to Safe war ds (n=16 wards) Allocated to p hysica l hea lth (n=15 war ds) Trust dec lined to pa rci pate (n= 1) Recei ved Safewards (n=16 wards)* Recei ved p hysi cal heal th (n=15 war ds) Substute war d at 1 hosp ital (n= 1)* Prop osed ward closure ann oun ced and su bstu te re cruit ed. However, cl osure deci sio n deferred so co nnued wit h both. Fig. 1. Trial proﬁle. Sample fr ame of eligible hosp ita ls (n=42) Trust dec lined to pa rci pate (n= 1) Excluded (n= 7): War d closures (n= 3) Service r econﬁguraon to coinci de with trial (n= 4 ) Randomised ( 31 war ds at 15 h osp itals) Allocated to Safe war ds (n=16 wards) Allocated to p hysica l hea lth (n=15 war ds) Recei ved Safewards (n=16 wards) Recei ved p hysi cal heal th (n=15 war ds) *Prop osed ward closure ann oun ced and su bstu te re cruit ed. However, cl osure deci sio n deferred so co nnued wit h both. Fig. 1. Trial proﬁle. Fig. 1. Trial proﬁle. L. Bowers et al. / International Journal of Nursing Studies 52 (2015) 1412–1422 Baseline outcome measures. Baseline outcome measures. p p b Positive ﬁgures represent increases or improvements on the experimental wards, negative ﬁgures increases or improvements on the control wards. Table 1 Table 1 Baseline outcome measures. Outcome Experimental Control Primary outcomes PCC conﬂict Overall event rate – mean (SD) 5.22 (6.32) 4.69 (4.6) Overall event rate – median (IQR) 3 (1–7) 4 (1–7) Risk of events – n/N 0.80 (1391/1607) 0.87 (1293/1609) PCC containment Overall event rate – mean (SD) 1.26 (1.93) 1.39 (1.94) Overall event rate – median (IQR) 0 (0–2) 1 (0–2) Risk of events – n/N 0.804 (938/1607) 0.866 (802/1609) Secondary outcomes WAS Order and organisation – mean (SD) 7.19 (2.27) 6.43 (2.53) Programme clarity – mean (SD) 7.4 (2.04) 7.18 (2.06) Staff control – mean (SD) 1.83 (1.55) 1.8 (1.4) SHAS Total – mean (SD) 78.79 (18.85) 80.16 (21.1) APDQ Enjoyment – mean (SD) 4.76 (0.7) 4.8 (0.7) Security – mean (SD) 5.11 (0.68) 5.09 (0.61) Acceptance – mean (SD) 5.35 (0.65) 5.38 (0.56) Purpose – mean (SD) 5.08 (0.79) 5.1 (0.85) Enthusiasm – mean (SD) 4.28 (0.99) 4.23 (0.93) SF-36v2 Physical health – mean (SD) 52.19 (7.79) 51.94 (7.31) Mental health – mean (SD) 50.24 (9.46) 50.74 (9.98) PCC Ch kli t WAS W d At h S l SHAS S lf H A ti th S l APDQ Attit d t P lit Di d Q ti i SF 36 2 Sh t 2. Results In the ﬁrst imputation strategy we assumed all missing observations had zero events. In contrast for the second strategy we assumed that all missing shift reports were due to higher than average event rates. Neither of these strategies changed the direction of the treatment effect, rate ratios were 0.85–0.9 for conﬂict events and 0.76–0.9 for containment. This was the case even if missing Of the 31 wards at 15 hospitals recruited to the study, 16 were assigned to the Safewards intervention package, and 15 to the physical health package. The mean number of beds per ward was 19 (SD 3.96) and most were generic acute wards (n = 24) with three triage/assessment wards and four psychiatric intensive care units. The majority L. Bowers et al. / International Journal of Nursing Studies 52 (2015) 1412–1422 1417 Table 1 eases or improvements on the experimental wards, negative ﬁgures increases or improvements on the control wards. in number of zero event shifts between baseline implementation and outcome periods. PCC, Checklist; WAS, Ward Atmosphere Scale; SHAS, Self-Harm Antipathy Scale; APDQ, Attitudes to Personality Disorder Questionnaire. SF-36v2, Short Form 36 Health Survey v2 o d e e ce u be o e o e e t s ts bet ee base e p e e tat o a d outco e pe ods. ve ﬁgures represent increases or improvements on the experimental wards, negative ﬁgures increases or improvements s represent increases or improvements on the experimental wards, negative ﬁgures increases or improvements Table 3 generated during implementation of the interventions which were checked and rated by the researchers. Evidence for the control interventions was much easier to observe than for the experimental package. Fidelity checklist scores were converted into percentage imple- mentation scores by regarding the maximum score as 100% and the minimum score as 0%. The mean ﬁdelity to the experimental intervention by ward during the outcome period was 38% (SD 8, range 27–54%, n = 271) and for the control intervention 90% (SD 9, range 69–99%, n = 209). Examination of scores over time for the experimental wards showed a linear increasing trend from the start of the implementation period at 0% through to the end of the outcome period at 50%. Fidelity was also assessed via the completion of an end of study questionnaire by partici- pants, in which they were asked whether they used each of the interventions. This was also converted into percentage ﬁdelity scores by ward. Mean ﬁdelity to the experimental intervention by ward by this measure was 89% (SD 11, range 62–100%, n = 79) and for the control intervention 73% (SD 19, range 39–100%, n = 74). Number of missing shift PCC reports by trial phase and treatment condition. The number missing was the same for both conﬂict and containment incidents as they were on the recorded on the same report. Control Treatment n % n % Baseline 913 36.23 1079 40.14 Implementation 1106 43.89 1250 46.50 Outcome 1384 54.92 1524 56.70 observations were imputed to have absurd rates of events, e.g. over 100 conﬂict and containment events in 1 shift (Table 5, SI). Next we assumed that missingness mechanisms were different in the two trial arms (despite no overall difference in the rates, Table 3) by imputing observations that were missing in addition to the baseline rate. Firstly, we considered that missing observations in control arm had the same rate of events as the present observations but those in the experimental condition were associated with increased rates of events. For conﬂict an increase of 2 events per shift was needed to reduce the rate ratio to 1 (RR for no effect, see Table 6, SI). For containment, a mean increase of 1 per missing shift was necessary. 2.5. Credibility of the control intervention, efﬁcacy of blinding, and contamination 2.5. Credibility of the control intervention, efﬁcacy of blinding, and contamination Both control and experimental ward staff were provided with a rationale as to why their interventions should reduce conﬂict and containment, and that the researchers were interested to discover which set of interventions worked best. Via the ‘end of study’ question- naire participants were asked whether they thought they were in the experimental or the control group. The majority in both groups thought they were in the experimental group, but that proportion was higher in the experimental group (88% vs. 74%). Comments during completion indicated that the question was often not understood, or was interpreted as ‘I took part in an experiment therefore I must have been in the experimental group’. Managers and staff were asked not to discuss the interventions in use on their own ward with people working on the ward in the opposing arm of the study. However via the ‘end of study’ questionnaire three quarters (116/147, 79%) admitted to engaging in such discussions. 2.3. Secondary outcomes For the secondary outcomes, treatment effects (also in Table 2) showed no difference between the control and Safewards intervention for the Ward Atmosphere Scale, Self-Harm Antipathy Scale, Attitudes to Personality Disor- der Questionnaire. On the SF-36 scale there was no difference in experimental group for the mental health measure, but there was an effect of group on physical health. The control group staff showed a 1.85 point (95% CI: 0.003–3.704) greater improvement in physical health than the Safewards intervention. Table 2 Table 2 Estimates of treatment effects for Safewards interventions relative to control for primary and secondary outcomes. Table 2 Estimates of treatment effects for Safewards interventions relative to control for primary and secondary outcomes. Estimates of treatment effects for Safewards interventions relative to control for primary and secondary outcomes. Outcome Estimate 95% CI p-Value Primary outcome PCC conﬂict Count rate ratio 0.850 0.763–0.943 0.001 Hurdle rate ratioa 1.139 0.915–1.426 0.234 PCC containment Count rate ratio 0.768 0.655–0.901 0.004 Hurdle rate ratioa 1.044 0.828–1.336 0.708 Secondary outcomesb WAS Order and organisation 0.315 0.792 to 0.163 0.197 Programme clarity 0.267 0.218 to 0.753 0.281 Staff control 0.196 0.568 to 0.176 0.301 SHAS Total 0.227 3.375 to 3.829 0.902 APDQ Enjoyment 0.023 0.13 to 0.176 0.768 Security 0.079 0.209 to 0.05 0.231 Acceptance 0.067 0.062 to 0.196 0.312 Purpose 0.087 0.28 to 0.1 0.388 Enthusiasm 0.031 0.178 to 0.24 0.772 SF-36 Physical health 1.85 3.702 to 0.003 0.05 Mental health 0.709 2.962 to 1.544 0.537 a Test for difference in number of zero event shifts between baseline implementation and outcome periods p-Value a Test for difference in number of zero event shifts between baseline implementation and outcome periods. b Positive ﬁgures represent increases or improvements on the experimental wards, negative ﬁgures increases or L. Bowers et al. / International Journal of Nursing Studies 52 (2015) 1412–1422 1418 Table 3 Given the large amount of missing data imputed, the need for a relatively large increase in event rates to change the direction of the experimental effect suggests our ﬁndings are robust. 3. Discussion The most stringent control over potential bias or other threats to validity such as trial registration, binding commitments to main out- comes and analysis plans in advance of results, trial steering committee management, fully independent ran- domisation, blinded independent analysis, blinding of participants, an active control intervention, outcome measures of proven validity and reliability, do not appear to have been utilised in any previous study in this ﬁeld. All previous studies relied upon ofﬁcial incident reports or forms that were only completed by staff once incidents had occurred. Using this method there never appears to be any missing data, and rates cannot be assessed for bias. Using the PCC meant that missing data rates became visible and potential bias could be assessed. The Safewards trial thus represents a new step in the rigour, scale and scope of research into patient and staff safety in inpatient psychiatry. p A fall in the rate of conﬂict also occurred on the control wards, albeit not as large at that on the experimental wards. This may represent a true placebo effect, the result of expectancy of a beneﬁcial change, but in that case containment should also have shown some decrease. The pedometer intervention on the control wards may have increased nurses’ movements around the ward, making them more available to patients, or enabling early intervention to avert conﬂict. Alternatively this effect may have been mediated by the ward team feeling valued (the health intervention focused on them personally rather than patients), however there was no mental health gain for control ward staff shown on the SF-36. Finally the gain on the control wards may have been a result of contamination from the experimental wards on the same site. Staff on the different wards were asked not to talk with each other about the interventions they were using, however there were unit managers who crossed over both wards, and the small size of many psychiatric units meant that staff often substituted for each other across wards to cover absence due to holidays or sickness, and some evidence for effects of one ward on another has been previously published (Bowers et al., 2008a,b). However, the generally low level of implementation on the experimental wards suggests that any contamination to the control wards would have been minimal. 3. Discussion To the extent that contamination did occur, it would have diluted the experimental interventions’ effect upon the outcomes relative to the control wards, thus making the reported positive ﬁndings more robust and an underestimate of their true size. The trial did yield an undesirably high level of missing data, despite the regular visits of researchers to the wards. Not all staff consented to the study, therefore on some occasions the nurse in charge did not feel obliged to complete the end of shift PCC. On other occasions staff were busy with other matters, the close of a nursing shift being when many reports are assembled and messages passed on in handover to the oncoming team of nurses. This moment both increases the likelihood of accuracy of a PCC (information is already being assembled) and decreases the likelihood it will be completed. The real issue with respect to the trial is whether the data were missing at random, and if not, did any bias undermine the ﬁndings? We believe our sensitivity analysis strongly supports the interpretation that this was not the case, however no such analysis can ever completely substitute for a full dataset. Some degree of caution must therefore be expressed about our ﬁndings. Some of these same factors may also have led to the improvements in the Attitude to Personality Disorder Questionnaire scores on both the experimental and the control wards. However the direction of causality is open to question: better attitudes to patients might have been the result of decreased rates of difﬁcult patient behaviour. Some previous research has indicated that this is more likely to be the case (Bowers et al., 2007a,b; Kellam et al., 1966). It was clear that the experimental interventions had marginally greater credibility amongst the participants, with a greater number on those wards correctly able to see through the attempt to blind them. It remains therefore possible that some proportion of our positive results were due to a greater expectancy of change on the part of staff on the experimental wards. It is hard to be certain about the degree of implemen- tation of the research interventions. The ‘end of study’ questionnaires were positive, but the numbers of these questionnaires represented a low response rate and the chance of response bias towards exaggerated ﬁdelity high. 3. Discussion Assessing ﬁdelity to the interventions was intrinsically difﬁcult. The ﬁnal target of most interventions was to organically change interactions between patients and between staff and patients. Detailed, intensive and laborious structured observation would have been neces- sary to capture this. As an alternative, researchers completed a checklist on every ward visit (2–3 times a week) scoring the presence of visible evidence of the use of each of the interventions, with a different checklist being used for the experimental and control wards. Each of the interventions provided differing degrees of visible evi- dence, some provided none. Often the ﬁdelity score reﬂected evidence that was on display, rather than the degree to which staff engaged with and used the displayed material. Other scores were dependent on staff records A large scale cluster RCT was conducted over a three month period in ﬁfteen hospitals in and around London, with the aim of testing a package of interventions to increase safety and reduce coercion. The trial was a complex undertaking, requiring a large number of research staff operating across multiple sites, with considerable planning and organisation. Completion was dependent on the support and willingness of busy nursing staff to engage with the trial and undertake new and additional activities. Nevertheless, no ward dropped out from the study, and the trial intervention proved to be effective in reducing both conﬂict and containment. There have been no previous randomised controlled trials of interventions to reduce conﬂict and containment L. Bowers et al. / International Journal of Nursing Studies 52 (2015) 1412–1422 1419 and continued, intensifying throughout the outcome phase. together and across all types of event. The trials that have taken place have been restricted to one or two items, usually violent incidents, or seclusion and mechanical restraint (Abderhalden et al., 2008; Prescott et al., 2007). They have been mostly limited to samples of wards at single hospitals (Van De Sande et al., 2011), inactive treatment as usual controls (Putkonen et al., 2013), or simply report change over time without any control comparison at all (Pollard et al., 2007). 3. Discussion Nevertheless the result does suggest that there may be real occupational health value in light but efﬁcacious health promotion interven- tions with staff, for which there is some other evidence (Dugdill et al., 2008). The location of the trial in the UK may also be seen as a limitation. At the time of the trial, acute inpatient psychiatric care in the UK was composed of three main types of wards. Generic acute wards, usually serving patients of both genders, triage or assessment wards catering speciﬁcally for new admissions with a view to fast discharge, usually with higher nurse stafﬁng levels, and psychiatric intensive care units, smaller wards with greater security and higher stafﬁng levels for the management of more disturbed patients. Psychiatric nursing in the UK is a specialist qualiﬁcation attained through a three year University based course. On average half of all ward staff are qualiﬁed nurses, the remainder unqualiﬁed healthcare assistants. These wards also beneﬁt from input from medical staff, occupational therapists and in some cases psychologists. Safewards ﬁndings are therefore most generalisable to similar settings. Applicability and efﬁcacy in other specialities (forensic, adolescent, or older people’s psychiatric wards) and in other countries with different ward types, stafﬁng compositions and care pathways, is therefore open to question. The limitations of the Safewards trial, stated above, were the large quantity of missing data and the limited degree to which the interventions could be implemented within the short time period of the study. The study had no third arm for treatment as usual, however this would have yielded no extra beneﬁt as the control arm of the trial controlled for the combined effect of natural change over time and the effect of participating in an experiment. Collecting data at the patient level might be seen as superior to the shift level data taken during the trial, however this would have required staff to complete at the end of every shift, in effect a matrix consisting of 22 different events and the numbers of patient on the ward concerned (on average 19), notwithstanding the problems of conﬁdentially identifying patients in the submitted research data. Whilst this approach was considered, it was rejected as likely to lead to catastroph- ically low levels of data return, and a likely high degree of data errors. Placing researcher non-participant observers on the wards was also considered and rejected on two grounds. 3. Discussion Firstly one observer only on the ward can only see what is happening where they are, thus producing a report which is not as comprehensive as that which can be collated by the whole nursing team. Secondly, cost, as an observational based trial would have required at least double the time and double the numbers of researchers. As it was, thirteen people were employed full time on the trial while it was running. It may be considered that all the items on the outcome measure (PCC) were not equally severe (e.g. a suicide attempt vs. and attempted abscond), and that some weighting method should have been used before calculating the trial outcome. However we previously conducted an expert rating exercise for the PCC, only to ﬁnd that the total scores produced from actual ward data were so very highly correlated with the unweighted score as to yield no extra beneﬁt or accuracy (Bowers et al., 2006). We therefore retained the more readily understandable unweighted scoring system. It might be argued that the study results may be biased by the presence of particularly difﬁcult patients at different times, skewing the rates of conﬂict and containment on those wards. However, if such a process was occurring, we would expect any effects to be randomly spread across experimental and control wards, and different periods of the study. It is worth noting that the trial covered more than 15 years of acute ward time (31 wards 6 months In contrast to the limitations, the strengths of the Safewards trial were a theoretically generated set of interventions, advance registration, an adequately pow- ered and credibly generalisable sample size, independent randomisation, active control for expectancy, experimen- tal effect and change over time, independent oversight, blinding of subjects, independent and blinded statistical analysis and a demonstrable impact on conﬂict and containment rates. Therefore, in the absence of any comparable quality of evidence on what makes psychiatric wards safer places, we recommend that the Safewards interventions are imple- mented on adult acute mental health wards, as the ﬁndings of this trial are that the gains for patients and staff may be signiﬁcant. 3. Discussion Objective observational measures taken by the researchers showed more modest implementation rates, however these did climb steadily over time. Acute psychiatric wards are busy and chaotic environments, with constant patient turnover and large teams of nurses and others working on a shift system seven days a week (Cleary et al., 2011). Asking them to implement ten interventions, however small, across the whole team and within an eight week period was a huge demand. A longer time period clearly would have allowed a greater degree of intervention implementation. As it was, implementation stretched out of the implementation phase of the research We cannot tell from our study what the sustainability of the experimental interventions would be over the longer term, and without researcher support. The interventions have different although related targets, some to change the attitudes of staff to patients and the ways they relate to them, others to permanently extend their interactions skill set in certain common circumstances. The ten interven- tions were derived from the Safewards Model and were a subset of many possible additional interventions, all of which are freely available online. Clinicians may also use the underlying model to construct their own interventions. The ambition is that healthcare organisations will use the L. Bowers et al. / International Journal of Nursing Studies 52 (2015) 1412–1422 1420 Safewards Model as a mechanism for continuing effort in improving safety and reducing coercion, and to that end Safewards has already been recommended in policy (Department of Health, 2014). each). In addition, in other previous studies we have sought to ascertain whether individual patients can skew PCC results. In a longitudinal study collecting PCC data on sixteen wards for two years we also conducted regular interviews of staff that speciﬁcally asked about particular problem patients. It was not possible to draw a connec- tion between particular patients and ﬂuctuations in conﬂict and containment rates (Papadopoulos et al., 2012a,b). The small physical health gains occurring on the control wards were a notable and welcome outcome, but were possibly expectancy or placebo effect. The experimental interventions were not a full control for physical health interventions, as staff on the experimental wards were not primed to expect an improvement in their health as a result of their changes to practice. References Abderhalden, C., Needham, I., Dassen, T., Halfens, R., Haug, H.J., Fischer, J.E., 2008. Structured risk assessment and violence in acute psychi- atric wards: randomised controlled trial. Br. J. Psychiatry 193, 44–50. Moos, R., 1974. Ward Atmosphere Scale Manual. Consulting Psychologists Press, Palo Alto, CA. i i i i f di f f l f Owiti, J., Bowers, L., 2011. A narrative review of studies of refusal of psychotropic medication in acute inpatient psychiatric care. J. Psy- chiatr. Ment. Health Nurs.. Abderhalden, C., Needham, I., Miserez, B., Almvik, R., Dassen, T., Haug, H., Fischer, J., 2004. Predicting inpatient violence in acute psychi- atric wards using Broset-Violence-Checklist: a multicentre pro- spective cohort study. J. Psychiatr. Ment. Health Nurs. 11, 422– 427. Papadopoulos, C., Bowers, L., Quirk, A., Khanom, H., 2012a. Events pre- ceding changes in conﬂict and containment rates on acute psychiatric wards. Psychiatr. Serv. 63 (1), 40–47. Papadopoulos, C., Ross, J., Stewart, D., Dack, C., James, K., Bowers, L., 2012b. The antecedents of violence and aggression within psychiatric in-patient settings. Acta Psychiatr. Scand. 125 (6), 425–439. Atkins, D.C., Baldwin, S., Zheng, C., Gallop, R.J., Neighbors, C., 2013. A tutorial on count regression and zero-altered count models for longitudinal substance use data. Psychol. Addict. Behav. 27, 166– 177. Paterson, B., Bradley, P., Stark, C., Saddler, D., Leadbetter, D., Allen, D., 2003. Deaths associated with restraint use in health and social care in the UK. The results of a preliminary survey. J. Psychiatr. Ment. Health Nurs. 10 (1), 3–15. Bowers, L., 2009. Association between staff factors and levels of conﬂict and containment on acute psychiatric wards in England. Psychiatr. Serv. 60 (2), 231–239. Patterson, P., Whittington, R., Bogg, J., 2007. Measuring nurse attitudes towards deliberate self harm: the Self Harm Antipathy Scale (SHAS). J. Psychiatr. Ment. Health Nurs. 14 (5), 438–445. Bowers, L., 2014. Safewards: a new model of conﬂict and containment on psychiatric wards. J. Psychiatr. Ment. Health Nurs. 21, 499–508. l d k ll l f b d Bowers, L., Alexander, J., Gaskell, C., 2003. A trial of an anti-absconding intervention in acute psychiatric wards. J. Psychiatr. Ment. Health Nurs. 10 (4), 410–416. Pollard, R., Yanasak, E., Rogers, S., 2007. Organisational and unit factors contributing to reduction in the use of seclusion and restraint pro- cedures on an acute psychiatric inpatient unit. Psychiatr. Q. 78, 73–81. Bowers, L., Allan, T., 2006. References The Attitude to Personality Disorder Question- naire: psychometric properties and results. J. Personal. Disord. 20 (3), 281–293. Prescott, D.L., Madden, L.M., Dennis, M., Tisher, P., Winget, C., 2007. Reducing mechanical restraints in acute psychiatric care settings using rapid response teams. J. Behav. Health Serv. Res. 34 (1), 96– 105. Bowers, L., Banda, T., Nijman, H., 2010. Suicide inside: a systematic review on inpatient suicide. J. Nerv. Ment. Dis. 198, 315–328. Bowers, L., Douzenis, A., Galeazzi, G., Forghieri, M., Tsopelas, C., Simpson, A., Allan, T., 2005. Disruptive and dangerous behaviour by patients on acute psychiatric wards in three European centres. Soc. Psychiatry Psychiatr. Epidemiol. 40, 822–828. Putkonen, A., Kuivalainen, S., Louheranta, O., Repo-Tiihonen, E., Ryynanen, O.P., Kautiainen, H., Tiihonen, J., 2013. Cluster-randomized controlled trial of reducing seclusion and restraint in secured care of men with schizophrenia. Psychiatr. Serv. 64 (9), 850–855. Bowers, L., Flood, C., Brennan, G., Allan, T., 2008a. A replication study of the city nurses intervention: reducing conﬂict and containment on three acute psychiatric wards. J. Psychiatr. Ment. Health Nurs. 15, 737–742. Simpson, A., Barlow, S., Cox, L., Jones, J., SUGAR, 2014. Adding SUGAR: collaborating with service users and carers in mental health nursing research. J. Psychosoc. Nurs. Health Serv. 52 (1), 22–30. Stewart, D., Bowers, L., 2011. Absconding and locking ward doors: evidence from the literature. J. Psychiatr. Ment. Health Nurs. 18, 89–93. Bowers, L., Flood, C., Brennan, G., LiPang, M., Oladapo, P., 2006. Prelimi- nary outcomes of a trial to reduce conﬂict and containment on acute psychiatric wards: city nurses. J. Psychiatr. Ment. Health Nurs. 13, 165–172. Stewart, D., Bowers, L., Simpson, A., Ryan, C., Tziggili, M., 2009. Manual restraint of adult psychiatric inpatients: a literature review. J. Psy- chiatr. Ment. Health Nurs. 16 (8), 749–757. Bowers, L., Hackney, D., Nijman, H., Grange, A., Allan, T., Simpson, A., Hall, C., Eyres, S., 2007a. A longitudinal study of conﬂict and containment on acute psychiatric wards. In: Report to the Department of Health- London, City University. Stewart, D., Van Der Merwe, M., Bowers, L., Simpson, A., Jones, J., 2010. A review of interventions to reduce mechanical restraint and seclusion among adult psychiatric inpatients. Issues Ment. Health Nurs. 31 (6), 413–424. Bowers, L., Jones, J., Simpson, A., 2008b. The demography of nurses and patients on acute psychiatric wards in England. J. Clin. Nurs. 18, 884– 892. L. Bowers et al. / International Journal of Nursing Studies 52 (2015) 1412–1422 L. Bowers et al. / International Journal of Nursing Studies 52 (2015) 1412–1422 1421 Bowers, L., Whittington, R., Nolan, P., Parkin, D., Curtis, S., Bhui, K., Hackney, D., Allan, T., Simpson, A., Flood, C., 2007b. The City 128 Study of Observation and Outcomes on Acute Psychiatric Wards. Report to the NHS SDO Programme. London, NHS SDO Programme. Acknowledgements This paper presents independent research funded by the National Institute for Health Research (NIHR) under its Programme Grants for Applied Research programme (RP-PG-0707-10081). Additional support was provided by the National Institute for Health Research (NIHR) Biomedical Research Centre for Mental Health at South London and Maudsley NHS Foundation Trust and the Institute of Psychiatry, King’s College London. The views expressed are those of the authors and not necessarily those of the NHS, the NIHR or the Department of Health. Cleary, M., Hunt, G.E., Horsfall, J., Deacon, M., 2011. Ethnographic research into nursing in acute adult mental health units: a review. Issues Ment. Health Nurs. 32 (7), 424–435. Dack, C., Ross, J., Papadopoulos, C., Stewart, D., Bowers, L., 2013. A review and meta-analysis of the patient factors associated with psychiatric in-patient aggression. Acta Psychiatr. Scand. 127 (4), 255–268. Department of Health, 2014. Positive and Proactive Care: Reducing the Need for Restrictive Interventions. Department of Health, London. Dugdill, L., Brettle, A., Hulme, C., McCluskey, S., Long, A.F., 2008. Work- place physical activity interventions: a systematic review. Int. J. Workplace Health Manag. 1 (1), 20–40. Flood, C., Bowers, L., Parkin, D., 2008. Estimating the costs of conﬂict and containment on adult acute inpatient psychiatric wards. Nurs. Econ. 26 (5), 325–330. Funding: National Institute of Health Research, RP-PG- 0707-10081. Hadﬁeld, J.D., 2010. MCMC methods for multi-response generalized linear mixed models: the MCMCglmm R package. J. Statist. Software 32 (2), 1–22. Conﬂicts of interest: None declared. Huckshorn, K.A., 2005. Six Core Strategies to reduce the use of Seclusion and Restraint Planning Tool. National Technical Assistance Center, National Association of State Mental Health Program Directors, pp. 1–10. Appendix A. Supplementary data James, K., Stewart, D., Bowers, L., 2012. Self-harm and attempted suicide within inpatient psychiatric services: a review of the literature. Int. J. Ment. Health Nurs. 21 (4), 301–309. Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j. ijnurstu.2015.05.001. ( ) Kellam, S.G., Durell, J., Sharder, R.I., 1966. Measurement of staff attitudes and the clinical courses of patients on a psychiatric ward. Am. J. Psychother. 20 (1), 169–183. Lee, S., Wright, S., Sayer, J., Parr, A., Gray, R., Gournay, K., 2001. Physical restraint training in English and Welsh psychiatric intensive care and regional secure units. J. Ment. Health 10, 151–162. 3. Discussion Decreased conﬂict means fewer injuries to patients and staff from violence, self-harm, suicide, etc., a better patient experience due to less frequent use of force and coercion by staff, including high risk procedures such as manual restraint (Paterson et al., 2003), and a signiﬁcant release of staff time from dealing with conﬂict and containment to more positive and productive activities (Flood et al., 2008). Full instructions on how to use the Safewards interventions are freely available online, supported by instructional videos, downloadable docu- ment templates, planning and implementation guidance, and a web based forum offering support (www.safewards. net). Independent replication of the results in a further trial would make them more secure, and we recommend good quality evaluations in places where Safewards is implemented. Van De Sande, R., Nijman, H.L.I., Noorthoorn, E.O., Wierdsma, A.I., Hel- lendoorn, E., van der Staak, C., Mulder, C.L., 2011. Aggression and Ware, J.E., Kosinski, M., Turner-Bowker, D.M., Gandek, B., 2002. How to Score Version 2 of the SF-121 Health Survey (With a Supplement Documenting Version 1). Quality Metric Incorporated, Lincoln, RI. White, I.R., Horton, N.J., Carpenter, J., Pocock, S.J., 2011. Strategy for intention to treat analysis in randomised trials with missing outcome data. Br. Med. J. (Clin. Res. Ed.) 342, d40. L. Bowers et al. / International Journal of Nursing Studies 52 (2015) 1412–1422 References Van De Sande, R., Nijman, H.L.I., Noorthoorn, E.O., Wierdsma, A.I., Hel- lendoorn, E., van der Staak, C., Mulder, C.L., 2011. Aggression and 1422 L. Bowers et al. / International Journal of Nursing Studies 52 (2015) 1412–1422 seclusion on acute psychiatric wards: effect of short-term risk assess- ment. Br. J. Psychiatry 199 (6), 473–478. Van Der Merwe, M., Muir-Cochrane, E., Jones, J., Tziggili, M., Bowers, L., 2013. Improving seclusion practice: implications of a review of staff and patient views. J. Psychiatr. Ment. Health Nurs. 20 (3), 203–215.
W4289543646.txt	https://cdr.lib.unc.edu/downloads/zk51vk10v	en		“That’s probably what my mama’s lungs look like”: how adolescent children react to pictorial warnings on their parents’ cigarette packs	Carolina Digital Repository (University of North Carolina at Chapel Hill)	2,018	cc-by	5,631	Brodar et al. BMC Public Health (2018) 18:1125 https://doi.org/10.1186/s12889-018-6011-7 RESEARCH ARTICLE Open Access “That’s probably what my mama’s lungs look like”: how adolescent children react to pictorial warnings on their parents’ cigarette packs Kaitlyn E. Brodar1,2 , M. Justin Byron2,3, Kathryn Peebles6, Marissa G. Hall2,4, Jessica K. Pepper5 and Noel T. Brewer2,4* Abstract Background: Pictorial cigarette pack warnings discourage smoking, but most evidence comes from studies of adults. Our qualitative study explored adolescents’ reactions to pictorial warnings on their parents’ cigarette packs. Methods: We interviewed 24 adolescents whose parents received pictorial warnings on their cigarette packs as part of a randomized clinical trial. We conducted a thematic content analysis of the interview transcripts. Results: Pictorial cigarette pack warnings led adolescents to imagine the depicted health effects happening to their parents, which elicited negative emotions. The warnings inspired adolescents to initiate conversations with their parents and others about quitting smoking. Adolescents believed the warnings would help smokers quit and prevent youth from starting smoking. Some current smokers said the warnings made them consider quitting. Conclusions: Conversations about the pictorial warnings may amplify their effectiveness for smokers, their adolescent children, and friends of the adolescent children. Cigarette pack warnings may reach a broad audience that includes adolescent children of smokers. Keywords: Tobacco, Cigarettes, Pictorial warnings, Adolescents, Qualitative research Background Adolescents, particularly those with parents who smoke, are a key population to target with tobacco prevention messages and policy interventions. Adolescents with a smoking parent are more likely to experiment with smoking at an early age and more likely to become regular smokers than those without a smoking parent [1]. There is a dose-response relationship, such that the adolescents’ likelihood of smoking increases with each additional year exposed to parental smoking. On the positive side, research indicates that parental smoking * Correspondence: ntb@unc.edu 2 Department of Health Behavior, Gillings School of Global Public Health, University of North Carolina, 325 Rosenau Hall CB7440, Chapel Hill, NC 27599, USA 4 Lineberger Comprehensive Cancer Center, University of North Carolina, Chapel Hill, NC 27599, USA Full list of author information is available at the end of the article cessation may lower the risk of adolescent smoking initiation [2, 3]. Pictorial (graphic) warnings on cigarette packs are a promising approach for changing antecedents to smoking behavior and reducing smoking [4–7]. Most research on pictorial warnings comes from studies of adults. The existing studies of adolescents have largely been conducted in countries with active pictorial warning regulations [8–12]. One study of US adolescents indicated that high-emotion pictorial warnings increase perceptions of risk and intentions to quit compared to text-only warnings [13]. Another found that strong negative emotions mediate the relationship between perceived graphicness of cigarette warning labels and increased negative beliefs about smoking [14]. One limitation of past research on adolescents’ responses to pictorial warnings is that studies have been conducted almost exclusively in laboratory settings [15]. Also, there is a lack of research that © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Brodar et al. BMC Public Health (2018) 18:1125 specifically investigates how the vulnerable population of children of smokers react to pictorial warning labels. Thus, a gap remains in understanding how US adolescents, especially those whose parents are smokers, react to and think about pictorial warnings when exposed to them in a natural setting. Our randomized clinical trial (RCT) of pictorial cigarette pack warnings among 2149 adult US smokers found that pictorial warnings led to more quit attempts and quitting than text-only warnings [5]. The RCT presented a unique opportunity to interview adolescent children of parents who participated in the trial. A quantitative study of these adolescents found that exposure to pictorial warnings was associated with more attention and greater recall [15]. Here, we examine the qualitative findings from in-depth interviews with a subset of these adolescents to understand their reactions to and experiences with pictorial warnings, advancing the science of how pictorial warning labels work. Methods Participants and procedures From December 2014 to September 2015, we recruited 24 adolescents with a parent, guardian, or other household member who participated in our RCT of pictorial cigarette pack warnings, which enrolled a convenience sample of adult smokers in California and North Carolina, US [16]. The RCT affixed pictorial-and-text versus text-only warning labels to participants’ cigarette packs for 4 weeks [5, 17]. A participant received the same warning for the duration of the study. The four pictorial warnings, that included text and imagery Fig. 1 Pictorial warnings used in the randomized clinical trial Page 2 of 8 proposed by the US Food and Drug Administration, showed healthy and unhealthy lungs, a person with a tracheotomy, a sick person in a hospital bed, and a diseased mouth (Fig. 1). We contacted RCT participants who had adolescents ages 13–17, usually the adolescent’s mother or father, referred to hereafter as the “parent.” We obtained verbal consent from the parent for the adolescent’s participation and then verbal assent from the adolescent. We stopped recruitment of adolescents once we reached saturation (i.e., when we stopped hearing new information). We conducted qualitative interviews by phone with adolescents who reported seeing the pictorial warnings. Both interviewers (KP, MGH) were graduate students trained in qualitative interviewing techniques. Interviewers followed a guide that began with an ice-breaker question about how participants felt about their parents’ smoking and then addressed four content areas: attention to and thoughts about pictorial warnings, emotional reactions to the warnings, conversations sparked by the warnings, and perceptions of warning effectiveness. The interview started with the following instructions: “For the time that we’re on the phone, I’d like to ask that you give me your full attention and not do other things that would distract you, like be on a computer or watch TV. Please also pick a place where you can talk privately.” Additional file 1: Appendix A contains the interview guide used with study participants. The average duration of the interviews was 26 min (range: 15–36 min). We digitally recorded the interviews and an independent company transcribed them. Brodar et al. BMC Public Health (2018) 18:1125 Participants received a $40 incentive for completing the interview. The University of North Carolina’s institutional review board approved the study. Data analysis We conducted a thematic content analysis of the data, creating emergent codes under our pre-specified research questions [18]. We applied Green and Thorogood’s five criteria for qualitative analysis [18]. First, the research team created a codebook built from the questions in the interview guide and then KB and MJB each read and coded the same three transcripts using NVivo Pro v. 11 (QSR International) to create codes relevant to each content area. Next, KB and MJB discussed the results and revised the codebook. KB used the revised codebook to code three new transcripts, and then the coders met to discuss and finalize the codebook. KB then used the final codebook to code all remaining transcripts. Finally, we employed quote matrices to summarize findings and highlight examples of key themes. We organized the findings into the categories of the Tobacco Warnings Model (i.e., attention to warnings, negative emotional reactions, social interactions, and discouragement from smoking), which describes the psychological processes involved in how cigarette pack warnings work [19]. Results Most adolescents (n = 16) were non-smokers (had never tried cigarettes), five were ever-smokers (had tried cigarettes but had not smoked within the past 30 days), and three were current smokers (smoked within the past 30 days, Table 1). Most participants were Black (n = 16) and living in low-income households (n = 15). Most adolescents reported seeing their parents’ labeled packs when their parents pulled out cigarettes to smoke or when the packs were left around the house. Two adolescents only viewed the warnings because their parent told them about being in the trial. Nearly all adolescents expressed concern and anxiety about their parents’ smoking. They worried that their parents could become sick or die from smoking-related diseases. Many described watching grandparents or other relatives who smoked die from cancer or lung disease and expressed fear that their parents could experience the same consequences. Teens had extensively thought about and formed opinions about their parents’ smoking. Some felt their parents should stop smoking in order to stay alive and raise their children: “I just wish my dad and my mom’d stop smoking. I need them to be here.” (female non-smoker, age 14). While still communicating concern for their parents’ health, a few of the teens said that they respected their parent’s choice to smoke or explained how their parent needed to smoke to manage their stress. For example, one teen described Page 3 of 8 Table 1 Participant characteristics (n = 24) Number Age, mean years (SD) 15 (1.5) 13 5 14 4 15 5 16 4 17 6 Gender Male 14 Female 9 Respondent stated “gender-neutral” 1 Ethnicity Not Hispanic 22 Hispanic 2 Race Asian 2 Black 16 White 4 Other/multiracial 2 Low income household (≤ 150% of Federal Poverty Level) No 9 Yes 15 Smoking status Non-smoker 16 Ever-smoker 5 Current smoker 3 Study site California 11 North Carolina 13 Time between parent’s completion of RCT and adolescent’s interview, mean days (SD) 20 (16) how smoking served as a coping mechanism for his dad, a veteran: “I mean, he’s been in the war and stuff. . . he’s smokin’ cigarettes for a better reason other than ‘I want to smoke.’” (male ever-smoker, 17). Teens paid attention to and thought about pictorial warnings When we asked about the pictorial warnings, teens described how the warnings caught their attention and were difficult to ignore: “It’s right there. It’s right in front of the box. You can’t miss it.” (male ever-smoker, 16). Teens believed smoking was dangerous before seeing the warnings, but the warnings reinforced their existing beliefs and increased the reality of the health consequences: “It made me think, like, this is really possible. This isn’t fiction.” (male non-smoker, 17). One teen Brodar et al. BMC Public Health (2018) 18:1125 described how the warnings strengthened her negative beliefs about smoking: You don’t really get to visualize it when someone just says, ‘you can get risks by smoking’ but when you actually see something like [the pictorial warning], you actually get a new perspective . . . it makes you worried and afraid and a little bit scared and nervous that maybe that could happen to you and you just think about how you don’t want that to happen to you and your parents and your family and friends. (female non-smoker, 13). Like many of the teens, this participant indicated that thinking about the warning label and its depicted health consequences happening to her family and friends stirred up negative emotions. Pictorial warnings elicited negative emotional reactions Teens described negative emotions including fear, disgust, shame, anxiety, and sadness after seeing the pictorial warnings. Most of the adolescents expressed feeling anxious or sad, concerned that the specific health consequence depicted in the warning could happen to them or to their parents. Some teens said that although they had previously seen and heard warnings about smoking, the pictorial images on the cigarette packs were more realistic than other warnings, enabling them to visualize the consequences of smoking and causing a stronger emotional reaction. For example, teens who saw the warning with unhealthy lungs expressed anxiety that their parents’ lungs might look like the lungs on the warning: . . . that’s probably what my mama’s lungs look like. (male non-smoker, 17). . . . it made me sad to think what [my dad’s] lungs already look like or what possible disease or cancers he has from smoking. (male non-smoker, 16). [My mom] could already be developing heart disease or lung cancer or something like that, and you know, that’s something that any kid would be scared of, to see their mother be diagnosed with some, like, crazy disease like that. (male ever-smoker, 16). Similarly, an adolescent exposed to the sick person warning explained that she had “just like an anxious feeling. ‘Cause you don’t want your parent to get cancer. Especially, like, if it’s something that they can control.” (female non-smoker, 15). Nearly every participant expressed concerns and fears evoked by the warning about their own health or their parents’ health. Teens also described Page 4 of 8 feelings of shame, disappointment, and embarrassment about their parents’ smoking after seeing the pictorial warnings. Many believed their parents had control over the decision to smoke and were frustrated that their parents continued smoking despite seeing the consequences so clearly depicted on their cigarette packs. For instance, one adolescent said that when her mother took out her cigarettes to smoke, people asked about the image, requiring the mother to explain the picture. The teen expressed frustration and embarrassment that her mother could explain why smoking was dangerous yet continued to smoke. One of the current smokers also reported feeling ashamed about their own personal smoking habits, and said this shame could be productive for quitting: Whenever I see warning labels, it kind of makes me feel a little ashamed to smoke so many cigarettes. ‘Cause, well, when you smoke a lot of cigarettes for a good amount of time you can feel your lungs and your throat begin to. . . you can feel the, like, smoker’s voice and the smoker’s cough. . . But [being ashamed] also does help because I am trying to quit and it is more motivation. (current smoker who identified as “gender-neutral,” 16). The negative emotions that resulted from thinking about the warnings seemed to motivate teens to both consider their own behavior and to talk with others about smoking. Pictorial warnings sparked conversations Most of the teens reported at least one instance in which they initiated a conversation with their parents because of the pictorial warning. In these conversations, teens expressed concerns that the health effects on the warnings could happen to the parents. The conversations frequently revolved around the teen’s desire for the parent to stop smoking. One teen described initiating a conversation after he saw his mother’s cigarette pack with a pictorial warning on the kitchen table. He said the image was a sad reminder of his grandfather, who was also a smoker, dying from lung cancer: I said, ‘Mom I really would like it if you would put that away ‘cause I don’t like seeing it because it bring back bad memories of my grandad passing and I really would like it if you stopped—I mean, if you would stop smoking.’ (male non-smoker, 17). Teens told their parents that what was on the warning could happen to the parents. In some cases, these conversations provided opportunities for children to Brodar et al. BMC Public Health (2018) 18:1125 discuss their parents’ progress toward quitting. For example, a 17-year-old ever-smoker male said, “I said that’s what your lungs is probably looking like. . . she said she would finally quit. So I’m. . . I think she’s getting the message.” Similarly, a 16-year-old nonsmoker male described how he encouraged his mother to cut back on smoking after being exposed to the warnings, saying, “Well she usually tells me how many she smoked… I’m like, ‘Yeah. Good job.’” One girl whose mother had quit smoking by the end of the trial felt the conversation she had with her mother was very influential: I showed her the picture and I said, ‘That could be, like, you one day, and there could be some serious, maybe even devastating disease.’ . . . She said, “I think that [quitting] it’s a really good idea, because this could be me, and I don’t want this to happen to me at all.” (female non-smoker, 13) In one case, the warning sparked a conversation focused on why the teen should avoid smoking in the future: (Interviewer) So what did your grandma think about the warning label? She told me she hoped I never smoke cigarettes. (Interviewer) Yeah. Did she say anything about the warning label? She said my lungs would look like that if I smoke. If I don’t smoke, my lungs would be healthy. (female current smoker, 14). Several teens reported talking to their friends about smoking after seeing the warnings. For instance, one teen talked about showing the warning to his friend to encourage him to quit smoking: My friend just kind of recently started smoking so I wanted him to see the pack. And when I did show him the pack of cigarettes, it made him kind of think about it and he kind of like put the pack down for a little bit once he seen it . . . seeing that skinny man, I guess, just like me, made him sad and made him like, ‘well I don’t want to become or look like that.’ (male non-smoker, 17). In sum, the pictorial warnings sparked numerous conversations between adolescents and their family members as well as their friends. These conversations often focused on the powerful images and the need to avoid or quit smoking. Page 5 of 8 Teens largely thought pictorial warnings were effective at discouraging smoking Teens felt that pictorial warnings would help discourage them and others from smoking. Non-smoking teens said the warnings reinforced their decision not to smoke: It has really made me think that I definitely don’t want to smoke at all seeing that cigarette pack. I don’t want to be around the smoke, I don’t want to smoke, I don’t want nobody I love to smoke. (male non-smoker, 17). I guess definitely it holds me back from smoking. It doesn't make me want to smoke at all, especially because I do athletics, too, and once I see the lungs, which I actually saw, it was just like wow, I really don't want to smoke at all. It's nasty to me, to be honest. (male non-smoker, 16). Two of the current smokers described how the warnings made them reconsider their own smoking. One talked about how the warnings were an “eye-opener,” (male current smoker, 15) and the other discussed how seeing the warnings added to previous messages: “I’ve seen warning labels like that before, but I guess the more I see them, the more I’m consciously aware that that’s what I’m doing to myself.” (current smoker who identified as “gender-neutral,” 16). When asked how the warning changed how she thought about smoking, another current smoker indicated that it made her want to quit smoking: “Just the way it looked. . . I don’t ever want to smoke.” (female current smoker, 14). The visual imagery resonated with the teens and appeared to heighten their awareness of the seriousness of the risks involved. Most adolescents felt that the warnings would convince smokers to quit and prevent people, especially teens, from starting to smoke. For example, a 17-year-old male who had recently quit after 5 years of smoking said that if his very first pack of cigarettes had displayed a similar picture, he probably would not have started smoking. He said the warnings allow someone to make a more conscious decision about smoking: . . . it kind of gives people a heads up so they’re not walking into war with a blindfold on. It’s kind of like helping people out in a way. Just it’s a fair warning: ‘Hey look, go ahead and smoke this but this could happen if you keep smoking it.’ (male ever-smoker, 17). Some adolescents said that the pictorial warnings would have more impact than the current Surgeon General warnings: Brodar et al. BMC Public Health (2018) 18:1125 Because you know how every pack has a warning label, that’s a warning label, but it doesn’t really give your mind the image. The actual image that shows you what can happen. It may give you that extra push, like, I don’t want to smoke a cigarette. (male non-smoker, 17). A few teens, however, expressed doubts about the warnings’ effectiveness. For instance, one 15-year-old current smoker said that the warning would not have prevented him from starting to smoke at age 12 because he “had it pretty hard growing up” so he “didn’t really care.” Another current smoker found the warning personally effective, but did not think the warning would affect her parents: “They’d just look at the pack, and [think] like ‘that ain’t nothing to me. My lungs is already black.’” One teen mentioned potential limitations to the effectiveness of the warnings for smokers, given the addictiveness of cigarettes: Well, I think it could potentially stop some people from smoking, but I think that if someone has already started to smoke, then the warning probably won’t make them stop smoking. I mean that it’s a very addictive thing. (male non-smoker, 13). Similarly, a few teens said that most smokers already know that smoking has negative consequences, so viewing the warning may not produce a change. The teens largely supported the idea of the government requiring pictorial warnings, although one 13-year-old non-smoker male thought that cigarette companies should be able to decide for themselves how to label packages: “That’s cigarette companies’ own business. . . they’re allowed to do what they want. But I do think it can be helpful. It can stop people from smoking. I don’t know how often, but I do think it can.” Despite expressing some doubts, teens generally believed that pictorial warnings would be worthwhile even if they did not deter smoking in every case. Discussion In interviews with US adolescent children of smokers, we found that teens worried about their parents’ smoking, and pictorial cigarette pack warnings heightened this worry. Teens paid attention to the pictorial warnings and experienced strong negative emotional reactions, which typically resulted from thinking about the health consequence depicted on the warning happening to their parent or to them. The pictorial warnings also inspired numerous conversations with family and friends about quitting smoking. Most of the teens in the study believed that the pictorial warnings would be effective. Several said that the warnings reinforced their existing negative Page 6 of 8 impression of smoking and some of the smokers said they made them consider quitting. Teens reported feeling anxiety, sadness, disgust, and shame in response to seeing the warnings. Prior research shows that negative emotions are an important part of how pictorial warnings exert their effects [20–22], acting as a mediator between the warnings and outcomes such as intentions and behavior [19, 23–25]. The current study extends these findings by describing the natural language adolescents use to describe the emotions they experience in response to pictorial warnings. It also provides insight about why adolescents experience these negative emotions when they view these warnings; namely, adolescents visualize the health effects shown in the image as happening to themselves, their parents, or other people they know who smoke. We found that pictorial warnings cued conversations among adolescents [15], building on previous research that pictorial warnings spark conversations among adults [5, 26–28] (Morgan JC, Golden SD, Noar SM, Ribisl KM, Southwell BG, Jeong M, Hall MG, Brewer NT: Conversations about pictorial cigarette pack warnings: Theoretical mechanisms of influence impact quit attempts, submitted). Pictorial warnings offered an important opportunity for adolescents to broach the topic of smoking cessation with their parents (and sometimes with their friends). In these conversations, teens expressed their fears and frustrations with their parents and encouraged them to quit smoking [29]. Researchers have described teen-initiated conversations about smoking as “parenting the parent.” In a qualitative study of Canadian youth, researchers found that expressing concerns that the parent might die was a strategy used by teens to encourage their parents to quit smoking. In contrast to that study, which found that most youth who initiated these conversations did not feel their efforts were worthwhile, in our study many teens thought that expressing their concerns would help parents quit. It may be that the pictorial warnings legitimate and make the concerns more concrete, potentially enhancing the effectiveness of teens’ conversations. Considering the addictiveness of cigarettes [30], however, future research should explore the impact of conversations about pictorial warnings initiated by adolescents on subsequent quitting among parents and friends. To our knowledge, ours is the first qualitative study about how US adolescent children of smokers react to pictorial warnings. Additionally, our study examined how teens reacted to pictorial warnings in their own natural environment rather than in a laboratory setting. Furthermore, our sample included a diverse group of teens with respect to race and poverty, although the generalizability of our findings remains to be established. Another limitation is that our sample contained only three current smokers; future research should examine reactions to pictorial warnings in a larger sample of adolescent smokers. We also conducted the interviews by phone, which may have reduced Brodar et al. BMC Public Health (2018) 18:1125 Page 7 of 8 the potential for rapport between interviewers and participants. Finally, because smoking is socially stigmatized, participant responses may not have been fully candid. Competing interests Dr. Brewer has served as a paid expert consultant in litigation against tobacco companies. The authors declare that they have no competing interests. Conclusions Adolescents who viewed pictorial warnings on cigarette packs felt negative emotions and visualized the health consequences of smoking. The warnings provided an important opportunity for adolescents to talk to their parents about quitting smoking. This finding suggests that pictorial warnings may have a broad reach, affecting not only smokers but also their children and potentially others in their social networks. The present study adds to the large body of evidence supporting pictorial warnings as a meaningful policy tool to reduce tobacco use. Publisher’s Note Additional file Additional file 1: Appendix A. Interview guide. This file contains the interview guide used with study participants. (DOCX 22 kb) Abbreviations RCT: Randomized clinical trial; US: United States Acknowledgments The authors thank the research participants for taking part in this study. We thank Seth Noar for his leadership role on the parent trial and Clare Barrington for her comments on an earlier draft. Funding Research reported in this publication was supported by grant number P50CA180907 from the National Cancer Institute and the FDA Center for Tobacco Products (CTP). T32-CA057726 from the National Cancer Institute of the National Institutes of Health supported MGH’s time writing the paper. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH or the Food and Drug Administration. The funding agencies played no role in the design of the study, data collection, data analysis and interpretation, or in the writing of the manuscript. Availability of data and materials Due to our university’s requirements on grant funded research, we can share the study data only with a signed data use agreement. Investigators wishing to access the data may contact the first or last author of the paper. Authors’ contributions NB designed the study. KP, MJB, and JKP developed the interview guide. KP and MGH conducted the interviews. KB and MJB conducted data analysis and KB drafted the manuscript. All authors provided substantial feedback on manuscript drafts. All authors read and approved the final manuscript. Ethics approval and consent to participate The University of North Carolina’s institutional review board approved the study (study number 14–2011). We obtained verbal consent from the parent for each adolescent’s participation and then verbal assent from each adolescent. The UNC IRB approved the procedure for obtaining verbal consent from participants. During the call, research staff reviewed the information in the verbal assent scripts with participants, adolescents provided verbal assent over the phone, and study staff documented the assent in a log. The rationale for the waiver of written consent came from a position paper from the Society for Adolescent Medicine stating that adolescents are capable of providing informed consent for low-risk studies. Consent for publication Not applicable. Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Author details 1 Department of Psychology, University of Miami, 5665 Ponce de Leon Drive, Coral Gables, FL 33146-0751, USA. 2Department of Health Behavior, Gillings School of Global Public Health, University of North Carolina, 325 Rosenau Hall CB7440, Chapel Hill, NC 27599, USA. 3Department of Family Medicine, School of Medicine, University of North Carolina, 590 Manning Dr., Chapel Hill, NC 27599, USA. 4Lineberger Comprehensive Cancer Center, University of North Carolina, Chapel Hill, NC 27599, USA. 5RTI International, 3040 E. Cornwallis Road, Research Triangle Park, NC 27709, USA. 6Department of Epidemiology, School of Public Health, University of Washington, Seattle, WA 98195, USA. Received: 6 June 2018 Accepted: 2 September 2018 References 1. Mays D, Gilman SE, Rende R, Luta G, Tercyak KP, Niaura RS. Parental smoking exposure and adolescent smoking trajectories. Pediatrics. 2014; 133(6):983–91. 2. Kandel DB, Griesler PC, Hu M-C. Intergenerational patterns of smoking and nicotine dependence among US adolescents. Am J Public Health. 2015; 105(11):E63–72. 3. Chassin L, Presson C, Rose J, Sherman SJ, Prost J. Parental smoking cessation and adolescent smoking. J Pediatr Psychol. 2002;27(6):485–96. 4. Noar SM, Hall MG, Francis DB, Ribisl KM, Pepper JK, Brewer NT. Pictorial cigarette pack warnings: a meta-analysis of experimental studies. Tob Control. 2016;25(3):341–54. 5. Brewer NT, Hall MG, Noar SM, et al. Effect of pictorial cigarette pack warnings on changes in smoking behavior: a randomized clinical trial. JAMA Intern Med. 2016;176(7):905–12. 6. Noar SM, Hall MG, Brewer NT. Pictorial cigarette pack warnings have important effects. Am J Public Health. 2015;105(3):e1. 7. Hammond D. Health warning messages on tobacco products: a review. Tob Control. 2011;20(5):327. 8. Moodie C, Mackintosh AM, Hastings G. Adolescents’ response to pictorial warnings on the reverse panel of cigarette packs: a repeat cross-sectional study. Tob Control. 2015;24(e1):e93–7. 9. Vardavas CI. Adolescents perceived effectiveness of the proposed European graphic tobacco warning labels. Eur J Pub Health. 2009;19(2):212–7. 10. White V, Webster B, Wakefield M. Do graphic health warning labels have an impact on adolescents smoking-related beliefs and behaviours? Addiction. 2008;103(9):1562–71. 11. McCool J, Webb L, Cameron LD, Hoek J. Graphic warning labels on plain cigarette packs: will they make a difference to adolescents? Soc Sci Med. 2012;74(8):1269–73. 12. Andrews JC, Netemeyer RG, Burton S, Kees J. Effects of plain package branding and graphic health warnings on adolescent smokers in the USA, Spain and France. Tob Control. 2016;25(e2):e120. 13. Evans AT, Peters E, Shoben AB, Meilleur LR, Klein EG, Tompkins MK, Romer D, Tusler M. Cigarette graphic warning labels are not created equal: they can increase or decrease smokers’ quit intentions relative to text-only warnings. Nicotine Tob Res. 2017;19:1155–62. 14. Netemeyer RG, Netemeyer RG, Burton S, Andrews JC, Kees J. Graphic health warnings on cigarette packages: the role of emotions in affecting adolescent smoking consideration and secondhand smoke beliefs. J Public Policy Mark. 2016;35(1):124–43. 15. Peebles K, Hall MG, Pepper JK, Byron MJ, Noar SM, Brewer NT. Adolescents’ responses to pictorial warnings on their parents’ cigarette packs. J Adolesc Health. 2016;59(6):635–41. 16. Brodar K, Hall M, Butler E, Parada H, Stein-Seroussi A, Hanley S, Brewer N. Recruiting diverse smokers: enrollment yields and cost. Int J Environ Res Public Health. 2016;13(12):1251. Brodar et al. BMC Public Health (2018) 18:1125 17. Brewer NT, Hall MG, Lee JGL, Peebles K, Noar SM, Ribisl KM. Testing warning messages on smokers’ cigarette packages: a standardised protocol. Tob Control. 2015;25(2):153–9. 18. Green J, Thorogood N. Qualitative methods for health research. 2nd ed. Los Angeles: SAGE; 2009. 19. Brewer NT, Parada H, Hall MG, Boynton MH, Noar SM, Ribisl KM. Understanding why pictorial cigarette pack warnings increase quit attempts. Ann Behav Med. 2018. Advance online publication. https://doi.org/10.1093/ abm/kay032. 20. Cameron LD, Pepper JK, Brewer NT. Responses of young adults to graphic warning labels for cigarette packages. Tob Control. 2013;24(0):e14–22. 21. Wang A-L, Lowen SB, Romer D, Giorno M, Langleben DD. Emotional reaction facilitates the brain and behavioural impact of graphic cigarette warning labels in smokers. Tob Control. 2015;24(3):225–32. 22. Kees J, Burton S, Andrews JC, Kozup J. Understanding how graphic pictorial warnings work on cigarette packaging. J Public Policy Mark. 2010;29(2): 265–76. 23. Hall MG, Sheeran P, Noar SM, Boynton MH, Ribisl KM, Parada H, Johnson TO, Brewer NT. Negative affect, message reactance, and perceived risk: how do pictorial cigarette pack warnings change quit intentions? Tob Control. 2017. Advance online publication. https://doi.org/10.1136/tobaccocontrol-2017053972. 24. Emery LF, Romer D, Sheerin KM, Jamieson KH, Peters E. Affective and cognitive mediators of the impact of cigarette warning labels. Nicotine Tob Res. 2014;16(3):263–9. 25. Evans AT, Peters E, Strasser AA, Emery LF, Sheerin KM, Romer D. Graphic warning labels elicit affective and thoughtful responses from smokers: results of a randomized clinical trial. PLoS One. 2015;10(12):e0142879. 26. Hall M, Peebles K, Bach L, Noar S, Ribisl K, Brewer N. Social interactions sparked by pictorial warnings on cigarette packs. Int J Environ Res Public Health. 2015;12(10):13195–208. 27. Thrasher JF, Abad-Vivero EN, Huang L, O'Connor RJ, Hammond D, BansalTravers M, Yong H-H, Borland R, Markovsky B, Hardin J. Interpersonal communication about pictorial health warnings on cigarette packages: policy-related influences and relationships with smoking cessation attempts. Soc Sci Med. 2016;164:141–9. 28. Morgan JC, Southwell BG, Noar SM, Ribisl KM, Golden SD, Brewer NT. Frequency and content of conversations about pictorial warnings on cigarette packs. Nicotine Tob Res. 2018;20(7):882-7. 29. Woodgate RL, Kreklewetz CM. Youth’s narratives about family members smoking: parenting the parent- it’s not fair. BMC Public Health. 2012; 12(1):965. 30. Hyland A, Li Q, Bauer JE, Giovino GA, Steger C, Cummings M. Predictors of cessation in a cohort of current and former smokers followed over 13 years. Nicotine Tob Res. 2004;6:S363–9. Page 8 of 8
https://openalex.org/W2884404516	https://dipot.ulb.ac.be/dspace/bitstream/2013/279039/1/doi_262666.pdf	English	null	Predictive gravity models of livestock mobility in Mauritania: The effects of supply, demand and cultural factors	PloS one	2,018	cc-by	12,366	RESEARCH ARTICLE Predictive gravity models of livestock mobility in Mauritania: The effects of supply, demand and cultural factors Gae¨lle Nicolas1, Andrea Apolloni2, Caroline Coste3, G. R. William Wint4, Renaud Lancelot2, Marius Gilbert1,3 1 Spatial Epidemiology Lab (SpELL), Universite´ Libre de Bruxelles, Brussels, Belgium, 2 International Center for Agronomic Research and Development, CIRAD, Montpellier, France, 3 Fonds National de la Recherche Scientifique, Brussels, Belgium, 4 Environmental Research Group Oxford (ERGO)—Department of Zoology, University of Oxford, Oxford, United Kingdom gaelle.nicolas6@gmail.com * gaelle.nicolas6@gmail.com * gaelle.nicolas6@gmail.com Editor: Rachata Muneepeerakul, University of Florida, UNITED STATES Editor: Rachata Muneepeerakul, University of Florida, UNITED STATES Received: December 20, 2017 Accepted: June 8, 2018 Published: July 18, 2018 Copyright: © 2018 Nicolas et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Received: December 20, 2017 Accepted: June 8, 2018 Published: July 18, 2018 Copyright: © 2018 Nicolas et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability Statement: Data are part of a previously published study referred as Apolloni et al. (2018) and available on: https://doi.org/10. 5061/dryad.v4m56. The full citation of the previously published work is: Apolloni A, Nicolas G, Coste C, EL Mamy AB, Yahya B, EL Arbi AS, Baba Gueya M, Baba D, Gilbert M, Lancelot R (2018). Towards the description of livestock mobility in Sahelian Africa: Some results from a survey in Mauritania. PLOS ONE 13(1): e0191565. Abstract Animal movements are typically driven by areas of supply and demand for animal products and by the seasonality of production and demand. As animals can potentially spread infec- tious diseases, disease prevention can benefit from a better understanding of the factors influencing movements patterns in space and time. In Mauritania, an important cultural event, called the Tabaski (Aïd el Kebir) strongly affects timing and structure of movements, and due to the arid and semi-arid climatic conditions, the season can also influence move- ment patterns. In order to better characterize the animal movements patterns, a survey was carried out in 2014, and those data were analysed here using social network analysis (SNA) metrics and used to train predictive gravity models. More specifically, we aimed to contrast the movements structure by ruminant species, season (Tabaski vs. Non-Tabaski) and mode of transport (truck vs. foot). The networks differed according to the species, and to the season, with a changed proportion of truck vs. foot movements. The gravity models were able to predict the probability of a movement link between two locations with moderate to good accuracy (AUC ranging from 0.76 to 0.97), according to species, seasons, and mode of transport, but we failed to predict the traded volume of those trade links. The significant predictor variables of a movement link were the human and sheep population at the source and origin, and the distance separating the locations. Though some improvements would be needed to predict traded volumes and better account for the barriers to mobility, the results provide useful predictions to inform epidemiological models in space and time, and, upon external validation, could be useful to predict movements at a larger regional scale. OPEN ACCESS Citation: Nicolas G, Apolloni A, Coste C, Wint GRW, Lancelot R, Gilbert M (2018) Predictive gravity models of livestock mobility in Mauritania: The effects of supply, demand and cultural factors. PLoS ONE 13(7): e0199547. https://doi.org/ 10.1371/journal.pone.0199547 On this occasion, young rams are slaughtered in most families on the 10th day of the month "dhou al-hija", a religious holiday during the last month of the Muslim (lunar) calendar. The date of this festival changes each year according to the Gregorian (solar) calendar and strongly structures the volume of traded sheep during the year. Annual and seasonal differences are thus observed in the sheep trade flows. The country, and the whole Sahel, was hit by severe droughts in the 1960’s, 70’s and 80’s, and more limited droughts later on till 2017 [15]. The drought of 1970 was the main climatic event for the area. Since then, the area has been mostly in deficit of rainfall [16]. The series of droughts had a profound impact on the livestock (affecting 2/3 of the production) and human population. Cattle population dropped, whilst the population of small ruminants increased, the latter being more robust than the former to harsh climatic conditions. Apolloni et al. [17] described the livestock trading mobility for the year 2014 in Mauritania, highlighting that the main trading peak related to the Tabaski took place between August and December. During this period, the price of male lambs sharply increased, and the high demand strongly affected the trading network structure. In this paper, we aimed to understand how the Tabaski festivity changed the trade network i M it i d t th t f th A ll i t l [17] id d h i Funding: This study was funded by EU grant FP7- 613996 VMERGE and is catalogued by the VMERGE Steering Committee as VMERGE000 (http://www.vmerge.eu). The contents of this publication are the sole responsibility of the authors and don’t necessarily reflect the views of the European Commission. Competing interests: The authors have declared that no competing interests exist. In this paper, we aimed to understand how the Tabaski festivity changed the trade network in Mauritania compared to the rest of the year. Apolloni et al. [17] provided a comprehensive study of the Mauritania survey data, characterizing the seasonal trade network and ruminant flows within Mauritania, and between Mauritania and the neighboring countries. Here, we first provided a complementary description of inner flows using social network analysis indi- cators, by contrasting the Tabaski and non-Tabaski periods, the different ruminant species and modes of transport. Livestock mobility in Mauritania distribution, growth and connectivity of human populations have also been changing rapidly as result of demographic transitions, conflicts or migrations. Similarly, the distribution and connectivity of traded animal populations are strongly influenced by agricultural intensifica- tion and changes in trade patterns. The combined effect of these societal and environmental changes taking place simultaneously is difficult to assess, but some, particularly mobility of livestock and traditional trading practices, have been associated with the emergence and the spread of infectious diseases [1–4], and can have strong socio-economic impacts [5,6]. In addi- tion to these long-term trends, culture and tradition strongly shape societies at national, regional and global scale. For example, human population movement during specific periods such as Chinese Spring Festival [7], annual holidays [8], or religious feast around Christmas, Ramadan, Thanksgiving or Hindu Holy feast are known to cause substantially affect global mobility [9] with significant economic and epidemiological implications [10–12]. Large move- ments of animal populations are also linked to changes in the spatial pattern of food demand, which is anticipated by the market. In low-income countries, such as Sahelian African coun- tries, rapid changes in demand for animal products linked to cultural and traditional events therefore leads to a large number of animals–notably sheep, being slaughtered to meet the seasonal food demand. As a consequence, in the months and weeks preceding such events, trading of live animals is particularly intense. Due to the dry ecoclimate of the Sahelian area, agriculture and breeding strongly depend on the amount of rainfall and the availability of pasture. As a consequence, successive droughts also dramatically affect the Livestock flow. Another example is Madagascar, where kapsile is a traditional practice consisting in an animal barters linked to the labour need in a period when breeders are cash-poor and which strongly affects the patterns of trade flows [3,13]. In West Africa, similar practices, called loans, have been described as a large number of short term animal exchanges for reproduction, food sup- ply through milk production, animal traction, etc [14]. In Mauritania, the Muslim feast of Tabaski or Aïd el-Kebir, is a major cultural and traditional event that strongly influence trad- ing patterns and could have major impact on the spread of diseases. Introduction Many factors that may influence the dynamics and transmission of infectious diseases have been rapidly changing over the last decades. Alongside climate and land use changes, often considered in emerging infectious diseases literature as main drivers, other factors such as the PLOS ONE \| https://doi.org/10.1371/journal.pone.0199547 July 18, 2018 1 / 21 PLOS ONE \| https://doi.org/10.1371/journal.pone.0199547 July 18, 2018 Data collection A survey among veterinary officers was conducted in June 2015 by the National Office for Live- stock Research and Development (ONARDEL) to collect their knowledge of ruminant trade flows as reported in Apolloni et al. [17]. The survey aimed to monitor the movement patterns during the year 2014. Its results were recorded as a series of trade flow events, with the following information: i) the origin, ii) the destination, iii) the type of trade movement (farm, transit, mar- ket), iv) the frequency (annual, monthly, weekly), v) the species (cattle–Bos indicus, sheep, goat or camel—Camelus dromedarius), vi) the number of heads, vii) the date of the starting of the event, viii) the transportation mode (by truck or on foot) and ix) the latitude/longitude coordi- nates of the origin and destination. Three types of movements were recorded: i) between pasto- ral areas for grazing and/or reproduction, ii) from farm to market, iii) or from market to market. Transhumance movements aiming to gradually move herds for suitable pasture areas were not included. The database was cross-checked against sanitary certificates, the scientific documents describing transhumance patterns, and the knowledge of veterinarian staff [17]. Both national and international trade-flows were recorded and the transboundary movements were double-checked through surveys on transit sites between Senegal and Mauritania. Study area Mauritania is situated in the hyper-arid (Sahara) and arid (Sahel) ecozones [18], with low annual rainfall (0–400 mm) concentrated in a short rainy season (June-September). In the northern part of the country, the driest one, only short-cycle plants grow. Livestock, mainly camels and small ruminants, are reared moving around available water points and grazing areas. The southern area, more humid and greener is mostly exploited by transhumant herds. Most of cattle population, being less mobile and demanding more water and nutrients, is con- centrated in the southern area, mainly in the region around the river Senegal. Because of the harsh conditions, mobility is a key aspect of animal rearing in Mauritania. Animals are moved almost continuously among grazing areas to optimize the consumption of good quality nutri- ents. In the absence of slaughterhouses, stocking facilities and road infrastructures, animals are traded alive and butchered at consumption markets. Past droughts indirectly contributed to the growth of cities, in particular Nouakchott, due to the migration of previous farmers and herders from the countryside to urban areas in search of jobs. Because of this, Nouakchott, the capital city has seen its population dramatically exploding during and following the drought years. As of today, almost one quarter of the total population lives in the capital city (National Bureau of statistics http://www.ons.mr/) Maurita- nia is still recovering from the latest food crisis in 2011, affecting almost 1 million of its habi- tants. In 2018, the drought indicators were at the same levels as those of 2012, indicating that the food emergency is not completely over. The inadequate levels of rainfall and the continu- ous threat of the droughts force herders to sell their livestock, in particular small ruminants, due to shortages of suitable feeding areas. Second, we developed predictive models to estimate the probability of a trade connection between two spatial units (areas around markets and/or farms, etc.) based PLOS ONE \| https://doi.org/10.1371/journal.pone.0199547 July 18, 2018 2 / 21 Livestock mobility in Mauritania on their potential production and demand characteristics and different measures of the cost distance between them. on their potential production and demand characteristics and different measures of the cost distance between them. Additional data In 2013, the Ministry for Rural Development and Environment reported a population of 16.8 million sheep and goats, 1.8 million cattle, and 1.4 million camels for Mauritania to FAOSTAT 3 / 21 PLOS ONE \| https://doi.org/10.1371/journal.pone.0199547 July 18, 2018 Livestock mobility in Mauritania Fig 1. Population of human and ruminants in Mauritania. The size of the circle are proportional to the number of head located within the Voronoï polygon related to the spatial location of the trading network. https://doi.org/10.1371/journal.pone.0199547.g001 Fig 1. Population of human and ruminants in Mauritania. The size of the circle are proportional to the number of head located within the Voronoï polygon related to the spatial location of the trading network. https://doi.org/10.1371/journal.pone.0199547.g001 https://doi.org/10.1371/journal.pone.0199547.g001 [19]. Accordingly, and because no finer data was available at national level, we used the most recent version of the Gridded Livestock of the World database (GLW), where the subnational livestock statistics for Mauritania dates back to 2007, and were adjusted to match the FAO- STAT 2010 national totals [20,21]. The WorldPop database was used for the human popula- tion [22]. Both databases were aggregated at a spatial resolution of 0.083333 decimal degrees (i.e. approximately 10 km at the equator) (Fig 1). To estimate the cost paths potentially affect- ing mobility between different localities, we considered two main data sources: the friction layer of Nelson accessibility map, which quantifies the time needed to travel through each pixel [23], and the elevation from the GTOPO30 database (https://lta.cr.usgs.gov/GTOPO30). PLOS ONE \| https://doi.org/10.1371/journal.pone.0199547 July 18, 2018 Analysis In this study, we only considered inner movements of cattle, camels, sheep and goats within Mauritania. Social network analysis (SNA) [24] have proved to be of significant interest in animals movements analyses in the past decade [25–28]. Here, it was first used to describe the trade networks according to the species, the transport modality (using truck vs. walking), and the season (Tabaski vs. non-Tabaski). These mentioned periods were defined as strongly influenc- ing the livestock flow within the country by Apolloni et al. [17] who investigated the dataset regarding the occurrence of the Muslim festival for the year 2014. A set of network parameters were estimated for the different networks using the simplified definition provided in Wasser- man and Faust, (1994) [24]: PLOS ONE \| https://doi.org/10.1371/journal.pone.0199547 July 18, 2018 4 / 21 Livestock mobility in Mauritania • Diameter: a network-level parameter representing the greatest number of links in the short- est path between two nodes. • Average path length: a network-level parameter measuring the average number of steps along the shortest paths of all possible nodes pairs, i.e. the average number of nodes an actor has to trade through to connect to any other node. • The clustering coefficient: a node-level parameter of the density of local ties. It measures the probability that neighboring nodes of a node are connected. • The density: a network-level parameter measuring the proportion of observed links among the possible links between nodes, and indicates how strongly a network is connected. • Average degree: a network-level parameter quantifying the average number of links con- nected with a node in a network. Besides these global measures, other centrality measures highlighting the prominent role of nodes in the network were considered: node’s in- and out-degree (the number of connection towards and from each node), node’s in and out- weight (the volume of animal towards and from each node); node’s betweenness (the num- ber of shortest path passing through the node) and node’s eigenvector centrality (scoring the importance of nodes). Network’s vulnerability to target removal of nodes based on centrality measures and esti- mates of the size of the largest connected component were tested. The removal of specific nodes, and their links, cause the network to fragment in a set of smaller subnetworks. PLOS ONE \| https://doi.org/10.1371/journal.pone.0199547 July 18, 2018 Analysis Eq (2) can be rewritten as a model (3) with an intercept (a flow that would still be pres- ent when populations are equal to zero), and a set of predictor variables xk characterizing the origin or destination, with their associated coefficients βk. MIGij ¼ pa i pb j dg ij ð1Þ ð1Þ logðMIGijÞ ¼ api þ bpj gdij ð2Þ GijÞ ¼ b0 þ b1pi þ b2pj þ b3dij þ PK k 4bkxk ð3Þ logðMIGijÞ ¼ api þ bpj gdij ð2Þ ð2Þ logðMIGijÞ ¼ b0 þ b1pi þ b2pj þ b3dij þ PK k¼4bkxk ð3Þ ð3Þ In this study, we first aimed to estimate the probability of a trade connection between two nodes. Therefore, log(MIGij) was replaced with the logit of this probability and logistic regres- sion was used to estimate the coefficients. As the response, all pairs of connected nodes were coded with 1, and all other pairs of nodes were coded with 0. The analysis was split according to the main structuring factors of the networks, i.e. species, transport modality and season. For each sub-model, we tested seven combinations of predictors, considering them both at the ori- gin and destination with inclusion of a distance estimator ((i) great-circle distance, (ii) cost- path distance based on accessibility friction surface or (iii) cost-path distance based on eleva- tion friction surface). The different combinations of predictors that were tested in the models are shown in Table 1. The human and animal populations at the origin and destination were extracted from the Worldpop and GLW raster layers within the Thiessens’ polygons around each node (Fig 1). Thiessen’s polygons represent areas consisting of all points closer to the node than to any other node. These were used because the nodes did not correspond to any particular administrative division that could have been used (i.e. many nodes per admin unit). Although Thiessen’s polygon can produce somewhat misleading long shapes in desertic areas, this is not necessarily a problem as the livestock and human population demographics would be low in these areas anyway. In each sub-model, we first included the extracted animal popu- lation of each species at both the source and destination (cattle, sheep, goats, and camels). We used stepwise regression based on Akaike information criterion (AIC) to select a more parsi- monious model, with the lowest AIC. Analysis The size of the largest component can be thought as the maximum extent a disease can spread after the implementation of the control measure (vaccination of animal in the areas surrounding the nodes, market closure, etc.) [29]. Volkova et al. (2010) [30] introduced the notion of epidemic threshold (q) in veterinarian epidemiology. This parameter estimates the (minimum) proba- bility for a disease to be transmitted from one node to another to trigger an epidemic. The low- est the epidemic threshold the higher is the risk of an epidemics. This quantity depends on the heterogeneity of the network and the weight’s distribution. In the case of a weighted network the epidemic threshold can be estimated as: q ¼ hwouti hwin wouti Where h i indicates the average value and win, wout indicate node’s in- and out-weight, respectively. Following the same procedure as in Lancelot et al. (2017) [31], the invasion threshold for each month were estimated. Highlight on the role of occasional links were given (connections appearing just once per year). We plotted path intersections between different species, transports modalities and seasons, to highlight possible common or specific links for different combinations. In addition, metrics quantifying these intersections were estimated, such as the pairwise percentage of common and specific paths between two networks, respectively. Gravity models were used to estimate the probability of a link between two distinct nodes according to their features. These models were developed in the field of socio-economics and human migration studies [32,33]. They provide estimates for the flows of goods or people between two nodes, as a function of node-level variables (e.g. population size, socio-economic factors, etc.), and of the distance or movement cost between these nodes. Its most general for- mulation (Eq 1) shows the analogy with Newton’s gravity law. MIGij is the flow between the origin i and destination j, pi and pj are the population at the origin and destination, dij is the distance between i and j, and α, β and γ are model parameters. The equation is linearized into PLOS ONE \| https://doi.org/10.1371/journal.pone.0199547 July 18, 2018 5 / 21 Livestock mobility in Mauritania Eq (2) by a log-transformation, and its coefficients can be estimated with generalized linear models. Analysis In a second step, this analysis was repeated using the number of animals traded between these two locations as the response variable. To better differentiate the factors influencing the trade probability from those influencing its volume, the latter analysis was restricted to pairs of Table 1. General equation of the tested models. MODEL GENERAL EQUATION E1 Hpi + Hpj + dijg E2 Lpi + Lpj + dijg E3 Hpi + Hpj + Lpi + Lpj + dijg E4 Hpi + Hpj + Lpi + Lpj + dija E5 Hpi + Hpj + Lpi + Lpj + dijel E6 Δ(Hpi, Hpj) + Δ(Lpi, Lpj) + dijg E7 E3 + mode + mode: pi + mode: pj + mode: dijg pi: population at the origin (L: livestock, H: human); pj: population at destination (L: livestock, H: human); dij: distance (g: great circle distance, a: costhpath distance based on accessibility friction surface, el: costhpath distance based on elevation friction surface), mode: transport modality (by truck or on foot). htt //d i /10 1371/j l 0199547 t001 Table 1. General equation of the tested models. MODEL GENERAL EQUATION E1 Hpi + Hpj + dijg E2 Lpi + Lpj + dijg E3 Hpi + Hpj + Lpi + Lpj + dijg E4 Hpi + Hpj + Lpi + Lpj + dija E5 Hpi + Hpj + Lpi + Lpj + dijel E6 Δ(Hpi, Hpj) + Δ(Lpi, Lpj) + dijg E7 E3 + mode + mode: pi + mode: pj + mode: dijg pi: population at the origin (L: livestock, H: human); pj: population at destination (L: livestock, H: human); dij: distance (g: great circle distance, a: costhpath distance based on accessibility friction surface, el: costhpath distance based on elevation friction surface), mode: transport modality (by truck or on foot). https://doi.org/10.1371/journal.pone.0199547.t001 Table 1. General equation of the tested models. https://doi.org/10.1371/journal.pone.0199547.t001 PLOS ONE \| https://doi.org/10.1371/journal.pone.0199547 July 18, 2018 Analysis MODEL GENERAL EQUATION E1 Hpi + Hpj + dijg E2 Lpi + Lpj + dijg E3 Hpi + Hpj + Lpi + Lpj + dijg E4 Hpi + Hpj + Lpi + Lpj + dija E5 Hpi + Hpj + Lpi + Lpj + dijel E6 Δ(Hpi, Hpj) + Δ(Lpi, Lpj) + dijg E7 E3 + mode + mode: pi + mode: pj + mode: dijg pi: population at the origin (L: livestock, H: human); pj: population at destination (L: livestock, H: human); dij: distance (g: great circle distance, a: costhpath distance based on accessibility friction surface, el: costhpath distance based on elevation friction surface), mode: transport modality (by truck or on foot). pi: population at the origin (L: livestock, H: human); pj: population at destination (L: livestock, H: human); dij: distance (g: great circle distance, a: costhpath distance based on accessibility friction surface, el: costhpath distance based on elevation friction surface), mode: transport modality (by truck or on foot). PLOS ONE \| https://doi.org/10.1371/journal.pone.0199547 July 18, 2018 6 / 21 Livestock mobility in Mauritania locations with an existing trade link. All analyses were coded and carried out using R [34]. The “sna” package was used to describe and analyse the trade network [35]. Table 2. Number of national and international traded animals in Mauritania. Values represent the number of animals moved for each given period. Movements are defined as national if both the origin and destination are located within the Mauritanian border. Small ruminants (SR) include sheep and goats. Some of the records did not differentiate the species individually and are counted as “Cattle and SR” or “Mixed”. Cattle Cattle and SR Camels SR Mixed Total Truck 67 240 43 760 1 105 870 1 216 870 Aug-Dec 34 398 19 119 843 047 896 565 international 2 400 2 120 410 500 415 020 national 31 998 16 999 432 547 481 545 Jan-July 32 842 24 641 262 822 320 305 international 9 600 3 280 980 13 860 national 23 242 21 361 261 842 306 445 Foot 1 801 846 1 660 598 034 3 334 755 96 563 5 832 859 Aug-Dec 342 444 277 112 740 716 645 19 740 1 191 847 international 283 403 28 011 518 420 829 835 national 59 041 277 84 729 198 225 19 740 362 012 Jan-July 1 459 402 1 383 485 294 2 618 110 76 823 4 641 012 international 119 1123 271 755 2 286 295 3 749 172 national 268 279 1 383 213 539 331 815 76 823 891 839 Total 1 869 086 1 660 641 794 4 440 625 96 563 7 049 729 https://doi.org/10.1371/journal.pone.0199547.t002 Table 2. Number of national and international traded animals in Mauritania. Values represent the number of animals moved for each given period. Movements are defined as national if both the origin and destination are located within the Mauritanian border. Small ruminants (SR) include sheep and goats. Some of the records did not differentiate the species individually and are counted as “Cattle and SR” or “Mixed”. locations with an existing trade link. All analyses were coded and carried out using R [34]. The “sna” package was used to describe and analyse the trade network [35]. PLOS ONE \| https://doi.org/10.1371/journal.pone.0199547 July 18, 2018 Results The dataset consisted in 2,219 trade movements involving 7.1 million head. The subset of national movements, the focus of this analysis, included 1,178 movement events involving 2.1 million head (Table 2). International movements involved around 5 million head, sold or bought to or from Senegal, Mali, Ivory Coast, Guinea Bissau and Morocco. As the destination of these international movements within these countries was unknown, they could not be included in the network analysis or gravity models. Within Mauritania, the trading network was composed of 65 nodes and 84 unique paths. Transport by foot was the most represented mode, corresponding to 83% of the animal flow (Tables 2 and 3). The foot and truck transport modalities presented contrasted patterns. The largest share of foot movements being short to medium distance (0–200 km), whereas the opposite was observed for movements by truck, where the largest share was represented by movements > 500 km (Fig 2). Similarly, those con- trasting patterns somewhat matched the seasonal pattern. The Tabaski period represented 41% of the inner animal movements, with 57% of them by truck. In contrast, only 25% of the head were moved by truck outside the Tabaski period (Table 2). The Mauritanian network was weakly connected. The density value indicated that 2% of possible node pairs were actually connected, and the network level centrality parameters and clustering coefficient were low (Table 3). However, each of the exchange networks (full net- work, or species-specific sub-networks) contained a single component in which the average length of the shortest path between node pairs was lower than 2 links, the maximum value (diameter) being 5 links. The goat and camel trading networks were smaller with a diameter of 2 and 3 links. On average, in the full species network, a given node was directly connected with approximately 2 other nodes on average (average degree). Both the in-degree and in-weight distributions are right-skewed. Only 5 locations from the livestock trading network attracted more than 50% of the connections, and are also the destination for more than 60% of the PLOS ONE \| https://doi.org/10.1371/journal.pone.0199547 July 18, 2018 7 / 21 Livestock mobility in Mauritania Fig 3. Network’s centrality measures of the Mauritanian’s livestock trade. Node size show the importance of the measured centrality values. in and out-weight measures were scaled on the total volume of traded livestock; eigenvector centrality (centrality measure) were scored from 0 to 1, betweenness was considered for the fraction of paths passing through the node. h //d i /10 13 1/j l 0199 4 003 Fig 3. Network’s centrality measures of the Mauritanian’s livestock trade. Node size show the importance of the measured centrality values. in and out-weight Fig 3. Network’s centrality measures of the Mauritanian’s livestock trade. Node size show the importance of the measured centrality values. in and out-weight measures were scaled on the total volume of traded livestock; eigenvector centrality (centrality measure) were scored from 0 to 1, betweenness was considered for the fraction of paths passing through the node. https://doi.org/10.1371/journal.pone.0199547.g003 traded volume. Nouakchott, the main urban consumption market of the country, acts as hub for livestock mobility with 18 links which concentrate around 1/3 of the traded animals’ total volume (Fig 3, Table 4). With the highest centrality score, the capital city is the most important node of the network. Almost 2/3 of the nodes has at most one outgoing link, whilst Aleg and traded volume. Nouakchott, the main urban consumption market of the country, acts as hub for livestock mobility with 18 links which concentrate around 1/3 of the traded animals’ total volume (Fig 3, Table 4). With the highest centrality score, the capital city is the most important node of the network. Almost 2/3 of the nodes has at most one outgoing link, whilst Aleg and Table 4. List of nodes with highest values for centrality measures. Each column corresponds to a specific network (all species or by single species) Each line relates to a specific centrality measure. Only the name of the node corresponding with the largest value for each measure is reported. In the case of multiple nodes with same value of the measure, all the names are reported. Table 4. List of nodes with highest values for centrality measures. Each column corresponds to a specific network (all species or by single species) Each line relates to a specific centrality measure. Only the name of the node corresponding with the largest value for each measure is reported. Livestock mobility in Mauritania ber of links of the national movement networks in Mauritania. Links are provided according to the transport modality and -Tabaski: Jan–July). Table 3. Network parameters and number of links of the national movement networks in Mauritania. Links are provided according to the transport modality and Tabaski periods (Tabaski: Aug–Dec; Non-Tabaski: Jan–July). ALL SPECIES CATTLE SR SHEEP GOAT CAMEL Network-level parameter Number of nodes 65 65 65 65 65 65 Number of links 84 49 56 54 32 34 Diameter 5 4 5 5 3 2 Clustering coefficient 0.19 0.15 0.18 0.18 0.15 0.03 Average path length 1.84 1.56 1.73 1.73 1.34 1.13 Density 0.020 0.012 0.013 0.013 0.0077 0.0082 Average degree 2.65 1.51 1.72 1.66 0.985 1.05 Number of link Transport modality Truck 33 13 28 28 13 14 Foot 55 37 30 28 20 22 Intersection 2 2 2 2 1 2 Trading period Non-Tabaski 73 45 46 44 29 31 Tabaski 57 31 39 39 19 21 Intersection 44 27 14 29 16 18 Fig 2. Great-circle distance of the national movements by truck (left) and on foot (right) recorded in the survey. Distance are given in kilometers for each species (thick line: small ruminants, dashed line: cattle, thick dash line: camel). https://doi org/10 1371/journal pone 0199547 g002 Fig 2. Great-circle distance of the national movements by truck (left) and on foot (right) recorded in the survey. Distance are given in kilometers for each species (thick line: small ruminants, dashed line: cattle, thick dash line: camel). https://doi.org/10.1371/journal.pone.0199547.g002 PLOS ONE \| https://doi.org/10.1371/journal.pone.0199547 July 18, 2018 8 / 21 In the case of multiple nodes with same value of the measure, all the names are reported. ALL SPECIES CATTLE SR CAMEL INDEGREE Nouakchott Nouakchott Nouakchott Nouakchott INWEIGHT Nouakchott Selibabi Nouakchott Zouerate OUTDEGREE Aleg Aleg Kiffa Aleg Kiffa Kiffa Mmat Laakarich II M. Lahjar OUTWEIGHT Tintane Kaedi Tintane Nbeika BETWEENNESS Boutilimit Adel-Bagrou Boutilimit Boutilimit Mbout Nbeika EIGENVECTOR CENTRALITY Nouakchott Kaedi Nouakchott Nbeika https://doi org/10 1371/journal pone 0199547 t004 9 / 21 Livestock mobility in Mauritania Table 5. Correlation coefficients among centrality measures. Pearson correlations coefficients among centrality measure. Only significant (p-value <0.05) coefficients are reported. INDEGREE OUTDEGREE INWEIGHT OUTWEIGHT INDEGREE 0.88 OUTDEGREE 0.63 INWEIGHT 0.88 OUTWEIGHT 0.63 https://doi.org/10.1371/journal.pone.0199547.t005 Kiffa, in the southern region, are connected to other 7 localities, and Tintane has the largest out-weight. Finally, the betweenness distribution is right-skewed and Boutilimit appears to be the node with highest betweenness. Few of the centrality measures are significantly correlated (p-value<0.05) and reported in Table 5. Results of the percolation analysis on networks cohesion are shown in Fig 4. Nodes are removed based on the centrality measures (indegree, outdegree, incoming and outgoing Fig 4. Effect of targeted removal on the connected component size. Values on the x axis indicate the percentage of nodes removed (cumulative), together with their links. Values on y axis indicate the percentage of nodes in the largest connected sub-network, after the removal. Color indicates the removal procedure based on centrality measure score (starting from the highest score nodes) or randomly. Shaded areas correspond to 95% CI for the random procedure. https://doi.org/10.1371/journal.pone.0199547.g004 Fig 4. Effect of targeted removal on the connected component size. Values on the x axis indicate the percentage of nodes removed (cumulative), together with their links. Values on y axis indicate the percentage of nodes in the largest connected sub-network, after the removal. Color indicates the removal procedure based on centrality measure score (starting from the highest score nodes) or randomly. Shaded areas correspond to 95% CI for the random procedure. https://doi.org/10.1371/journal.pone.0199547.g004 Fig 4. Effect of targeted removal on the connected component size. Values on the x axis indicate the percentage of nodes removed (cumulative), together with their links. Values on y axis indicate the percentage of nodes in the largest connected sub-network, after the removal. Color indicates the removal procedure based on centrality measure score (starting from the highest score nodes) or randomly. Shaded areas correspond to 95% CI for the random procedure. https://doi.org/10.1371/journal.pone.0199547.g004 PLOS ONE \| https://doi.org/10.1371/journal.pone.0199547 July 18, 2018 10 / 21 Livestock mobility in Mauritania Fig 5. Species-level movement networks. Each diagram represents the movement links that are specific to the species (red: by truck, dark grey: by foot, both truck and foot: blue). The last plot shows the entire network of all species. https://doi.org/10.1371/journal.pone.0199547.g005 Fig 5. Species-level movement networks. All species: Association of small ruminants, Cattle and Camel trading networks. Each diagram represents the movement links that are specific to the species (red: by truck, dark grey: by foot, both truck and foot: blue). The last plot shows the entire network of all species. Fig 5. Species-level movement networks. Each diagram represents the movement links that are specific to the species (red: by truck, dark grey: by foot, both truck and foot: blue). The last plot shows the entire network of all species. https://doi.org/10.1371/journal.pone.0199547.g005 https://doi.org/10.1371/journal.pone.0199547.g005 volume, betweenness, eigenvector centrality) and, for comparison purposes, randomly. volume, betweenness, eigenvector centrality) and, for comparison purposes, randomly. Removing nodes in order of their incoming connections, incoming volume and centrality result as the most effective strategies of fragmentation. In the first case, removing less than 20% of the nodes (13 nodes) results in decomposing network in a set of subnetworks whom the largest one contains less than 10% of the nodes (7 nodes). Removing nodes in order of their incoming connections, incoming volume and centrality result as the most effective strategies of fragmentation. In the first case, removing less than 20% of the nodes (13 nodes) results in decomposing network in a set of subnetworks whom the largest one contains less than 10% of the nodes (7 nodes). In total, 94% of the goats trading network paths were shared with the sheep (Fig 5, Table 6). These two networks were merged into a single small ruminant trading network in the gravity models. A high degree of overlap between pairs of species networks was highlighted (Table 6). However, overlapping only involved 20% of the full set of trading links (Fig 5, Table 6). the different species. The values correspond to the percentage (italic) and the number (bracket) of common links and of species. Table 6. Comparison of the trading networks of the different species. The values correspond to the percentage (italic) and the number (bracket) of common links and difference between the trading network of the pair of species. network SPECIES 1 network SPECIES 2 INTERSECTION % (n) DIFFERENCE SP1/SP2 % (n) DIFFERENCESP2/SP1 % (n) CATTLE/SR 49 56 65 (32) 35 (17) 43 (24) CAMEL/SR 34 56 65 (22) 35 (12) 61 (34) CAMEL/CATTLE 34 49 47 (16) 53 (18) 67 (33) GOAT/SHEEP 32 54 94 (30) 6 (2) 44 (24) ALL SPECIES 84 - 20 (17) Table 6. Comparison of the trading networks of the different species. ation of small ruminants, Cattle and Camel trading networks. https://doi.org/10.1371/journal.pone.0199547.t006 All species: Association of small ruminants, Cattle and Camel trading networks. https://doi.org/10.1371/journal.pone.0199547.t006 The values correspond to the percentage (italic) and difference between the trading network of the pair of species. PLOS ONE \| https://doi.org/10.1371/journal.pone.0199547 July 18, 2018 PLOS ONE \| https://doi.org/10.1371/journal.pone.0199547 July 18, 2018 11 / 21 Livestock mobility in Mauritania Among the remaining links, many were species-specific: 30% of the small ruminant (17/56), 33% of the cattle (16/49) and 32% of the camel (11/34) links were not included in any of the other networks. Very few links were shared between the truck and foot networks and were identical for all species (2/84, Table 6). The role of nodes could change depending on the species considered. Previous results about most central nodes hold when we consider the “small ruminants only” network. Whilst Nouakchott remains always the most connected node, most of the volume of cattle and camels are directed towards Selibabi and Zouerate, respectively. The markets of Kaedi, for cattle, and Nbeika, for camels, become more central in the network (see S1 File and Table 4). In term of cohesiveness, specie-specific network are more vulnerable to the target removal of nodes. As in the all species case target removal based on indegree and inweight are the most efficient pro- cedures. In fact, in both cases, removing a limited quantity of nodes, 8 for camel’s network 10 for cattle’s one and 12 for small ruminants one, results in decomposing the respective networks in a set of subnetworks whom the largest one contains at most 8 nodes. This is particularly rel- evant for small ruminants and the cattle networks, whose largest subnetworks consist of 7 and 10 nodes, respectively, after the removal of only 3 most connected nodes. There was a high number of shared links between the Tabaski and non-Tabaski periods (44/84, Table 6). Both Tabaski and non-Tabaski networks included truck and foot movements, but their relative proportion was different. During the Tabaski period, there were fewer foot movements and higher truck movements (Fig 6). Very few trade connections involved both truck and foot movements in the two periods (Table 6, Figs 5 and 6). Fig 6. Movement networks according to the species and seasons. Note that for each season the links present in both defined transportation mode are not plotted and are the same for both defined season (red: by truck, dark grey: by foot, nf: number of movement links by foot, Jan-July: Non-Tabaski period; Aug-Dec: Tabaski period). Fig 7. Variations of network quantities along the year. Bottom: monthly epidemic threshold variations along the year; Centre Volume of animals traded during the month; Top number of links active. For each month, we have considered 2 networks: all network including occasional links appearing on that month; backbone, excluding occasional link. The different colors correspond to the quantity evaluated for the specific network. https://doi.org/10.1371/journal.pone.0199547.g007 Fig 7. Variations of network quantities along the year. Bottom: monthly epidemic threshold variations along the year; Centre Volume of animals traded during the month; Top number of links active. For each month, we have considered 2 networks: all network including occasional links appearing on that month; backbone, excluding occasional link. The different colors correspond to the quantity evaluated for the specific network. https://doi.org/10.1371/journal.pone.0199547.g007 As the network changes along the year so its proneness to diffuse diseases. Fig 7, present the variation of the epidemic threshold (denoted q), estimated every month, along the year in comparison with the number of links active and the volume of livestock traded. Since the net- work can change along the year, particularly new active links around Tabaski, we considered the backbone network from Apolloni et al. [17] (containing links present more than 2 months) and the total network containing all the link active that month to elicit the role that Tabaski plays on the risk of transmission. The epidemic threshold is at the lowest values between the March and June, when the volume and the number of exchanges (links) is at maximum, and around Tabaski, when a second peak of movements (mainly small ruminants) whose volume is almost equal to the first peak, is observed. Occasional links, appearing only for the Tabaski reason, decreases the invasion threshold and consequently the risk of disease spreading is higher in this period. Table 7 presents the different sets of gravity models that were applied to the full network (binary outcome: 1 if two nodes were connected and 0 otherwise), and to the networks broken down by species, period and transport modality. The best results were obtained with models including human and animal populations at the origin and destination, with a great-circle distance (model E3, Tables 1 and 7). Replacing this distance with cost-distance functions of accessibility (E4) or elevation (E5) did not improve the results. Fig 3 should be considered to see the redundancy between transportation mode (blue edge). Fig 6. Movement networks according to the species and seasons. Note that for each season the links present in both defined transportation mode are not plotted and are the same for both defined season (red: by truck, dark grey: by foot, nf: number of movement links by foot, Jan-July: Non-Tabaski period; Aug-Dec: Tabaski period). Fig 3 should be considered to see the redundancy between transportation mode (blue edge). https://doi.org/10.1371/journal.pone.0199547.g006 12 / 21 PLOS ONE \| https://doi.org/10.1371/journal.pone.0199547 July 18, 2018 Livestock mobility in Mauritania Fig 7. Variations of network quantities along the year. Bottom: monthly epidemic threshold variations along the year; Centre Volume of animals traded during the month; Top number of links active. For each month, we have considered 2 networks: all network including occasional links appearing on that month; backbone, excluding occasional link. The different colors correspond to the quantity evaluated for the specific network. https://doi.org/10.1371/journal.pone.0199547.g007 PLOS ONE \| https://doi.org/10.1371/journal.pone.0199547 July 18, 2018 AIC model n link n unique link E1 E2 E3 E4 E5 E6 E7 ALL SPECIES All 1178 140 1501.8 1604.1 1464.1 1488.9 1485.1 1461.9 1342.7 NTab 518 53 1339.8 1418.6 1303.3 1323.7 1323.6 1303.3 Tab 660 87 995.6 1133.7 979.8 983.1 982.9 976.1 Truck 735 121 427.2 601.1 405.0 408.9 407.3 416.9 Truck—NTab 443 91 315.7 470.7 302.4 309.1 305.4 307.8 Truck—Tab 257 40 175.2 280.5 172.0 174.9 172.2 175.2 Foot 129 21 899.3 885.9 877.5 951.1 934.7 885.6 Foot—NTab 478 81 843.8 832.2 821.8 886.8 874.0 832.1 Foot—Tab 182 41 460.4 454.7 455.2 498.3 484.9 455.7 SR All 620 58 589.5 639.7 572.0 578.9 576.7 570.3 NTab 374 46 482.7 530.3 470.1 474.1 476.5 470.7 Tab 246 40 412.9 473.1 404.0 404.2 403.7 403.7 Truck 324 31 210.3 290.5 204.5 205.7 205.7 205.9 Foot 296 27 315.3 308.2 304.7 331.3 327.0 307.9 CT All 341 47 494.1 526.9 480.6 474.9 487.2 489.4 NTab 214 43 465.0 492.4 451.3 444.5 455.8 459.3 Tab 127 31 342.3 384.2 329.3 322.5 328.0 340.2 Truck 109 13 107.2 146.7 105.6 105.9 106.0 106.0 Foot 232 34 305.7 295.8 296.0 304.1 307.9 307.7 CM All 217 35 411.5 425.1 399.2 397.4 399.5 395.8 NTab 147 32 390 393.9 377.0 375.9 377.6 373.4 Tab 70 20 241.7 265.8 236.1 235.7 236.1 233.4 Truck 113 13 116.6 156.3 109.9 110.8 108.9 112.3 Foot 104 22 266.6 258.7 261.3 264.9 265.9 258.6 https://doi.org/10.1371/journal.pone.0199547.t007 Table 7. AIC values of the logistic regression models. The models are broken down by species (SR: small ruminants, CT: cattle, CM: camels), period (Tab: Tabaski, NTab: Non-Tabaski) and transport modality. E1-E6 correspond to different models described in Table 1. Table 7. AIC values of the logistic regression models. The models are broken down by species (SR: small ruminants, CT: cattle, CM: camels), period (Tab: Tabaski, NTab: Non-Tabaski) and transport modality. E1-E6 correspond to different models described in Table 1. Table 7. AIC values of the logistic regression models. The models are broken down by species (SR: small ruminan NTab: Non-Tabaski) and transport modality. E1-E6 correspond to different models described in Table 1. Table 7. AIC values of the logistic regression models. The models are broken down by species (SR: small ruminants, CT: cattle, CM: camels), period (Tab: Tabaski, NTab: Non-Tabaski) and transport modality. E1-E6 correspond to different models described in Table 1. AIC model n link n unique link E1 E2 E3 E4 E5 E6 E7 ALL SPECIES All 1178 140 1501.8 1604.1 1464.1 1488.9 1485.1 1461.9 1342.7 NTab 518 53 1339.8 1418.6 1303.3 1323.7 1323.6 1303.3 Tab 660 87 995.6 1133.7 979.8 983.1 982.9 976.1 Truck 735 121 427.2 601.1 405.0 408.9 407.3 416.9 Truck—NTab 443 91 315.7 470.7 302.4 309.1 305.4 307.8 Truck—Tab 257 40 175.2 280.5 172.0 174.9 172.2 175.2 Foot 129 21 899.3 885.9 877.5 951.1 934.7 885.6 Foot—NTab 478 81 843.8 832.2 821.8 886.8 874.0 832.1 Foot—Tab 182 41 460.4 454.7 455.2 498.3 484.9 455.7 SR All 620 58 589.5 639.7 572.0 578.9 576.7 570.3 NTab 374 46 482.7 530.3 470.1 474.1 476.5 470.7 Tab 246 40 412.9 473.1 404.0 404.2 403.7 403.7 Truck 324 31 210.3 290.5 204.5 205.7 205.7 205.9 Foot 296 27 315.3 308.2 304.7 331.3 327.0 307.9 CT All 341 47 494.1 526.9 480.6 474.9 487.2 489.4 NTab 214 43 465.0 492.4 451.3 444.5 455.8 459.3 Tab 127 31 342.3 384.2 329.3 322.5 328.0 340.2 Truck 109 13 107.2 146.7 105.6 105.9 106.0 106.0 Foot 232 34 305.7 295.8 296.0 304.1 307.9 307.7 CM All 217 35 411.5 425.1 399.2 397.4 399.5 395.8 NTab 147 32 390 393.9 377.0 375.9 377.6 373.4 Tab 70 20 241.7 265.8 236.1 235.7 236.1 233.4 Truck 113 13 116.6 156.3 109.9 110.8 108.9 112.3 Foot 104 22 266.6 258.7 261.3 264.9 265.9 258.6 https://doi.org/10.1371/journal.pone.0199547.t007 rather than their absolute values (E6) did not lead to improved models for the different break- downs. In all models, and for all species networks, when the animal population was kept as a predictor, the number of sheep was the most important predictor. rather than their absolute values (E6) did not lead to improved models for the different break- downs. In all models, and for all species networks, when the animal population was kept as a predictor, the number of sheep was the most important predictor. Table 8 presents the details of the final models broken down by species, season and trans- port mode with the human and sheep population at the origin and destination and great-circle distance as predictors. The predictive power of the models were moderate to very good, according to the species, season and mode of transport, with AUC values ranging from 0.76 to 0.97. Similarly, although we noted some improvements for some combinations of species and transport modalities, the use of the population difference between origin and destination PLOS ONE \| https://doi.org/10.1371/journal.pone.0199547 July 18, 2018 13 / 21 PLOS ONE \| https://doi.org/10.1371/journal.pone.0199547 July 18, 2018 Livestock mobility in Mauritania rather than their absolute values (E6) did not lead to improved models for the different break- downs. In all models, and for all species networks, when the animal population was kept as a predictor, the number of sheep was the most important predictor. Table 8 presents the details of the final models broken down by species, season and trans- port mode with the human and sheep population at the origin and destination and great-circle distance as predictors. The predictive power of the models were moderate to very good, according to the species, season and mode of transport, with AUC values ranging from 0.76 to 0.97. Considering all species, seasons and transport modes, positive associations were found between the probability of a trade connection and: i) low human population at the origin, ii) high human population at the destination, iii) high sheep population at the origin, iv) low sheep population at the destination, and v) a low great-circle distance (Table 8). The results were similar for the small ruminants’ network model, whereas for the cattle model, the human population at the origin and the sheep population at the destination were not significant. For the camel network, the human population at the origin and the great-circle distance were not significant. The sign of the significant effects was coherent across the sub-models, i.e. a higher probability of trade event was always associated with low human population at the origin, or high human population at the destination, a high population of sheep at the origin, a low pop- ulation of sheep at the destination and a low great-circle distance, or a combination of these effects. In addition, meaningful differences were noticed in some sub-models. For example, in small ruminant models split by transport mode, human population at the destination, and sheep population at the origin were not significant in the foot-movement sub-model. In Table 7. AIC values of the logistic regression models. The models are broken down by species (SR: small ruminants, CT: cattle, CM: camels), period (Tab: Tabaski, NTab: Non-Tabaski) and transport modality. E1-E6 correspond to different models described in Table 1. Multivariate linear model and significance of the parameters given for the selected model (E3). g Hpop: Human population at the origin (i) and at destination (j); Spop: Sheep population at the origin (i) and at destination (j). https://doi org/10 1371/journal pone 0199547 t008 g Hpop: Human population at the origin (i) and at destination (j); Spop: Sheep population at the origin (i) and at destination (j). Hpop: Human population at the origin (i) and at destination (j); Spop: Sheep population at the origin (i) and at destination (j). https://doi.org/10.1371/journal.pone.0199547.t008 contrast, both predictors at the origin were not significant in the truck-movement sub-model. Great-circle distance and both human and sheep population were not significant for the latter, though. The observed and estimated trade links are illustrated in Fig 8 for these small rumi- nant models for the Tabaski and non-Tabaski periods, and movements by trucks or on foot. During the non-Tabaski period, the fitted values correctly captured the co-existence of short- and long-distance movements, whereas long-distance movements were prominent during the Tabaski period. Similarly, the prediction of the truck or foot movements corresponded to their respective long or short distances. None of the models correctly fitted the south/north movements. None of the gravity models with the flow of animals as the response, and the same set of predictors as in the presence/absence model detailed above, showed any significant association. Considering all species, seasons and transport modes, positive associations were found between the probability of a trade connection and: i) low human population at the origin, ii) high human population at the destination, iii) high sheep population at the origin, iv) low sheep population at the destination, and v) a low great-circle distance (Table 8). The results were similar for the small ruminants’ network model, whereas for the cattle model, the human population at the origin and the sheep population at the destination were not significant. For the camel network, the human population at the origin and the great-circle distance were not significant. The sign of the significant effects was coherent across the sub-models, i.e. a higher probability of trade event was always associated with low human population at the origin, or high human population at the destination, a high population of sheep at the origin, a low pop- ulation of sheep at the destination and a low great-circle distance, or a combination of these effects. In addition, meaningful differences were noticed in some sub-models. For example, in small ruminant models split by transport mode, human population at the destination, and sheep population at the origin were not significant in the foot-movement sub-model. In PLOS ONE \| https://doi.org/10.1371/journal.pone.0199547 July 18, 2018 14 / 21 PLOS ONE \| https://doi.org/10.1371/journal.pone.0199547 July 18, 2018 Livestock mobility in Mauritania Table 8. Multivariate linear model and significance of the parameters given for the selected model (E3). Intercept Hpopi Hpopj Spopi Spopj Great-circle distance AUC All mode All species -4.239 -7.622 10−6 4.667 10−6 3.905 10−6 -5.104 10−6 -2.278 10−3 0.836 Small ruminants -3.895 -1.0 10−5 4.92 10−6 2.45 10−6 -4.00 10−6 -2.129 10−3 0.840 Cattle -4.619 -8.928 10−6 5.596 10−6 2.118 10−5 7.090 10−6 -4.416 10−3 0.884 Camel -5.402 -1.966 10−6 4.189 10−6 4.699 10−6 -6.374 10−6 -5.632 10−4 0.790 All non-Tabaski All species -4.210 -1.069 10−5 4.555 10−6 4.196 10−6 -5.356 10−3 -2.415 10−3 0.840 Small ruminants -3.809 -1.751 10−5 5.08 10−6 2.634 10−6 -3.575 10−6 -2.628 10−3 0.863 Cattle -4.599 -1.239 10−5 5.417 10−6 2.244 10−5 4.982 10−6 -4.095 10−3 0.885 Camel -5.563 -1.559 10−6 3.852 10−6 5.167 10−6 -6.139 10−6 -5.802 10−4 0.781 All Tabaski All species -5.355 -3.679 10−6 5.424 10−6 3.563 10−6 -4.107 10−6 -1.595 10−3 0.869 Small ruminants -4.680 -5.409 10−6 5.408 10−6 1.985 10−6 -4.637 10−6 -1.378 10−3 0.863 Cattle -6.179 -3.851 10−6 6.665 10−6 2.269 10−5 1.598 10−5 -3.287 10−3 0.901 Camel -6.391 -4.458 10−7 5.102 10−6 4.862 10−6 -7.167 10−6 -2.454 10−4 0.837 Truck All period All species -5.841 -7.730 10−8 6.876 10−6 3.242 10−6 -2.072 10−5 1.126 10−3 0.934 Small ruminants -5.432 -9.748 10−7 6.600 10−6 1.634 10−6 -8.942 10−6 8.968 10−4 0.91 Cattle -2.574 -1.350 10−7 3.652 10−5 1.903 10−5 9.883 10−6 -1.049 10−2 0.945 Camel -6.015 8.635 10−7 7.230 10−6 4.261 10−6 -3.349 10−5 1.590 10−3 0.966 Foot All period All species -2.665 -1.396 10−5 -1.435 10−6 4.702 10−6 -3.418 10−6 -5.783 10−3 0.830 Small ruminants -2.225 -1.963 10−5 3.289 10−6 3.289 10−6 -2.996 10−6 -6.409 10−3 0.853 Cattle -6.470 9.398 10−7 6.589 10−6 1.636 10−5 -1.105 10−4 8.720 10−4 0.908 Camel -4.200 -3.373 10−6 -4.043 10−6 -4.940 10−6 -4.366 10−6 -1.982 10−3 0.759 Table 8. Discussion Though international movements were not addressed in this study, it is noteworthy that in 2014, the largest sheep exportation peak was observed during the “soudure”, i.e. the period sepa- rating the end of familial cereal reserve saved after the previous harvest (millet, sorghum. . .) from the next harvest. This was also the hot, dry season, when forage and surface water resources were finished, and ruminant livestock starved. Therefore, the most obvious option for the Mau- ritanian livestock farmers was to sell most of the offspring, only keeping the core of reproductive PLOS ONE \| https://doi.org/10.1371/journal.pone.0199547 July 18, 2018 15 / 21 Livestock mobility in Mauritania Fig 8. Observed and predicted movements links from the gravity model applied to the small ruminant networks. Models were applied to movements occurring within the non-Tabaski period (A), the Tabaski period (B), without distinction of period (C-D): movements by truck (C) and by foot (D). https://doi.org/10.1371/journal.pone.0199547.g008 Fig 8. Observed and predicted movements links from the gravity model applied to the small ruminant networks. Models were applied to movements occurring within the non-Tabaski period (A), the Tabaski period (B), without distinction of period (C-D): movements by truck (C) and by foot (D). https://doi.org/10.1371/journal.pone.0199547.g008 https://doi.org/10.1371/journal.pone.0199547.g008 https://doi.org/10.1371/journal.pone.0199547.g008 ewes and she-goats. In addition, many of the latter spent this season in the closest areas with more abundant pastoral resources, i.e. in Senegal and Mali. Short- and mid-distance move- ments–most of them by foot, thus allowed pastoralists to exploit more suitable environmental conditions and reduce the economic cost of feeding the animals. In future years, when the Tabaski feast occurs during the soudure period, the relative importance of truck vs. foot ewes and she-goats. In addition, many of the latter spent this season in the closest areas with more abundant pastoral resources, i.e. in Senegal and Mali. Short- and mid-distance move- ments–most of them by foot, thus allowed pastoralists to exploit more suitable environmental conditions and reduce the economic cost of feeding the animals. In future years, when the Tabaski feast occurs during the soudure period, the relative importance of truck vs. foot PLOS ONE \| https://doi.org/10.1371/journal.pone.0199547 July 18, 2018 16 / 21 Livestock mobility in Mauritania movements may change and add to each other. Moreover, as the Tabaski is a mobile festivity which is celebrated annually among Muslims worldwide 70 days after Ramadan, a time slip of 10 days upstream occurs each year. Discussion An overlap with specific diseases vectors may thus occurs depending on the occurrence of the feast (year-dependence of the epidemic risk). Around Sene- gal river, Aedes mosquitoes show a peak in the periods of July-August and September during the rainy season. Thanks to the happening of the Tabaski, held in the last years between September and October, the risk of infection was elevated [36]. The risk remains high every year due to the large volume of moved animals and the traders’ preference for truck transportation and its fast- ness which allow viremic animal introduction on remote locations. The low values for the national network diameter, the presence of hubs and the low values of the epidemic threshold, indicate that the network could be prone to transmission of dis- eases. This means that also a lowly transmitted disease, once introduced in the national net- work, could reach all nodes (the network’s single component) in a short amount of time (small diameter). On the other hand, independently of the species considered, the mobility network is prone to fragmentation due to targeted intervention based on nodes activity (in-degree and in-weight). After the intervention, the network is decomposed in a set of smaller subnetworks, and virus can circulate only among nodes of the same sub-network. Vaccinating animals in largest markets (nodes with largest number of incoming animals) or closing these markets, could result in a very effective way of controlling the epidemic spread. Regarding internal movements, the gravity models correctly predicted the probability of a trade connection and their interpretation was straightforward. Locations with few sheep and high human populations, i.e. urban consumption centers, acted as movement sinks. Con- versely, areas with high sheep and low human populations, i.e. rural livestock farming areas, acted as movement sources. For a similar level of production and demand, short distance movements were more likely that long distance ones. In addition, Nouakchott, the capital city located on the coast, strongly influenced the network structures. Almost a quarter of the total population of Mauritania lives in Nouakchott (according to the National Statistical Office http://www.ons.mr/) and because of this, it is the largest terminal market with an incoming volume of almost 30% of the national one. In preparation for Tabaski, almost 50% of the total traded small ruminants are sold in Nouakchott. The largest majority of livestock provisioning Nouakchott markets comes from the South-Western area of Mauritania. PLOS ONE \| https://doi.org/10.1371/journal.pone.0199547 July 18, 2018 Discussion The cost-distance was estimated using the accessibility friction map which is based on the road net- work and land-use data, two expected drivers of the livestock trade networks. However, the friction layer might not be adapted to the specific constrains of animal movements in arid and hyper-arid environments. For example, the presence of water points or stopover feeding sites along the roads might be more relevant factors than those included in the global friction sur- face. Therefore, further work is needed to build friction surfaces better suited to the specific constrains of animal movement in this environment. The movement survey database also included volumes of the livestock trade flows between locations. However, these quantitative data were not correctly predicted by any of the models. This failure might be related to two non-exclusive factors. First, when a link was established, the volume of traded animals did not vary much, and not proportionally to the deficit in demand or to the distance. Second, there might be noise in the data related to inaccurate replies during the interviews with field veterinary officers. However, considering the outer trade data (international movements), estimates from this survey closely matched importation estimates from the Senegalese Veterinary Services for the Tabaski period. Because these data were of crucial importance to stabilize sheep price (and thus prevent social troubles), all efforts were made for an accurate monitoring of sheep importation. Therefore, this good match pro- vided a partial validation for the quality of data produced by this survey, as well as the good predictive power obtained with the presence/absence models. An important question underlying these analyses is the role that animal mobility might play in the spread of animal and zoonotic diseases in the region. For example, animal move- ments may contribute to the spread of Rift Valley fever (RVF) in the Sahelian region of Mauri- tania, and from Mauritania to Senegal [37,38]. The network structure predicted by this model may provide input for an epidemiological model of RVF or other important diseases affecting the region, such as Peste des petits ruminants (PPR) [6]. Gravity models are an important method in economic analyses, used mainly to predict bilateral flow of population and goods between two distant locations [39]. They were recently adapted to describe the spread of biological agents [33,40–42]. To our knowledge, this study is a first attempt to predict livestock mobility patterns. Discussion A continuous flux of animals is ensured by stockists who collect animals at collection market and transfer them by truck to the capital city. Here, stockists buy imported goods, arrived at the international port, to sell at collection markets. It might look odd that the sheep population at the origin and destination of movements was the best predictor for all species-level networks. A possible explanation might be that the spatial distributions of all ruminant species were positively correlated, thus making the sheep population a confounding factor for the other ruminant populations. This assumption was corroborated by the dominance of small ruminant movements: 1.3 million vs. 0.39 million for cattle and 0.37 million for camels. Therefore, the small ruminant network probably influenced all other ruminant trade activities. The seasonal models had good predictive power for the models of both truck or foot move- ments. These models captured the dominance of long-distance movements (truck movement network), of short-distance movements (foot movement network) and different combinations of short and long-distance movements (Tabaski vs non-Tabaski periods). However, the models did not capture some of the trade links. For example, they failed to predict the link to the northern city of Zouerate, which was involved in the small ruminant and camel trading net- works. Located in the desert and with a low accessibility, this city has a large iron mining industry, and virtually no local production. The working population there may be much higher PLOS ONE \| https://doi.org/10.1371/journal.pone.0199547 July 18, 2018 17 / 21 Livestock mobility in Mauritania than the population estimates in the Worldpop population database, which maps people according to their residence, not their working place. Such high populations of workers may generate high demand for small ruminant and camel meat, and camel milk. Armed conflicts may also influence the pattern of demand with drivers not accounted for in the models. For example, an important flow of human population was reported in south-eastern Mauritania coming from Mali after the terrorist attack of January 2013. These refugees established camps close to the border with Mali: their population may have influenced the pattern of demand and production in a way that was not captured by the model. The use of cost-distance instead of great-circle distance did not improve the models. Discussion Besides obvious applications in the field of pastoral economics, it opens new perspectives for predicting the transmission of pathogens such as PPR or RVF viruses in animal meta-populations, or extending existing models of post- vaccination immunity persistence at the population level [43]. Also, similar models might be used at the regional level–e.g, between Maghreb or Sahel countries, or between Sahel and Maghreb regions, etc.–to validate and compliment (un)available information on transbound- ary animal movements. Acknowledgments This study was funded by EU grant FP7-613996 VMERGE and is catalogued by the VMERGE Steering Committee as VMERGE000 (http://www.vmerge.eu). The contents of this publication are the sole responsibility of the authors and don’t necessarily reflect the views of the European Commission. Supporting information S1 File. Small ruminant, cattle and camel trading networks and related centrality measures for Mauritania in 2014. Node size show the importance of the measured centrality values. in 18 / 21 PLOS ONE \| https://doi.org/10.1371/journal.pone.0199547 July 18, 2018 Livestock mobility in Mauritania and out-weight measures were scaled on the total volume of traded livestock; eigenvector cen- trality (centrality measure) were scored from 0 to 1, betweenness was considered for the frac- tion of paths passing through the node. (DOCX) and out-weight measures were scaled on the total volume of traded livestock; eigenvector cen- trality (centrality measure) were scored from 0 to 1, betweenness was considered for the frac- tion of paths passing through the node. (DOCX) Author Contributions Conceptualization: Gae¨lle Nicolas, Marius Gilbert. Data curation: Gae¨lle Nicolas. Formal analysis: Gae¨lle Nicolas. Formal analysis: Gae¨lle Nicolas. Methodology: Gae¨lle Nicolas, Marius Gilbert. Methodology: Gae¨lle Nicolas, Marius Gilbert. Supervision: Marius Gilbert. Supervision: Marius Gilbert. Validation: Gae¨lle Nicolas. Validation: Gae¨lle Nicolas. Visualization: Gae¨lle Nicolas. Visualization: Gae¨lle Nicolas. Writing – original draft: Gae¨lle Nicolas. Writing – original draft: Gae¨lle Nicolas. Writing – review & editing: Gae¨lle Nicolas, Andrea Apolloni, Caroline Coste, G. R. William Wint, Renaud Lancelot, Marius Gilbert. PLOS ONE \| https://doi.org/10.1371/journal.pone.0199547 July 18, 2018 References 1. Jones KE, Patel NG, Levy MA, Storeygard A, Balk D, Gittleman JL, et al. Global trends in emerging infectious diseases. Nature. 2008; 451: 990–993. https://doi.org/10.1038/nature06536 PMID: 18288193 2. Black P, Nunn M. Impact of climate change and environmental changes on emerging and reemerging animal disease and animal production. Conf OIE. 2009; 15–25. 3. Nicolas G, Durand B, Duboz R, Rakotondravao R, Chevalier V. Description and analysis of the cattle trade network in the Madagascar highlands: Potential role in the diffusion of Rift Valley fever virus. Acta Trop. 2013; 126: 19–27. https://doi.org/10.1016/j.actatropica.2012.12.013 PMID: 23295911 4. Grace D, Mutua F, Ochungo P, Kruska R, Jones K, Brierley L, et al. Mapping of poverty and likely zoo- noses hotspots. Zoonoses Project 4. Report to the UK Department for International Development. Nai- robi, Kenya: ILRI [Internet]. ILRI; 2012 Jul. Available: https://cgspace.cgiar.org/handle/10568/21161 5. Fèvre EM, Bronsvoort BM de C, Hamilton KA, Cleaveland S. Animal movements and the spread of infectious diseases. Trends Microbiol. 2006; 14: 125–131. https://doi.org/10.1016/j.tim.2006.01.004 PMID: 16460942 6. Jones BA, Rich KM, Mariner JC, Anderson J, Jeggo M, Thevasagayam S, et al. The Economic Impact of Eradicating Peste des Petits Ruminants: A Benefit-Cost Analysis. PLOS ONE. 2016; 11: e0149982. https://doi.org/10.1371/journal.pone.0149982 PMID: 26900944 7. Wang X, Liu C, Mao W, Hu Z, Gu L. Tracing The Largest Seasonal Migration on Earth. ResearchGate. 2014; Available: https://www.researchgate.net/publication/267811689_Tracing_The_Largest_ Seasonal_Migration_on_Earth 8. Deville P, Linard C, Martin S, Gilbert M, Stevens FR, Gaughan AE, et al. Dynamic population mapping using mobile phone data. Proc Natl Acad Sci. 2014; 111: 15888–15893. https://doi.org/10.1073/pnas. 1408439111 PMID: 25349388 19 / 21 PLOS ONE \| https://doi.org/10.1371/journal.pone.0199547 July 18, 2018 Livestock mobility in Mauritania 9. Chapman M, Prothero RM. Circulation in Population Movement (Routledge Revivals): Substance and concepts from the Melanesian case. Reprint edition. London: Routledge; 2013. 10. Clarkson L, Crawford M. Feast and FamineFood and Nutrition in Ireland 1500–1920 [Internet]. 2001. Available: http://www.oxfordscholarship.com/view/10.1093/acprof:oso/9780198227519.001.0001/ acprof-9780198227519 11. Colizza V, Barrat A, Barthe´lemy M, Vespignani A. The role of the airline transportation network in the prediction and predictability of global epidemics. Proc Natl Acad Sci U S A. 2006; 103: 2015–2020. https://doi.org/10.1073/pnas.0510525103 PMID: 16461461 12. Colizza V, Barthe´lemy M, Barrat A, Vespignani A. Epidemic modeling in complex realities. C R Biol. 2007; 330: 364–374. https://doi.org/10.1016/j.crvi.2007.02.014 PMID: 17502293 13. Nicolas G, Chevalier V, Tantely LM, Fontenille D, Durand B. References A Spatially Explicit Metapopulation Model and Cattle Trade Analysis Suggests Key Determinants for the Recurrent Circulation of Rift Valley Fever Virus in a Pilot Area of Madagascar Highlands. PLoS Negl Trop Dis. 2014; 8: e3346. https://doi.org/10. 1371/journal.pntd.0003346 PMID: 25474116 14. Moritz M. Chapter 13—Networks and Hazard Adaptation Among West African Pastoralists. Social Net- work Analysis of Disaster Response, Recovery, and Adaptation. Butterworth-Heinemann; 2017. pp. 193–207. Available: http://www.sciencedirect.com/science/article/pii/B9780128051962000133 15. Ibra T, Ickowicz A, Wane A, Garba I, Gerber P. Eds. Atlas of Trends in pastoral systems in the Sahel 1970–2012. SIPSA. FAO-CIRAD, 32 pages(http://umr-selmet.cirad.fr/publications-et-ressources/ documents-techniques). [Internet]. 2012. Available: https://www.researchgate.net/publication/ 261842575_Atlas_of_Trends_in_pastoral_systems_in_the_Sahel_1970-2012_SIPSA_FAO-CIRAD_ 32_pages_httpumr-selmetciradfrpublications-et-ressourcesdocuments-techniques 16. L’hoˆte Y, Mahe´ G, Some´ B, Triboulet JP. Analysis of a Sahelian annual rainfall index from 1896 to 2000; the drought continues. Hydrol Sci J. 2002; 47: 563–572. https://doi.org/10.1080/02626660209492960 17. Apolloni A, Nicolas G, Coste C, El Mamy ABO, Lancelot R, Gilbert M. Towards the description of live- stock mobility in Sahelian Africa: some results from a survey in Mauritania. Transbound Emerg Dis. 2018; 13: e0191565. https://doi.org/10.1371/journal.pone.0191565 PMID: 29364989 18. Metzger MJ, Bunce RGH, Jongman RHG, Sayre R, Trabucco A, Zomer R. A high-resolution bioclimate map of the world: a unifying framework for global biodiversity research and monitoring. Glob Ecol Bio- geogr. 2013; 22: 630–638. https://doi.org/10.1111/geb.12022 19. FAOSTAT [Internet]. [cited 12 Feb 2018]. Available: http://www.fao.org/faostat/en/#data/QA 20. Nicolas G, Robinson TP, Wint GRW, Conchedda G, Cinardi G, Gilbert M. Using Random Forest to Improve the Downscaling of Global Livestock Census Data. PLOS ONE. 2016; 11: e0150424. https:// doi.org/10.1371/journal.pone.0150424 PMID: 26977807 21. Robinson TP, Wint GRW, Conchedda G, Van Boeckel TP, Ercoli V, Palamara E, et al. Mapping the Global Distribution of Livestock. PLoS ONE. 2014; 9: e96084. https://doi.org/10.1371/journal.pone. 0096084 PMID: 24875496 22. Stevens FR, Gaughan AE, Linard C, Tatem AJ. Disaggregating census data for population mapping using random forests with remotely-sensed and ancillary data. PloS One. 2015; 10: e0107042. https:// doi.org/10.1371/journal.pone.0107042 PMID: 25689585 23. Nelson A. Travel time to major cities: A global map of Accessibility [Internet]. Ispra Italy: Global Envi- ronment Monitoring Unit—Joint Research Centre of the European Commission; 2008. Available: Avail- able at http://gem.jrc.ec.europa.eu/ 24. Wasserman S, Faust K. Social Network Analysis: Methods and Applications. Cambridge University Press; 1994. 25. Beaune´e G, Vergu E, Ezanno P. Modelling of paratuberculosis spread between dairy cattle farms at a regional scale. Vet Res. 2015; 46. https://doi.org/10.1186/s13567-015-0247-3 PMID: 26407894 26. PLOS ONE \| https://doi.org/10.1371/journal.pone.0199547 July 18, 2018 References Moslonka-Lefebvre M, Gilligan CA, Monod H, Belloc C, Ezanno P, Filipe JAN, et al. Market analyses of livestock trade networks to inform the prevention of joint economic and epidemiological risks. J R Soc Interface. 2016; 13. https://doi.org/10.1098/rsif.2015.1099 PMID: 26984191 27. Lichoti JK, Davies J, Kitala PM, Githigia SM, Okoth E, Maru Y, et al. Social network analysis provides insights into African swine fever epidemiology. Prev Vet Med. 2016; 126: 1–10. https://doi.org/10.1016/ j.prevetmed.2016.01.019 PMID: 26848113 28. Relun A, Grosbois V, Alexandrov T, Sa´nchez-Vizcaı´no JM, Waret-Szkuta A, Molia S, et al. Prediction of Pig Trade Movements in Different European Production Systems Using Exponential Random Graph Models. Front Vet Sci. 2017; 4. https://doi.org/10.3389/fvets.2017.00027 PMID: 28316972 29. Kiss IZ, Green DM, Kao RR. The network of sheep movements within Great Britain: network properties and their implications for infectious disease spread. J R Soc Interface. 2006; 3: 669–677. https://doi.org/ 10.1098/rsif.2006.0129 PMID: 16971335 20 / 21 PLOS ONE \| https://doi.org/10.1371/journal.pone.0199547 July 18, 2018 Livestock mobility in Mauritania 30. Volkova VV, Howey R, Savill NJ, Woolhouse MEJ. Sheep Movement Networks and the Transmission of Infectious Diseases. PLOS ONE. 2010; 5: e11185. https://doi.org/10.1371/journal.pone.0011185 PMID: 20567504 31. Lancelot R, Be´ral M, Rakotoharinome VM, Andriamandimby S-F, He´raud J-M, Coste C, et al. Drivers of Rift Valley fever epidemics in Madagascar. Proc Natl Acad Sci U S A. 2017; 114: 938–943. https://doi. org/10.1073/pnas.1607948114 PMID: 28096420 32. Simini F, Gonza´lez MC, Maritan A, Baraba´si A-L. A universal model for mobility and migration patterns. Nature. 2012; 484: 96–100. https://doi.org/10.1038/nature10856 PMID: 22367540 33. Balcan D, Colizza V, Gonc¸alves B, Hu H, Ramasco JJ, Vespignani A. Multiscale mobility networks and the spatial spreading of infectious diseases. Proc Natl Acad Sci. 2009; 106: 21484–21489. https://doi. org/10.1073/pnas.0906910106 PMID: 20018697 34. R Development Core Team. R: A language and Environment for Statistical Computing. Vienna Austria R Found Stat Comput. 2015; 35. Butts CT. sna: Tools for Social Network Analysis [Internet]. 2016. Available: https://cran.r-project.org/ web/packages/sna/index.html 36. Biteye B, Fall AG, Ciss M, Seck MT, Apolloni A, Fall M, et al. Ecological distribution and population dynamics of Rift Valley fever virus mosquito vectors (Diptera, Culicidae) in Senegal. Parasit Vectors. 2018; 11. https://doi.org/10.1186/s13071-017-2591-9 PMID: 29316967 37. El Mamy ABO, Baba MO, Barry Y, Isselmou K, Dia ML, Hampate B, et al. Unexpected Rift Valley Fever Outbreak, Northern Mauritania. Emerg Infect Dis. 2011; 17: 1894–1896. https://doi.org/10.3201/ eid1710.110397 PMID: 22000364 38. PLOS ONE \| https://doi.org/10.1371/journal.pone.0199547 July 18, 2018 References El Mamy AB, Lo MM, Thiongane Y, Diop M, Isselmou K, Doumbia B, et al. Comprehensive Phyloge- netic Reconstructions of Rift Valley Fever Virus: The 2010 Northern Mauritania Outbreak in the Came- lus dromedarius Species. Vector Borne Zoonotic Dis. 2014; 14: 856–861. https://doi.org/10.1089/vbz. 2014.1605 PMID: 25514121 39. Truscott J, Ferguson NM. Evaluating the Adequacy of Gravity Models as a Description of Human Mobil- ity for Epidemic Modelling. PLoS Comput Biol. 2012; 8. https://doi.org/10.1371/journal.pcbi.1002699 PMID: 23093917 40. Xia Y, Bjørnstad ON, Grenfell BT. Measles Metapopulation Dynamics: A Gravity Model for Epidemio- logical Coupling and Dynamics. Am Nat. 2004; 164: 267–281. https://doi.org/10.1086/422341 PMID: 15278849 41. Barrios JM, Verstraeten WW, Maes P, Aerts J-M, Farifteh J, Coppin P. Using the Gravity Model to Esti- mate the Spatial Spread of Vector-Borne Diseases. Int J Environ Res Public Health. 2012; 9: 4346– 4364. https://doi.org/10.3390/ijerph9124346 PMID: 23202882 42. Sarzynska M, Udiani O, Zhang N. A study of gravity-linked metapopulation models for the spatial spread of dengue fever. ArXiv13084589 Q-Bio Stat. 2013; Available: http://arxiv.org/abs/1308.4589 43. Hammami P, Lancelot R, Lesnoff M. Modelling the Dynamics of Post-Vaccination Immunity Rate in a Population of Sahelian Sheep after a Vaccination Campaign against Peste des Petits Ruminants Virus. PloS One. 2016; 11: e0161769. https://doi.org/10.1371/journal.pone.0161769 PMID: 27603710 21 / 21
https://openalex.org/W2095027540	https://europepmc.org/articles/pmc3603679?pdf=render	English	null	Tanshinone IIA and Cryptotanshinone Prevent Mitochondrial Dysfunction in Hypoxia-Induced H9c2 Cells: Association to Mitochondrial ROS, Intracellular Nitric Oxide, and Calcium Levels	Evidence-based complementary and alternative medicine	2,013	cc-by	7,828	Hindawi Publishing Corporation Evidence-Based Complementary and Alternative Medicine Volume 2013, Article ID 610694, 11 pages http://dx.doi.org/10.1155/2013/610694 Hindawi Publishing Corporation Evidence-Based Complementary and Alternative Medicine Volume 2013, Article ID 610694, 11 pages http://dx.doi.org/10.1155/2013/610694 Hindawi Publishing Corporation Evidence-Based Complementary and Alternative Medicine Volume 2013, Article ID 610694, 11 pages http://dx.doi.org/10.1155/2013/610694 Academic Editor: Kashmira Nanji Academic Editor: Kashmira Nanji Copyright © 2013 H.-J. Jin and C.-G. Li. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. The protective actions of tanshinones on hypoxia-induced cell damages have been reported, although the mechanisms have not been fully elucidated. Given the importance of nitric oxide (NO) and reactive oxygen species (ROS) in regulation of cell functions, the present study investigated the effects of two major tanshinones, Tanshinone IIA (TIIA) and cryptotanshinone (CT), on hypoxia- induced myocardial cell injury and its relationships with intracellular NO and ROS, calcium, and ATP levels in H9c2 cells. Chronic hypoxia significantly reduced cell viability which accompanied with LDH release, increase in mitochondrial ROS, intracellular NO and calcium levels, decrease in superoxide dismutase (SOD) activity, and cellular ATP contents. TIIA and CT significantly prevented cell injury by increasing cell viability and decreasing LDH release. The protective effects of tanshinones were associated with reduced mitochondrial superoxide production and enhanced mitochondrial SOD activity. Tanshinones significantly reduced intracellular NO and Ca2+ levels. ATP levels were also restored by TIIA. These findings suggest that the cytoprotective actions of tanshinones may involve regulation of intracellular NO, Ca2+, ATP productions, mitochondrial superoxide production, and SOD activity, which contribute to their actions against hypoxia injuries. Hyou-Ju Jin1 and Chun-Guang Li1,2 Hyou-Ju Jin1 and Chun-Guang Li1,2 1 Traditional & Complementary Medicine Program, RMIT Health Innovations Research Institute, School of Health Sciences, RMIT University, Bundoora, VIC 3083, Australia 2 Center for Complementary Medicine Research, National Institute of Complementary Medicine, University of Western Sydney, Campbelltown Campus, Penrith, NSW 2751, Australia 1 Traditional & Complementary Medicine Program, RMIT Health Innovations Research Institute, School of Health Sciences, RMIT University, Bundoora, VIC 3083, Australia 2 Center for Complementary Medicine Research, National Institute of Complementary Medicine, University of Western Sydney, Campbelltown Campus, Penrith, NSW 2751, Australia 1 Traditional & Complementary Medicine Program, RMIT Health Innovations Research Institute, School of Health Sciences, RMIT University, Bundoora, VIC 3083, Australia 2 Center for Complementary Medicine Research, National Institute of Complementary Medicine, University of Western Sydney, Campbelltown Campus, Penrith, NSW 2751, Australia 1 Traditional & Complementary Medicine Program, RMIT Health Innovations Research Institute, School of Health Sciences, RMIT University, Bundoora, VIC 3083, Australia 2 Center for Complementary Medicine Research, National Institute of Complementary Medicine, University of Western Sydney, Campbelltown Campus, Penrith, NSW 2751, Australia Correspondence should be addressed to Hyou-Ju Jin; hyouju.jin@outlook.com Received 12 August 2012; Accepted 27 January 2013 2. Materials and Methods 2.1. Chemicals. Tansinone IIA (TIIA) and cryptotanshi- none (CT) were purchased from the National Institute for the Control of Pharmaceutical and Biological Products (>99% purity) (Beijing, China). Dulbecco’s Modified Eagle’s Medium (DMEM), fetal bovine serum (FBS), penicillin, and strep-tomycin were purchased from Gibco BRL (Grand Island, NY, USA). GasPak EZ Anaerobe Container System Sachets with Indicator and GasPak EZ Standard Incubation Container were from Becton Dickinson and company (Syd- ney, NSW, Australia). Trypsin-EDTA solution, (3-(4,5-dim- ethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide), 2󸀠,7󸀠- dichlorodihydrofluorescein diacetate, Superoxide dismutase assay kit, dihydroethidium, diphenyleneiodonium chloride, 4-hydroxy-TEMPO (TEMPOL), rotenone, antimycin A and nitro-L-arginine methyl ester (L-NAME) were from Sigma- Aldrich (St. Louis, MO, USA). Fura-2 AM and MitoSOX were from Molecular Probes (S. San Francisco, CA, USA). Lucigenin and MnTBAP were from Santa Cruz Biotech- nology (CA, USA). CytoTox96 NonRadioactive Cytotoxicity assay kit and ENLITEN ATP Assay System Bioluminescence Detection Kit were from Promega (Madison, WI, USA). 4, 5- Diaminofluorescein (DAF-2) was purchased from Sapphire Bioscience Biochemicals (Sydney, NSW, Australia). Mito- chondria Isolation Kit for Cultured Cells was purchased from Thermo Scientific (Rockford, USA). 2.4. LDH Release Measurement. LDH release was deter- mined by CytoTox 96 NonRadioactive Cytotoxicity Assay kit according to the manufacturer’s instructions (Promega). After 8 hr hypoxia, the supernatant was collected and placed in 96 wells and 50 𝜇L of reconstitute substrate mixture was added in each well. After 30 mins incubation, 50 𝜇L of stop solution was added and absorbance was measured at 490 nm using Flexstation multiplate reader (Molecular Devices). 2.5. Cellular ATP Content Measurement. Cellular ATP con- tent was measured by ENLITEN ATP Assay System Biolu- minescence Detection Kit according to the manufacturer’s instructions (Promega). After hypoxia, The cells were washed with PBS and lysated, and supernatants were collected. Pro- teins (10 𝜇g/20 𝜇L) were added in white optiplate and initiated action by adding reconstituted reagent. Then, luminescence was measured in POLARstar OPTIMA microplate reader (BMG LabTech). 2.6. NADPH Oxidase Activity. NADPH oxidase activity was measured by lucigenin chemiluminescence as described pre- viously with minor modification [20]. After 8 hr hypoxia, the cells were collected and centrifuged at 750 g for 10 mins at 4∘C. The supernatant was discarded and the pellet was resus- pended in lysis buffer (50 mM KH2PO4, pH 7.0, 1 mM EGTA, 10 𝜇g/mL aprotinin, 0.5 𝜇g/mL leupeptin, 1 𝜇g/mL pepstatin, and 0.5 mM PMSF). Next, the cells were homogenized by quick freeze and thaw step. 1. Introduction The experimental condition was established from a preliminary study involving different concentrations of tanshinones (0.1–10 𝜇M) at different (2 and 24 hr) pre- and posthypoxia incubation periods. In some experiments, MnT- BAP (1 𝜇M), rotenone (10 𝜇M), antimycin A (AA: 10 𝜇M), TEMPOL (10 mM), and L-NAME (1 mM) were treated 1 hr before inducing hypoxia as positive controls. 2.3. MTT Assay. Cell viability was determined by MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bro- mide) assay as described previously with a modification [19]. The cells (1 × 104 cells/well) were seed in 96 wells. At the end of hypoxia period, MTT solution was added into plates at a final concentration of 0.5 mg/mL and incubated for 2 hr at 37∘C. Then, the culture medium was discarded and 150 𝜇L DMSO was added to each well to dissolve dark blue formazan crystals. The absorbance was read at 570 nm using POLARstar OPTIMA microplate reader (BMG LabTech). 1. Introduction by hypoxia can result in mitochondrial ROS increase by inhibiting mitochondrial electron transport chain function, which in turn promotes peroxynitrite formation and cell apoptosis [6, 7]. On the other hand, hypoxia may modulate NO production by regulating intracellular calcium which is important for Ca2+/calmodulin-dependent eNOS and nNOS activity, and NO increase in turn may inhibit mitochondrial complex IV [8]. This indicates an interaction among NO, ROS, intracellular calcium, and regulation of ATP synthesis in mitochondria. Understanding the relationship of these factors may help to interpret the mechanisms of cellular injury in hypoxia condition [9, 10]. It has been established that chronic hypoxia is associated with cardiac dysfunctions in certain pathological conditions such as ischemia reperfusion, myocardial infarction (MI), and hypertrophy [1]. Hypoxia causes changes of various cel- lular mechanisms related to mitochondrial dysfunction and oxidative stress [2]. Among these, hypoxia-induced changes of ROS and NO productions, intracellular calcium, and ATP levels may have particular importance, given the role of these molecules in regulation of cell functions in general [3]. For example, a recent study shows that hypoxia-increased mitochondrial superoxide anion (O2 ∙−), not cytosolic O2 ∙−, plays an important role in hypoxia-induced cell apoptosis [4, 5]. Studies have also found that excess NO production Tanshinones are a group of bioactive compounds isolated from Salvia miltiorrhiza (Danshen), a traditionally medicinal plant used in management of angina pectoris, atherosclerosis, 2 Evidence-Based Complementary and Alternative Medicine Evidence-Based Complementary and Alternative Medicine 2 Evidence-Based Complementary and Alternative Medicine and MI [11]. Among these, tanshinone IIA (TIIA) and cryp- totanshinone (CT) are two major bioactive tanshinones [12]. They have been reported to have actions against oxidative stress, myocardial infarction, and myocardial ischemia reper- fusion injury [13]. For example, studies in vitro have revealed antioxidant actions of TIIA by attenuating intracellular ROS level and enhancing antioxidant enzymes activity [14, 15]. TIIA and CT have also been shown to influence vasodilation by regulating NO and intracellular Ca2+ levels in endothelial cells [16, 17]. However, the actions of TIIA and CT on ROS and NO pathways under hypoxic conditions are still not clear. Thus, the present study was conducted to investigate the effects of TIIA and CT on hypoxia-induced cardiac injury and their regulations of intracellular NO, ROS, calcium levels, and ATP contents in H9c2 cells. hypoxia period. 2. Materials and Methods 10 𝜇g of proteins were added in optiwhite 96-well plates with the 100 𝜇L assay buffer (50 mM KH2PO4 (pH 7.0), 150 mM Sucrose, 100 𝜇M NADPH, and 1 mM EGTA). Then, the reaction was started by 5 𝜇M luci- genin. 5 𝜇M DPI was added as an inhibitor. Chemilumines- cence was measured with POLARstar OPTIMA microplate reader (BMG LabTech). 2.2. Cells Culture and Hypoxia. The H9c2 embryonic rat heart-derived the cells were obtained from American Type Culture Collection (ATCC; Manassas, VA) and main- tained in Dulbecco’s modified Eagle’s medium supplemented with 10% v/v fetal bovine serum and 100 𝜇g/mL peni- cillin/streptomycin at 37∘C in a humidified atmosphere con- taining 5% CO2 (passage 25–35). To mimic hypoxia condition, cells were placed in a Gas- Pak EZ Gas generating Pouch System (Becton-Dickinson) for 8 hr and incubated with serum-free and glucose-free DMEM as described previously [18]. As normoxia control, serum- free DMEM was added to cells and incubated for 8 hr in normoxia condition (21% O2). For the treatment groups, TIIA or CT (3 𝜇M) was added 2 hr before and during the 2.7. Intracellular Hydrogen Peroxide/Peroxynitrite Produ- ction. Intracellular hydrogen peroxide/peroxynitrite (H2O2/ ONOO−) production was measured by using 2󸀠,7󸀠-di- chlorofluoresceindiacetate (DCFH-DA). The nonfluores- Evidence-Based Complementary and Alternative Medicine 3 3 Cell viability (% of normoxia) 0 20 40 60 80 100 Hypoxia + treatments Normoxia Hypoxia TIIA 3 𝜇M CT 3 𝜇M ∗∗∗ ∗∗∗/## ∗∗∗/#/# (a) LDH release (%) ( by normoxia) 0 50 100 150 200 250 Hypoxia + treatments Normoxia Hypoxia TIIA 3 𝜇M CT 3 𝜇M ∗∗∗ ∗∗/## ###/Δ (b) ATP contents (by normoxia) 0 0.3 0.6 0.9 1.2 Hypoxia + treatments Normoxia Hypoxia TIIA 3 𝜇M CT 3 𝜇M ∗ ∗ #/Δ (c) Figure 1: Effects of TIIA and CT on hypoxia-induced H9c2 cell injury. (a) Cell viability by MTT assay. The cell viability of normoxia was adjusted to 100% (𝑛= 5), ##𝑃< 0.01 versus hypoxia, ∗∗∗𝑃< 0.001 versus normoxia. (b) LDH release. The LDH release of normoxia was adjusted to 100% (𝑛= 3), Δ𝑃< 0.05 versus CT, ##𝑃< 0.01 versus hypoxia, and ###𝑃< 0.001 versus hypoxia, ∗∗∗𝑃< 0.001 versus normoxia. (c) Cellular ATP contents. The cellular ATP contents of normoxia was adjusted to 1 (𝑛= 6), ∗𝑃< 0.05 versus normoxia, ∗𝑃< 0.05 versus normoxia, #𝑃< 0.05 versus hypoxia, and Δ𝑃< 0.05 versus CT. 2. Materials and Methods Cell viability (% of normoxia) 0 20 40 60 80 100 Hypoxia + treatments Normoxia Hypoxia TIIA 3 𝜇M CT 3 𝜇M ∗∗∗ ∗∗∗/## ∗∗∗/#/# LDH release (%) ( by normoxia) 0 50 100 150 200 250 Hypoxia + treatments Normoxia Hypoxia TIIA 3 𝜇M CT 3 𝜇M ∗∗∗ ∗∗/## ###/Δ (a) (b) ATP contents (by normoxia) 0 0.3 0.6 0.9 1.2 Hypoxia + treatments Normoxia Hypoxia TIIA 3 𝜇M CT 3 𝜇M ∗ ∗ #/Δ (c) (c) Figure 1: Effects of TIIA and CT on hypoxia-induced H9c2 cell injury. (a) Cell viability by MTT assay. The cell viability of normoxia was adjusted to 100% (𝑛= 5), ##𝑃< 0.01 versus hypoxia, ∗∗∗𝑃< 0.001 versus normoxia. (b) LDH release. The LDH release of normoxia was adjusted to 100% (𝑛= 3), Δ𝑃< 0.05 versus CT, ##𝑃< 0.01 versus hypoxia, and ###𝑃< 0.001 versus hypoxia, ∗∗∗𝑃< 0.001 versus normoxia. (c) Cellular ATP contents. The cellular ATP contents of normoxia was adjusted to 1 (𝑛= 6), ∗𝑃< 0.05 versus normoxia, ∗𝑃< 0.05 versus normoxia, #𝑃< 0.05 versus hypoxia, and Δ𝑃< 0.05 versus CT. 1 and gain of 2 using laser-based focusing. Images were captured using a DAPI filter (350/70 nm Ex, 470/50 nm Em for DAPI) and Cy3 filter (550/35 nm Ex, 570/30 nm Em for DHE and MitoSOX). The cell images were analysed by Meta express software (Molecular Devices). cent DCFH-DA readily diffuses into the cells, where it is hydrolysed to the polar derivative DCFH, which is oxidized in the presence of H2O2 or ONOO−to the highly fluorescent DCF. At the End of hypoxia, the cells were incubated with 20 𝜇M DCFH-DA for 20 mins at 37∘C. The fluorescence was measured at an excitation wavelength of 488 nm and emis- sion at 530 nm in Flexstation multiplate reader (Molecular Devices). The fluorescence intensity was normalized to total protein contents and expressed as arbitrary units per mg protein. 2.9. Superoxide Dismutase Activity. Superoxide dismutase activity was determined by superoxide dismutase assay kit according to the manufacturer’s instructions (Sigma- Aldrich). Briefly, cytosolic and mitochondria fractions were prepared by Mitochondria Isolation kit (Thermofisher). The protein amount was measured by Bradford assay. 20 𝜇g of protein was added in 96-well plates and then the reaction was initiated by adding enzyme solution. Absorbance was measured at 450 nm with POLARstar OPTIMA microplate reader (BMG LabTech). 2.9. Superoxide Dismutase Activity. 2. Materials and Methods Superoxide dismutase activity was determined by superoxide dismutase assay kit according to the manufacturer’s instructions (Sigma- Aldrich). Briefly, cytosolic and mitochondria fractions were prepared by Mitochondria Isolation kit (Thermofisher). The protein amount was measured by Bradford assay. 20 𝜇g of protein was added in 96-well plates and then the reaction was initiated by adding enzyme solution. Absorbance was measured at 450 nm with POLARstar OPTIMA microplate reader (BMG LabTech). 2.8. Intracellular and Mitochondrial Superoxide Production. Intracellular and mitochondrial superoxide production was measured by loading cells with 20 𝜇M dihydroethidium (DHE) and 2 𝜇M MitoSOX, respectively, by following method described previously with minor modification [21]. End of hypoxia, DHE and MitoSOX were added and incubated for 30 mins at 37∘C including 10 mins DAPI (1 𝜇M) staining. After the incubation, the cells were washed with PBS. The cell images were obtained using Image Xpress MICRO system (Molecular Devices) at 20X magnification with binning of 2.10. Intracellular Nitric Oxide Production. Direct measure- ment of intracellular nitric oxide production was performed by loading 4,5-diaminofluorescein (DAF-2) [22]. The cells 4 Evidence-Based Complementary and Alternative Medicine Hypoxia Normoxia (a) Fluorescence intensity (a.u.) 0 100 200 300 Hypoxia + treatments Normoxia Hypoxia TIIA 3 𝜇M CT 3 𝜇M DPI 5 𝜇M ∗∗∗ ∗∗∗ ∗∗ ∗∗ (b) 0 3000 6000 9000 12000 Relative light units (RLU) Hypoxia + treatments Normoxia Hypoxia TIIA 3 𝜇M CT 3 𝜇M DPI 5 𝜇M ∗∗∗ ∗∗ ∗∗ ∗∗ (c) Figure 2: Effects of TIIA and CT on hypoxia-induced decrease in intracellular superoxide level and NADPH oxidase activity. (a) Images of ells labelled with DHE in normoxia and hypoxia groups. (b) The quantified value of DHE fluorescence intensity. Represented data are mean value of 500 each cells with 4 independent experiments. (c) Quantitative value of NADPH oxidase activity. Data shown are representative of 4 independent experiments. ∗∗𝑃< 0.01 versus normoxia, ∗∗∗𝑃< 0.001 versus normoxia. (a) 0 3000 6000 9000 12000 Relative light units (RLU) Hypoxia + treatments Normoxia Hypoxia TIIA 3 𝜇M CT 3 𝜇M DPI 5 𝜇M ∗∗∗ ∗∗ ∗∗ ∗∗ (c) Fluorescence intensity (a.u.) Relative light units (RLU) (b) (c) Figure 2: Effects of TIIA and CT on hypoxia-induced decrease in intracellular superoxide level and NADPH oxidase activity. (a) Images of cells labelled with DHE in normoxia and hypoxia groups. (b) The quantified value of DHE fluorescence intensity. 2. Materials and Methods Represented data are mean value of 500 each cells with 4 independent experiments. (c) Quantitative value of NADPH oxidase activity. Data shown are representative of 4 independent experiments. ∗∗𝑃< 0.01 versus normoxia, ∗∗∗𝑃< 0.001 versus normoxia. Fluorescence intensity (a.u./mg protein) 0 3 6 9 12 Hypoxia + treatments Normoxia Hypoxia MnTBAP 1 𝜇M Rotenone 10 𝜇M AA 10 𝜇M TIIA 3 𝜇M CT 3 𝜇M ∗∗ ∗∗ # # ## Figure 3: Effects of TIIA and CT on hypoxia-induced increase in H2O2/ONOO−production. Quantitative value of DCFH-DA fluorescence intensity (𝑛= 4). #𝑃< 0.05 versus hypoxia, ##𝑃< 0.01 versus hypoxia, and ∗∗𝑃< 0.01 versus normoxia. Fluorescence intensity (a.u./mg protein) 0 3 6 9 12 Hypoxia + treatments Normoxia Hypoxia MnTBAP 1 𝜇M Rotenone 10 𝜇M AA 10 𝜇M TIIA 3 𝜇M CT 3 𝜇M ∗∗ ∗∗ # # ## Figure 3: Effects of TIIA and CT on hypoxia-induced increase in H2O2/ONOO−production. Quantitative value of DCFH-DA fluorescence intensity (𝑛= 4). #𝑃< 0.05 versus hypoxia, ##𝑃< 0.01 versus hypoxia, and ∗∗𝑃< 0.01 versus normoxia. were incubated with 5 𝜇M DAF-2 for 1 hr after hypoxia. Fol- lowing the incubation, the cells were washed twice with PBS and visualized in Image Xpress MICRO system (Molecular Devices) at 20X magnification with binning of 1 and gain of 2 using laser-based focusing. Images were captured using a DAPI filter (350/70 nm Ex, 470/50 nm Em for DAPI) and GFP filter (490/40 nm Ex, 510/50 nm Em for DAF-2). The cell images were analysed by Meta Express software (Molecular Devices). 2.11. Intracellular Calcium Level Measurement. Intracellular calcium level was determined by Fura-2AM as described pre- viously with minor modification [23]. After hypoxia, the cells were rinsed twice with PBS and detached by trypsinization. The detached cells were centrifuged and washed with PBS once. Then, the cells were incubated with HBSS with Ca2+ buffer (140 mM NaCl, 4.2 mM KCl, 1 mM CaCl2, 0.4 mM MgSO4, 0.4 mM Na2HPO4, 0.5 mM NaH2PO4, 0.3 mM MgCl2, 5 mM glucose, and 0.2% bovine serum albumin, pH 7.4) supplemented with 2 𝜇M Fura-2AM for 30 mins at 37∘C. After the incubation, the cells were resuspended to HBSS buffer only and incubated for 30 mins at room temperature. were incubated with 5 𝜇M DAF-2 for 1 hr after hypoxia. 2. Materials and Methods Fol- lowing the incubation, the cells were washed twice with PBS and visualized in Image Xpress MICRO system (Molecular Devices) at 20X magnification with binning of 1 and gain of 2 using laser-based focusing. Images were captured using a DAPI filter (350/70 nm Ex, 470/50 nm Em for DAPI) and GFP filter (490/40 nm Ex, 510/50 nm Em for DAF-2). The cell images were analysed by Meta Express software (Molecular Devices). 2.11. Intracellular Calcium Level Measurement. Intracellular calcium level was determined by Fura-2AM as described pre- viously with minor modification [23]. After hypoxia, the cells were rinsed twice with PBS and detached by trypsinization. The detached cells were centrifuged and washed with PBS once. Then, the cells were incubated with HBSS with Ca2+ buffer (140 mM NaCl, 4.2 mM KCl, 1 mM CaCl2, 0.4 mM MgSO4, 0.4 mM Na2HPO4, 0.5 mM NaH2PO4, 0.3 mM MgCl2, 5 mM glucose, and 0.2% bovine serum albumin, pH 7.4) supplemented with 2 𝜇M Fura-2AM for 30 mins at 37∘C. After the incubation, the cells were resuspended to HBSS buffer only and incubated for 30 mins at room temperature. Figure 3: Effects of TIIA and CT on hypoxia-induced increase in H2O2/ONOO−production. Quantitative value of DCFH-DA fluorescence intensity (𝑛= 4). #𝑃< 0.05 versus hypoxia, ##𝑃< 0.01 versus hypoxia, and ∗∗𝑃< 0.01 versus normoxia. 5 5 Evidence-Based Complementary and Alternative Medicine Merge MitoSOX DAPI Normoxia Hypoxia TEMPOL Rotenone L-NAME TIIA 3 CT 3 (a) F 4 C ti d DAPI Normoxia Hypoxia TEMPOL Rotenone L-NAME TIIA 3 CT 3 (a) Figure 4: Continued. Normoxia Hypoxia TEMPOL Rotenone L-NAME TIIA 3 CT 3 (a) Figure 4: Continued. L-NAME TIIA 3 (a) Figure 4: Continued. Figure 4: Continued. 6 Evidence-Based Complementary and Alternative Medicine 6 Fluorescence intensity (a.u.) 0 50 100 150 200 Hypoxia + treatments Normoxia Hypoxia TEMPOL 10 mM Rotenone 10 𝜇M L-NAME 1 mM TIIA 3 𝜇M CT 3 𝜇M ∗∗∗ ## ## ### ### # (b) Fluorescence intensity (a.u.) 0 50 100 150 200 Hypoxia + treatments Normoxia Hypoxia TEMPOL 10 mM Rotenone 10 𝜇M L-NAME 1 mM TIIA 3 𝜇M CT 3 𝜇M ∗∗∗ ## ## ### ### # (b) Figure 4: Effects of TIIA and CT on hypoxia-induced increase in mitochondrial superoxide production. (a) Cell images illustrate MitoSOX (red) and DAPI (blue) in each group. (b) The quantified value of MitoSOX fluorescence intensity. 2. Materials and Methods Represented data are mean value of 500 each cells with 5 independent experiments. #𝑃< 0.05 versus hypoxia, ##𝑃< 0.01 versus hypoxia, ###𝑃< 0.001 versus hypoxia, and ∗∗∗𝑃< 0.001 versus normoxia. Fluorescence intensity (a.u.) (b) Figure 4: Effects of TIIA and CT on hypoxia-induced increase in mitochondrial superoxide production. (a) Cell images illustrate MitoSOX (red) and DAPI (blue) in each group. (b) The quantified value of MitoSOX fluorescence intensity. Represented data are mean value of 500 each cells with 5 independent experiments. #𝑃< 0.05 versus hypoxia, ##𝑃< 0.01 versus hypoxia, ###𝑃< 0.001 versus hypoxia, and ∗∗∗𝑃< 0.001 versus normoxia. 0 20 40 60 80 100 SOD activity (%) Hypoxia + treatments Normoxia Hypoxia TIIA 3 𝜇M CT 3 𝜇M (a) SOD activity (%) 0 20 40 60 80 100 Hypoxia + treatments Normoxia Hypoxia TIIA 3 𝜇M CT 3 𝜇M ∗∗ ## ### (b) Figure 5: Effects of TIIA and CT on SOD enzyme activity. (a) Cytosolic SOD activity. (b) Mitochondrial SOD activity. Data shown are representative of four independent experiments. ∗∗𝑃< 0.01 versus normoxia, ##𝑃< 0.01 versus hypoxia, and ###𝑃< 0.001 versus hypoxia. SOD activity (%) 0 20 40 60 80 100 Hypoxia + treatments Normoxia Hypoxia TIIA 3 𝜇M CT 3 𝜇M ∗∗ ## ### (b) (b) (a) Figure 5: Effects of TIIA and CT on SOD enzyme activity. (a) Cytosolic SOD activity. (b) Mitochondrial SOD activity. Data shown are representative of four independent experiments. ∗∗𝑃< 0.01 versus normoxia, ##𝑃< 0.01 versus hypoxia, and ###𝑃< 0.001 versus hypoxia. Prism Software version 5.0. P ≤0.05 was considered signifi- cant. 1 × 105 cells were transferred to 96-well plates, and then Fura-2AM fluorescence was obtained by alternate excitation at 340 and 380 nm and the emission was detected at 510 nm. The fluorescence maximum was determined by lysing cells with 0.2% Triton X-100 and fluorescence minimum was obtained by recording fluorescence following addition of 40 mM EDTA. The calcium concentration was calculated by equation according to what previously described [24]. 3. Results 3.1. Effects of TIIA and CT on Hypoxia-Induced Cell Injury. Cells exposed to a 8 hr hypoxia exhibited a significant decrease in cell viability (𝑃< 0.001), measured by MTT assay, which was significantly inhibited by pretreatment of TIIA and CT (3 𝜇M) (P < 0.01) (Figure 1(a)). 8 hr hypoxia also significantly increased LDH release (to 220.0%, P < 0.001), which was significantly inhibited by 3 𝜇M TIIA and CT (P < 0.001). The value of LDH release in TIIA-treated group was 2.12. Statistical Analysis. Results were expressed as means ± SEM. Statistical differences among groups were analysed by one-way analysis of variance (ANOVA) using GraphPad 7 Evidence-Based Complementary and Alternative Medicine 7 Merge DAF-2 DAPI Normoxia Hypoxia L-NAME TIIA 3 CT 3 (a) Merge DAF-2 DAPI Normoxia Hypoxia L-NAME TIIA 3 CT 3 (a) Fluorescence intensity (a.u.) 0 30 60 90 120 150 180 Hypoxia + treatments Normoxia Hypoxia L-NAME TIIA 3 𝜇M CT 3 𝜇M ## ## ## ∗∗∗ (b) ffects of TIIA and CT on hypoxia-induced increase in intracellular nitric oxide production (a) Cell images Normoxia Hypoxia L-NAME TIIA 3 CT 3 (a) Fluorescence intensity (a.u.) 0 30 60 90 120 150 180 Hypoxia + treatments Normoxia Hypoxia L-NAME TIIA 3 𝜇M CT 3 𝜇M ## ## ## ∗∗∗ (b) : Effects of TIIA and CT on hypoxia-induced increase in intracellular nitric oxide production. (a) Cell images illustra nd DAPI (blue) in each group. (b) Quantitative DAF-2 fluorescence intensity. Represented data are mean value of 500 each ndent experiments (𝑛= 3). ##𝑃< 0.01 versus hypoxia, ∗∗∗𝑃< 0.001 versus normoxia. Normoxia Hypoxia L-NAME TIIA 3 (a) Fluorescence intensity (a.u.) 0 30 60 90 120 150 180 Hypoxia + treatments Normoxia Hypoxia L-NAME TIIA 3 𝜇M CT 3 𝜇M ## ## ## ∗∗∗ (b) Fluorescence intensity (a.u.) 0 30 60 90 120 150 180 Hypoxia + treatments Normoxia Hypoxia L-NAME TIIA 3 𝜇M CT 3 𝜇M ## ## ## ∗∗∗ (b) Figure 6: Effects of TIIA and CT on hypoxia-induced increase in intracellular nitric oxide production. (a) Cell images illustrate DAF-2 (green) and DAPI (blue) in each group. (b) Quantitative DAF-2 fluorescence intensity. Represented data are mean value of 500 each cells with 3 independent experiments (𝑛= 3). ##𝑃< 0.01 versus hypoxia, ∗∗∗𝑃< 0.001 versus normoxia. 3. Results 8 Evidence-Based Complementary and Alternative Medicine 8 0 50 100 150 200 Hypoxia + treatments Normoxia Hypoxia TIIA 3 𝜇M CT 3 𝜇M ∗∗∗ ∗/#/# ∗/#/# [Ca2+]𝑖(nM) Figure 7: Effects of TIIA and CT on hypoxia-induced increase in intracellular calcium level. Data shown are representative of four independent experiments. ∗𝑃< 0.05 versus normoxia, ##𝑃< 0.01 versus hypoxia, and ∗∗∗𝑃< 0.001 versus normoxia. 0 50 100 150 200 Hypoxia + treatments Normoxia Hypoxia TIIA 3 𝜇M CT 3 𝜇M ∗∗∗ ∗/#/# ∗/#/# [Ca2+]𝑖(nM) fluorescence intensity to 154.3 ± 19.9 arbitrary units (a.u) while normoxia showed 88.2 ± 5.3 a.u. 1 hr pretreatment with 10 mM TEMPOL (SOD mimic), 10 𝜇M rotenone (complex I inhibitor), and 1 mM L-NAME (NO synthase inhibitor) significantly decreased the mitochondrial superoxide pro- duction. In the presence of TIIA and CT (3 𝜇M), the mito- chondrial superoxide production was significantly reduced to 111.4 ± 30.2 a.u (𝑃< 0.01) and 122.5 ± 29.7 a.u (𝑃< 0.05), respectively (Figure 4(b)). 3.5. Effects of TIIA and CT on SOD Activity. Cytosolic SOD activity in hypoxia group (78.8 ± 4.1%) did not significantly (𝑃> 0.05) change compared to normoxia control (81.3 ± 2.1%). In contrast, the cells in hypoxia showed a significant decrease in mitochondrial SOD activity by 12.6% compared to normoxia control. In the presence of TIIA and CT (3 𝜇M), the cytosolic SOD activity did not significantly change com- pared to hypoxia control. The decrease in mitochondrial SOD activity by hypoxia was restored when the cells were pre- treated with 3 𝜇M TIIA or 3 𝜇M CT. There was no statistically significant difference in mitochondrial SOD activity between tanshinones treated groups and normoxia control (Figures 5(a) and 5(b)). 3.5. Effects of TIIA and CT on SOD Activity. Cytosolic SOD activity in hypoxia group (78.8 ± 4.1%) did not significantly (𝑃> 0.05) change compared to normoxia control (81.3 ± 2.1%). In contrast, the cells in hypoxia showed a significant decrease in mitochondrial SOD activity by 12.6% compared to normoxia control. In the presence of TIIA and CT (3 𝜇M), the cytosolic SOD activity did not significantly change com- pared to hypoxia control. The decrease in mitochondrial SOD activity by hypoxia was restored when the cells were pre- treated with 3 𝜇M TIIA or 3 𝜇M CT. There was no statistically significant difference in mitochondrial SOD activity between tanshinones treated groups and normoxia control (Figures 5(a) and 5(b)). 3. Results Figure 7: Effects of TIIA and CT on hypoxia-induced increase in intracellular calcium level. Data shown are representative of four independent experiments. ∗𝑃< 0.05 versus normoxia, ##𝑃< 0.01 versus hypoxia, and ∗∗∗𝑃< 0.001 versus normoxia. significantly less than that of LDH release in CT group (P < 0.05) (Figure 1(b)). After 8 hr hypoxia, a significant reduction of the cellular ATP contents (by 20.9%, 𝑃 < 0.05) was observed compared to normoxia control, which was restored by pretreatment with TIIA. The cellular ATP contents in TIIA-treated group were significantly higher (𝑃< 0.05) than those in CT group (Figure 1(c)). 3.6. Effects of TIIA and CT on Hypoxia-Induced Increase in Intracellular Nitric Oxide Production. Figure 6(a) illustrates the fluorescence images of cells stained with DAF-2. When compared with normoxia control, distinct intensification in DAF-2 fluorescence was observed in hypoxia group. The quantitative values of the florescence intensity of images were presented in Figure 6(b). Pretreatment with TIIA and CT significantly decreased the intracellular NO production. There was no statistical difference between tanshinones- treated groups and normoxia control. L-NAME (1 mM, NOS inhibitor) significantly (𝑃< 0.01) reduced the intracellular NO production compared to hypoxia control. 3.2. Effects of TIIA and CT on Hypoxia-Induced Decrease in Intracellular Superoxide Generation and NADPH Oxidase Activity. 8 hr hypoxia significantly decreased intracellular superoxide generation (𝑃< 0.001) and NADPH oxidase activity (𝑃< 0.01) compared to normoxia control. TIIA-, CT- and DPI-treated groups did not significantly affect the intracellular superoxide generation and NADPH oxidase activity compared to hypoxia control (Figures 2(b) and 2(c)). 3.7. Effects of TIIA and CT on Hypoxia-Induced Increase in Intracellular Ca2+ Level. Intracellular Ca2+ ([Ca2+]𝑖) level was significantly elevated in hypoxia group (𝑃< 0.001) compared to normoxia control. Pretreatment with 3 𝜇M of TIIA and CT significantly (𝑃< 0.01) prevented the [Ca2+]𝑖elevation compared to hypoxia control (Figure 7). 3.3. Effects of TIIA and CT on Hypoxia-Induced Increase in 𝐻2𝑂2/𝑂𝑁𝑂𝑂−Production. After 8 hr hypoxia, the DCF fluorescence intensity was significantly elevated to 9.57 ± 0.50 a.u/mg protein (𝑃< 0.01), presenting 59% increase compared to normoxia group. The hypoxia-induced increase in DCF fluorescence was abolished by 1 𝜇M MnTBAP (per- oxynitrite inhibitor), 10 𝜇M rotenone (complex I inhibitor), and antimycin A (AA) (complex III inhibitor), presenting 70.1%, 79.2%, and 91.4% inhibition rates, respectively. Also, 3 𝜇M TIIA reduced the increase in DCF fluorescence by 28.9% without statistical significance difference compared to hypoxia control (Figure 3). 4. Discussion The main finding of the present study is that TIIA and CT protect against chronic hypoxia-induced H9c2 cells injury by restoring cellular ATP contents, decreasing mitochondrial superoxide, intracellular NO, and calcium levels in H9c2 cells. This is consistent with previous observations that hypoxia-induced apoptosis was associated with ROS, NO, and calcium in myocardial cells [5, 25]. Additionally, cellular ATP contents, NO, and calcium are closely associated in mitochondrial ROS production and this suggests that chronic hypoxia-induced cell damages are related to mitochondrial dysfunction. 3.4. Effects of TIIA and CT on Hypoxia-Induced Increase in Mitochondrial Superoxide Production. Figure 4(a) illustrates the fluorescence images of cells stained with MitoSOX. When compared with normoxia group, distinct intensifica- tion in fluorescence was observed in hypoxia group. The cells exposed to hypoxia significantly increased MitoSOX 9 Evidence-Based Complementary and Alternative Medicine Evidence-Based Complementary and Alternative Medicine Myocardial hypoxia is a main cause of cardiac dys- function due to its triggering cell injury, apoptosis, and/or necrosis [1, 26]. The present study showed that the main cause of cell injury or death under the chronic hypoxia condition was associated with mitochondrial dysfunction with accompanying LDH release and cellular ATP depletion, which is consistent with previous reports [5, 27]. The protec- tive actions of tanshinones against chronic hypoxia-induced cell injury indicate that these compounds may conserve mitochondrial function. Previous studies have reported car- dioprotective effects of TIIA on H2O2-induced cell injury [28] and doxorubicin-induced cell apoptosis [29] in neonatal car- diomyocytes by protecting DNA integrity mitochondrial pro- teins and reducing intercellular ROS production. Similarly, the antiapoptotic effect of CT has been shown previously with actions of preventing mitochondrial-dependent apoptosis in nitric oxide induced neuroblastoma cells apoptosis [30]. The increase of ATP level by TIIA may be related to its protection of mitochondrial electron transport chain (ETC) function as ATP is mostly generated by oxidative phosphorylation, a process translocating protons by complex I/III/IV and subse- quently uptake of the protons by ATP synthase accompanying the synthesis of ATP, in mitochondria ETC [31]. The finding that CT was less effective than TIIA in restoring cellular ATP contents may be related to a previous observation that CT enhanced AMP-activated protein kinase (AMPK) [32], as it has been known that AMPK is associated with energy homeostasis, mitochondrial function, and cell survival [33]. It will be interesting to investigate further the effects of tanshinones on AMPK activity in chronic hypoxia condition. (nanomolar). 4. Discussion However, at higher concentrations NO can oxi- dize ubiquinol of ubiquinol-cytochrome c-reductase (Com- plex III) to increase unstable ubisemiquinone, which pro- duces superoxide by univalent electron transfer to O2 [36]. Additionally, exposure to higher concentrations of NO can increase peroxynitrite formation which causes an inhibition of mitochondrial respiration at multiple sites (complex I, complex II, cytochrome c oxidase, the ATP synthase, aconi- tase, MnSOD, and creatine kinase) [37]. This implies that restoring electron transport chain function by reducing NO production, in addition to antioxidant enzyme activity, may help to reduce mitochondrial superoxide production during chronic hypoxia condition. Since cytosolic ROS may not play major role in hypoxia- induced cell damages, it is not surprising to observe the lack of effect of TIIA and CT in intracellular ROS and NADPH oxidase activity. The important finding in this study is that the mitochondria superoxide generation was increased by hypoxia. This increase was significantly inhibited by TIIA and CT treatments, indicating that mitochondrial ROS plays a major role in cell damage-induced hypoxia. Interestingly, NO synthase inhibitor L-NAME also significantly inhib- ited mitochondrial superoxide generation, which suggests that endogenous NO may regulate the ROS production in hypoxia condition. It is possible that ROS may be generated from mitochondrial nitric oxide syntheses which may be uncoupled under hypoxic condition [38]. The observation of increase in mitochondrial superoxide dismutase activity and decrease in intracellular NO level by TIIA and CT in the present study is consistent with previous studies showing actions of tanshinones on regulating NO level and SOD activity in H2O2-induced cell injury and inflammation- induced cell death in endothelial cells [16, 39].i y yp Previous studies on hypoxia-induced ROS generation have shown conflicting results. This could be due to a con- fusion of ROS examined (cytosolic and mitochondria). It has been shown that hypoxia decreased cytosolic superoxide gen- eration but increased mitochondrial superoxide generation [34]. Consistent with this, a significant decrease in cytosolic superoxide generation, but increase in mitochondrial super- oxide generation after hypoxia, was observed in the present study. Decreased cytosolic superoxide generation may be associated with lower oxygen level during hypoxia condition and/or decreased NADPH oxidase activity [2, 21]. Interest- ingly, the activity of cytosolic antioxidant enzyme superoxide dismutase was not significantly changed after hypoxia, indi- cating that this cytosolic antioxidant enzyme may not play a major role in cell injury and death pathway during chronic hypoxia. 4. Discussion On the other hand, there was a significant increase of intracellular hydrogen peroxide/peroxynitrite production, as indicated by DCFH-DA fluorescence probe, suggesting that a mitochondrial-derived ROS component may be involved as shown by the effects of complex I and III inhibitors (rotenone and antimycin A).h Interestingly, TIIA and CT significantly decreased intra- cellular NO and mitochondrial superoxide generations, but not peroxynitrite/hydrogen peroxide levels. Previous studies using the same DCF-DH probe found that TIIA significantly inhibited ROS generation induced by doxorubicin [14, 15]. However, it is not clear if the ROS in those studies is perox- ynitrite/hydrogen peroxide specific as no specific inhibitors were used to validate the species of ROS observed. One possible explanation is that ROS labelled with DCF-DH may mainly be peroxynitrite as specific peroxynitrite inhibitor MnTBAP markedly reduced ROS generation (about 70%) in the present study. Thus, TIIA and CT may have a capacity to direct ROS production from peroxynitrite to hydrogen peroxide, as both compounds showed no significant effects on peroxynitrite/hydrogen peroxide production; even they significantly reduced NO and superoxide productions which theoretically should reduce peroxynitrite formation. Partial supporting evidence is that TIIA and CT increased mitochon- drial SOD activity, which may result in increase in hydrogen peroxide formation. Further study is required to confirm this hypothesis. The effects of tanshinones on other antioxidant enzymes such as glutathione peroxidase and catalyse were not examined in this study which also requires further investigation. The present result is in line with a previous report showing that increased NO and ONOO−generations resulted in enhanced mitochondrial superoxide generation by block- ing mitochondrial electron transport chain [35]. NO can act as a physiological regulator of respiration by reversibly inhibiting cytochrome c-oxidase at the low concentrations Evidence-Based Complementary and Alternative Medicine Evidence-Based Complementary and Alternative Medicine 10 Changes in intracellular calcium level during hypoxia are important in mitochondrial functions, especially in mito- chondrial membrane permeability transition pore opening [40]. Decreased intracellular ATP by hypoxia can decrease cellular pH by glycolysis activation and this elicits imbalanc- ing in intracellular ion exchange and, subsequently, increases in intracellular calcium level [41]. This is consistent with the present finding showing that hypoxia-induced ATP depletion was accompanied with increased intracellular calcium level. The finding of inhibition of intracellular calcium by TIIA and CT is consistent with previous reports in neonatal car- diomyocytes and rat coronary artery [42, 43]. 4. Discussion The increased intracellular calcium level is likely due to ATP depletion caused by hypoxia. Thus, it is possible that TIIA may regulate intracellular calcium through affecting ATP level. However, CT reduced intracellular calcium without affecting ATP levels. This suggests that other mechanisms such as endoplasmic reticulum-related stress, which also regulate intracellular calcium production [44], may also be involved. Additionally, studies have shown that increased intracellular calcium may increase mitochondrial ROS production [45, 46]. Therefore, tanshinones may have multiple targets of reducing mitochondrial ROS production.i [6] G. A. Walford, R. Moussignac, A. W. Scribner, J. Loscalzo, and J. A. Leopold, “Hypoxia potentiates nitric oxide-mediated apop- tosis in endothelial cells via peroxynitrite-induced activation of mitochondria-dependent and -independent pathways,” Journal of Biological Chemistry, vol. 279, no. 6, pp. 4425–4432, 2004. [7] J.-X. Chen and B. Meyrick, “Hypoxia increases Hsp90 binding to eNOS via PI3K-Akt in porcine coronary artery endothelium,” Laboratory Investigation, vol. 84, no. 2, pp. 182–190, 2004. [8] D. X. Zhang and D. D. Gutterman, “Mitochondrial reactive oxy- gen species-mediated signaling in endothelial cells,” American Journal of Physiology, vol. 292, no. 5, pp. H2023–H2031, 2007. [9] X.-B. Dong, C.-T. Yang, D. D. Zheng et al., “Inhibition of ROS-activated ERK1/2 pathway contributes to the protection of H2S against chemical hypoxia-induced injury in H9c2 cells,” Molecular and Cellular Biochemistry, vol. 362, no. 1-2, pp. 149– 157, 2012. [10] H. W. Chen, C. T. Chien, S. L. Yu, Y. Lee, and W. Chen, “Cyclosporine A regulate oxidative stress-induced apoptosis in cardiomyocytes: mechanisms via ROS generation, iNOS and Hsp70,” British Journal of Pharmacology, vol. 137, no. 6, pp. 771– 781, 2002. [11] B. Wu, M. Liu, and S. Zhang, “Dan Shen agents for acute ischaemic stroke,” Cochrane Database of Systematic Reviews, no. 2, Article ID CD004295, 2007. [12] L. Zhou, Z. Zuo, and M. S. S. Chow, “Danshen: an overview of its chemistry, pharmacology, pharmacokinetics, and clinical use,” Journal of Clinical Pharmacology, vol. 45, no. 12, pp. 1345–1359, 2005. In summary, the findings from the present study indicate that TIIA and CT protect H9c2 cells via preserving mitochon- dria function by reducing excess production of mitochondrial superoxide, SOD activity, intracellular NO, and calcium levels and restoring cellular ATP contents. These molecular mechanisms may be involved in the cardioprotective actions of TIIA and CT in hypoxic injuries. [13] J.-Y. Han, J.-Y. Fan, Y. Acknowledgments [14] J. Gao, G. Yang, R. Pi et al., “Tanshinone IIA protects neona- tal rat cardiomyocytes from adriamycin-induced apoptosis,” Translational Research, vol. 151, no. 2, pp. 79–87, 2008. This work was supported by grants from The National Institute of Complementary Medicine and RMIT University. H.-J. Jin was supported by a RMIT University International Research Scholarship. [15] H.-J. Hong, J.-C. Liu, P.-Y. Chen, J.-J. Chen, P. Chan, and T.-H. Cheng, “Tanshinone IIA prevents doxorubicin-induced cardiomyocyte apoptosis through Akt-dependent pathway,” International Journal of Cardiology, vol. 157, no. 2, pp. 174–179, 2012. 4. Discussion Horie et al., “Ameliorating effects of compounds derived from Salvia miltiorrhiza root extract on microcirculatory disturbance and target organ injury by ischemia and reperfusion,” Pharmacology & Therapeutics, vol. 117, no. 2, pp. 280–295, 2008. Evidence-Based Complementary and Alternative Medicine Evidence-Based Complementary and Alternative Medicine Evidence-Based Complementary and Alternative Medicine 11 [21] W. Wu, O. Platoshyn, A. L. Firth, and J. X.-J. Yuan, “Hypoxia divergently regulates production of reactive oxygen species in human pulmonary and coronary artery smooth muscle cells,” American Journal of Physiology, vol. 293, no. 4, pp. L952–L959, 2007. [36] C. I. Jones III, Z. Han, T. Presley et al., “Endothelial cell respiration is affected by the oxygen tension during shear expo- sure: role of mitochondrial peroxynitrite,” American Journal of Physiology, vol. 295, no. 1, pp. C180–C191, 2008. [37] G. C. Brown and V. Borutaite, “Nitric oxide, mitochondria, and cell death,” IUBMB Life, vol. 52, no. 3–5, pp. 189–195, 2002. [22] H.-W. Chen, C.-T. Chien, S.-L. Yu, Y.-T. Lee, and W.-J. Chen, “Cyclosporine A regulate oxidative stress-induced apoptosis in cardiomyocytes: mechanisms via ROS generation, iNOS and Hsp70,” British Journal of Pharmacology, vol. 137, no. 6, pp. 771– 781, 2002. [38] M. C. Verhaar, P. E. Westerweel, A. J. van Zonneveld, and T. J. Rabelink, “Free radical production by dysfunctional eNOS,” Heart, vol. 90, no. 5, pp. 494–495, 2004. [39] R. Lin, W.-R. Wang, J.-T. Liu, G.-D. Yang, and C.-J. Han, “Protective effect of tanshinone IIA on human umbilical vein endothelial cell injured by hydrogen peroxide and its mecha- nism,” Journal of Ethnopharmacology, vol. 108, no. 2, pp. 217– 222, 2006. [23] D. J. McConkey and L. Nutt, “Measurement of changes in intracellular calcium during apoptosis,” Methods in Molecular Biology, vol. 282, pp. 117–130, 2004. [24] G. Grynkiewicz, M. Poenie, and R. Y. Tsien, “A new generation of Ca2+ indicators with greatly improved fluorescence proper- ties,” Journal of Biological Chemistry, vol. 260, no. 6, pp. 3440– 3450, 1985. [40] A. P. Halestrap and P. Pasdois, “The role of the mitochondrial permeability transition pore in heart disease,” Biochimica et Biophysica Acta, vol. 1787, no. 11, pp. 1402–1415, 2009. [41] P. S. Brookes, Y. Yoon, J. L. Robotham, M. W. Anders, and S. Sheu, “Calcium, ATP, and ROS: a mitochondrial love-hate triangle,” American Journal of Physiology, vol. 287, no. 4, pp. C817–C833, 2004. [25] A. J. Patterson, D. Xiao, F. Xiong, B. Dixon, and L. Zhang, “Hypoxia-derived oxidative stress mediates epigenetic repres- sion of PKC𝜀gene in foetal rat hearts,” Cardiovascular Research, vol. 93, no. 2, pp. 302–310, 2012. [42] P. Yang, Y.-H. Jia, J. Li, L.-J. Li, and F.-H. Evidence-Based Complementary and Alternative Medicine Zhou, “Study of anti- myocardial cell oxidative stress action and effect of tanshinone IIA on prohibitin expression,” Journal of Traditional Chinese Medicine, vol. 30, no. 4, pp. 259–264, 2010. [26] F. Jung, U. Weiland, R. A. Johns, C. Ihling, and S. Dimmeler, “Chronic hypoxia induces apoptosis in cardiac myocytes: a pos- sible role for Bcl-2-like Proteins,” Biochemical and Biophysical Research Communications, vol. 286, no. 2, pp. 419–425, 2001. [43] F. F. Y. Lam, J. H. K. Yeung, K. M. Chan, and M. Y. O. Penelope, “Mechanisms of the dilator action of cryptotanshinone on rat coronary artery,” European Journal of Pharmacology, vol. 578, no. 2-3, pp. 253–260, 2008. [27] M. J. Kim, C. Moon, M. H. Kim, S. H. Lee, E. J. Baik, and Y. Jung, “Role of PKC-𝛿during hypoxia in heart-derived H9c2 cells,” Japanese Journal of Physiology, vol. 54, no. 4, pp. 405–414, 2004. [28] J. Fu, H. Huang, J. Liu, R. Pi, J. Chen, and P. Liu, “Tanshinone IIA protects cardiac myocytes against oxidative stress-triggered damage and apoptosis,” European Journal of Pharmacology, vol. 568, no. 1–3, pp. 213–221, 2007. [44] I. J. Park, M. J. Kim, O. J. Park et al., “Cryptotanshinone induces ER stress-mediated apoptosis in HepG2 and MCF7 cells,” Apoptosis, vol. 17, no. 3, pp. 248–257, 2012. [45] T. Kaminishi and K. J. Kako, “Sensitivity to oxidants of mitochondrial and sarcoplasmic reticular calcium uptake in saponin-treated cardiac myocytes,” Basic Research in Cardiol- ogy, vol. 84, no. 3, pp. 282–290, 1989. [29] J. Gao, G. Yang, R. Pi et al., “Tanshinone IIA protects neona- tal rat cardiomyocytes from adriamycin-induced apoptosis,” Translational Research, vol. 151, no. 2, pp. 79–87, 2008. [46] H.-Y. Sun, N.-P. Wang, F. Kerendi et al., “Hypoxic post- conditioning reduces cardiomyocyte loss by inhibiting ROS generation and intracellular Ca2+ overload,” American Journal of Physiology, vol. 288, no. 4, pp. H1900–H1908, 2005. [30] R. Mahesh, H. W. Jung, G. W. Kim, Y. S. Kim, and Y.- K. Park, “Cryptotanshinone from salviae miltiorrhizae radix inhibits sodium-nitroprusside-induced apoptosis in neuro-2a cells,” Phytotherapy Research, vol. 26, no. 8, pp. 1211–1219, 2012. [31] B. Kadenbach, “Intrinsic and extrinsic uncoupling of oxidative phosphorylation,” Biochimica et Biophysica Acta, vol. 1604, no. 2, pp. 77–94, 2003. [32] J. K. Eun, S. Jung, H. S. Kun et al., “Antidiabetes and antiobesity effect of cryptotanshinone via activation of AMP-activated protein kinase,” Molecular Pharmacology, vol. 72, no. 1, pp. 62– 72, 2007. [33] K. Terai, Y. References [16] Z. Zhou, S.-Q. Wang, Y. Liu, and A.-D. Miao, “Cryptotan- shinone inhibits endothelin-1 expression and stimulates nitric oxide production in human vascular endothelial cells,” Biochim- ica Et Biophysica Acta, vol. 1760, no. 1, pp. 1–9, 2006. [1] J. Cassavaugh and K. M. Lounsbury, “Hypoxia-mediated bio- logical control,” Journal of Cellular Biochemistry, vol. 112, no. 3, pp. 735–744, 2011. [17] C. Pan, L. Lou, Y. Huo et al., “Salvianolic acid B and Tanshinone IIA attenuate myocardial ischemia injury in mice by NO production through multiple pathways,” Therapeutic Advances in Cardiovascular Disease, vol. 5, no. 2, pp. 99–111, 2011. [2] C. X. C. Santos, N. Anilkumar, M. Zhang, A. C. Brewer, and A. M. Shah, “Redox signaling in cardiac myocytes,” Free Radical Biology and Medicine, vol. 50, no. 7, pp. 777–793, 2011. [18] M. J. Kim, C.-H. Moon, M.-Y. Kim et al., “KR-32570, a novel Na+ / H+ exchanger-1 inhibitor, attenuates hypoxia-induced cell death through inhibition of intracellular Ca2+ overload and mitochondrial death pathway in H9c2 cells,” European Journal of Pharmacology, vol. 525, no. 1–3, pp. 1–7, 2005. [3] G. Solaini, A. Baracca, G. Lenaz, and G. Sgarbi, “Hypoxia and mitochondrial oxidative metabolism,” Biochimica et Biophysica Acta, vol. 1797, no. 6-7, pp. 1171–1177, 2010. [4] Q. Gao and M. S. Wolin, “Effects of hypoxia on relationships between cytosolic and mitochondrial NAD(P)H redox and superoxide generation in coronary arterial smooth muscle,” American Journal of Physiology, vol. 295, no. 3, pp. H978–H989, 2008. [19] R. D. Rakhit, R. J. Edwards, J. W. Mockridge et al., “Nitric oxide- induced cardioprotection in cultured rat ventricular myocytes,” American Journal of Physiology, vol. 278, no. 4, pp. H1211–H1217, 2000. [5] R. T. Kolamunne, M. Clare, and H. R. Griffiths, “Mitochondrial superoxide anion radicals mediate induction of apoptosis in cardiac myoblasts exposed to chronic hypoxia,” Archives of Biochemistry and Biophysics, vol. 505, no. 2, pp. 256–265, 2011. [20] K. K. Griendling, C. A. Minieri, J. D. Ollerenshaw, and R. W. Alexander, “Angiotensin II stimulates NADH and NADPH oxidase activity in cultured vascular smooth muscle cells,” Circulation Research, vol. 74, no. 6, pp. 1141–1148, 1994. Evidence-Based Complementary and Alternative Medicine Hiramoto, M. Masaki et al., “AMP-activated protein kinase protects cardiomyocytes against hypoxic injury through attenuation of endoplasmic reticulum stress,” Molecular and Cellular Biology, vol. 25, no. 21, pp. 9554–9575, 2005. [34] Q. Gao and M. S. Wolin, “Effects of hypoxia on relationships between cytosolic and mitochondrial NAD(P)H redox and superoxide generation in coronary arterial smooth muscle,” American Journal of Physiology, vol. 295, no. 3, pp. H978–H989, 2008. [35] G. Ilangovan, S. Osinbowale, A. Bratasz et al., “Heat shock regu- lates the respiration of cardiac H9c2 cells through upregulation of nitric oxide synthase,” American Journal of Physiology, vol. 287, no. 5, pp. C1472–C1481, 2004.
https://openalex.org/W2137893869	https://europepmc.org/articles/pmc2635756?pdf=render	English	null	Integration of Motor Proteins – Towards an ATP Fueled Soft Actuator	International journal of molecular sciences	2,008	cc-by	8,431	Graduate School of Science, Hokkaido University, Sapporo 060-0810, Japan Received: 18 June 2008; in revised form: 8 August 2008 / Accepted: 27 August 2008 / Published: 4 September 2008 Abstract: We present a soft bio-machine constructed from biological motors (actin/myosin). We have found that chemically cross-linked polymer-actin complex gel filaments can move on myosin coated surfaces with a velocity as high as that of native F- actin, by coupling to ATP hydrolysis. Additionally, it is shown that the velocity of polymer-actin complex gel depends on the species of polycations binding to the F-actins. Since the design of functional actuators of well-defined size and morphology is important, the structural behavior of polymer-actin complexes has been investigated. Our results show that the morphology and growth size of polymer-actin complex can be controlled by changes in the electrostatic interactions between F-actins and polycations. Our results indicate that bio actuators with desired shapes can be created by using a polymer-actin complex. Keywords: Biological motors, self-assembly, hierarchical structure, soft-bio-machine. Int. J. Mol. Sci. 2008, 9, 1685-1703; DOI: 10.3390/ijms9091685 Int. J. Mol. Sci. 2008, 9, 1685-1703; DOI: 10.3390/ijms9091685 International Journal of Molecular Sciences ISSN 1422-0067 www.mdpi.org/ijms/ OPEN ACCESS International Journal of Molecular Sciences ISSN 1422-0067 www.mdpi.org/ijms/ OPEN ACCESS 1. Introduction There are two basic differences between the motion in a man-made machine and in a biological motor. One is in their principles. The motion of a man-made machine, which is constructed from hard and dry materials such as metals, ceramics or plastics, is realized by the relative displacement of the macroscopic constituent parts of the machine. In contrast to this, the motion of a living organism, which consists of soft and wet protein and tissues, is caused by a molecular deformation that is integrated to a macroscopic level through its hierarchical structure [1–3]. The other is in their energy Int. J. Mol. Sci. 2008, 9 1686 sources. The man-made machine is fueled by electrical or thermal energy, which is used with an efficiency of around 30%, but a biological motor is driven by direct conversion of chemical energy with much higher efficiency than that possible with a man-made machine [4]. In order to create biomimetic systems, polymer gels have been employed due to their reversible size and shape change, thereby realizing the motion by integrating the deformation on a molecular level. The contractile collagen fiber prepared by Katchalsky et al. are the earliest man-made example of gel actuators [5, 6]. Later on, Osada et al. and others developed several gel machines based on synthetic polymers [7-13]. Recently, various types of peptide-based gels as well as hybrid gels have been developed [14-18]. The specific properties of these gels might be beneficial to constructing soft actuators based on the concept of molecular machines [19, 20]. Polypeptides from living organisms are attractive materials because of their high functionality, such as enzyme activity, ability of self-organization, or selective recognition. Some groups have constructed devices by depositing or arranging polypeptides on substrates. For example, thin films composed of bacteriorhodopsin [20] have been successfully used as photoelectric devices, and thin films of glucose oxidase as glucose-response biosensors [21]. Actin-based metallic nano-wires and actuators were also constructed recently [22-25]. Furthermore, various kinds of motor proteins, such as actin myosin [26-28] or microtubule kinesin [29], as well as other proteins that generate rotational motion like F1-ATPase [30], have been used as micro-motors [31]. Thus micro-machines and micro-devices driven by bio-molecules have been received considerable attention. However, most of these devices have exploited only in one aspect of biological function, partly because the polypeptides were immobilized or arranged two-dimensionally on the surface of substrate. 1. Introduction By using the chemical cross-linking technique, polypeptides could have a highly ordered structure such as three-dimensional gels. Actins and myosins are major components of muscle proteins and play an important role in dynamic motion of animals that is caused by the molecular deformation using the chemical energy released by hydrolysis of ATP. Recently, we have found that F-actins can be self-assembled into large bundles in the presence polycations through polymer complex formation. These polycation-actin complexes, several tens of times the length of native actin filaments, move along a chemically cross- linked myosin fibrous gel with a velocity as high as that of native actins, by coupling to ATP hydrolysis [22-24, 32]. This result indicates that muscle proteins can be tailored into desired shape and size without sacrificing their bioactivities, and can be used as biomaterials. 2. Actin gel formed from polymer-actin complexes [32] Under physiological conditions, globular actin (G-actin) monomers are self-assembled into linear filaments (F-actin). F-Actins also assemble into parallel bundles or form a cross-linked network in the presence of actin linker proteins. It is believed that specific linker proteins are responsible for the morphology of actin assemblies [33] However, it has been reported that the various morphologies of actin bundles such as Drosophila bristle, nurse cell strut, stereocilia, and the acrosomal process are independent of the specific type of linker proteins. In addition, it has been reported that the binding of some proteins, including calponin, dystrophin, and MARKS peptide, to F-actins is due to electrostatic interaction between negatively charged F-actin and positively charged binding proteins without specific binding sites [34-36]. These facts indicate that the morphology of actin assemblies is Int. J. Mol. Sci. 2008, 9 Int. J. Mol. Sci. 2008, 9 1687 determined not only by specific linker proteins but also by other factors such as the concentration of the linker proteins, environmental conditions and the kinetics of actin-linker protein interactions. Here, we employed a model system consisting of F-actin and synthetic cationic polymers to create assemblies with various morphologies. Figure 1. Fluorescence microscope images of polymer-actin complexes formed by mixing F-actin and various cationic polymers at room temperature. (a) p-Lys, (b) 3,3-ionene, (c) 6,6-ionene, (d) 6,12-ionene, (e) F-actin only. The molar ratio of ammonium cation of polymer to monomeric actin was kept constant at 30:1 for x,y-ionene polymers and 100:1 for p-Lys, which correspond to weight ratios of [3,3-ionene]/[actin]) 0.41 g/g, [6,6- ionene]/[actin]) 0.61 g/g, [6,12-ionene]/[actin]) 0.81 g/g, [p-Lys]/[actin]) 0.35 g/g. Actin concentration was 0.001 mg/mL. Taken from ref. [32] with permission. Since the isoelectric point of actins is pH 4.7, F-actins are negatively charged in neutral buffers. Therefore, they have been assumed to form complexes with cationic polymers through electrostatic interaction. Figures 1(a)–(d) show some examples of fluorescence microscope images of polymer-actin complexes obtained by mixing F-actins with poly-L-Lysine (p-Lys) [Figure 1(a)] and x,y-ionene polymers, which have a -[-(CH2)x-N+Br-(CH3)2-(CH2)y-N+Br-(CH3)2-]n- structure [Figures 1(b)-(d)], for 120 min. The morphology of the complexes depends on the chemical structure of the polycations [32]. One can see that large filamentous, stranded and branched complexes of 20–30 μm in size are formed in the presence of p-Lys, 3,3- 6,6- and 6,12-ionene polymers and their morphological nature, both size and shape, are remarkably different from that of native F-actin [Figure 1(e)]. 2. Actin gel formed from polymer-actin complexes [32] 2008, 9 1688 polymer gives the smallest complexes. These results imply that hydrophobicity and charge density of the ionene polymers are important in complex formation. In the case of 6, 4- and 6, 10-ionene polymer, after reached maximum, decrease in the average length of the complexes is observed. This is attributed to the compacting of the F-actins in the complexes. The effect of electrostatic interactions is proved, since no complexation was observed with negative controls such as p-Glu, DNA and neutral polymer, such as PEG. The average lengths of polymer-actin complexes are shown in Figure 2(a) (b). Figure 2. (a) Time courses of polymer-actin complexes growth. (●): 3,3-ionene-actin complexes, (▲): 6,4-ionene-actin complexes, (■): 6,6-ionene-actin complexes, (○): 6,10- ionene-actin complexes, (△): 6,12-ionene-actin complexes, (□): p-Lys-actin complexes, (◆): p-Glu-actin complexes, (×): [DNA-actin complexes, (+): PEG-actin complexes. (b) Average length of polymer–actin complexes observed from fluorescence microscope images (white columns) and from transmission electron microscope (TEM) images (shade columns) at 210–300 min. The molar ratio of ammonium cation to monomeric actin was 30:1 for x,y-ionene polymers and 100:1 for p-Lys. The corresponding weight ratios were as follows: [3,3-ionene] /[actin]=0.41 g/g, [6,4-ionene]/[actin]=0.54 g/g, [6,6-ionene]/[actin]= 0.61 g/g, [6,10-ionene]/[actin]=0.74 g/g, [6,12-ionene]/[actin]=0.81 g/g, [p-Lys]/ [actin]=0.35 g/g, [p-Glu]/[actin]=0.36 g/g, [DNA]/[actin]= 0.77 g/g, [PEG]/[actin]=0.10 g/g. Actin concentration: 0.001 mg/mL. Taken from ref. [32] with permission. Concerning the analysis of the lateral structure of the polymer-actin complexes, transmission electron microscopy (TEM) accompanied with negative staining technique was employed. Although the samples must be exposed to high vacuum in a dry state in this technique, the higher resolution offers advantage compared to analyzing from the images of fluorescent microscopes concerning such a precise structure. The TEM measurements show that the polymer-actin complexes are bundles that consist of 3–20 filaments [24]. The average width of the p-Lys-actin complex is 21.0 nm, with a standard deviation of 2.6 nm. Comparing with the native F-actins that have an average width of 12.1 nm with a standard deviation of 1.2 nm, p-Lys-actin complexes are only slightly thicker than that of the native F-actin with almost the same width scattering. 3,3-ionene complexes also showed a very thin and homogeneous wire-like morphology, showing an average width of 16.1 nm with a standard deviation of 1.7 nm. Concerning the analysis of the lateral structure of the polymer-actin complexes, transmission electron microscopy (TEM) accompanied with negative staining technique was employed. 2. Actin gel formed from polymer-actin complexes [32] The relationship between the morphology of polycation-actin complex and the concentrations of components is discussed in Section 3. Since the isoelectric point of actins is pH 4.7, F-actins are negatively charged in neutral buffers. Therefore, they have been assumed to form complexes with cationic polymers through electrostatic interaction. Figures 1(a)–(d) show some examples of fluorescence microscope images of polymer-actin complexes obtained by mixing F-actins with poly-L-Lysine (p-Lys) [Figure 1(a)] and x,y-ionene polymers, which have a -[-(CH2)x-N+Br-(CH3)2-(CH2)y-N+Br-(CH3)2-]n- structure [Figures 1(b)-(d)], for 120 min. The morphology of the complexes depends on the chemical structure of the polycations [32]. One can see that large filamentous, stranded and branched complexes of 20–30 μm in size are formed in the presence of p-Lys, 3,3- 6,6- and 6,12-ionene polymers and their morphological nature, both size and shape, are remarkably different from that of native F-actin [Figure 1(e)]. The relationship between the morphology of polycation-actin complex and the concentrations of components is discussed in Section 3. of actins is pH 4.7, F-actins are negatively charged in neutral buffers. Since the isoelectric point of actins is pH 4.7, F-actins are negatively charged in neutral buffers. Therefore, they have been assumed to form complexes with cationic polymers through electrostatic interaction. Figures 1(a)–(d) show some examples of fluorescence microscope images of polymer-actin complexes obtained by mixing F-actins with poly-L-Lysine (p-Lys) [Figure 1(a)] and x,y-ionene polymers, which have a -[-(CH2)x-N+Br-(CH3)2-(CH2)y-N+Br-(CH3)2-]n- structure [Figures 1(b)-(d)], for 120 min. The morphology of the complexes depends on the chemical structure of the polycations [32]. One can see that large filamentous, stranded and branched complexes of 20–30 μm in size are formed in the presence of p-Lys, 3,3- 6,6- and 6,12-ionene polymers and their morphological nature, both size and shape, are remarkably different from that of native F-actin [Figure 1(e)]. The relationship between the morphology of polycation-actin complex and the concentrations of components is discussed in Section 3. The number-average length of fluorescence image of F-actins is 2.14 μm, with a standard deviation of 0.11 μm (average over 784 samples) in the F-buffer that is known to cause a transformation to the fibrous polymerized state of actin (F-actin) from the globular monomer actin (G-actin). However, polymer-actin complexes grow with time and reach as large as 5–20 μm within 1 or 2 h, which is about 2–10 times larger than that of native F-actin. p-Lys produces a large complex, whereas 3,3-ionene Int. J. Mol. Sci. 2. Actin gel formed from polymer-actin complexes [32] Although the samples must be exposed to high vacuum in a dry state in this technique, the higher resolution offers advantage compared to analyzing from the images of fluorescent microscopes concerning such a precise structure. The TEM measurements show that the polymer-actin complexes are bundles that consist of 3–20 filaments [24]. The average width of the p-Lys-actin complex is 21.0 nm, with a standard deviation of 2.6 nm. Comparing with the native F-actins that have an average width of 12.1 nm with a standard deviation of 1.2 nm, p-Lys-actin complexes are only slightly thicker than that of the native F-actin with almost the same width scattering. 3,3-ionene complexes also showed a very thin and homogeneous wire-like morphology, showing an average width of 16.1 nm with a standard deviation of 1.7 nm. 1689 Int. J. Mol. Sci. 2008, 9 In the case of x,y-ionene polymers, they form bundles of thin filaments with F-actin above certain critical polymer concentration that are dependent on each polymer. The effect of x,y- of the ionene polymers was observed in the width of complexes. Actin- 6,4-, 6,6-, 6,10-, and 6,12-ionene complexes have average widths of 79.0, 59.3, 38.7 and 66.1 nm, with a standard deviation of 60, 29, 21, and 27 nm, respectively (Figure 3). From the large scattering in the width of actin-6,4-, 6,6-, 6,10-, and 6,12- ionene complexes, the morphology of complexes seems to have randomness quantitatively. Indeed, a ring-shaped complex (nano-ring) is observed in a 6,6-ionene-actin complex occasionally[32]. Figure 3. Average width of polymer–actin complexes obtained at 240 min by TEM images. Error bar means standard deviation. Taken from ref. [32] with permission. Figure 3. Average width of polymer–actin complexes obtained at 240 min by TEM images. Error bar means standard deviation. Taken from ref. [32] with permission. 3. Polymorphism of actin complexes [37] 3. Polymorphism of actin complexes [37] In order to investigate the polymorphism, the complexation of F-actin with poly-N-[3-(dimethyl- amino)propyl]-acrylamide(PDMAPAA-Q), which has positive charges on its side chain, were further investigated systematically. The polymer-actin complexes exhibit a rich polymorphism in a wide range of the polycation concentration (CP) and KCl salt concentration (CS), as elucidated by the fluorescent analysis and TEM analysis, which show micro- and nano-scale images, respectively. There are five characteristic phases in the CP-CS phase diagram (Figure 4). Figure 5 shows the fluorescent images and TEM images for the polymer-actin complexes in the five phases. 2. Actin gel formed from polymer-actin complexes [32] In phase I, F-actin does not grow. In phase II, we observe the coexistence of native F-actin and polymer-actin complexes. In the coexistence phase (II), the fraction of native F-actins increases with the increase of CS. The borderline between phases I and II shifts to a higher CP with the increase of CS. This can be explained by screening effect of salts on the electrostatic interaction between F-actins and polycations. As CP and CS increase, the F- actins form complexes with polycations and exhibit various structures. The polymer- actin complexes evolve from the cross-linked structure dominant phase (III), to the branched-structure dominant phase (IV), and then to the parallel-bundle dominant phase (V) as shown by the TEM images in Figure 5. 3. Polymorphism of actin complexes [37] 3. Polymorphism of actin complexes [37] In order to investigate the polymorphism, the complexation of F-actin with poly-N-[3-(dimethyl- amino)propyl]-acrylamide(PDMAPAA-Q), which has positive charges on its side chain, were further investigated systematically. The polymer-actin complexes exhibit a rich polymorphism in a wide range of the polycation concentration (CP) and KCl salt concentration (CS), as elucidated by the fluorescent analysis and TEM analysis, which show micro- and nano-scale images, respectively. There are five characteristic phases in the CP-CS phase diagram (Figure 4). Figure 5 shows the fluorescent images and TEM images for the polymer-actin complexes in the five phases. In phase I, F-actin does not grow. In phase II, we observe the coexistence of native F-actin and polymer-actin complexes. In the coexistence phase (II), the fraction of native F-actins increases with the increase of CS. The borderline between phases I and II shifts to a higher CP with the increase of CS. This can be explained by screening effect of salts on the electrostatic interaction between F-actins and polycations. As CP and CS increase, the F- actins form complexes with polycations and exhibit various structures. The polymer- actin complexes evolve from the cross-linked structure dominant phase (III), to the branched-structure dominant phase (IV), and then to the parallel-bundle dominant phase (V) as shown by the TEM images in Figure 5. 1690 Int. J. Mol. Sci. 2008, 9 The fluorescent images in Figure 5 show that the polymer-actin complexes of the cross-linked structure dominant phase (III) are in compact globule state, whereas those of the parallel bundle dominant phase are in the extended state (V). The globule size of 15–20μm is attributed to the persistence length of F-actin, which is assumed as ca. 10μm. The morphological change of polymer- actin complexes from the compact globule state to the extended bundle state is due to the increase of bending rigidity that increases with the thickness D of actin bundles, varying as D [38, 39]. Figure 4. Phase diagram for the morphology of polymer-actin complexes, in which the phase behavior is summarized as a function of PDMAPAA-Q concentration CP and KCl concentration CS for a constant actin concentration CA = 0.01 mg/mL (2.32 × 10-7M). ○, F-actins; □, coexistence (polymer-actin complex and native F-actins); ▲, cross-linked structure dominant phase; , branched structure dominant phase; and ■, parallel bundle dominant phase. The dotted line shows the possible borderline between the native F-actin and F-actin with charge inversion. Taken from ref. 3. Polymorphism of actin complexes [37] [37] with permission. 4. Anisotropic nucleation growth mechanism for actin bundle [40] 4. Anisotropic nucleation growth mechanism for actin bundle [40] A model system consisting of F-actin and p-Lys is used to investigate the physical origin of the well–defined thickness D and length L of the actin bundles. This system reveals that D growth is nearly completed in the initial stage of bundle growth, while dramatic L growth starts later on, after completion of D growth. Additionally, D decreases with the increase in the polycation mediated attraction between F-actins but is hardly influenced by the actin concentration (CA), while L increases with the increase of CA. From these results, an anisotropic nucleation-growth model is proposed, in which D is determined by the critical nucleus size D0, while the length L is determined by free actin concentration relative to nucleus concentration. Here, the factor to determine D in detail is discussed. The free energy, ∆G of forming bundles can be expressed as ∆G = ∆Gvol + ∆GSurf = −∆gV + γ S, where ∆g is the energy gain due to bundle formation of F-actins per volume, V is the volume of actin bundles, γ is the surface energy, and S is the surface area of actin bundles. For the nucleation-growth process, spontaneous growth occurs only when stable nuclei are formed, which occurs when the favorable volume term ∆Gvol surpasses the 008, 9 1691 Int. J. Mol. Sci. 2008, 9 Int. J. Mol. Sci. 2008, 9 1691 unfavorable surface term ∆GSurf. It is presupposed that the bundle is a cylindrical object with a length L and a thickness D. The free energy ∆G (D, L) for the formation of the bundle is: γ π π π 2 ) 2 ( 4 2 2 DL D g L D G + + Δ − = Δ ( ) L D L gD γ π 4 4 2 − Δ − ≈ (1) (1) where it is assumed that for a newborn nucleus, D<< g Δ / γ << L. The critical thickness D0 is determined by the free energy barrier ∆G* where ∆G is maximum. From , we have 0 = ∂ Δ ∂ D G g D Δ ≈ γ 2 0 (2) g L G Δ ≈ Δ 2 πγ * (3) (2) (3) Figure 5. 3. Polymorphism of actin complexes [37] Typical morphologies of polymer-actin complexes in phases I, II, III, IV, and V as observed by TEM images and fluorescence images. Scale bars present 200 nm for TEM images and 25 μm for fluorescence images. Taken from ref. [37] with permission. Figure 5. Typical morphologies of polymer-actin complexes in phases I, II, III, IV, and V as observed by TEM images and fluorescence images. Scale bars present 200 nm for TEM images and 25 μm for fluorescence images. Taken from ref. [37] with permission. Int. J. Mol. Sci. 2008, 9 1692 Eq. (2) and (3) indicate that D0 and ∆G* are reciprocally proportional to the energy gain per volume for bundle formation ∆g. This anisotropic growth of actin bundles is originated from the rod-like polyelectrolytes nature of F-actins. Therefore, this model not only provides insight into physical origins of how the growth of cellular actin bundles is determined but also can be broadly applied to the self-assembly of rod-like polyelectrolytes. 5. Three dimensional structure of actin bundles formed with polycations [41] In Sections 2 and 3, the thickness D of actin bundle is evaluated by using TEM projection images, which smear the cross-sectional structure information. However recently developed transmission electron microtomography (TEMT) allows to obtain the three dimensional structure image of actin bundles. Figure 6. Fluorescence images and cross-sectional images of the actin bundles formed with PDMAPAA-Q at various concentrations of polymer (CP) and KCl (CS) obtained by fluorescence microscopy and transmission electron microtomography (TEMT). Arrows along the x-y plane of the TEMT images represent the positions of the images taken along the y-z plane. The fluorescence and cross-sectional images presented are drawn to scales of 10 µm and 50 nm, respectively. Taken from ref. [41] with permission. Int. J. Mol. Sci. 2008, 9 1693 Figure 6 shows the fluorescent and TEMT images of actin bundles formed with PDMAPAA-Q at various CS and CP. In the absence of polycations, F-actins at a concentration of 0.01 mg/mL are present as single filaments of thickness ∼10 nm; the filaments exhibit a polydisperse length distribution of 1-10 μm, with an average of approximately 5 μm. The actin bundles show a ribbon-like cross-sectional morphology with lateral packing of the actin filaments in the presence of polycations at a low CS. The cross-sectional morphology changes to cylindrical with hexagonal packing at a high CS. The width and height of the bundles formed at various CP and CS are summarized in Figure 7. At a constant CS of 0.01M, the width increases with an increase in CP, while the height barely depends on the CP, showing a value that is the same as that of F-actin. In contrast, at a constant CP of 10-5 M, both the width and height increase with an increase in CS from 0.01 to 0.3 M (triangular symbols). The fluorescent image in Figure 6 reveals that the actin bundles undergo morphological changes from a compact globule structure to an extended structure. These morphological changes might partially be attributed to the cross-sectional changes in the structure of actin bundles form a ribbon-like structure that is flexible to a cylindrical structure that is rigid. Figure 7. The width (closed symbols) and height (open symbols) of the actin bundles formed at various PDMAPAA-Q concentration (CP) and KCl concentration (CS). (●, ○), CS = 0.01 M; (S,U), CS = 0.3 M. 5. Three dimensional structure of actin bundles formed with polycations [41] The width and height, defined in the illustration, are measured along the guide line in the schematic views of Figure 1. The size of the bundle formed at CP = 10−5 M, CS = 0.01 M and CP = 10−1 M, CS= 0.01 M was the average of two and three samples, respectively. Other data were measured for one sample. The diameter of F-actin is 11.8 nm, as determined from Figure 1, which corresponds well with our previous result [32]. The relative size to the F-actin diameter, which means the ratio of the bundle width or height to the F-actin diameter, is also displayed in the figure as the right vertical axis. Taken from ref. [41] with permission. Int. J. Mol. Sci. 2008, 9 1694 Thus, cross-sectional morphology of actin bundles depends on the ionic strength but is insensitive to the concentration of polycation. The anisotropy in the cress-sectional structure of actin bundles may originate from an anisotropic electrostatic interaction between F-actins. At present, no satisfactory explanation on how this anisotropy is induced has been given. According to the anisotropic nucleation growth model, which states that D is determined at the nucleation stage, the anisotropy in the lateral direction should occur in the initial nucleation stage when F-actins form nuclei mediated by polycations [40]. 6. Oriented myosin gel formed under shear flow [22, 23] Chemically cross-linked myosin gel with its oriented filament array can be obtained by reacting scallop myosin at pH 7.0 using transglutaminase (TG) under shear flow-induced stretching. The oriented myosin gel is semi-transparent, showing a swelling degree of ca. 100, and a Young modulus of 190 Pa in the oriented direction, which is more than two times larger than that of the myosin gel prepared without stretching. The orientation of myosin fibers in the gels was analyzed by scanning electron microscopy (SEM) and atomic force microscopy (AFM) (Figure 8). From the images, distinct bundles of regularly oriented filaments ca. 1.5 μm in diameter were observed, indicating that the rod- like myosin molecules are self-organized with orientation to form a hierarchical structure. The molecular orientation within the filaments was confirmed by the strong IR dichroism of carbonyl absorption at 1600 cm-1, which could not be observed in the absence of stretching. The chemically cross-linked myosin gel shows an ATPase activity as high as that of native myosins in the presence of 0.5 w/w native actin. Myosin gels cross-linked by other cross-linking agents, such as glutaraldehyde and 1-ethyl-3-(3-dimethylaminoprolyl)carbodiimide, also showed an ATPase activity, though not as high as those cross-linked by TG. The motion of these myosin gels will be discussed in next section. Figure 8. The orientation of myosin fibers in the gels was analyzed by scanning electron microscopy (SEM) and atomic force microscopy (AFM). Figure 8. The orientation of myosin fibers in the gels was analyzed by scanning electron microscopy (SEM) and atomic force microscopy (AFM). 7. Motility assay of F-actin on oriented myosin gel [23] F-actins showed a preferential motion along the axis of oriented myosin gel as elucidated by the degree of anisotropy (D.A.), which is defined as the ratio of the square-root average velocity in the fiber direction to that perpendicular to the fiber direction. 7. Motility assay of F-actin on oriented myosin gel [23] F-actins showed a preferential motion along the axis of oriented myosin gel as elucidated by the degree of anisotropy (D.A.), which is defined as the ratio of the square-root average velocity in the fiber direction to that perpendicular to the fiber direction. Int. J. Mol. Sci. 2008, 9 1695 y x A D = . . (4) y x A D = . . 6. Oriented myosin gel formed under shear flow [22, 23] (4) N N 2 2 (4) where, ∑ ∑ = = = = N i i N i i N y y N x x 1 2 1 2 , (5) (5) (5) The D.A. measured on the non-oriented myosin gel was 1.1 (average over 66 samples), and that on the oriented myosin gel was 1.7 (average over 91 samples). The mean velocity on the non-oriented myosin gel was 0.69 μm/s with a standard deviation of 0.24 μm/s, while that on the oriented myosin gel was 0.83 μm/s with a standard deviation of 0.30 μm/s. Thus, F-actin filaments prefer to move along the axis of the oriented myosin gel with an increased velocity. 8. Motility assay of polymer-actin complex gel [22] The p-Lys-actin complex can be cross-linked to obtain a stable structure. The complex gel cross- linked with TG also shows a high motility on the oriented myosin gel in spite of its large dimension. Figure 9(a). The sequential fluorescent images of p-Lys-actin complex gels on the oriented myosin gel oriented myosin gel Here, it should be noted that if no cross-linking was performed on the polymer -actin complexes, disassembling of them into F-actins is observed. The p-Lys-actin complex gels move preferentially along the axis of the oriented myosin gel almost without path deviations. Figure 9 shows the sequential fluorescent image (a) and velocity distributions (b) of p-Lys-actin complex gels on the oriented myosin gel as a function of the individual filament size. The p-Lys-actin complex gels, about four times larger than that native F-actin, move with an average velocity of 1μm/s, almost the same as that of native F- actins on the oriented myosin gel (opened circles). Some of the p-Lys-actin complex gels move as fast as 2.0 μm/s. In addition, D.A. of the p-Lys-actin complex gels was 2.2 (average over 38 samples) on the oriented myosin gel, which was higher than that of the native F-actin (D.A. = 1.7), indicating an enhanced directional preference along the axis of the oriented myosin gel. Thus, despite its increased mass, the p-Lys-actin complex gels, several tens to hundreds times the volume of the native F-actins, move on the covalently cross-linked myosin gel, with an increased velocity. This is rather surprising, since the interaction between the myosin gel and the actin gel can only occur at the two-dimensional Here, it should be noted that if no cross-linking was performed on the polymer -actin complexes, disassembling of them into F-actins is observed. The p-Lys-actin complex gels move preferentially along the axis of the oriented myosin gel almost without path deviations. Figure 9 shows the sequential fluorescent image (a) and velocity distributions (b) of p-Lys-actin complex gels on the oriented myosin gel as a function of the individual filament size. The p-Lys-actin complex gels, about four times larger than that native F-actin, move with an average velocity of 1μm/s, almost the same as that of native F- actins on the oriented myosin gel (opened circles). Some of the p-Lys-actin complex gels move as fast as 2.0 μm/s. In addition, D.A. 8. Motility assay of polymer-actin complex gel [22] of the p-Lys-actin complex gels was 2.2 (average over 38 samples) on the oriented myosin gel, which was higher than that of the native F-actin (D.A. = 1.7), indicating an enhanced directional preference along the axis of the oriented myosin gel. Thus, despite its increased mass, the p-Lys-actin complex gels, several tens to hundreds times the volume of the native F-actins, move on the covalently cross-linked myosin gel, with an increased velocity. This is rather surprising, since the interaction between the myosin gel and the actin gel can only occur at the two-dimensional Here, it should be noted that if no cross-linking was performed on the polymer -actin complexes, disassembling of them into F-actins is observed. The p-Lys-actin complex gels move preferentially along the axis of the oriented myosin gel almost without path deviations. Figure 9 shows the sequential fluorescent image (a) and velocity distributions (b) of p-Lys-actin complex gels on the oriented myosin gel as a function of the individual filament size. The p-Lys-actin complex gels, about four times larger than that native F-actin, move with an average velocity of 1μm/s, almost the same as that of native F- actins on the oriented myosin gel (opened circles). Some of the p-Lys-actin complex gels move as fast as 2.0 μm/s. In addition, D.A. of the p-Lys-actin complex gels was 2.2 (average over 38 samples) on the oriented myosin gel, which was higher than that of the native F-actin (D.A. = 1.7), indicating an enhanced directional preference along the axis of the oriented myosin gel. Thus, despite its increased mass, the p-Lys-actin complex gels, several tens to hundreds times the volume of the native F-actins, move on the covalently cross-linked myosin gel, with an increased velocity. This is rather surprising, since the interaction between the myosin gel and the actin gel can only occur at the two-dimensional Int. J. Mol. Sci. 2008, 9 Int. J. Mol. Sci. 2008, 9 1696 interface and due to cross-linking a considerable number of actin and myosin molecules are not involved in the sliding motion [22]. These results indicate that actins can be tailored into the desired shape and size without sacrificing their bioactivities by using complex formation with synthetic polymers. Figure 9(b). Individual velocities of native F-actins (left; opened circle) and p-Lys-actin gels (right; closed circle) of various length on the oriented myosin gel. 8. Motility assay of polymer-actin complex gel [22] 6 8 10 12 14 Actin gel length (μm) 0 0.5 1 1.5 2 2.5 2 3 4 5 Velocity (μm/s) F-actin length (μm) 6 8 10 12 14 Actin gel length (μm) 9 Polarity of the actin in complexes [24] 9. Polarity of the actin in complexes [24] On a glass surface where myosin molecules are simply immobilized, F-actin is known to show motility. We found that all these polymer-actin complex gels, including those from p-Lys, PDMAPAA-Q, and x,y-ionene bromide polymers (x=3or 6; y=3,4,or10), also exhibit motility and the velocity of the complexes were comparable to that of native F-actin (0.77±0.32μm/s) [24]. The polarity of polymer-actin complexes, which is considered to be essential in the sliding motion of polymer-actin complex gels, has not been completely clarified yet. This question is important for understanding the cooperative motion by the actin assembly. An arrowhead-like pattern can be observed under transmission electron microscopy (TEM) when heavy meromyosin (HMM) was decorated on native F-actin [Figure 10(a)], indicating that F-actin has a well-defined polarity by self- organization. The pointed end of arrowhead and the opposite end of arrowhead are called the p-end and b-end, respectively. To evaluate the polarity of actin complexes, we also attempted to decorate the actin complexes with HMM. Figure 10(a) shows some examples of TEM images of HMM-decorated PDMAPAA-Q-actin complexes and 6,4-ionene-actin complexes. Different from native F-actin that is a single strand, the polymer-actin complexes are bundles that consist of 3–20 filaments. Arrowhead structures within a filament of the bundle pointed in the same direction, although some defects are observed occasionally. However, arrowhead directions of filaments within a bundle are not completely the same. The complex polarity P defined by Eq. (6) was estimated. Int. J. Mol. Sci. 2008, 9 1697 2 1 2 1 / n n n n P + − = 2 1 2 1 / n n n n P + − = 2 1 2 1 / n n n n P + − = (6) Here n1 and n2 are the number of filaments pointed in the opposite directions. Figure 10. (a) Polarity of polymer-actin complexes decorated with HMM. White arrows indicate the direction of arrowhead structures of decorated filaments. (b) Histograms of complex polarity distributions. Cited from ref. [24] with permission. The average polarity of the actin complex is shown in Table 2. As seen in this data, polarity depends on the chemical structure of polycations, and PDMAPAA-Q-actin complexes show the highest polarity as 0.89 (average over 23 samples), while 6,4-actins show the lowest value as 0.42 average over 17 samples). Thus, PDMAPAA-Q prefers to form actin filament bundles having uni-polarity. 9. Polarity of the actin in complexes [24] Among all these polymer actin complexes the PDMAPAA Q actin complex which has the highest The average polarity of the actin complex is shown in Table 2. As seen in this data, polarity depends on the chemical structure of polycations, and PDMAPAA-Q-actin complexes show the highest polarity as 0.89 (average over 23 samples), while 6,4-actins show the lowest value as 0.42 (average over 17 samples). Thus, PDMAPAA-Q prefers to form actin filament bundles having a uni-polarity. The average polarity of the actin complex is shown in Table 2. As seen in this data, polarity depends on the chemical structure of polycations, and PDMAPAA-Q-actin complexes show the highest polarity as 0.89 (average over 23 samples), while 6,4-actins show the lowest value as 0.42 (average over 17 samples). Thus, PDMAPAA-Q prefers to form actin filament bundles having a uni-polarity. Among all these polymer-actin complexes, the PDMAPAA-Q-actin complex, which has the highest polarity, also shows the highest motility, with a velocity of 1.3 μm/s. Dendritic complexes, which are occasionally observed when F-actin is mixed with 6,4- ionene, did not exhibit a translational motion but instead migrate around their barycentric position. When the velocity evaluated at 3.3s intervals is plotted agains the polarity, a linear relationship is observed, i.e. the velocity of the complexes is proportional to the polarity. These results suggest that the polarity of the polymer-actin complex is essential in producing a high motility. It is known that a native F-actin always moves toward one direction without moving back to the opposite direction. The direction of the F-actin motion is associated with its polarity. For the polymer-actin bundles that are assembly of F-actin, the whole Int. J. Mol. Sci. 2008, 9 1698 polarity is determined by the polar direction of F-actin. If the F-actin filaments assemble in an anti- parallel way, the whole polarity is cancelled and leads to no sliding motion. This explains why the correlation between polarity and velocity is observed. However, the difference between the velocity of native F-actin and that of the polymer-actin complex gels with a polarity close to 1 (p-Lys-actin and PDMAPAA-Q-actin) cannot be explained only in terms of polarity. The higher velocity observed for the complex gels should, therefore, be attributed to two possible factors: 1) the arrowhead structure that depends on the structures of polycations, and 2) the bundle formation. 9. Polarity of the actin in complexes [24] If the change of the arrowhead structure is attributed to the change of the helix pitch of an actin filament, an elongation of arrowhead pattern will cause the extensions of helix structure of the actin filament. In consequence, the elasticity of the actin filament might increase and cause the effective dynamic-interaction with myosin which favors the motion. On the other hand, by forming an actin bundle, the bending fluctuation is eliminated, and this also leads to an effective integration of driving-forces from each myosin molecules. To elucidate the randomness of the motion, we also investigated the trajectory of the motion [24]. From this characterization, it is found that all polymer-actin complexes show more translational motion than that of native F-actin. Because polymer-actin complexes are bundles formed from actin filaments, they are less flexible than native F-actin. The less random motion of polymer-actin complexes is attributed to the less structural flexibility. 10. Spatiotemporal control of actin assembly [42, 43] In Sections 7 and 8, the motility system composed of the biomolecular motor (actin/myosin) assemblies was described. To enhance their potential application, introduction of hierarchical structure into the assemblies is inevitable. Within living cells, in order to organize their body with ordered structure, the bio-molecule (actins or other motor proteins) assembly process is spatially controlled by the localization of binding proteins, and also temporally regulated by the coupling of the binding proteins to signal molecules [44-46]. Intrinsic polarity of cells leads the binding proteins and signal molecules to exhibit a strong localization within a cell. It can be demonstrated that under asymmetric conditions by unidirectional diffusion of polycations (PDMAPAA-Q) to F-actin solution F-actins are effectively assembled into a globally linked network even at a low actin concentration that only small actin bundles are formed under homogeneous mixing of F-actins with polycations (Figure 11). This fact indicates that the asymmetric spatial distribution of polycations is essential for the architectural variety of F-actin bundles. Meanwhile, temporal control of the actin assembly formation can be achieved by applying a photoresponsible polycation composed mainly of DMAPAA-Q and partially of photochromic molecules (triphenylmethane leucohydroxide) that can be cationized in aqueous solution by ultraviolet (UV) irradiation. As shown in Figure 12, the photo-induced actin assembly formation is observed following UV irradiation. The localization of UV irradiation by using an optical tip made of a glass capillary enables to control the formation of F-actin/photoresponsive polycation complex spatiotemporally. As shown in the fluorescence micrograph in Figure 13, prior to UV irradiation, F- actin is in its native state. Following UV irradiation through the glass capillary tip, the large aggregated structure of the F-actin/photoresponsive polycation complex is observed only at the tip. Thus, we can remotely trigger and confine the formation of F-actin assemblies to any location of Int. J. Mol. Sci. 2008, 9 1699 Int. J. Mol. Sci. 2008, 9 interest within a complex system. This technique will provide a method to create F-actin assembly with highly ordered structure by 3D drawing system with UV light. Figure 11. (a): Schematic illustration of microchamber for the uni-directional diffusion of polycation to F-actin solution. (b): Time-lapse images of actin network formation after the first 10 min of the uni-directional diffusion of PDMAPAA-Q (CP = 0.1 M) to F- actin solution (CA = 0.01 mg/mL in microchamber. 10. Spatiotemporal control of actin assembly [42, 43] (c): Actin bundles formed by homogeneous mixing of F-actin solution (CA = 0.01mg/mL) with PDMAPAA-Q solution (CP = 0.1 M) and then incubating for 60 min. (d, e): Schematics of actin network formation under uni-directional diffusion of polycations (d), actin bundles formation under homogeneous mixing with polycations (e). Black rods and the intensity of gray background represent F-actins and polycation concentration, respectively. Scale bar presents 50μm. Cited from ref. [42] with permission. 11. Conclusions 11. Conclusions Biopolymers including F-actin have a lot of potential as functional materials in biotechnology and biomaterials science. It has been revealed that actin and myosin vary their functionality depending on the orientation within the complexes. In the case of myosin, it has been found to have self-assembly ability. On the other hand, actin has shown an ability to form a functional complex with polycations by electrostatic interaction. Moreover, the cross-linked actin complex gels also have shown motility on the myosin gels. The motility of actin complex gels depends on the structure that is determined by the conditions such as the concentrations or the structures of the polycations. This may also be applicable Int. J. Mol. Sci. 2008, 9 1700 Int. J. Mol. Sci. 2008, 9 for the other functional biopolymers. To develop more functional biomaterials hereafter, how to orientate the functional domain or the polarity, may become the most significant issue. The basic analysis and technical development to form and stabilizing the biopolymers by cross-linking may contribute to promote the possibility of the biopolymers as functional materials in biotechnology and biomaterials science. Figure 12. Fluorescent micrographs of the photo-induced bundle formation of F- actin/photoresponsive polycation complex, where [F-actin]/[photoresponsive polycation] is fixed at 1.0 g/g, and [F-actin] = 5×10-3 mg/mL. The scale bars are 10 μm. Figure 13. Photo-induced local formations of actin complex at the UV photon emitting tip-edge grass capillary connected to the optical fiber. In the brightfield image of fluorescent microscope, the block object is the tip-edge of grass capillary. The scale bar is 100 μm. The observation was performed before and after the local UV irradiation in 30 min. Figure 13. Photo-induced local formations of actin complex at the UV photon emitting tip-edge grass capillary connected to the optical fiber. In the brightfield image of fluorescent microscope, the block object is the tip-edge of grass capillary. The scale bar is 100 μm. The observation was performed before and after the local UV irradiation in 30 min. On the other hand, many studies on the mechanism of muscle contraction have been performed both on macroscopic level using the physiological technique and on molecular level using optical microscope, optical tweezers, and etc. 11. Conclusions However, the approach from a mesoscopic level, from between On the other hand, many studies on the mechanism of muscle contraction have been performed both on macroscopic level using the physiological technique and on molecular level using optical microscope, optical tweezers, and etc. However, the approach from a mesoscopic level, from between Int. J. Mol. Sci. 2008, 9 Int. J. Mol. Sci. 2008, 9 1701 micro-level and macro-level, has not been studied completely. New architectures made from muscle protein will also provide a mesoscopic insight into the mechanism of muscle contraction including the role of hierarchical structure and the mechanism of cooperative sliding motion micro-level and macro-level, has not been studied completely. New architectures made from muscle protein will also provide a mesoscopic insight into the mechanism of muscle contraction including the role of hierarchical structure and the mechanism of cooperative sliding motion References Gregg, B.A.; Heller, A. Cross-Linked Redox Gels Containing Glucose Oxidase for Amperometric Biosensor Applications. Anal. Chem. 1990, 62, 258-263. 22. Kakugo, A.; Sugimoto, S.; Gong, J.P.; Osada, Y. Gel machines constructed from chemically cross-linked actins and myosins. Adv. Mater. 2002, 14, 1124-1126. 23. Kakugo, A.; Sugimoto, S.; Shikinaka, K.; Gong, J.P.; Osada, Y. Characteristics of chemically cross-linked myosin gels. J. Biomater. Sci. Polymer Ed. 2005, 16, 203-218. 24. Kakugo, A.; Shikinaka, K.; Takekawa, N.; Sugimoto, S.; Osada, Y.; Gong, J.P. Polarity and motility of large polymer-actin complexes. Biomacromolecules, 2005, 6, 845-849. 25. Patolsky, F.; Weizmann, Y.; Willner, I. Actin-based metallic nanowires as bio-nanotransporters. Nat. Mater. 2004, 3, 692-695. 26. Uyeda, T.Q.P.; Warrich, H.M.; Kron, S.J.; Spudich, J.A. Quantized velocities at low myosin densities in an in vitro motility assay. Nature 1991, 352, 307-311. 27. Huxley, H.E. Sliding filaments and molecular motile systems. J. Biol. Chem. 1990, 265, 8347-8350. 28. Kron, S.J.; Spudich, J.A. Fluorescent actin filaments move on myosin fixed to a glass surface. Proc. Natl. Acad. Sci. USA 1986, 83, 6272-6276. 29. Vale, R.D.; Reese, T.S.; Sheetz, M.P. Identification of a novel force-generating protein, kinesin, involved in microtubule-based motility. Cell 1985, 42, 39-50. 30. Noji, H.; Yasuda, R.; Yoshida, M.; Kinosita, K. Jr. Direct observation of the rotation of F1- ATPase. Nature 1997, 386, 299-302. 31. Suzuki, H.; Oiwa, K.; Yamada, A.; Sakakibara, H.; Nakayama, H.; Mashiko, S. Linear arrangement of motor protein on a mechanically deposited fluorepolymer thin film. Jpn. J. Appl. Phys. 1995, 34, 3937-3941. 32. Kakugo, A.; Shikinaka, K.; Matsumoto, K.; Gong, J.P.; Osada, Y. Growth of large polymer-actin complexes. Bioconjugate Chem. 2003, 14, 1185-1190. 33. Lodish, H.; Berk, A.; Zipursky, S.L.; Matsudaira, P.; Baltimore, D.; Darnell, J.E. Molecular Cell Biology, 4th Ed.; Freeman: San Francisco, CA, 1999. 34. Tang, J.X.; Janmey, P.A. J. Biol. Chem. 1996, 271, 8556-8563. 35. Tang, J.X.; Szymanski, P.; Janmey, P.A.; Tao, T. Eur. J. Biochem. 1997, 247, 43 36. Amann, K.J.; Renley, B.A.; Ervasti, J.M. J. Biol. Chem. 1998, 273, 28419. 37. Kwon, H.J.; Kakugo, A.; Shikinaka, K.; Osada, Y.; Gong, J.P. Morphology of Actin Assemblies in Response to Polycation and Salts. Biomacromolecules 2005, 6, 3005-3009. 38. Meyer, R.K.; Aebi, U. Bundling of actin filaments by alpha-actinin depends on its molecular length. J. Cell Biol. 1990, 110, 2013-2024. g 39. Landau, L.D.; Lifshitz. E.M. Theory of Elasticity, 3rd Ed.; Pergamon: Oxford, UK, 1986. 40. References 1. Huxley, H.E. The mechanism of muscular contraction. Science 1969, 164, 1356-136 2. Huxley, A.F. Reflections on muscle. Liverpool University Press: Liverpool, 1980. 3. Pollack, G.H. Cells, gels and the engines of life. Ebner and Sons Publishers: Seattle, WA, 2001. 4. Kitamura, K.; Tokunaga, M.; Iwane, A.H.; Yanagida, T. A single myosin head moves along an actin filament with regular steps of 5.3 nanometres. Nature 1999, 397, 129-134. 5. Steinberg, I.Z.; Oplatka, A.; Katchalsky, A. Mechanochemical engines. Nature 1966, 210, 568-571. 5. Steinberg, I.Z.; Oplatka, A.; Katchalsky, A. Mechanochemical engines. Nature 1966, 210, 568-571. 6. Sussman, M.V.; Katchalsky, A. Mechanochemical turbine: A new power cycle. Science 1970, 167, 45-47. 7. Osada, Y.; Okuzaki, H.; Hori, H. A polymer gel with electrically driven motility. Nature 1992, 355, 242-244. 8. Mitsumata, T.; Ikeda, K.; Gong, J.P.; Osada, Y. Solvent-driven chemical motor. Appl. Phys. Lett. 1998, 73, 2366-2368. 9. Osada, Y.; Matsuda, A. Shape memory in hydrogels. Nature 1995, 376, 219-220. 10. Osada, Y.; Ross-Murphy, S.B. Intelligent gels. Sci. Am. 1993, 268, 82-87. 10. Osada, Y.; Ross-Murphy, S.B. Intelligent gels. Sci. Am. 1993, 268, 82-87. 11. Jager, E.W.H.; Smela, E.; Inganas, O. Microfabricating conjugated polymer actuators. Science 2000, 290, 1540-1546. 12. Beebe, D.J.; Moore, J.S.; Bauer, J.M.; Yu, Q.; Liu, R.H.; Devadoss, C.; Jo, B.H. Functional hydrogel structures for autonomous flow control inside microfluidic channels. Nature 2000, 404, 588-590. 13. Quake, S.R.; Scherer, A. From micro- to nanofabrication with soft materials. Science 2000, 290, 1536-1540. 14. Haines, L.A.; Rajagopal, K.; Ozbas, B.; Salick, D.A.; Pochan, D.J.; Schneider, J.P. Light- activated hydrogel formation via the triggered folding and self-assembly of a designed peptide. J. Am. Chem. Soc. 2005, 127, 17025-17029. Rajagopal, K.; Ozbas, B.; Pochan, D.J.; Schneider, J.P. Biopolymers 2005, 80, 487-487 16. Nowak, A.P.; Breedveld, V.; Pakstis, L.; Ozbas, B.; Pine, D.J.; Pochan, D.; Deming, T.J. Rapidly recovering hydrogel scaffolds from self-assembling diblock copolypeptide amphiphiles. Nature 2002, 417, 424-428. 17. Zohuriaan-Mehr, M.J.; Motazedi, Z.; Kabiri, K.; Ershad-Langroudi, A. New Superabsorbing Hydrogel Hybrid from Arabic Gum and Acrylic Monomers. J. Macromol. Sci., Part A: Pure Appl. Chem. 2005, A42, 1655-1666. 18. Balzani, V.; Credi, A.; Venturi, M. Molecular Devices and Machines; Wiley-VCH: Weinheim, 2003. 19. Stoddart, J.F. Molecular machines. Acc. Chem. Res. 2001, 34, 410-411. 20. Koyama, T.; Yamaguchi, N.; Miyasaka, T. Antibody-mediated bacteriorhodopsin orientation for molecular device architectures. Science 1994, 265, 762-765. Int. J. Mol. Sci. 2008, 9 1702 21. References Kwon, H.J.; Tanaka, Y.; Kakugo, A.; Shikinaka, K.; Furukawa, H.; Osada, Y.; Gong, J.P. Anisotropic nucleation-growth of actin bundle: A model for determining the well defined thickness of bundles. Biochemistry 2006, 45, 10313-10318. 41. Shikinaka, K.; Kwon, H.J.; Kakugo, A.; Furukawa, H.; Osada, Y.; Gong, J.P.; Aoyama, Y.; Nishioka, H.; Jinnai, H.; Okajima, T. Observation of the Three-Dimensional Structure of Actin Bundles formed with Polycations. Biomacromolecles 2008, 9, 537-542. Int. J. Mol. Sci. 2008, 9 1703 42. Kwon, H.J.; Kakugo, A.; Ura, T.; Okajima, T.; Tanaka, Y.; Furukawa, H.; Osada, Y.; Gong, J.P. Actin network formation by uni-directional polycation diffusion. Langmuir 2007, 23, 6257-6262. 43. Misu, M.; Furukawa, H.; Kwon, H.J.; Shikinaka, K.; Kakugo, A.; Satoh, T.; Osada, Y.; Gong, J.P. Photoinduced in situ formation of various F-actin assemblies with a photoresponsive polycation. J. Biomed. Mater. Res. Part A in press. 44. Chung, C.Y.; Funamoto, S.; Firtel, R. Signaling pathways controlling cell polarity and chemotaxis. Trends Biochem. Sci. 2001, 26, 557-566. 45. Kurokawa, K.; Nakamura, T.; Aoki, K.; Matsuda, M. Mechanism and role of localized activation of Rho-family GTPases in growth factor-stimulated fibroblasts and neuronal cells. Biochem. Soc. Trans. 2005, 33, 631-634. 46. Svitkina, T.; Borisy, G. Arp2/3 Complex and Actin Depolymerizing Factor/Cofilin in Dendritic Organization and Treadmilling of Actin Filament Array in Lamellipodia. J. Cell. Biol. 1999, 145, 1009-1026. © 2008 by the authors; licensee Molecular Diversity Preservation International, Basel, Switzerland. This article is an open-access article distributed under the terms and conditions of the Creative Commons Attribution license (http://creativecommons.org/licenses/by/3.0/). © 2008 by the authors; licensee Molecular Diversity Preservation International, Basel, Switzerland. This article is an open-access article distributed under the terms and conditions of the Creative Commons Attribution license (http://creativecommons.org/licenses/by/3.0/).
https://openalex.org/W3209748921	https://www.medrxiv.org/content/medrxiv/early/2021/10/26/2021.10.20.21264950.full.pdf	English	null	Genome-wide rare variant score associates with morphological subtypes of autism spectrum disorder	medRxiv (Cold Spring Harbor Laboratory)	2,021	cc-by	18,888	Affiliations 1 1The Centre for Applied Genomics, Genetics and Genome Biology, The Hospital for Sick Children, Toronto, ON, Canada; 1The Centre for Applied Genomics, Genetics and Genome Biology, The Hospital for Sick Children, Toronto, ON, Canada; 2Genetics and Genome Biology, The Hospital for Sick Children, Toronto, ON, Canada 3Ted Rogers Centre for Heart Research, The Hospital for Sick Children, Toronto, ON, Canada; ON, Canada; 4Provincial Medical Genetics Program, Eastern Health, St. John’s, NL, Canada; 4Provincial Medical Genetics Program, Eastern Health, St. John’s, NL, Canada; 5Department of Psychiatry, The Hospital for Sick Children, Toronto, ON, Canada; 4Provincial Medical Genetics Program, Eastern Health, St. John’s, NL, Canada; 5Department of Psychiatry, The Hospital for Sick Children, Toronto, ON, Canada; 6Department of Psychiatry, University of Toronto, Toronto, ON, Canada; 7D f M l l G i U i i f T T ON C d Provincial Medical Genetics Program, Eastern Health, St. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: edRxiv preprint of autism spectrum disorder Ada J.S. Chan1,2, Worrawat Engchuan1,2, Miriam S. Reuter3, Zhuozhi Wang1,2, Bhooma Thiruvahindrapuram1,2, Brett Trost1,2, Thomas Nalpathamkalam1,2, Carol Negrijn4, Sylvia Lamoureux1,2, Giovanna Pellecchia1,2, Rohan Patel1,2, Wilson W.L. Sung1,2, Jeffrey R. MacDonald1,2, Jennifer L. Howe1,2, Jacob Vorstman1,5,6, Neal Sondheimer7-9, Nicole Takahashi10, Judith H. Miles10, Evdokia Anagnostou9,11, Kristiina Tammimies12, Mehdi Zarrei1,2, Daniele Merico1,13, Dimitri J. Stavropoulos14,15, Ryan K.C. Yuen1,2,7, Bridget A. Fernandez4,16,17, Stephen W. Scherer1,2,7,18 Affiliations 1 John s, NL, Canada; 5Department of Psychiatry, The Hospital for Sick Children, Toronto, ON, Canada; 6Department of Psychiatry, University of Toronto, Toronto, ON, Canada; Department of Psychiatry, The Hospital for Sick Children, Toronto, ON, Canada; 6Department of Psychiatry, University of Toronto, Toronto, ON, Canada; 7Department of Molecular Genetics, University of Toronto, Toronto, ON, Canada; 8Division of Clinical and Metabolic Genetics, Department of Pediatrics, The Department of Molecular Genetics, University of Toronto, Toronto, ON, Canada; 8Division of Clinical and Metabolic Genetics, Department of Pediatrics, The Hospital for Sick Children, Toronto, ON, Canada; p 9Department of Pediatrics, University of Toronto, Toronto, ON, Canada; 10 p y 10Thompson Center for Autism and Neurodevelopmental Disorders, University of Missouri, Columbia, MO, USA; 11H ll d Bl i Kid R h bili i H i l 10Thompson Center for Autism and Neurodevelopmental Disorders, Univ Thompson Center for Autism and Neurodevelopmental Disorders, University of Missouri, Columbia, MO, USA; 11H ll d Bl i Kid R h bilit ti H it l Missouri, Columbia, MO, USA; 11Holland Bloorview Kids Rehabilitation Hospital; 12Department of Women's and Children's Health, Karolinska Institutet, 12Department of Women's and Children's Health, Karolinska Institutet, Stockholm Sweden; 12Department of Women s and Children s Health, Karolinska Institutet, Stockholm, Sweden; Stockholm, Sweden; 13Deep Genomics Inc., Toronto, ON, Canada; p 14Department of Paediatric Laboratory Medicine, Genome Diagnostics, The Hospital for Sick Children Toronto ON Canada; 14Department of Paediatric Laboratory Medicine, Genome Diagnostics, Th Department of Paediatric Laboratory Medicine, Ge Hospital for Sick Children, Toronto, ON Canada; Hospital for Sick Children, Toronto, ON Canada; Hospital for Sick Children, Toronto, ON Canada; 15Department of Paediatric Laboratory Medicine, Genome Diagnostics, Hospital for Sick Children, Toronto, Ontario, Canada; 16Department of Pediatrics and The Saban Research Institute, Children's H it l L A l K k S h l f M di i f U i it f S th p 15Department of Paediatric Laboratory Medicine, Genome Diagnostics, Hospital for Sick Children, Toronto, Ontario, Canada; 16Department of Pediatrics and The Saban Research Institute, Children's 16Department of Pediatrics and The Saban Research Institute, Children's Hospital Los Angeles, Keck School of Medicine of University of Southern California, Los Angeles CA USA g 17Discipline of Genetics, Faculty of Medicine, Memorial University of g 17Discipline of Genetics, Faculty of Medicine, Memorial University of Newfoundland, St. John's NL Canada 17Discipline of Genetics, Faculty of Medicine, Memorial University of Newfoundland, St. 1 NOTE: This preprint reports new research that has not been certified by peer review and should not be used to guide clinical practice. Affiliations 1 ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: edRxiv preprint Abstract Defining different genetic subtypes of autism spectrum disorder (ASD) can enable the prediction of developmental outcomes. Based on minor physical and major congenital anomalies, we categorized 325 Canadian children with ASD into dysmorphic and nondysmorphic subgroups. We developed a method for calculating a patient-level, genome-wide rare variant score (GRVS) from whole- genome sequencing (WGS) data. GRVS is a sum of the number of variants in morphology-associated coding and non-coding regions, weighted by their effect sizes. Probands with dysmorphic ASD had a significantly higher GRVS compared to those with nondysmorphic ASD (P= 0.027). Using the polygenic transmission disequilibrium test, we observed an over-transmission of ASD- associated common variants in nondysmorphic ASD probands (P= 2.9×10-3). These findings replicated using WGS data from 442 ASD probands with accompanying morphology data from the Simons Simplex Collection. Our results provide support for an alternative genomic classification of ASD subgroups using morphology data, which may inform intervention protocols. Affiliations 1 John's NL Canada 18McLaughin Centre, University of Toronto, Toronto, ON, Canada 18McLaughin Centre, University of Toronto, Toronto, ON, Canada Co-correspondence: 1 NOTE: This preprint reports new research that has not been certified by peer review and should not be used to guide clinical practice. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: edRxiv preprint Stephen W. Scherer (stephen.scherer@sickkids.ca, 416-813-7613, The Centre for Applied Genomics, Peter Gilgan Centre for Research and Learning, 686 Bay Street Room 13.9705 Toronto, Ontario, M5G 0A4, Canada) Bridget A. Fernandez (bfernandez@chla.usc.edu, 323-361-1007, Department of Pediatrics and The Saban Research Institute, Children's Hospital Los Angeles, 4650 Sunset Blvd MS #001, Los Angeles CA USA) Stephen W. Scherer (stephen.scherer@sickkids.ca, 416-813-7613, The Centre for Applied Genomics, Peter Gilgan Centre for Research and Learning, 686 Bay Street Room 13.9705 Toronto, Ontario, M5G 0A4, Canada) Bridget A. Fernandez (bfernandez@chla.usc.edu, 323-361-1007, Department of Pediatrics and The Saban Research Institute, Children's Hospital Los Angeles, 4650 Sunset Blvd MS #001, Los Angeles CA USA) Stephen W. Scherer (stephen.scherer@sickkids.ca, 416-813-7613, The Centre for Applied Genomics, Peter Gilgan Centre for Research and Learning, 686 Bay Street Room 13.9705 Toronto, Ontario, M5G 0A4, Canada) Bridget A. Fernandez (bfernandez@chla.usc.edu, 323-361-1007, Department of Pediatrics and The Saban Research Institute, Children's Hospital Los Angeles, 4650 Sunset Blvd MS #001, Los Angeles CA USA) 2 2 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. Main ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint Here, from two independent cohorts, we used whole-genome sequences (WGS) and detailed clinical morphology data to: 1) develop a genome-wide rare variant score (GRVS) to measure the relationship between rare variants and morphology, and 2) examine the contribution of rare and common variants in morphological ASD subtypes (Figure 1 and Supplementary figure 1). For our discovery cohort, we used a population-based sample of 325 unrelated children with Autism Diagnostic Observation Scale (ADOS)-confirmed ASD. Following clinical examination, a total morphology score was assigned to each case based on the number of MPAs and MCAs9,10. The cohort was then stratified into three subtypes of increasing morphologic severity: 187 essential ASD (57.5%), 57 equivocal ASD (17.5%) and 81 complex ASD (24.9%) (Supplementary Table 1). We further stratified these samples into two subtypes by combining complex and equivocal ASD into a single dysmorphic ASD grouping and redefining essential ASD as nondysmorphic ASD. We performed WGS on 795 genomes (325 probands and 470 parents) and detected all classes of variation (SNV, indel, CNV and structural variants (SVs) (Figure 1, Supplementary Table 2-4). Using the American College of Medical Genetics and Genomics guidelines16,17, we identified a total of 46 clinically significant variants (CSVs) in 46 of 325 probands (14.1%) (Supplementary Tables 5-7). The proportion of dysmorphic ASD cases with a CSV (25.9%; 35/135) was significantly higher than nondysmorphic ASD (5.8%, 11/190) (P= 3.2×10-7, one-sided Fisher’s test), consistent with our previous findings10. We also identified 29 variants of uncertain significance (VUS) in 26 probands that were of interest, including tandem repeat expansions in previously reported ASD candidate loci18; three probands each had two VUS (Supplementary Tables 5-7 and Supplementary Note). To further investigate the contribution of rare variants among morphological ASD subtypes, we first conducted a rare variant burden analysis and multiple test correction using the Benjamini Hochberg approach (BH-FDR) (see methods). We found a significantly higher prevalence of rare coding deletions >10kb in probands with more dysmorphic features (P= 5.00 × 10-4 and BH-FDR= 5.00 × 10-3, Supplementary Table 8). Rare coding duplications >10kb and ≤10kb, genic deletions ≤10kb, loss-of-function (LoF) and missense variants were not significantly different among subtypes (Supplementary Table 8). We then performed enrichment and burden analyses to identify gene sets or noncoding regions, respectively, that were differentially affected by rare or de novo variants between the morphological ASD subtypes. Main Autism spectrum disorder (ASD), which is diagnosed on the basis of behavioral assessments that reveal social communication deficits and repetitive behaviors, is often associated with traits including major congenital anomalies (MCAs), minor physical anomalies (MPAs)1,2 and intellectual disability3-5. Increasingly, penetrant variants of diagnostic value6,7 and lesser impact common variants are being implicated in the etiology of ASD4,8. Autistic individuals who are more dysmorphic (complex ASD) tend to have lower intelligence quotients (IQ) and more brain and other major congenital anomalies9,10 compared with those who are less dysmorphic (essential ASD), leading to poorer developmental outcomes. Individuals with complex ASD are also less likely to have a family history of ASD, suggesting that morphological subtypes can reveal informative genetic differences among ASD subgroups9. Genetic liability to ASD can be quantified using a polygenic risk score (PRS), which is a weighted sum of ASD-associated common variants, using effect sizes drawn from genome-wide association studies11. A similar score for rare variants remains to be established. Rare variant studies use burden analyses to compare the frequency of rare variants, equally weighted, between cases and controls or among ASD subtypes3,12,13. Quadratic tests have also been used in rare variant association tests and typically weigh variants by minor allele frequency14,15. However effect sizes depend on the affected gene and variant type and these However, effect sizes depend on the affected gene and variant type, and these variables should be considered in rare variant analyses. 3 3 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: edRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. Main The 67 gene sets and noncoding regions studied have been previously associated with ASD13,19-23. After multiple-testing correction (permutation-based false discovery rate (FDR) <20%), 20 significant gene sets or noncoding regions were identified (Supplementary Tables 9 and 10). We observed that probands with dysmorphic features had higher burdens of deletions and missense variants impacting genes 4 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: edRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint responsible for various neuronal functions and duplications >10kb impacting brain-expressed genes (Figure 2 and Supplementary Table 9). Dysmorphic probands also had a significantly higher prevalence of rare deletions ≤10kb overlapping promoters of long noncoding genes and duplications (larger and smaller than 10kb) overlapping active brain enhancers (Supplementary Figure 2, Supplementary Table 10). We then tested the collective contribution of rare variants in morphology- associated regions, while considering the effect size of each variant, which varies depending on the variant type and morphology-associated region. We developed a GRVS for each proband, which is a weighted sum of the number of rare variants in morphology-associated regions identified from gene set enrichment and noncoding burden tests (Supplementary Table 11). We weighed the number of rare variants in each morphology-associated region as well as the variant type (i.e., coding or noncoding deletions and duplications >10kb or ≤10kb, loss-of- function variants, missense variants, and noncoding SNVs and indels) using the coefficients from logistic regression models. To calculate GRVSs for each proband in the discovery cohort, we used a 10-fold cross validation strategy to reduce over-fitting (Supplementary Figure 1a). Main These findings suggest that variants in morphology-associated regions that are not CSVs also significantly contribute to morphological outcomes in ASD. probands with CSVs, those with dysmorphic ASD still had significantly higher average GRVSs than those with nondysmorphic ASD (P= 0.048, one-sided Wilcoxon rank sum test, Figure 3b). These findings suggest that variants in morphology-associated regions that are not CSVs also significantly contribute to morphological outcomes in ASD. To explore the contribution of common (minor allele frequency >0.05) ASD- associated variants in ASD subtypes, we calculated polygenic risk scores (PRS) for ASD and body mass index (BMI)8 (Online Methods and Supplementary Table 12). We then compared these scores across the morphologic groups using the polygenic transmission disequilibrium test (pTDT)8, which compares the PRS of the proband to parents’ mean PRS. We found a significant over-transmission of common ASD-associated variants in probands with nondysmorphic ASD (P= 2.9×10-3, one-sided t-test) and no significant over-transmission in probands with dysmorphic ASD (P= 0.3) (Figure 4). PRS for BMI was selected as a negative control because there is no genetic correlation between BMI and ASD24, and we did not find over-transmission of PRS for BMI in either subtype (Figure 4). IQ is often negatively correlated with the burden of rare variants3,4,13,25,26. We therefore examined our probands with dysmorphic ASD and determined they had a significantly lower mean IQ compared to nondysmorphic ASD (P= 0.013, one- sided t-test, Figure 5a and Supplementary Table 12). Probands with a CSV had significantly lower IQ compared to probands without a CSV (P= 2.2 ×10-4, one- sided t-test, Figure 5b). However, IQ was not significantly correlated with GRVS (rho= -0.042, P= 0.64, Figure 5c) or PRS (rho= -0.15, P= 0.12, Figure 5d). We repeated our analysis on a replication cohort of relevant samples from the Simons Simplex Collection (442 ADOS-confirmed affected probands and 355 unaffected siblings)27. The affected probands had been categorized into two morphological subtypes (400 nondysmorphic and 42 dysmorphic cases)27 using the Autism Dysmorphology Measure (ADM)28. In contrast to the discovery cohort, the SSC probands were classified by targeted physical examinations performed by individuals without expert training in dysmorphogy, and the classification did not incorporate the presence or absence of major congenital anomalies. To compare the two cohorts, we reclassified a subset of the original discovery cohort based on minor anomalies alone using the ADM algorithm (203 nondysmorphic and 73 dysmorphic cases, Online Methods). Main We used Nagelkerke’s R2 to determine the optimal P value threshold (P<0.1) to identify morphology-associated regions (Supplementary Figure 3a and Online Methods). GRVS can be calculated for probands regardless of whether their parents have been sequenced. However, there would be a systematic difference in GRVSs in this cohort if all probands were used because those probands whose parents have been sequenced would include scores from de novo variants, whereas those without sequenced parents would not have de novo variant scores. To avoid this, GRVS was calculated only for probands with two sequenced parents (n= 235) (Figure 3a and Supplementary Table 12). Probands with dysmorphic ASD had significantly higher average GRVSs than those with nondysmorphic ASD (P= 0.027, one-sided Wilcoxon rank sum test) (Figure 3b). Most probands (95.7%, 225/235) had more than one variant impacting morphology-associated regions (Supplementary Table 12). Rare coding CNVs had the highest effect size; rare noncoding SNVs and indels had the lowest (Figure 3c and Supplementary Table 11). Using the GRVS formula, we calculated a score for CSVs that overlapped an ASD relevant, morphology-associated region (so that effect size was available for calculation) and that occurred in probands with sequencing data from both parents. 17 of the 46 CSVs met these criteria. No score was calculated for the remaining 29 variants because 15 were identified in probands where both parents were not available for sequencing, and 14 variants were not located in or encompassed by one of the 20 morphology-associated regions. (Online Methods). In 47% of samples with CSV scores (8/17 probands, Supplementary Table 13), CSVs contributed >50% of the total GRVS. When we excluded the 5 5 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: edRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint probands with CSVs, those with dysmorphic ASD still had significantly higher average GRVSs than those with nondysmorphic ASD (P= 0.048, one-sided Wilcoxon rank sum test, Figure 3b). Main In both cohorts, probands with ADM-defined dysmorphic ASD had significantly higher GRVSs (Pdiscovery= 3.6 ×10-6 and Preplication= 2.7 ×10-4, one-sided Wilcoxon rank sum test, Figure 6a) and yield of CSVs (Pdiscovery= 2.7 ×10-7 and Preplication= 2.1 ×10-3, one-sided Wilcoxon rank sum test, Figure 6b and Supplementary Tables 17 and 18) compared to ADM-defined nondysmorphic ASD, consistent with our findings using the gold-standard dysmorphology classification. In the replication cohort, unaffected siblings had a significantly lower GRVS compared to ADM-defined dysmorphic ASD (P= 7.7 ×10-4 one-sided Wilcoxon rank sum test) but did not have a significantly lower GRVS compared to ADM-defined nondysmorphic ASD (Figure 6a). Furthermore, unaffected siblings of nondysmorphic probands did not have a significantly lower GRVS compared unaffected siblings of dysmorphic probands (P= 0.75, one-sided Wilcoxon rank sum test, data not shown). In both cohorts we also found a significant over-transmission of common ASD- associated SNPs in ADM-defined nondysmorphic ASD (Pdiscovery= 6.7×10-3 and Preplication= 6.3 ×10-3, one-sided Wilcoxon rank sum test, Figure 6c). In results similar to Weiner et al.8, we did not observe over-transmission in unaffected siblings in the replication cohort (P= 0.88, one-sided Wilcoxon rank sum test). Individuals with ADM-defined dysmorphic ASD or with CSVs had a significantly lower IQ compared to ADM-defined nondysmorphic ASD or those without CSVs, respectively (Supplementary Figure 4a and b, Supplementary Tables 12, 14-16). Although there was no correlation between IQ and GRVS in the discovery cohort when the subtype classification was done by either gold standard dysmorphology examination (rho= -0.042, P= 0.64, Figure 5c) or using the ADM (rho= 0.12, P= 0.21, Supplementary Figure 4c), a significant negative correlation was found in the replication cohort (rho= -0.14, P= 4.4×10-3, Supplementary Figure 4c). We did not find significant correlations between IQ and PRS (Figure 5d and Supplementary Figure 4d), or PRS and GRVS in either cohort (Supplementary Figure 5 and Supplementary Tables 12 and 16). Differences in the correlation between GRVS and IQ between the cohorts might be attributable to differences in ascertainment. The discovery cohort was assembled using a population-based recruitment strategy, and the average IQ of the cohort is 105 similar to the population average of 100. In contrast, individuals with comorbid ID or low IQ are found in SSC29, consistent with the replication cohort having a significantly lower IQ compared to the discovery cohort (mean IQdiscovery= 105±23, mean IQreplication= 82±27, P= 1.1 × 10-21, two-sided t-test). Main We calculated new GRVSs for the ADM-reclassified discovery cohort using a 10-fold cross-validation approach (143 nondysmorphic and 48 dysmorphic cases met criteria for inclusion in this analysis, Supplementary Figure 1a, Supplementary Tables 12 and 14, and Online Methods). We used Nagelkerke’s R2 to determine the optimal P-value threshold and identified 32 morphology-associated regions, which largely overlapped with our original analysis (Supplementary Figure 3b). The morphology-associated regions (P< 0.1, Supplementary Table 15) identified in the reclassified discovery cohort were used to calculate GRVSs for the 6 6 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: edRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint replication cohort (Supplementary Figure 1b, Supplementary Table 16, and Online Methods). In both cohorts, probands with ADM-defined dysmorphic ASD had significantly higher GRVSs (Pdiscovery= 3.6 ×10-6 and Preplication= 2.7 ×10-4, one-sided Wilcoxon rank sum test, Figure 6a) and yield of CSVs (Pdiscovery= 2.7 ×10-7 and Preplication= 2.1 ×10-3, one-sided Wilcoxon rank sum test, Figure 6b and Supplementary Tables 17 and 18) compared to ADM-defined nondysmorphic ASD, consistent with our findings using the gold-standard dysmorphology classification. In the replication cohort, unaffected siblings had a significantly lower GRVS compared to ADM-defined dysmorphic ASD (P= 7.7 ×10-4 one-sided Wilcoxon rank sum test) but did not have a significantly lower GRVS compared to ADM-defined nondysmorphic ASD (Figure 6a). Furthermore, unaffected siblings of nondysmorphic probands did not have a significantly lower GRVS compared unaffected siblings of dysmorphic probands (P= 0.75, one-sided Wilcoxon rank sum test, data not shown). Main Inconsistent findings between ASD cohorts have also been observed when examining gender differences in IQ30, where findings from cohorts with specific selection criteria (e.g., simplex families) may not be generalizable to the ASD population. 7 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: edRxiv preprint Our data suggest that while both dysmorphic and nondysmorphic ASD demonstrate over-transmission of common ASD-associated variants, there is a significantly higher burden of rare variants in dysmorphic ASD than nondysmorphic ASD. GRVS methods may add further specificity to identifying clinically informative endophenotypes but exquisitely phenotyped cohorts will be required. While dysmorpholgy classification by expert clinical examination is not highly scalable, the use of automated tools for 2 and 3-dimensional imaging31 may make it feasible to perform high throughput dysmorphology classification. This will allow GRVSs to be more widely used, potentially in combination with one or more early clinical biomarkers. Online Methods Subjects and Methods j Subject enrolment – Discovery Cohort j Subject enrolment – Discovery Cohort j Subject enrolment – Discovery Cohort CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: edRxiv preprint ventricular septal defect). A single experienced dysmorphologist (B.A.F.) performed a detailed morphological examination of the child and (if possible) parents documenting minor physical anomalies (MPAs), height, weight, head circumference and anthropometric measurements of the head, face, hands, and feet. As described by Miles, et al.9, each proband was assigned an MPA score; one point was given for each embryologically unrelated MPA or for each measurement greater than two standard deviations above or below the population mean and that was absent from the parents if they were available for examination. Each child was also assigned a major congenital anomaly (MCA) score (two points were given for each MCA), and a total morphology score (MPA + MCA scores). Using the total morphology score, we classified each child into essential (total morphology score 0-3), equivocal (total morphology score 4-5) or complex (total morphology score ≥6) groups. We used the final classification for comparing the yield of CSVs and for performing the rare and common variant analyses. Autism Dysmorphology Measure – Discovery and Replication Cohorts Our replication cohort consisted of a subset of samples from the Simons Simplex Collection27. This subset of samples had already been categorized into two morphological groups (400 nondysmorphic and 42 dysmorphic cases) by multiple non-geneticist examiners using the Autism Dysmorphology Measure28. In brief, the Autism Dysmorphology Measure is a decision tree-based classifier that assigns cases into nondysmorphic and dysmorphic groups based the presence or absence of minor physical anomalies of 12 body areas. It was designed to be used by clinicians who do not have expert training in dysmorphology and the assessment is limited to the craniofacies, hands and feet of the child. The ADM decision tree was trained on expert-derived consensus classification of 222 ASD cases who had gold standard examinations of all body areas by clinical geneticists with expertise in dysmorphology9,37. The latter was the approach we used for the initial morphologic classification of our discovery cohort into essential, equivocal and complex groups9. j Subject enrolment – Discovery Cohort j y The cohort consists of children residing in the Canadian province of Newfoundland and Labrador, recruited from one of three developmental team assessment clinics between 2010 and 2018. Assessment through one of these clinics was required for a child with ASD to qualify for provincially funded home Applied Behavioural Analysis (ABA) therapy. Families were invited to participate after their child received an ASD diagnosis from the multidisciplinary team which was led by a developmental paediatrician. Probands met ASD criteria according to the Diagnostic and Statistical Manual of Mental Disorders (Fourth or Fifth Edition, Text Revision)32-34 and all diagnoses were confirmed by an Autism Diagnostic Observation Schedule35 assessment. Most probands also had an Autism Diagnostic Interview-Revised36 assessment consistent with ASD. Children were not excluded from the study based on syndromic features or the presence of a known syndrome. Parents or guardians of the children provided written informed consent. The study was approved by Newfoundland’s Health Research Ethics Boards (HREB# 2003.027) and SickKids Research Ethics Board (REB#0019980189). Subject enrolment – Replication Cohort The replication cohort consisted of a subset of samples from the Simons Simplex Collection, including 442 affected probands with dysmorphology and WGS data along with their unaffected siblings (n= 355). Clinical Assessment and Morphological Examination – Discovery Cohort Clinical assessments, morphological examinations and classification were performed as previously described9,10. In brief, the team reviewed the child’s family history and medical records, including radiology and electroencephalogram (EEGs). EEGs were ordered if there was a clinical suspicion of seizures. Other screens for birth defects were arranged based on standard physical examination of the proband, which included a cardiovascular examination (e.g. echocardiogram for a proband with a murmur consistent with a 8 8 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint . j Subject enrolment – Discovery Cohort CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: edRxiv preprint Thus, the final number of ASD cases in the discovery cohort used for analysis was 276, of which 203 had nondysmorphic ASD and 73 had dysmorphic ASD according to ADM. j Subject enrolment – Discovery Cohort In contrast to the Autism Dysmorphology Measure, the morphological scores used to classify the discovery cohort factored in major congenital anomalies as well as MPAs, and MPAs were documented for the entire body including areas not assessed by the ADM (for example the thorax, arms, legs and skin). In order to align the type of morphologic data that was used to classify the discovery and replication cohorts, we reclassified the discovery cohort using the Autism Dysmorphology Measure, yielding 248 nondysmorphic and 77 dysmorphic cases. Of the 248 ADM-defined nondysmorphic cases, 18 cases were clearly In contrast to the Autism Dysmorphology Measure, the morphological scores used to classify the discovery cohort factored in major congenital anomalies as well as MPAs, and MPAs were documented for the entire body including areas not assessed by the ADM (for example the thorax, arms, legs and skin). In order to align the type of morphologic data that was used to classify the discovery and replication cohorts, we reclassified the discovery cohort using the Autism Dysmorphology Measure, yielding 248 nondysmorphic and 77 dysmorphic cases. Of the 248 ADM-defined nondysmorphic cases, 18 cases were clearly dysmorphic upon further review by an experienced dysmorphologist (B.A.F). The Autism Dysmorphology Measure is reported to have an 82% sensitivity28, and the sensitivity for the discovery cohort is similar at 80%. Thus, we excluded the 18 individuals with a false negative dysmorphic ADM classification to make the discovery cohort as clean as possible. We also included only samples that were sequenced on Illumina platforms to be consistent with the replication cohort27. y p , y dysmorphic upon further review by an experienced dysmorphologist (B.A.F). The Autism Dysmorphology Measure is reported to have an 82% sensitivity28, and the sensitivity for the discovery cohort is similar at 80%. Thus, we excluded the 18 individuals with a false negative dysmorphic ADM classification to make the discovery cohort as clean as possible. We also included only samples that were sequenced on Illumina platforms to be consistent with the replication cohort27. 9 9 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint . Thus, the final number of ASD cases in the discovery cohort used for analysis was 276, of which 203 had nondysmorphic ASD and 73 had dysmorphic ASD according to ADM. Whole-genome sequencing and variant detection We extracted DNA from whole blood or lymphoblast-derived cell lines and assessed the DNA quality with PicoGreenTM and gel electrophoresis. We sequenced 795 genomes (325 probands and 470 parents) with one of the following WGS technologies/sites as previously described4: Complete Genomics (Mountain View, CA, n= 33 probands, 64 parents), Illumina HiSeq2000 by The Centre for Applied Genomics (TCAG) (Toronto, ON, n= 24 probands, 48 parents), or Illumina HiSeq X by Macrogen (Seoul, South Korea, n= 182 probands, 250 parents) or TCAG (n= 86 probands, 108 parents). We used KING38 to confirm familial relationships and ADMIXTURE39 and EIGENSOFT40 to confirm ancestries (Supplementary Table 12). Alignment and variant calling for genomes sequenced by Complete Genomics were conducted as previously described41. For samples sequenced on Illumina platforms, each WGS site aligned WGS reads to the human reference genome assembly hg19 (GRCh37) using Burrows-Wheeler Aligner v.0.7.1242 (TCAG) or Isaac v.2.0.1343 (Macrogen). For each genome, we performed local realignment and quality recalibration and detected SNVs and small indels using the Genome Analysis Toolkit (GATK) Haplotype Caller44 v.3.4.6 without genotype refinement. We detected CNVs using ERDS (estimation by read depth with single nucleotide variants)45 and CNVnator46 as previously described47. We detected SVs using Manta v.0.29.648. When supported by the variant caller (i.e. GATK and Manta), trio-based joint variant calling was conducted for each family. To identify uniparental isodisomies (isoUPDs), we calculated the ratio of the number of homozygous or hemizygous SNPs to the number of SNPs per chromosome, for each sample. Samples with a ratio greater than 0.55 had a putative isoUPD on the corresponding chromosome. We examined CNV and kinship data to rule out confounding factors (i.e., large CNVs or consanguinity). For each sample with a ratio greater than 0.55, we examined plots of B-allele frequency per chromosome; those with runs of homozygosity > 10Mb on one chromosome were considered to have a putative isoUPD49. We examined the inheritance of homozygous SNPs within the region of the putative isoUPD via visual inspection of BAM files and experimentally validated one of the SNPs to confirm the isoUPD and inheritance. We systematically detected aneuploidies by calculating a ratio of the average read depth per chromosome to that for the entire sample. Ratios ≤0.5 and ≥1.5 were considered a loss or gain, respectively. For Complete Genomics data, we identified aneuploidies by looking for an excess of large CNVs for each chromosome per sample. Thus, the final number of ASD cases in the discovery cohort used for analysis was 276, of which 203 had nondysmorphic ASD and 73 had dysmorphic ASD according to ADM. 10 10 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: edRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint Tandem repeats were detected from samples with PCR-free DNA library preparation and sequenced on the Illumina HiSeq X platforms using ExpansionHunter Denovo50 with default parameters. We detected tandem repeat expansions in the discovery cohort using ExpansionHunterDenovo size cutoffs as previously described51. Sample quality control procedures were performed as previously described51. Variant Annotation ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: edRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint 2,573 parents in MSSNG (fourth version)4 to decrease batch and cross platform effects; and 4) allele frequency < 5% from 250 parents from this study aligned with Isaac to decrease alignment-specific artifacts. To further minimize cross platform and batch effects, we required heterozygous SNVs and indels to have an alternative allele fraction of 0.3-0.7 (inclusive) and homozygous/hemizygous SNVs and indels to have an alternative allele fraction >0.7 for variants from Complete Genomics. For Illumina variants, we also required heterozygous SNVs and indels to have a genotype quality score of at least 99 and 90, respectively, and homozygous SNVs and indels to have a genotype quality score of at least 25. We retained CNVs >2kb that had <70% overlap with gaps, centromeres, telomeres, and segmental duplications. For CNVs from Illumina platforms, we defined stringent CNVs as those called by both ERDS and CNVnator (with 50% reciprocal overlap). We defined CNVs as rare if the allelic frequency was < 1% in parents from the MSSNG database4 and < 5% in parents of this cohort that were aligned with Isaac. We retained as rare SVs, those with an allelic frequency of < 1% in parents analyzed with Manta from the MSSNG database and <5% in parents in this cohort that were aligned with Isaac. Pairs of entries with identical non-zero first numbers in the MATEID tag were retained as one inversion. Entries with identical MATEID values were retained as complex SVs. On average per sample, we detected ~3.7 million SNPs, 36,514 rare single nucleotide variants (SNVs), 4,113 small insertions and deletions (indels), 13 rare copy number variants (CNVs), 390 rare structural variations (Supplementary Table 2). Detection of de novo variants We determined de novo SNVs and indels from Complete Genomics data as previously described41. Variant Annotation Variant Annotation We annotated SNVs and indels with information on population allele frequency, variant impact predictors, and putative pathogenicity and disease association, using a custom pipeline based on ANNOVAR52 as previously described4. For non-genic regions, we annotated whether the variant overlapped reported ASD- associated non-coding regions19-23 (Supplementary Table 19). These included transcription start sites, fetal brain promoters and enhancers of LoF intolerant genes20, histone modification (H3K27ac) sites in fetal and adult brain21, splice sites, 3’- and 5’-untranslated regions (UTRs)23, binding sites predicted by DeepSEA22 to cause LoF, as well as conserved promoters of any genes, developmental delay-associated genes, and long non-coding RNA genes19. We tested three additional functional sites that have not been previously associated with ASD. These included boundaries of topologically associating domains53, CTCF binding sites54, and brain enhancers from Roadmap Epigenomics chromatin states (15-states chromHMM)55. We annotated CNVs and SVs with a custom pipeline using RefSeq gene models, with repeat regions, gaps, centromeres, telomeres and segmental duplications relative to University of California at Santa Cruz genome assembly hg19. Similar to our non-genic annotations for SNVs, we annotated whether a CNV overlapped promoters of genes19, H3K27ac sites19, 3’UTR and 5’UTR23 (Supplementary Table 19). We retained CNVs overlapping such regions, but not exonic regions. We also annotated the frequency of each CNV and SV from among 3,107 parents in the MSSNG database4 (fifth version) and the putative pathogenicity and disease association [from Human and Mouse Phenotype Ontologies56,57 (HPO and MPO), ClinGen Genome Dosage Sensitivity Map58, Online Mendelian Inheritance in Man, and Database of genomic variation and phenotype in humans using ensemble resources (DECIPHER)59]. We annotated mitochondrial variants using Annovar-based custom scripts with annotations from MitoMaster (April 2019) and Ensembl v96. Detection of rare variants We extracted high quality rare data for SNVs and indels after applying the following filters: 1) FILTER is PASS or varQuality is VQHIGH or PASS; 2) population allele frequencies < 1% in 1000 Genome Project60, NHLBI-ESP61, Exome Aggregation Consortium62, The Genome Aggregation Database63, and internal Complete Genomics control databases; 3) reference and alternative allele frequency > 95% and < 1%, respectively, based on allele frequencies of 11 11 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. Variant Annotation For Illumina WGS data, we also used DenovoGear64 (version 0.5.4) to detect de novo SNVs and indels. We extracted variants inconsistent with Mendelian inheritance (present in offspring but not in parents) with FILTER= PASS and defined rare, as above. To identify high confidence de novo SNVs, we applied the following quality filters: 1) pp_DNM score ≥ 0.9 from DenovoGear64; 2) overlap GATK44 calls with genotype quality scores ≥ 99 for heterozygous SNVs. We defined high confidence de novo indels as those called by DenovoGear and GATK with the same start site. We retained de novo SNVs and indels with a ratio of sequenced reads supporting the alternative call to the total number of reads at the position of 0.3-0.7, or > 0.7 for X- and Y-linked variants not in the pseudoautosomal regions in male subjects. We defined putative de novo CNVs as rare stringent CNVs (see “Detection of rare variants”) that were inconsistent with Mendelian inheritance. For CNVs that did not have a conclusive inheritance pattern (i.e., CNV in child and parent were not the same size), we defined putative de novo CNV as those with a CNV length ratio between child and parent of > 2. For each putative de novo CNV from 12 12 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: edRxiv preprint Illumina platforms, we calculated a read depth ratio of the CNV with the surrounding region in each family member, as previously described47. Ratios of 0.35-0.65 were considered heterozygous deletions, <0.35 as homozygous/hemizygous deletions, >= 1.4 as duplications and 0.9-1.1 as a normal copy number. Variant Annotation Putative de novo CNVs were considered de novo if the copy number status based on ratios were inconsistent with Mendelian inheritance. For the 40 regions with ratios that did not meet the afore-mentioned criteria, we visualized the WGS reads to determine the inheritance status for samples sequenced by Illumina. To determine the inheritance status for samples sequenced by Complete Genomics, we examined the read depth coverage of the CNV relative to that of Complete Genomics controls65 and its flanking regions in each family member. On average per sample, we detected 73.4 de novo SNVs, 7.3 de novo indels, and 0.1 de novo CNVs (Supplementary Table 2) Validation of variants We randomly selected a subset of all high quality exonic de novo SNVs, all de novo indels and all CSVs for validation in probands and available parents. We used Primer366 to design primers to span at least 100 bp upstream and downstream of a putative variant, avoiding regions of known SNPs, repetitive elements, and segmental duplications. DNA from whole blood, if available, was used to amplify candidate regions by polymerase chain reaction and to assay with Sanger Sequencing. For CNVs, we validated all high confidence de novo exonic and all clinically significant CNVs in whole blood DNA (if available) of probands and available parents using TaqManTM Copy Number Assay (Applied Biosystems), SYBR ® Green qPCR (Thermofisher) or digital droplet PCR (BioRad). Experimental validation rates were 94.8%, 85.7%, and 87.5%, respectively, for de novo SNVs, indels, and CNVs (Supplementary Tables 3 and 4). Mitochondrial variant detection For the samples sequenced by Illumina platforms, reads aligning to the mitochondrial genome were extracted and realigned to the revised Cambridge Reference Sequence (NC_012920) in b37 using BWA v0.7.8. Pileups were generated with samtools mpileup v1.1 requiring the program to include duplicate reads in the analysis and retaining all positions in the output. Custom scripts were developed to parse the mpileup output to determine the most frequently occurring non-reference base at each position in the mitochondrial genome. The heteroplasmic fractions were calculated and vcf files were generated. Fasta files with the most frequently occurring base at every position were also generated and used as input for the program HaploGrep v2.1.1 for haplogoup prediction. The vcf files were annotated using Annovar based custom scripts with annotations from MitoMaster (April 2019) and Ensembl v96. Variant Annotation For the samples that were sequenced by Complete Genomics, the mitochondrial variants called by the proprietary software were extracted. Fasta files were generated using custom scripts replacing mitochondrial reference bases with 13 13 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: edRxiv preprint alternative bases at heteroplasmic sites and the files were used as input for the program HaploGrep v2.1.1 for haplogoup prediction. The vcf files were annotated using Annovar based custom scripts with annotations from MitoMaster (April 2019) and Ensembl v96. alternative bases at heteroplasmic sites and the files were used as input for the program HaploGrep v2.1.1 for haplogoup prediction. The vcf files were annotated using Annovar based custom scripts with annotations from MitoMaster (April 2019) and Ensembl v96. Positions with heteroplasmic fraction less than 5% or greater than 95% and common in certain haplogroups (greater than 5%) were excluded from downstream analysis. All variants were manually reviewed, and a list of artefacts was compiled and excluded. To identify pathogenic mitochondrial variants, the following variants were considered: any MitoMaster pathogenic variants at 5- 100% heteroplasmy, variants between 10-90% heteroplasmy, and variants between 5-100% heteroplasmy and seen <2% of the time in the individual’s haplogroup. Variant detection for replication cohort For the replication cohort, CRAM files and sequence-level variants were downloaded from Globus (https://www.globus.org/). We detected CNVs using ERDS 45 and CNVnator46, as previously described47. Rare variants were filtered as described for the discovery cohort. We identified de novo SNVs and indels using DeNovoGear64. Variant Annotation Therefore, we also manually inspected WGS data when we identified CSVs that did not pass filtering criteria for high confidence variants. Clinically significant variants classified as pathogenic or likely pathogenic or that were considered clinically relevant (i.e., prompting further clinical management) were reviewed by a medical geneticist in the context of the patient’s phenotype and family history. Relevant findings were reported back to families through a clinical geneticist. Differences in the yield of CSVs among the morphological groups were calculated using Fisher’s exact test. To identify CSVs from the affected probands in the replication cohort, the afore- mentioned approach was applied to de novo LoF, damaging missense and CNVs. CSVs from the replication cohort were confirmed by manual inspection of WGS reads. Rare variant burden analysis in gene sets and noncoding regions For the discovery cohort, we performed two ASD subtype comparisons for each rare variant burden analysis as follows: 1) comparing complex, equivocal and essential ASD using ordinal regression tests and 2) comparing complex and equivocal ASD (i.e. dysmorphic ASD) to essential ASD using logistic regression tests. The test was done by regressing an event (e.g., number of genes impacted by rare deletions per subject) capturing a particular genomic region (i.e., coding, gene sets, or noncoding regions) on the phenotype outcome (e.g., complex vs. essential ASD). The events tested in this study were the number of LoF, missense, and predicted deleterious variants for sequence-level variants and the number of genes or noncoding regions for CNVs. Tier 1 and 2 missense variants consist of all or only predicted damaging missense variants, respectively, as defined in Yuen, et al.23 The CNVs were grouped into two size bins, small CNVs (2-10kb) and large CNVs (10kb to 3Mb) due to greater proportion of these CNVs overlapping coding or noncoding regions, respectively. The number of genes impacted by other CNVs was based on their overlap with the coding regions of each gene. However, the number of genes impacted by small CNVs were based on genic overlap since there were not enough small coding CNVs for the gene set enrichment analysis. We compiled a list of 37 gene sets related to neuronal function, brain expression, mouse phenotypes from MPO, or human phenotypes from HPO that have been previously associated with ASD or used as negative control gene sets when comparing ASD to control groups (Supplementary Table 20) 71-81. Variant Annotation Allele frequencies from the Simons Simplex Collection were calculated and de novo variants with internal frequencies <1% were excluded. De novo SNVs and indels at poorly sequenced or highly variable sites were also excluded from further analysis. The remaining de novo variants were filtered as described for the discovery cohort with the exception of using a PP_DNM <0.95 threshold for de novo SNVs. Variants were annotated as described above for the discovery cohort. Variant prioritization and molecular diagnosis To identify CSVs from the discovery cohort, we prioritized rare and de novo LoF and damaging (as predicted by at least five/seven predictors23) missense variants, and variants reported by ClinVar67 or the Human Gene Variant Database68. We also prioritized rare and de novo CNVs and SVs, including those overlapping syndromic regions in DECIPHER59 or ClinGen Genome Dosage Sensitivity Map58 databases. Genes affected by such variants were compared to ASD candidate genes3,4,13,69,70, candidate genes for neurodevelopmental disorders69, and genes implicated in neurodevelopmental or behavioural phenotypes according to HPO57 and MPO56. Additionally, we considered the mode of inheritance from the Online Mendelian Inheritance in Man and Clinical Genomics Database70, segregation and genotype-phenotype correlations. We classified the variants as pathogenic, likely pathogenic, variants of uncertain significance, likely benign, or benign, based on the American College of Medical Genetics and Genomics Guidelines16,17. Variants of unknown significance in known or candidate ASD genes with emerging evidence were further categorized into three ASD candidate variant categories (Supplementary Note and Supplementary Tables 5-7). Although applying quality filters for high confidence variants is important to minimize false positives for burden analysis, this can 14 14 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: edRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint increase false negatives. Therefore, we also manually inspected WGS data when we identified CSVs that did not pass filtering criteria for high confidence variants. increase false negatives. Variant Annotation For the gene sets and noncoding regions burden test, total variant count (for SNVs and noncoding CNVs) or total gene count (for CNVs) was also included as one of the covariates to get rid of a global burden bias that might inflate the test P value. The coefficients, however, were calculated from the model with all the covariates except the total variant count or the total gene count for the actual magnitude of their impact. Permutation-based FDR correction (1000 permutations) corrected for the multiple comparison. Since different gene sets and non-coding regions consist of different number of genes or regions, we calculated the coefficients using z-scores for the number of features in each gene set/region to compare the coefficients across morphology-associated regions. When examining the burden of rare variants using logistic regression models, we used all probands from the discovery cohort (n=325). Since some probands did not have their parents sequenced, we used a subset of the discovery cohort (n= 235) when examining de novo variants. covariates and another with all both covariates and target variable as previously described82. Global burden analysis was performed to compare the total number of LOF variants, missense variants, predicted deleterious variants for sequence- level variants, and genes impacted by CNVs. The coefficients reported were obtained from the model with the covariates. Multiple test correction for global burden tests was done using Benjamini Hochberg approach (BH-FDR). For the gene sets and noncoding regions burden test, total variant count (for SNVs and noncoding CNVs) or total gene count (for CNVs) was also included as one of the covariates to get rid of a global burden bias that might inflate the test P value. The coefficients, however, were calculated from the model with all the covariates except the total variant count or the total gene count for the actual magnitude of their impact. Permutation-based FDR correction (1000 permutations) corrected for the multiple comparison. Since different gene sets and non-coding regions consist of different number of genes or regions, we calculated the coefficients using z-scores for the number of features in each gene set/region to compare the coefficients across morphology-associated regions. When examining the burden of rare variants using logistic regression models, we used all probands from the discovery cohort (n=325). Since some probands did not have their parents sequenced, we used a subset of the discovery cohort (n= 235) when examining de novo variants. Variant Annotation For non-coding regions, we compiled a list of regions reported to be associated with ASD (Supplementary Table 19)19-21. We also included a score that predicts the impact of a variant on transcription factor binding as one of the non-coding regions tested22. Logistic regression and ordinal regression were applied for two subtypes and three subtypes comparison, respectively. Sex, genotyping platform, and three principal components from population stratification were included in the model as covariates to correct for any biases caused by sex difference, platforms, or ethnicity. Deviance test P value was calculated by comparing residuals from two regression models; one with just the Rare variant burden analysis in gene sets and noncoding regions For the discovery cohort, we performed two ASD subtype comparisons for each rare variant burden analysis as follows: 1) comparing complex, equivocal and essential ASD using ordinal regression tests and 2) comparing complex and equivocal ASD (i.e. dysmorphic ASD) to essential ASD using logistic regression tests. The test was done by regressing an event (e.g., number of genes impacted by rare deletions per subject) capturing a particular genomic region (i.e., coding, gene sets, or noncoding regions) on the phenotype outcome (e.g., complex vs. 15 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: edRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint covariates and another with all both covariates and target variable as previously described82. Global burden analysis was performed to compare the total number of LOF variants, missense variants, predicted deleterious variants for sequence- level variants, and genes impacted by CNVs. The coefficients reported were obtained from the model with the covariates. Multiple test correction for global burden tests was done using Benjamini Hochberg approach (BH-FDR). Variant Annotation ch individual, the GRVS was calculated using the formula below 𝛽𝑋𝑛𝑘 For each individual, the GRVS was calculated using the formula below 𝛽𝑖𝑖𝑋𝑖𝑖𝑛𝑘 GRVS = ෍෍𝛽𝛽𝑖𝑖𝑖𝑖𝑋𝑋𝑖𝑖𝑖𝑖 𝑛𝑛 𝑖𝑖=1 𝑘𝑘 𝑗𝑗=1 GRVS = ෍෍𝛽𝛽𝑖𝑖𝑖𝑖𝑋𝑋𝑖𝑖𝑖𝑖 𝑛𝑛 𝑖𝑖=1 𝑘𝑘 𝑗𝑗=1 𝑗𝑗 where n is the number of significant (P < 0.1) gene sets or regions for a particular variant type, j, k is the number of variant types (e.g., de novo missense variants), βi is a regression coefficient of a significant gene set or region, i, and Xi is the number of variants (for SNVs and indels) or number of genes or regions (for CNVs) that are found in the significant gene set or region in the sample. where n is the number of significant (P < 0.1) gene sets or regions for a particular variant type, j, k is the number of variant types (e.g., de novo missense variants), βi is a regression coefficient of a significant gene set or region, i, and Xi is the number of variants (for SNVs and indels) or number of genes or regions (for CNVs) that are found in the significant gene set or region in the sample. To examine the GRVS in the discovery cohort, we used a 10-fold cross validation strategy to avoid over-fitting. Using this strategy, the discovery cohort was randomly divided into 10 equally sized subsamples (stratified by subtypes). We calculated the GRVS of each sample in each subset using the effect sizes determined in the remaining nine subsets. To minimize stochasticity in the GRVS calculation, we repeated this procedure 30 times and the average GRVS and average number of variants for each sample were used for subsequent subtypes comparisons (Supplementary Figure 1a). For the replication cohort, we calculated GRVSs using significant gene sets and effect sizes derived from the discovery cohort (Supplementary Figure 1b). GRVS can be calculated for probands regardless of whether their parents have been sequenced. However, there would be a systematic difference in GRVSs in the discovery cohort if all probands were used because those whose parents have been sequenced includes scores from de novo variants, whereas probands whose parents have not been sequenced do not have scores from de novo variants. To ensure that the same variant types (including de novo variants) were included in each score for probands in the discovery cohort, GRVS was calculated only for probands whose parents had also both been sequenced. Variant Annotation (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: edRxiv preprint as the number of rare or de novo SNVs or indels or the number of genes or noncoding regions impacted by rare CNVs. For rare variants, we used all probands in the discovery cohort. Since some probands did not have their parents sequenced, we used a subset of the discovery cohort when examining de novo variants. To determine the optimal P value threshold to identify significant gene sets, we calculated the Nagelkerke’s R2 at different P value thresholds (P < 0.001, 0.005, 0.01, 0.05, 0.1. 0.5, and 1) using the discovery cohort and 10-fold cross-validation strategy. The optimal P value threshold was at P<0.1 (Supplementary Figure 3). To minimize the redundancy in significant gene sets and noncoding regions, we retained the most significant gene sets and noncoding regions with a Jaccard index < 0.75. We used the regression coefficients (βi) of significant gene sets or noncoding regions (P < 0.1) as a weight for the number of variants in those gene sets or regions in the GRVS calculation. as the number of rare or de novo SNVs or indels or the number of genes or noncoding regions impacted by rare CNVs. For rare variants, we used all probands in the discovery cohort. Since some probands did not have their parents sequenced, we used a subset of the discovery cohort when examining de novo variants. To determine the optimal P value threshold to identify significant gene sets, we calculated the Nagelkerke’s R2 at different P value thresholds (P < 0.001, 0.005, 0.01, 0.05, 0.1. 0.5, and 1) using the discovery cohort and 10-fold cross-validation strategy. The optimal P value threshold was at P<0.1 (Supplementary Figure 3). To minimize the redundancy in significant gene sets and noncoding regions, we retained the most significant gene sets and noncoding regions with a Jaccard index < 0.75. We used the regression coefficients (βi) of significant gene sets or noncoding regions (P < 0.1) as a weight for the number of variants in those gene sets or regions in the GRVS calculation. Variant Annotation Genome-wide rare variant score In addition to identifying relevant gene sets or regions that were differentially enriched among ASD morphologic subgroups, we developed a procedure to calculate a genome-wide rare variant score (GRVS) for each subject. This allowed the contribution of different variant types towards phenotype severity to be assessed together. The procedure involved two main steps: i) identification of relevant, differentially enriched gene sets or noncoding regions for each variant type along with an estimation of their effect sizes in the discovery cohort, and ii) calculation of the score for each subject in the target cohorts. To estimate the effect sizes in the discovery cohort, we first fitted a logistic regression model by regressing platform, sex and first three principal components from population stratification on the dysmorphology classification (nondysmorphic= 0 and dysmorphic= 1, or essential= 0, equivocal= 1, complex= 2). We then used the regression coefficients of these covariates and the intercept in the second logistic regression model, where a feature representing a particular gene set or region was tested. Therefore, regression coefficients of all the gene sets and regions were corrected for those possible biases from the covariates equally. The two models can be notated as below Y= ɑ + βC Y= ɑ + βC + βiXi where Y is the outcome variable of dysmorphology classification, ɑ is an intercept, β is a regression coefficients of covariates, C is a vector of covariates, βi is the regression coefficient of a morphology-associated region, i, and Xi is the number of features found in a morphology-associated region. A feature is defined 16 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Variant Annotation GRVSs were standardized within each cohort and subtyping method. We tested whether GRVS is higher in dysmorphic ASD compared to nondysmorphic ASD using a one-sided Wilcoxon’s Signed Ranked Test. 17 17 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: edRxiv preprint We used our ADM-reclassified cohort as the discovery cohort for several reasons: 1) In contrast to the MPAs (dysmorphology data) from SSC which were identified by multiple non-geneticist examiners, MPAs in the discovery cohort were documented by a single dysmorphologist with over 20 years of clinical experience (B.A.F.). MPA’s for children in the discovery cohort were then put through the ADM algorithim and the cases were classified as ADM-dysmorphic or ADM-nondysmorphic. This strategy allowed us to use very uniformly collected phenotypic data to derive the morphology-associated regions and effect sizes for GRVS calculation. 2) Our discovery cohort also contains more dysmorphic probands than SSC, which gives more power to identify morphology-associated regions (enriched in dysmorphic ASD). 3) Lastly, the discovery cohort was assembled using a population-based recruitment strategy so that the morphology-associated regions identified come from a patient collection representative of ASD as it exists at the level of primary care providers. In contrast there are ascertainment biases in SSC (e.g., simplex families and exclusion of severely affected/ syndromic probands) which might limit the generalizability of effect sizes and morphology-associated regions in a population-based cohort30. Variant Annotation ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint with iPSYCH-ASD and GIANT-BMI SNPs passing suggested a p-value threshold (P value < 0.1 for ASD and P value < 0.2 for BMI) by Weiner et al.2, respectively. We then calculated PRSs using PRSice84 (parameters used: clump-kb 250, clump-p 1.000000, clump-r2 0.100000, info-base 0.9) using a p-value threshold of 0.1 for iPSYCH and 0.2 for GIANT BMI, as suggested by Weiner et al.8. After clumping, only 18,549 SNPs and 38,245 SNPs remained for PRS calculation for ASD and BMI, respectively. Using standard methods as previously described11, we calculated PRS for ASD for the SSC replication cohort using 26,067 SNPs with P value < 0.1 after the clumping step. The PRSs in both cohorts were standardized (with a mean of zero and standard deviation of one). We used the pTDT method8 and one-sided t-test to examine the over-transmission of common variants associated with ASD susceptibility among subtypes. Probands were used in the analysis if the probands were of European ancestry and if sequencing data was available from both parents. with iPSYCH-ASD and GIANT-BMI SNPs passing suggested a p-value threshold (P value < 0.1 for ASD and P value < 0.2 for BMI) by Weiner et al.2, respectively. We then calculated PRSs using PRSice84 (parameters used: clump-kb 250, clump-p 1.000000, clump-r2 0.100000, info-base 0.9) using a p-value threshold of 0.1 for iPSYCH and 0.2 for GIANT BMI, as suggested by Weiner et al.8. After clumping, only 18,549 SNPs and 38,245 SNPs remained for PRS calculation for ASD and BMI, respectively. Using standard methods as previously described11, we calculated PRS for ASD for the SSC replication cohort using 26,067 SNPs with P value < 0.1 after the clumping step. The PRSs in both cohorts were standardized (with a mean of zero and standard deviation of one). We used the pTDT method8 and one-sided t-test to examine the over-transmission of common variants associated with ASD susceptibility among subtypes. Probands were used in the analysis if the probands were of European ancestry and if sequencing data was available from both parents. Acknowledgements g We thank the families for participation and The Centre for Applied Genomics for their analytical and technical support. We thank Lisa Strug, Andrew Paterson, and Delnaz Roshandel for analytical assistance. This work was funded by Autism Speaks, Autism Speaks Canada, the University of Toronto McLaughlin Centre, the Canada Foundation for Innovation, the Canadian Institutes of Health Research (CIHR), Genome Canada/Ontario Genomics Institute, the Government of Ontario, Brain Canada, Ontario Brain Institute Province of Ontario Neurodevelopmental Disorders (POND), and The Hospital for Sick Children Foundation. A.J.S.C. was supported throughout this research by Ontario Graduate Scholarship from the Government of Ontario, Restracomp Research Fellowship from The Hospital of Sick Children, and Autism Research Training Award and Frederick Banting and Charles Best Scholarship from CIHR. S.W.S holds the Northbridge Chair in Paediatric Research at the Hospital for Sick Children. Variant Annotation We calculated a score for CSVs using the GRVS formula if the CSV was identified in a proband with two sequenced parents, and if the variant occurred in or overlapped one of the morphology-associated gene sets or noncoding regions so that an effect size was available for that variant. 46 CSVs were identified in 46 probands and17 of these met the above criteria allowing us to calculate a score for the variant. Of the remaining 29 CSVs, 15 were identified in probands where sequencing data was not available from both parents and 14 variants did not overlap a morphology-associated region. Common variant and PRS analysis We examined the contribution of common SNPs among ASD subtypes. We calculated the PRS for each sample by deriving ASD summary statistics from a population-based genome-wide association study (GWAS) of 13,076 cases and 22,664 controls from the iPSYCH project 11. We calculated the PRS for BMI, which was a negative control due to its lack of association with ASD24, using BMI summary statistics from a population-based GWAS of 322,154 individuals of European descent from the GIANT Consortium83. We preprocessed the GWAS summary tables to fix the effect allele mismatch (swapped A1 and A2 alleles and converted its odds ratio) and to remove ambiguous SNPs (i.e., SNPs with A to T and C to G variations) and multi-allelic SNPs. We conducted joint genotyping of BMI- and ASD-associated SNPs only on samples sequenced on Illumina platforms (200 probands and 400 parents). We could not re-genotype Complete Genomics data, so the samples were excluded from further analysis. We retained SNPs with a minor allele frequency > 0.05 and genotyping rate > 90%, of which 349,682 SNPs and 428,364 SNPs intersected 18 18 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: edRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. Author Contributions A.J.S.C., R.K.C.Y., S.W.S., and B.A.F. conceived and designed experiments. B.A.F, C.N., T.N.T., and J.H.M. managed, recruited, diagnosed and examined participants. E.A. and R.P. helped with interpreting phenotype data. Z.W., B. Thiruvahindrapuram, B.Trost, T.N., G.P., W.S., and J.M. processed whole- genome sequencing data. A.J.S.C., W.E., R.K.C., D.M., D.R. and M.Z. conducted or interpreted different components of whole genome sequencing analyses. A.J.S.C., M.S.R., D.J.S., N.S. and K.T. performed variant interpretation. A.J.S.C. and S.L. performed experiments for variant characterization and validation. A.J.S.C., B.A.F., S.W.S., and R.K.C.Y. conceived and coordinated the project and wrote the manuscript. Data availability y Access to FASTQ data for samples in the discovery cohort that were consented for MSSNG can be obtained by completing the data access agreement: https://research.mss.ng. Access to FASTQ data for samples in the discovery cohort not consented for MSSNG, as well as VCF files for sequence-level variants for all samples in the discovery cohort are available at European Genome-Phenome Archive (pending EGA link and accession number. Submission in process.). Access to data for the replication cohort can be obtained by completing data access agreement (htt // f i / / f i b ) d f thi t d y p g g (https://www.sfari.org/resource/sfari-base), as was done for this study. Code availability Code used in this manuscript is available at GitHub (http://github.com/naibank/GRVS_ASD). References 1. Ozgen, H.M., Hop, J.W., Hox, J.J., Beemer, F.A. & van Engeland, H. Minor physical anomalies in autism: a meta-analysis. Mol Psychiatry 15, 300-7 (2010). 2. Timonen-Soivio, L. et al. The association between congenital anomalies and autism spectrum disorders in a Finnish national birth cohort. Dev Med Child Neurol 57, 75-80 (2015). 3. Sanders, S.J. et al. Insights into Autism Spectrum Disorder Genomic Architecture and Biology from 71 Risk Loci. Neuron 87, 1215-33 (2015). 4. Yuen, R. et al. Whole genome sequencing resource identifies 18 new candidate genes for autism spectrum disorder. Nat Neurosci 20, 602-611 (2017). 5. Douard, E. et al. Effect Sizes of Deletions and Duplications on Autism Risk Across the Genome. Am J Psychiatry 178, 87-98 (2021). 6. Schaaf, C.P. et al. A framework for an evidence-based gene list relevant to autism spectrum disorder. Nat Rev Genet 21, 367-376 (2020). 7. Fernandez, B.A. & Scherer, S.W. Syndromic autism spectrum disorders: moving from a clinically defined to a molecularly defined approach. Dialogues Clin Neurosci 19, 353-371 (2017). 8. Weiner, D.J. et al. Polygenic transmission disequilibrium confirms that common and rare variation act additively to create risk for autism spectrum disorders. Nat Genet 49, 978-985 (2017). 9. Miles, J.H. et al. Essential versus complex autism: definition of fundamental prognostic subtypes. Am J Med Genet A 135, 171-80 (2005). 10. Tammimies, K. et al. Molecular Diagnostic Yield of Chromosomal Microarray Analysis and Whole-Exome Sequencing in Children With Autism Spectrum Disorder. JAMA 314, 895-903 (2015). 11. Grove, J. et al. Identification of common genetic risk variants for autism spectrum disorder. Nat Genet 51, 431-444 (2019). 12. Iossifov, I. et al. The contribution of de novo coding mutations to autism spectrum disorder. Nature 515, 216-21 (2014). 13. Pinto, D. et al. Convergence of genes and cellular pathways dysregulated in autism spectrum disorders. Am J Hum Genet 94, 677-94 (2014). 1. Ozgen, H.M., Hop, J.W., Hox, J.J., Beemer, F.A. & van Engeland, H. Minor physical anomalies in autism: a meta-analysis. Mol Psychiatry 15, 300-7 (2010). 1. Ozgen, H.M., Hop, J.W., Hox, J.J., Beemer, F.A. & van Engeland, H. Minor physical anomalies in autism: a meta-analysis. Mol Psychiatry 15, 300-7 (2010). 2. Timonen-Soivio, L. et al. The association between congenital anomalies and autism spectrum disorders in a Finnish national birth cohort. Dev Med Child Neurol 57, 75-80 (2015). ( ) 3. Sanders, S.J. et al. Competing Interests statement p g S.W.S. is on the Scientific Advisory Committees of Deep Genomics, Population Bio and an Academic Consultant for the King Abdulaziz University. 19 19 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity (which was not certified by peer review) The copyright holder for this pre this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint References Insights into Autism Spectrum Disorder Genomic Architecture and Biology from 71 Risk Loci. Neuron 87, 1215-33 (2015). 4. Yuen, R. et al. Whole genome sequencing resource identifies 18 new candidate genes for autism spectrum disorder. Nat Neurosci 20, 602-611 (2017). 5. Douard, E. et al. Effect Sizes of Deletions and Duplications on Autism Risk Across the Genome. Am J Psychiatry 178, 87-98 (2021). y y ( ) Schaaf, C.P. et al. A framework for an evidence-based gene list relevant to autism spectrum disorder. Nat Rev Genet 21, 367-376 (2020). 6. Schaaf, C.P. et al. A framework for an evidence-based gene list relevant to autism spectrum disorder. Nat Rev Genet 21, 367-376 (2020). 7. Fernandez, B.A. & Scherer, S.W. Syndromic autism spectrum disorders: moving from a clinically defined to a molecularly defined approach. Dialogues Clin Neurosci 19, 353-371 (2017). ( ) 8. Weiner, D.J. et al. Polygenic transmission disequilibrium confirms that common and rare variation act additively to create risk for autism spectrum disorders. Nat Genet 49, 978-985 (2017). 8. Weiner, D.J. et al. Polygenic transmission disequilibrium confirms that common and rare variation act additively to create risk for autism spectrum disorders. Nat Genet 49, 978-985 (2017). , ( ) 9. Miles, J.H. et al. Essential versus complex autism: definition of fundamental prognostic subtypes. Am J Med Genet A 135, 171-80 (2005). ( ) 9. Miles, J.H. et al. Essential versus complex autism: definition of fundamental prognostic subtypes. Am J Med Genet A 135, 171-80 (2005). p g yp ( ) 10. Tammimies, K. et al. Molecular Diagnostic Yield of Chromosomal Microarray Analysis and Whole-Exome Sequencing in Children With Autism Spectrum Disorder. JAMA 314, 895-903 (2015). p g yp ( ) 10. Tammimies, K. et al. Molecular Diagnostic Yield of Chromosomal Microarray Analysis and Whole-Exome Sequencing in Children With Autism Spectrum Disorder. JAMA 314, 895-903 (2015). 11. Grove, J. et al. Identification of common genetic risk variants for autism spectrum disorder. Nat Genet 51, 431-444 (2019). 11. Grove, J. et al. Identification of common genetic risk variants for autism spectrum disorder. Nat Genet 51, 431-444 (2019). 12. Iossifov, I. et al. The contribution of de novo coding mutations to autism spectrum disorder. Nature 515, 216-21 (2014). 12. Iossifov, I. et al. The contribution of de novo coding mutations to autism spectrum disorder. Nature 515, 216-21 (2014). 13. Pinto, D. et al. References Convergence of genes and cellular pathways dysregulated in autism spectrum disorders. Am J Hum Genet 94, 677-94 (2014). 13. Pinto, D. et al. Convergence of genes and cellular pathways dysregulated in autism spectrum disorders. Am J Hum Genet 94, 677-94 (2014). 20 20 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: edRxiv preprint 14. Madsen, B.E. & Browning, S.R. A groupwise association test for rare mutations using a weighted sum statistic. PLoS Genet 5, e1000384 (2009). g g ( ) 15. Wu, M.C. et al. Rare-variant association testing for sequencing data with the sequence kernel association test. Am J Hum Genet 89, 82-93 (2011). 16. Richards, S. et al. Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology. Genet Med 17, 405-24 (2015). ( ) 17. Riggs, E.R. et al. Technical standards for the interpretation and reporting of constitutional copy-number variants: a joint consensus recommendation of the American College of Medical Genetics and Genomics (ACMG) and the Clinical Genome Resource (ClinGen). Genet Med 22, 245-257 (2020). ( ) ( ) 18. Trost, B. et al. Genome-wide detection of tandem DNA repeats that are expanded in autism. Nature (2020). 19. An, J.Y. et al. Genome-wide de novo risk score implicates promoter variation in autism spectrum disorder. Science 362(2018). 20. Brandler, W.M. et al. Paternally inherited cis-regulatory structural variants are associated with autism. Science 360, 327-331 (2018). 21. Liu, Y. et al. References (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: edRxiv preprint 35. Lord, C., Rutter, M., DiLavore, P. D., & Risi, S. Autism Diagnostic Observation Schedule, (Western Psychological Services, Los Angeles, CA, 2001). 36. Le Couteur, A., Lord, C., & Rutter, M. . The Autism Diagnostic Interview— Revised (ADI-R), (Western Psychological Services, Los Angeles, CA, 2003). f 37. Miles, J.H. & Hillman, R.E. Value of a clinical morphology examination in autism. Am J Med Genet 91, 245-53 (2000). 38. Manichaikul, A. et al. Robust relationship inference in genome-wide association studies. Bioinformatics 26, 2867-73 (2010). 38. Manichaikul, A. et al. Robust relationship inference in genome-wide association studies. Bioinformatics 26, 2867-73 (2010). ( ) 39. Alexander, D.H., Novembre, J. & Lange, K. Fast model-based estimation o ancestry in unrelated individuals. Genome Res 19, 1655-64 (2009). 40. Price, A.L. et al. Principal components analysis corrects for stratification in genome-wide association studies. Nat Genet 38, 904-9 (2006). 41. Yuen, R.K. et al. Whole-genome sequencing of quartet families with autism spectrum disorder. Nat Med 21, 185-91 (2015). 42. Li, H. & Durbin, R. Fast and accurate long-read alignment with Burrows-Wheele transform. Bioinformatics 26, 589-95 (2010). 43. Raczy, C. et al. Isaac: ultra-fast whole-genome secondary analysis on Illumina sequencing platforms. Bioinformatics 29, 2041-3 (2013). 44. McKenna, A. et al. The Genome Analysis Toolkit: a MapReduce framework for analyzing next-generation DNA sequencing data. Genome Res 20, 1297-303 (2010). ( ) 45. Zhu, M. et al. Using ERDS to infer copy-number variants in high-coverage genomes. Am J Hum Genet 91, 408-21 (2012). g ( ) 46. Abyzov, A., Urban, A.E., Snyder, M. & Gerstein, M. CNVnator: an approach to discover, genotype, and characterize typical and atypical CNVs from family and population genome sequencing. Genome Res 21, 974-84 (2011). 47. Trost, B. et al. A Comprehensive Workflow for Read Depth-Based Identification of Copy-Number Variation from Whole-Genome Sequence Data. Am J Hum Genet 102, 142-155 (2018). 48. Chen, X. et al. Manta: rapid detection of structural variants and indels for germline and cancer sequencing applications. Bioinformatics 32, 1220-2 (2016). 48. Chen, X. et al. Manta: rapid detection of structural variants and indels for germline and cancer sequencing applications. Bioinformatics 32, 1220-2 (2016). 49. Bis, D.M. et al. References A Statistical Framework for Mapping Risk Genes from De Novo Mutations in Whole-Genome-Sequencing Studies. Am J Hum Genet 102, 1031- 1047 (2018). ( ) 22. Zhou, J. et al. Whole-genome deep-learning analysis identifies contribution of noncoding mutations to autism risk. Nat Genet 51, 973-980 (2019). g 23. Yuen, R.K. et al. Genome-wide characteristics o NPJ Genom Med 1, 160271-1602710 (2016). 23. Yuen, R.K. et al. Genome-wide characteristics of de novo mutations in autism. NPJ Genom Med 1, 160271-1602710 (2016). 24. Bulik-Sullivan, B. et al. An atlas of genetic correlations across human diseases and traits. Nat Genet 47, 1236-41 (2015). 25. Krumm, N. et al. Excess of rare, inherited truncating mutations in autism. Nat Genet 47, 582-8 (2015). 26. Samocha, K.E. et al. A framework for the interpretation of de novo mutation in human disease. Nat Genet 46, 944-50 (2014). 27. Fischbach, G.D. & Lord, C. The Simons Simplex Collection: a resource for identification of autism genetic risk factors. Neuron 68, 192-5 (2010). 28. Miles, J.H. et al. Development and validation of a measure of dysmorphology: useful for autism subgroup classification. Am J Med Genet A 146a, 1101-16 (2008). 29. Jensen, M., Smolen, C. & Girirajan, S. Gene discoveries in autism are biased towards comorbidity with intellectual disability. J Med Genet 57, 647-652 (2020). 30. Howe, Y.J., Yatchmink, Y., Viscidi, E.W. & Morrow, E.M. Ascertainment and gender in autism spectrum disorders. J Am Acad Child Adolesc Psychiatry 53, 698-700 (2014). 31. Myers, L. et al. Clinical versus automated assessments of morphological variants in twins with and without neurodevelopmental disorders. Am J Med Genet A 182, 1177-1189 (2020). 32. Association, A.P. Diagnostic and statistical manual of mental disorders. 5th Edition. (2013). 33. American Psychiatric Association. Diagnostic and statistical manual of mental disorders (4th ed., Text Revision.), (Washington, DC, 2000). 34. American Psychiatric Association. Diagnostic and statistical manual of mental disorders (5th ed.), (Arlington, VA, 2013). 21 21 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. References Uniparental disomy determined by whole-exome sequencing in a t f t di d t i M l G t G i M d 5 49. Bis, D.M. et al. Uniparental disomy determined by whole-exome sequencing in a spectrum of rare motoneuron diseases and ataxias. Mol Genet Genomic Med 5, 280-286 (2017). ( ) 50. Dolzhenko, E. et al. ExpansionHunter Denovo: a computational method for locating known and novel repeat expansions in short-read sequencing data. Genome Biol 21, 102 (2020). ( ) 51. Trost, B. et al. Genome-wide detection of tandem DNA repeats that are expanded in autism. Nature 586, 80-86 (2020). 52. Wang, K., Li, M. & Hakonarson, H. ANNOVAR: functional annotation of genetic variants from high-throughput sequencing data. Nucleic Acids Res 38, e164 (2010). ( ) 53. Dixon, J.R. et al. Topological domains in mammalian genomes identified by analysis of chromatin interactions. Nature 485, 376-80 (2012). y , ( ) 54. Sloan, C.A. et al. ENCODE data at the ENCODE portal. Nucleic Acids Res 44, D726-32 (2016). y 54. Sloan, C.A. et al. ENCODE data at the ENCODE portal. Nucleic Acids Res 44, D726-32 (2016). 55. Roadmap Epigenomics, C. et al. Integrative analysis of 111 reference human epigenomes. Nature 518, 317-30 (2015). 55. Roadmap Epigenomics, C. et al. Integrative analysis of 111 reference human epigenomes. Nature 518, 317-30 (2015). p g ( ) 56. Blake, J.A. et al. The Mouse Genome Database: integration of and access to knowledge about the laboratory mouse. Nucleic Acids Res 42, D810-7 (2014). 22 22 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: edRxiv preprint 57. Kohler, S. et al. The Human Phenotype Ontology project: linking molecular biology and disease through phenotype data. Nucleic Acids Res 42, D966-74 (2014). ( ) 58. Rehm, H.L. et al. ClinGen--the Clinical Genome Resource. N Engl J Med 372, 2235-42 (2015). 59. Firth, H.V. et al. References DECIPHER: Database of Chromosomal Imbalance and Phenotype in Humans Using Ensembl Resources. Am J Hum Genet 84, 524-33 (2009). ( ) 60. Genomes Project, C. et al. A global reference for human genetic variation. Nature 526, 68-74 (2015). ( ) 61. NHLBI GO Exome Sequencing Project. Exome Variant Server. 62. Lek, M. et al. Analysis of protein-coding genetic variation in 60,706 humans. Nature 536, 285-91 (2016). 63. Karczewski, K.J. et al. The mutational constraint spectrum quantified from variation in 141,456 humans. bioRxiv, 531210 (2020). 64. Ramu, A. et al. DeNovoGear: de novo indel and point mutation discovery and phasing. Nat Methods 10, 985-7 (2013). 65. Drmanac, R. et al. Human genome sequencing using unchained base reads on self-assembling DNA nanoarrays. Science 327, 78-81 (2010). 66. Koressaar, T. et al. Primer3_masker: integrating masking of template sequence with primer design software. Bioinformatics 34, 1937-1938 (2018). p g , ( ) 67. Landrum, M.J. et al. ClinVar: public archive of interpretations of clinically relevant variants. Nucleic Acids Res 44, D862-8 (2016). ( ) 68. Stenson, P.D. et al. Human Gene Mutation Database (HGMD): 2003 update. Hum Mutat 21, 577-81 (2003). 69. Gonzalez-Mantilla, A.J., Moreno-De-Luca, A., Ledbetter, D.H. & Martin, C.L. A Cross-Disorder Method to Identify Novel Candidate Genes for Developmental Brain Disorders. JAMA Psychiatry 73, 275-83 (2016). y y ( ) 70. Solomon, B.D., Nguyen, A.D., Bear, K.A. & Wolfsberg, T.G. Clinical genomic database. Proc Natl Acad Sci U S A 110, 9851-5 (2013). ( ) 71. Ascano, M., Jr. et al. FMRP targets distinct mRNA sequence elements to regulate protein expression. Nature 492, 382-6 (2012). g p p ( ) 72. Bayes, A. et al. Characterization of the proteome, diseases and evolution of the human postsynaptic density. Nat Neurosci 14, 19-21 (2011). p y p y ( ) 73. Cabili, M.N. et al. Integrative annotation of human large intergenic noncoding RNAs reveals global properties and specific subclasses. Genes Dev 25, 1915-27 (2011). ( ) 74. Darnell, J.C. et al. FMRP stalls ribosomal translocation on mRNAs linked to synaptic function and autism. Cell 146, 247-61 (2011). 75. Fabregat, A. et al. Reactome pathway analysis: a high-performance in-memory approach. BMC Bioinformatics 18, 142 (2017). 76. Harris, M.A. et al. The Gene Ontology (GO) database and informatics resource. Nucleic Acids Res 32, D258-61 (2004). ( ) 77. Hawrylycz, M.J. et al. An anatomically comprehensive atlas of the adult human brain transcriptome. Nature 489, 391-399 (2012). References CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: edRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint Figure Legends Figure 1: Project workflow. Figure 1: Project workflow. Summary of phenotype stratification, whole-genome sequencing workflow, and genomic analyses performed in this study. ASD, Autism spectrum disorder; ADM, Autism dysmorphology measure; CNVs, copy number variants; SNVs, single nucleotide variants; indels, insertions and deletions; ERDS, estimation by read depth with single nucleotide variants; GATK-HC, Genome Analysis Toolkit- Haplotype Caller; SNPs, single nucleotide polymorphisms; ACMG, American College of Medical Genetics and Genomics; IQ, Intelligence quotient. Unaffected siblings were used for GRVS and PRS analyses. Excluding samples with false negative ADM-defined nondysmorphic ASD. We also included only samples sequenced on Illumina platforms to be consistent with replication cohort. For variant calling, on average per sample, we detected ~3.7 million SNPs, 36,514 rare single nucleotide variants (SNVs), 4,113 small insertions and deletions (indels), 13 rare copy number variants (CNVs), 390 rare SVs, 73.4 de novo SNVs, 7.3 de novo indels, and 0.1 de novo CNVs (Supplementary Table 2). Experimental validation rates were 94.8%, 85.7%, and 87.5%, respectively, for de novo SNVs, indels, and CNVs (Supplementary Tables 3 and 4). Using GRVS, we were able to quantify and validate the contribution of morphology-associated, rare sequence-level and copy number variants to morphological ASD subtypes. While we can call other SVs from the WGS, there needs to be higher-quality data before these can be effectively incorporated into GRVS. g j Summary of phenotype stratification, whole-genome sequencing workflow, and genomic analyses performed in this study. ASD, Autism spectrum disorder; ADM, Autism dysmorphology measure; CNVs, copy number variants; SNVs, single nucleotide variants; indels, insertions and deletions; ERDS, estimation by read depth with single nucleotide variants; GATK-HC, Genome Analysis Toolkit- Haplotype Caller; SNPs, single nucleotide polymorphisms; ACMG, American College of Medical Genetics and Genomics; IQ, Intelligence quotient. College of Medical Genetics and Genomics; IQ, Intelligence quotient. Unaffected siblings were used for GRVS and PRS analyses. Excluding samples with false negative ADM-defined nondysmorphic ASD. We also included only samples sequenced on Illumina platforms to be consistent with replication cohort. For variant calling, on average per sample, we detected ~3.7 million SNPs, 36,514 rare single nucleotide variants (SNVs), 4,113 small insertions and deletions (indels), 13 rare copy number variants (CNVs), 390 rare SVs, 73.4 de novo SNVs, 7.3 de novo indels, and 0.1 de novo CNVs (Supplementary Table 2). Experimental validation rates were 94.8%, 85.7%, and 87.5%, respectively, for de novo SNVs, indels, and CNVs (Supplementary Tables 3 and 4). Figure Legends Using GRVS, we were able to quantify and validate the contribution of morphology-associated, rare sequence-level and copy number variants to morphological ASD subtypes. While we can call other SVs from the WGS, there needs to be higher-quality data before these can be effectively incorporated into GRVS. Figure 2: Gene sets for which de novo and rare coding variants are significantly more prevalent in some subtypes of ASD. References 78. Kanehisa, M., Sato, Y., Kawashima, M., Furumichi, M. & Tanabe, M. KEGG as a reference resource for gene and protein annotation. Nucleic Acids Res 44, D457- 62 (2016). 79. Schaefer, C.F. et al. PID: the Pathway Interaction Database. Nucleic Acids Res 37, D674-9 (2009). 79. Schaefer, C.F. et al. PID: the Pathway Interaction Database. Nucleic Acids Res 37, D674-9 (2009). 23 23 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint 80. Su, A.I. et al. A gene atlas of the mouse and human protein-encoding transcriptomes. Proc Natl Acad Sci U S A 101, 6062-7 (2004). 81. Uddin, M. et al. Indexing Effects of Copy Number Variation on Genes Involved in Developmental Delay. Sci Rep 6, 28663 (2016). 82. Marshall, C.R. et al. Contribution of copy number variants to schizophrenia from a genome-wide study of 41,321 subjects. Nat Genet 49, 27-35 (2017). 83. Locke, A.E. et al. Genetic studies of body mass index yield new insights for obesity biology. Nature 518, 197-206 (2015). 84. Euesden, J., Lewis, C.M. & O'Reilly, P.F. PRSice: Polygenic Risk Score software. Bioinformatics 31, 1466-8 (2015). 85. Bourgeron, T. From the genetic architecture to synaptic plasticity in autism spectrum disorder. Nat Rev Neurosci 16, 551-63 (2015). y gy , ( ) Euesden, J., Lewis, C.M. & O'Reilly, P.F. PRSice: Polygenic Risk Score ( ) Bourgeron, T. From the genetic architecture to synaptic plasticity in autism spectrum disorder. Nat Rev Neurosci 16, 551-63 (2015). 24 24 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint . Figure 2: Gene sets for which de novo and rare coding variants are significantly more prevalent in some subtypes of ASD. g y p yp We define events as (a,b) genes impacted by CNVs or (c) as variants for SNVs and indels. The coefficient is the relationship between the number of events in each gene set and the ASD subtypes; it reflects the effect size of a variant type and gene set among different ASD subtypes. Positive coefficients indicate more events in individuals with ASD and more dysmorphic features; negative coefficients indicate more events in individuals with ASD and fewer dysmorphic features. We show only gene sets for which a,b) rare CNVs, or c) de novo missense variants are significantly more prevalent in different subtypes of ASD. Tier 1 and 2 missense variants consist of all or only predicted damaging missense variants, respectively, as defined in Yuen, et al.23 Symbol shapes indicate the subtype comparisons that were conducted for each combination of gene set and variant type. Two subtype comparison= nondysmorphic vs. dysmorphic ASD. Three subtype comparison= essential vs. equivocal vs. complex ASD. Coloured shapes indicate significant signals after multiple test correction by permutation-based FDR. Error bars indicate 95% confidence intervals. 25 25 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: edRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint Figure 3: Genome-wide rare variant score in ASD subtypes. Figure 3: Genome-wide rare variant score in ASD subtypes. Events are comprised of variants for SNVs and indels or genes impacted by CNVs. For each sample, the GRVS is the sum of rare and de novo events in morphology-associated regions, weighted by effect size (estimated from the coefficients in the regression model). GRVSs were generated 30 times for each sample (see methods), yielding an average score and average number of variants. CSVs, clinically significant variants. a) Distribution of standardized GRVS for the discovery cohort (n=235). b) GRVSs for the whole cohort (left plot, n=235) or the whole cohort excluding the 17 probands with clinically significant variants (right plot, n=218), were ordered and ranked by percentile. Note that while 46 probands in the discovery cohort (n=325) had CVSs, only 17 of them had two sequenced parents meeting inclusion criterion for the GRVS group (n=235). Violin plots show the distributions of the samples’ GRVS percentiles; box plots contained within show the median and quartiles of the percentiles for each subtype. P values denote the probability that the GRVS in dysmorphic ASD is not greater than nondysmorphic ASD (one-sided, Wilcoxon rank sum test). c) Rare variants have different effect sizes. The mean coefficient reflects the effect size of a variant type. Coefficients of deletions and duplications of the same size bin were averaged together. Coefficients of predicted LoF variants, missense variants, and predicted damaging missense variants were averaged together. Error bars indicate mean ± standard deviation. The number of morphology- associated regions for each variant type is indicated the y-axis with “n= ”. 26 26 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: edRxiv preprint Figure 5: Relationship between IQ, morphological ASD subtypes and genetic variants. a) Comparison of IQ among morphological ASD subtypes. b) Comparison of IQ between probands with and without a CSV. a,b) Violin plots show the distributions of the probands’ IQ; box plots contained within show the median and quartiles of IQ for each subtype. P values denote the probability that the mean IQ of nondysmorphic ASD or probands without CSVs is not greater than dysmorphic ASD or probands with CSVs, respectively (one-sided, t-test). Correlation between IQ and c) GRVS and d) PRS is shown. c,d) Each dot represents the IQ and GRVS or PRS percentile of a sample. The linear regression line indicates the linear correlation between IQ and GRVS or PRS percentiles. Correlation coefficient is quantified by Spearman’s rho correlation. P values indicate the probability that the correlation is occurred due to chance. Figure 6: Replication of rare and common genetic findings in subset of Simons Simplex Collection cohort. Figure 4: Inheritance of polygenic risk for ASD and BMI in morphologic subtypes. Differences in polygenic risk score (PRS) for ASD and BMI between subjects and their respective mid-parent score. Box plots depict the median and quartiles of polygenic transmission disequilibrium test (pTDT) deviation. Dots represent pTDT deviations of subjects. P values for each subgroup indicate the probability that the mean of the pTDT deviation distribution is not greater than zero (one- sided, t-test), as depicted by the dotted line. Figure 6: Replication of rare and common genetic findings in subset of Simons Simplex Collection cohort. a) GRVSs for each cohort were ordered and ranked by percentile. Violin plots show the distributions of the probands’ GRVS percentiles; box plots contained within show the median and quartiles of the percentiles for each subtype. P values denote the probability that the GRVS in ADM-defined dysmorphic ASD is not greater than ADM-defined nondysmorphic ASD (one-sided, Wilcoxon rank sum test). b) Yield of CSVs between dysmorphic and nondysmorphic subtypes in discovery and replication cohorts. P values indicate the probability that the yield of CSVs in nondysmorphic ASD is not lower than that of dysmorphic ASD (one- sided, t-test). c) Inheritance of polygenic risk for ASD in dysmorphic and nondysmorphic ASD subtypes in discovery and replication cohorts. Box plots depict the median and quartiles of pTDT deviation. Dots represent pTDT deviations of subjects. P values for each subgroup indicate the probability that the mean of the pTDT deviation distribution is not greater than zero (one-sided, t- test), as depicted by the dotted line. The finding of no significant over- transmission in dysmorphic ASD did not replicate in SSC, which might be due to lack of statistical power (i.e., at least 100 dysmorphic samples are needed to achieve 80% power if PRS explains 2.45% of phenotypic variance11) and/or ascertainment differences between the discovery and replication cohorts. The discovery cohort included data about major congenital anomalies in morphologic classification, whereas the replication cohort did not. While our discovery cohort was population-based, the Simons Simplex Collection excluded probands with 27 27 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity (which was not certified by peer review) The copyright holder for this pre this version posted October 26, 2021. Figure 6: Replication of rare and common genetic findings in subset of Simons Simplex Collection cohort. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint medically significant perinatal diseases, severe neurological deficits, and certain genetic syndromes27. This likely decreased the proportion of probands with excess MPAs and birth defects, potentially leading to a lower burden of common ASD-associated variants85. 28 28 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: edRxiv preprint Figure 1: Project workflow. Figure 6: Replication of rare and common genetic findings in subset of Simons Simplex Collection cohort. Variant Calling • SNV & indels (DenovoGear & GATK-HC) • CNVs (CNVnator & ERDS) • Structural variants (Manta) • Anneuploidies (in-house script) • Uniparental isodisomy (in-house script) • Mitochondrial variants (in-house script) • Tandem repeat expansions (Expansion Hunder denovo) PHENOTYPE STRATIFICATION GENOMIC ANALYSIS Identify clinically significant variants using ACMG guidelines for variant interpretation (all rare & de novo variants) Gene-set based rare variant score (rare & de novo SNVs, indels, CNVs) Polygenic risk score and polygenic transmission disequilibrium test (common SNPs) Relationship between genetic findings and IQ 1.Burden analysis 2.Gene-set/noncoding region enrichment analysis Discovery cohort: Population- based recruitment of 325 cases Clinical assessment of minor physical anomalies (MPAs) and major congenital anomalies (MCAs) Essential ASD (n=187) Equivocal ASD (n=57) Complex ASD (n=81) Whole genome sequencing (WGS) Gold standard dysmorphology (MPA+MCA score) ADM-defined non-dysmorphic ASD (n=203) ADM-defined dysmorphic ASD (n=73)** ADM Alignment Variant Calling • SNV & indels (DenovoGear & GATK-HC) • CNVs (CNVnator & ERDS) ADM on affected probands ADM-defined nondysmorphic ASD (n=400) ADM-defined dysmorphic ASD (n=42) Unaffected siblings (n=355) CRAM files from WGS Replication cohort: Subset of cases from Simons Simplex Collection (N= 442 affected probands and 355 unaffected siblings) PHENOTYPE STRATIFICATION Replication cohort: Subset of cases from Simons Simplex Collection (N= 442 affected probands and 355 unaffected siblings) Discovery cohort: Population- based recruitment of 325 cases Clinical assessment of minor physical anomalies (MPAs) and major congenital anomalies (MCAs) ADM on affected probands Gold standard dysmorphology (MPA+MCA score) ADM-defined nondysmorphic ASD (n=400) ADM-defined dysmorphic ASD (n=42) Unaffected siblings (n=355) Variant Calling • SNV & indels (DenovoGear & GATK-HC) • CNVs (CNVnator & ERDS) Polygenic risk score and polygenic transmission disequilibrium test (common SNPs) Identify clinically significant variants using ACMG guidelines for variant interpretation (all rare & de novo variants) 1.Burden analysis 2.Gene-set/noncoding region enrichment analysis Gene-set based rare variant score (rare & de novo SNVs, indels, CNVs) Figure 1: Project workflow. Relationship between genetic findings and IQ Relationship between ge findings and IQ 29 29 Figure 2: Gene sets for which de novo and rare variants are significantly more prevalent in some subtypes of ASD. Figure 2: Gene sets for which de novo and rare varian ne sets for which de novo and rare variants are significantly more prevalent in some subtypes of Figure 2: Gene sets for which de novo and rare variants are significantly more prevalent in some subtypes of ASD. 30 . Figure 6: Replication of rare and common genetic findings in subset of Simons Simplex Collection cohort. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: edRxiv preprint Figure 3: Genome-wide rare variant score in ASD subtypes. Figure 3: Genome-wide rare variant score in ASD subtypes. Figure 3: Genome-wide rare variant score in ASD subtypes. 31 31 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: edRxiv preprint Figure 4: Inheritance of polygenic risk for ASD and BMI in morphologic ASD subtypes. Figure 4: Inheritance of polygenic risk for ASD and BMI in morphologic ASD subtypes. 32 32 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: edRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint Figure 5: Relationship between IQ, morphological ASD subtypes and genetic variants. Figure 5: Relationship between IQ, morphological ASD subtypes and genetic variants. 33 33 . Figure 6: Replication of rare and common genetic findings in subset of Simons Simplex Collection cohort. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: edRxiv preprint . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.20.21264950 doi: medRxiv preprint Figure 6: Replication of common and rare genetic findings in subset of Simons Simplex Collection cohort. Figure 6: Replication of common and rare genetic findings in subset of Simons Simplex Collection cohort. 34
https://openalex.org/W3183406834	https://pub.uni-bielefeld.de/download/2956361/2956364/fsurg-08-681797.pdf	English	null	Intraluminal Monitoring of Micro Vessels. A Surgical Feasibility Study	Frontiers in surgery	2,021	cc-by	3,380	ORIGINAL RESEARCH published: 21 July 2021 doi: 10.3389/fsurg.2021.681797 Edited by: Pavel Dulguerov, Geneva University Hospitals (HUG), Switzerland Results: In all cases, surgical placement was uneventful. Measurement of intra-venous and intra-arterial pressure was possible and stable over the whole measurement period of an hour. Reviewed by: Maria Grosheva, University of Cologne, Germany Anastasios Kanatas, Leeds Teaching Hospitals NHS Trust, United Kingdom Reviewed by: Maria Grosheva, University of Cologne, Germany Anastasios Kanatas, Leeds Teaching Hospitals NHS Trust, United Kingdom Conclusion: Fiber optic pressure measurement in microvessels is possible and surgically feasible. An application to monitor the perfusion of free ﬂaps seems possible. Correspondence: Ingo Todt todt@gmx.net Correspondence: Ingo Todt todt@gmx.net Keywords: microvessel, pressure monitoring, ﬁber optic pressure measurement, anastomosis, ﬂap monitoring device Intraluminal Monitoring of Micro Vessels. A Surgical Feasibility Study Leonard Walle 1, Holger Sudhoff 2, Onno Frerichs 1 and Ingo Todt 2* 1 Department of Otolaryngology, Head and Neck Surgery, Medical School OWL, Bielefeld University, Bielefeld, Germany, 2 Department of Plastic Surgery, Medical School OWL, Bielefeld University, Bielefeld, Germany Leonard Walle 1, Holger Sudhoff 2, Onno Frerichs 1 and Ingo Todt 2 Objective: Monitoring of vessel perfusion is of high clinical importance in vascular anastomosis of free ﬂaps. Current sensor systems are based on different principles and show limitations in validity and accuracy. Fiber optic pressure sensors exhibit high accuracy and are small in size. The aim of the present study was to evaluate the surgical feasibility of intraluminal pressure (ILP) measurements with a ﬁber optic pressure sensor in an animal model. Methods: In a microsurgical setting we sedated 10 Wistar rats with weight adapted phenobarbital, xylazine, and fentanyl. We performed a surgical approach to A. carotis communis and V. jugularis and introduced a 600 µm ﬁber optic pressure sensor into the vessels followed by measuring the ILP. The sensor was stabilized by the surrounding tissue, and the vessels were closed. INTRODUCTION Specialty section: This article was submitted to Otorhinolaryngology - Head and Neck Surgery, a section of the journal Frontiers in Surgery Monitoring of vessel pressure is of central importance in various clinical ﬁelds, such as the evaluation of vital parameters in patients at intensive care units. This is often performed using intraluminal catheters. Thrombosis in conjunction with these catheters is rare (1, 2). In another clinical ﬁeld, vessel monitoring is performed by plastic, maxillofacial, or otolaryngological surgeons in conjunction with micro-anastomosis and evaluation of the perfusion of diﬀerent types of microsurgery ﬂaps. The monitoring is mainly applied on the venous side of the responsible vessels (3). Received: 17 March 2021 Accepted: 28 June 2021 Published: 21 July 2021 Citation: Conventional free ﬂap monitoring techniques require an external component, whereas an implantable monitor readily indicates changes in free ﬂap perfusion, especially in buried ﬂaps used in head and neck reconstruction. Therefore, diﬀerent approaches have been described (4). Walle L, Sudhoff H, Frerichs O and Todt I (2021) Intraluminal Monitoring of Micro Vessels. A Surgical Feasibility Study. Front. Surg. 8:681797. doi: 10.3389/fsurg.2021.681797 In fact, two implantable monitoring systems are in clinical use. The ﬁrst is based on a cough electrode, which circumvents the venous vessel (Cook Swartz doppler probe, Cook Medical, July 2021 \| Volume 8 \| Article 681797 Frontiers in Surgery \| www.frontiersin.org 1 Pressure Monitoring in Microvessels Walle et al. measurements at values of ±300 mmHg. The size of the conducting cable is 155 µm, and therefore, it can even be used for conduction out of a very small vessel. The small size and high accuracy of the sensor make it a possible candidate for intraluminal monitoring of vessel pressure. Theoretically, such a system would combine high accuracy related to the intraluminal position and high reliability against dislocations of the sensor itself related to the intraluminal position. Figure 1 is an exemplary picture of a pressure sensor (200 µm). Bloomington, USA). The technological principle of this system is based on a Doppler sensor with an extraluminal placement. It is surgically placed during the suturing of the veins and contains a cable for an external monitoring unit. After a few days, the cable is cut and persists in the local area (5). Diﬀerent studies have shown that this is a reliable technique for postoperative monitoring (6). Nevertheless, problems were observed related to the removal of the cable. Even dislocated electrodes with failure of blood ﬂow measurement have been described (7). Bloomington, USA). The technological principle of this system is based on a Doppler sensor with an extraluminal placement. It is surgically placed during the suturing of the veins and contains a cable for an external monitoring unit. After a few days, the cable is cut and persists in the local area (5). Diﬀerent studies have shown that this is a reliable technique for postoperative monitoring (6). Nevertheless, problems were observed related to the removal of the cable. Even dislocated electrodes with failure of blood ﬂow measurement have been described (7). Animals Five female rats and ﬁve male Wistar rats (482–616 g) were sedated with a weight-adapted amount of phenobarbital, xylazine, and fentanyl. Then, a surgical approach to the A. carotis communis and V. jugularis was performed on the left side. All experiments were covered by the allowance of the state oﬃce (LANUV; ﬁle number: 84–02.05.40.16.035). Fiber optic sensors are highly accurate and sizeable down to 0.2 mm/ 0.8 Char (FISO, Quebec, Canada). The physical principle of measurements is based on a Fabry- Pérot chamber with a ﬂexible membrane on top and allows Citation: The second concept is based on an ultrasonic Doppler sensor placed inside a coupler device for end-to-end microsurgical venous anastomosis (Flow coupler, Synovis, Birmingham, USA). The coupler system is intended to detect blood ﬂow and conﬁrm vessel patency intra- and postoperatively at the anastomotic site. The results with this device show a variability in outcome, and these devices are expensive. A comparative study between these two systems observed no signiﬁcant diﬀerence in outcome (8). The aim of the present study was to evaluate, the surgical feasibility of handling, introducing, and using a ﬁber optic sensor for an intraluminal pressure (ILP) measurement in an animal model of anastomosis. Sensor The ILP was measured using a micro-optical pressure sensor, FOP- MIV (FISO, Quebec, Canada). The tip of the pressure sensor (600 µm) is a hollow glass tube sealed on one end by a thin plastic ﬁlm diaphragm coated with a reﬂective surface of evaporated gold. The optical ﬁber is located in the glass tube, with a short distance (50–100 µm) to the diaphragm tip. The optical ﬁber is attached to an LED light source and a photodiode sensor. Light from the LED source reaches the sensor tip of the optical ﬁber, fans out as it exits the ﬁber, and is reﬂected by the gold-covered ﬂexible diaphragm. The photodiode senses the reﬂected light, and small pressure-induced distance displacements of the diaphragm modulate the intensity of the reﬂected light. The sensor is connected to a module that is linked to a computer. Evolution software was used to record the ILP. The time sensitivity of the sensor was 300 measurements per second. For a better understanding Figure 2 shows the set up. Surgical Procedure FIGURE 1 \| Fiber optic pressure sensor 200 µm (Courtesy of FISO). g According to Simsek et al. (9) the rats were placed in a supine position. The legs were ﬁxed by staples. The head was freely moveable and a surgical access to the left neck was performed. Surgical clips were placed on the A. carotis communis and V. jugularis, and a slot was made in the vessel wall with an 11-blade scalpel (Figure 4). The microsensor was introduced (Figure 5), and the vessel was closed with Ethilon 10-0 (Figure 6). Finally, a drop of ﬁbrin glue was placed on the entrance of the cable. Finally, the clips were removed. The cable was additionally sutured in the surrounding tissue to prevent dislocation. After an hour, we conducted the milking test to evaluate the patency of the vein and arteria distal to the anastomosis (10). Anticoagulant medication was not used. FIGURE 1 \| Fiber optic pressure sensor 200 µm (Courtesy of FISO). FIGURE 2 \| Schematic description of experimental set up. Sensor is introduced in the animal region of interest. Light signal is transferred to the controller and translated into an electrical signal for the laptop. FIGURE 2 \| Schematic description of experimental set up. Sensor is introduced in the animal region of interest. Light signal is transferred to the controller and translated into an electrical signal for the laptop. July 2021 \| Volume 8 \| Article 681797 Frontiers in Surgery \| www.frontiersin.org 2 2 Walle et al. Pressure Monitoring in Microvessels FIGURE 3 \| Exemplary arterial pressure signal (A. Carotis) generated by a FOPS x-axis: mmHg, y-axis: time. FIGURE 4 \| Exemplary venous pressure signal (V. Jugularis) generated by a FOPS x-axis: mmHg, y-axis: time. FIGURE 4 \| Exemplary venous pressure signal (V. Jugularis) generated by a FOPS x-axis: mmHg, y-axis: time. FIGURE 3 \| Exemplary arterial pressure signal (A. Carotis) generated by a FOPS x-axis: mmHg, y-axis: time. FIGURE 4 \| Exemplary venous pressure signal (V. Jugularis) generated by a FOPS x-axis: mmHg, y-axis: time. FIGURE 3 \| Exemplary arterial pressure signal (A. Carotis) generated by a FOPS x-axis: mmHg, y-axis: time. FIGURE 5 \| Intraoperative situs. Blue clips are visible, Ring indicates vessel. Arrow indicates slot. Opening of the vessel. RESULTS FIGURE 5 \| Intraoperative situs. Blue clips are visible, Ring indicates vessel. Arrow indicates slot. Opening of the vessel. In all cases, introduction of the sensor and closure of the vessel were uneventful and without any leakage around the sensor access. Over the period of measurements, no decrease in the pressure signal was observed indicating a clotting or thrombus around the sensor. Measured intraluminal mean maximum arterial amplitude was 13.6 mmHg SD 1.9; mean venous amplitude was 5.2 mmHg SD 3.3 (Table 1), with a ratio of 2.6. Exemplary arterial and venous measurements are shown in Figures 3, 4. Figures 5–7 shows the surgical procedure of introduction of the 600 µm ﬁber optic pressure sensor (FOPS). Frontiers in Surgery \| www.frontiersin.org July 2021 \| Volume 8 \| Article 681797 3 Walle et al. Pressure Monitoring in Microvessels FIGURE 6 \| Intraoperative situs. Arrow indicates sensor and cable, Ring indicates vessel. Insertion of the sensor. FIGURE 7 \| Intraoperative situs. Arrow indicates entrance of the sensor into the vessel, Ring indicates vessel. Sutured vessel with inserted sensor. FIGURE 6 \| Intraoperative situs. Arrow indicates sensor and cable, Ring indicates vessel. Insertion of the sensor. FIGURE 6 \| Intraoperative situs. Arrow indicates sensor and cable, Ring indicates vessel. Insertion of the sensor. FIGURE 6 \| Intraoperative situs. Arrow indicates sensor and cable, Ring indicates vessel. Insertion of the sensor. FIGURE 7 \| Intraoperative situs. Arrow indicates entrance of the sensor into the vessel, Ring indicates vessel. Sutured vessel with inserted sensor. FIGURE 7 \| Intraoperative situs. Arrow indicates entrance of the sensor into the vessel, Ring indicates vessel. Sutured vessel with inserted sensor. FIGURE 7 \| Intraoperative situs. Arrow indicates entrance of the sensor into the vessel, Ring indicates vessel. Sutured vessel with inserted sensor. TABLE 1 \| Individual maximum pressure amplitude measured intraluminally by ﬁber optic pressure sensoring. v1 v2 v3 v4 v5 V6 v7 v8 v9 v10 ARTERIAL mmHg 15 16 14 15 15 10 12 12 15 12 VENOUS mmHg 2.3 1.8 2 2.1 1.9 7.5 9 8 9 8 V stands for each animal and the according A. carotis and V. jugularis measurement. revision can save the ﬂap in some cases. Therefore, monitoring is of high importance. Common monitoring systems, in contrast to visual control of the ﬂap, have the advantage of providing a timely alert before substantial damage to the ﬂap may occur (11). RESULTS Current monitoring systems have the disadvantage of a certain degree of false negative or false positive results in terms of alerting the surgeon (12), leading to a search for diﬀerent systems to improve the monitoring (13, 14). Therefore, systems with a higher accuracy are of high interest. Accuracy is determined by more exact measurement values and a lower degree of sensor dislocation. FOPSs are a newer form to observe pressure changes based on the principle of a Fabry-Pérot interferometer with a ﬂexible embodiment (15). So far, they have been used in experimental settings to evaluate pressure changes in cardiology (16), neurosurgery (17), and otology (18), as well as for the evaluation of inner ear pressure changes (19). During our measurement periods (1 h), a stable signal was evaluated, indicating no event of clotting or thrombosis. After an hour, the milking test showed no abnormalities. REFERENCES 9. Sim¸sek T, Eroglu L, Engin MS, Kaplan S, Yildiz L. End-to-end microvascular anastomosis in the rat carotid artery using continuous horizontal mattress sutures. J Reconstr Microsurg. (2006) 22:631–40. doi: 10.1055/s-2006-9 56237 1. Oh H, Choe SH, Kim YJ, Yoon HK, Lee HC, Park HP. Intraarterial catheter diameter and dynamic response of arterial pressure monitoring system: a randomized controlled trial. J Clin Monit Comput. (2021). doi: 10.1007/s10877-021-00663-7. [Epub ahead of print]. 10. Podeur F, Peyrachon B, Nokovitch L, Dammacco MA, Qassemyar Q, Deneuve S. Predictive value of the milking patency test when performing the arterial microanastomosis in head and neck surgery. Head Neck. (2019) 41:3328–33. doi: 10.1002/hed. 25841 2. Wicky S, Pinto EG, Oklu R. Catheter-directed thrombolysis of arterial thrombosis. Semin Thromb Hemost. (2013) 39:441– 5. doi: 10.1055/s-0033-1334482 3. Klifto KM, Milek D, Gurno CF, Seal SM, Hultman CS, Rosson GD, et al. Comparison of arterial and venous implantable Doppler postoperative monitoring of free ﬂaps: systematic review and meta-analysis of diagnostic test accuracy. Microsurgery. (2020) 40:501–11. doi: 10.1002/micr.30564 11. Rozen WM, Chubb D, Whitaker IS, Acosta R. The eﬃcacy of postoperative monitoring: a single surgeon comparison of clinical monitoring and the implantable Doppler probe in 547 consecutive free ﬂaps. Microsurgery. (2010) 30:105–10. doi: 10.1002/micr.20706 4. Abdel-Galil K, Mitchell D. Postoperative monitoring of microsurgical free- tissue transfers for head and neck reconstruction: a systematic review of current techniques–part II. Invasive techniques. Br J Oral Maxillofac Surg. (2009) 47:438–42. doi: 10.1016/j.bjoms.2008.12.002 12. Lineaweaver W. Techniques of monitoring buried fasciocutaneous free ﬂaps. Plast Reconstr Surg. (2009) 124:1729–31. doi: 10.1097/PRS.0b013e3181b 98d34 13. Rosenberg JJ, Fornage BD, Chevray PM. Monitoring buried free ﬂaps: limitations of the implantable Doppler and use of color duplex sonography as a conﬁrmatory test. Plast Reconstr Surg. (2006) 118:109–13. doi: 10.1097/01.prs.0000221113.78244.8c 5. Jones NF. Postoperative monitoring of microsurgical free tissue transfers for head and neck reconstruction. Microsurgery. (1988) 9:159–64. doi: 10.1002/micr.1920090219 14. Hölzle F, Loeﬀelbein DJ, Nolte D, WolﬀKD. Free ﬂap monitoring using simultaneous non-invasive laser Doppler ﬂowmetry and tissue spectrophotometry. J Craniomaxillofac Surg. (2006) 34:25–33. doi: 10.1016/j.jcms.2005.07.010 6. Schmulder A, Gur E, Zaretski A. Eight-year experience of the Cook-Swartz Doppler in free-ﬂap operations: microsurgical and reexploration results with regard to a wide spectrum of surgeries. Microsurgery. (2011) 31:1– 6. doi: 10.1002/micr.20816 15. Poeggel S, Tosi D, Duraibabu D, Leen G, McGrath D, Lewis E. Optical ﬁbre pressure sensors in medical applications. Sensors. (2015) 15:17115– 48. doi: 10.3390/s150717115 7. ETHICS STATEMENT Although not observed in our measurements, clotting could not be fully excluded. Further observations in a long-term measurement interval would be of high importance. The animal study was reviewed and approved by LANUV. DISCUSSION An application in the ﬁeld of anastomosis to evaluate microsurgical quality control has not been performed to the best of our knowledge. Monitoring of vascular anastomosis is a helpful tool to control the perfusion of free ﬂaps. In cases of decreased perfusion, a timely July 2021 \| Volume 8 \| Article 681797 Frontiers in Surgery \| www.frontiersin.org 4 Pressure Monitoring in Microvessels Walle et al. with central venous catheters. These options are a matter for further evaluation. Related to its small size and high accuracy, an intraluminal measurement, which eliminates the problem of dislocation of external sensors, seems possible. CONCLUSION Our experiments showed that with microsurgical skills, handling of the ﬁber optic sensor is easy to perform and was uneventful. Pressure measurements could be performed without any complications. Fiber optic pressure measurement in microvessels is possible and surgically feasible. An application to monitor the perfusion of free ﬂaps seems possible. One limitation of the study was the size of the sensor. The higher dimension of our sensor in contrast to the possible dimensions that could be produced make its use even easier. AUTHOR CONTRIBUTIONS Surface coating for clotting prevention might solve this theoretical problem (20). Handling of the sensor after the successful control of a persisting anastomosis is a second issue. From our point of view, two options might be chosen: (1). The cable is cut, and the small sensor stays in the vessel; (6) the sensor is pulled out. A 0.2 mm lesion caused by the pullout should be closed immediately as known from the clinical routine LW and IT consultant, idea, surgeon, and writing. HS head of Bielefeld ENT and co-writing. OF head of Bielefeld Plastic Surgery and co-writing. All authors contributed to the article and approved the submitted version. DATA AVAILABILITY STATEMENT Further limitations is that this study is at an very early stage. Further comprehensive animal studies are necessary before a human application can be performed. The important limitations are e.g., the sensors are not approved for human intraoperative use and the ﬁxation of the sensor is an unsolved problem as well as a long term observation is needed. The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author/s. 19. Mittmann P, Ernst A, Mittmann M, Todt I. Optimisation of the round window opening in cochlear implant surgery in wet and dry conditions: impact on intracochlear pressure changes. Eur Arch Otorhinolaryngol. (2016) 273:3609–13. doi: 10.1007/s00405-016-3993-6 Frontiers in Surgery \| www.frontiersin.org 18. SudhoﬀH, Mittmann P, Todt I. In vivo measurement of middle ear pressure changes during balloon Eustachian tuboplasty. Biomed Res Int. (2018) 2018:9519204. doi: 10.1155/2018/95 19204 July 2021 \| Volume 8 \| Article 681797 20. Tchouta LN, Bonde PN. The quest for nonthrombotic surface modiﬁcations to achieve hemocompatibility of implantable devices. ASAIO J. (2015) 61:623– 34. doi: 10.1097/MAT.0000000000000283 REFERENCES Paprottka FJ, Klimas D, Krezdorn N, Schlarb D, Trevatt AEJ, Hebebrand D. Cook-Swartz Doppler probe surveillance for free ﬂaps-deﬁning pros and cons. Surg J. (2020) 6:e42–e6. doi: 10.1055/s-0040-17 02922 16. Konieczny G, Opilski Z, Pustelny T, Maciak E. Fiber pressure sensor for the Polvad prosthesis. Mol Quantum Acoust. (2008) 29:135–41 8. Um GT, Chang J, Louie O, Colohan SM, Said HK, Neligan PC, et al. Implantable Cook-Swartz Doppler probe versus Synovis Flow Coupler for the post-operative monitoring of free ﬂap breast reconstruction. J Plast Reconstr Aesthet Surg. (2014) 67:960–6. doi: 10.1016/j.bjps.2014.03.034 17. Chavko M, Koller WA, Prusaczyk WK, McCarron RM. Measurement of blast wave by a miniature ﬁber optic pressure transducer in the rat brain. J Neurosci Methods. (2007) 159:277–81. doi: 10.1016/j.jneumeth.2006.07.018 July 2021 \| Volume 8 \| Article 681797 Frontiers in Surgery \| www.frontiersin.org 5 Walle et al. Pressure Monitoring in Microvessels 18. SudhoﬀH, Mittmann P, Todt I. In vivo measurement of middle ear pressure changes during balloon Eustachian tuboplasty. Biomed Res Int. (2018) 2018:9519204. doi: 10.1155/2018/95 19204 Conﬂict of Interest: The authors declare that the research was conducted in the absence of any commercial or ﬁnancial relationships that could be construed as a potential conﬂict of interest. 19. Mittmann P, Ernst A, Mittmann M, Todt I. Optimisation of the round window opening in cochlear implant surgery in wet and dry conditions: impact on intracochlear pressure changes. Eur Arch Otorhinolaryngol. (2016) 273:3609–13. doi: 10.1007/s00405-016-3993-6 Copyright © 2021 Walle, Sudhoﬀ, Frerichs and Todt. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. 20. Tchouta LN, Bonde PN. The quest for nonthrombotic surface modiﬁcations to achieve hemocompatibility of implantable devices. ASAIO J. (2015) 61:623– 34. doi: 10.1097/MAT.0000000000000283 20. Tchouta LN, Bonde PN. The quest for nonthrombotic surface modiﬁcations to achieve hemocompatibility of implantable devices. ASAIO J. (2015) 61:623– 34. doi: 10.1097/MAT.0000000000000283 July 2021 \| Volume 8 \| Article 681797 Frontiers in Surgery \| www.frontiersin.org 6
https://openalex.org/W1912151326	https://www.scielo.br/j/pab/a/B8cZV8yrTgkrwpqHMLmLmhG/?lang=pt&format=pdf	Portuguese	null	Herdabilidade e correlação fenotípica de caracteres relacionadosà produtividade de grãos e à morfologia da canola	Pesquisa Agropecuária Brasileira	2,011	cc-by	5,214	Heritability and phenotypic correlation of characters related to grain yield and morphology of canola Abstract – The objective of this work was to estimate the heritability and to evaluate the phenotypic correlations of canola (Brassica napus) traits related to grain yield components and plant architecture. Three experiments were carried out using 0.20, 0.40 and 0.60 m row spacing. During two crop years (2008 and 2009), the genotypes 'Hyola 432' and 'Hyola 61' were evaluated at plant densities of 20, 40, 60 and 80 plants per square meter in each experiment. A complete block experimental design was used in a 2x2x4 factorial arrangement (years x genotypes x densities), with four replicates. Evaluations were done for components linked to production (grain yield per area and per plant, number of pods per plant, number of grains per pod and per plant, and pod mass), and to the morphology of canola (pod length, number of secondary branches, height of insertion of the secondary branch, branch length and number of tertiary branches). Grain yield per area and per plant have higher heritability in the narrower row spacing. The number of pods and of seeds per plant are the components with higher direct and positive correlation with yield. Index terms: Brassica napus, morphological traits, yield components, planting density, row spacing. Herdabilidade e correlação fenotípica de caracteres relacionados à produtividade de grãos e à morfologia da canola Cleusa Adriane Menegassi Bianchi Krüger(1), José Antonio Gonzalez da Silva(1), Sandro Luis Petter Medeiros (2), Genei Antonio Dalmago(3) e Juliano Gaviraghi(1) 1)Universidade Regional do Noroeste do Estado do Rio Grande do Sul, Departamento de Estudos Agrários, Rua do Comércio, no 3.000, CEP 98700‑000 Ijuí, RS. E‑mail: cleusa_bianchi@yahoo.com.br, jagsfaem@yahoo.com.br, gaviraghi_juli@hotmail.com 2)Universidade Federal de Santa Maria, Centro de Ciências Rurais, Departamento de Fitotecnia, CEP 97105‑900 Santa Maria, RS. E‑mail: slpmedeiros@yahoo.com.br (3)Embrapa Trigo, Caixa Postal 451, CEP 99001‑970 Passo Fundo, RS. E‑mail: dalmago@cnpt.embrapa.br Resumo – O objetivo deste trabalho foi estimar a herdabilidade e avaliar a correlação fenotípica entre caracteres de canola (Brassica napus) relacionados à produtividade de grãos e à arquitetura de plantas. Foram realizados três experimentos, com espaçamento entre linhas de 0,20, 0,40 e 0,60 m. Durante dois anos de cultivo (2008 e 2009), os genótipos 'Hyola 432' e 'Hyola 61' foram avaliados em quatro densidades de plantio (20, 40, 60 e 80 plantas por metro quadrado), em cada experimento. Empregou-se o delineamento de blocos ao acaso, em arranjo fatorial 2x2x4 (anos x genótipos x densidades), com quatro repetições. Foram avaliados componentes ligados à produção (produtividade de grãos por área e por planta, número de síliquas por planta, número de grãos por síliqua e por planta, e massa de síliqua) e à morfologia da canola (comprimento de síliqua, número de ramos secundários, altura de inserção do ramo secundário, comprimento de ramo e número de ramos terciários). A produtividade de grãos por área e por planta apresenta maior herdabilidade no menor espaçamento entre linhas. O número de síliquas e o de grãos por planta são os componentes de produção com maior correlação direta e positiva com a produtividade de grãos. Termos para indexação: Brassica napus, caracteres morfológicos, componentes de produção, densidade de plantio, espaçamento entre linhas. Termos para indexação: Brassica napus, caracteres morfológicos, componentes de produção, densidade de plantio, espaçamento entre linhas. Pesq. agropec. bras., Brasília, v.46, n.12, p.1625-1632, dez. 2011 C.A.M.B. Krüger et al. C.A.M.B. Krüger et al. Em estudos que visam incrementar a produtividade de grãos, é necessário observar a expressão dos componentes de produção e a relação entre eles, bem como estimar os parâmetros genéticos responsáveis pela manifestação desses caracteres (Carvalho et al., 2001). Componentes diretos e indiretos de produção têm sido estudados em diversas culturas como canola, aveia, girassol e soja (Tourino et al., 2002; Coimbra et al., 2004; Hartwig et al., 2007; Amorin et al., 2008). Esses trabalhos visam determinar quais componentes são mais sensíveis a variações nas condições ambientais. Nesse contexto, é ainda necessário determinar os efeitos da redução do espaçamento entre linhas recomendado para canola (Thomas, 2003; Shahin & Valiollah, 2009), sobre os componentes de produção. Agrários, da Universidade Regional do Noroeste do Estado do Rio Grande do Sul (Unijuí), no Município de Augusto Pestana, RS, a 28º26'30"S, 54º00'58"W e à altitude média de 298 m. O solo da área experimental é classificado como Latossolo Vermelho distroférrico típico (Santos et al., 2006). O clima da região, segundo classificação de Köppen, é do tipo Cfa, com verão quente sem estação seca. q O trabalho envolveu a realização de três experimentos, com três espaçamentos entre linhas: 0,20, 0,40 e 0,60 m. Em cada experimento, empregou- se o delineamento de blocos ao acaso, com quatro repetições, em arranjo fatorial 2x2x4. Os fatores de variação considerados foram: ano de cultivo (2008 e 2009), genótipo ('Hyola 432' e 'Hyola 61' ) e densidade de plantas (20, 40, 60 e 80 plantas por metro quadrado). A unidade experimental foi constituída por cinco linhas de cinco metros de comprimento. Portanto, a dimensão das parcelas experimentais foi alterada de acordo com o espaçamento entre linhas adotado, com parcelas de 5 (0,20 m), 10 (0,40 m) e 15 m2 (0,60 m). Medidas de correlação entre os caracteres dos componentes de produção de grãos podem auxiliar no aumento de produtividade das culturas, pois possibilitam o conhecimento dos reflexos da alteração de um caráter sobre a expressão de outros (Benin et al., 2009). A correlação fenotípica entre os caracteres é a única possível de se medir diretamente, pode auxiliar na descoberta de componentes de produção que podem ser alterados para maior produtividade de grãos (Coimbra et al., 2004). A correção e a adubação foram feitas de acordo com a análise de solo, para uma expectativa de produtividade de grãos de aproximadamente 1.500 kg ha‑1. C.A.M.B. Krüger et al. A semeadura foi realizada de forma manual, na terceira semana de maio, com quantidade de sementes superior à mínima necessária para cada densidade avaliada. No estádio de duas a três folhas, as plantas foram desbastadas de forma a se obter em as densidades de plantas desejadas. Como a correlação fenotípica inclui simultanea- mente partes atribuídas a efeitos genéticos e de ambiente, a inclusão de fontes de variação, para ambos os efeitos, pode permitir conhecer com maior fidelidade a força das relações entre os caracteres. Assim, conhecer se a expressão das variáveis é mais fortemente influenciada por efeitos genéticos ou ambientais aumenta a consistência da estimativa da herdabilidade, pelo efeito cumulativo de fontes de variação. A produtividade de grãos (PG) foi determinada pela colheita manual total da parcela, com posterior conversão para quilogramas por hectare. Para as avaliações das plantas, individualmente, foram colhidas aleatoriamente cinco plantas, a fim de se compor a média dos seguintes caracteres: produtividade de grãos por planta (PGP, em gramas por planta); número de síliquas por planta (NSP); número de grãos por síliquas (NGS), obtido pela média da contagem dos grãos de 20 síliquas por planta escolhida ao acaso; número de grãos por planta (NGP), obtido pelo produto entre o NSP e o NGS; massa de síliqua (MS, g), obtida pela média da pesagem das 20 síliquas com grãos coletadas ao acaso; comprimento de síliqua (CS, cm), obtido a partir de medição com régua do comprimento médio das 20 síliquas amostradas para a determinação do NGS; número de ramos secundários (NRS), obtido pela contagem de todos os ramos inseridos no caule O objetivo deste trabalho foi estimar a herdabilidade e avaliar a correlação fenotípica entre caracteres de canola relacionados à produtividade de grãos e à arquitetura de plantas, como forma de prever a estabilidade fenotípica a partir de efeitos que envolvam, de forma conjunta, anos, genótipos e densidades de plantio. Introdução planta, na uniformidade de maturação das síliquas e na produtividade de grãos.i A produtividade da canola (Brassica napus L.) está relacionada aos componentes diretos de produção densidade de plantas, número de síliquas por planta, número de grãos por síliqua e massa de grãos (Thomas, 2003). Assim, alterações promovidas nesses componentes ou em outros, de efeito indireto sobre a produtividade, como comprimento e número de ramos, podem proporcionar modificações na arquitetura da Se a herdabilidade é conhecida, pode-se quantificar a intensidade com que as variações de ambiente afetam a expressão dos caracteres (Amorim et al., 2008). Segundo Carvalho et al. (2001), estimativas de herdabilidade podem ser obtidas com base nos componentes da análise da variância, em experimentos com repetições com igual número de plantas ou de parcelas. Pesq. agropec. bras., Brasília, v.46, n.12, p.1625-1632, dez. 2011 1626 Pesq. agropec. bras., Brasília, v.46, n.12, p.1625-1632, dez. 2011 Material e Métodos Os experimentos foram realizados nos anos agrícolas de 2008 e 2009, no Instituto Regional de Desenvolvimento Rural, do Departamento de Estudos Pesq. agropec. bras., Brasília, v.46, n.12, p.1625-1632, dez. 2011 Herdabilidade e correlação fenotípica de caracteres 1627 principal da planta; altura de inserção do primeiro ramo secundário (AIRS, cm), obtido pela medição com régua entre o ponto de corte, na superfície do solo, até a inserção da primeira ramificação; comprimento de ramo secundário (CR, cm), obtido pela medição com régua da inserção do primeiro ramo no caule principal até sua extremidade; número de ramos terciários (NRT), pela contagem dos ramos provenientes dos ramos secundários; e estatura da planta (EST, cm), medida com régua do ponto de corte da planta, na superfície do solo, até o ápice da planta. a mudanças no ambiente. Portanto, as elevadas magnitudes de h2 para esses caracteres mostram que a redução do espaçamento entre linhas promoveu maior uniformidade de plantas, já que, caso contrário, a desuniformidade no ponto de colheita, em razão do crescimento indeterminado da espécie, teria causado grande perda de grãos pela abertura natural da síliquas (Silva et al., 2011), o que tornaria difícil a obtenção dos altos valores de h2 observados para esses caracteres. Para NGS e MS, foram observadas magnitudes intermediárias de herdabilidade com valores de 0,56 e 0,55, respectivamente, no espaçamento de 0,20 m (Tabela 1). Os caracteres de arquitetura de planta, como NRS, AIRS e NRT, também mostraram valores médios de h2, ao contrário de CR e EST, que apresentaram valores reduzidos de h2. Estas observações indicam que, sob condições de menor espaçamento entre linhas, CR e EST são as variáveis de maior tendência à alteração. Os dados foram submetidos ao modelo de correlação de Pearson para obtenção da magnitude e sentido de associações entre as variáveis estudadas, conforme Falconer & Mackay (1996). A significância das correlações foi avaliada a 5% de probabilidade, tendo- se adotado o teste t, descrito por Steel & Torrie (1980), com n ‑ 2 graus de liberdade, de acordo com o modelo t = r/[(1 ‑ r2)/(n ‑ 2)]0,5, em que r é o coeficiente de correlação entre os caracteres X e Y, e n o número de graus de liberdade nos níveis de tratamentos considerados, no total de n = 64 unidades de observação: dois anos, dois genótipos e quatro densidades de plantio, com quatro repetições. Material e Métodos A partir daí, foi realizada a análise de variância dos tratamentos, que englobaram o efeito cumulativo de ano, genótipo e densidade de plantas, para obtenção dos valores de quadrado médio necessários à determinação dos parâmetros genéticos, como forma de se conhecer a estabilidade dos caracteres agronômicos mensurados, segundo Carvalho et al. (2001). Para estas determinações, foi empregado o programa Genes (Cruz, 2001). No espaçamento de 0,40 m, foram obtidas h2 médias para PG (0,60) e PGP (0,62), em comparação ao espaçamento de 0,20 m (Tabela 1). Quanto aos demais componentes de produção, como NSP, NGS e NGP, também foram observados menores valores em comparação ao espaçamento de 0,20 m. Portanto, no espaçamento de 0,40 m, o ambiente teve maior participação na variação fenotípica. As variáveis CR e NRT mostraram maior estabilidade fenotípica (pouco afetado pela variação ambiental), o que indica efeitos mais diretamente ligados ao genótipo, com elevados valores de h2. Em canola, os efeitos negativos que espaçamentos menores podem exercer sobre alguns caracteres, geralmente, são compensados pela maior PG (Johnson & Hanson, 2003). Além disso, espaçamentos maiores determinam maior competição por fotoassimilados em uma mesma planta, para produzir ramos mais finos e em maior número (Thomas, 2003), o que reduz o NSP (Shahin & Valiollah, 2009). Como a PG está diretamente relacionada à capacidade da planta de converter energia solar em órgãos e grãos, a eficiência de uso da radiação também pode ser alterada via manipulação na arquitetura de plantas (Argenta et al., 2001). Pesq. agropec. bras., Brasília, v.46, n.12, p.1625-1632, dez. 2011 C.A.M.B. Krüger et al. Todavia, ressalta-se a elevada h2 encontrada para o componente de produção direto NGS (0,75). O CS também mostrou elevada participação genética em sua expressão (h2 = 0,71). Herdabilidade elevada também foi observada para NRT. que indica que a plasticidade fenotípica da espécie atua, principalmente por meio desses caracteres, no ajuste do arranjo de plantas e no aproveitamento de área de exploração (Tabela 1). Benin et al. (2005) comentam a necessidade de se identificar, nas espécies, as modificações que maximizam o aproveitamento de estímulos do ambiente. Em canola de inverno, em Foram encontrados valores reduzidos de herdabili- dade para NRS (0,30), AIRS (0,37) e EST (0,30), o Tabela 1. Quadrados médios da análise de variância, médias, parâmetros genéticos de componentes ligados à produção e à arquitetura de plantas de canola. Pesq. agropec. bras., Brasília, v.46, n.12, p.1625-1632, dez. 2011 Resultados e Discussão As estimativas de herdabilidade (h2) para PGP e PG foram mais elevadas no espaçamento entre linhas de 0,20 m, em comparação aos demais espaçamentos (Tabela 1). Os componentes diretos de produtividade de grãos, NSP e NGP, também apresentaram maiores h2 nesse espaçamento (0,87 e 0,80, respectivamente). Em estudos com feijão e girassol, a herdabilidade de caracteres diretamente ligados à produção de grãos foi baixa (Londero et al., 2006; Amorin et al., 2008). Os autores ressaltaram que esses caracteres são de herança quantitativa e, portanto, são governados por grande número de genes de pequeno efeito cumulativo para a expressão do caráter e fortemente responsivos A PG e PGP, no espaçamento entre linhas 0,60 m, apresentaram os menores valores de herdabilidade, 0,39 e 0,50, respectivamente (Tabela 1). O NSP também evidenciou maior participação do ambiente em sua expressão fenotípica (h2 = 0,48), nesse espaçamento. Pesq. agropec. bras., Brasília, v.46, n.12, p.1625-1632, dez. 2011 C.A.M.B. Krüger et al. 1628 C.A.M.B. Krüger et al. Caráter(1) Análise de variância Média Parâmetro genético QME QMT σ2 a σ2 f σ2 g h2 Espaçamento entre linhas de 0,20 m PG 22.218,00 478.754,00 1.010,00 22.218,00 136.352,00 114.134,00 0,84 PGP 5,10 124,01 9,87 5,11 34,84 29,73 0,85 NSP 1.653,00 45.261,00 216,28 165,00 12.555,00 10.902,00 0,87 NGS 2,30 13,78 16,16 2,30 5,17 2,87 0,56 NGP 1.114.528,00 18.694.117,00 3.684,00 1.114.528,00 5.509.425,00 4.394.897 0,80 MS 0,06 0,35 0,08 0,06 0,13 0,07 0,55 CS 0,10 0,32 5,70 0,11 0,16 0,05 0,32 NRS 0,20 0,83 3,92 0,22 0,37 0,15 0,41 AIRS 22,55 143,20 32,01 22,55 52,71 30,16 0,57 CR 46,15 119,18 72,29 46,15 64,41 18,26 0,28 NRT 1,76 14,25 4,93 1,76 4,88 3,12 0,64 EST 67,59 98,00 102,83 67,59 75,19 7,60 0,10 Espaçamento entre linhas de 0,40 m PG 29.442,00 204.226,00 941,00 29.442,00 73.138,00 43.696,00 0,60 PGP 12,14 91,39 11,61 12,14 31,95 19,81 0,62 NSP 5.035,00 22158 237,37 5035 9316 4281 0,46 NGS 2,23 10,94 16,30 2,23 4,41 2,18 0,49 NGP 1.582.436,00 7.488.526,00 4.223,00 1.582.436,00 3.058.959,00 1.476.523,00 0,48 MS 0,28 0,76 0,09 0,28 0,40 0,12 0,30 CS 0,17 0,72 5,75 0,17 0,31 0,14 0,45 NRS 0,18 0,91 3,99 0,18 0,36 0,18 0,50 AIRS 33,12 132,81 34,03 33,12 58,04 24,92 0,43 CR 37,8 394,58 73,72 37,80 127,00 89,20 0,70 NRT 1,21 23,95 5,28 1,21 6,90 5,69 0,82 EST 75,63 214,67 75,63 75,63 110,39 34,76 0,31 Espaçamento entre linhas de 0,60 m PG 17.305,00 61.443,00 848,00 17.305,00 28.340,00 11.035,00 0,39 PGP 20,22 115,53 17,60 20,22 44,05 23,83 0,54 NSP 7.475,00 35.615,00 331,00 7.475,00 14.510,00 7.035,00 0,48 NGS 2,11 27,90 16,52 2,11 8,56 6,45 0,75 NGP 2.400.429,00 10511770,00 6.273,00 2.400.429,00 4.428.264,00 2.027.835,00 0,46 MS 0,05 0,33 0,09 0,05 0,12 0,07 0,58 CS 0,06 0,64 5,78 0,06 0,21 0,15 0,71 NRS 0,36 0,97 4,28 0,36 0,51 0,15 0,30 AIRS 29,85 99,83 31,49 29,85 47,35 17,50 0,37 CR 57,62 405,96 78,64 57,62 144,71 87,09 0,60 NRT 2,01 56,44 7,05 2,01 15,62 13,61 0,87 EST 74,14 198,52 109,06 74,14 105,24 31,10 0,30 (1)PG, produtividade de grãos (kg ha-1); PGP, produtividade de grãos por planta (gramas por planta); NSP, número de síliquas por planta; NGS, número de grãos por síliqua; NGP, número de grãos por planta; MS, massa de síliqua (g); CS, comprimento de síliqua (cm); NR, número de ramos secundários; AIRS, altura de inserção do ramo secundário (cm); CR, comprimento de ramo (cm); NRT, número de ramos terciários; EST, estatura de planta (cm). ( )PG, produtividade de grãos (kg ha ); PGP, produtividade de grãos por planta (gramas por planta); NSP, número de síliquas por planta; NGS, número de grãos por síliqua; NGP, número de grãos por planta; MS, massa de síliqua (g); CS, comprimento de síliqua (cm); NR, número de ramos secundários; AIRS, altura de inserção do ramo secundário (cm); CR, comprimento de ramo (cm); NRT, número de ramos terciários; EST, estatura de planta (cm). QME, quadra- do médio do erro; QMT, quadrado médio de tratamento; σ2a, variância de ambiente; σ2f, variância fenotípica; σ2g, variância genética; h2, herdabilidade. C.A.M.B. Krüger et al. QME, quadra- os da análise de variância, médias, parâmetros genéticos de componentes ligados à produção e à l Tabela 1. Quadrados médios da análise de variância, médias, parâmetros genéticos de componentes ligados à produção e à arquitetura de plantas de canola. Pesq. agropec. bras., Brasília, v.46, n.12, p.1625-1632, dez. 2011 Herdabilidade e correlação fenotípica de caracteres 1629 correlação significativa e positiva com PG, indício de que o favorecimento na expressão desses caracteres pode aumentar a produtividade de grãos. Observou- se correlação negativa entre NGS e PG. De acordo com Thomas (2003), a maior demanda de energia pela planta, para a manutenção do óvulo fertilizado, tem efeito negativo sobre PG. Assim, em razão da competição por fotoassimilados, a planta diminui NSP para garantir que as síliquas formadas suportem o enchimento de vários grãos, já que, segundo Silva et al. (2011), condições de alta densidade de plantas, Leach et al. (1999) não constataram alterações significativas em PG e nos componentes diretos de produção, em razão de diferentes densidades populacionais. Na análise de correlação entre caracteres, no espaçamento de 0,20 m, a PGP apresentou baixa correlação com a PG (Tabela 2). No entanto, a significância da correlação suporta a possibilidade de utilização de PGP como subsídio para a estimativa de PG. Os caracteres NSP e NGP também apresentaram Tabela 2. Coeficientes de correlação fenotípica de componentes ligados à produção e à arquitetura em canola. C.A.M.B. Krüger et al. Caráter(1) PGP NSP NGS NGP MS CS NRS AIRS CR NRT EST Espaçamento entre linhas de 0,20 m PG 0,36* 0,31* -0,44* 0,26* -0,28* -0,19 0,23* 0,07 -0,15 -0,27* 0,01 PGP 0,96* 0,18 0,95* 0,06 -0,04 0,18 -0,33* 0,18 0,08 -0,02 NSP 0,17 0,97* 0,01 -0,06 0,14 -0,29* 0,15 0,02 -0,01 NGS 0,21 0,47* 0,52* -0,18 0,15 0,22 0,24* 0,24* NGP 0,01 -0,07 0,10 -0,29* 0,18 0,02 0,07 MS 0,59* 0,11 -0,31* 0,48* 0,42* 0,30* CS 0,12 -0,30* 0,38* 0,31* 0,18 NRS -0,47* 0,49* 0,37* 0,08 AIRS -0,66* 0,48* 0,25* CR 0,51* 0,34* NRT 0,09 Espaçamento entre linhas de 0,40 m PG 0,61* 0,57* -0,36* 0,51* -0,15 -0,01 0,22 0,21 -0,29* -0,31* -0,07 PGP 0,93* -0,18 0,92* 0,17 0,17 -0,63* -0,24* 0,18 0,08 0,11 NSP -0,27* 0,98* 0,11 -0,04 0,52* -0,22 0,17 -0,01 0,04 NGS -0,20 0,40* 0,59* -0,13 -0,09 0,35* 0,52* 0,39* NGP 0,12 -0,03 0,54* -0,25* 0,20 0,01 0,05 MS 0,44* 0,24* -0,35* 0,55* 0,62* 0,36* CS 0,23* -0,14 0,29* 0,43* 0,38* NRS -0,37* 0,35* 0,18 0,15 AIRS -0,64* -0,49* -0,01 CR 0,74* 0,63* NRT 0,46* Espaçamento entre linhas de 0,60 m PG 0,45* 0,40* -0,21 0,43* 0,08 0,15 0,13 0,24 -0,14 -0,21 -0,02 PGP 0,81* -0,02 0,84* 0,23* 0,44* 0,39* -0,10 0,21 0,18 0,19 NSP -0,17 0,97* 0,07 0,20 0,47* -0,05 0,16 0,14 0,13 NGS -0,13 0,67* 0,70* 0,22 -0,50* 0,60* 0,62* 0,46* NGP 0,08 0,22 0,45* -0,05 0,17 0,12 0,15 MS 0,80* 0,39* -0,39* 0,53* 0,59* 0,43* CS 0,37* -0,37* 0,63* 0,60* 0,53* NRS -0,56* 0,56* 0,59* 0,25* AIRS -0,67* -0,68* -0,12 CR 0,81* 0,72* NRT 0,46* (1)PG, produtividade de grãos (kg ha-1); PGP, produtividade de grãos por planta(gramas por planta); NSP, número de síliquas por planta; NGS, número de grãos por síliqua; NGP, número de grãos por planta; MS, massa de síliqua (g); CS, comprimento de síliqua (cm); NR, número de ramos secundários; AIRS, altura de inserção do ramo secundário (cm); CR, comprimento de ramo (cm); NRT, número de ramos terciários; EST, estatura de planta (cm). Significativo pelo teste t, a 5% de probabilidade. C.A.M.B. Krüger et al. A ausência de associação entre NGS e PG, no espaçamento de 0,60 m, pode estar relacionada à deiscência da síliqua no momento da colheita, em razão da desuniformidade de maturação proporcionada pelo incremento do comprimento de ramos; Coimbra et al. (2004) relatam que esse componente é estreitamente relacionado a PG. a massa de grãos é o caráter com maior estabilidade de expressão. Quanto aos caracteres morfológicos, NRS apresentou correlação positiva com PG, e NRT negativa. A PGP apresentou, como era de se esperar, elevadas correlações com NSP e NGP (Tabela 2). De acordo com Diepenbrok (2000), cada síliqua formada produz, no mínimo, 15 grãos. No entanto, observou-se correlação negativa entre esse caráter e AIRS. O CR apresentou elevada correlação positiva com NRT e EST. A EST correlacionou-se, sempre positivamente, com um maior número de componentes de produção, no espaçamento de 0,60 m. Silva et al. (2011) comentam que espaçamentos maiores em canola podem promover a maior expressão de componentes ligados à arquitetura da planta, como forma de ajuste no dossel, tendendo, porém, a alterações negativas nos componentes diretos ligados à produtividade de grãos. Foram observadas correlações positivas de NGS com MS, CS, NRT e EST. É interessante observar que caracteres que conferem efeito compensatório por área, como a EST e NRT, contribuem também para a expressão de NGS. A variável MS foi positivamente influenciada pelos componentes morfológicos CS, CR, NRT e EST e negativamente por AIRS. Portanto, NRT e EST mostraram efeito direto e positivo sobre os componentes diretos NGS e MS. Contudo, o NRT foi mais efetivo em promover alterações positivas nesses caracteres do que a EST. g Em feijão‑caupi sob diferentes espaçamentos entre linhas e densidades na linha, não foram observadas modificações nos caracteres número de nós no ramo principal, número de grãos por vagem e peso de 100 grãos; porém, densidades maiores promoveram reduções no número de ramos laterais e número de vagens e de grãos por planta (Bezerra et al., 2009). Em soja, Tourino et al. (2002) observaram modificações na arquitetura de plantas e na PG, quando foi alterado o arranjo de plantas, com forte ligação do genótipo na modificação da arquitetura de planta. Em milho, aumentos na produtividade de grãos foram constatados com a redução do dossel, e o número de grãos por área foi o componente de produção maximizado com a diminuição do espaçamento entre linhas (Strieder et al., 2008). Pesq. agropec. bras., Brasília, v.46, n.12, p.1625-1632, dez. 2011 C.A.M.B. Krüger et al. (1)PG, produtividade de grãos (kg ha-1); PGP, produtividade de grãos por planta(gramas por planta); NSP, número de síliquas por planta; NGS, número de grãos por síliqua; NGP, número de grãos por planta; MS, massa de síliqua (g); CS, comprimento de síliqua (cm); NR, número de ramos secundários; AIRS, altura de inserção do ramo secundário (cm); CR, comprimento de ramo (cm); NRT, número de ramos terciários; EST, estatura de planta (cm). Significativo pelo teste t, a 5% de probabilidade. (1)PG, produtividade de grãos (kg ha-1); PGP, produtividade de grãos por planta(gramas por planta); NSP, número de síliquas por planta; NGS, número de grãos por síliqua; NGP, número de grãos por planta; MS, massa de síliqua (g); CS, comprimento de síliqua (cm); NR, número de ramos secundários; AIRS, altura de inserção do ramo secundário (cm); CR, comprimento de ramo (cm); NRT, número de ramos terciários; EST, estatura de planta (cm). *Significativo pelo teste t, a 5% de probabilidade. Pesq. agropec. bras., Brasília, v.46, n.12, p.1625-1632, dez. 2011 1630 C.A.M.B. Krüger et al. C.A.M.B. Krüger et al. Em trigo, o uso de diferentes densidades de cultivo pôde determinar incrementos na produtividade de grãos, e a distribuição equidistante de sementes por área foi a variável que contribuiu para a maior produtividade (Silveira et al., 2010). Em canola, Coimbra et al. (2004) observaram que, com elevada densidade de plantas, os componentes número de plantas por unidade de área e de grãos por planta apresentam maior relação com a produtividade de grãos. No espaçamento de 0,40 m, observou-se incremento da correlação de PG com PGP, NSP e NGP (Tabela 2). Nesse espaçamento, no entanto, MS e NRS deixaram de ter efeito significativo sobre PG. A correlação negativa de PG com NGS diminuiu, e a com NRT aumentou. Além disso, o CR também passou a ter efeito negativo sobre PG. A PGP voltou a mostrar elevadas correlações com o NSP e NGP e correlação negativa com AIRS. No espaçamento de 0,40 m foi observada correlação negativa de alta magnitude entre PGP e NRS. No espaçamento de 0,40 m, observou-se relação positiva entre NGP e NRS (Tabela 2). Esse resultado indica que a tendência de incremento em NRS, em resposta ao aumento no espaçamento, aumentou NSP e, por consequência, também o NGP. Além disto, a EST mostrou efeitos positivos e significativos em NGS, MS, CS, CR e NRT, nesse espaçamento. No espaçamento entre linhas de 0,60 m, observaram- se correlações intermediárias de PG com PGP, NSP e NGP, em comparação aos espaçamentos anteriores (Tabela 2). Além disso, observaram-se menores valores de correlação de PGP com NSP e NGP. O NRS foi o único caráter ligado à arquitetura da planta correlacionado a NSP e NGP, nesse espaçamento. Para PGP, o CS também apresentou correlação significativa e positiva. O NGS apresentou os maiores valores de correlação direta com MS e CS, em comparação aos valores observados nos demais espaçamentos. Na análise conjunta dos três espaçamentos entre linhas, a partir das médias gerais de densidades de planta e anos de cultivo, observou-se diminuição de PG com o aumento do espaçamento entre linhas (Tabelas 1 e 2). Para a PGP, no entanto, observou-se o contrário, o que também ocorreu com os demais Pesq. agropec. bras., Brasília, v.46, n.12, p.1625-1632, dez. 2011 Herdabilidade e correlação fenotípica de caracteres 1631 caracteres diretos de produção, com exceção de NGS, que apresentou comportamento estável. COIMBRA, J.L.M.; GUIDOLIN, A.F.; ALMEIDA, M.L. de; SANGOI, L.; ENDER, M.; MEROTTO JÚNIOR, A. C.A.M.B. Krüger et al. Análise de trilha dos componentes do rendimento de grãos em genótipos de canola. Ciência Rural, v.34, p.1421‑1428, 2004. As variáveis ligadas à arquitetura MS, AIRS, CR e NRT mostraram tendência de aumento com o acréscimo no espaçamento entre linhas. Ressalta-se que, com o maior espaçamento, PG e PGP também mostraram redução nos valores de herdabilidade, indício de que são mais estáveis nos menores espaçamentos, ao contrário de CS e NRT. CRUZ, C.D. Programa Genes: versão Windows: aplicativo computacional em genética e estatística. Viçosa: UFV, 2001. 648p. DIEPENBROK, W. Yield analysis of winter oilseed rape (Brassica napus L.): a review. Field Crops Research, v.67, p.35‑49, 2000. FALCONER, D.S.; MACKAY, T.F.C. Introduction to quantitative genetics. Harlow: Longman, 1996. 464p. As correlações positivas entre PG e os componentes NSP e NGP foram observadas nos três espaçamentos estudados. A PGP e os caracteres NSP e NGP também apresentaram correlações positivas e de elevada magnitude. A relação de NSP com PG mostrou magnitude média nos espaçamentos de 0,20 e 0,40 m e ausência de correlação com PGP. HARTWIG, I.; CARVALHO, F.I.F. de; OLIVEIRA, A.C. de; VIEIRA, E.A.; SILVA, J.A.G. da; BERTAN, I.; RIBEIRO, G.; FINATTO, T.; REIS, C.E.S. dos; BUSATO, C.C. Estimativa de coeficientes de correlação e trilha em gerações segregantes de trigo hexaplóide. Bragantia, v.66, p.203‑218, 2007. JOHNSON, B.L.; HANSON, B.K. Row‑spacing interactions on spring canola performance in the northern great plains. Agronomy Journal, v.95, p.703‑708, 2003. Conclusões LEACH, J.E.; STEVENSON, H.J.; RAINBOW, A.J.; MULLEN, L.A. Effects of high plant populations on the growth and yield of winter oilseed rape (Brassica napus). Journal of Agricultural Science, v.132, p.173‑180, 1999. 1. A produtividade de grãos por área e por planta apresenta maior herdabilidade no menor espaçamento entre linhas. LONDERO, P.M.G.; RIBEIRO, N.D.; CARGNELUTTI FILHO, A.; RODRIGUES, J. de A.; ANTUNES, I.F. Herdabilidade dos teores de fibra alimentar e rendimento de grãos em populações de feijoeiro. Pesquisa Agropecuária Brasileira, v.41, p.51‑58, 2006. 2. Os caracteres número de síliquas e número de grãos por planta apresentam maior correlação direta e positiva com a produtividade de grãos. SANTOS, H.G. dos; JACOMINE, P.K.T.; ANJOS, L.H.C. dos; OLIVEIRA, V.A. de; OLIVEIRA, J.B. de; COELHO, M.R.; LUMBRERAS, J.F.; CUNHA, T.J.F. (Ed.). Sistema brasileiro de classificação de solos. 2.ed. Rio de Janeiro: Embrapa Solos, 2006. 306p. Pesq. agropec. bras., Brasília, v.46, n.12, p.1625-1632, dez. 2011 Referências AMORIM, E.P.; RAMOS, N.P.; UNGARO, M.R.G.; KIIHL, T.A.M. Correlações e análise de trilha em girassol. Bragantia, v.67, p.307‑316, 2008. SHAHIN, Y.; VALIOLLAH, R. Effects of row spacing and seeding rates on some agronomical traits of spring canola (Brassica napus L.) cultivars. Journal of Central European Agriculture, v.10, p.115‑122, 2009. ARGENTA, G.; SILVA, P.R.F. da; SANGOI, L. Arranjo de plantas em milho: análise do estado‑da‑arte. Ciência Rural, v.31, p.1075‑1084, 2001. SILVA, J.A.G. da; MOTTA, M.B. da; WUNSCH, J.A.; CRESTANI, M.; FERNANDES, S.B.V.; BERTO, J.L.; GAVIRAGHI, F.; MARTINS, J.A.K.; WAGNER, J.F.;VALENTINI, A.P.F.; ZAMBONATO, F. Dessecação em pré‑colheita como estratégia de manejo na redução de perdas por fatores de ambiente em canola. Revista Brasileira de Agrociência, v.16, p.151‑158, 2011. BENIN, G.; CARVALHO, F.I.F. de; OLIVEIRA, A.C.; LORENCETTI, C.; VIEIRA, E.A.; COIMBRA, J.L.M.; VALÉRIO, I.P.; FLOSS, E.L.; BERTAN, I.; SILVA, G.O. da. Adaptabilidade e estabilidade em aveia em ambientes estratificados. Ciência Rural, v.35, p.295‑302, 2005. BENIN, G.; SILVA, G.O. da; PAGLIOSA, E.S.; LEMES, C.; SIGNORINI, A.; BECHE, E.; CAPELIN, M.A. Capacidade de combinação em genótipos de trigo estimada por meio de análise multivariada. Pesquisa Agropecuária Brasileira, v.44, p.1145‑1151, 2009. SILVEIRA, G. da; CARVALHO, F.I.F. de; OLIVEIRA, A.C. de; VALÉRIO, I.P.; BENIN, G.; RIBEIRO, G.; CRESTANI, M.; LUCHE, H. de S.; SILVA, J.A.G. da. Efeito da densidade de semeadura e potencial de afilhamento sobre a adaptabilidade e estabilidade em trigo. Bragantia, v.69, p.63‑70, 2010. BEZERRA, A.A. de C.; TÁVORA, F.J.A.; FREIRE FILHO, F.R.; RIBEIRO, V.Q. Características de dossel e de rendimento em feijão‑caupi ereto em diferentes densidades populacionais. Pesquisa Agropecuária Brasileira, v.44, p.1239‑1245, 2009. STEEL, R.G.D.; TORRIE, J.H. Principles and procedures of statistics: a biometrical approach. 2nd ed. New York: McGraw‑Hill, 1980. 631p. STRIEDER, M.L.; SILVA, P.R.F. da; RAMBO, L.; SANGOI, L.; SILVA, A.A.; ENDRIGO, P.C.; JANDREY, D.B. Crop management systems and maize grain yield under narrow row spacing. Scientia Agricola, v.65, p.346‑353, 2008. CARVALHO, F.I.F. de; SILVA, S.A.; KUREK, A.J.; MARCHIORO, V.S. Estimativas e implicações da herdabilidade como estratégia de seleção. Pelotas: UFPel, 2001. 99p. Pesq. agropec. bras., Brasília, v.46, n.12, p.1625-1632, dez. 2011 1632 C.A.M.B. Krüger et al. TOURINO, M.C.C.; REZENDE, P.M. de; SALVADOR, N. Espaçamento, densidade e uniformidade de semeadura na produtividade e características agronômicas da soja. Pesquisa Agropecuária Brasileira, v.37, p.1071‑1077, 2002. THOMAS, P. Canola grower’s manual. Winnipeg: Canola Council of Canada, 2003. Available at: <http://www. canolacouncil.org/canola_growers_manual.aspx>. Accessed on: 15 Sept. 2010. Recebido em 17 de junho de 2011 e aprovado em 21 de outubro de 2011
https://openalex.org/W2925366671	https://hal.archives-ouvertes.fr/hal-02294221/file/ogst180267.pdf	English	null	Palaeostructure, evolution and tight oil distribution of the Ordos Basin, China	Oil & gas science and technology	2,019	cc-by	7,607	To cite this version: Shengli Gao, Jinxia Yang. Palaeostructure, evolution and tight oil distribution of the Ordos Basin, China. Oil & Gas Science and Technology - Revue d’IFP Energies nouvelles, 2019, 74, pp.35. 10.2516/ogst/2019010. hal-02294221 Received: 11 August 2018 / Accepted: 15 February 2019 Received: 11 August 2018 / Accepted: 15 February 2019 Received: 11 August 2018 / Accepted: 15 February 2019 Abstract. Whether or not the tight oil in the Triassic Yanchang Formation of the Ordos Basin is controlled by structural factors is a controversial issue, the relationship between the structural factors of the strata and the distribution of tight oil is limited to the study of current structures. The traditional view is that structural factors have no obvious control over the formation and distribution of the oil reservoir. Taking the Chang 8 member of the Triassic Yanchang Formation in the Ordos Basin as an example, this paper studies respectively the burial of strata-hydrocarbon generation history of the individual well and the structural evolution history of strata in the basin by using software tools of the Genex burial-hydrocarbon generation history restoration and TemisFlow evolution of stratigraphic structures. It is considered that the hydrocarbon generation period of the source rock of the Triassic Yanchang Formation in the Ordos Basin is from early Middle Jurassic to end of Early Cretaceous. By reconstructing the evolution and structure of the Chang 8 member during the hydrocar- bon accumulation period, combined with a comprehensive analysis on the distributional characteristics of the Chang 8 oil reservoir, we found the palaeoslopes and palaeohighs of the Chang 8 reservoir to represent areas in which tight oils were distributed. Palaeo-structural characteristics of the target layer exhibit control over the Chang 8 reservoir. The new theory underlying tight oil exploration, which is based on the recovery of the palaeogeomorphology of the target layer during the hydrocarbon generation period, incorporates the vital roles of key controlling factors over tight oil accumulation, so that the mind-set on tight oil exploration in the Ordos Basin has evolved. In settings over large distributional areas and with source reservoir contact, reservoir formation and distribu- tion are mostly controlled by sedimentary microfacies and the physical properties of the reservoir (Ren et al., 2014; Yang et al., 2017) or by the hydrocarbon source and reser- voir (Zhao et al., 2012). The evolution of the thermal struc- ture of the basin, especially structural thermal events during the Mesozoic, exhibits an important control over oil accumulation (mineralization) (Gao and Ren, 2006; Ren et al., 2014, 2017). Received: 11 August 2018 / Accepted: 15 February 2019 What should be emphasized in par- ticular, is that the controlling effect of tectonism on the formation and distribution of the tight oil reservoir in the Triassic Yanchang Formation has increasingly attracted attention of researchers, and the traditional theories applied to this reservoir have gradually changed. Early researchers ignored the effects of structural factors on the accumulation of tight oil. For many years, researchers focused on current structural characteristics; based on the traditional mind-set that “the slope in northern Shaanxi is a gentle west-inclined monocline with simple structure, and lacks structural traps such as anticline”, it has been assumed that the formation and distribution of oil reservoirs are dominantly controlled This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Oil & Gas Science and Technology - Rev. IFP Energies nouvelles 74, 35 (2019) Available online at: S. Gao & J. Yang, published by IFP Energies nouvelles, 2019 ogst.ifpenergiesnouvelles.fr https://doi.org/10.2516/ogst/2019010 Oil & Gas Science and Technology - Rev. IFP Energies nouvelles 74, 35 (2019) S. Gao & J. Yang, published by IFP Energies nouvelles, 2019 https://doi.org/10.2516/ogst/2019010 Oil & Gas Science and Technology - Rev. IFP Energies nouvelles 74, 35 (2019) S. Gao & J. Yang, published by IFP Energies nouvelles, 2019 https://doi.org/10.2516/ogst/2019010 * Corresponding author: gshl@xsyu.edu.cn Available online at: ogst.ifpenergiesnouvelles.fr HAL Id: hal-02294221 https://hal.science/hal-02294221v1 Submitted on 23 Sep 2019 L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés. HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. Palaeostructure, evolution and tight oil distribution of the Ordos Basin, China Shengli Gao1,, and Jinxia Yang2 Shengli Gao1,, and Jinxia Yang2 Shengli Gao1,*, and Jinxia Yang2 1 Xi’an Shiyou University, Xi’an, 710065 Shaanxi, China 2 School of Mathematics and Information Science, Shaanxi Normal University, Xi’an, 710062 Shaanxi, China 1 Xi’an Shiyou University, Xi’an, 710065 Shaanxi, China 2 School of Mathematics and Information Science, Shaanxi N 1 Xi’an Shiyou University, Xi’an, 710065 Shaanxi, China 2 School of Mathematics and Information Science, Shaanxi Normal University, Xi’an, 710062 Shaanxi, China 1 Introduction The overall structural form of the basin presents as a north–south, rectangular basin, composed of a large, asymmetrical syncline, with a wide and gentle eastern wing, a steep and narrow western wing, and an area of 25 104 km2 (He, 2003; Lai et al., 2016). ( , ; , ) Before the late Triassic, the Ordos Basin was part of the western margin of the North China Platform, inheriting and developing the Palaeozoic ocean-land phases of conti- nental sedimentation. In the late Triassic, due to compres- sion from west to east and the left-lateral shearing of the western margin of the North China Platform, the platform basement in the Ordos Basin was bent downward during the unbalanced uplift and subsidence. The southern region, which had continuously been a depression over a long period of time, gradually began separating from the North China Platform, until it evolved into an independent, large, inland lacustrine basin. During the Indosinian Orogeny at the end of the Late Triassic, the basin rapidly rose, and this lake basin disappeared and was denuded. During the Jurassic, the basin began to sink again, developing a broad distribution of ﬂood plain deposits. By this time, within the 25 104 km2 area of the main body of the Ordos Basin, the structural and sedimentary evolution provided the time, space and abundant material needed for the formation, migration and accumulation of Mesozoic oils. After the Early Cretaceous, the basin began to uplift and the overly- ing strata in local areas suffered from denudation. Since then, it has been maintained as the current structural features. In a word, the control factors of tight oil accumulation need further investigation (Hu et al., 2016; Yao et al., 2013; Yang et al., 2017). Few studies have been conducted on the control of the formation and distribution of tight oil by the paleostructure during hydrocarbon ﬁlling and accumulation. Recently, the role of structural conditions in controlling hydrocarbon accumulation is emphasized (Guo, 2010). In this paper, the relationship between strati- graphic structural features and tight oil distribution in the reservoir forming period is revealed through the restoration of paleotectonics in the Chang 8 Formation of the Triassic system. 1 Introduction With the acceleration of tight oil exploration, the dis- covery of the Chang 6 and Chang 8 oil reservoirs in the Huaqing region of the Ordos Basin has attracted attention to the control of structural factors (i.e., the geomorphology of bottom of the lake basin) throughout the oil reservoir. The sedimentary periods of target layers in the Chang 6 and Chang 8 reservoirs are in a deep to somewhat deep lake environment. Based on conventional exploration theories, this area is a “forbidden zone” for tight oil exploration. In-depth studies have suggested that the sedimentary period of the Chang 6 oil-bearing formation in this area included the large-scale development of a delta. Large-scale, low-permeability reservoirs were formed under the joint control of favourable sedimentary facies and diagenetic facies, and the setting of a multi-slope break bottom geo- morphology of the lake basin was the structural condition needed for the formation of the reservoirs (Yang et al. 2012). Focusing on the important role of structural factors during the formation of tight oil makes it possible to directly search for and evaluate the sweet spots of tight oil. Exploring tight oil in the “deep water zone” is acceler- ated (Bai et al., 2017; Xu et al., 2016). The basin can be divided into six ﬁrst-class structural units, namely the central Yishan Slope, the eastern Jinxi Flexure Belt, the western Tianhuan Depression, the west- ern marginal thrust belt, the northern Yimeng Uplift and the southern Weibei Uplift (Fig. 1). Fractures and folds along the edge of the basin are relatively well-developed (He, 2003), while the internal structure of the basin is com- paratively simple; the strata are gently sloping by generally less than 10 m/km. There is no secondary structure on the Yishan Slope, and the tertiary structure is mostly composed of nose-shaped folds, anticlinal structural developments, with large amplitudes and relatively good traps are rarely found. According to the drilling and seismic data in the study area, the cross section was made. In the Ordos Basin, the oldest strata revealed by drilling are Proterozoic strata (Pt) (Fig. 1c). Most of the Yishan Slope was formed during the Early Cretaceous. Presenting as a gentle single slope inclined to the west, with an inclination angle of only 0.5–1, and the stratigraphic thickness varies little (Figs. 1c and 2). 1 Introduction The exploration of tight oil in the Ordos Basin has been aimed at the deep water zone. The search for a zone for high-quality reservoir development and the sweet spot of the tight oil reservoir remains in the exploration stage, and will be the key research direction to pursue in the future. Current studies on the factors controlling tight oil accumulation are mostly based on comprehensive analyses of various geological factors, among which structural factors have gradually attracted the attention of researchers. This change also reﬂects the fact that researchers have been focused on discerning the distributions of the high-quality reservoirs and the sweet spots of tight oil reservoirs. Regard- ing the factors controlling oil accumulation, most research- ers believe that the high-quality hydrocarbon source rock of the 7th segment of the Yanchang Formation in the Ordos Basin is the major hydrocarbon generation layer in the basin. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. S. Gao and J. Yang: Oil & Gas Science and Technology - Rev. IFP Energies nouvelles 74, 35 (2019) 2 by their sedimentary facies, while other factors, including structure, impart no control over the formation and distri- bution of reservoirs in the Yanchang Formation (Yang et al., 2007, 2013). With the notable increase in drilling den- sity, the uplifted structure of the Yanchang Formation, especially the low-amplitude, nose-shaped uplifted struc- ture, has been found to be more developed than previously thought, and it has substantial control over the formation and enrichment of Triassic oil reservoirs (Zhao et al., 2006). However, these structural characteristics refer to the present structure after the accumulation period. What were the structural characteristics during the accumulation period? How did those structural characteristics evolve and become retained as the current ones? How were the struc- tures during the accumulation period (i.e., the tectonic geo- morphology of the bottom of the lake basin) and the period of oil and gas charging formed? These were unknown in the past, and there have been limited studies more recently. In the south, it faces the Weihe Basin across from the Weibei Flexure Belt. In the north, it is adjacent to the Hetao Basin from across the Wulinger bulge, and it has a rectangular outline. 1 Introduction Nose-shaped uplifts of different scales and sizes, and tilting from east to west, are developed along the slope belt. 3 Materials and methods A geological model of the basin was established based on a systematic sorting of various data, such as structural, sedi- mentary, mud logging and well logging data in the study area, with the palaeo-surface morphology, erosion thickness and compaction coefﬁcients taken into account. Simulation parameters were selected, and the initial and boundary con- ditions were deﬁned, which provided a basis for recovering the evolutionary processes at work in the basin and the evo- lutionary history of regional ﬂuid dynamics. The Temis3D v.4.0 and Temis2D v.4.0.4 basin simulation software devel- oped by the French Petroleum Research Institute was used to recover the post-hydrocarbon structure and evolution of the target layers in the Chang 8 reservoir (Including plane and proﬁle features). With the distributional characteristics of the Chang 8 reservoir taken into consideration, a study on the control of the bottom palaeostructure and the evolution of the target layers of the tight oil reservoir was conducted. The mainly eastward subduction occurred in the north- west of the North China ancient land, under the inﬂuence of the Paleotethys in the Late Triassic (Indosinian period) and acted on the Alxa ancient land. In this period (Indosinian), the North China ancient land still had the northeastward compressive force acting on the Qilian-North Qinling fold belt because of the impact of the Yangtze plate tectonic activity. Under the combined action of the forces in the two directions mentioned above, the push-eastward occurred, and a foreland basin was formed in the western part of the Ordos Basin due to the subduction. During the Indosinian-Yanshanian period, the China plate was affected by the Tethys and Paciﬁc plate activities. The force acting on the North China paleo-continent is mainly represented by the westward compression stress. Affected by this stress, the North China ancient land grad- ually uplifted from east to west, forming the structural features of the Yanchang Formation strata in the Ordos Basin, which are “high in the east and low in the west”. From the Late Cretaceous to the Early Tertiary, the stress of westward compression gradually weakened and disap- peared. Finally, the present Ordos Basin was formed, Genex v. 4.0.3 burial history recovery software was used to recover the single-well burial and hydrocarbon genera- tion histories. 2 Geological setting 2. Two-dimensional seismic proﬁle of the H115781 in the Ordos Basin, China (See Fig. 1a for the position of seismic line). Fig. 2. Two-dimensional seismic proﬁle of the H115781 in the Ordos Basin, China (See Fig. 1a for th tectonic movements reforming the North China ancient land and through the multi-stage reformation of Indochina movement and Yanshan movement (Fig. 4). basically being a large westward inclined monoclinic struc- ture, constituting a structural feature from high to low east- to-west, with a depression-thrust belt in the western part of the basin (Liu et al., 1986; Sun et al., 1985; Ye, 1983). To the south of the prototype basin in Late Triassic was the Qilian-North Qinling Caledonian fold belt formed in the Caledonian period, with compression direction of north- west, and the Kunlun-Qinling Variscan-Indosinian fold belts formed in Variscan-Indosinian movement period, with compression direction of northwest. To the north of the prototype basin is the Tianshan-Xingmeng Variscan fold belt formed in the Variscan period, with compression direc- tion of northwest, and the Alxa ancient land in the Yinshan mountain system and the northwest region of Ordos Basin. To the east is a large piece of North China ancient land. 2 Geological setting The Chang 8 oil layer is one of the major oil layers of the Triassic Yanchang Formation in the Ordos Basin. The Yanchang Formation is divided into ﬁve members accord- ing to the sedimentary sequence, and further divided into 10 oil layers from top to bottom, according to the vertical The Ordos Basin is located in central western China, and echoes the Jinxi Flexure Belt and the Luliang Uplift in the east. In the west, it stands opposite to the Liupanshan and Yinchuan basins through a thrust structural belt. Gao and J. Yang: Oil & Gas Science and Technology - Rev. IFP Energies nouvelles 74, 35 (2019) 3 f the oil layers. Namely, these layers are the Chang 10, among which, Chang 1, Chang 2, ang 6 and Chang 8 are regional industrial oil . The Ordos Basin was developed in the the Mesozoic. During this period, the Ordos was part of the North China ancient land ( formed under the background of Caledoni cation of the Ordos Basin in China; (b) Tectonic units in the Ordos Basin; (c) Stratigraphic cor 1. Fig. 1. (a) Location of the Ordos Basin in China; (b) Tectonic units in the Ordos Basin; (c) Stratigraphic correlation proﬁle of Majiatan-Zk9-1. Fig. 1. (a) Location of the Ordos Basin in China; (b) Tectonic units in the Ordos Basin; (c) Stratigraphic correlation proﬁle of Majiatan-Zk9-1. The Ordos Basin was developed in the Late Triassic of the Mesozoic. During this period, the Ordos prototype basin was part of the North China ancient land (Ye, 1983). It was formed under the background of Caledonian and Variscan distribution of the oil layers. Namely, these layers are the Chang 1 to Chang 10, among which, Chang 1, Chang 2, Chang 3, Chang 6 and Chang 8 are regional industrial oil layers (Fig. 3). distribution of the oil layers. Namely, these layers are the Chang 1 to Chang 10, among which, Chang 1, Chang 2, Chang 3, Chang 6 and Chang 8 are regional industrial oil layers (Fig. 3). 4 S. Gao and J. Yang: Oil & Gas Science and Technology - Rev. IFP Energies nouvelles 74, 35 (2019 4 S. Gao and J. Yang: Oil & Gas Science and Technology - Rev. IFP Energies nouvelles 74, 35 (2019) Fig. 2. Two-dimensional seismic proﬁle of the H115781 in the Ordos Basin, China (See Fig. 1a for the position of seismic line). Fig. 3 Materials and methods When applying the TemisFlow basin simulation soft- ware to recover the evolution of the palaeostructure, the major parameters involved were the mudstone compaction coefﬁcient, erosional thickness, stratigraphic thickness, sand-to-ground ratio and the structure of the top surface Gao and J. Yang: Oil & Gas Science and Technology - Rev. IFP Energies nouvelles 74, 35 (2019) S. Gao and J. Yang: Oil & Gas Science and Technology - Rev. IFP Energies nouvelles 74, 35 (2019) 5 ition of these parameters was based on the many petroleum exploration wells in the stratigraphic thickness, sand-to-ground surface structural characteristics was achiev n table and comprehensive stratigraphic column in the Ordos Basin. Fig. 3. Formation table and comprehensive stratigraphic column in the Ordos Basin. stratigraphic thickness, sand-to-ground ratio and top surface structural characteristics was achieved by a compre- hensive application of logging curves, such as those for sonic wave time difference, electric resistivity and density. In the process of palaeo-structural simulations, the acquisition of structure. Acquisition of these parameters was based on the logging curves of many petroleum exploration wells in the study area, and the sonic wave time difference curve was used to recover the mudstone compaction coefﬁcient and the erosional thickness, while the acquisition of the S. Gao and J. Yang: Oil & Gas Science and Technology - Rev. IFP Energies nouvelles 74, 35 (2019 6 Fig. 4. The geodynamic setting of the development of Late Triassic in the Ordos Basin, China (modiﬁed from Ye, 1983; Sun et a 1985; Liu et al., 1986). Fig. 4. The geodynamic setting of the development of Late Triassic in the Ordos Basin, China (modiﬁed from Ye, 1983; Sun et al 1985; Liu et al., 1986). Fig. 4. The geodynamic setting of the development of Late Triassic in the Ordos Basin, China (modiﬁed from Ye, 1983; Sun et al., 1985; Liu et al., 1986). C9-Jt600, revealing the evolutionary characteristics of the target strata. C9-Jt600, revealing the evolutionary characteristics of the target strata. other simulation parameters, such as the hydrocarbon generation kinetic parameter and temperature ﬁeld param- eter, were conducted by referring to previously published research. Other researchers have previously conducted many studies on the geothermal ﬁeld and its evolution in the Ordos Basin, and have made many important discover- ies. 4 Results TemisFlow basin simulation software was used to recon- struct the structure of the target layers of the Chang 8 reservoir and their evolution after the hydrocarbon genera- tion period, comprehensive basin burial-hydrocarbon gener- ation history analysis, in order to reveal the intrinsic coupling between the palaeostructure and evolution of the target layer during the hydrocarbon generation period and the distribution of the oil reservoir. Abundant log data of interval transit times were applied, the overpressure of the individual wells were calcu- lated and a plane distribution map was drafted. Meanwhile, abundant drilling data were used to produce a sand-ground ratio distribution map of the Chang 8 member. Comprehen- sive analyses were performed on the tectonic migration and accumulation conditions of the sand body by combining the evolutionary characteristics of the structure of tie well 3 Materials and methods After recovering the bottom structure and evolution of the target layers of the Chang 8 reservoir after the hydro- carbon period, the distribution of Chang 8 oil reservoirs that have been discovered in the basin were also taken into consideration. Subsequently, the bottom palaeostructure and evolution of the Chang 8 reservoir, as well as the distri- bution of tight oils, were revealed. 4.1 Reconstruction of the burial history and thermal evolution of hydrocarbon source rocks of a single well The results showed that the hydrocarbon source rock of the Triassic Yanchang Formation reached the hydrocarbon generation threshold in the Early Cretaceous, and entered 7 S. Gao and J. Yang: Oil & Gas Science and Technology - Rev. IFP Energies nouvelles 74, 35 (2019) Fig. 5. Single well burial history and hydrocarbon generation history restoration of Y_1 well. Fig. 5. Single well burial history and hydrocarbon generation history restoration of Y_1 well. Fig. 5. Single well burial history and hydrocarbon generation history restoration of Y_1 well. Fig. 6. Structural morphology of Chang 8 member bottom at the end of the Middle Jurassic and reservoir distribution of the Ordos Basin. Fig. 5. Single well burial history and hydrocarbon generation history restoration of Y_1 well. the peak period of oil generation by the end of the Early Cretaceous. By this time, major lithologic traps and struc- tural-lithologic composite traps in the basin had also been formed; after the Early Cretaceous, the basin overall was uplifted, and no large fault systems were formed inside the basin. Therefore, the Mesozoic low-permeability and ultra-low permeability reservoirs, especially the ﬂuids in the Yanchang Formation reservoirs, effectively remained in a closed environment, so the gas and oil reservoirs that had been formed were well preserved (Fig. 5). Fig. 6. Structural morphology of Chang 8 member bottom at the end of the Middle Jurassic and reservoir distribution of the Ordos Basin. structural pattern of high in the east and northeast and low in the west and southwest. Compared with the bottom structure of the Chang 8 reservoir at the end of the Middle Jurassic, an obvious inversion in tectonic movement had occurred. Overall, four large and small convex belts that converged toward the north-western part of the basin devel- oped. Three convex belts in the south central region extended approximately along a northwest–southeast direc- tion, and the northern part extended approximately from west–east. The difference in the elevations of the southwest- ern area and the northern part of the basin was >300 m. The eastern part of the basin and the area between the two convex belts in the south-eastern part of the basin were depression belts, among which a large depression was dis- tributed across the eastern part of the basin, extending from north–south. 4.2 Late Middle Jurassic tectonic geomorphology of concave and convex structures of plane distribution characteristics at the base of Chang 8 At the end of the Middle Jurassic, the bottom structure of the Chang 8 reservoir evolved into a structural pattern that was high in the south and southeast and low in the north and northwest. In the south-eastern part of the basin, there were three large convex belts extending from northwest–southwest. The convex belts in the southeast and southwest were large, while the convex belt in the south central region was relatively small. The difference in the elevation of the convex belt in the southwest was >860 m, that in the southwest >900 m and in the south central part, it was >300 m. Large areas of the central and north-western parts of the basin were depression zones, among which some partially low-amplitude, nose-shaped convexes were developed (Fig. 6). It can be seen from Figure 4 that the Chang 8 reservoir is almost completely distributed in the Late Jurassic convex region in the southwest. 4.1 Reconstruction of the burial history and thermal evolution of hydrocarbon source rocks of a single well The Early Cretaceous concave and convex tectonic features of the Chang 8 sub-face were generally inherited from tectonic features developed in the Middle Jurassic. It also can be concluded from Figure 4 that the Chang 8 reservoir is dominantly distributed across the southwestern convex region in the Early Cretaceous basin (Fig. 7). 4.3 Early Cretaceous tectonic geomorphology of concave and convex structures of plane distribution characteristics at the base of Chang 8 The current bottom geomorphology of the Chang 8 reser- voir presents an overall large-scale pattern that includes one convex area and one concave area. Convex belts are dis- tributed in the southeast of the basin. Eastern areas, and the difference in elevation of the convex area is >700 m. At the end of the Early Cretaceous, the overall geomorphol- ogy of the base of the Chang 8 reservoir presented a S. Gao and J. Yang: Oil & Gas Science and Technology - Rev. IFP Energies nouvelles 74, 35 (2019 8 Fig. 8. Structural morphology of well connection proﬁle in the Ordos Basin. (a) At the end of the Early Cretaceous; (b) At the end of the Middle Jurassic. The area in the western part of the basin is a depression. Between the concave and convex sections are slope areas, Fig. 7. Structural morphology of Chang 8 member bottom at the end of the Early Cretaceous and reservoir distribution of the Ordos Basin. Fig. 8. Structural morphology of well connection proﬁle in the Ordos Basin. (a) At the end of the Early Cretaceous; (b) At the end of the Middle Jurassic Fig. 7. Structural morphology of Chang 8 member bottom at the end of the Early Cretaceous and reservoir distribution of the Ordos Basin. Fig. 7. Structural morphology of Chang 8 member bottom at the end of the Early Cretaceous and reservoir distribution of the Ordos Basin. Fig. 8. Structural morphology of well connection proﬁle in the Ordos Basin. (a) At the end of the Early Cretaceous; (b) At the end of the Middle Jurassic. The area in the western part of the basin is a depression. Between the concave and convex sections are slope areas, whose structures change more gently. Additionally, it is obvious that in the entire slope setting, smaller, partially convex areas are developed, such as in the south-eastern region (Fig. 1c). It can be seen from Figure 6 that the Chang 8 reservoir remains almost completely distributed on the slope, where the structure changes relatively gently. 4.7 Distribution of the sand-ground ratio in the Chang 8 member The higher sand-ground ratio of the Chang 8 member exhibits a zonal distribution, and generally features a northwest–southeast trend, revealing that sediments are mainly provided to the basin from the northeast and southwest. At present, explorations have shown that the Chang 8 reservoir is not completely distributed within the region with higher sand-ground ratios, but is found in the region with low sand-ground ratios. Recent studies have shown that a high-quality reservoir may have been formed in the palaeoslope background region during the sedimenta- tion phase, including within the turbidite sand body (Fig. 10). 4.5 Evolutionary characteristics and structure of the well-connected section C9-Jt600 The structural features of the well-connected section, C9-Jt600, indicate that the tectonic position of the Yanchang Formation is gradually rising in the north- eastern direction. The formation appears to exhibit a nearly single-dip structure. Meanwhile, the section of the well tie during the hydrocarbon generation period (i.e., the deepest period of burial at the end of the Late Cretaceous) is repre- sented by a tectonic arch developed in region W47-Dt6254 (Fig. 8). 5.2 Gentle slope of the target layers during the hydrocarbon generation period are necessary for the formation of a high-quality tight oil reservoir Fig. 9. Distribution characteristics of overpressure between Chang 7 and Chang 8 member in the Ordos Basin. The high-quality reservoir condition in the Ordos Basin is one of the basic geological conditions necessary for tight oil accumulation. However, the formation of high-quality reservoirs of tight oil also requires certain structural condi- tions. The edges of sub-level depressions (or sedimentation centres) in lacustrine basins, the leading edge slopes of palaeo-highlands and the edges of palaeohighs are often the structural conditions needed for the formation of high-quality reservoirs (Yang et al., 2012). However, not until very recently has this understanding attracted much attention in the Ordos Basin. At present, studies on the characteristics of the bottom geomorphology of lake basin, especially quantitative studies, are very rare, and the recov- ery of the bottom geomorphology of the lake basin remains based on the traditional analyses of sedimentary facies, which only roughly reﬂect the outline of the bottom geo- morphology of the lake basin. Regarding the nature of the Mesozoic lake basin on the stable craton in the Ordos Basin, it is difﬁcult to clearly resolve important issues, such as the bottom geomorphology of the lake basin and the sedimenta- tion slope break using the stratum ﬁlling structure and mor- phology revealed by current lithologic units, as well as the monotonous, almost parallel structure reﬂected by seismic data. This is also one of the reasons why the relationship between palaeostructures and oil reservoir distributions remain poorly understood. Basin simulation technologies provide an effective way for quantitatively investigating the characteristics of the bottom geomorphology of the lake basin, making it possible to further study the relationship between the bottom geomorphology and high-quality reser- voir, and the underlying causes and distributions of high quality reservoirs (Figs. 5 and 6). Fig. 9. Distribution characteristics of overpressure between Chang 7 and Chang 8 member in the Ordos Basin. Fig. 10. Distribution of the sand-ground ratio in the Chang 8 member in the Ordos Basin. 5.1 Distribution of oil reservoirs along gentle slopes of target layers during the hydrocarbon generation period In the north-western and western Ordos Basin, there are two high-value zones of differential overpressure between Chang 8 and Chang 7 strata, with the overpressure ranging from 10–20 Mpa. If the Chang 7 member is the hydrocarbon source rock, then the high overpressure region is provided by the dynamic condition of downward movement of the Chang 7 oil source (Fig. 9). Both the Chang 8 and Chang 6 oil reservoirs are located on the gentle slopes of a palaeohigh. The Chang 8 oil reservoirs are mostly distributed in the south-western part of the basin, while the Chang 6 oil reservoirs are mostly distributed 9 S. Gao and J. Yang: Oil & Gas Science and Technology - Rev. IFP Energies nouvelles 74, 35 (2019) 6 Conclusion The area where the palaeotectonic background is developed in the Chang 8 target layer during the hydrocarbon genera- tion stage is the priority area for exploration. Within this area, combining the palaeostructure with other reservoir- forming geological conditions is a feasible exploration method. The target layer structure of the Chang 8 reservoir and its evolutionary characteristics indicated that the bottom convex structure (i.e., the nose-shaped, uplifted structure) is oil-rich. Most Chang 8 tight oil reservoirs are distributed on the slopes of both sides of the bottom convex structure. This structural slope belt is a necessary condition for the formation of high-quality reservoirs (such as turbidites). The strata of the current Yanchang Formation present as a west-inclined monocline, with a dense reservoir lithol- ogy. This is also what has stiﬂed the exploration of oil reservoirs in the Yanchang Formation for so long. The results of this study show that models of the palaeo-anticli- nal structural setting of the Yanchang Formation during the high-volume hydrocarbon generation period, when con- trolled by this setting and combined with other reservoir forming conditions (e.g., downward migration and accumu- lation dynamics, reservoir, etc.), can effectively predict the tight oil distribution zones in the Ordos Basin. The tectonic geomorphology of the concave and convex structure of the bottom of the Chang 8 target layer in each hydrocarbon generation period showed that the bottom convex structure was inherited and played a decisive role in the formation of large oil reservoirs in later stages. During the hydrocarbon generation period, this prominent, nose- shaped uplift controlled the distribution of large oil ﬁelds, while the current structure has little relationship to the oil reservoirs. The palaeotectonic geomorphology of the target layers during the hydrocarbon generation period had an obvious control over the distribution of the reservoir. 5.3 Favourable structural traps allowed oil reservoir preservation in the current westward-inclined monoclines Therefore, for the Mesozoic low-permeability and ultra-low-permeability reservoirs, the overall uplift-driven erosion was just a reduc- tion of the ﬂuid pressure in the reservoir, as the ﬂuid in the reservoir of the Yanchang Formation was still essentially in a closed environment. Oil and gas reservoirs that had been formed could thus be well-preserved in the monoclinal oil reservoir of today. conditions in China and other nations, a certain structural setting (i.e., one that is conducive to the long-term accumu- lation of oil and gas and to the development of natural fractures) and ﬂuid ﬂow are prerequisites for the formation of liquid hydrocarbon sweet spots in shale layers (Yang et al., 2015). ) In the speciﬁc exploration practice, the area with paleotectonic background developed in the Chang 8 target layer in the hydrocarbon generation stage is the priority area. Within the scope of this area, combining with other reservoir forming geological conditions is a feasible explo- ration idea and method. 5.3 Favourable structural traps allowed oil reservoir preservation in the current westward-inclined monoclines Taking the well tie section C9-Jt600 as an example, Temis2D v.4.3 was used to simulate the stratigraphic evo- lution of this section. The simulation results of the section of the well tie indicate that the C9-Jt600 well tie section shows the structural position of the current strata of the Yanchang Formation is gradually uplifted in the northeast direction, and the strata present an almost single-dip struc- ture. However, during the hydrocarbon generation period (namely, the period when the strata were buried at the maximum depth, which was during the Late Cretaceous, the well tie section developed a large uplifted structure in the W47-Dt6254 region (Fig. 8). Fig. 10. Distribution of the sand-ground ratio in the Chang 8 member in the Ordos Basin. in the north-eastern and south-western parts of the basin. This convincingly indicates that the bottom characteristics of the palaeohigh during the evolution of the lake basin in the Ordos Basin controlled the distribution of oil reservoirs, and that the distributions of different oil layer groups exhi- bit differential aggregation. The reason for this difference is the difference between the palaeostructures of the Chang 8 and Chang 6 reservoirs during the hydrocarbon generation period. This geological phenomenon initially attracted the attention of researchers from around the world. Now, researchers are gradually focusing on tight oil exploration in deep lacustrine areas (Figs. 5 and 6). The Chang 7 member of the Yanchang Formation in the Ordos Basin is the major hydrocarbon generation layer of the Mesozoic. In the period of maximum burial depth of the Yanchang Formation, the Chang 7 member had already begun to generate large quantities of hydrocarbon (Fig. 4). Oil and gas usually tend to migrate upward. Therefore, the hydrocarbons generated by Chang 7 hydrocarbon S. Gao and J. Yang: Oil & Gas Science and Technology - Rev. IFP Energies nouvelles 74, 35 (2019) 10 generation layer must have exhibited downward migration dynamics in order to gather and accumulate in the underly- ing Chang 8. In areas with large differences in pressure between Chang 7 and Chang 8 strata, hydrocarbons could have migrated and accumulated downward (assuming that reservoir conditions were also available) (Fig. 10). By using the logging data of acoustic time difference, it is possible to calculate the excess stratigraphic pressure at the maximum burial depth. 5.3 Favourable structural traps allowed oil reservoir preservation in the current westward-inclined monoclines For the W47-Dt6254 well tie zone, which has a relatively large anticlinal structure that developed at the end of the Late Cretaceous, when large quantities of hydro- carbons were generated, the excess pressure difference between Chang 7 and Chang 8 strata was relatively high at 4–10 MPa (Fig. 9). This excess pressure difference might have caused Chang 7 hydrocarbons to migrate downward. The migration channel would mainly have been the connected pores in the reservoir, and it was possible for the hydrocarbons to accumulate in the more favourable reservoir segments of the Chang 8 member. At the end of the Cretaceous, the Ordos Basin began to rise and no large fault systems were formed inside of the basin. Therefore, for the Mesozoic low-permeability and ultra-low-permeability reservoirs, the overall uplift-driven erosion was just a reduc- tion of the ﬂuid pressure in the reservoir, as the ﬂuid in the reservoir of the Yanchang Formation was still essentially in a closed environment. Oil and gas reservoirs that had been formed could thus be well-preserved in the monoclinal oil reservoir of today. generation layer must have exhibited downward migration dynamics in order to gather and accumulate in the underly- ing Chang 8. In areas with large differences in pressure between Chang 7 and Chang 8 strata, hydrocarbons could have migrated and accumulated downward (assuming that reservoir conditions were also available) (Fig. 10). By using the logging data of acoustic time difference, it is possible to calculate the excess stratigraphic pressure at the maximum burial depth. For the W47-Dt6254 well tie zone, which has a relatively large anticlinal structure that developed at the end of the Late Cretaceous, when large quantities of hydro- carbons were generated, the excess pressure difference between Chang 7 and Chang 8 strata was relatively high at 4–10 MPa (Fig. 9). This excess pressure difference might have caused Chang 7 hydrocarbons to migrate downward. The migration channel would mainly have been the connected pores in the reservoir, and it was possible for the hydrocarbons to accumulate in the more favourable reservoir segments of the Chang 8 member. At the end of the Cretaceous, the Ordos Basin began to rise and no large fault systems were formed inside of the basin. References Yang H., Fu J., He H., Liu X., Zhang Z., Deng X. (2012) Formation and distribution of large low-permeability litho- logic oil regions in Huaqing, Ordos Basin, Petrol. Explor. Dev. 39, 641–648. Bai H., Pang X., Kuang L., Wan Z., Pang H., Wang X., Jia X., Song X. (2017) Depositional environment, hydrocarbon gener- ation and expulsion potential of the middle Permian Pingdi- quan source rocks based on geochemical analyses in the eastern Junggar Basin, NW China, Aust. J. Earth Sci. 64, 497–518. Yang H., Liang X.W., Niu X.B., Feng S.B., You Y. (2017) Geological conditions for continental tight oil formation and the main controlling factors for the enrichment: A case of Chang 7 Member, Triassic Yanchang Formation, Ordos Basin, NW China, Petrol. Explor. Dev. 44, 12–20. Gao S., Ren Z. (2006) Restoration of eroded thickness and its inﬂuence on thermal evolution of Upper Paleozoic source rocks in Ordos Basin, Oil Gas Geol. 02, 180–186. Yang H., Liu X., Zhang C., Han T., Hui X. (2007) The main controlling factors and distribution of low permeability lithologic reservoirs of Triassic Yanchang Formation in Ordos Basin, Lithol. Reserv. 19, 1–6. Guo S. (2010) The inﬂuence of the pre-existing topography on the depositionary systems, the development of the Lower Jurassic reservoirs and hydrocarbon accumulation in Central Western Ordos Basin, J. Petrol. Sci. Eng. 75, 129–134. Yang W., Liu G., Liu X., Feng Y., Du Y., Cheng D. (2013) The accumulation mechanism and accumulation models of oil in low permeability reservoir of Yanchang Formation in Longdong Area, Ordos Basin, Earth Sci. Front. 20, 132–139. (China University of Geosciences (Beijing); Peking University). He Z. (2003) Evolution and oil and gas in Ordos Basin, Oil Industry Press, Beijing, China. Hu T., Pang X., Wang X., Pang H., Liu Y., Wang Y., Tang L., Chen L., Pan Z., Xu J., Pang Y. (2016) Tight oil play characterisation: the lower–middle Permian Lucaogou For- mation in the Jimusar Sag, Junggar Basin, Northwest China, Austral. J. Earth Sci. 63, 349–365. Yang Z., Hou L., Tao S., Cui J., Wu S., Lin S., Pan S. (2015) Formation conditions and “sweet spot” evaluation of tight oil and shale oil, Petrol. Explor. Dev. 42, 555–565. Yao J.L., Deng X.Q., Zhao Y.D., Han T.Y., Chu M.J., Pang J.L. (2013) Characteristics of tight oil in Triassic Yanchang Formation, Ordos Basin, Petrol. Explor. Dev. 02, 150–158. Lai J., Wang G., Chai Y., Ran Y. 5.4 Tight oil exploration based on palaeostructure during the hydrocarbon generation period elucidates the relationship between the geological and causal mechanisms in hydrocarbon accumulation The characteristics of lake basin bottom geomorphology during the accumulation period controlled the formation and distribution of large tight oil reservoirs. The palaeo- slope and palaeo-uplift areas of the lake basin during the accumulation period were tight oil areas, which convinc- ingly demonstrate that conditions for high-quality reservoir formation existed in this area. Combined with the favour- able raw reservoir allocation and better subsequent diage- netic environment, this was an important area for the development of tight oil reservoirs. At present, the fundamental hydrocarbon generation condi- tions for tight oil in the Ordos Basin have been ascertained, and the effective allocation of hydrocarbon source rocks and other accumulation factors during the accumulation period will be the focus of future studies. Paying attention to the important role of structural factors (here, the bottom geomorphology of the lake basin) in the accumulation of tight oils makes it possible to directly discover and evalu- ate the sweet spot of tight oil. It has been suggested in previous studies that two major factors control the forma- tion of sweet spots. First, maturity plays an important role in controlling the distribution of liquid hydrocarbon sweet spots in shale layer, and is the primary inﬂuencing factor. Second, under the current economic and technical It is useful to clarify the key factors controlling tight oil accumulation in the Ordos Basin, and reservoir conditions are clearly key to reservoir formation. In this study, the identiﬁcation of the location of a high-quality reservoir and exploration of the relationship between its formation and the bottom geomorphology of lake basin were ﬁrst pursued. A tight oil reservoir was evaluated within palaeo- slope and uplift areas (i.e., high-quality reservoir forma- tion areas), which is conducive to guiding changes in the 11 S. Gao and J. Yang: Oil & Gas Science and Technology - Rev. IFP Energies nouvelles 74, 35 (2019) on mind-set and enriching the theory of tight oil on. Ren Z The exploration mind-set and enriching the theory of tight oil exploration. exploration mind-set and enriching the theory of tight oil exploration. Ren Z., Yu Q., Cui J., Qi K., Chen Z., Cao Z., Yang P. (2017) Thermal history and its controls on oil and gas of the Ordos Basin, Earth Sci. Front. 24, 137–148. (China University of Geosciences (Beijing); Peking University). Acknowledgments. 5.4 Tight oil exploration based on palaeostructure during the hydrocarbon generation period elucidates the relationship between the geological and causal mechanisms in hydrocarbon accumulation The authors would like to acknowledge the editorial ofﬁce of OGST and the reviewers for their comments and suggestions of this contribution. Thanks to professors Qu Hongjun and Pu Renhai from the department of geology of Northwestern University of China for their helpful suggestions during the revision of the paper. We thank the Research Insti- tute of Changqing Oilﬁeld, CNPC for software platform support. Sun G., Liu J., Liu K., Yuan W. (1985) Evolution of a major mesozoic continental basin within huabei plate and its geodynamic seting, Oil Gas Geol. 6, 280–287+350. Xu Q., Shi W., Xie X., Manger W., McGuire P., Zhang X., Wang R., Xu Z. (2016) Deep-lacustrine sandy debrites and turbidites in the lower Triassic Yanchang Formation, southeast Ordos Basin, central China: Facies distribution and reservoir quality, Mar. Pet. Geol. 77, 1095–1107. References (2016) Prediction of diagenetic facies using well logs: Evidences from upper triassic yanchang formation Chang 8 sandstones in Jiyuan Region, Ordos Basin, China, Oil Gas Sci. Technol. - Rev. IFP Energies nouvelles 71, 34. Ye L. (1983) North China platform sedimentary formation, Science Press, Beijing, China. Liu J., Xie S., Liu K., Yuan W. (1986) Evolution stacking and hydrocarbon potential of the Ordos Basin, Oil Gas Geol. 7, 356–367. Zhao J., Bai Y., Cao Q., Er C. (2012) Quasi-continuous hydrocarbon accumulation: A new pattern for large tight sand oilﬁelds in the Ordos, Oil Gas Geol. 33, 811–827. Ren Z., Li W., Liang Y., Wu X., Yu Q., Ren L., Wang W. (2014) Tight oil reservoir formation conditions and main controlling factors of Yanchang Formation in southeastern Ordos Basin, Oil Gas Geol. 35, 190–198. Zhao J., Yang X., Wu F., Shi B. (2006) Controlling of uplifts on the Triassic petroleum accumulation and distribution in North Shaanxi Slope, Ordos Basin, Acta Geol. Sin. 80, 648–655.
https://openalex.org/W4385705043	https://www.frontiersin.org/articles/10.3389/fimmu.2023.1195659/pdf	English	null	Characteristics of immunotherapy trials for nasopharyngeal carcinoma over a 15-year period	Frontiers in immunology	2,023	cc-by	7,681	TYPE Original Research PUBLISHED 09 August 2023 DOI 10.3389/fimmu.2023.1195659 OPEN ACCESS EDITED BY Takumi Kumai, Asahikawa Medical University, Japan REVIEWED BY Ying Guo, Third Afﬁliated Hospital of Sun Yat-sen University, China Xinmao Song, Fudan University, China CORRESPONDENCE Tongyu Lin linty@sysucc.org.cn He Huang huanghe@sysucc.org.cn †These authors have contributed equally to this work and share ﬁrst authorship ‡These authors have contributed equally to this work and share last authorship RECEIVED 28 March 2023 ACCEPTED 24 July 2023 PUBLISHED 09 August 2023 CITATION Huageng Huang 1†, Yuyi Yao 1†, Xinyi Deng 2†, Huawei Weng 1, Zegeng Chen 1, Le Yu 3, Zhao Wang 1, Xiaojie Fang 1, Huangming Hong 3, He Huang 1‡ and Tongyu Lin 1,3‡ Huageng Huang 1†, Yuyi Yao 1†, Xinyi Deng 2†, Huawei Weng 1, Zegeng Chen 1, Le Yu 3, Zhao Wang 1, Xiaojie Fang 1, Huangming Hong 3, He Huang 1‡ and Tongyu Lin 1,3*‡ 1Department of Oncology, Sun Yat-sen University Cancer Center, State Key Laboratory of Oncology in South China, Collaborative Innovation Center for Cancer Medicine, Guangdong Key Laboratory of Nasopharyngeal Carcinoma Diagnosis and Therapy, Guangzhou, China, 2Department of Dermatology, The First Afﬁliated Hospital of Guangzhou Medical University, Guangzhou, China, 3Department of Oncology, Senior Ward and Phase I Clinical Trial Ward, Sichuan Cancer Hospital and Institute, Sichuan Cancer Center, School of Medicine, University of Electronic Science and Technology of China, Chengdu, China Background: Immunotherapy has been a hotspot in nasopharyngeal carcinoma (NPC) in recent years. This study aimed to provide a comprehensive landscape of the characteristics of immunotherapy clinical trials in NPC and to determine whether contemporary studies are of sufﬁcient quality to demonstrate therapeutic value. Huang H, Yao Y, Deng X, Weng H, Chen Z, Yu L, Wang Z, Fang X, Hong H, Huang H and Lin T (2023) Characteristics of immunotherapy trials for nasopharyngeal carcinoma over a 15-year period. Front. Immunol. 14:1195659. doi: 10.3389/fimmu.2023.1195659 Methods: This is a cross-sectional analysis of NPC trials registered on ClinicalTrials.gov in the last 15 years (Jan 1, 2008-Nov 20, 2022). Only interventional trials with a primary purpose of treatment were included in the ﬁnal analysis. Characteristics of immunotherapy trials were compared with those of other NPC trials. Chronological shifts in NPC immunotherapy trials were also analyzed. COPYRIGHT © 2023 Huang, Yao, Deng, Weng, Chen, Yu, Wang, Fang, Hong, Huang and Lin. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). Data source and selection criteria This is a cross-sectional analysis of immunotherapy trials for NPC. We searched ClinicalTrials.gov on Nov 20, 2022 using the keyword “nasopharyngeal carcinoma”. In total, 803 registered clinical studies were identiﬁed and downloaded. We restricted our selection to interventional trials with a primary purpose of treatment that were registered between Jan 1, 2008 and Nov 20, 2022 (n = 440) (Figure 1). This study was considered exempt by the institutional review board of Sun Yat-sen University Cancer Center because it did not involve human participants. In recent years, immunotherapy has sparked a revolution in the clinical management of cancer (5, 6). NPC is regarded as a typical “immune-hot” tumor due to the expression of EBV antigen and CD4+/CD8+ T-cell target proteins (7, 8), massive lymphocytic inﬁltration (9), the expression of programmed death ligand-1 (PD- L1) up to 89-95% (10), and the presence of several key immune molecules (CD40, CD70, CD80, and CD86) that regulate T-cell activation (11). Several clinical studies on NPC immunotherapy have shown early successes (12–14). Nevertheless, aside from individual reports, the overall characteristics of NPC immunotherapy clinical trials and whether contemporary studies are of sufﬁcient quality to demonstrate the therapeutic value of immunotherapy in personalized NPC practice are unclear. Study data accuracy was ensured by independent veriﬁcation of all data by three investigators. Two oncologists (H.-G.H. and Y.- Y.Y.) manually and independently reviewed all of the selected trials, and a third author (T.-Y.L.) adjudicated any disagreements. Trials were then categorized according to treatment type, identiﬁed by the term “Intervention/treatment”, “Brief Summary”, or “Ofﬁcial Title”. If the treatment type was not clear, other registration information (e.g., “detailed description” and “eligibility”) was reviewed. We deﬁned immunotherapy trials as studies that (1) added immunotherapy to the standard of care (2); compared any treatment regimens with or without immunotherapy; (3) investigated novel immunotherapy regimens, such as new agents, usages, or dosages; (4) compared different immunotherapy regimens; and (5) evaluated interventions for immunotherapy- related complications. Although EBV-speciﬁc monoclonal antibody is a type of immunotherapy tool, some of them work in a more targeted way and partly overlap with ICIs and targeted therapy. Therefore, we excluded EBV-speciﬁc monoclonal antibody trials from the immunotherapy study. ClinicalTrials.gov, a publicly available registry and results database for human clinical studies, provides the most comprehensive clinical study information worldwide. Introduction 2022). We compared the fundamental characteristics of NPC trials focusing on immunotherapy with the characteristics of other non- immunotherapy NPC trials, and we evaluated the changes over time. Nasopharyngeal carcinoma (NPC) is an Epstein-Barr virus (EBV)-related cancer that is particularly prevalent in South East Asia and Southern China (1). Unlike other head and neck cancers, NPC is susceptible to radiotherapy (RT), which has become the mainstay of treatment for this disease. Despite advances in RT techniques and optimization of chemotherapy regimens, about 20% of patients with locally advanced NPC (LANPC) will recur (2). Moreover, recurrent and/or metastatic (R/M) NPC remains the most serious challenge because the median overall survival (OS) of these patients is only 15.7 months (3). Current conventional treatments, including RT, chemotherapy and surgery, are often accompanied by serious adverse effects and limited efﬁcacy (4). Therefore, there is an urgent need for novel treatment strategies to improve the prognosis of patients with NPC. OPEN ACCESS The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Results: Of the 440 NPC studies selected, 161 (36.6%) were immunotherapy trials and 279 (63.4%) were other NPC trials. NPC immunotherapy trials were more likely than other NPC trials to be phase 1-2 (82.6% vs. 66.7%, P < 0.001), single- arm (51.3% vs. 39.6%, P = 0.020), non-randomized (64.8% vs. 44.4%, P < 0.001), and enroll fewer than 50 participants (46.3% vs. 34.4%, P = 0.015). Blinding was used in 8.8% of NPC immunotherapy trials. Also, 90.7% of NPC immunotherapy trials were recruited nationally and 82.6% were Asia-centric. Although academic institutions and governments (72.7%) were the major sponsors of NPC trials, immunotherapy trials were more likely to be industry-funded than other NPC trials (34.2% vs. 11.5%, P < 0.001). The number of NPC immunotherapy trials increased exponentially after 2017, attributed to the exploration of immune checkpoint inhibitors. Immunotherapy combined with chemotherapy was the most commonly investigated regimen. 01 frontiersin.org Frontiers in Immunology Huang et al. 10.3389/fimmu.2023.1195659 10.3389/fimmu.2023.1195659 Conclusion: NPC immunotherapy trials over a 15-year period were predominantly exploratory. To generate high-quality evidence and advance the clinical application of immunotherapy in NPC, more attention and concerted efforts are needed. nasopharyngeal carcinoma, immunotherapy, Clinicaltrials.gov, clinical trial, immune checkpoint inhibitor Frontiers in Immunology Characteristics of included trials We extracted the following information for each trial: (1) whether the trial was registered before participant enrollment; (2) study phase; (3) sample size; (4) number of arms; (5) masking; (6) allocation methods; (7) number of centers; (8) national or international recruitment; (9) age selection; (10) funding source; (11) site location; and (12) recruitment status. As previously described (17–20), if a trial reported only one treatment arm, the allocation methods (if missing) were classiﬁed as non-randomized, and the blinding category (if missing) was classiﬁed as open-label. Funding sources were assigned as an industry, National Institutes of Health (NIH), and other academic institutions or governments based on the recorded lead sponsor and/or collaborator for each clinical trial. A trial was classiﬁed as industry-funded if its lead sponsor or one of its collaborators was from the industry with no NIH involvement or NIH-funded if its lead sponsor or one of its collaborators was from the NIH with no industry involvement (21). All other trials were classiﬁed as other- funded studies. We extracted the following information for each trial: (1) whether the trial was registered before participant enrollment; (2) study phase; (3) sample size; (4) number of arms; (5) masking; (6) allocation methods; (7) number of centers; (8) national or international recruitment; (9) age selection; (10) funding source; (11) site location; and (12) recruitment status. As previously described (17–20), if a trial reported only one treatment arm, the allocation methods (if missing) were classiﬁed as non-randomized, and the blinding category (if missing) was classiﬁed as open-label. Of the 440 NPC trials eligible for analysis, 161 (36.6%) were immunotherapy trials and 279 (63.4%) were other NPC trials (Figure 1). Table 1 shows the trial characteristics of immunotherapy and other NPC trials included in this study. Immunotherapy trials were more likely than other NPC trials to be registered before participant enrollment (116 of 161 [72.0%] vs. 128 of 279 [45.9%], P < 0.001). In addition, immunotherapy trials tended to have more phase 1-2 studies (133 of 161 [82.6%] vs. 164 of 246 [66.7%], P < 0.001) and less likely to be phase 3 studies (23 of 161 [14.3%] vs. 68 of 246 [27.6%], P = 0.002) than the other NPC trials. Furthermore, immunotherapy trials were more likely to be single-arm (80 of 156 [51.3%] vs. 109 of 275 [39.6%], P = 0.020), non-randomized (103 of 159 [64.8%] vs. Statistical analysis gene-3 (LAG-3) inhibitors; (2) adoptive cell therapy (ACT), including adoptive cell transfer of autologous cytotoxic T lymphocytes (CTLs), tumor-inﬁltrating lymphocytes (TILs), cytokine-induced killer (CIK) cells, and genetically modiﬁed cellular immunotherapy such as chimeric antigen receptor- modiﬁed T (CAR-T) cell therapy and T cell receptor-engineered T (TCR-T) cell therapy; (3) vaccines; and (4) immunomodulators, including cytokines and oncolytic viruses. The remaining eligible studies constituted the other NPC trials, investigating RT, chemotherapy, surgery, targeted therapy, etc. Descriptive statistics were primarily used to summarize the clinical trial characteristics. Categorical variables were reported as frequencies and percentages. Missing values were excluded from the analyses unless they could be inferred from other relevant data. Trial characteristics were compared using the Pearson c2 test, as well as Fisher’s exact test, if indicated. The statistical signiﬁcance level was set at P < 0.05 (two-sided). Analyses were undertaken using SPSS, version 25.0 (IBM Corp). Data source and selection criteria In 2004, the International Committee of Medical Journal Editors (ICMJE) announced a policy as a prerequisite for publication that requires the registration of clinical trials before enrolling participants (15, 16). As of Nov 20, 2022, ClinicalTrials.gov contains detailed information on more than 430 000 clinical trials conducted in over 200 countries. ClinicalTrials.gov is recognized as a promising information source for facilitating the systematic evaluation of clinical trials (16). Immunotherapy is categorized into four types in this study: (1) immune checkpoint inhibitors (ICIs), including but not limited to programmed cell death protein-1 (PD-1)/PD-L1/cytotoxic T- lymphocyte associated antigen-4 (CTLA-4)/lymphocyte activation In this study, we examined all of the interventional NPC studies registered on ClinicalTrials.gov in 15 years (Jan 1, 2008-Nov 20, Frontiers in Immunology 02 frontiersin.org Huang et al. 10.3389/fimmu.2023.1195659 FIGURE 1 Flowchart identifying trials registered on ClinicalTrials.gov from Jan 1, 2008 to Nov 20, 2022. NPC, nasopharyngeal carcinoma. FIGURE 1 Flowchart identifying trials registered on ClinicalTrials.gov from Jan 1, 2008 to Nov 20, 2022. NPC, nasopharyngeal carcinoma. Characteristics of included trials 123 of 277 [44.4%], P < 0.001), and enroll fewer than 50 participants (74 of 160 [46.3%] vs. 96 of 279 [34.4%], P = 0.015) compared with other NPC trials. Blinding was used in Funding sources were assigned as an industry, National Institutes of Health (NIH), and other academic institutions or governments based on the recorded lead sponsor and/or collaborator for each clinical trial. A trial was classiﬁed as industry-funded if its lead sponsor or one of its collaborators was from the industry with no NIH involvement or NIH-funded if its lead sponsor or one of its collaborators was from the NIH with no industry involvement (21). All other trials were classiﬁed as other- funded studies. Frontiers in Immunology 03 frontiersin.org Huang et al. 10.3389/fimmu.2023.1195659 TABLE 1 Characteristics of immunotherapy vs. other nasopharyngeal carcinoma trials. TABLE 1 Characteristics of immunotherapy vs. other nasopharyngeal carcinoma trials. Characteristic No./Total No. (%) P valueb Immunotherapy NPC trials (n = 161)a Other NPC trials (n = 279)a Registration before participant enrollment 116/161 (72.0) 128/279 (45.9) < 0.001 Phase Early Phase 1 1/161 (0.6) 2/246 (0.8) 0.002 Phase 1 28/161 (17.4) 26/246 (10.6) Phase 1/Phase 2 19/161 (11.8) 14/246 (5.7) Phase 2 86/161 (53.4) 124/246 (50.4) Phase 2/Phase 3 4/161 (2.5) 7/246 (2.9) Phase 3 23/161 (14.3) 68/246 (27.6) Phase 4 0/161 (0) 5/246 (2.0) Enrollment, No. of patients < 50 74/160 (46.3) 96/279 (34.4) 0.040 50 - 100 25/160 (15.6) 61/279 (21.9) > 100 61/160 (38.1) 122/279 (43.7) No. of study arms 1 80/156 (51.3) 109/275 (39.6) 0.001 2 59/156 (37.8) 151/275 (54.9) ≥3 17/156 (10.9) 15/275 (5.5) Masking Open-label 145/159 (91.2) 244/276 (88.4) 0.420 Blind 14/159 (8.8) 32/276 (11.6) Allocation Randomized 56/159 (35.2) 154/277 (55.6) < 0.001 Non-randomized 103/159 (64.8) 123/277 (44.4) No. of centers Single 93/161 (57.8) 183/279 (65.6) 0.102 Multiple 68/161 (42.2) 96/279 (34.4) Recruitment National 146/161 (90.7) 266/279 (95.3) 0.061 International 15/161 (9.3) 13/279 (4.7) Excludes children (aged < 18 y) 150/161 (93.2) 262/279 (93.9) 0.840 Excludes elderly (aged > 65 y) 29/161 (18.0) 65/279 (23.3) 0.227 Funding source Industry 55/161 (34.2) 32/279 (11.5) < 0.001 NIH 9/161 (5.6) 24/279 (8.6) (Continued) 04 04 Frontiers in Immunology frontiersin.org Huang et al. 10.3389/fimmu.2023.1195659 Characteristic No./Total No. Characteristics of included trials (%) P valueb Immunotherapy NPC trials (n = 161)a Other NPC trials (n = 279)a Otherc 97/161 (60.2) 223/279 (79.9) Locationsd US/Canada 32/161 (19.9) 40/279 (14.3) 0.142 Europe 10/161 (6.2) 9/279 (3.2) 0.150 Asia 133/161 (82.6) 236/279 (84.6) 0.593 Othere 4/161 (2.5) 4/279 (1.4) 0.472 Recruitment status Ongoingf 118/161 (73.3) 89/279 (31.9) < 0.001 Stopped earlyg 6/161 (3.7) 24/279 (8.6) 0.075 Completed 19/161 (11.8) 77/279 (27.6) < 0.001 Unknown 18/161 (11.2) 89/279 (31.9) < 0.001 NPC, nasopharyngeal carcinoma; NIH, National Institutes of Health; US, United States. aDifferent denominators were the number of trials with available data for different variables. bCalculated using the c2 test or the Fisher exact test if indicated. cOther Funding sources included individuals, universities, and organizations. dThe sum of the percentages may exceed 100% because categories are not mutually exclusive. eOther regions included South America, North America other than US/Canada, Central America, Oceania, and Africa. fThis status includes trials that were “not yet recruiting”, “recruiting”, “enrolling by invitation”, “active, not recruiting”, or “suspended” in the database. gThis status includes trials that were “terminated” or “withdrawn” in the database. 8.8% (14 of 159) of NPC immunotherapy trials and 90.7% (146 of 161) of NPC immunotherapy trials recruited nationally. NPC immunotherapy trials was distributed in China (120 of 161 [74.5%]), followed by the US (31 of 161 [19.3%]) and Singapore (16 of 161 [9.9%]) (Figure 2) Although other funding sources accounted for the highest proportion of immunotherapy and other NPC trials (97 of 161 [60.2%] vs. 223 of 279 [79.9%], P < 0.001), immunotherapy trials were more likely to be industry-funded than the other NPC trials (55 of 161 [34.2%] vs. 32 of 279 [11.5%], P < 0.001). Asia was the most common study location for the NPC immunotherapy trials (133 of 161 [82.6%]), followed by the United States (US)/Canada (32 of 161 [19.9%]). The most commonly recruited population for With regard to recruitment status, immunotherapy trials were more likely to be ongoing (118 of 161 [73.3%] vs. 89 of 279 [31.9%], P < 0.001) and less likely to be completed (19 of 161 [11.8%] vs. 77 of 279 [27.6%], P < 0.001) than other NPC trials. Despite the marginal difference, immunotherapy trials had a lower proportion of trials that stopped early than the other NPC trials (6 of 161 [3.7%] vs. 24 of 279 [8.6%], P = 0.075). FIGURE 2 Population distribution of nasopharyngeal carcinoma immunotherapy trials. Chronological shifts in the number of NPC immunotherapy trials (n = 123, 76.4%) (Table 2). Compared to 2008-2017, a higher proportion of immunotherapy trials were registered before ﬁrst participant enrollment in 2018-2022 (94 of 123 [76.4%] vs. 22 of 38 [57.8%], P = 0.038). Immunotherapy trials were more likely to be phase 2-3 in 2018- 2022 than in 2008-2017 (96 of 123 [78.0%] vs. 17 of 38 [44.7%], P < 0.001). Despite the marginal difference, more immunotherapy trials had a sample size of more than 100 patients in 2018-2022 than in 2008-2017 (52 of 123 [42.3%] vs. 9 of 37 [24.3%], P = 0.055). The two periods’ basic trial characteristics remained unchanged (all P > 0.05). (n = 123, 76.4%) (Table 2). Compared to 2008-2017, a higher proportion of immunotherapy trials were registered before ﬁrst participant enrollment in 2018-2022 (94 of 123 [76.4%] vs. 22 of 38 [57.8%], P = 0.038). Immunotherapy trials were more likely to be phase 2-3 in 2018- 2022 than in 2008-2017 (96 of 123 [78.0%] vs. 17 of 38 [44.7%], P < 0.001). Despite the marginal difference, more immunotherapy trials had a sample size of more than 100 patients in 2018-2022 than in 2008-2017 (52 of 123 [42.3%] vs. 9 of 37 [24.3%], P = 0.055). The two periods’ basic trial characteristics remained unchanged (all P > 0.05). Figure 3A shows chronological shifts in the number of NPC immunotherapy trials. Between 2008 and 2017, the number of NPC immunotherapy trials remained relatively stable, ranging between 2 and 7 annually. The number of NPC immunotherapy trials increased from 15 in 2018 to 32 in 2021 (P = 0.001). As of Nov 20, 2022, the number of NPC immunotherapy trials in 2022 had reached 32, the same as in 2021. The number of industry-funded NPC immunotherapy trials increased marginally from 8 of 38 studies (21.1%) in 2008-2017 to 47 of 123 studies (38.2%) in 2018-2022 (P = 0.077), but the number of NIH-funded immunotherapy trials decreased signiﬁcantly from 7 of 38 studies (18.4%) in 2008-2017 to 2 of 123 studies (1.6%) in 2018-2022 (P = 0.001). The proportion of other-funded immunotherapy trials remained stable at approximately 60%. Chronological shifts in the number of NPC immunotherapy trials In terms of study locations, there was a signiﬁcant decrease in US/ Canada centric from 16 of 38 studies (42.1%) in 2008-2017 to 16 of 123 studies (13.0%) in 2018-2022 (P < 0.001) but a signiﬁcant increase in Asia centric from 22 of 38 studies (57.9%) in 2008-2017 to 111 of 123 studies (90.2%) in 2018-2022 (P < 0.001). Furthermore, we looked at the chronological shifts in the number of different types of immunotherapy trials for NPC (Figure 3B). From 2008 to 2017, the numbers of ICIs, ACT, vaccine and immunomodulator trials were relatively stable (fewer than 5 annually). Notably, the number of ICI trials rapidly increased to 14 in 2018 and 2019 and doubled to approximately 30 from 2020 to 2022. Toripalimab was the most commonly investigated ICIs in NPC (30 of 121 [24.8%]), followed by camrelizumab (20 of 121 [16.5%]) (Figure 4). But the numbers of ACT, vaccine and immunomodulator trials stayed stagnant. Characteristics of included trials FIGURE 2 Population distribution of nasopharyngeal carcinoma immunotherapy trials. 05 05 Frontiers in Immunology frontiersin.org frontiersin.org Huang et al. 10.3389/fimmu.2023.1195659 10.3389/fimmu.2023.1195659 Immunotherapy usage in NPC trials Because NPC immunotherapy trials increased exponentially in number after 2017, we analyzed chronological shifts in the characteristics of NPC immunotherapy trials in the two periods Jan 1, 2008 to Dec 31, 2017 (n = 38, 23.6%) and Jan 1, 2018 to Nov 20, 2022 Among the 161 NPC immunotherapy trials, 46 (28.6%) evaluated single agents and 115 (71.4%) were designed to A B FIGURE 3 The number of (A) immunotherapy trials and (B) different types of immunotherapy trials for nasopharyngeal carcinoma registered on ClinicalTrials.gov between 2008 and 2022. ACT, adoptive cell therapy; ICIs, immune checkpoint inhibitors. a Observed period was Jan 1, 2022 to Nov 20, 2022. FIGURE 3 The number of (A) immunotherapy trials and (B) different types of immunotherapy trials for nasopharyngeal carcinoma registered on ClinicalTrials.gov between 2008 and 2022. ACT, adoptive cell therapy; ICIs, immune checkpoint inhibitors. a Observed period was Jan 1, 2022 to Nov 20, 2022. FIGURE 3 The number of (A) immunotherapy trials and (B) different types of immunotherapy trials for nasopharyngeal carcinoma registered on ClinicalTrials.gov between 2008 and 2022. ACT, adoptive cell therapy; ICIs, immune checkpoint inhibitors. a Observed period was Jan 1, 2022 to Nov 20, 2022. 06 Frontiers in Immunology frontiersin.org Huang et al. 10.3389/fimmu.2023.1195659 FIGURE 4 Speciﬁc drug ratios in nasopharyngeal cancer clinical trials involving immune checkpoint inhibitors. The sum of the percentages may exceed 100% because categories are not mutually exclusive. FIGURE 4 Speciﬁc drug ratios in nasopharyngeal cancer clinical trials involving immune checkpoint inhibitors. The sum of the percentages may exceed 100% because categories are not mutually exclusive. trials were predominantly phase 1-2 trials of limited sample size and tended to be single-arm, non-randomized and industry-funded. Blinding was rarely used. Asia was the major study location and clinical trials with international collaboration were lacking as well. The number of NPC immunotherapy trials increased exponentially after 2017, attributed to the exploration of ICIs. But the progress in trial design over time was slow and the basic trial characteristics largely remained unchanged. These ﬁndings raise concerns that trials evaluating the therapeutic role of immunotherapy in NPC may not be received the attention or efforts necessary to generate high-quality data. As a result, this orientation toward a less robust design may compromise evidence-based care for NPC. investigate immunotherapy combination strategies (Table 3). Immunotherapy usage in NPC trials Monotherapy (34 of 46 [73.9%]) was the most commonly explored immunotherapy regimen in single usage, followed by immunotherapy maintenance after standard treatment (11 of 46 [23.9%]). The highest proportion of immunotherapy combination strategies investigated was combination chemotherapy (39 of 115 [33.9%]), followed by radiochemotherapy (27 of 115 [23.5%]) and targeted therapy (19 of 115 [16.5%]). Frontiers in Immunology Discussion Well-designed clinical trials are desperately needed to validate the clinical applications of immunotherapy in NPC, given its promising efﬁcacy. However, with an overall low incidence rate worldwide for its unique epidemiology, NPC does not attract much attention from most research. To the best of our knowledge, this is the ﬁrst study assessing the critical characteristics of NPC immunotherapy trials over a 15-year period. By evaluating a comprehensive landscape, we found that NPC immunotherapy As an EBV-associated malignancy, NPC is frequently inﬁltrated with varied stromal cells, making its microenvironment a highly heterogeneous and suppressive harbor that protects NPC cells from drug penetration and immune attack and promotes tumor progression (22, 23). This general immune landscape of NPC renders patients suitable for immunotherapy. In the past 15 years, immunotherapy trials accounted for 36.6% of all NPC trials. Unfortunately, 82.6% of NPC immunotherapy trials were phase Frontiers in Immunology 07 frontiersin.org Huang et al. 10.3389/fimmu.2023.1195659 TABLE 2 Trend changes in characteristics of immunotherapy trials for nasopharyngeal carcinoma registered on ClinicalTrials.gov between Jan 1, 2008 to Dec 31, 2017, and Jan 1, 2018 to Nov 20, 2022. es in characteristics of immunotherapy trials for nasopharyngeal carcinoma registered on ClinicalTrials.gov between Jan 1, and Jan 1, 2018 to Nov 20, 2022. Characteristic No./Total No. (%) P valueb 2008-2017 (n = 38)a 2018-2022 (n = 123)a Registration before participant enrollment 22/38 (57.8) 94/123 (76.4) 0.038 Phase Early Phase 1 0/38 (0) 1/123 (0.8) 0.003 Phase 1 13/38 (34.2) 15/123 (12.2) Phase 1/Phase 2 8/38 (21.1) 11/123 (9.0) Phase 2 15/38 (39.5) 71/123 (57.7) Phase 2/Phase 3 0/38 (0) 4/123 (3.2) Phase 3 2/38 (5.2) 21/123 (17.1) Enrollment, No. of patients < 50 25/37 (67.6) 49/123 (39.8) 0.012 50-100 3/37 (8.1) 22/123 (17.9) > 100 9/37 (24.3) 52/123 (42.3) No. of study arms 1 18/33 (54.6) 62/123 (50.4) 0.813 2 11/33 (33.3) 48/123 (39.0) ≥3 4/33 (12.1) 13/123 (10.6) Masking Open-label 35/36 (97.2) 110/123 (89.4) 0.194 Blind 1/36 (2.8) 13/123 (10.6) Allocation Randomized 9/36 (25.0) 47/123 (38.2) 0.168 Non-randomized 27/36 (75.0) 76/123 (61.8) No. Discussion (%) Single usage 46/161 (28.6) Monotherapy 34/46 (73.9) Versus ST 1/46 (2.2) Maintenance after ST 11/46 (23.9) Combined usagea 115/161 (71.4) IT + CT 39/115 (33.9) IT + RT 7/115 (6.1) IT + surgery 2/115 (1.7) IT + TT 19/115 (16.5) Multiple IT combination 9/115 (7.9) IT + CT + RT 27/115 (23.5) IT + CT + surgery 2/115 (1.7) IT + CT +TT 8/115 (7.0) IT + CT + RT + TT 2/115 (1.7) ST, standard treatment; IT, immunotherapy; CT, chemotherapy; RT, radiotherapy; TT, targeted therapy. aTrials combining immunotherapy with other therapies simultaneously. Characteristics No./Total No. (%) Single usage 46/161 (28.6) Monotherapy 34/46 (73.9) Versus ST 1/46 (2.2) Maintenance after ST 11/46 (23.9) Combined usagea 115/161 (71.4) IT + CT 39/115 (33.9) IT + RT 7/115 (6.1) IT + surgery 2/115 (1.7) IT + TT 19/115 (16.5) Multiple IT combination 9/115 (7.9) IT + CT + RT 27/115 (23.5) IT + CT + surgery 2/115 (1.7) IT + CT +TT 8/115 (7.0) IT + CT + RT + TT 2/115 (1.7) ST, standard treatment; IT, immunotherapy; CT, chemotherapy; RT, radiotherapy; TT, targeted therapy. aTrials combining immunotherapy with other therapies simultaneously. Establishing international collaborative groups to foster research networks is an effective way to enroll more participants and improve the power of a study. However, 90.7% of NPC immunotherapy trials were conducted in only one region without sufﬁcient international collaboration. Furthermore, in contrast to the ﬁndings that the US leads global immunotherapy research with stable growth (24), Asia (82.6%) is the major study location for NPC immunotherapy trials. And the proportion of US/Canada-centric decreased from 42.1% to 13.0% (P < 0.001) while the proportion of Asia-centric increased from 57.9% to 90.2% (P < 0.001) over the two periods. It’s not surprising because the Asian centricity is concordant with the unique epidemiology of NPC as a predominantly Asian disease. The patterns of NPC (incidence, histology) are different in South East Asia and the rest of the world. It would be helpful if clinical trials could address this discrepancy in future study designs. g py Trials combining immunotherapy with other therapies simultaneously. 1-2 studies and tended to be single-arm, non-randomized, and enrolled less than 50 participants. Frontiers in Immunology Discussion of centers Single 23/38 (60.5) 70/123 (56.9) 0.693 Multiple 15/38 (39.5) 53/123 (43.1) Recruitment National 33/38 (86.8) 113/123 (91.9) 0.349 International 5/38 (13.2) 10/123 (8.1) Excludes children (aged < 18 y) 32/38 (84.2) 118/123 (95.9) 0.022 Excludes elderly (aged > 65 y) 3/38 (7.9) 26/123 (21.1) 0.089 Funding source Industry 8/38 (21.1) 47/123 (38.2) 0.001 NIH 7/38 (18.4) 2/123 (1.6) Otherc 23/38 (60.5) 74/123 (60.2) (Continued) 08 08 Frontiers in Immunology frontiersin.org Huang et al. 10.3389/fimmu.2023.1195659 Characteristic No./Total No. (%) P valueb 2008-2017 (n = 38)a 2018-2022 (n = 123)a Locationsd US/Canada 16/38 (42.1) 16/123 (13.0) < 0.001 Europe 5/38 (13.2) 5/123 (4.1) 0.057 Asia 22/38 (57.9) 111/123 (90.2) < 0.001 Othere 0/38 (0) 4/123 (3.3) 0.574 Recruitment status Ongoingf 8/38 (21.1) 110/123 (89.4) < 0.001 Stopped earlyg 2/38 (5.2) 4/123 (3.3) 0.627 Completed 16/38 (42.1) 3/123 (2.4) < 0.001 Unknown 12/38 (31.6) 6/123 (4.9) < 0.001 NIH, National Institutes of Health; US, United States. aDifferent denominators were the number of trials with available data for different variables. bCalculated using the c2 test or the Fisher exact test if indicated. cOther Funding sources included individuals, universities, and organizations. dThe sum of the percentages may exceed 100% because categories are not mutually exclusive. eOther regions included South America, North America other than US/Canada, Central America, Oceania, and Africa. fThis status includes trials that were “not yet recruiting”, “recruiting”, “enrolling by invitation”, “active, not recruiting”, or “suspended” in the database. gThis status includes trials that were “terminated” or “withdrawn” in the database. TABLE 3 Immunotherapy usage in nasopharyngeal carcinoma clinical trials. they were studying more highly innovative expensive cellular-based studies where funding was often inadequate for larger studies with more patients. In addition, the well-deﬁned geographic distribution of NPC might further limit clinicians from conducting large-scale immunotherapy trials. Similarly, Xu et al. (24) tracked the evolving landscape of global immuno-oncology trials in 2007-2019 and found that most immunotherapy trials worldwide were phase 2 studies. Fortunately, NPC immunotherapy trials in 2018-2022 were more likely to be phase 2-3 (78.0% vs. 44.7%, P < 0.001) and had a sample size of more than 100 patients (42.3% vs. 24.3%, P = 0.055) than in 2008-2017. However, the other basic trial characteristics did not improve in an obvious manner over time. Characteristics No./Total No. Discussion Actually, the high proportion of single-arm, non-randomized, early-phase studies may either be because these studies are exploratory, hypotheses generating to fuel future randomized trials involving more patients, or because Generally, the lengthy duration and high cost of immunotherapy trials may suppress industry enthusiasm. However, our ﬁndings showed that NPC immunotherapy trials Frontiers in Immunology 09 frontiersin.org Huang et al. 10.3389/fimmu.2023.1195659 Limitations of this study should also be acknowledged. First, not all investigators choose ClinicalTrials.gov to register their projects. There are many alternative registries available around the world (39). Nevertheless, ClinicalTrials.gov is the most robust database to date, accounting for 70–80% of the unique clinical trials recorded by the World Health Organization (39). Second, partial NPC trials have not yet been registered on ClinicalTrials.gov, which hindered us from more fully reﬂecting current global trends in NPC immunotherapy trials. Third, the National Library of Medicine, which operates ClinicalTrials.gov, is unable to validate all registered data. The accuracy of the data relies on the study sponsor. Fourth, we did not include noninterventional trials in our analysis. were more likely to be industry-funded compared with other NPC trials (34.2% vs. 11.5%, P < 0.001) and the proportion has increased over time (21.1% vs. 38.2%, P = 0.077). It indicates a large market potential in the ﬁeld of NPC immunotherapy, thus raising ﬁnancial interests and industrial enthusiasm for sponsorship of such trials. Moreover, 60.2% of NPC immunotherapy trials were other-funded, and this proportion has remained stable over time. It implies that academic institutions and governments continue to play an important role in supporting immunotherapy clinical research for NPC and shoulder vital public health responsibility. Still, allocating more resources to NPC immunotherapy from all relevant parties is essential to improve the effective leveraging of the constrained resources. In conclusion, NPC Immunotherapy trials over a 15-year period have been largely exploratory. Advancing the clinical application of immunotherapy in NPC requires more attention and concerted efforts to improve the quality of trials. It is noteworthy that there was an increasing number of NPC immunotherapy trials after 2017. Actually, this reﬂects more recent successes in other major tumor types, and therefore there is an increasing interest in studying this intervention in an EBV-related tumor type like NPC that does not have many mutational targets. However, only the number of ICI trials increased signiﬁcantly, while the numbers of ACT, vaccine and immunomodulator trials remained stagnant. Discussion A potential explanation is that ICIs as pan- cancerous antitumor agents were found to have equally promising efﬁcacy in NPC, thus spurring enthusiasm for research. Furthermore, the recognition of ICI-based immunotherapy by the 2018 Nobel Prize in Physiology or Medicine might have further increased researchers’ interest in the exploration of ICIs in NPC. In contrast, the exploration and further application of EBV-speciﬁc ACTs and vaccines and immunomodulators were hampered by the lack of speciﬁc and effective targets, generally low and transient immune responses, technical limitations and ﬁnancial shortages. According to the results of published NPC studies (Supplementary Table 1), ICIs monotherapy achieves 17.1-34.0% of the objective response rate in the second or later-line treatment of R/M NPC (14, 25–30). In the ﬁrst-line treatment, the addition of ICIs to chemotherapy also signiﬁcantly improved progression-free survival and OS in R/M NPC (31–33). Further studies are needed to assess the therapeutic value of ICIs in LANPC and early-stage disease. Notably, CAR-T/TCR-T cell therapy and antibody-drug conjugates may be another promising immunotherapy modality for NPC, as they have shown promising efﬁcacy in a variety of other cancers (34, 35). Therefore, concerted efforts by oncologists, sponsors and other concerned parties are still needed to advance the development of immunotherapy for NPC. Data availability statement Publicly available datasets were analyzed in this study. This data can be found here: https://clinicaltrials.gov/. Funding This study is supported by the National Natural Science Foundation of China-Science and Technology Development Fund, Macau SAR (No. 81661168011), and the Regional Innovation and Cooperation Project of Sichuan Province (No. 2021YFQ0037). Integration with conventional treatment modalities is one of the trends in immunotherapy. In this study, we found that the most commonly investigated immunotherapy regimen in NPC was combination chemotherapy, followed by combination radiochemotherapy. A recently published study reported on the promising antitumor activity and a manageable toxicity proﬁle of immunotherapy combined with antiangiogenic therapy in R/M NPC (36). In addition, future NPC studies could consider more novel combination strategies to enhance the clinical responses, for example, ICIs combined with ACT (37) or CAR-T cell therapy combined with the oncolytic virus (38). Author contributions TL had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. HGH, YY, and XD contributed equally to this work. Concept and design: HGH, YY, XD, HH, and TL. Acquisition, analysis, or interpretation of data: HGH, YY, XD, and TL. Drafting of the manuscript: HGH, YY, and XD. Critical revision of the manuscript for important intellectual content: All authors. Statistical analysis: HGH and XD. Obtained funding: HH and TL. Administrative, technical, or material support: HH and TL. Supervision: HH and TL. All authors contributed to the article and approved the submitted version. Publisher’s note The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/ﬁmmu.2023.1195659/ full#supplementary-material All claims expressed in this article are solely those of the authors and do not necessarily represent those of their afﬁliated organizations, or those of the publisher, the editors and the References 17. Chen YP, Lv JW, Liu X, Zhang Y, Guo Y, Lin AH, et al. The landscape of clinical trials evaluating the theranostic role of PET imaging in oncology: insights from an analysis of clinicalTrials. gov Database. Theranostics (2017) 7(2):390–9. doi: 10.7150/thno.17087 1. Wong KCW, Hui EP, Lo KW, Lam WKJ, Johnson D, Li L, et al. Nasopharyngeal carcinoma: an evolving paradigm. Nat Rev Clin Oncol (2021) 18(11):679–95. doi: 10.1038/s41571-021-00524-x 1. Wong KCW, Hui EP, Lo KW, Lam WKJ, Johnson D, Li L, et al. Nasopharyngeal carcinoma: an evolving paradigm. Nat Rev Clin Oncol (2021) 18(11):679–95. doi: 10.1038/s41571-021-00524-x 2. Huang CL, Guo R, Li JY, Xu C, Mao YP, Tian L, et al. Nasopharyngeal carcinoma treated with intensity-modulated radiotherapy: clinical outcomes and patterns offailure among subsets of 8th AJCC stage IVa. Eur Radiol (2020) 30(2):816–22. doi: 10.1007/ s00330-019-06500-5 18. Califf RM, Zarin DA, Kramer JM, Sherman RE, Aberle LH, Tasneem A. Characteristics of clinical trials registered in ClinicalTrials. gov 2007-2010. JAMA (2012) 307(17):1838–47. doi: 10.1001/jama.2012.3424 19. Hirsch BR, Califf RM, Cheng SK, Tasneem A, Horton J, Chiswell K, et al. Characteristics of oncology clinical trials: insights from a systematic analysis of ClinicalTrials. gov. JAMA Internal Med (2013) 173(11):972–9. doi: 10.1001/ jamainternmed.2013.627 3. Prawira A, Oosting SF, Chen TW, Delos Santos KA, Saluja R, Wang L, et al. Systemic therapies for recurrent or metastatic nasopharyngeal carcinoma: a systematic review. Br J cancer. (2017) 117(12):1743–52. doi: 10.1038/bjc.2017.357 20. Liu X, Zhang Y, Tang LL, Le QT, Chua MLK, Wee JTS, et al. Characteristics of radiotherapy trials compared with other oncological clinical trials in the past 10 years. JAMA Oncol (2018) 4(8):1073–9. doi: 10.1001/jamaoncol.2018.0887 4. Poh SS, Soong YL, Sommat K, Lim CM, Fong KW, Tan TW, et al. Retreatment in locally recurrent nasopharyngeal carcinoma: Current status and perspectives. Cancer Commun (London England). (2021) 41(5):361–70. doi: 10.1002/cac2.12159 21. Liu X, Zhang Y, Li WF, Vokes E, Sun Y, Le QT, et al. Evaluation of oncology trial results reporting over a 10-year period. JAMA network Open (2021) 4(5):e2110438. doi: 10.1001/jamanetworkopen.2021.10438 5. McNutt M. Cancer immunotherapy. Science (2013) 342(6165):1417. doi: 10.1126/ science.1249481 6. Hargadon KM, Johnson CE, Williams CJ. Immune checkpoint blockade therapy for cancer: An overview of FDA-approved immune checkpoint inhibitors. Int immunopharmacol (2018) 62:29–39. doi: 10.1016/j.intimp.2018.06.001 22. Tsang CM, Lui VWY, Bruce JP, Pugh TJ, Lo KW. Translational genomics of nasopharyngeal cancer. Semin Cancer Biol (2020) 61:84–100. doi: 10.1016/ j.semcancer.2019.09.006 7. Conﬂict of interest The authors declare that the research was conducted in the absence of any commercial or ﬁnancial relationships that could be construed as a potential conﬂict of interest. Frontiers in Immunology frontiersin.org 10 Huang et al. Huang et al. 10.3389/fimmu.2023.1195659 reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher. The reviewer YG declared a shared parent afﬁliation with the authors HGH, YY, HW, ZC, ZW, XF, HH, and TL to the handling editor at the time of review. References Delayed-type hypersensitivity (DTH) immune response related with EBV-DNA in nasopharyngeal carcinoma treated with autologous dendritic cell vaccination after radiotherapy. J immunother (Hagerstown Md: 1997). (2013) 36(3):208–14. doi: 10.1097/CJI.0b013e31828bd87b 12. Li F, Song D, Lu Y, Zhu H, Chen Z, He X. Delayed-type hypersensitivity (DTH) immune response related with EBV-DNA in nasopharyngeal carcinoma treated with autologous dendritic cell vaccination after radiotherapy. J immunother (Hagerstown Md: 1997). (2013) 36(3):208–14. doi: 10.1097/CJI.0b013e31828bd87b 28. Wang FH, Wei XL, Feng J, Li Q, Xu N, Hu XC, et al. Efﬁcacy, safety, and correlative biomarkers of toripalimab in previously treated recurrent or metastatic nasopharyngeal carcinoma: A phase II clinical trial (POLARIS-02). J Clin Oncol (2021) 39(7):704–12. doi: 10.1200/jco.20.02712 29. Even C, Wang HM, Li SH, Ngan RK, Dechaphunkul A, Zhang L, et al. Randomized study of spartalizumab (PDR001), an anti-PD-1 antibody, versus chemotherapy in patients with recurrent/metastatic nasopharyngeal cancer. Clin Cancer Res (2021) 27(23):6413–23. doi: 10.1158/1078-0432.Ccr-21-0822 13. Smith C, Lee V, Schuessler A, Beagley L, Rehan S, Tsang J, et al. Pre-emptive and therapeutic adoptive immunotherapy for nasopharyngeal carcinoma: Phenotype and effector function of T cells impact on clinical response. Oncoimmunology (2017) 6(2): e1273311. doi: 10.1080/2162402x.2016.1273311 13. Smith C, Lee V, Schuessler A, Beagley L, Rehan S, Tsang J, et al. Pre-emptive and therapeutic adoptive immunotherapy for nasopharyngeal carcinoma: Phenotype and effector function of T cells impact on clinical response. Oncoimmunology (2017) 6(2): e1273311. doi: 10.1080/2162402x.2016.1273311 14. Ma BBY, Lim WT, Goh BC, Hui EP, Lo KW, Pettinger A, et al. Antitumor activity of nivolumab in recurrent and metastatic nasopharyngeal carcinoma: an international, multicenter study of the mayo clinic phase 2 consortium (NCI-9742). J Clin Oncol (2018) 36(14):1412–8. doi: 10.1200/jco.2017.77.0388 14. Ma BBY, Lim WT, Goh BC, Hui EP, Lo KW, Pettinger A, et al. Antitumor activity of nivolumab in recurrent and metastatic nasopharyngeal carcinoma: an international, multicenter study of the mayo clinic phase 2 consortium (NCI-9742). J Clin Oncol (2018) 36(14):1412–8. doi: 10.1200/jco.2017.77.0388 30. Shi Y, Qin X, Peng X, Zeng A, Li J, Chen C, et al. Efﬁcacy and safety of KL-A167 in previously treated recurrent or metastatic nasopharyngeal carcinoma: a multicenter, single-arm, phase 2 study. Lancet regional Health Western Paciﬁc (2023) 31:100617. doi: 10.1016/j.lanwpc.2022.100617 31. Yang Y, Qu S, Li J, Hu C, Xu M, Li W, et al. References Smith C, Wakisaka N, Crough T, Peet J, Yoshizaki T, Beagley L, et al. Discerning regulation of cis- and trans-presentation of CD8+ T-cell epitopes by EBV-encoded oncogene LMP-1 through self-aggregation. Blood (2009) 113(24):6148–52. doi: 10.1182/blood-2009-02-203687 23. Chen YP, Yin JH, Li WF, Li HJ, Chen DP, Zhang CJ, et al. Single-cell transcriptomics reveals regulators underlying immune cell diversity and immune subtypes associated with prognosis in nasopharyngeal carcinoma. Cell Res (2020) 30 (11):1024–42. doi: 10.1038/s41422-020-0374-x 8. Münz C, Bickham KL, Subklewe M, Tsang ML, Chahroudi A, Kurilla MG, et al. Human CD4(+) T lymphocytes consistently respond to the latent Epstein-Barr virus nuclear antigen EBNA1. J Exp Med (2000) 191(10):1649–60. doi: 10.1084/ jem.191.10.1649 24. Xu C, Zhang S, Zhang Y, Tang SQ, Fang XL, Zhu GL, et al. Evolving landscape and academic attitudes toward the controversies of global immuno-oncology trials. Int J cancer. (2021) 149(1):108–18. doi: 10.1002/ijc.33503 9. Lee AZE, Tan LSY, Lim CM. Cellular-based immunotherapy in Epstein-Barr virus induced nasopharyngeal cancer. Oral Oncol (2018) 84:61–70. doi: 10.1016/ j.oraloncology.2018.07.011 25. Hsu C, Lee SH, Ejadi S, Even C, Cohen RB, Le Tourneau C, et al. Safety and antitumor activity of pembrolizumab in patients with programmed death-ligand 1- positive nasopharyngeal carcinoma: results of the KEYNOTE-028 study. J Clin Oncol (2017) 35(36):4050–6. doi: 10.1200/jco.2017.73.3675 10. Outh-Gauer S, Alt M, Le Tourneau C, Augustin J, Broudin C, Gasne C, et al. Immunotherapy in head and neck cancers: A new challenge for immunologists, pathologists and clinicians. Cancer Treat Rev (2018) 65:54–64. doi: 10.1016/ j.ctrv.2018.02.008 26. Fang W, Yang Y, Ma Y, Hong S, Lin L, He X, et al. Camrelizumab (SHR-1210) alone or in combination with gemcitabine plus cisplatin for nasopharyngeal carcinoma: results from two single-arm, phase 1 trials. Lancet Oncol (2018) 19(10):1338–50. doi: 10.1016/s1470-2045(18)30495-9 11. Agathanggelou A, Niedobitek G, Chen R, Nicholls J, Yin W, Young LS. Expression of immune regulatory molecules in Epstein-Barr virus-associated nasopharyngeal carcinomas with prominent lymphoid stroma. Evidence for a functional interaction between epithelial tumor cells and inﬁltrating lymphoid cells. Am J pathology. (1995) 147(4):1152–60. 27. Yang Y, Zhou T, Chen X, Li J, Pan J, He X, et al. Efﬁcacy, safety, and biomarker analysis of Camrelizumab in Previously Treated Recurrent or Metastatic Nasopharyngeal Carcinoma (CAPTAIN study). J Immunother Cancer (2021) 9(12): e003790. doi: 10.1136/jitc-2021-003790 12. Li F, Song D, Lu Y, Zhu H, Chen Z, He X. Frontiers in Immunology 33. Yang Y, Pan J, Wang H, Zhao Y, Qu S, Chen N, et al. Tislelizumab plus chemotherapy as ﬁrst-line treatment for recurrent or metastatic nasopharyngeal cancer: A multicenter phase 3 trial (RATIONALE-309). Cancer Cell (2023) 41 (6):1061–72.e4. doi: 10.1016/j.ccell.2023.04.014 36. Ding X, Zhang WJ, You R, Zou X, Wang ZQ, Ouyang YF, et al. Camrelizumab plus apatinib in patients with recurrent or metastatic nasopharyngeal carcinoma: an 34. Norberg SM, Hinrichs CS. Engineered T cell therapy for viral and non-viral epithelial cancers. Cancer Cell (2023) 41(1):58–69. doi: 10.1016/j.ccell.2022.10.016 35. Thomas A, Teicher BA, Hassan R. Antibody-drug conjugates for cancer therapy. Lancet Oncol (2016) 17(6):e254–e62. doi: 10.1016/s1470-2045(16)30030-4 multicenter randomized phase 3 trial. Nat Med (2021) 27(9):1536–43. doi: 10.1038/ s41591-021-01444-0 33. Yang Y, Pan J, Wang H, Zhao Y, Qu S, Chen N, et al. Tislelizumab plus chemotherapy as ﬁrst-line treatment for recurrent or metastatic nasopharyngeal cancer: A multicenter phase 3 trial (RATIONALE-309). Cancer Cell (2023) 41 (6):1061–72.e4. doi: 10.1016/j.ccell.2023.04.014 34. Norberg SM, Hinrichs CS. Engineered T cell therapy for viral and non-viral epithelial cancers. Cancer Cell (2023) 41(1):58–69. doi: 10.1016/j.ccell.2022.10.016 35. Thomas A, Teicher BA, Hassan R. Antibody-drug conjugates for cancer therapy. Lancet Oncol (2016) 17(6):e254–e62. doi: 10.1016/s1470-2045(16)30030-4 36. Ding X, Zhang WJ, You R, Zou X, Wang ZQ, Ouyang YF, et al. Camrelizumab plus apatinib in patients with recurrent or metastatic nasopharyngeal carcinoma: an multicenter randomized phase 3 trial. Nat Med (2021) 27(9):1536–43. doi: 10.1038/ s41591-021-01444-0 37. Smith C, McGrath M, Neller MA, Matthews KK, Crooks P, Le Texier L, et al. Complete response to PD-1 blockade following EBV-speciﬁc T-cell therapy in metastatic nasopharyngeal carcinoma. NPJ Precis Oncol (2021) 5(1):24. doi: 10.1038/ s41698-021-00162-7 39. Zarin DA, Tse T, Williams RJ, Califf RM, Ide NC. The ClinicalTrials.gov results database–update and key issues. New Engl J Med (2011) 364(9):852–60. doi: 10.1056/ NEJMsa1012065 38. Rezaei R, Esmaeili Gouvarchin Ghaleh H, Farzanehpour M, Dorostkar R, Ranjbar R, Bolandian M, et al. Combination therapy with CAR T cells and oncolytic viruses: a new era in cancer immunotherapy. Cancer Gene Ther (2022) 29(6):647–60. doi: 10.1038/s41417-021-00359-9 Huang et al. open-label, single-arm, phase II study. J Clin Oncol (2023), Jco2201450. doi: 10.1200/ jco.22.01450 References Camrelizumab versus placebo in combination with gemcitabine and cisplatin as ﬁrst-line treatment for recurrent or metastatic nasopharyngeal carcinoma (CAPTAIN-1st): a multicentre, randomised, double-blind, phase 3 trial. Lancet Oncol (2021) 22(8):1162–74. doi: 10.1016/s1470- 2045(21)00302-8 15. De Angelis C, Drazen JM, Frizelle FA, Haug C, Hoey J, Horton R, et al. Clinical trial registration: a statement from the International Committee of Medical Journal Editors. New Engl J Med (2004) 351(12):1250–1. doi: 10.1056/NEJMe048225 15. De Angelis C, Drazen JM, Frizelle FA, Haug C, Hoey J, Horton R, et al. Clinical trial registration: a statement from the International Committee of Medical Journal Editors. New Engl J Med (2004) 351(12):1250–1. doi: 10.1056/NEJMe048225 16. De Angelis C, Drazen JM, Frizelle FA, Haug C, Hoey J, Horton R, et al. Clinical trial registration: a statement from the International Committee of Medical Journal Editors. Lancet (London England). (2004) 364(9438):911–2. doi: 10.1016/s0140-6736 (04)17034-7 16. De Angelis C, Drazen JM, Frizelle FA, Haug C, Hoey J, Horton R, et al. Clinical trial registration: a statement from the International Committee of Medical Journal Editors. Lancet (London England). (2004) 364(9438):911–2. doi: 10.1016/s0140-6736 (04)17034-7 32. Mai HQ, Chen QY, Chen D, Hu C, Yang K, Wen J, et al. Toripalimab or placebo plus chemotherapy as ﬁrst-line treatment in advanced nasopharyngeal carcinoma: a 32. Mai HQ, Chen QY, Chen D, Hu C, Yang K, Wen J, et al. Toripalimab or placebo plus chemotherapy as ﬁrst-line treatment in advanced nasopharyngeal carcinoma: a Frontiers in Immunology 11 frontiersin.org Huang et al. 10.3389/fimmu.2023.1195659 open-label, single-arm, phase II study. J Clin Oncol (2023), Jco2201450. doi: 10.1200/ jco.22.01450 37. Smith C, McGrath M, Neller MA, Matthews KK, Crooks P, Le Texier L, et al. Complete response to PD-1 blockade following EBV-speciﬁc T-cell therapy in metastatic nasopharyngeal carcinoma. NPJ Precis Oncol (2021) 5(1):24. doi: 10.1038/ s41698-021-00162-7 38. Rezaei R, Esmaeili Gouvarchin Ghaleh H, Farzanehpour M, Dorostkar R, Ranjbar R, Bolandian M, et al. Combination therapy with CAR T cells and oncolytic viruses: a new era in cancer immunotherapy. Cancer Gene Ther (2022) 29(6):647–60. doi: 10.1038/s41417-021-00359-9 39. Zarin DA, Tse T, Williams RJ, Califf RM, Ide NC. The ClinicalTrials.gov results database–update and key issues. New Engl J Med (2011) 364(9):852–60. doi: 10.1056/ NEJMsa1012065 39. Zarin DA, Tse T, Williams RJ, Califf RM, Ide NC. The ClinicalTrials.gov results database–update and key issues. New Engl J Med (2011) 364(9):852–60. doi: 10.1056/ NEJMsa1012065 12 12 Frontiers in Immunology frontiersin.org
https://openalex.org/W3023624730	https://zenodo.org/record/5760132/files/10%2021720.pdf	English	null	Novel switching technique for five leg inverter in dual motor control	Indonesian journal of electrical engineering and computer science	2,020	cc-by	4,162	Indonesian Journal of Electrical Engineering and Computer Science Vol. 19, No. 2, August 2020, pp. 644~651 ISSN: 2502-4752, DOI: 10.11591/ijeecs.v19.i2.pp644-651 Indonesian Journal of Electrical Engineering and Computer Science Vol. 19, No. 2, August 2020, pp. 644~651 ISSN: 2502-4752, DOI: 10.11591/ijeecs.v19.i2.pp644-651 Indonesian Journal of Electrical Engineering and Computer Science Vol. 19, No. 2, August 2020, pp. 644~651 ISSN: 2502-4752, DOI: 10.11591/ijeecs.v19.i2.pp644-651  644  644  ABSTRACT This work presents a novel switching technique for five leg inverter in dual motor control. As the technology advances in industry, requirements in reducing the cost plays an important role with reliable product design. In conventional method, the six legs are used in 2- three phase Voltage Source Inverters (VSI) to control 2 motors. This proposed technique will give the improved performance of speed control for dual motor control using Five Leg Inverter (FLI). New proposed method suggests to use 5-inverter legs instead of 6-inverter legs to control 2 induction motors. New Switching technique proposed in FLI system is designed in effective way that improved performance and Total Harmonic Distortion of ~23% achieved. The load sharing on common leg is called Common Mode (CM) of operation. In this new method, closed loop control designed by using space vector pulse width modulation (SVPWM) and Direct Torque Control (DTC) in FLI Technology. With this new method smooth speed regulation is achieved when load torque is changed. THD% for CM-FLI is reduced when compared with convetional FLI technique. The new Switching technique is controlled in effective way that the common leg is not overloaded and able to drive both the induction motors independently at required speeds. Proposed switching technique verified at different operating speeds with No load and rated torque. Simulation results computed using MATLAB/SIMULINK Software. Corresponding Author: Jayaprakash Sabarad, Research Scholar, Department of Electrical Engineering, VTU RRC Belagavi, PO#590018 Karnataka, Belagavi, India. Email: jpsabarad@gmail.com Corresponding Author: Jayaprakash Sabarad, Research Scholar, Department of Electrical Engineering, VTU RRC Belagavi, PO#590018 Karnataka, Belagavi, India. Email: jpsabarad@gmail.com Novel switching technique for five leg inverter in dual motor control Jayaprakash Sabarad1, G.H. Kulkarni2 1Department of Electrical Engineering, Visvesvaraya Technological University, India 2Department of Electrical Engineering, Jain College of Engineering, India Jayaprakash Sabarad1, G.H. Kulkarni2 1Department of Electrical Engineering, Visvesvaraya Technological University, India 2Department of Electrical Engineering, Jain College of Engineering, India Copyright © 2020 Institute of Advanced Engineering and Science. All rights reserved. 1. INTRODUCTION Present trend in industries are looking for the advanced control methodology with precise speed control. The industries looking for minimising the cost with improved performance in drive control system. In this regard the proposed work is carried out to reduce cost and to give improved performance in the field of 3-phase Induction Motor (IM) control. As the number of inverter switching devices increases more, then it leads to complex control design [1, 2] and also affects reliability of the system. The proposed solution helps in reducing the number of switching elements, thus reduces the cost, and size of the system [3, 4]. The proposed solution is having the 5-legs inverters instead of conventional 6-legs inverter for dual motor control. The main aim of this five leg is to have reduced switching elements from 12 to 10 [5]. The new technology uses only 10 switching elements with switching combination is up to 25. Closed loop control is done by interfacing FLI -SVPWM with DTC of induction motor [6, 7]. The main drawback of those conventional methodologies are, in case of dual voltage modulation (DVM) two consecutive switching periods are applied for each motor, due to this DC voltage restricted to half. In case of modulation block method (MBM) applied reference voltage shall be known prior to Journal homepage: http://ijeecs.iaescore.com ISSN: 2502-4752 645 Indonesian J Elec Eng & Comp Sci  application. Whereas in inversion table method (IVM) method bigger look up table need to be designed and difficult to be implemented on digital processors [8-10]. application. Whereas in inversion table method (IVM) method bigger look up table need to be designed and difficult to be implemented on digital processors [8-10]. In this novel design of switching technique in FLI is common mode of operation. This common leg shares load of each motor and also supply voltage to both motors. Hence it is important to have the optimal switching technique to be used in such a manner that the common leg does not get overloaded. The novel switching technique is called CM-FLI technology gives improved performance of speed control and is explained in detail in below section. 2. CONTROL TECHNIQUE AND SWITCHING METHODOLOGY 2.1. Mathematical equations for DTC of induction motor In Conventional DTC the stator flux and torque are regulated within limited hysteresis bands and which in turn switch the VSI to apply appropriate stator voltage. In conventional method the flux and torque errors are directly used to generate the switching pulses to VSI [11-13]. An improved DTC is proposed here is reference stator voltage vectors are produced by torque and flux errors as given in below (1) to (12), which in turn modulated by means of SVPWM technique [14, 15]. The field component is aligned to d-axis and torque component is aligned to q-axis and both are orthogonal to each other. The (1) to (6) gives the motor stator voltage from stator reference frame. The flux and torque with respect to load torque and inertia is given as below. (1) ( ) (2) (3) (4) (5) (6) (1) ( ) (2) (3) (4) (5) (6) ( ) (6) Where, Where, is stator applied voltage. and are rotor and stator resistances. are stator flux linkages. are stator and rotor current. Mutual inductance is . is stator applied voltage. and are rotor and stator resistances. are stator flux linkages. are stator and rotor current. Mutual inductance is . are stator and rotor self-inductances. , are load torque and electromagnetic torque and, are reference field and rotor speed. Model equations can be represented in stationary stator reference frames to estimate the Torque and flux linkages as below. In stationary reference frame α is magnetic linkages of stator coil resistance R. The estimation of stator flux as, ∫ (7) Magnitude of stator flux as, \| \| √ + ) (8) Estimation of position vector as, Estimation of position vector as, (9) (9) Estimation of motor torque as, Estimation of motor torque as, - ) (10) - ) (10) e, Te is the reference electromagnetic torque estimated from the motor. Where, Te is the reference electromagnetic torque estimated from the motor. Where, Te is the reference electromagnetic torque estimated from the motor. Novel switching technique for five leg inverter in dual motor control (Jayaprakash Sabarad) Novel switching technique for five leg inverter in dual motor control (Jayaprakash Sabarad)  ISSN: 2502-4752 646 (11) (12) (11) (12) The flux error and torque error are used to produce the decoupled stator reference vector by means of PI regulator. With the above information the reference voltage then be converted in to voltage vector by means of stator co-ordinates reference co-ordinate transformation. DTC technique used in variable speed control of asynchronous motor fed by VSI. Using Clark transformation, measured 3-phase stator currents and voltages are then converted to 2-phase co-ordinates (α, β). The 2-phase co-ordinates then converted to co-ordinates [16-18]. 3. COMMON MODE-FIVE LEG INVERTER (CM-FLI) USING SVPWM The switching techniques for the 3-leg inverter is common in many applications. In FLI -VSI method the critical part is computing common mode voltage , which shares common leg between two inverters. Common leg voltage shares the voltage between two motors. The reference signal for both the inverters are computed in such a way that reference signals does not include the zero sequence component. The reference voltages are respectively. For reference voltages, the zero sequence voltages are getting added to reduce lower currents harmonics and also to increase the available modulation index given in (19) to (24). ( ) (20) ( ) ( ⁄ ) (21) ( ) ( ⁄ ) (22) ( ) (20) ( ) ( ⁄ ) (21) ( ) ( ⁄ ) (22) (20) (21) (22) Similarly for the second inverter machine the reference signal is computed by, Similarly for the second inverter machine the reference signal is computed by, ( ) (23) ( ) ( ⁄ ) (24) ( ) ( ⁄ ) (25) ( ) (23) ( ) ( ⁄ ) (24) ( ) ( ⁄ ) (25) (23) Where, are voltage input signals fed to SVPWM technique. Now the six output voltages have to be converted to five called five leg inverter output. The main switching technique is called common mode technique. In this common mode technique, inverter output voltages are only five and one common leg is present between the two inverters. The load on the common leg inverter will be more, so in order to avoid the extra load, the switching technique used here is combination of SVPWM and also common mode techniques. This common leg to share the load current and which results in little more current flows in common leg than the other four legs. A novel switching technique of five leg VSI is shown in Figure 2. The common mode of operation of 2 three-phase modulators are added in such a manner, that number of switching signals gets reduced from six to five as shown in Figure 2. Where are the switching pulses for the 5-leg inverter. SVPWM and CM mode combined to achieve the right inverter pulse to FLI. The five- leg VSI with common leg mode of operation for dual three phase induction motor is shown in Figure 3. 2.2. SVPWM technique The 2-level SVPWM technique is applied to control the inverter switching pulses. This method has proven better than the other PWM techniques like SPWM [19, 20]. This SVPWM technique is used for precise speed control and optimal switching technique. Due to this the reduction in the switching harmonics and utilization of higher DC bus volts. The DTC and SVPWM combined to achieve the precise torque control [21-23]. In the first sector is voltage vector, which is reference voltage as shown in the Figure 1. The voltage is selected for two zero vectors and active voltage vectors, or , , respectively. Vectors are selected in anticlockwise direction. Reference vector can be anywhere in the above six sectors. Each time sector selection and timing of the reference selection is computed [24, 25]. Figure 1. Space vector representation for voltage vector V1 100 V2 110 V3 010 V4 011 V5 001 V6 101 V7 000 V8 111 Figure 1. Space vector representation for voltage vector The switching of inverter is made such that, zero state space vectors or for period, space vector for period and space vector for period. Let , represents duty cycles of ⁄ , and respectively. The relationship between timing and voltage is given by (13), (14), (15), (16) and (17). (13) (14) (13) here, M is Mdoulation index, and are calculated as below, Where, M is Mdoulation index, and are calculated as below, ( ) (15) (16) (17) ( ) ( ) (15) (16) (17) the duty cycles for the voltage vectors are given by (18), Indonesian J Elec Eng & Comp Sci, Vol. 19, No. 2, August 2020 : 644 - 651 Indonesian J Elec Eng & Comp Sci, Vol. 19, No. 2, August 2020 : 644 - 651 ISSN: 2502-4752  647  647 647 Indonesian J Elec Eng & Comp Sci ⁄ ⁄ (18) Defining the modulation index by (19), Defining the modulation index by (19), \| ⁄ \| (19) \| ⁄ \| (19) \| ⁄ \| (19) is switching period, and θ is angle bewteen stator volatge and common null vectors . Figure 3. Complete block diagram of Five-leg VSI - CM of operation for dual motor Figure 3. Complete block diagram of Five-leg VSI - CM of operation for dual motor 4. SIMULATION RESULTS AND DISCUSSIONS Modelling and Simulation plays a very important role in the field of drives control. The effectiveness of the five leg inverter techniques proved by running various simulation results are computed using Matlab/Simulink. Induction motor ratings are given as in Table 1. 3. COMMON MODE-FIVE LEG INVERTER (CM-FLI) USING SVPWM From Figure 3, complete connection of dual motors in FLI system is shown and common leg C is connected to both motors. The switching technique is designed in such a way that the current limit is maintained and is also called common mode voltage. Novel switching technique for five leg inverter in dual motor control (Jayaprakash Sabarad)  648 ISSN: 2502-4752  648 Figure 2. Novel switching technique of CM-FLI using SVPWM + + Five Leg VSI VA VB VC VD VE SVPWM Switching Technique Zero Sequence Voltage + + + + + + + + + + + + + + + + + + + + Figure 2. Novel switching technique of CM-FLI using SVPWM Five Leg – Voltage Source Inverter (FL-VSI) IM1 IM2 Torque and Flux Estimation of IM1 & IM2 PI PI +- +- PI +- synchro stationary New CM-FLI Switching Technique Wfeedback (IM1 & IM2) Wref (IM1 & IM2) φref (IM1 & IM2) φestimated (IM1 & IM2) Testimated (IM1 & IM2) VA VB VC VD VE Wref (IM1 & IM2) Table 1. Induction motor ratings Induction Motors IM1 Ratings: Induction Motors IM2 Ratings: 7.5KW (10HP), 50Hz, 400V, 1500rpm. 7.5KW (10HP), 50Hz, 400V, 1500rpm FLI - IM 1 at No-Load torque and IM2 at load torque for independent speed control Figure 8. FLI-IM 1 phase currents (amps) based on Figure 7 of independent speed Figure 9. FLI-IM2 phase currents (amps) based Figure 7 of independent speed Figure 8. FLI-IM 1 phase currents (amps) based on Figure 7 of independent speed Figure 9. FLI-IM2 phase currents (amps) based Figure 7 of independent speed Figure 8. FLI-IM 1 phase currents (amps) based on Figure 7 of independent speed From Figure 7, IM1 speed commanded to operate at 850rpm and IM2 speed commanded to operate 1100rpm, during this period the IM1 and IM2 motor phase currents are shown in Figure 8 and Figure 9 respectively. From this simulation results, independent speed control of dual motor operation is achieved at different operating speed. Table 1. Induction motor ratings Induction Motors IM1 Ratings: Induction Motors IM2 Ratings: 7.5KW (10HP), 50Hz, 400V, 1500rpm. 7.5KW (10HP), 50Hz, 400V, 1500rpm The results are obtained and proved from the novel switching SVPWM based fed to FLI system for dual motor speed control results found satisfactorily. Both the motors are controlled independently and verified at rated load torque at different motor speeds. Both motor running at different speed verified and shown in output waveforms. The output waveforms from Figure 4 to Figure 9 shows the independent operation of dual induction motors. Figure 4, shows the motor phase voltages and Figure 5, shows the motor line voltages. In Figure 6, both motors are commanded to operate at 1400rpm and 1440 rpm respectively. The independent control of dual motor speeds are shown in Figure 7. Where initially both motors are commanded to run at the same speed and after that both motors operating at different speed. The proposed novel CM-FLI switching technique shows the improved performance of the system with respect to existing switching techniques as compared in Table 2. FLI-VSI Induction Motor Output Waveforms at Rated Load Torque. Indonesian J Elec Eng & Comp Sci, Vol. 19, No. 2, August 2020 : 644 - 651 ISSN: 2502-4752 Indonesian J Elec Eng & Comp Sci 649  g p Figure 4. Five leg inverter output phase volts (volt) Figure 5. Five leg inverter output line volts (volt) Figure 6. FLI - IM 1 at No-Load torque and IM2 at load torque for independent speed control Figure 7. FLI-IM 1 and IM2 independent speeds (rpm) and torque (N-m) control operation Figure 8. FLI-IM 1 phase currents (amps) based on Figure 7 of independent speed Figure 9. FLI-IM2 phase currents (amps) based Figure 7 of independent speed From Figure 7, IM1 speed commanded to operate at 850rpm and IM2 speed commanded to operate 1100rpm, during this period the IM1 and IM2 motor phase currents are shown in Figure 8 and Figure 9 respectively. From this simulation results, independent speed control of dual motor operation is achieved at different operating speed. 4.1. THD comparison From Table 2, New Swicthing technique CM-FLI holds good from the motor thermal stand point as Figure 5. Five leg inverter output line volts (volt) Figure 5. Five leg inverter output line volts (volt) Figure 4. Five leg inverter output phase volts (volt) Figure 7. FLI-IM 1 and IM2 independent speeds (rpm) and torque (N-m) control operation Figure 6. 4.1. THD comparison p From Table 2, New Swicthing technique CM-FLI holds good from the motor thermal stand point as THD is lower than the conventional PWM technique. Total Harmonic Distortion (THD) is compared with standard single motor Vs FLI techniques. Thus the New switching technique (CM-FLI) holds good from operating stand point. This proposed technique will give the improved performance of ~23% for dual motor control using Five Leg Inverter (FLI). Novel switching technique for five leg inverter in dual motor control (Jayaprakash Sabarad) 650 ISSN: 2502-4752  Table 2. Induction motor THD comparison Three Leg Inverter, Single Induction Motor (SPWM) Five Leg Inverter, Coventional SPWM with Dual Motor Five Leg Inverter, New PWM (CM-FLI) with Dual Motor Output Phase Current THD % 11.59% 22.03% 14.31% REFERENCES [1] Rashid M.H, “Power Electronics, Circuits, Devices, and Applications”, PPH of India Pvt Ltd, New Delhi, 3rd Edition -2013. [2] Peter Vas “Sensorless Vector and Direct Torque Control”, 1998- Oxford University Press. er Vas “Sensorless Vector and Direct Torque Control”, 199 [3] Cheng T, Dan X, John E. F, Rahman M.F, “Carrier-Based PWM Methods With Common-Mode Voltage Reduction for Five-Phase Coupled Inductor Inverter” IEEE Transactions Industrial Electronics-Jan 2016, vol. 63, no. 1. pp. 526-537, 2015. pp [4] Jayaprakash Sabarad, Dr. G.H. Kulkarni, and Sushma Sattigeri, "Dual Three Phase Induction Motor Control Using Five Leg Inverter", IEEE- ICSPACE 2017. [5] Wei W, Jinghao Z, Ming C, “A Dual-Level Hysteresis Current Control for One Five-Leg VSI to Control Two PMSMs” IEEE Transactions Power Electronics, vol. 32, no. 1, Jan 2017. M. Jones, S.N. Vukosavic, D. Dujic, P. Wright, E. Levi, “Five-leg inverter PWM technique for reduced swit count two-motor constant power applications”, EPA, vol. 2, no. 5, September 2008. [6] M. Jones, S.N. Vukosavic, D. Dujic, P. Wright, E. Levi, “Five-leg inverter PW count two-motor constant power applications”, EPA, vol. 2, no. 5, September 2008 [7] Arkan A. Kadum “ PWM control techniques for three phase three level inverter drives",TELKOMNIKA (Telecommunication, Computing, Electronics and Control), vol. 18, no. 1, pp. 519-529, Feb 2020. ( , p g, ), , , pp , [8] Martin J, Drazen D, Emil Levi “A Performance Comparison of PWM Techniques for Five-Leg VSIs Supplying Two-Motor Drives” IEEE, 2008. [9] Ankit D, Nikhil, S.K.Sinha, and Priyanka S. “A Review on Different PWM Techniques for Five Leg Voltage Source Inverter”, IEEE- lCAESM 2012. 0] Gerardo Escobar Valderrama, Gerardo Vazquez Guzman, Erick I. Pool-Mazún, Panfilo Raymundo M, "A Singl Phase Asymmetrical T-Type Five-Level Transformerless PV Inverter" IEEE Journal of ESTPE vol. 6, no. 1, 2018 y yp [11] Wei Wang, Jinghao Zhang, Ming Cheng, "A Dual-Level Hysteresis Current Control for One Five-Leg VSI to Control Two PMSMs," IEEE Transaction Power Electronics, vol. 32, no. 1, 2017. SMs," IEEE Transaction Power Electronics, vol. 32, no. 1 [12] Nicolás T García1 Yeison Alberto G Gomez2 Fredy E Hoyos Velasco on "Parameter estimation of three-phase linear induction motor by a DSP-based electric-drives system", International Journal of Electrical and Computer Engineering (IJECE), vol. 10, no. 1, pp. 626-636, February 2020. 5. CONCLUSIONS From the above results, novel switching technique of CM-FLI using SVPWM plays an important role in switching the five leg inverter. This new CM-FLI switching technique uses only 10 switching elements with switching combination is up to 25. From the simulation results the proposed technique will give the improved performance of ~23% for dual motor control using Five Leg Inverter (FLI). The THD % of the New CM-FLI system reduced to 14.31% from conventional 22.03% for the Motor phase currents. The direct torque control for dual induction motor controlled independently with different load torque at different speed of operation. Switching methodology is verified for no rated load torque operation. The common DC link voltage is applied to both motors are same. It is verified that the common leg is able to share the both motor load current at different operating speeds. From this CM-FLI using SVPWM system is found to be most cost productive, reduced complex control algorithm and used in precise speed control of dual three phase induction motors independently. This work can be extended for different control methodology for the induction motor control. REFERENCES Kulkarni,“ Dual Motor Control using Double Zero Sequence Injection Method in Five Leg Voltage Source Inverter” International Journal of recent technology and Engineering (IJRTE), ISSN: 2277-3878, vol. 8 no. 5, January 2020. y [25] Jinglai Wu, et.al "Efficiency comparison of electric vehicles powertrains with dual motor and single motor input” Elsevier Mechanism and Machine Theory Journal 2018. REFERENCES g g pp y [13] Young-Seol Lim, June-Seok Lee, Kyo-Beum Lee, "Advanced Speed Control for a Five-Leg Inverter Driving a Dual-Induction Motor System", IEEE Transactions on IE, vol. 66, no. 1, 2019. y [14] Yashan Hu, Shoudao Huang, Xuan Wu, Xuefei Li, "Control of Dual Three-Phase Permanent Magnet Synchronous Machine Based on Five-Leg Inverter" IEEE Transactions on Power Electronics, vol. 34, no. 11, 2019. 5] O.C. Kivanc ; S.B. Ozturk, "Sector determination for SVPWM based four-switch three-phase VSI" IEE Electronics Letters, vol. 5, no. 5, 2017. [16] Yong-Chao Liu ; Xinglai Ge ; Qidi Tang ; Bin Gou "Two modified SVPWM algorithms for common-mode voltage reduction in eight-switch three-phase inverters" IEEE Electronics Letters, vol. 53, no. 10, 2017. g p [17] Jay K. Pandit, Mohan V. Aware, Ronak V. Nemade, Emil Levi by "Direct Torque Control Scheme for a Six-Phase Induction Motor With Reduced Torque Ripple" IEEE Transactions on Power Electronics, vol. 32, no. 9, 2017. [18] R. Palanisamy, Th. Ankit Singh, Anamika Ranjan, Jatin Singh "BLDC motor driven electric skateboard using SVPWM", International Journal of Electrical and Computer Engineering (IJECE), vol. 10, no. 1, pp. 711-718, Feb 2020. [19] Ibrahim Mohd Alsofyani, Yeongsu Bak, Kyo-Beum Lee "Fast Torque Control and Minimized Sector-Flux Droop for Constant Frequency Torque Controller Based DTC of Induction Machines" IEEE Transactions on Power Electronics, vol. 34, no. 12, 2019. donesian J Elec Eng & Comp Sci, Vol. 19, No. 2, August 2020 : 644 - 651  Indonesian J Elec Eng & Comp Sci 651 ISSN: 2502-4752 [20] Gopinathan Vivek, Jayanta Biswas, Meenu D. Nair, Mukti Barai "Simplified double switching SVPWM implementation for three-level VSI" IEEE The Journal of Engineering, vol. 2019, no. 11, 2019. p f g g, , , [21] P. Krishna Reddy,Deepak Ronanki, P. Parthiban "Direct torque and flux control of switched reluctance moto enhanced torque per ampere ratio and torque ripple reduction" IEEE Journl Electronics letter, vol. 55, no. 8, 2 [22] Yuan Ren, Z. Q. Zhu, James E. Green, Yun Li, Shiwu Zhu,Zijian Li "Improved Duty-Ratio-Based Direct Torque Control for Dual Three-Phase Permanent Magnet Synchronous Machine Drives" IEEE Transactions on Industry Applications, vol. 55, no. 6, 2019. pp [23] Jayaprakash Sabarad and Dr. G.H. Kulkarni, "Comparative Analysis of SVPWM and SPWM Techniques for Multilevel Inverter", IEEE Conference - ICPACE 2015. f [24] Jayaprakash Sabarad and Dr. G.H. Novel switching technique for five leg inverter in dual motor control (Jayaprakash Sabarad) BIOGRAPHIES OF AUTHORS Jayaprakash Sabarad, He obtained Bachelor of Engineering in Electrical and Electronics Engineering in year 2002 from V.T.U. Belagavi, India and M.Tech. in Power and Energy Systems in year 2006 from NITK Surathkal, India. Presently he is Research Scholar in the Department of Electrical Engineering at V.T.U. Belagavi, India. His research interests include Drives Control System and Power Electronics. Email: jpsabarad@gmail.com G.H. Kulkarni, He received the Bachelor of Engineering in Electrical Engineering in year 1985 from Karnataka University, Dharwad, India. Master of Technology in year 1992 in the domain of Power Systems from University of Mysore, Karnataka, India and Ph.D. in ANN Applications to Power System from J.N.T.U. Hyderabad, India in 2007. Presently he is working as a HOD of Electrical and Electronics Engineering at Jain College of Engineering Belagavi, India. His areas of interest include ANN application to Power System. Email: eeejceghk@gmail.com G.H. Kulkarni, He received the Bachelor of Engineering in Electrical Engineering in year 1985 from Karnataka University, Dharwad, India. Master of Technology in year 1992 in the domain of Power Systems from University of Mysore, Karnataka, India and Ph.D. in ANN Applications to Power System from J.N.T.U. Hyderabad, India in 2007. Presently he is working as a HOD of Electrical and Electronics Engineering at Jain College of Engineering Belagavi, India. His areas of interest include ANN application to Power System. Email: eeejceghk@gmail.com Novel switching technique for five leg inverter in dual motor control (Jayaprakash Sabarad)
https://openalex.org/W2147550150	https://europepmc.org/articles/pmc2714068?pdf=render	English	null	Sin Nombre Virus and Rodent Species Diversity: A Test of the Dilution and Amplification Hypotheses	PloS one	2,009	cc-by	7,551	Abstract This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: Research support was provided by NSF-NIH grant (EF 0326999) and University of Utah seed grant to M.D.D.; an NIH Training Grant to E.M.L. (AI055434- 01A1); a University of Utah Graduate Research Fellowship and American Society of Mammalogists Grant-in-Aid of Research to C.A.C. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of this manuscript. Competing Interests: The authors have declared that no competing interests exist. * E-mail: cclay@westminstercollege.edu Received November 18, 2008; Accepted June 23, 2009; Published July 31, 2009 pyright: 2009 Clay et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, whic e, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: Research support was provided by NSF-NIH grant (EF 0326999) and University of Utah seed grant to M.D.D.; an NIH Training Grant to E.M.L. (AI055434- 01A1); a University of Utah Graduate Research Fellowship and American Society of Mammalogists Grant-in-Aid of Research to C.A.C. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of this manuscript. Competing Interests: The authors have declared that no competing interests exist. * E-mail: cclay@westminstercollege.edu Christine A. Clay1*, Erin M. Lehmer2, Stephen St. Jeor3, M. Denise Dearing4 1 Department of Biology, Westminster College, Salt Lake City, Utah, United States of America, 2 Department of Biology, Fort Lewis College, Durango, Colorado, United States of America, 3 School of Medicine, University of Nevada Reno, Reno, Nevada, United States of America, 4 Department of Biology, University of Utah, Salt Lake City, Utah, United States of America PLoS ONE \| www.plosone.org Abstract Background: Species diversity is proposed to greatly impact the prevalence of pathogens. Two predominant hypotheses, the ‘‘Dilution Effect’’ and the ‘‘Amplification Effect’’, predict divergent outcomes with respect to the impact of species diversity. The Dilution Effect predicts that pathogen prevalence will be negatively correlated with increased species diversity, while the Amplification Effect predicts that pathogen prevalence will be positively correlated with diversity. For many host-pathogen systems, the relationship between diversity and pathogen prevalence has not be empirically examined. Methodology/Principal Findings: We tested the Dilution and Amplification Effect hypotheses by examining the prevalence of Sin Nombre virus (SNV) with respect to diversity of the nocturnal rodent community. SNV is directly transmitted primarily between deer mice (Peromyscus maniculatus). Using mark-recapture sampling in the Spring and Fall of 2003–2005, we measured SNV prevalence in deer mice at 16 landscape level sites (3.1 hectares each) that varied in rodent species diversity. We explored several mechanisms by which species diversity may affect SNV prevalence, including reduced host density, reduced host persistence, the presence of secondary reservoirs and community composition. We found a negative relationship between species diversity and SNV prevalence in deer mice, thereby supporting the Dilution Effect hypothesis. Deer mouse density and persistence were lower at sites with greater species diversity; however, only deer mouse persistence was positively correlated with SNV prevalence. Pinyon mice (P. truei) may serve as dilution agents, having a negative effect on prevalence, while kangaroo rats (Dipodomys ordii), may have a positive effect on the prevalence of SNV, perhaps through effects on deer mouse behavior. Conclusions/Significance: While previous studies on host-pathogen systems have found patterns of diversity consistent with either the Dilution or Amplification Effects, the mechanisms by which species diversity influences prevalence have not been investigated. Our study indicates that changes in host persistence, coupled with interspecific interactions, are important mechanisms through which diversity may influence patterns of pathogens. Our results reveal the complexity of rodent community interactions with respect to SNV dynamics. Citation: Clay CA, Lehmer EM, Jeor SS, Dearing MD (2009) Sin Nombre Virus and Rodent Species Diversity: A Test of the Dilution and Amplification Hypotheses. PLoS ONE 4(7): e6467. doi:10.1371/journal.pone.0006467 Editor: Matthew Baylis, University of Liverpool, United Kingdom Received November 18, 2008; Accepted June 23, 2009; Published July 31, 2009 Copyright: 2009 Clay et al. Species diversity, deer mouse density and SNV prevalence Deer mouse densities (estimated using DISTANCE) ranged from 1.5 to 42.5 mice/ha during the study (Table 1). We found a negative relationship between deer mouse density and species diversity (LMM, estimate = 228.50, Z = 25.28, P,0.01). We did not find a significant relationship between deer mouse density and SNV prevalence (LMM, estimate = 0.001, Z = 0.33, P = 0.74). Introduction predicts increased pathogen prevalence with greater species diversity, through increased encounters between hosts [7], or through the presence of secondary hosts [7,9]. Support for these models has been investigated primarily in vector-borne pathogens, such as Borrelia burgdorferi (the causative agent of Lyme disease) and Louping ill virus [6,8]. Other investigations in plant-pathogen systems and directly transmitted wildlife diseases suggest that these phenomena may be widespread in pathogen dynamics [1,10–13]. Despite the body of theoretical research outlining the patterns of the Dilution or Amplification Effect in host-pathogen systems, the fundamental mechanisms by which species diversity influences pathogen prevalence have not been thoroughly investigated. Species diversity greatly impacts many ecosystem functions such as productivity, resistance to invasion, and stability [1–5]. Recently, it has been suggested that species diversity also plays a key role in governing pathogen prevalence. The two leading hypotheses (‘‘Dilution Effect’’ and ‘‘Amplification’’ or ‘‘Rescue Effect’’) are similar in that they suggest that changes in species diversity alter host-pathogen dynamics and, in turn, affect pathogen prevalence [6–8]. However, these hypotheses predict divergent outcomes. The Dilution Effect predicts that species diversity decreases pathogen prevalence through mechanisms such as decreased host density, reduced encounters between hosts, or reduced host survival [7]. In contrast, the Amplification Effect We tested whether the prevalence of Sin Nombre virus (SNV), a hantavirus directly transmitted by its mammalian host, could be PLoS ONE \| www.plosone.org July 2009 \| Volume 4 \| Issue 7 \| e6467 1 Diversity Dilutes SNV Diversity Dilutes SNV explained by either the Dilution or Amplification Effect. The primary reservoir for SNV is the deer mouse (Peromyscus maniculatus), though other Peromyscus species [14,15] and the desert woodrat (Neotoma lepida)[16] potentially serve as secondary reservoirs. SNV transmission between rodents is primarily by direct contact, during aggressive behaviors such as biting and scratching, as evidenced by the strong correlation between SNV infection and scarring [17–20]. SNV infection in rodents is chronic, resulting in lifelong antibody production [21]. Rodents shed the virus in urine, saliva and feces [22,23]. Human infection with SNV typically occurs through inhalation of aerosolized virus from rodent excrement [24], and can progress to hantavirus cardio-pulmonary syndrome (HCPS), a disease with 38% mortality [25–27]. possibly due to the presence of more secondary reservoirs or due to increased host encounters. Potential Dilution or Amplification Agents Principle components analysis of the rodent communities indicated that pinyon mice and Ord’s kangaroo rats have a significant influence on the dynamics of SNV prevalence in Great Basin rodent communities. For all prevalence categories, Factor 1 explained 35.8% to 65.6% of variance, whereas 16.8% to 29.1% was explained by Factor 2 (Fig. 1). Overall these factors explained between 62.0% and 82.4% of total model variance. For all prevalence categories, the total number of pinyon mice and total number of kangaroo rats appeared to have a strong influence on overall model variance, as these species segregated independently from all other species at all levels of prevalence. There was a negative relationship between the number of pinyon mice per site and SNV prevalence (GLM, coeffi- cient = 20.54, F1, 39 = 5.55, P = 0.02). In this model there was also a significant relationship between season and SNV prevalence on each site with prevalence generally increasing from spring to fall (GLM, F5, 39 = 2.78, P = 0.03). The relationship between the number of kangaroo rats per site and SNV prevalence was marginally significant (GLM, coefficient = 0.48, F1, 39 = 3.86, P = 0.06). We explored mechanisms such as changes in host density, altered host persistence (via survival or dispersal), as well as the presence of secondary reservoirs, by which species diversity may affect SNV prevalence in deer mice. We predicted that if SNV followed the Dilution Effect model, SNV prevalence would be highest in low diversity communities, possibly due to increased density or increased persistence of deer mice, the primary SNV host. Conversely, if SNV followed the Amplification Effect model, SNV prevalence would be highest in high diversity communities, Introduction Additionally, we examined the relationship between SNV prevalence and the abundance of non-deer mice to identify potential dilution or amplification agents, i.e., rodent species other than deer mice that may affect the dynamics of SNV. Species diversity, deer mouse persistence and SNV prevalence The proportion of deer mice that survived across seasons ranged from 0% to 27.5% during this study. We found a negative relationship between species diversity in one season and the proportion of deer mice that survived to the following season (e.g., Spring 2004 to Fall 2004; LMM, estimate = 20.147, Z = 22.738, P,0.01). There was a positive relationship between deer mouse persistence and SNV prevalence in deer mice (LMM, esti- mate = 0.44, Z = 2.40, P = 0.02). , , p [ , ] We tested the Dilution Effect and Amplification Effect models by measuring SNV prevalence and rodent species diversity at several sites in Great Basin Desert, Utah. Previous studies suggest that species diversity reduces the prevalence of hantaviruses, including SNV [12,13,38,39], although these studies did not identify the mechanisms driving such patterns. The juniper-sagebrush habitat of the Great Basin desert is particularly well suited for studies of this nature, as SNV prevalence (0–50%) spans the range seen across all habitat types [15,17,21,40–42]. Within the Great Basin, rodent species diversity also varies considerably between sites within the same habitat type [43]. Low-diversity communities may consist of one or two species, including deer mice, while diverse communities have seven or more nocturnal rodent species of varying abundance. Secondary reservoirs such as pinyon mice (P. truei) and desert woodrats (N. lepida) co-occur with deer mice and may serve to ‘‘amplify’’ the prevalence of SNV [6]. However, the competence of these secondary reservoirs is currently unknown and is difficult to test given the Biosafety Level 4 facility required to work with animals infected with SNV. Conversely, heteromyid rodents such as Ord’s kangaroo rat (Dipodomys ordii) and Western pocket mouse (P. parvus) also co-occur with deer mice; these rodents are not SNV reservoirs and can achieve relatively high population densities. Heteromyids are frequently present in high diversity communities and may ‘‘dilute’’ SNV prevalence. PLoS ONE \| www.plosone.org Results At 16 sites, we captured 2,855 rodents of eight species over three years (Table 1). Gini-Simpson Index values ranged from 0.18 at a site comprised of deer mice and pocket mice, to 0.79 at a site consisting of seven rodent species (Table 1). During the study, SNV prevalence across sites ranged from 0.0% to 51.3% within a single season (Table 1). There was a negative relationship between species diversity and SNV prevalence in deer mice (LMM, estimate = 20.150, Z = 21.893, P = 0.05). The composition and relative abundance of species within a community can have profound effects on the behavior and density of deer mice, thereby altering the prevalence of SNV. For example, deer mice shift microhabitat use to avoid encounters with other species such as kangaroo rats (Dipodomys spp.), pinyon mice (P. truei) and pocket mice (Perognathus parvus) [28–33]. When encounters between deer mice and other species occur, they often result in fighting and other aggressive interactions [28–30,32]. Long-term experiments manipulating species composition demonstrate that the presence of other species can greatly depress deer mouse density through competition for limited resources [34–36]. Such changes in deer mouse density are likely to affect the frequency and type of encounters between deer mice and with other rodent species in the community, which could then alter SNV prevalence. Furthermore, competition for limited resources in high diversity communities could result in reduced survival or increased dispersal of deer mice; such changes in the persistence of the primary SNV host at a site could, in turn, result in a reduction of SNV prevalence [7,37]. Discussion The goal of our study was to investigate the effect of species diversity on pathogen dynamics in nature. Our findings indicate PLoS ONE \| www.plosone.org July 2009 \| Volume 4 \| Issue 7 \| e6467 2 Diversity Dilutes SNV Table 1. Summary of species diversity, prevalence of Sin Nombre virus and density of deer mice (Peromyscus maniculatus) at 16 sites in central Utah sampled 2003–2005. Table 1. Summary of species diversity, prevalence of Sin Nombre virus and density of deer mice (Peromyscus maniculatus) at 16 sites in central Utah sampled 2003–2005. Discussion p Site Spring 2003 Fall 2003 Species Diversity (D) SNV Prevalence Deer Mouse Density Species Diversity (D) SNV Prevalence Deer Mouse Density TJ-3 .30 13.6 22.0 .18 9.1 27.5 TJ-4 .29 7.0 42.5 .22 5.9 25.5 LS-5 .73 18.5 13.5 .77 6.7 15.4 LS-6 .68 0.0 5.0 .64 0.0 3.0 LS-7 .46 17.7 24.41 .58 5.9 18.0 LS-8 .44 16.7 12.0 .56 10.0 10.0 LS-9 - - - .67 9.1 11.0 LS-10 - - - .66 7.7 13.0 LS-11 - - - .68 11.1 4.5 UL-13 - - - .54 0.0 10.0 UL-14 - - - .77 0.0 3.0 Site Spring 2004 Fall 2004 Species Diversity (D) SNV Prevalence Deer Mouse Density Species Diversity (D) SNV Prevalence Deer Mouse Density TJ-3 .20 29.3 17.5 .43 37.1 18.0 TJ-4 .20 16.3 24.5 .27 23.9 33.5 LS-5 .76 11.5 13.0 .73 4.8 12.0 LS-6 .66 0.0 4.0 - - - LS-7 .70 25.9 13.5 .49 19.5 20.5 LS-8 .57 21.7 11.5 .56 29.2 12.0 LS-9 .56 5.0 .47 4.8 10.5 LS-10 .64 10.0 11.0 .77 15.0 10.0 LS-11 .63 0.0 1.5 - - - UL-13 .66 15.4 3.8 - - - UL-14 .69 30.4 4.5 - - - TJ-15 - - - .42 15.0 10.0 TJ-16 - - - .43 33.3 7.5 TJ-17 - - - .70 10.0 10.0 TJ-18 - - - .74 13.3 7.5 TJ-19 - - - .79 0.0 7.5 Site Spring 2005 Fall 2005 Species Diversity (D) SNV Prevalence Deer MouseDensity Species Diversity (D) SNV Prevalence Deer MouseDensity TJ-3 .63 39.1 11.5 .67 11.5 12.5 TJ-4 .37 51.3 19.0 .59 19.5 21.0 LS-5 .70 10.3 14.5 .74 20.0 18.0 LS-6 - - - - - - LS-7 .42 14.3 17.5 .65 2.6 19.0 LS-8 .49 11.5 13.0 .57 15.0 20.0 LS-9 .65 0.0 4.4 .68 24.1 14.5 LS-10 .64 5.3 10.0 .61 25.0 4.4 TJ-15 .27 30.4 11.5 .36 9.7 30.5 TJ-16 .27 18.8 8.0 .20 7.0 28.0 TJ-17 .68 25.0 6.0 .54 25.0 16.0 TJ-18 .65 29.4 8.5 .46 9.2 32.5 TJ-19 .75 5.9 8.5 .73 7.4 13.5 Site identifier includes geographic location (TJ=Tintic Junction; LS=Little Sahara; UL=Utah Lake) and the numeric code (3–19). Species diversity was estimated using Simpson’s Index (D) for each site per season. Site identifier includes geographic location (TJ=Tintic Junction; LS=Little Sahara; UL=Utah Lake) and the numeric code (3–19). Species diversity was estimated using Simpson’s Index (D) for each site per season. Prevalence of SNV was estimated as the number of deer mice positive for SN viral antibodies divided by the total number of deer mice sampled at each site per season x 100%. Density of deer mice (no./ha) was estimated using program DISTANCE. Dashes indicate sites that were not sampled in a particular season. doi:10.1371/journal.pone.0006467.t001 Discussion Prevalence of SNV was estimated as the number of deer mice positive for SN viral antibodies divided by the total number of deer mice sampled at each site per season x 100%. Density of deer mice (no./ha) was estimated using program DISTANCE. Dashes indicate sites that were not sampled in a particular season. doi:10.1371/journal.pone.0006467.t001 July 2009 \| Volume 4 \| Issue 7 \| e6467 3 3 Diversity Dilutes SNV Figure 1. Principle components analysis (PCA) of rodent communities grouped by mean SNV prevalence. Total number of individuals of each species was entered as components. Factor loadings plots were standardized to two factors. Figure 1.a. Factor loadings plot for low (0–5%) prevalence communities. Figure 1.b. Factor loadings plot for moderate (7–14%) prevalence communities. Figure 1.c. Factor loadings plot for high (.17%) prevalence communities. PEMA = deer mouse, PETR = pinyon mouse, DIOR = Ord’s kangaroo rat, NELE = desert woodrat, REME = western harvest mouse, PEPA = Great Basin pocket mouse, ONLE = Northern Grasshopper mouse. doi:10.1371/journal.pone.0006467.g001 Diversity Dilutes SNV Figure 1. Principle components analysis (PCA) of rodent communities grouped by mean SNV prevalence. Total number of individuals of each species was entered as components. Factor loadings plots were standardized to two factors. Figure 1.a. Factor loadings plot for low (0–5%) prevalence communities. Figure 1.b. Factor loadings plot for moderate (7–14%) prevalence communities. Figure 1.c. Factor loadings plot for high (.17%) prevalence communities. PEMA = deer mouse, PETR = pinyon mouse, DIOR = Ord’s kangaroo rat, NELE = desert woodrat, REME = western harvest mouse, PEPA = Great Basin pocket mouse, ONLE = Northern Grasshopper mouse. doi:10.1371/journal.pone.0006467.g001 directly, their study underscores the importance of rodent community composition with respect to disease dynamics in humans. In Paraguay, antibodies for Laguna Negra virus (a hantavirus) were most prevalent in rodent communities with the greatest proportion of vesper mice (Calomys laucha), the primary reservoir [13]. Lastly, a longitudinal study of multiple SNV reservoir species in habitats across the southwestern US also reported a negative relationship between SNV prevalence and species diversity [38]. None of these studies addressed the mechanisms by which diversity may influence hantavirus preva- lence. that species diversity is negatively correlated with SNV prevalence in deer mice, following the predictions of the Dilution Effect hypothesis. Discussion This outcome could not have been predicted a priori in this system because other species can potentially host SNV and thus species diversity could have increased SNV prevalence. Deer mouse density and deer mouse persistence were lower at sites with higher species diversity, suggesting that these two mechanisms were important means through which species diversity may have shaped SNV prevalence. Below we discuss the mechanisms in greater detail, as well as the potential significance of particular species in the rodent community that may influence the dynamics of SNV. PLoS ONE \| www.plosone.org Species diversity, deer mouse density and SNV prevalence Our findings corroborate other studies that implicate a negative relationship between species diversity and hantavirus prevalence. In a study of Choclo virus (a hantavirus found in Panama) [12], sites with reduced rodent species diversity (relative to a control) were found to have greater numbers of human cases of this strain of hantavirus, compared to other sites. Although they did not determine the prevalence of Choclo virus in the rodent community Results of our study indicate that species diversity is negatively correlated with deer mouse density, but we did not find a significant relationship between deer mouse density and SNV prevalence. Other studies have also found that changes in rodent community composition had marked effects on deer mouse density July 2009 \| Volume 4 \| Issue 7 \| e6467 4 Diversity Dilutes SNV We found that the number of pinyon mice in a community was negatively correlated with prevalence, indicating that they may serve as dilution agents. A priori, we predicted pinyon mice were likely amplification agents since they co-occur with deer mice in relatively high densities in the Great Basin and they are potential reservoirs for SNV (S. St. Jeor, unpublished). Measuring the competence of pinyon mice for SNV was beyond the scope of our study as it requires a Biosafety Level 4 facility. Perhaps pinyon mice serve as dilution agents for SNV by altering encounters between deer mice through changes in deer mouse movement, behavior or maturation. For example, Ovadia et al. [45] found a significant behavioral change within a species when that species was in the presence of another species. Male and female Allenby’s gerbils (Gerbillus andersoni allenbyi) engaged in fewer aggressive interactions when in the presence of a larger-bodied competitor, G. pyramidum. Similarly, the presence of pinyon mice (a larger-bodied competitor) may decrease encounters between deer mice, and therefore reduce opportunities for SNV transmission. Alternative- ly, if pinyon mice are less competent reservoirs for SNV, then their encounters with deer mice may not result in successful transmis- sion events, which would decrease prevalence, particularly if the total number of encounters regardless of species is fixed for deer mice. Lastly, if pinyon mice either directly or indirectly delayed the onset of reproduction in deer mice such that the population consisted of a greater proportion of younger individuals, SNV prevalence in deer mice could be reduced because juveniles are less likely to engage in behaviors that promote transmission [45,46]. Species diversity, persistence of deer mice and SNV prevalence However, path analysis is beyond the scope of the present study, as path analyses are more appropriate for larger data sets such as long-term longitudinal studies (e.g., .5 years, 25+ sites). Studies with larger and longer data sets may consider path analysis as a means to further understand the relationship between community ecology and the dynamics of pathogens in the wild. The results of our study demonstrate the importance of community ecology in the dynamics of SNV prevalence in deer mice. In particular, our findings indicate that species diversity is negatively correlated with SNV prevalence and this relationship may be mediated by changes in deer mouse persistence and indirectly by deer mouse density. Our findings highlight the importance of understanding the behavior and ecology of disease reservoirs in the wild, to both predict future outbreaks and minimize the risk for human infection. We acknowledge the limitations of this study in that it was an observational study on free-ranging animals. However, it lays the necessary groundwork for future more manipulative studies that require substantial investment in enclosure infrastructure to execute experiments with replication at a landscape scale to address issues at the level of rodent communities. Furthermore, for pathogens such as SNV, that require BioSafety Level 4 laboratory facilities for experiments with infected animals, field studies can quickly advance our understanding of host-pathogen systems because the degree of safety precautions required in the laboratory are not needed for field studies. LoGiudice and colleagues [9] suggested that pathogen preva- lence could be influenced by a species other than the primary host, and coined this phenomenon a ‘dilution host’ A dilution host may be a species with low reservoir competency, transmitting the pathogen ineffectively and thereby reducing prevalence [9]. However, a dilution host, regardless of reservoir competency, may also reduce prevalence by altering the behavior of the primary host, such that contacts between hosts are infrequent and prevalence is therefore reduced [7]. Because these species can still affect prevalence without hosting the pathogen, we term them ‘dilution agents’, rather than hosts. Similarly, what we have termed an ‘amplification agent’ may be a species with high reservoir competency, transmitting the pathogen frequently and increasing overall prevalence [9]. Species diversity, persistence of deer mice and SNV prevalence Our study suggests that species diversity may influence SNV prevalence by reducing the persistence of deer mice at high diversity sites, as species diversity was negatively correlated with recapture probabilities of deer mice in subsequent seasons. Because this was a field study on free-ranging animals, we cannot definitively discriminate between two alternatives that could be driving this pattern, i.e. differential survival versus dispersal. However, further analysis of the data indicated that there was not a relationship between deer mouse body mass (a proxy for age) and species diversity, thus, differential survival seems less likely than differential dispersal (LMM, estimate = 0.88, Z = 1.09, P = 0.27). Fewer mice persisting at high diversity sites across seasons, indicates that these sites are less likely to have infected individuals remaining to maintain SNV in the population. We recognize that these conclusions are based on correlative studies and that further studies with the capacity for experimental manipulation are required to thoroughly test this hypothesis. However, these results are a necessary first step towards understanding the role of species diversity on SNV and identifying potential mechanisms. Using a two-species mathematical model, Peixoto and Abram- son [39] proposed that species diversity ‘‘dilutes’’ hantavirus prevalence through competitive pressure exerted by the non-host species on the host. The model differs from our findings, as their model implicates reduced host density as the mechanism by which species diversity modulates SNV prevalence whereas we found no direct connection with density. However, the possibility exists that competitive pressure from non-host species alters the retention (survival or dispersal) of deer mice on a site thereby impacting SNV prevalence. Our study also suggests that kangaroo rats, which do not host SNV, had a positive, though not statistically significant (P = 0.06), relationship with prevalence of SNV. Kangaroo rats may serve as amplification agents because they cause deer mice to have more interactions with one another. Falkenberg and Clarke [29] found that deer mice significantly shifted habitat use from open and closed microhabitats to almost exclusively closed microhabitats in the presence of kangaroo rats. The concentration of deer mouse activity within a microhabitat could increase SNV prevalence through the increased frequency of encounters. Path analysis would provide further insight into the dynamics of the relationship between species diversity, deer mouse density and persistence, and SNV prevalence. Species diversity, deer mouse density and SNV prevalence These hypotheses could be addressed in future field studies. [34,36,44]. Such changes in density may result from interspecific competition for nesting sites or food resources, from increased deer mouse dispersal away from sites, or increased predation risk as deer mouse movement and other behaviors change in the presence of other species [28–30]. Species diversity, persistence of deer mice and SNV prevalence However, an amplification agent may also be a species with low reservoir competency, yet the presence of this species may change the behavior of the primary host such that contacts between hosts are more frequent and prevalence subsequently increases [7]. Study sites and sampling Deer mice were sampled from 16 sites near the West Tintic Mountains in the Great Basin Desert of central Utah (Juab and Utah Counties) on public lands. To maintain independence, all sites were located .700 m apart. Study sites were dominated by big sagebrush (Artemisia tridentata) and Utah juniper (Juniperus osteosperma). We selected sites that varied in habitat heterogeneity based on shrub cover and bare ground, as both have been linked to diversity in rodent community composition [47–52]. Sites ranged from 1.2% shrub cover and 62.2% bare ground to 48.1% shrub cover and 6.1% bare ground. More site details can be found in Lehmer et al. [53]. Rodent sampling occurred in ‘‘Spring’’ (May and June) and ‘‘Fall’’ (late August and September, October) of 2003, 2004 and 2005, during 15 consecutive day periods that coincided with the new moon. Each site was trapped for three consecutive nights per season (Spring versus Fall). Although we monitored a total of 16 sites during our study, only 12 sites could be sampled during any single season due to time limitations. Of the 16 sites, five sites were sampled in every season and the remaining 12 sites were sampled in some, but not all seasons. Statistical analysis We first examined the relationship between SNV prevalence and species diversity using linear mixed models (LMM). LMM do not require balanced sampling of longitudinal repeated data (SYSTAT, Version 10, 2000; [59]), and thus was appropriate considering our unbalanced study design (e.g., sites repeatedly sampled, but not equally). In this analysis, species diversity was treated as a fixed effect, whereas site was considered a random effect to account for repeated sampling. Season could not be included as a factor in the analyses due to sample size (N = 16) limitations. SNV antibody detection and estimates of prevalence SNV antibody detection and estimates of prevalence Enzyme linked immunosorbent assays (ELISA) were used to screen deer mouse blood for IgG antibodies to SNV. Because SNV is a chronic infection in deer mice, antibodies are a reliable indicator of SNV infection [57]. Briefly, wells of polyvinyl chloride microtiter plates (Dynatech Laboratories) were coated overnight at 4uC with recombinant nucleocapsid antigen diluted 1:2000 in phosphate buffered saline. A non-hantavirus recombinant antigen was used as a negative control. After incubation, unbound antigen was removed from wells by washing 36 with wash buffer. Deer mouse sera were heat inactivated in a 55uC water bath for 30 minutes. Heat inactivated sera were diluted 1:100 in serum- dilution buffer, containing powdered non-fat milk, Tween 20 and 106PBS in a 1:1:20 ratio. The diluted serum solution was added to the antigen-coated wells and plates were then incubated at 37uC for 60 minutes. Plates were then washed 3 times with wash buffer (1:20 Tween and 106PBS) and incubated at 37uC for 30 minutes with 100 ul of ABTS Microwell Peroxidase Substrate Solution (Kirkegaard and Perry Laboratories, Inc.) [58]. Absorbance (405 nm) was recorded with a Versa Max Tunable Microplate Reader (VWR International) and values .3 standard deviations of the negative control wells on each plate were considered positive for anti - SNV antibodies [58]. Prevalence was estimated by dividing the number of unique seropositive deer mice by the total number of deer mice captured within a sampling period and this quotient was multiplied by 100. To eliminate the problem of overestimating prevalence through the inclusion of uninfected juveniles carrying maternal antibodies, prevalence was only calculated on the adults (.14 g) in the population [53]. The Holy Grail: Contact Rates For directly transmitted pathogens such as SNV, contact rates between infected and susceptible individuals represent the underlying mechanism through which diversity is predicted to PLoS ONE \| www.plosone.org PLoS ONE \| www.plosone July 2009 \| Volume 4 \| Issue 7 \| e6467 5 Diversity Dilutes SNV Diversity Dilutes SNV act. Both the Dilution and Amplification hypotheses predict that species diversity will change the rate of contact between individuals. For many animals such as small, nocturnal animals contact rates are not easily obtained. While radiotelemetry may seem an obvious technology, the weight of the transmitters and the presence of observers can alter the behavior of rodents. Additionally telemetry location estimates are usually too large (within meters) to define probable contacts between small animals like rodents and the costs prohibit marking an entire rodent community. However, new technologies such as radio-frequency identification (RFID) devices such as passive integrative transpon- ders coupled with inexpensive surveillance systems will make it possible in the future to document actual contact rates of such difficult to observe animals. These data will be extremely useful for the further assessment of the Dilution and Amplification hypotheses. estimated using the program DISTANCE (version 4.1) a software program designed to better estimate animal densities on web trapping designs than using the raw values [56]. We estimated persistence of deer mice as the percentage of individuals known to be alive from one trapping season to subsequent seasons on each site [40]; we were unable to estimate survival using programs such as MARK because it is often unreliable when recapture frequency is very low, as in the case of our study (often ,10%). Rodent Sampling At each site, animals were live-trapped using 148 traps (H.B. Sherman Traps, Inc.) distributed in a ‘‘web’’ configuration over 3.1 hectares, following the methods of Mills et al. [54]. Upon capture, animals were identified to species; sex and weight were recorded. All animals were marked with uniquely numbered ear tags. Approximately 0.2 ml of blood was collected via the retro- orbital sinus of P. maniculatus at the time of initial capture for each trapping season. Blood was stored immediately on dry ice until transfer to an 280uC freezer. After blood collection, all animals were released at the point of capture. Some animals were captured multiple times during a single season (e.g., two of the three nights in spring), however only the data from the initial capture per season was used in the analysis. All workers implemented the recommended precautions for handling animals potentially infected with hantavirus [27] and all techniques for capturing and handling animals were approved by the Institutional Animal Care and Use Committee of the University of Utah (IACUC #0203011, #0503011). We next investigated the two potential mechanisms by which species diversity may influence SNV prevalence, also using LMM. In independent analyses, we evaluated the impact of species diversity on deer mouse density, as well as the effect of species diversity on the persistence of deer mice. In both analyses, species diversity was considered a fixed effect and site a random effect. Finally, we examined whether these two mechanisms, deer mouse density and persistence were directly correlated with SNV prevalence. We evaluated the response of SNV prevalence to deer mouse density (fixed effect) with site as a random effect, using LMM. In a separate analysis, we evaluated the response of SNV prevalence to host persistence (fixed effect) with site as a random effect. We could not simultaneously evaluate the response of SNV to these two variables due to sample size limitations. Species diversity, density and persistence estimates Species diversity was estimated using Gini-Simpson Index (D = 1-g p2 i; [53]). Density of deer mice at each site were PLoS ONE \| www.plosone.org July 2009 \| Volume 4 \| Issue 7 \| e6467 6 Diversity Dilutes SNV elimination was used to reduce general models to the most parsimonious version. We used principle components analysis (PCA) to isolate key rodent species in our study system that had potential to influence SNV prevalence. References 1. Knops JMH, Tilman D, Haddad NM, Naeem S, Mitchell CE, et al. (1999) Effects of plant species richness on invasion dynamics, disease outbreaks, insect abundances and diversity. Ecology Letters 2: 286–293. 20. Mills JN, Ksiazek TG, Peters CJ, Childs JE (1999) Long-term studies of hantavirus reservoir populations in the southwestern United States: a synthesis. Emerging Infectious Diseases 5: 135–142. y gy 2. Naeem S, Li CS (1997) Biodiversity enhances ecosystem reliability. Nature 390: 507–509. 21. Douglass RJ, Wilson T, Semmens WJ, Zanto SN, Bond CW, et al. (2001) Longitudinal studies of Sin Nombre virus in deer-mouse dominated ecosystems of Montana. American Journal of Tropical Medicine and Hygiene 65: 33–41. 3. Raffaelli D, van der Putten WH, Persson L, Wardle DA, Petchey OL, et al. (2002) Multi-trophic dynamics and ecosystem processes, in Biodiversity and Ecosystem Functioning. Loreau M, Naeem S, Inchausti P, eds. Oxford: Oxford University Press. 22. Botten J, Mirowsky K, Kusewitt D, Bharadwaj M, Yee J, et al. (2000) Experimental infection model for Sin Nombre hantavirus in the deer mouse (Peromyscus maniculatus). Proceedings of the National Academy of Sciences of the United States of America 97: 10578–10583. 4. Smith MD, Knapp AK (1999) Exotic plant species in a C4-dominated grassland: invasibility, disturbance, and community structure. Oecologia 120: 605–612. 23. Yamada T, Hjelle B, Lanzi R, Morris B, Anderson B, et al. (1995) Antibody responses to Four Corners hantavirus infection in the deer mouse (Peromyscus maniculatus): Identification of an immunodominant region of the viral nucleocapsid protein. Journal of Virology 69: 1939–1943. 5. Tilman D, Knops JMH, Wedin D, Reich PB, Ritchie ME, et al. (1997) The influence of functional diversity and composition on ecosystem processes. Science 277: 1300–1302. 6. Gilbert L, Norman R, Laurenson KM, Reid HW, Hudson PJ (2001) Disease persistence and apparent competition in a three-host community: an empirical and analytical study of large-scale, wild populations. Journal of Animal Ecology 70: 1053–1061. 24. Doyle TJ, Bryan RT, Peters CJ (1998) Viral hemorrhagic fever and hantavirus infections in the Americas. Infectious Disease Clinics of North America 12: 95. 25. Hjelle B, Jenison S, Torrez-Martinez N, Yamada T, Nolte K, et al. (1994) A novel hantavirus associated with an outbreak of fatal respiratory disease in the southwestern United States: evolutionary relationships to known hantaviruses. Journal of Virology 68: 592–6. 7. Keesing F, Holt RD, Ostfeld RS (2006) Effects of species diversity on disease risk. Ecology Letters 9: 485–498. 8. References Brown JH, Munger JC (1985) Experimental manipulation of a desert rodent community: food addition and species removal. Ecology 66: 1545–1563. 14. Abbott KD, Ksiazek TG, Mills J (1999) Long-term hantavirus persistence in rodent populations in Central Arizona. Emerging Infectious Diseases 5: 102–112. community: food addition and species removal. Ecology 66: 1545 35. Ernest S, Brown JH (2001) Delayed compensation for missing keystone species by colonization. Science 292: 101–103. 15. Otteson EW, Riolo J, Rowe JE, Nichol ST, Ksiazek TG, et al. (1996) Occurrence of hantavirus within the rodent population of northeastern California and Nevada. American Journal of Tropical Medicine and Hygiene 54: 127–33. 36. Heske EJ, Brown JH, Mistry S (1994) Long-term experimental study of a Chihuahuan desert rodent community: 13 years of competition. Ecology 75: 438–445. 37. Calisher CH, Mills JN, Sweeney WP, Choate JR, Sharp DE, et al. (2001) Do unusual site-specific population dynamics of rodent reservoirs provide clues to the natural history of hantaviruses? Journal of Wildlife Diseases 37: 280–288. 16. Dearing MD, Mangione AM, Karasov WH, Morzunov S, Otteson E, et al. (1998) Prevalence of hantavirus in four species of Neotoma from Arizona and Utah. Journal of Mammalogy 79: 1254–1259. 38. Mills JN (2005) Regulation of rodent-borne viruses in the natural host: implications for human disease. Archives of Virology 19: 45–57. 17. Boone JD, Otteson EW, McGwire KC, Villard P, Rowe JE, et al. (1998) Ecology and demographics of hantavirus infection in rodent populations in the Walker River Basin of Nevada and California. American Journal of Tropical Medicine and Hygiene 59: 445–451. 39. Peixoto ID, Abramson G (2006) The effect of biodiversity on the Hantavirus epizootic. Ecology 87: 873–879. 40. Kuenzi AJ, Morrison ML, Swann DE, Hardy PC, Downard GT (1999) A longitudinal study of Sin Nombre virus prevalence in rodents, southeastern Arizona. Emerging Infectious Diseases 5: 113–118. 18. Calisher CH, Sweeney W, Mills JN, Beaty BJ (1999) Natural history of Sin Nombre virus in Western Colorado. Emerging Infectious Diseases 5: 126–134. 19. Mills JN, Ksiazek TG, Ellis BA, Rollin PE, Nichol ST, et al. (1997) Patterns of association with host and habitat: antibody reactive with Sin Nombre virus in small mammals in the major biotic communities of the southwestern United States. American Journal of Tropical Medicine and Hygiene 56: 273–84. 41. Mackelprang R, Dearing MD, St. Jeor SJ (2001) High prevalence of Sin Nombre virus in rodent populations, central Utah: A consequence of human disturbance? References Ostfeld R, Keesing F (2000) The function of biodiversity in the ecology of vector- borne zoonotic diseases. Canadian Journal of Zoology-Revue Canadienne De Zoologie 78: 2061–2078. 26. Kilpatrick ED, Terajima M, Koster FT, Catalina MD, Cruz J, et al. (2004) Role of specific CD8+ T cells in the severity of a fulminant zoonotic viral hemorrhagic fever, Hantavirus Pulmonary Syndrome. Journal of Immunology 172: 3297–3304. g 9. LoGiudice K, Ostfeld RS, Schmidt KA, Keesing F (2003) The ecology of infectious disease: Effects of host diversity and community composition on Lyme disease risk. Proceedings of the National Academy of Sciences of the United States of America 100: 567–571. 27. Mills J, Childs J, Ksiazek T, Peters C, Velleca W (1995) Methods for trapping and sampling small mammals for virologic testing. Centers for Disease Control and Prevention. 10. Mitchell CA, Reich PB, Tilman D, Groth JV (2003) Effects of elevated CO2 nitrogen deposition, and decreased species diversity on foliar fungal plant disease. Global Change Biology 9. 28. Ambrose RF, Meehan TE (1977) Aggressive Behavior of Perognathus parvus and Peromyscus maniculatus. Journal of Mammalogy 58: 665–668. 29. Falkenberg JC, Clarke JA (1998) Microhabitat use of deer mice: effects of interspecific interaction risks. Journal of Mammalogy 79: 558–565. 11. Mitchell CA, Tilman D, Groth JV (2002) Effects of grassland plant species diversity, abundance, and composition on foliar fungal disease. Ecology 83: 1713–1726. 30. Kritzman EB (1974) Ecological relationships of Peromyscus maniculatus and Perognathus parvus in eastern Washington. Journal of Mammalogy 55: 172–188. 12. Ruedas LA, Salazar-Bravo J, Tinnin DS, Armien B, Caceres L, et al. (2004) Community ecology of small mammal populations in Panama following an outbreak of Hantavirus pulmonary syndrome. Journal of Vector Ecology 29: 177–191. 31. Larsen E (1986) Competitive release in microhabitat use among existing desert rodents: a natural experiment. Oecologia 69: 231–237. 32. Llewellyn JB (1980) Seasonal changes in the aggressive behavior of Peromyscus maniculatus inhabiting a pinyon-juniper woodland in western Nevada. Journal of Mammalogy 61: 341–345. 13. Yahnke CJ, Meserve PL, Ksiazek TG, Mills J (2001) Patterns of infection with Laguna Negra virus in wild populations of Calomys laucha in the central Paraguayan Chaco. American Journal of Tropical Medicine and Hygiene 65: 768–776. gy 33. Llewellyn JB, Jenkins SH (1987) Patterns of niche shift in mice: seasonal changes in microhabitat breadth and overlap. The American Naturalist 129: 365–381. 34. Acknowledgments We thank Dr. Matthew Baylis and two reviewers for insightful comments on this manuscript. Thanks to S. O’Grady, S. Haley and M. Skopec for comments on previous drafts of this manuscript. Heartfelt thanks to the numerous field crewmembers for invaluable assistance with many aspects of this research. Based on the patterns identified by PCA, we used general linear models (GLM) to investigate the relationship between SNV prevalence and key rodent species. The total number of pinyon mice, kangaroo rats, and deer mice captured at each site were treated as independent continuous variables and site and season were independent categorical variables. Backward stepwise Rodent Sampling In this analysis, we used the untransformed number of individuals of each species as components, since density estimates calculated by the program DISTANCE are not accurate for small samples sizes. Sites were grouped by categories according to mean prevalence across all sampling seasons (low (0–5%), moderate (7–14%), high (.17%)). We specified two factors for each analysis to standardize the factor loadings plots. elimination was used to reduce general models to the most parsimonious version. Author Contributions Conceived and designed the experiments: CAC EML MDD. Performed the experiments: CAC EML. Analyzed the data: CAC EML. Contributed reagents/materials/analysis tools: SSJ. Wrote the paper: CAC EML MDD. Conceived and designed the experiments: CAC EML MDD. Performed the experiments: CAC EML. Analyzed the data: CAC EML. Contributed reagents/materials/analysis tools: SSJ. Wrote the paper: CAC EML MDD. References Emerging Infectious Diseases 7: 480–481. PLoS ONE \| www.plosone.org 7 July 2009 \| Volume 4 \| Issue 7 \| e6467 Diversity Dilutes SNV Diversity Dilutes SNV 52. Valone TJ, Sauter P (2005) Effects of long-term cattle exclosure on vegetation and rodents at a desertified arid grassland site. Journal of Arid Environments 61: 161–170. 42. Mills JN, Johnson JM, Ksiazek TG, Ellis BA, Rollin PE, et al. (1998) A survey of hantavirus antibody in small-mammal populations in selected United States national parks. American Journal of Tropical Medicine and Hygiene 58: 525–32. 53. Lehmer EM, Clay CA, Pearce-Duvet J, Jeor SS, Dearing MD (2008) Differential regulation of pathogens: the role of habitat disturbance in predicting prevalence of Sin Nombre virus. Oecologia 155: 429–439. 43. Brown JH (1973) Species diversity of seed-eating desert rodents in sand dune habitats. Ecology 54: 775–787. 44. Munger JC, Brown JH (1981) Competition in desert rodents: an experiment with semiperimeable exclosures. Science 211: 510–512. 54. Mills J, Yates T, Ksiazek TG, Peters CJ, Childs JE (1999) Long-term studies of hantavirus reservoir populations in the southwestern United States: Rationale, potential and methods. Emerging Infectious Disease 5: 95–101. semiperimeable exclosures. Science 211: 510–512. 45. Klein SL, Zink MC, Glass GE (2004) Seoul virus increases aggressive behaviour in male Norway rats. Animal Behaviour 67: 421–429. in male Norway rats. Animal Behaviour 67: 421–4 55. Gini C (1912) Variabilita e mutabilita. Studi Economico-Giuridici Fac, ed. Cagliari A Vol. III, parte II. 46. Klein S, Bird B, Nelson R, Glass G (2002) Environmental and physiological factors associated with Seoul virus infection among urban populations of Norway rats. Journal of Mammalology 83: 478–488. 56. Thomas L, Laake JL, Strindherg S, Marques FFC, Buckland ST, et al. (2004) Distance 4.1 Release 2. Research Unit for Wildlife Population Assessment, University of St. Andrews, UK. 47. Ovadia O, Abramsky Z, Kotler BP, Pinshow B (2005) Inter-specific competitors reduce inter-gender competition in Negev Desert gerbils. Oecologia (Berlin) 142: 480–488. 57. Botten J, Mirowsky K, Kusewitt D, Ye C, Gottlieb K, et al. (2003) Persistent Sin Nombre infection in the deer mouse (Peromyscus maniculatus) model: sites of replication and strand-specific expression. Journal of Virology 77: 1540–1550. 48. Alain B, Gilles P, Yannick D (2006) Factors driving small rodents assemblages from field boundaries in agricultural landscapes of western France. Landscape Ecology 21: 449–461. 58. Borucki MK, Boone JD, Rowe JE, Bohlman MC, Kuhn EA, et al. PLoS ONE \| www.plosone.org References (2000) Role of maternal antibody in natural infection of Peromyscus maniculatus with Sin Nombre virus. Journal of Virology 74: 2426–2429. gy 49. Cui Q, Jiang Z, Lian X, Zgang T, Su J (2005) Factors influencing habitat selection of root voles (Microtus oeconomus). Acta Theriologica Sinica 25: 45–51. 59. Cnaan A, Laird NM, Slasor P (1997) Using the general linear mixed model to analyse unbalanced repeated measures and longitudinal data. Stat Med 16: 2349–80. ( ) g 50. Flowerdew JR, Ellwood SA (2001) Impacts of woodland deer on small mammal ecology. Forestry (Oxford) 74: 277–287. gy y ( ) 51. Rogovin KA (1996) Vegetation as a factor of species diversity in desert rodent communities. Izvestiya Akademii Nauk Seriya Biologicheskaya (Moscow) 0: 743–748. PLoS ONE \| www.plosone.org July 2009 \| Volume 4 \| Issue 7 \| e6467 8
https://openalex.org/W3017033974	https://europepmc.org/articles/pmc7221622?pdf=render	English	null	A Novel Biocompatible Titanium–Gadolinium Quantum Dot as a Bacterial Detecting Agent with High Antibacterial Activity	Nanomaterials	2,020	cc-by	12,914	Received: 18 March 2020; Accepted: 16 April 2020; Published: 17 April 2020 Abstract: In this study, the titanium–gadolinium quantum dots (TGQDs) were novel, ﬁrst of its type to be synthesized, and fully characterized to date. Multiple physical characterization includes scanning electron microscopy (SEM), scanning electrochemical microscope (SCEM), x-ray ﬂuorescence, spectrophotometry, and dynamic light scattering were carried out. The obtained results conﬁrmed appropriate size and shape distributions in addition to processing optical features with high quantum yield. The synthesized TGQD was used as a ﬂuorescent dye for bacterial detection and imaging by ﬂuorescent microscopy and spectrophotometry, where TGQD stained only bacterial cells, but not human cells. The signiﬁcant antibacterial activities of the TGQDs were found against a highly pathogenic bacterium (Staphylococcus aureus) and its antibiotic resistant strains (vancomycin and methicillin resistant Staphylococcus aureus) using growth curve analysis and determination of minimum inhibitory concentration (MIC) analysis. Live/dead cell imaging assay using phase-contrast microscope was performed for further conﬁrmation of the antibacterial activity. Cell wall disruption and release of cell content was observed to be the prime mode of action with the reduction of cellular oxygen demand (OD). Keywords: bacterial resistance; titanium–gadolinium quantum dots; bacterial detection; antibacterial activity; SECM nanomaterials nanomaterials nanomaterials Article A Novel Biocompatible Titanium–Gadolinium Quantum Dot as a Bacterial Detecting Agent with High Antibacterial Activity Vishma Pratap Sur 1,2,,† , Aninda Mazumdar 1,2,†, Amirmansoor Ashraﬁ1,2 , Atripan Mukherjee 1,2, Vedran Milosavljevic 1,2 , Hana Michalkova 1, Pavel Kopel 3 , Lukáš Richtera 1,2 and Amitava Moulick 1,2, 1 Department of Chemistry and Biochemistry, Faculty of AgriSciences, Mendel University in Brno, Zemedelska 1/1665, CZ-61300 Brno, Czech Republic; xmazumda@mendelu.cz (A.M.); amirmansoor.ashraﬁ@mendelu.cz (A.A.); atripan@gmail.com (A.M.); grizlidripac@gmail.com (V.M.); hanabuchtelova@gmail.com (H.M.); oliver@centrum.cz (L.R.) 1 Department of Chemistry and Biochemistry, Faculty of AgriSciences, Mendel University in Brno, Zemedelska 1/1665, CZ-61300 Brno, Czech Republic; xmazumda@mendelu.cz (A.M.); amirmansoor.ashraﬁ@mendelu.cz (A.A.); atripan@gmail.com (A.M.); grizlidripac@gmail.com (V.M.); hanabuchtelova@gmail.com (H.M.); oliver@centrum.cz (L.R.) g 2 Central European Institute of Technology (CEITEC), Brno University of Technology, Purkynova 123 CZ-62100 Brno-Královo Pole, Czech Republic 3 Department of Inorganic Chemistry, Faculty of Science, Palacky University, 17. listopadu 12, CZ-771 46 Olomouc, Czech Republic; paulko@centrum.cz * Correspondence: vishmapratapsur@gmail.com or xsur@mendelu.cz (V.P.S.); amitava.moulick@mendelu.cz or amitavamoulick@gmail.com (A.M.); Tel.: +420-545-133-350 (V.P.S.); +420-545-133-350 (A.M.) * Correspondence: vishmapratapsur@gmail.com or xsur@m (V.P.S.); amitava.moulick@mendelu.cz or amitavamoulick Tel.: +420-545-133-350 (V.P.S.); +420-545-133-350 (A.M.) † These authors contributed equally to this work. www.mdpi.com/journal/nanomaterials Nanomaterials 2020, 10, 778; doi:10.3390/nano10040778 1. Introduction Nanoparticles exert various kinds of toxicity, they can be cancerous, aﬀect our immune system, liver, spleen, can generate cellular ROS, damage our DNA, or they can also aﬀect our metabolism through blocking or disrupting various enzymatic pathways [21,22]. The detection and imaging of bacteria by staining has been a reliable method for more than a century. Commercial unavailability of any bacteria speciﬁc ﬂuorescent stain has limited the process of ﬂuorescent labelling. Most of the existing stains follow a general mode of action, where it stains almost all types of cells including mammalian cells, or it is artiﬁcially designed for speciﬁc target staining. For example, BacLight green is a bacterial speciﬁc stain that contains SYTO9, frequently used in live/dead cell imaging of bacteria, but SYTO9 is an intercalating membrane permeable green stain that stains all cells containing nucleic acid [23]. Moreover the vital stain, ﬂuorescein diacetate (FDA) is based on intracellular hydrolysis of FDA and in the case of living cells, the non-ﬂuorescent FDA is converted into ﬂuorescein, a green ﬂuorescent compound, indicating the cell viability [24,25]. The detection of bacteria in human tissue during infection development, or the host pathogen interaction study by using a ﬂuorescent dye is quite challenging due to the generalized mode of action of commercialized organic dyes. Thus, the necessity of developing a bacteria speciﬁc stain to improve the bacterial detection and imaging is required and long awaited. Nanoparticles (NPs) can be derived from various sources in a desirable nanometer (nm) size. Due to their enhanced retention and permeability, NPs have a broad range of clinical applications along with their ability to accumulate the sites of infection by enhancing retention and the permeability eﬀect [26]. NPs with few modiﬁcations help in site speciﬁc drug delivery by reducing target toxicity [27]. Quantum dots (QDs) are a type of nanoparticles with the ability to ﬂuoresce, which is helping in the development of new detection and imaging techniques [27–29]. A few novel QDs have been reported to have antibacterial [30–32] and anticancer properties [33,34], which elevates their value even more in the ﬁeld of biomedicine with some drawbacks like toxicity, inﬂammation, and stability [27]. 1. Introduction Antibiotic resistance in bacteria is a major challenge to medical sciences and the priority list by World Health Organization (WHO) categorizes it into critical, high, and medium [1]. Research for the past two decades have produced only two new types of antibiotics: lipopeptides and oxazolidinones, although majorly overcoming antibiotic resistance has not been achieved yet. The present situation of bacterial infections is considered to be a threat to human society due to the evolving strains of Nanomaterials 2020, 10, 778; doi:10.3390/nano10040778 www.mdpi.com/journal/nanomaterials 2 of 18 Nanomaterials 2020, 10, 778 Staphylococcus aureus with resistance to broad-range antibiotics such as methicillin and vancomycin [2]. The infections caused by the resistant bacteria are usually treated with existing antibiotics, which usually exerts side eﬀects with their incomplete dosage and improper use. The overuse and misuse of antibiotics are the key factors contributing to antibiotic resistance [3]. The overuse of antibiotics, especially taking antibiotics even when they are not the appropriate treatment, promotes antibiotic resistance [4]. On the other hand side, the eﬀects can be varied, for example, vomiting, nausea (feeling like one may vomit), diarrhea, bloating and indigestion, abdominal pain, loss of appetite, a raised, itchy skin rash (urticaria, or hives), coughing, wheezing, tightness of the throat, feeling lightheaded or faint, breathing diﬃculties such as fast, shallow breathing, wheezing, a fast heartbeat, clammy skin, confusion and anxiety, collapsing or losing consciousness [5]. Furthermore, some antibacterial agents can elicit adverse eﬀects against the hepatic system, where the types of liver damage induced by antibacterial agents cover cytotoxic, injury, cholestatic injury, mixed cytotoxic and cholestatic injury, steatosis, chronic, active hepatitis, and cirrhosis [6]. This further increases the complexity in the development of the resistance against these molecules, assisting it in being a global health problem [7,8]. Thus, the importance of developing new treatment strategies or alternatives for antibiotics against these multidrug resistant bacterial strains are the major demands in 21st century medical science [9–11]. To overcome the antibiotic resistance and absolute necessity of ﬁnding an alternative to existing antibiotics with negligible toxicity, proper biocompatibility, and fulﬁl the needs of clinical research, novel antibacterial agents like nanoparticles, peptides, metabolites, oligonucleotides, and other biogenic or chemical compounds have been introduced [8,12–17]. Nanoparticles, nowadays, are the focus of medical sciences with various applications like drug delivery, tracking, and alternatives to antibiotics such as nanomedicine [18–20]. However, nanoparticles also have adverse eﬀects toward human health. 1. Introduction Titanium dioxide (TiO2) based nanocomposites have been reported to have antimicrobial activity against Escherichia coli, Staphylococcus aureus, Streptococcus sobrinus, Pseudomonas putida, Pseudomonas aeruginosa, and Listeria innoculim due to the production of reactive oxygen species (ROS), which can inactivate organic and inorganic pollutants and consequently inactivate the microorganisms, even 3 of 18 Nanomaterials 2020, 10, 778 where environmental and economic factors have raised their volume [35–38]. The gadolinium based nanoparticles also have antimicrobial activities against Bacillus subtilis, E. coli, and Staphylococcus aureus [39,40]. Though both gadolinium and TiO2 nanoparticles are biologically active, no study has tested their real implications and eﬀects against vancomycin resistant Staphylococcus aureus (VRSA) and methicillin resistant Staphylococcus aureus (MRSA) yet. In this study, we used titanium and gadolinium together to synthesize a titanium–gadolinium quantum dot (TGQG), the ﬁrst of its kind, and its characterization was undertaken using dynamic light scattering, spectrophotometry, X-ray ﬂuorescence, scanning electrochemical microscope (SECM), and scanning electron microscope (SEM). Thereafter, TGQD was used as a stain in the detection of VRSA using a ﬂuorescence microscope. Furthermore, the antibacterial activity was tested against the pathogenic strains of vancomycin resistant S. aureus (VRSA), methicillin resistant S. aureus (MRSA), and the mechanism of action was also investigated. Finally, the cytotoxicity test was performed and conclusions were drawn. 2.2. Chemicals and Synthesis of Titanium–Gadolinium Quantum Dots (TGQDs) 2.2. Chemicals and Synthesis of Titanium–Gadolinium Quantum Dots (TGQDs) The preparation of the gadolinium Schiﬀbase was according to our previous studies with necessary modiﬁcation [28]. Initially, 1.08 mL of diethylene triamine and 1.9 mL of 2-pyridinecarboxaldehyde were mixed and heated under reﬂux in 35 mL of methanol (MeOH) for 6 h. To prepare the desired solution of Schiﬀbase [(2-[(E)-2-pyridylmethyleneamino]-N-[2-[(E)-2-pyridylmethylene-amino]ethyl] -ethanamine)], the solution was cooled and the volume was made up to 50 mL using MeOH. In a separate beaker, 5 mL of gadolinium nitrate aqueous solution (90 mg/mL) was mixed with 10 mL of MeOH, which was subsequently mixed with 5 mL solution of the Schiﬀbase. Finally, the solutions were mixed on hot plate magnetic stirrer for 2 h at 40 ◦C and the volume was made up to 100 mL with deionized water. The Gadolinium (Gd)–Schiﬀbase solution was stored at 25 ◦C. To prepare the titanium dioxide (TiO2) solution, 50 mL of isopropanol and 50 µL of titanium butoxide was put into a glass beaker and placed on a magnetic stirrer for 48 h, and the clear solution turned into a milky white solution without precipitation. Then, the aqueous phase synthesis of quantum dots was carried out by mixing the Gd–Schiﬀbase solution with TiO2 in a beaker. The Gd–Schiﬀbase and TiO2 solutions were mixed in the ratio of 1:1. Later, 2 mL of the solution was taken in a glass tube and heated for 10 min at 80 ◦C, 300 W, (ramping time, 10 min) under microwave irradiation (Multiwave 3000, Anton Paar GmbH, Graz, Austria) to prepare the TG quantum dots. The samples were ﬁltered using 3 kDa (Amicon Ultra 0.5 mL centrifugal ﬁlters) ﬁlters and further ﬁltered through a 0.22 µm membrane. To remove the unreacted initiators, the solution was dialyzed against deionized water several times, then the TGQD were dried under a vacuum dry system. 2.1. Chemicals 2-Pyridinecarboxaldehyde, diethylene triamine, gadolinium (III) nitrate, titanium (IV) butoxide, isopropanol, and methanol were obtained from Sigma-Aldrich (St. Louis, MO, USA). Muller Hinton broth (MH) was obtained from HiMedia Laboratories Pvt. Ltd., Thane, India. 2.3. Characterization of TGQD The prepared TGQDs were visualized under a UV transilluminator at excitation wavelengths (λex) of 312 nm and 270 nm (Transilluminator Multiband TFX-35.MC, Torcy, France). The ﬂuorescence and absorbance spectra of the TG QDs were obtained using a microplate (UV plate, 96 well; Corning Incorporate, Corning, NY, USA) in Tecan Inﬁnite m200 PRO (Männedorf, Switzerland). A total of 100 µL of the samples were used for the measurements. The ﬂuorescence spectrum was measured using an excitation wavelength (λex) of 230 nm, the emission wavelength (λem) range was 280–850 4 of 18 Nanomaterials 2020, 10, 778 nm, and the absorbance spectrum of the TGQDs was measured from 230 to 1000 nm. The QDs were observed under a ﬂuorescence microscope with a UV ﬁlter for visualization. The photoluminescence quantum yield of the TGQDs was determined using the reference as rhodamine 6G according to a reported protocol [41]. The scanning electron microscopy (SEM) instrument (TESCAN Company, Brno, Czech Republic, EU) was used to observe the TGQDs under the following conditions: high vacuum mode (10–3 Pa), voltage of 15 kV, and work distance of 3 mm [28].The scanning electrochemical microscopy (SECM) measurements were performed using a CHI 900 setup (CH Instrument Inc., Austin, TX, USA). A Pt ultramicroelectrode (d = 10 mm) was the SECM tip (RG factor = 10), a bare glassy carbon (GC) covered with 10 µL Naﬁon 1%, and a GC immobilized with TGQDs and covered with 10 µL Naﬁon 1% served as the SECM substrates. The tip was positioned near the substrate using the probe-approach curve (PAC) technique and a potential of 500 mV was applied to the tip with a scan rate of 0.5 µm/s. A platinum wire was the counter electrode and the reference electrode was Ag/AgCl (3 M KCl). A solution of 1 mM of ferrocenemethanol (FcOH) in 0.1 M KCl was used as the redox mediator. All potentials were referred to the reference electrode. TGQDs were also characterized using an elemental analyzer SPECTRO XEPOS energy dispersive x-ray ﬂuorescence (ED-XRF) spectrometer (SPECTRO Analytical Instruments GmbH, Kleve, Germany) equipped with a 10 mm2 Si-Drift Detector with Peltier cooling and a 75 µm Be side window was employed. The instrument uses a Pd-target end window tube at a maximum power of 50 W and a maximum voltage of 50 kV. Spectral resolution of the instrument (FWHM) was <170 eV for Mo Kα (measured under input count rate 10,000 pulses). 2.3. Characterization of TGQD SPECTRO XEPOS was operated and data were evaluated by means of the software Spectro X-Lab Pro, Version 2.5, Kleve, Germany. For the excitation of light elements (Mg–V, 25 kV), a HOPG (highly oriented pyrolithic graphite) crystal target was used. For the determination of heavier elements, a Mo secondary target (Cr–Y, Hf–U, 45 kV) and Al2O3 polarization target (Zr–Ce, 49.5 kV) were used. The sample (1.0 mL) was dried directly at 70 ◦C in a sample cup (32 mm in diameter) on polypropylene thin-ﬁlm (Specac Ltd., Kent, UK) and measured in vacuum using the so-called Turboquant method (fundamental parameters method) [42]. Fourier transform infrared (FTIR) spectra were measured using a Thermo Scientiﬁc Nicolet iS5 spectrometer (Thermo Fisher Scientiﬁc, Waltham, MA, USA) equipped with an iD5 Diamond ATR accessory over a wave number range of 4000–550 cm–1. The average size of the TG QDs, the size distribution, and zeta potential were determined by quasielastic laser dynamic light scattering (DLS) with a Malvern Zetasizer (NANO-ZS, Malvern Instruments Ltd., Worcestershire, UK). Initially, 1.5 mL of an aqueous solution of TGQDs (1 mg/mL) was poured into a polystyrene latex cell and measured at a temperature of 25 ◦C with a detector angle of 173◦, a wavelength of 633 nm, a refractive index of 0.30, and a real refractive index of 1.59 [28]. 2.4. Application of TGQD on Bacteria for Detection and Killing The Petri dish was coated with Poly-L-Lysine and then the VRSA cells were added to the Petri dish for their attachment with the Poly-L-Lysine coated Petri dish surface. The Petri dish was ﬁlled with MH Broth, which was composed of beef infusion form 300.00 g/L, casein acid hydrolysate 17.50 g/L, starch 1.50 g/L, and KCl 7.45 g/L. The probe electrode (Pt 10 µm) was inserted in the Petri dish and the cyclic voltammetry (CV) was recorded. Standard isolates of S. aureus, VRSA, and MRSA were cultivated and used for this experiment. All bacterial isolates were grown in Mueller Hinton broth in a shaking incubator at 37 ◦C. The susceptibility of bacterial cultures against TGQD detection was performed by the unaided eye [45,46,51,52]. The minimum inhibitory concentration (MIC) is the lowest concentration that inhibits bacterial growth. The MIC of TGQD was obtained by adding TGQD from lower to higher concentrations in microplate wells, mixed with bacterial cultures (0.1 O.D. equivalent to 0.5 MacFarland), and incubated at 37 ◦C for 24 h. The final working concentration gradient of TGQD in the microplate wells were 45 µg/mL, 55 µg/mL, 62.5 µg/mL, 85 µg/mL, and 125 µg/mL. The lowest concentration of TGQD added in the microplate well and the microplate well showed almost no bacterial growth (transparent medium with no turbidity), which was counted as the MIC value of TGQD against that specific bacterium. The control was the bacteria without TGQD treatment. The growth curve analysis for S. aureus, VRSA, and MRSA were performed for further conﬁrmation of antibacterial activity of TGQD. Standard isolates of S. aureus, VRSA, and MRSA were cultivated and used for this experiment. All bacterial isolates were grown in MH broth in a shaking incubator at 37 ◦C. The control was the bacterial culture with no TGQD treatment. The antibacterial activity of the TGQDs were measured by Bioscreen C MBR (Oy Growth Curves Ab Ltd., Helsinki, Finland) for 24 h at 37 ◦C using a multichannel pipette system in the 10 × 10 Honeycomb optical microplate well systems (Oy Growth Curves Ab Ltd., Helsinki, Finland) [51,53–56]. Diﬀerent concentrations of TGQDs (45 µg/mL, 55 µg/mL, 62.5 µg/mL, 85 µg/mL, 125 µg/mL) were added to bacterial cultures (0.1 absorbance equivalent to 0.5 McFarland) in the microplate wells. 2.4. Application of TGQD on Bacteria for Detection and Killing The bacterial strains Staphylococcus aureus (S. aureus) (NCTC 8511) [43,44], vancomycin resistant Staphylococcus aureus (CCM 1767) [19,45], and methicillin resistant Staphylococcus aureus (MRSA) ST239:SCCmec IIIA) [46] were obtained from the Czech Collection of Microorganisms, Faculty of Science, Masaryk University, Brno, Czech Republic and from England, with the cooperation of the University of Cambridge [19,43–46]. The bacterial cells were cultivated in Muller Hinton (MH) broth medium, pH 7.4. The bacterial cultures were cultivated overnight at 37 ◦C in a shaking incubator. Bacterial optical density was adjusted to 0.1 absorbance (0.5 MacFarland standards) at 600 nm for the subsequent experiments [45–50]. The absorbance spectral scan was recorded of the bacterial sample for their detection by using Tecan Inﬁnite m200 PRO immediately after preparing the samples. Bacterial culture was adjusted up to 0.012 absorbance (OD600) and incubated with TGQD for 5 min, then the bacterial sample incubated with TGQD was placed on a spectrophotometric plate and the plate was scanned under the Tecan Inﬁnite m200 PRO. 5 of 18 Nanomaterials 2020, 10, 778 Furthermore, the bacterial detection via TGQD was carried out through microscopic analysis. The bacterial sample (VRSA) was stained with TGQD (15 µg/mL) and Images were obtained using an optical Olympus BX51 ﬂuorescence microscope equipped with 40× phase contrast lens. Thereafter, further conﬁrmation was obtained by treating bacteria, bacteria–human cell co-culture with TGQD, and microscopic images were captured. For the co-culture and the bacterial sample, VRSA was added in a 24-well cell culture plate in the presence of PNT1A (human cells) cells with 60% conﬂuency, incubated for 1 h at 37 ◦C. The co-culture system containing bacteria and human cells together was stained by TGQD (15 µg/mL), then the plate was used for microscopy. The microscopy was performed by using an microscopic inverted Olympus IX 71S8F3 ﬂuorescence microscope (Olympus Corporation, Tokyo Japan), which was equipped with an Olympus UIS2 series objective LUCplanFLN 40X (N.A. 0.6, WD 2.7–4 mm, F.N.22), and a mercury arc lamp X-cite 12 (120W; Lumen Dynamics, Mississauga, ON, Canada) was used. As the control, bacterial cells and the human cell line PNT1A were also stained with TGQD and observed under a microscope. p The bactericidal eﬀect of TGQD was analyzed by scanning electrochemical microscopy (SECM). where a Petri dish containing the VRSA treated with TGQD was subjected to the SECM study. 2.5. In Vitro Cytotoxicity Testing Assay The MDA-MB-231 (mammary gland adenocarcinoma cells) and HBL-100 (mammary gland epithelial cells) human cell line were used to study the cytotoxicity of TGQDs. The MTT 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) assay was used to study the cell viability. The cells were maintained in RPMI-1640 medium with 10% fetal bovine serum, supplemented with penicillin (100 u/mL) and streptomycin (0.1 mg/mL). In a microtiter plate, each well was ﬁlled with 5000 cells in 50 µL medium, followed by 24 h incubation at 37 ◦C with 5% CO2. After 24 h incubation with TGQDs, 10 µL of MTT (5 mg/mL in PBS) was added, and incubated for 4 h at 37 ◦C with 5% CO2. When this incubation period was over, MTT containing medium was replaced and 100 µL of 99.9% dimethyl sulfoxide (v/v) was added for 5 min. The absorbance was taken at 570 nm by an Inﬁnite m200 PRO reader [46,62] 2.4. Application of TGQD on Bacteria for Detection and Killing The microplate with the samples was then incubated in the growth curve analyzer (Bioscreen C MBR) for 24 h and the results were evaluated accordingly [46]. Furthermore, the cell leakage assay was carried out. The bacterial sample was treated with TGQD and incubated at 37 ◦C for 4 h, which showed the cells treated with the TGQD were precipitated after 4 h. The treated samples as well as the untreated samples were centrifuged for cell debris removal and the supernatant was separately kept in a micro-centrifuge tube for quantiﬁcation of DNA and RNA by a spectrophotometer (Tecan Inﬁnite m200 PRO). The wavelength of 260 nm was used for the DNA and RNA quantiﬁcation. For the DNA quantiﬁcation, the samples were treated with RNase, and for the RNA quantiﬁcation samples, we treated them with DNase, and puriﬁed them in a column puriﬁcation system. After column puriﬁcation, the absorbance spectra were recorded at 6 of 18 Nanomaterials 2020, 10, 778 260 nm. The presence of DNA and RNA were calculated according to a 260/280 ratio. According to the 260/280 ratio, ~1.8 is generally accepted as pure for DNA and ~2.0 is generally accepted as pure for RNA [57–61]. The DNA in the supernatant was used as a template with 16S rRNA primers (16S Forward-ACTGGGATAACTTCGGGAAAC, and 16S reverse-CAGCGCGGATCCATCTATAA) to perform PCR ampliﬁcation, and the conﬁrmation was done using agarose gel electrophoresis with 1.3% agarose gel [45]. The positive control was genomic DNA and the negative control was the supernatant from the untreated bacterial sample. p To conﬁrm the presence of all types of live cells (bacterial and human cells), SYTO9 was used. Thereafter, an optical Olympus BX51 ﬂuorescence microscope equipped with a 40× phase contrast lens was used to study the antibacterial activity of the TGQD against VRSA. Initially, the samples were incubated with bacteria and TGQDs (respective MIC) for 4 h at 37 ◦C in a shaking incubator. Two diﬀerent kinds of ﬂuorescent dyes were used: SYTO9 and propidium iodide (PI) for live and dead cell staining in equal proportion. Fluorescence dyes were added to the sample and observed under microscope. 3.1. Synthesis and Characterization of TGQDs The characterization was initiated by measuring the ﬂuorescence and absorbance spectra of the TGQDs, as shown in Figure 1a. The Gd–Schiﬀbase was mixed with TiO2 in a 1:1 ratio and cooked at 80 ◦C to form TGQDs, which when observed under microscopy showed a blue ﬁeld due to its high ﬂuorescence as observed under a UV ﬁlter (Figure 1b). The TGQD under UV transilluminator (λ = 312 nm) showed high ﬂuorescence intensity with a bluish white color (Figure 1c). The ﬂuorescence emission maximum and the absorbance maximum of the TGQDs were observed at λ = 375 nm and λ = 262 nm, respectively. Furthermore, it was characterized by x-ray ﬂuorescence spectrophotometry using the secondary target as Mo. The results indicate speciﬁc peaks that showed the presence of Gd and Ti in the sample, as shown in Figure 1d. Moreover, the particle size distribution, average particle size, and the zeta potential of the prepared TGQDs were analyzed using SEM and DLS (Figure 2). The results revealed that the size of the TGQDs were found to be in the range of 45 ± 2 nm to 95 ± 2 nm (Figure 2a) and the value of the zeta potential showed that the value was highest at 58.7 ± 0.13 mV (Figure 2c). The size determined by DLS was in good agreement with the SEM image and data, as shown in Figure 2b. Finally, the FTIR spectrum of TGQD (Figure S3) indicates the presence of the diﬀerent vibrations of the functional group, showing the presence of all the bond and functional groups present in the gadolinium nitrate and titanium butoxide, initial component of TGQD (Figures S1 and S2). In titanium, the functional group NO3 has asymmetric and symmetric stretching from 1369 to 1381 cm−1 and 1340 to 1267 cm−1 [63]. In the case of titanium butoxide absorptions at 1125 cm−1 (Ti–O–C vibration), the intensity of the bands was from 1490 to 1340 cm−1 [64]. The SECM images obtained are shown in 7 of 18 7 of 18 Nanomaterials 2020, 10, 778 Figure 3. The scale of the current measured on the bare glassy carbon (GC) was from 0.09 to 0.2 nA. The GC immobilized with TGQD showed a current range mainly from 0.6 to 0.7 nA. Thus, it can be concluded that the current interval in both cases was the same (almost 1 nA). Hence, the inhomogeneity of both electrodes is comparable. 3.1. Synthesis and Characterization of TGQDs It can be concluded that the homogeneity of the both electrodes is mainly aﬀected by the Naﬁon layer. However, the current increased signiﬁcantly in the presence of the TGQD. This can be attributed to the good electrochemical conductivity of the synthesized TGQD, which is due to low electrochemical resistance of their metallic components. Moreover, the charge of the QDs may inﬂuence their electrochemical conductivity. However, we found that the TGQDs were water soluble and showed high stability (ﬂuorescence measurement) in solution, even after four months from its prepared time. Nanomaterials 2020, 10, x FOR PEER REVIEW 7 of 19 found that the TGQDs were water soluble and showed high stability (fluorescence measurement) in solution, even after four months from its prepared time. Figure 1. (a) The fluorescence spectrum is given in the blue line and the absorbance spectrum is given in the green line. (b) Titanium-Gadolinium Quantum dot (TGQD) observed under the UV filter of the fluorescence microscope. (c) TGQD shows a bluish white color under UV transilluminator; (d) X–ray fluorescence (XRF) spectrum confirms that the synthesized TGQDs contained both gadolinium and titanium metal Figure 1. (a) The ﬂuorescence spectrum is given in the blue line and the absorbance spectrum is given in the green line. (b) Titanium-Gadolinium Quantum dot (TGQD) observed under the UV ﬁlter of the ﬂuorescence microscope. (c) TGQD shows a bluish white color under UV transilluminator; (d) X–ray ﬂuorescence (XRF) spectrum conﬁrms that the synthesized TGQDs contained both gadolinium and titanium metal. Figure 1. (a) The fluorescence spectrum is given in the blue line and the absorbance spectrum is given in the green line. (b) Titanium-Gadolinium Quantum dot (TGQD) observed under the UV filter of the fluorescence microscope. (c) TGQD shows a bluish white color under UV transilluminator; (d) X–ray fluorescence (XRF) spectrum confirms that the synthesized TGQDs contained both gadolinium and Figure 1. (a) The ﬂuorescence spectrum is given in the blue line and the absorbance spectrum is given in the green line. (b) Titanium-Gadolinium Quantum dot (TGQD) observed under the UV ﬁlter of the ﬂuorescence microscope. (c) TGQD shows a bluish white color under UV transilluminator; (d) X–ray ﬂuorescence (XRF) spectrum conﬁrms that the synthesized TGQDs contained both gadolinium and titanium metal. 8 of 18 Nanomaterials 2020, 10, 778 Nanomaterials 2020, 10, x FOR PEER REVIEW 8 of 19 Figure 2. (a) Zeta size of the TGQD. 3.1. Synthesis and Characterization of TGQDs (b) SEM analysis of TGQD with 20 µm scale bar;.(c) Zeta potential of TGQDs. Data represent the mean ± SD, n = 5. Figure 2. (a) Zeta size of the TGQD. (b) SEM analysis of TGQD with 20 µm scale bar;.(c) Zeta potential of TGQDs. Data represent the mean ± SD, n = 5. Nanomaterials 2020, 10, x FOR PEER REVIEW 9 of 19 Figure 2. (a) Zeta size of the TGQD. (b) SEM analysis of TGQD with 20 µm scale bar;.(c) Zeta potential of TGQDs. Data represent the mean ± SD, n = 5. Figure 2. (a) Zeta size of the TGQD. (b) SEM analysis of TGQD with 20 µm scale bar;.(c) Zeta potential of TGQDs. Data represent the mean ± SD, n = 5. Nanomaterials 2020, 10, x FOR PEER REVIEW 9 of 19 Figure 3. (a) Scanning electrochemical microscopy (SECM) image of glassy carbon electrode covered with Nafion. (b) SECM image of glassy carbon immobilized with QDs and covered with Nafion. The applied parameters: the tip electrode scan rate: 20 µm/s, the applied potential on tip: 0.5 V, the tip— substrate distance: 10 µm, quiet time: 30 s, the scanned surface area 100 µm × 100 µm. The measurements were carried out in a solution of 1 mm FcOH and 0.1 M KCl. 3.2. Detection of Bacteria Figure 3. (a) Scanning electrochemical microscopy (SECM) image of glassy carbon electrode covered with Naﬁon. (b) SECM image of glassy carbon immobilized with QDs and covered with Naﬁon. The applied parameters: the tip electrode scan rate: 20 µm/s, the applied potential on tip: 0.5 V, the tip—substrate distance: 10 µm, quiet time: 30 s, the scanned surface area 100 µm × 100 µm. The measurements were carried out in a solution of 1 mm FcOH and 0.1 M KCl. Figure 3. (a) Scanning electrochemical microscopy (SECM) image of glassy carbon electrode covered with Nafion. (b) SECM image of glassy carbon immobilized with QDs and covered with Nafion. The applied parameters: the tip electrode scan rate: 20 µm/s, the applied potential on tip: 0.5 V, the tip— substrate distance: 10 µm, quiet time: 30 s, the scanned surface area 100 µm × 100 µm. The measurements were carried out in a solution of 1 mm FcOH and 0.1 M KCl. 3 2 D i f B i Figure 3. 3.2. Detection of Bacteria This experimental data was further validated by control cell imaging (Figure S6) where TGQD stained VRSA cells showed high ﬂuorescence under a ﬂuorescence microscope, but the PNT1A cells were unstained and no ﬂuorescence was observed. ) p Q g g the UV filter of the fluorescence microscope (Figure S5). Detection (through imaging) of the VRSA in the presence of PNT1A (coculture imaging) was carried out under fluorescence microscopy. The samples were stained with a very small amount (10 to 15 µg/mL) of TGQD. The VRSA cells in the presence of PNT-1A were stained with TGQD and SYTO9 to understand the dye specificity (whether TGQD is specific toward bacteria or not) toward bacteria and human cells. The bright field image (Figure 4a) showed the presence of bacteria and human cells (indicated by arrow) in the field of observation. Figure 4b shows that TGQD positively stained the bacterial cells, but PNT1A cells were unstained and not visible in the field. In Figures 4c,e (merged image of Figures 4a,c), it can be seen that the SYTO9 stained both the VRSA and the PNT1A cells, which confirmed that SYTO9 does not have a cell specific staining nature. In contrast, TGQD specifically stained the bacterial cells, which was clearly visible in Figures 4b, d (merged image of Figures 4a,b) and f (merged image of Figures 4b,c). Finally, from Figure 4f (the merged image of Figures 4b,c), it can be clearly concluded that STYO9 is neither specific toward bacteria nor human cells (because both bacterial and human cells were stained), whereas the TGQD is a bacteria specific stain (only bacterial cells were stained). According to the best of our knowledge, there is no such TGQD that can stain only bacteria and not the human cells. This experimental data was further validated by control cell imaging (Figure S6) where TGQD stained VRSA cells showed high fluorescence under a fluorescence microscope, but the PNT1A cells were unstained and no fluo e e e a ob e ed Figure 4. The microscopic images. (a) VRSA in presence of PNT1A cells; (b) VRSA in presence of PNT1A cells stained with TGQDs and imaged under UV filter; (c) VRSA mixed with PNT1A cells stained with SYTO9 and imaged under fluorescence microscope through 485/498 filter; (d) Merged image of (a) and (b); (e) Merged image of (a) and (c); (f) Merged image of (b) and (c). 3.3. Antibacterial Activities Figure 4. 3.2. Detection of Bacteria Furthermore, according to the spectrophotometry it is clear that the bacterial cells were detected by applying TGQDs, where bacterial cells presented a 0.012 O.D. value, whereas absorbance from only bacteria and only QDs were lower than the QD mixed sample. Lower concentrations (0.012 O.D. concentration) of the bacterial sample were detected through the spectrophotometric method (Figure S4). The bacterial sample was stained with TGQD, which gave a high bluish white ﬂuorescence under the UV ﬁlter of the ﬂuorescence microscope (Figure S5). Nanomaterials 2020, 10, 778 9 of 18 only bacteria and only QDs were lower than the QD mixed sample. Lower concentrations (0.012 O.D. concentration) of the bacterial sample were detected through the spectrophotometric method (Figure S4). The bacterial sample was stained with TGQD, which gave a high bluish white fluorescence under g Detection (through imaging) of the VRSA in the presence of PNT1A (coculture imaging) was carried out under ﬂuorescence microscopy. The samples were stained with a very small amount (10 to 15 µg/mL) of TGQD. The VRSA cells in the presence of PNT-1A were stained with TGQD and SYTO9 to understand the dye speciﬁcity (whether TGQD is speciﬁc toward bacteria or not) toward bacteria and human cells. The bright ﬁeld image (Figure 4a) showed the presence of bacteria and human cells (indicated by arrow) in the ﬁeld of observation. Figure 4b shows that TGQD positively stained the bacterial cells, but PNT1A cells were unstained and not visible in the ﬁeld. In Figure 4c,e (merged image of Figure 4a,c), it can be seen that the SYTO9 stained both the VRSA and the PNT1A cells, which conﬁrmed that SYTO9 does not have a cell speciﬁc staining nature. In contrast, TGQD speciﬁcally stained the bacterial cells, which was clearly visible in Figure 4b,d (merged image of Figure 4a,b) and f (merged image of Figure 4b,c). Finally, from Figure 4f (the merged image of Figure 4b,c), it can be clearly concluded that STYO9 is neither speciﬁc toward bacteria nor human cells (because both bacterial and human cells were stained), whereas the TGQD is a bacteria speciﬁc stain (only bacterial cells were stained). According to the best of our knowledge, there is no such TGQD that can stain only bacteria and not the human cells. 3.1. Synthesis and Characterization of TGQDs (a) Scanning electrochemical microscopy (SECM) image of glassy carbon electrode covered with Naﬁon. (b) SECM image of glassy carbon immobilized with QDs and covered with Naﬁon. The applied parameters: the tip electrode scan rate: 20 µm/s, the applied potential on tip: 0.5 V, the tip—substrate distance: 10 µm, quiet time: 30 s, the scanned surface area 100 µm × 100 µm. The measurements were carried out in a solution of 1 mm FcOH and 0.1 M KCl. 9 of 18 Nanomaterials 2020, 10, 778 3.3. Antibacterial Activities N t i l 2020 10 778 The antibacterial activity of TGQDs was initiated by the determination of the MIC of TGQD, which was performed by the broth micro-dilution method against VRSA, MRSA, and S. aureus. The MIC of the TGQD against VRSA, MRSA, and S. aureus was obtained as 62.5, 62.5, and 55 µg/mL, as shown in Table 1. However, the TGQD concentrations below the respective MIC value showed turbidity in the solutions, proving the presence of bacteria in the solutions. Thereafter, diﬀerent concentrations of the TGQDs (125 µg/mL, 85 µg/mL, 62.5 µg/mL, 55 µg/mL, and 45 µg/mL) were used to further understand the growth curves of respective bacteria in the presence of TGQD [45,51]. The 24 h growth curve of S. aureus showed more than 97% inhibition up to 55 µg/mL, but concentrations below showed lower inhibitory eﬀects, as shown in Figure 5a. However, in the case of VRSA and MRSA, the inhibition was signiﬁcant up to more than 98% at 62.5 µg/mL, but concentrations below showed no inhibitory eﬀects, as shown in Figure 5b,c. Depending on the previous research, vancomycin resistant strains (contains VanA clustered gene) growth was inhibited with the use of 512 µg/mL of vancomycin [65,66], but this is not suﬃcient for preventing or curing infections by vancomycin resistant strains. Thus, it can be concluded that our synthesized TGQDs are novel and biologically active antibacterial compounds that can inhibit the growth of resistant strains of S. aureus almost completely by using 62.5 µg/mL concentrations. Thereafter, the mechanism of action of TGQD was studied. Nanomaterials 2020, 10, 778 10 of 18 MIC of the TGQD against VRSA, MRSA, and S. aureus was obtained as 62.5, 62.5, and 55 µg/mL, as shown in Table 1. However, the TGQD concentrations below the respective MIC value showed turbidity in the solutions, proving the presence of bacteria in the solutions. Thereafter, different concentrations of the TGQDs (125 µg/mL, 85 µg/mL, 62.5 µg/mL, 55 µg/mL, and 45 µg/mL) were used to further understand the growth curves of respective bacteria in the presence of TGQD [45,51]. The 24 h growth curve of S. aureus showed more than 97% inhibition up to 55 µg/mL, but concentrations below showed lower inhibitory effects, as shown in Figure 5a. However, in the case of VRSA and MRSA, the inhibition was significant up to more than 98% at 62.5 µg/mL, but concentrations below showed no inhibitory effects, as shown in Figures 5b,c. 3.2. Detection of Bacteria The microscopic images. (a) VRSA in presence of PNT1A cells; (b) VRSA in presence of PNT1A cells stained with TGQDs and imaged under UV ﬁlter; (c) VRSA mixed with PNT1A cells stained with SYTO9 and imaged under ﬂuorescence microscope through 485/498 ﬁlter; (d) Merged image of (a) and (b); (e) Merged image of (a) and (c); (f) Merged image of (b) and (c). Figure 4. The microscopic images. (a) VRSA in presence of PNT1A cells; (b) VRSA in presence of PNT1A cells stained with TGQDs and imaged under UV filter; (c) VRSA mixed with PNT1A cells stained with SYTO9 and imaged under fluorescence microscope through 485/498 filter; (d) Merged image of (a) and (b); (e) Merged image of (a) and (c); (f) Merged image of (b) and (c). Figure 4. The microscopic images. (a) VRSA in presence of PNT1A cells; (b) VRSA in presence of PNT1A cells stained with TGQDs and imaged under UV ﬁlter; (c) VRSA mixed with PNT1A cells stained with SYTO9 and imaged under ﬂuorescence microscope through 485/498 ﬁlter; (d) Merged image of (a) and (b); (e) Merged image of (a) and (c); (f) Merged image of (b) and (c). 10 of 18 10 of 18 Nanomaterials 2020, 10, 778 3.3. Antibacterial Activities N t i l 2020 10 778 Depending on the previous research, vancomycin resistant strains (contains VanA clustered gene) growth was inhibited with the use of 512 µg/mL of vancomycin[65,66] , but this is not sufficient for preventing or curing infections by vancomycin resistant strains. Thus, it can be concluded that our synthesized TGQDs are novel and biologically active antibacterial compounds that can inhibit the growth of resistant strains of S. aureus almost completely by using 62.5 µg/mL concentrations. Thereafter, the mechanism of action of TGQD was studied. Figure 5. Analysis of growth curve for S. aureus (a), VRSA (b), and MRSA (c) treated with TGQD. Data represent the mean ± SD, n = 5. VRSA = vancomycin resistant S. aureus. Figure 5. Analysis of growth curve for S. aureus (a), VRSA (b), and MRSA (c) treated with TGQD. Data represent the mean ± SD, n = 5. VRSA = vancomycin resistant S. aureus. Figure 5. Analysis of growth curve for S. aureus (a), VRSA (b), and MRSA (c) treated with TGQD. Figure 5. Analysis of growth curve for S. aureus (a), VRSA (b), and MRSA (c) treated with TGQD. Data represent the mean ± SD, n = 5. VRSA = vancomycin resistant S. aureus. Data represent the mean ± SD, n = 5. VRSA = vancomycin resistant S. aureus. Table 1 Determining the Minimum Inhibitory Concentration (MICs) of TGQD by the broth Table 1. Determining the Minimum Inhibitory Concentration (MICs) of TGQD by the broth microdilution method. Data represent the mean ± SD, n = 5. VRSA = vancomycin resistant S. aureus. Table 1 Determining the Minimum Inhibitory Concentration (MICs) of TGQD by the broth Table 1. Determining the Minimum Inhibitory Concentration (MICs) of TGQD by the broth microdilution method. g y ( ) microdilution method. Bacterial strains TGQD (µg/mL) S. aureus 55 MRSA 62.5 VRSA 62 5 Bacterial Strains TGQD (µg/mL) S. aureus 55 MRSA 62.5 VRSA 62.5 y ( ) terial strains TGQD (µg/mL S. aureus 55 MRSA 62.5 Bacterial Strains TGQD (µg/mL) S. aureus 55 MRSA 62.5 VRSA 62.5 3.4. Mechanism of Actions 3.4. Mechanism of Actions (b) Cell wall damage and leakage assay where (1) 100 bp ladder, (2) PCR product of 16S rRNA of supernatant of VRSA cells treated with QDs, (3) PCR product of 16S rRNA of supernatant of VRSA cells not treated with TGQDs, and (4) PCR product of 16S rRNA of genomic DNA of VRSA cells treated with TGQDs. As shown in Figure 7a, a reduction peak could be observed around −0.8 V, which was attributed to the reduction in dissolved oxygen. This reduction peak is suitable to track the biological species that consume oxygen [67]. Thus, the potential was set at −0.8 V and the probe approach curve was recorded on a part that was removed from the bacteria. Since the Petri dish is an insulating material, a decrease in the current magnitude was expected when moving the probe toward the Petri dish. When the current dropped to 75% of its initial magnitude, the probe was stopped (Figure 7b). Theoretically, at this current magnitude, the distance between the probe to the bottom of the Petri dish is around 10 µm [68]. When the probe was positioned at an appropriate distance to the Petri dish, it was moved to part of the Petri dish with bacteria. Then, the surface scanning (200 µm × 200 µm) was carried out. As seen in Figure 7c, the current magnitude was mainly between the range of 3.5 to 6.0 nA. Furthermore, the surface seems to be rough, which is due to the presence of the bacteria. It must be noted that due to the oxygen consumption, the current magnitude decreased when the probe was placed above the bacteria, hence the roughness of the surface was due to the presence of bacteria and its rough texture. The effect of the TGQDs on the bacteria was also studied by the addition of TGQDs (50 µg/mL) to the same Petri dish containing the bacteria. After 5 min, when the interaction of the TGQDs and the bacteria was ensured, the same surface of the Petri dish was scanned. As can be observed in Figure 7d, the current magnitude ranged from 3.5 nA to 11.0 nA and with scanning, the current gradually increased. The increase in the current magnitude was attributed to the increase in dissolved oxygen concentration. It can be then assumed that the interaction of the TGQDs with the bacteria resulted in the death of bacteria. 3.4. Mechanism of Actions Then, the bacteria no longer consume dissolved oxygen. Moreover, the surface showed less roughness when compared to that in Figure 7c. When the bacteria die, they detach from the surface of the Petri dish and float in the buffer. Thus, the surface of the Petri dish, which was less rough, was actually imaged after the addition of TGQDs. As shown in Figure 7a, a reduction peak could be observed around −0.8 V, which was attributed to the reduction in dissolved oxygen. This reduction peak is suitable to track the biological species that consume oxygen [67]. Thus, the potential was set at −0.8 V and the probe approach curve was recorded on a part that was removed from the bacteria. Since the Petri dish is an insulating material, a decrease in the current magnitude was expected when moving the probe toward the Petri dish. When the current dropped to 75% of its initial magnitude, the probe was stopped (Figure 7b). Theoretically, at this current magnitude, the distance between the probe to the bottom of the Petri dish is around 10 µm [68]. When the probe was positioned at an appropriate distance to the Petri dish, it was moved to part of the Petri dish with bacteria. Then, the surface scanning (200 µm × 200 µm) was carried out. As seen in Figure 7c, the current magnitude was mainly between the range of 3.5 to 6.0 nA. Furthermore, the surface seems to be rough, which is due to the presence of the bacteria. It must be noted that due to the oxygen consumption, the current magnitude decreased when the probe was placed above the bacteria, hence the roughness of the surface was due to the presence of bacteria and its rough texture. The eﬀect of the TGQDs on the bacteria was also studied by the addition of TGQDs (50 µg/mL) to the same Petri dish containing the bacteria. After 5 min, when the interaction of the TGQDs and the bacteria was ensured, the same surface of the Petri dish was scanned. As can be observed in Figure 7d, the current magnitude ranged from 3.5 nA to 11.0 nA and with scanning, the current gradually increased. The increase in the current magnitude was attributed to the increase in dissolved oxygen concentration. It can be then assumed that the interaction of the TGQDs with the bacteria resulted in the death of bacteria. 3.4. Mechanism of Actions 3.4. Mechanism of Actions Bacterial cells are negatively charged and our synthesized QDs were highly positively charged, for this reason, TGQD had a high binding affinity with the bacterial cells. Therefore, to study and understand the mechanism of action, VRSA cells were incubated with 62.5 µg/mL of TGQD for 4 h at 37 °C in a shaking incubator and the cell leakage assay was performed. The supernatant was used to detect the presence of DNA and RNA after treatment with TGQD. The presence of DNA and RNA was detected by spectrophotometry, as shown in Figure 6a, and further confirmed by the presence of bands around 109 bps in the agarose gel after PCR amplification of the 16S rRNA sequence using the supernatant as the DNA template (Figure 6b). The positive control showed no band, whereas the negative control showed the presence of a band in the same place as seen in the case of the treated Bacterial cells are negatively charged and our synthesized QDs were highly positively charged, for this reason, TGQD had a high binding aﬃnity with the bacterial cells. Therefore, to study and understand the mechanism of action, VRSA cells were incubated with 62.5 µg/mL of TGQD for 4 h at 37 ◦C in a shaking incubator and the cell leakage assay was performed. The supernatant was used to detect the presence of DNA and RNA after treatment with TGQD. The presence of DNA and RNA was detected by spectrophotometry, as shown in Figure 6a, and further conﬁrmed by the presence of bands around 109 bps in the agarose gel after PCR ampliﬁcation of the 16S rRNA sequence using the supernatant as the DNA template (Figure 6b). The positive control showed no band, whereas the negative control showed the presence of a band in the same place as seen in the case of the treated samples. Thus, the result showed that the treatment of VRSA with TGQD caused leakage of the cells, Nanomaterials 2020, 10, 778 11 of 18 Nanomaterials 2020, 10, 778 which may by interacting with the cell membrane, causing the release of inner components containing nucleic acid into the supernatant. As a result, the PCR ampliﬁcation showed ampliﬁed DNA product. Nanomaterials 2020, 10, 778 11 of 18 Figure 6. (a) DNA and RNA concentration measurement from the untreated sample and TGQD treated sample. 3.4. Mechanism of Actions (b) Cell wall damage and leakage assay where (1) 100 bp ladder, (2) PCR product of 16S rRNA of supernatant of VRSA cells treated with QDs, (3) PCR product of 16S rRNA of supernatant of VRSA cells not treated with TGQDs, and (4) PCR product of 16S rRNA of genomic DNA of VRSA cells treated with TGQDs. Figure 6. (a) DNA and RNA concentration measurement from the untreated sample and TGQD treated sample. (b) Cell wall damage and leakage assay where (1) 100 bp ladder, (2) PCR product of 16S rRNA of supernatant of VRSA cells treated with QDs, (3) PCR product of 16S rRNA of supernatant of VRSA cells not treated with TGQDs, and (4) PCR product of 16S rRNA of genomic DNA of VRSA cells treated with TGQDs. Figure 6. (a) DNA and RNA concentration measurement from the untreated sample and TGQD treated sample. (b) Cell wall damage and leakage assay where (1) 100 bp ladder, (2) PCR product of 16S rRNA of supernatant of VRSA cells treated with QDs, (3) PCR product of 16S rRNA of supernatant of VRSA cells not treated with TGQDs, and (4) PCR product of 16S rRNA of genomic DNA of VRSA cells treated with TGQDs. Figure 6. (a) DNA and RNA concentration measurement from the untreated sample and TGQD treated sample. (b) Cell wall damage and leakage assay where (1) 100 bp ladder, (2) PCR product of 16S rRNA of supernatant of VRSA cells treated with QDs, (3) PCR product of 16S rRNA of supernatant of VRSA cells not treated with TGQDs, and (4) PCR product of 16S rRNA of genomic DNA of VRSA cells treated with TGQDs. Figure 6. (a) DNA and RNA concentration measurement from the untreated sample and TGQD treated sample. (b) Cell wall damage and leakage assay where (1) 100 bp ladder, (2) PCR product of 16S rRNA of supernatant of VRSA cells treated with QDs, (3) PCR product of 16S rRNA of supernatant of VRSA cells not treated with TGQDs, and (4) PCR product of 16S rRNA of genomic DNA of VRSA cells treated with TGQDs. Figure 6. (a) DNA and RNA concentration measurement from the untreated sample and TGQD treated sample. 3.4. Mechanism of Actions Then, the bacteria no longer consume dissolved oxygen. Moreover, the surface showed less roughness when compared to that in Figure 7c. When the bacteria die, they detach from the surface of the Petri dish and ﬂoat in the buﬀer. Thus, the surface of the Petri dish, which was less rough, was actually imaged after the addition of TGQDs. 12 of 18 12 of 18 12 of 18 12 of 18 Nanomaterials 2020, 10, 778 Nanomaterials 2020 10 778 Figure 7. (a) dissolved oxygen reduction; (b) drop down percentage of probe; (c) represents high oxygen consumption by the bacteria and a decrease in the magnitude; (d) increased magnitude due t th b (d th) f b t i l l ti Figure 7. (a) dissolved oxygen reduction; (b) drop down percentage of probe; (c) represents high oxygen consumption by the bacteria and a decrease in the magnitude; (d) increased magnitude due to the absence (death) of bacterial population. Figure 7. (a) dissolved oxygen reduction; (b) drop down percentage of probe; (c) represents high oxygen consumption by the bacteria and a decrease in the magnitude; (d) increased magnitude due h b (d h) f b i l l i Figure 7. (a) dissolved oxygen reduction; (b) drop down percentage of probe; (c) represents high oxygen consumption by the bacteria and a decrease in the magnitude; (d) increased magnitude due to the absence (death) of bacterial population. ( ) p p Finally, the VRSA was treated with TGQD to observe the viability of the cells by a live/dead cell assay using a fluorescence microscope equipped with phase contrast. The samples were stained with SYTO9 and propidium iodide (PI), which helps to differentiate the live and dead bacterial cells. The samples were treated with TGQD (125 µg/mL) and observed under the microscope after a time interval of 0 h and after 4 h of treatment. The images of the micro-centrifuge tubes had considerable change in the samples after 4 h of treatment, showing the precipitation of dead VRSA cells along with TGQDs at the bottom of the tube with a clear solution above. Furthermore, the microscopic images also agreed with the precipitated samples after 4 h of treatment. 3.4. Mechanism of Actions The phase contrast image for 0 h showed the presence of the VRSA in bright field, similarly, an identical image of the field under UV excitation proved the interaction of the VRSA with TGQD (both separately and in merged field), without any aggregations of bacteria, as seen in the micro-centrifuge tube. Moreover, the green fluorescence of the bacteria, which was due to the presence of SYTO9, proved the VRSA cells were in a viable state at 0 h. In the case of the 4 h treated sample, it showed aggregation of the bacterial cells in the phase contrast under bright field with the TGQD and under UV excitation also showed the bacteria in the identical position in that field. Further PI emitting red fluorescence confirmed that the cells were dead in that specific field, as shown in Figure 8. Thus, the TGQD not only detects both live and dead bacterial cells by staining both live and dead VRSA cells, but also showed considerable Finally, the VRSA was treated with TGQD to observe the viability of the cells by a live/dead cell assay using a ﬂuorescence microscope equipped with phase contrast. The samples were stained with SYTO9 and propidium iodide (PI), which helps to diﬀerentiate the live and dead bacterial cells. The samples were treated with TGQD (125 µg/mL) and observed under the microscope after a time interval of 0 h and after 4 h of treatment. The images of the micro-centrifuge tubes had considerable change in the samples after 4 h of treatment, showing the precipitation of dead VRSA cells along with TGQDs at the bottom of the tube with a clear solution above. Furthermore, the microscopic images also agreed with the precipitated samples after 4 h of treatment. The phase contrast image for 0 h showed the presence of the VRSA in bright ﬁeld, similarly, an identical image of the ﬁeld under UV excitation proved the interaction of the VRSA with TGQD (both separately and in merged ﬁeld), without any aggregations of bacteria, as seen in the micro-centrifuge tube. Moreover, the green ﬂuorescence of the bacteria, which was due to the presence of SYTO9, proved the VRSA cells were in a viable state at 0 h. 3.4. Mechanism of Actions (a) TGQD treated VRSA at 0 h where (I) stands for the phase contrast image of VRSA, (II) stands for UV excitation image, (III) denotes ‘I’ and ‘II’ merged cells, and (IV) denotes SYTO9 (living cells green ﬂuorescence) and propidium iodide (dead cells red ﬂuorescence) stained cells. (b) TGQD treated VRSA in 4 h where (I) stands for the phase contrast image of VRSA, (II) stands for UV excitation image in phase contrast microscope, (III) denotes ‘I’ and ‘II’ merged cells, and (IV) SYTO9 and propidium iodide stained cells ﬂuorescence image. 3.4. Mechanism of Actions In the case of the 4 h treated sample, it showed aggregation of the bacterial cells in the phase contrast under bright ﬁeld with the TGQD and under UV excitation also showed the bacteria in the identical position in that ﬁeld. Further PI emitting red ﬂuorescence conﬁrmed that the cells were dead in that speciﬁc ﬁeld, as shown in Figure 8. Thus, the TGQD not only detects both live and dead bacterial cells by staining both live and dead VRSA cells, but also showed considerable antibacterial eﬀects against VRSA with time by interacting with the negatively charged cell wall of the bacteria. 13 of 18 13 of 18 Nanomaterials 2020, 10, 778 Figure 8. (a) TGQD treated VRSA at 0 h where (I) stands for the phase contrast image of VRSA, (II) stands for UV excitation image, (III) denotes ‘I’ and ‘II’ merged cells, and (IV) denotes SYTO9 (living cells green fluorescence) and propidium iodide (dead cells red fluorescence) stained cells. (b) TGQD treated VRSA in 4 h where (I) stands for the phase contrast image of VRSA, (II) stands for UV excitation image in phase contrast microscope, (III) denotes ‘I’ and ‘II’ merged cells, and (IV) SYTO9 Figure 8. (a) TGQD treated VRSA at 0 h where (I) stands for the phase contrast image of VRSA, (II) stands for UV excitation image, (III) denotes ‘I’ and ‘II’ merged cells, and (IV) denotes SYTO9 (living cells green ﬂuorescence) and propidium iodide (dead cells red ﬂuorescence) stained cells. (b) TGQD treated VRSA in 4 h where (I) stands for the phase contrast image of VRSA, (II) stands for UV excitation image in phase contrast microscope, (III) denotes ‘I’ and ‘II’ merged cells, and (IV) SYTO9 and propidium iodide stained cells ﬂuorescence image. Figure 8. (a) TGQD treated VRSA at 0 h where (I) stands for the phase contrast image of VRSA, (II) stands for UV excitation image, (III) denotes ‘I’ and ‘II’ merged cells, and (IV) denotes SYTO9 (living cells green fluorescence) and propidium iodide (dead cells red fluorescence) stained cells. (b) TGQD treated VRSA in 4 h where (I) stands for the phase contrast image of VRSA, (II) stands for UV excitation image in phase contrast microscope, (III) denotes ‘I’ and ‘II’ merged cells, and (IV) SYTO9 Figure 8. 4. Conclusions In the present study, the synthesis of the TGQDs was itself a novelty, as it was the ﬁrst of its kind to be synthesized to date. The characterizations of the TGQD helped to obtain the biophysical and biochemical information, stating a size below 100 nm, the charge of the TGQD was positive with the presence of both Ti and Gd in the QD, and the ﬂuorescence and absorbance maxima of the TGQD was in the UV region with a high quantum yield of 29.3%. Subsequently, the TGQD was used to detect the VRSA in a co-culture of VRSA and PNT1A cells. The limitations of ﬂuorescence dye (SYTO9) to stain speciﬁc bacteria were overcome by TGQD because there is no bacteria speciﬁc stain that can stain only bacteria and no other cells, but SYTO9 stains both bacteria and human cells. The TGQD stained the VRSA cells to emit bluish white ﬂuorescence without staining the PNT1A cells, but SYTO9 stained both types of cells emitting green ﬂuorescence. Furthermore, the TGQD inhibits by interacting with the bacterial cells causing cell leakage, and ultimately the release of the inner components. The antibacterial activities of the TGQD against S. aureus and its resistant strains showed strong inhibitory activities, which were validated by microscopic evidence. The mode of action was cell disruption and leakage of cytoplasmic content, which led to the precipitation and accumulation of bacterial cells with the amount of oxygen demand decreasing. However, it was also observed that the presence of any bacteria live or dead was detected by TGQD through ﬂuorescence microscopic analysis. Finally, the cytotoxicity test of the TGQD showed no toxicity and proved the TGQD was biocompatible for normal human cells in vitro. Thus, the examined TGQD with biocompatibility and non-toxicity can be used as potential theranostics to detect and inhibit the presence of bacteria in diﬀerent types of co-cultures or mixed samples with signiﬁcant antibacterial activity. To the best of our knowledge, there is no such staining agent present that can stain only bacterial cells. In various research purposes like host pathogen interaction, study requires a speciﬁc staining agent. Thus, TGQD can be very useful for host pathogen interaction study. and propidium iodide stained ce 3.5. In Vitro Cytotoxicity Analysis 3.5. In Vitro Cytotoxicity Analysis The cytotoxicity of TGQDs was estimated using the 3-(4,5-dimethylthiazol-2-yl)-2,5- diphenyltetrazolium bromide (MTT) assay with mammary gland epithelial cells (HBL-100) and mammary gland adenocarcinoma cells (MDA-MB-231) [69–71]. The viability was studied using different concentrations of TGQD, as shown in Figures S7 and S8. In the case of normal epithelial cells (HBL-100), the TGQDs did not show any sign of toxicity at any concentration tested. Furthermore, the adenocarcinoma cells showed a slight reduction in viability, which was almost negligible and thus no prominent cytotoxicity was seen due to the TGQDs. Thus, the results showing the high viability of normal cells, but slight reduction in viability of cancer epithelial cells in the presence of ThecytotoxicityofTGQDswasestimatedusingthe3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay with mammary gland epithelial cells (HBL-100) and mammary gland adenocarcinoma cells (MDA-MB-231) [69–71]. The viability was studied using different concentrations of TGQD, as shown in Figures S7 and S8. In the case of normal epithelial cells (HBL-100), the TGQDs did not show any sign of toxicity at any concentration tested. Furthermore, the adenocarcinoma cells showed a slight reduction in viability, which was almost negligible and thus no prominent cytotoxicity was seen due to the TGQDs. Thus, the results showing the high viability of normal cells, but slight reduction in viability of cancer epithelial cells in the presence of TGQD at a higher concentration, proved that they are biocompatible and non-toxic toward normal human cells in vitro with some effects against cancer cells. Nanomaterials 2020, 10, 778 14 of 18 14 of 18 4. Conclusions Supplementary Materials: The following are available online at http://www.mdpi.com/2079-4991/10/4/778/s1, Figure S1: FTIR spectrum for gadolinium nitrate, Figure S2: FTIR spectrum for titanium butoxide, Figure S3: Synthesized TGQDs FTIR spectrum, Figure S4: Bacterial detection in low concentration, Figure S5: Bacterial cells stained with TGQD, Figure S6: Bacterial cells and Human PNT1A cells were stained with SYTO9 and TGQD,; Figure S7: MTT assay results for HBL-100, Figure S8: MTT assay results for MDA-MB-231. Author Contributions: All authors researched the data, wrote the manuscript, and reviewed and edited the article substantially. V.P.S. performed most of the experiments including designing the experiments, synthesis of TGQDs, and microbiological tests. A.M (Aninda Mazumdar) wrote most of the manuscript. A.A. did the SECM experiments and wrote the SECM part, A.M. (Atripan Mukherjee) helped to process the SECM sample preparation and writing. V.M. helped in the DLS, Zeta sizer, and SEM pictures. H.M. did the MTT assay, P.K. helped to synthesize the Schiﬀbase, L.R. did the XRF experiment, and A.M. (Amitava Moulick) supervised the whole experiment. All authors have read and agreed to the published version of the manuscript. Funding: This work was ﬁnancially supported by CEITEC 2020 (LQ1601) and by the EFRR project “Multidisciplinary research to increase application potential of nanomaterials in agricultural practice” (No. CZ.02.1.01/0.0/0.0/16_025/0007314). Conﬂicts of Interest: The authors declare no conﬂicts of interest. Conﬂicts of Interest: The authors declare no conﬂicts of interest. References Health Res. Rev. 2008, 9, 227–235. [CrossRef] 12. Edwards-Jones, V. Chapter 1—Alternative Antimicrobial Approaches to Fighting Multidrug-Resistant Infections. In Fighting Multidrug Resistance with Herbal Extracts, Essential Oils and Their Components; Rai, M.K., Kon, K.V., Eds.; Academic Press: San Diego, CA, USA, 2013; pp. 1–9. [CrossRef] 13. Kuete, V. Chapter 3—Bioactivity of Plant Constituents against Vancomycin-Resistant Enterococci. In Fighting Multidrug Resistance with Herbal Extracts, Essential Oils and Their Components; Rai, M.K., Kon, K.V., Eds.; Academic Press: San Diego, CA, USA, 2013; pp. 23–30. [CrossRef] 14. Ramírez Rueda, R.Y. Chapter 2—Natural Plant Products Used against Methicillin-Resistant Staphylococcus aureus. In Fighting Multidrug Resistance with Herbal Extracts, Essential Oils and Their Components; Rai, M.K., Kon, K.V., Eds.; Academic Press: San Diego, CA, USA, 2013; pp. 11–22. [CrossRef] 15. Barros, C.H.N.; Fulaz, S.; Stanisic, D.; Tasic, L. Biogenic Nanosilver against Multidrug-Resistant Bacteria (MDRB). Antibiotics 2018, 7, 69. [CrossRef] 6. Pal, I.; Bhattacharyya, D.; Kar, R.K.; Zarena, D.; Bhunia, A.; Atreya, H.S. A Peptide-Nanoparticle Sys with Improved Eﬃcacy against Multidrug Resistant Bacteria. Sci. Rep. 2019, 9, 4485. [CrossRef] [PubM 16. Pal, I.; Bhattacharyya, D.; Kar, R.K.; Zarena, D.; Bhunia, A.; Atreya, H.S. A Peptide-Nanoparticle System with Improved Eﬃcacy against Multidrug Resistant Bacteria. Sci. Rep. 2019, 9, 4485. [CrossRef] [PubMed] 17. Hemeg, H.A. Nanomaterials for alternative antibacterial therapy. Int. J. Nanomed. 2017, 12, 8211–8225. [CrossRef] [PubMed] 7. Hemeg, H.A. Nanomaterials for alternative antibacterial therapy. Int. J. Nanomed. 2017, 12, 8211–8 [CrossRef] [PubMed] 18. Patra, J.K.; Das, G.; Fraceto, L.F.; Campos, E.V.R.; Rodriguez-Torres, M.D.P.; Acosta-Torres, L.S.; Diaz-Torres, L.A.; Grillo, R.; Swamy, M.K.; Sharma, S.; et al. Nano based drug delivery systems: Recent developments and future prospects. J. Nanobiotechnol. 2018, 16, 71. [CrossRef] [PubMed] 19. Sur, V.P.; Kominkova, M.; Buchtova, Z.; Dolezelikova, K.; Zitka, O.; Moulick, A. CdSe QD Biosynthesis in Yeast Using Tryptone-Enriched Media and Their Conjugation with a Peptide Hecate for Bacterial Detection and Killing. Nanomaterials 2019, 9, 1463. [CrossRef] 20. Thurn, K.T.; Brown, E.; Wu, A.; Vogt, S.; Lai, B.; Maser, J.; Paunesku, T.; Woloschak, G.E. Nanoparticles for applications in cellular imaging. Nanoscale Res. Lett. 2007, 2, 430–441. [CrossRef] pp g g 21. Bahadar, H.; Maqbool, F.; Niaz, K.; Abdollahi, M. Toxicity of Nanoparticles and an Overview of Current Experimental Models. Iran. Biomed. J. 2016, 20, 1–11. [CrossRef] 22. Crisponi, G.; Nurchi, V.M.; Lachowicz, J.I.; Peana, M.; Medici, S.; Zoroddu, M.A. Chapter 18—Toxicity of Nanoparticles: Etiology and Mechanisms. References 1. Tacconelli, E.; Carrara, E.; Savoldi, A.; Harbarth, S.; Mendelson, M.; Monnet, D.L.; Pulcini, C.; Kahlmeter, G.; Kluytmans, J.; Carmeli, Y.; et al. Discovery, research, and development of new antibiotics: The WHO priority list of antibiotic-resistant bacteria and tuberculosis. Lancet Infect. Dis. 2018, 18, 318–327. [CrossRef] 1. Tacconelli, E.; Carrara, E.; Savoldi, A.; Harbarth, S.; Mendelson, M.; Monnet, D.L.; Pulcini, C.; Kahlmeter, G.; Kluytmans, J.; Carmeli, Y.; et al. Discovery, research, and development of new antibiotics: The WHO priority list of antibiotic-resistant bacteria and tuberculosis. Lancet Infect. Dis. 2018, 18, 318–327. [CrossRef] 2. Foster, T.J. Antibiotic resistance in Staphylococcus aureus. Current status and future prospects. FEMS Microbiol. Rev. 2017, 41, 430–449. [CrossRef] [PubMed] 2. Foster, T.J. Antibiotic resistance in Staphylococcus aureus. Current status and future prospects. FEMS Microbiol. Rev. 2017, 41, 430–449. [CrossRef] [PubMed] 3. Barker, A.K.; Brown, K.; Ahsan, M.; Sengupta, S.; Safdar, N. Social determinants of antibiotic misuse: A qualitative study of community members in Haryana, India. BMC Public Health 2017, 17, 333. [CrossRef] [PubMed] 3. Barker, A.K.; Brown, K.; Ahsan, M.; Sengupta, S.; Safdar, N. Social determinants of antibiotic misuse: A qualitative study of community members in Haryana, India. BMC Public Health 2017, 17, 333. [CrossRef] [PubMed] 4. Antibiotics: Are You Misusing Them? Available online: https://www.mayoclinic.org/healthy-lifestyle/ consumer-health/in-depth/antibiotics/art-20045720 (accessed on 16 April 2020). 4. Antibiotics: Are You Misusing Them? Available online: https://www.mayoclinic.org/healthy-lifestyle/ consumer-health/in-depth/antibiotics/art-20045720 (accessed on 16 April 2020). 15 of 18 15 of 18 Nanomaterials 2020, 10, 778 5. NHS U. Side Eﬀects of Antibiotics. Available online: https://www.nhs.uk/conditions/antibiotics/side-eﬀects/ (accessed on 16 April 2020). 6. Westphal, J.F.; Vetter, D.; Brogard, J.M. Hepatic side-eﬀects of antibiotics. J. Antimicrob. Chemother. 1994, 33, 387–401. [CrossRef] 7. Dang, C.N.; Prasad, Y.D.M.; Boulton, A.J.M.; Jude, E.B. Methicillin-resistant Staphylococcus aureus in the diabetic foot clinic: A worsening problem. Diabet. Med. 2003, 20, 159–161. [CrossRef] 8. Baptista, P.V.; McCusker, M.P.; Carvalho, A.; Ferreira, D.A.; Mohan, N.M.; Martins, M.; Fernandes, A.R. Nano-Strategies to Fight Multidrug Resistant Bacteria-“A Battle of the Titans”. Front. Microbiol. 2018, 9, 1441. [CrossRef] 9. Habiba, K.; Bracho-Rincon, D.P.; Gonzalez-Feliciano, J.A.; Villalobos-Santos, J.C.; Makarov, V.I.; Ortiz, D.; Avalos, J.A.; Gonzalez, C.I.; Weiner, B.R.; Morell, G. Synergistic antibacterial activity of PEGylated silver–graphene quantum dots nanocomposites. Appl. Mater. Today 2015, 1, 80–87. [CrossRef] 10. Alanis, A.J. Resistance to Antibiotics: Are We in the Post-Antibiotic Era? Arch. Med. Res. 2005, 36, 697–705. [CrossRef] 11. Sang, Y.; Blecha, F. Antimicrobial peptides and bacteriocins: Alternatives to traditional antibiotics. Anim. References 2011, 47, 1482–1484. [CrossRef] 35. Han, C.; Lalley, J.; Namboodiri, D.; Cromer, K.; Nadagouda, M.N. Titanium dioxide-based antibacterial surfaces for water treatment. Curr. Opin. Chem. Eng. 2016, 11, 46–51. [CrossRef] 36. Haugen, H.; Lyngstadaas, S. Antibacterial eﬀects of titanium dioxide in wounds. In Wound Healing Biomaterials; Elsevier: Amsterdam, The Netherlands, 2016; pp. 439–450. 37. Kubacka, A.; Diez, M.S.; Rojo, D.; Bargiela, R.; Ciordia, S.; Zapico, I.; Albar, J.P.; Barbas, C.; Dos Santos, V.A.M.; Fernández-García, M. Understanding the antimicrobial mechanism of TiO 2-based nanocomposite ﬁlms in a pathogenic bacterium. Sci. Rep. 2014, 4, 4134. [CrossRef] 38. Besinis, A.; De Peralta, T.; Handy, R.D. The antibacterial eﬀects of silver, titanium dioxide and silica dioxide nanoparticles compared to the dental disinfectant chlorhexidine on Streptococcus mutans using a suite of bioassays. Nanotoxicology 2014, 8, 1–16. [CrossRef] [PubMed] 39. Karthikeyan, G.; Mohanraj, K.; Elango, K.P.; Girishkumar, K. Synthesis, spectroscopic characterization and antibacterial activity of lanthanide–tetracycline complexes. Transit. Met. Chem. 2004, 29, 86–90. [CrossRef] 39. Karthikeyan, G.; Mohanraj, K.; Elango, K.P.; Girishkumar, K. Synthesis, spectroscopic characterization and antibacterial activity of lanthanide–tetracycline complexes. Transit. Met. Chem. 2004, 29, 86–90. [CrossRef] 40. Peng, P.; Jiang, W.-P.; Li, S.-M.; Chen, C.-Z.; Liu, G.-H. Synthesis, characterization and antibacterial activity of l f d li i i did ith thi A l Ch I d 2011 1 100 103 40. Peng, P.; Jiang, W.-P.; Li, S.-M.; Chen, C.-Z.; Liu, G.-H. Synthesis, characterization and antibacterial activity of complex for gadolinium iodide with thiourea. Appl. Chem. Ind. 2011, 1, 100–103. 41. Pradhan, N.; Peng, X. Eﬃcient and Color-Tunable Mn-Doped ZnSe Nanocrystal Emitters: Control of Optical Performance via Greener Synthetic Chemistry. J. Am. Chem. Soc. 2007, 129, 3339–3347. [CrossRef] [PubMed] 42. Moulick, A.; Blazkova, I.; Milosavljevic, V.; Fohlerova, Z.; Hubalek, J.; Kopel, P.; Vaculovicova, M.; Adam, V.; Kizek, R. Application of CdTe/ZnSe Quantum Dots in In Vitro Imaging of Chicken Tissue and Embryo. Photochem. Photobiol. 2015, 91, 417–423. [CrossRef] [PubMed] 43. Jelinkova, P.; Koudelkova, Z.; Milosavljevic, V.; Horky, P.; Kopel, P.; Adam, V. Utilization of Selenium Nanoparticles with SchiﬀBase Chitosan as Antibacterial Agents. MendelNet 2016, 989–993. Available online: https://pdfs.semanticscholar.org/6ba5/800ec37f37ab26508fefcafd3ac3056f3201.pdf (accessed on 16 April 2020). 44. Herathge, N.D.S.; George, J.T.; Rowley, D.A. Diﬀerential antimicrobial activities of Human Beta-Defensins against Methicillin Resistant (MRSA) and Methicillin sensitive (MSSA) Staphylococcus aureus. In Science and Technology Against Microbial Pathogens; World Scientiﬁc: Singapore, 2011; pp. 9–12. [CrossRef] 45. References In Antimicrobial Nanoarchitectonics; Grumezescu, A.M., Ed.; Elsevier: Amsterdam, The Netherlands, 2017; pp. 511–546. [CrossRef] 23. Stocks, S.M. Mechanism and use of the commercially available viability stain, BacLight. Cytom. Part A 2004, 61, 189–195. [CrossRef] 24. Datta, S.; Sherman, J.M.; Bravard, M.A.; Valencia, T.; Gilman, R.H.; Evans, C.A. Clinical Evaluation of Tuberculosis Viability Microscopy for Assessing Treatment Response. Clin. Infect. Dis. 2014, 60, 1186–1195. [CrossRef] 25. Kanade, S.; Nataraj, G.; Ubale, M.; Mehta, P. Fluorescein diacetate vital staining for detecting viability of acid-fast bacilli in patients on antituberculosis treatment. Int. J. Mycobacteriol. 2016, 5, 294–298. [CrossRef] 16 of 18 16 of 18 Nanomaterials 2020, 10, 778 26. Azzopardi, E.A.; Ferguson, E.L.; Thomas, D.W. The enhanced permeability retention eﬀect: A new paradigm for drug targeting in infection. J. Antimicrob. Chemother. 2012, 68, 257–274. [CrossRef] 27. Anderson, D.; Mj, S. Nanotechnology: The Risks and Beneﬁts for Medical Diagnosis and Treatment. J. Nanomed. Nanotechnol. 2016, 7. [CrossRef] 28. Moulick, A.; Heger, Z.; Milosavljevic, V.; Richtera, L.; Barroso-Flores, J.; Merlos Rodrigo, M.A.; Buchtelova, H.; Podgajny, R.; Hynek, D.; Kopel, P.; et al. Real-Time Visualization of Cell Membrane Damage Using Gadolinium–SchiﬀBase Complex-Doped Quantum Dots. ACS Appl. Mater. Interfaces 2018, 10, 35859–35868. [CrossRef] [PubMed] 29. Ristic, B.Z.; Milenkovic, M.M.; Dakic, I.R.; Todorovic-Markovic, B.M.; Milosavljevic, M.S.; Budimir, M.D.; Paunovic, V.G.; Dramicanin, M.D.; Markovic, Z.M.; Trajkovic, V.S. Photodynamic antibacterial eﬀect of graphene quantum dots. Biomaterials 2014, 35, 4428–4435. [CrossRef] [PubMed] 30. Li, H.; Huang, J.; Song, Y.; Zhang, M.; Wang, H.; Lu, F.; Huang, H.; Liu, Y.; Dai, X.; Gu, Z.; et al. Degradable Carbon Dots with Broad-Spectrum Antibacterial Activity. ACS Appl. Mater. Interfaces 2018, 10, 26936–26946. [CrossRef] [PubMed] 31. Garcia, I.M.; Leitune, V.C.B.; Visioli, F.; Samuel, S.M.W.; Collares, F.M. Inﬂuence of zinc oxide quantum dots in the antibacterial activity and cytotoxicity of an experimental adhesive resin. J. Dent. 2018, 73, 57–60. [CrossRef] 32. Meikhail, M.S.; Abdelghany, A.M.; Awad, W.M. Role of CdSe quantum dots in the structure and antibacterial activity of chitosan/poly ϵ-caprolactone thin ﬁlms. Egypt. J. Basic Appl. Sci. 2018, 5, 138–144. [CrossRef] 33. Belletti, D.; Riva, G.; Luppi, M.; Tosi, G.; Forni, F.; Vandelli, M.A.; Ruozi, B.; Pederzoli, F. Anticancer drug-loaded quantum dots engineered polymeric nanoparticles: Diagnosis/therapy combined approach. Eur. J. Pharm. Sci. 2017, 107, 230–239. [CrossRef] 34. Liu, Z.; Lin, Q.; Huang, Q.; Liu, H.; Bao, C.; Zhang, W.; Zhong, X.; Zhu, L. Semiconductor quantum dots photosensitizing release of anticancer drug. Chem. Commun. References Mazumdar, A.; Haddad, Y.; Milosavljevic, V.; Michalkova, H.; Guran, R.; Bhowmick, S.; Moulick, A. Peptide-Carbon Quantum Dots conjugate, Derived from Human Retinoic Acid Receptor Responder Protein 2, against Antibiotic-Resistant Gram Positive and Gram Negative Pathogenic Bacteria. Nanomaterials 2020, 10, 325. [CrossRef] 17 of 18 17 of 18 Nanomaterials 2020, 10, 778 46. Jelinkova, P.; Splichal, Z.; Jimenez, A.M.J.; Haddad, Y.; Mazumdar, A.; Sur, V.P.; Milosavljevic, V.; Kopel, P.; Buchtelova, H.; Guran, R.; et al. Novel vancomycin-peptide conjugate as potent antibacterial agent against vancomycin-resistant Staphylococcus aureus. Infect. Drug Resist. 2018, 11, 1807–1817. [CrossRef] 47. Chudobova, D.; Cihalova, K.; Dostalova, S.; Ruttkay-Nedecky, B.; Merlos Rodrigo, M.A.; Tmejova, K.; Kopel, P.; Nejdl, L.; Kudr, J.; Gumulec, J.; et al. Comparison of the eﬀects of silver phosphate and selenium nanoparticles on Staphylococcus aureus growth reveals potential for selenium particles to prevent infection. FEMS Microbiol. Lett. 2014, 351, 195–201. [CrossRef] 48. Richter, S.G.; Elli, D.; Kim, H.K.; Hendrickx, A.P.A.; Sorg, J.A.; Schneewind, O.; Missiakas, D. Small molecule inhibitor of lipoteichoic acid synthesis is an antibiotic for Gram-positive bacteria. Proc. Natl. Acad. Sci. USA 2013, 110, 3531–3536. [CrossRef] 49. Vukomanovic, M.; Torrents, E. High time resolution and high signal-to-noise monitoring of the bacterial growth kinetics in the presence of plasmonic nanoparticles. J. Nanobiotechnol. 2019, 17, 21. [CrossRef] 50. Stevenson, K.; McVey, A.F.; Clark, I.B.N.; Swain, P.S.; Pilizota, T. General calibration of microbial growth in microplate readers. Sci. Rep. 2016, 6, 38828. [CrossRef] [PubMed] 51. van Sorge, N.M.; Beasley, F.C.; Gusarov, I.; Gonzalez, D.J.; von Köckritz-Blickwede, M.; Anik, S.; Borkowski, A.W.; Dorrestein, P.C.; Nudler, E.; Nizet, V. Methicillin-resistant Staphylococcus aureus Bacterial Nitric-oxide Synthase Aﬀects Antibiotic Sensitivity and Skin Abscess Development. J. Biol. Chem. 2013, 288, 6417–6426. [CrossRef] [PubMed] 52. Schumacher, A.; Vranken, T.; Malhotra, A.; Arts, J.J.C.; Habibovic, P. In vitro antimicrobial susceptibility testing methods: Agar dilution to 3D tissue-engineered models. Eur. J. Clin. Microbiol. Infect. Dis. 2018, 37, 187–208. [CrossRef] [PubMed] 53. Stasiak-Ró˙za´nska, L.; Bła˙zejak, S.; Gientka, I. Eﬀect of glycerol and dihydroxyacetone concentrations in the culture medium on the growth of acetic acid bacteria Gluconobacter oxydans ATCC 621. Eur. Food Res. Technol. 2014, 239, 453–461. [CrossRef] 54. Hogenkamp, A.; Herías, M.V.; Tooten, P.C.J.; Veldhuizen, E.J.A.; Haagsman, H.P. Eﬀects of surfactant protein D on growth, adhesion and epithelial invasion of intestinal Gram-negative bacteria. Mol. Immunol. 2007, 44, 3517–3527. [CrossRef] [PubMed] 5. Li, R.C.; Nix, D.E.; Schentag, J.J. References New turbidimetric assay for quantitation of viable bacterial dens Antimicrob. Agents Chemother. 1993, 37, 371–374. [CrossRef] 56. Sacar, M.; Sacar, S.; Cevahir, N.; Onem, G.; Teke, Z.; Asan, A.; Turgut, H.; Adali, F.; Kaleli, I.; Susam, I.; et al. Comparison of antimicrobial agents as therapy for experimental endocarditis: Caused by methicillin-resistant Staphylococcus aureus. Tex Heart Inst. J. 2010, 37, 400–404. 57. Popova, M.; Martin, C.; Morgavi, D.P. Improved protocol for high-quality Co-extraction of DNA and RNA from rumen digesta. Folia Microbiol. 2010, 55, 368–372. [CrossRef] 58. Welsh, S.; Peakman, T.; Sheard, S.; Almond, R. Comparison of DNA quantiﬁcation methodology used in the DNA extraction protocol for the UK Biobank cohort. BMC Genom. 2017, 18, 26. [CrossRef] 59. Quantiﬁcation of DNA and RNA: A Spectrophotometric Method. In eLS; Wiley: New York, NY, USA, 2003. [CrossRef] 60. Jorgez, C.J.; Dang, D.D.; Simpson, J.L.; Lewis, D.E.; Bischoﬀ, F.Z. Quantity versus quality: Optimal methods for cell-free DNA isolation from plasma of pregnant women. Genet. Med. 2006, 8, 615–619. [CrossRef] 61. Wilﬁnger, W.W.; Mackey, K.; Chomczynski, P. Eﬀect of pH and Ionic Strength on the Spectrophotometric Assessment of Nucleic Acid Purity BioTechniques 1997 22 474 481 [CrossRef] 60. Jorgez, C.J.; Dang, D.D.; Simpson, J.L.; Lewis, D.E.; Bischoﬀ, F.Z. Quantity versus quality: Optimal methods for cell-free DNA isolation from plasma of pregnant women. Genet. Med. 2006, 8, 615–619. [CrossRef] 1. Wilﬁnger, W.W.; Mackey, K.; Chomczynski, P. Eﬀect of pH and Ionic Strength on the Spectrophotome Assessment of Nucleic Acid Purity. BioTechniques 1997, 22, 474–481. [CrossRef] 62. Heger, Z.; Merlos Rodrigo, M.A.; Michalek, P.; Polanska, H.; Masarik, M.; Vit, V.; Plevova, M.; Pacik, D.; Eckschlager, T.; Stiborova, M.; et al. Sarcosine Up-Regulates Expression of Genes Involved in Cell Cycle Progression of Metastatic Models of Prostate Cancer. PLoS ONE 2016, 11, e0165830. [CrossRef] 63. Trivedi, M.K.; Branton, A.; Trivedi, D.; Nayak, G.; Bairwa, K.; Jana, S. Spectroscopic Characterization of Disodium Hydrogen Orthophosphate and Sodium Nitrate after Bioﬁeld Treatment. J. Chromatogr. Sep. Tech. 2015, 6, 5. [CrossRef] 64. Perrin, F.X.; Nguyen, V.; Vernet, J.L. FT-IR Spectroscopy of Acid-Modiﬁed Titanium Alkoxides: Investigations on the Nature of Carboxylate Coordination and Degree of Complexation. J. Sol-Gel Sci. Technol. 2003, 28, 205–215. [CrossRef] 18 of 18 Nanomaterials 2020, 10, 778 18 of 18 65. Sondi, I.; Salopek-Sondi, B. Silver nanoparticles as antimicrobial agent: A case study on E. coli as a model for Gram-negative bacteria. J. Colloid Interface Sci. 2004, 275, 177–182. References [CrossRef] [PubMed] 66. Diana, P.; Sigifredo, O.; Jaime, R.; María Claudia, V.; Julieta, P.; Orville, H.; Jinnethe, R.; César, A.A. First Characterization of a Cluster of VanA-Type Glycopeptide-Resistant Enterococcus faecium Colombia. Emerg. Infect. Dis. J. 2002, 8, 961. [CrossRef] 67. Qureshi, N.K.; Yin, S.; Boyle-Vavra, S. The role of the Staphylococcal VraTSR regulatory system on vancomycin resistance and vanA operon expression in vancomycin-resistant Staphylococcus aureus. PLoS ONE 2014, 9, e85873. [CrossRef] 68. Liu, B.; Cheng, W.; Rotenberg, S.A.; Mirkin, M.V. Scanning electrochemical microscopy of living cells: Part 2. Imaging redox and acid/basic reactivities. J. Electroanal. Chem. 2001, 500, 590–597. [CrossRef] g g 69. Bard, A.J.; Mirkin, M.V. Scanning Electrochemical Microscopy; CRC Press: Florida, FL, USA, 2012. 70. Kim, S.R.; Park, M.J.; Lee, M.K.; Sung, S.H.; Park, E.J.; Kim, J.; Kim, S.Y.; Oh, T.H.; Markelonis, G.J.; Kim, Y.C. Flavonoids of Inula britannica protect cultured cortical cells from necrotic cell death induced by glutamate. Free Radic. Biol. Med. 2002, 32, 596–604. [CrossRef] 71. Shoemaker, M.; Cohen, I.; Campbell, M. Reduction of MTT by aqueous herbal extracts in the absence of cells. J. Ethnopharmacol. 2004, 93, 381–384. [CrossRef] [PubMed] © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/). © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
https://openalex.org/W3134924666	https://www.researchsquare.com/article/rs-508043/latest.pdf	English	null	Socio-ecological Systems Analysis of the Prevention and Control of Dengue in two Districts of Dar es Salaam City, Tanzania	Research Square (Research Square)	2,021	cc-by	7,844	Socio-ecological Systems Analysis of the Prevention and Control of Dengue in two Districts of Dar es Salaam City, Tanzania Dar es Salaam City, Ta Leonard E.G. Mboera (  lmboera@gmail.c Sokoine University of Agriculture https:// Calvin Sindato National Institute for Medical Research Irene R. Mremi National Institute for Medical Research Susan F. Rumisha National Institute for Medical Research Janeth George SACIDS Foundation for One Health Renfrid Ngolongolo SACIDS Foundation for One Health Gerald Misinzo Sokoine University of Agriculture Esron D. Karimuribo Sokoine University of Agriculture Mark M. Rweyemamu SACIDS Foundation for One Health Najmul Haider The Royal Veterinary College Muzamil Abdel Hamid University of Khartoum Richard Kock Leonard E.G. Mboera (  lmboera@gmail.com ) Sokoine University of Agriculture https://orcid.org/0000-0001-5746-3776 Calvin Sindato National Institute for Medical Research Irene R. Mremi National Institute for Medical Research Susan F. Rumisha National Institute for Medical Research Susan F. Rumisha National Institute for Medical Research Janeth George SACIDS Foundation for One Health Renfrid Ngolongolo SACIDS Foundation for One Health Gerald Misinzo Sokoine University of Agriculture License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Research Article Page 1/21 Keywords: Dengue, socio-ecological systems, readiness, surveillance, Tanzania Posted Date: June 2nd, 2021 DOI: https://doi.org/10.21203/rs.3.rs-508043/v1 Page 1/21 Page 1/21 License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License License:   This work is licensed under a Creative Commons Attribution 4.0 International License. Read Full License Page 2/21 Abstract Background: Since 2010, Tanzania has been experiencing frequent outbreaks of dengue. The objective of this study was to carry out a socio-ecological systems analysis and assess the readiness in dengue prevention and control in Kinondoni and Ilala districts of Dar es Salaam City, Tanzania. Methods: Twenty-seven key district officials responsible for human and animal health were involved in a socio-ecological systems framework analysis as regards to dengue. In addition, the readiness of the districts to respond to dengue outbreaks and the performance of the disease surveillance system was assessed. Results: The two districts were characterized by both urban and peri-urban ecosystems, with a mixture of planned and unplanned settlements which support breeding and proliferation of Aedes mosquitoes. The results indicate inadequate levels of readiness in the management and control of dengue outbreaks, in terms of clinical competence, diagnostic capacities, surveillance system and control/prevention measures. Mosquito breeding sites, especially discarded automobile tyres, were reported to be scattered in the districts. Constraining factors in implementing disease surveillance included both intrapersonal and interpersonal factors, lack of case management guidelines, difficult language used in standard case definitions, inadequate laboratory capacity, lack of appropriate rapid response teams, inadequate knowledge on outbreak investigation and inadequate capacities in data management. Conclusion: The two districts had limited readiness in the management and control of dengue, in terms of clinical competence, diagnostic capacities, surveillance system and prevention and control measures. These challenges require the immediate attention by the authorities, as they compromise the effectiveness of the national strategy for community health support. Introduction Tanzania has experienced several outbreaks of mosquito-borne viral diseases (MBVDs) including dengue, chikungunya and Rift Valley fever over the last two decades [1–4]. These diseases are ecologically sensitive, since environmental conditions, such as temperature and precipitation, affect both the infectious pathogens and the vectors that transmit them [5–6]. MBVDs have received little attention, most likely due to inadequate knowledge, awareness and inadequate diagnostic capacity. The ecology and transmission of MBVDs are closely related to environmental resource management, social and behavioural patterns [7]. Thus, the mechanisms underlying MBVDs require exploring complex interactions within and among coupled natural and human systems, considering human and natural ecologies, and the multiple layers of interaction among them [8–10]. Usually, MBVD risk increases as a result of the effect of heavy rainfall providing habitat suitability for mosquito breeding opportunities [5, 11]. Furthermore, repeated drought conditions and inadequate/unreliable water supply encourage storage of water for human and livestock consumption, the practices which provide the breeding sites for Aedes aegypti, the primary vector of dengue [12]. Furthermore, repeated drought conditions and inadequate/unreliable water supply encourage storage of water for human and livestock consumption, the practices which provide the breeding sites for Aedes aegypti, the primary vector of dengue [12]. Page 3/21 Page 3/21 A better understanding of the relationship between development transitions and emerging and re- emerging infectious diseases is critical for improving capability to predict and respond to epidemics. The complex nature of emerging and re-emerging infectious diseases highlights the need for new approaches to understand the mechanisms and processes underlying pathogen emergence to improve disease prevention and control. However, there are limited studies that have attempted to apply socio-ecological systems (SES) framework analysis on emerging and re-emerging infectious diseases in low-and middle- income countries [13]. It is important to understand how transformations in socio-ecological systems caused by multifaceted interactions between anthropogenic environmental changes such as urbanization, livelihoods activities, and natural habitat alterations, produce feedbacks that affect natural communities and ultimately their pathogens, animal and human populations [14–15]. The socio-ecological theory provides a framework for examining how individuals, their health, and their surrounding physical and social environments interact at multiple levels of a health problem and their interdependency [5, 16]. Introduction The levels are: (i) Intrapersonal factors such as individual attitudes, behaviours, knowledge, and skills; (ii) Interpersonal processes and community factors, such as social networks made of family, friends, or colleagues that provide support; (iii) Institutional factors such as formal or informal organizations that may have rules or expectations that impact health behaviours; (iv) laws, policies or regulations that promote or inhibit certain health practices that impact disease management, control or prevention [17–20]. The development of a framework for analysing social-ecological models aims to understand how these systems function, and the processes through which they interact to influence disease outbreaks. To our knowledge, only a single study carried out in Tanzania has so far documented socio-ecological factors as regards to dengue epidemics [21]. Moreover, the readiness of national and sub-national levels in dengue and other mosquito-borne viral disease mitigations has not been assessed in Tanzania. Identifying community and district levels health determinant factors for dengue transmission will provide specific targets for public health interventions. The objectives of this study were to (i) carry out a socio-ecological systems analysis on dengue to identify risk factors and interventions at the district level; and (ii) assess the readiness of the district in the prevention and control of dengue. Study area This study was carried out during October 2019 and involved Ilala and Kinondoni districts of Dar es Salaam city in eastern Tanzania. The city with an area of 1,339km2, is made up of five administrative districts, namely Ilala, Kigamboni, Kinondoni, Temeke and Ubungo. According to the 2012 national census, the city had a population of 4,364,541 (Ilala = 1,220,611; Kinondoni = 1,775,049), with an annual growth rate of 5.6% [22]. The climate of Dar es Salaam is generally hot and humid with small seasonal and daily variations in temperature. The mean daily temperature is 26oC, the mean seasonal range is 4oC, and the mean daily range is about 8oC. There are two rainy seasons, long (March-June) and short Page 4/21 Page 4/21 (October-December) with total annual rainfall averages 1,100mm. The relative humidity is generally high, reaching 100% almost every night throughout the year, but falling to 60% during the day [23–24]. (October-December) with total annual rainfall averages 1,100mm. The relative humidity is generally high, reaching 100% almost every night throughout the year, but falling to 60% during the day [23–24]. Dar es Salaam has been facing continued dengue outbreaks with increased frequency and occurrence during the past 10 years. The occurrence of dengue outbreaks was reported in 2010, 2012, 2013, 2014, 2018 and 2019 [2, 3, 25–26]. During the 2019 outbreak, about 6,917 cases including 13 deaths were reported between January and July 2019 in Tanzania [27]. Two districts of Dar es Salaam, namely Kinondoni and Ilala were selected for this study because of their experience in dengue outbreaks since 2010 [2–3]. We hypothesized that following experience of several dengue outbreaks, the districts would have plans and built their capacities on outbreak preparedness and response. Data collection Key officials of the departments responsible for human and animal health of the Dar es Salaam city and Ilala and Kinondoni districts were invited to a SES framework analysis workshop. The workshop comprised of series of presentations, group exercises and plenary sessions. Presentations included general overview of MBVDs in Tanzania, SES framework analysis and principles and functions of public and animal health surveillance systems. The group working sessions focused on specific assignments on SES framework analysis on dengue. A mapping exercise was also carried out to analyse the performance of the disease surveillance system at district and facility levels. Socio-ecological system analysis The workshop brought together 27 participants comprising of regional and district medical/ health officers, regional and district health management information systems officers, veterinary officers, information and communication technology specialists, surveillance officers and researchers. A number of observations and issues were raised as regards to the capacity of the country in addressing MBVDs. It was explained that until early 2000s, the Ministry of Health had a department responsible for the coordination of all issues pertaining to vector-borne diseases. However, the department does not exist anymore; and the activities related to vector-borne diseases were being implemented separately through vertical programmes. Such programmes focused on specific diseases such as malaria, lymphatic filariasis, trachoma and onchocerciasis. The two districts were characterized by both urban and peri-urban ecosystems, with mostly medium to high density human populations. A mixture of planned and unplanned settlements was common in both districts. Backyard poultry, pig and dairy farming was common in Ilala, but limited in Kinondoni district. Housing was characterized by iron roofing, with some of the houses having roof rain gutters. The main economic activities in Kinondoni included small-scale industries, business and extractive quarrying mining activities. In Ilala they included medium to large-scale food and non-food industries, business, fishing activities, crop production and extractive quarrying of sand. Seasonal cropping of vegetables and pot flower gardening were common practices in the two districts (Fig. 1). The sources of energy in both districts included electricity, charcoal, firewood, solar, gas and kerosene. Though solid waste management and disposal was centralized, improper household waste disposal was common in the two districts. District SES framework analysis This exploratory study utilized a cross-sectional purposive selection of key stakeholders responsible for disease surveillance and response in human and animal populations. The stakeholders were brought together to foster interdisciplinary collaboration and stimulate transdisciplinary approaches to improve prevention and control of MBVDs. We applied SES framework [28–31] to identify drivers and construct perceived thematic causal explanations of the dengue outbreaks in the study districts. The SES approach allows the application of a variety of methodological process to study the field of interest [32]. Building on a conceptual framework developed by ecologists to address natural resource management issues [28–31], we adapted the steps to the interface of human and natural systems to assess the capacity of the districts in addressing MBVDs. Participants worked in two groups, each for the respective district. Group works were followed by group presentations and discussions. District readiness to epidemics: A semi-structured questionnaire was used to assess and document the districts’ clinical competence of healthcare providers, diagnostic capacity, surveillance and control interventions. The assessment of the surveillance system covered the following items: structures of the surveillance system, staff and resources, responsibility units and capacity to respond to infectious disease outbreaks. Disease surveillance, response and control programmes and networks were assessed to evaluate the capacity to detect, notify, assess and respond to known, new, and unknown infectious disease threats. The gaps were identified and solutions to enhance utilization of data for relevant sources Page 5/21 for decision making process were proposed. Exploration was made to map how the districts were working and what actions would be needed to improve surveillance and response strategies in the districts. Data management and analysis We used data from the in-depth interviews to assess levels, practices and factors influencing socio- ecological system, epidemic readiness, disease surveillance capacity at health facility level and district level, and dengue prevention and control. Content analysis technique was applied using manifest surface structure approach [33] to group different expressions provided by participants into discrete categorical responses reflecting the specific themes defined by different variables in the SES framework analysis tool used. Data was summarized using a narrative structure for SES and were interpreted based on existing related literature, theories and knowledge/experience of the researchers. Epidemic readiness Overall, none of the two districts reported adequate readiness in the management and control of dengue, both in terms of clinical competence, diagnostic capacities, surveillance and control/prevention Page 6/21 measures. The challenges related to clinical competence included lack of standard case definition (SCD), standard operating procedures, inadequate rapid diagnostic tests, lack of case management guidelines, inadequate human resource and lack of job aids. The major challenges in surveillance were poor data quality (completeness and consistency), timeliness, inadequate data analysis capacity and unavailability of data sharing policy and practices. Community-based disease surveillance programme was not part of the surveillance system in the two districts. None of the districts had a contingency plan for dengue. Treatment guidelines for dengue were available, but not disseminated widely at all levels in the districts (Table 1). Page 7/21 Table 1 District readiness in the management and control of dengue Variable Ilala Kinondoni Clinical competence of healthcare providers • Standard case definition Not available Available but not used • Case management guidelines Not available Not available • Human resource availability Inadequate Inadequate Diagnostic capacity • Availability of rapid diagnostic test Inadequate Inadequate • Availability of other related diagnostic test (FBP, ESR, Platelets) Inadequate Inadequate • Availability of guidelines/standard operation procedures/job aids Not available Not available Surveillance system in place and functional • Data quality (completeness, consistency) • Data analytical capacity (pattern, demographic, hot spots identified from surveillance data) • Data sharing policy and practices Some data quality Inadequate capacity Not in place Poor Weak Not in place Control interventions available • Contingency plan available None None • Community involvement None None • Health education and promotion Limited Limited • Source reduction programme None None • Vector control Larviciding as malaria vector intervention Larviciding as malaria vector intervention • Insecticide-treated mosquito net High coverage High coverage Partners collaborators Not available Not available Key: FBP= full blood picture; ESR= erythrocyte sedimentation rate Disease surveillance Facility level. At the facility level (which include district hospital, health centre and dispensary), participants from Kinondoni identified availability of forms for reporting and routine meetings with local Page 8/21 authorities as the only facilitating factors in implementing surveillance programme. The constraining factors reported included lack of guidelines, dissemination and translation of SCDs (which were in English) and low diagnostic capacities for dengue. Lack of smooth communication network and appropriate rapid response team were reported. Epidemic readiness Low capacity on data management, including analysis, was reported to negatively affect the decision-making process at health facility level. Inadequate knowledge on outbreak investigation among health facility workers and shortage of sample collection equipment were identified as critical gaps. In Ilala, although the SCD were available, most of the clinician reported to have not received orientation on their use. Data analysis and interpretation at health facility level was not been done properly due to inadequate knowledge and skills. Like in Kinondoni, the participants reported that facility level health care workers were not part of the outbreak investigation team. District level In Kinondoni the identified constraining factors were (i) lack of data collection tools, shortage of staff and insufficient laboratory reagents; (ii) external and internal transfer of trained staff; insufficient knowledge and unreliable internet connectivity; (iii) inadequate communication tools; (iv) delays in receiving reports from the facilities or community levels; (vi) lack of periodic supportive supervision from higher levels; (vi) low capacity of the district outbreak response team; and (vii) inadequate budget for surveillance activities. In Ilala, participants reported on lack of feedback on data analysis as the most critical challenge in the use of surveillance data. They were not able to compare current and previous pattern (trends) of the disease due to poor quality data. Unavailability of sample collection tools and unclear responsibilities for health care workers were underscored as the key limiting factors in the outbreak investigation. It was pointed out that although the district took part in disease outbreak response, its staff lacked training on emergency preparedness and responses due to inadequate budget. They also reported that an Epidemic Response Committee in the district was not functioning. Most of the guidelines were reported to be in English, voluminous and not handy. The guidelines were perceived by most participants to be difficult to follow. It was proposed that the guidelines be translated into Kiswahili (the national language) and made available in simple user-friendly formats. In the two districts, analysis tools for surveillance data were not in place. Moreover, it was reported that the surveillance officers were not skilled enough on data management especially, analysis and interpretation. There were a number of health information systems for district and facility to use. They included Health Management Information System (HMIS), National System for Government Health Officials, Government of Tanzania Hospital Management Information System and others. All these were described to bring confusion to health workers. There was no information, knowledge, skills and experience sharing practices on data management between surveillance and HMIS officers at the district level. Page 9/21 Dengue prevention and control During the discussion, an issue on unchecked mosquito breeding sites in Dar es Salaam city was raised and discussed. Mosquito breeding sites especially discarded automobile tyres were reported to be scattered and found in almost all areas of the city. The automobile tyres were described to contribute to over two-thirds of the breeding sites for Aedes mosquitoes. Participants were concerned with the lack of appropriate actions as regards to disposal of discarded automobile tyres and water-carrying containers. It was reported that in the old days, the destruction of used automobile tyres was done by burning. However, the programme failed, as it required burning chambers, which were not available. It was noted that although a malaria larviciding programme was been implemented in the city, it was not appropriately executed; and was perceived to be ineffective against Aedes mosquitoes. Participants noted that the malaria mosquito larviciding project focused on large water bodies and left small water bodies and containers unchecked. Moreover, the programme was being implemented in only 60% of the wards. The participants raised their concerns for the districts inadequate budget for outbreaks response. The low level of awareness on dengue and Rift Valley fever was reported for both districts (Fig. 2). The following were proposed as the way forward in addressing the challenges in the control of dengue: (i) revising the larviciding programme to cover much larger areas of the city and target the priority mosquito- borne disease vectors; (ii) enhancing capacity in dengue diagnosis and case management; and (iii) developing a strategic plan to address advocacy to the regional and district government authorities as regards to the importance of outbreak preparedness. Discussion Dengue has become an increasingly important public health problem in the tropical world in terms of morbidity and mortality. With rapidly changing human population and activities, the ecology of both the pathogens and vectors, climate variability/change coupled with weak health systems, dengue outbreaks are likely to become more frequent. In our approach, we addressed the reality that effective prevention and control of dengue cannot be achieved from narrow perspectives, but would largely benefit from a holistic approach that considers an improved understanding of SES and health system readiness. The basic idea of SES is to be explicit in system thinking and linking together the ‘human system’ and the ‘natural system’ in a two-way feedback relationship. The interactions between humans, pathogens and environment, suggest that SES is crucial in disease occurrence and transmission dynamics. Moreover, the readiness of national and sub-national levels in dengue mitigations is central in the designing and planning of appropriate intervention measures. It is against this background that we conducted this analysis to identify socioecological factors and readiness of the districts in the control and prevention of dengue. In our data collection process, we adopted a consultative workshop method, which has been found to foster engagement through collaborative discussions and constructive feedback between participants from different organizations and facilitators on a particular topic [34–35]. The approach helped us to obtain information-rich data and Page 10/21 Page 10/21 persistent observation on dengue management and control in the study districts from purposively selected key stakeholders, which is similar to a consideration reported elsewhere [36]. In the SES framework analysis, we found that the ecological descriptors of Ilala and Kinondoni were characterised by both urban and peri-urban ecosystems with planned and unplanned settlements. Although in our study we did not establish the impact of urbanization on the occurrence of dengue, urban and peri-urban areas have been reported to be vulnerable to the disease, the outcome which is driven by rapid and uncontrolled urbanization that support Aedes mosquito productivity [2, 37–38]. The intrapersonal factors such as in-house water storage, improper household waste water disposal were common practices in the districts. The presence of potential mosquito breeding sites including discarded automobile car tyres, flower pots and water stagnation were reported to support habitat suitability for mosquito breeding and thereby increasing the risk of occurrence of dengue in the two districts [2, 21]. Discussion It was reported that to date, there have been no effective measures in the disposal of used automobile tyres. As a result, they were improperly discarded and continued to stockpile in the districts. Tyre stockpiles create a great health risk as they provide permissive breeding grounds for mosquitoes [39]. Used automobile tyres and domestic water storage containers have been reported as the most important larval habitats for Ae. aegypti in Dar es Salaam and Zanzibar [40–41]. These observations suggest that the solution to proper management of mosquito breeding sites does not fall clearly within the boundaries of a single discipline or sector but rather a multi-sectoral approach. Vulnerability to dengue was considered disproportionately distributed within each of the study districts. This suggests that interventions to address them should be repackaged based on specific-area risk profiles. However, as we have highlighted on the inefficiency of surveillance systems based on district- level SES framework analysis, we cannot account explicitly on the representation of vulnerability of sub- district levels to dengue. SES framework analysis at the sub-district levels is likely to complement our understanding on the vulnerability profiles at different ecological levels. One of the strategic mosquito prevention measures in the districts was larviciding. However, in the two districts, larviciding, which was introduced for malaria control [42], was implemented without consideration of the evidence to guide appropriate timing, areas and potential breeding sites of focus. We suggest that in order to effectively prevent dengue, it is important to implement the integrated vector management measures. This should be implemented by targeting the most suitable mosquito habitats to suppress mosquito larval stages thereby reducing adult emergence. Although larviciding has been recommended as a complementary intervention to control malaria [43], we suggest to conduct monitoring and evaluation of such interventions against the Aedes mosquitoes. It has been established that indices of mosquito and climate factors are the main determinants of dengue [44–46]. Since microclimate dynamics is largely determined by human activities it is imperative to suggest that a key strategy to prevent and control dengue is to focus on addressing the anthropogenic factors. A number of ecological, biological and social factors are involved in mosquito breeding and pathogen transmission [47]. A study in Sri Lanka revealed that dengue outbreaks were significantly Page 11/21 Page 11/21 associated with ecological, socio-economic and demographic factors; including the presence of built-up area, and areas with higher human population density [48]. Discussion Existing evidence on the complexity of eco- bio-social contexts repeatedly argues that dengue control necessitates sound intersectoral approaches that combine environmental management practices with community mobilization [49–51]. Probably intensification of public health education on proper management of water storage containers, discarded car tyres, periodic draining or removal of artificial containers could be the most effective strategy of reducing mosquito breeding habitats. Based on SES framework analysis results, we hypothesise that interventions on the management of potential mosquito breeding sites in domestic and peri-domestic environments and ecological health could generate data to guide sustainable community-based prevention strategies against dengue and other Aedes-borne viral diseases. On assessing the capacity in responding to outbreaks, none of the two districts reported the readiness in the management of dengue outbreak. The availability of disease surveillance and management guidelines in English language was reported as a utilization impediment and participants proposed that they should be translated into Kiswahili. In our study, the guidelines were reported to be voluminous and not user-friendly. Since a number of guidelines continue to proliferate on the same or similar subjects, it is likely impractical for the practitioners to use them efficiently. The successful implementation of guidelines depends on many factors including participatory development, dissemination and implementation strategy as well as evaluation of their effectiveness [52]. The presence of guidelines may not guarantee their implementation or utility, and some studies have reported failure of guidelines to influence the implementation of health programmes [53–55]. The reasons that have been associated with failure of guidelines to achieve their objectives, include inadequate consultation, lack of consideration of technical capacity, attitude and behaviour of health professionals and lack of training on the use of the guidelines [56–58]. Adoption of a web- or mobile-based electronic platform to enhance, on the one hand data collection especially at community level and on the other, prompt searching by topic of interest or based on prevailing needs is likely to be a breakthrough. We believe that with increasing occurrence of infectious disease epidemics the efficiency of disease surveillance, early detection and early warning will require use of such web or mobile technologies [59]. This can be enhanced through application of machine learning data-mining integration with socio-ecological and geospatial analyses [60], specifically designed for low- and middle- income countries like Tanzania. Discussion The capacities of the districts in dengue epidemic readiness was limited, and this calls for concerted efforts to improve resource availability, and training. Measuring readiness in epidemic preparedness and response is likely to present some challenges. This is because in this study epidemic preparedness focused mainly on human resource, infrastructure, and surveillance. The study did not document the relevant processes in epidemic preparedness or the activities executed during an epidemic response, including isolation and quarantine, public communication, and others [61]. Generally, the overall performance of the surveillance system in the two districts was not satisfactory in detecting and or monitoring trends of dengue. We recorded inadequate skills and capacity in outbreak investigation, insufficient laboratory supplies, poor surveillance data quality, inadequate data Page 12/21 management including data analysis, interpretation and use, and poor information/data sharing practices between sectors. The surveillance systems were reported to be inefficient with inadequate involvement of community and therefore not tailored for early detection and response. Limited or no evidence of routine data analysis at sub-national levels mainly due to lack of clear guidelines on how and when to analyse data has been reported in previous studies [62–65]. It has been described that when the health facilities do not analyse and use the data, the utility of the surveillance system becomes minimal, which makes the system too weak to pick outbreaks early that could guide prompt response. The absence of data analysis, interpretation and utilization for local action seen in the present study is in line with findings of studies in Ethiopia and Nigeria [66–68]. Skill gap in data management system, weak supervision and feedback system, low or no legal enforcement to the surveillance activities, lack of incentives, lack of continued capacity building training, and lack of sense of ownership have been reported as factors affecting analysis and use of surveillance data [66]. Conclusions Using the SES framework analysis, through a consultative workshop data collection approach, we have managed to identify important features, the processes and assess the functioning of multifaceted and complex systems involved in management 0f dengue in the two districts of Tanzania. The findings illustrate that the districts have limited readiness in the management and control of dengue outbreaks. The surveillance system is faced by lack of or inappropriate format of guidelines, dissemination and translation for use of SCDs and insufficient reagents for laboratory confirmation of cases. Low capacity in data management, analysis and use was common in both districts. Inadequate knowledge on outbrea investigation among facility workers and shortage of sample collection and processing tools were identified as critical gaps. Individual and community practices and behaviours contribute to the challenges towards response to dengue outbreaks. The capacities of the districts in dengue epidemic readiness is limited. To strengthen epidemic preparedness at the district level, a well-functioning supervised surveillance system, with appropriate systems of analysis and feedback covering critical variables, is essential. Abbreviations ESR= Erythrocyte sedimentation rate; FBP= Full blood picture; HMIS= Health Management Information System; LLIN= Long lasting insecticide nets; MBVD= Mosquito-borne viral diseases; SCD= Standard case definition; SES= Socio-ecological system Funding This study received financial support from the European and Developing Countries Clinical Trials Partnership (EDCTP2) Programme (Grant RIA2016E-1609) which is supported under Horizon 2020, the European Union’s Framework Programme for Research and Innovation. Competing interests The authors declare that they have no competing interests Acknowledgements We would like to thank the Dar es Salaam City Council, Ilala and Kinondoni Municipal Councils for their enthusiasm and support to the accomplishment of this study. Page 13/21 Consent for publication Not applicable Ethics approval and consent to participate This project received ethical approval by the Medical Research Coordinating Committee of the National Institute for Medical Research (Ref. NIMR/HQ/R.8c/Vol 1/1168). Authors’ contributions LEGM, CS contributed to the study design. IRM, JG, RN, gathered the data and SFR, CS, JG and LEGM analysed the data. NH, MM, RK, MMR, GM and EDK contributed to drafting of the manuscript. All authors read and approved the final version of the manuscript. Availability of supporting data All data sets on which the findings and conclusions of this study rely are presented in the paper. However, data is available from the corresponding author upon reasonable request. References 1. Sindato C, Karimuribo ED, Pfeiffer DU, Mboera, LEG, Kivaria, F, et al. Spatial and Temporal Pattern of Rift Valley Fever Outbreaks in Tanzania; 1930 to 2007. PLoS ONE. 2014; 9(2): e88897. 1. Sindato C, Karimuribo ED, Pfeiffer DU, Mboera, LEG, Kivaria, F, et al. Spatial and Temporal Pattern of Rift Valley Fever Outbreaks in Tanzania; 1930 to 2007. PLoS ONE. 2014; 9(2): e88897. 2. Mboera LEG, Mweya CN, Rumisha SF, Tungu PK, Stanley G, et al. The risk of dengue virus transmission in Dar es Salaam, Tanzania during an epidemic period of 2014. PLoS Negl Trop Dis. 2016; 10(1): e004313. doi: 10.1371/journal.pntd.0004313. 2. Mboera LEG, Mweya CN, Rumisha SF, Tungu PK, Stanley G, et al. The risk of dengue virus transmission in Dar es Salaam, Tanzania during an epidemic period of 2014. PLoS Negl Trop Dis. 2016; 10(1): e004313. doi: 10.1371/journal.pntd.0004313. Page 14/21 Page 14/21 3. Vairo F, Mboera LEG, De Nardo P, Oriyo NM, Meschi S, Rumisha SF, et al. Clinical, virologic, and epidemiologic characteristics of dengue outbreak, Dar es Salaam, Tanzania, 2014. Emerg Infect Dis. 2016; 22 (5): 895-899. doi: 3201/eid2205.151462. 4. Kinimi E, Shayo M, Bisimwa P, Angwenyi S, Kasanga C, Weyer J. et al. Evidence of chikungunya virus infection among febrile patients seeking healthcare in selected districts of Tanzania. Infect Ecol Epidemiol. 2018; 8: https://doi.org/10.1080/20008686.2018.1553460. 5. Cao Z, Liu T, Li X, Wang J, Lin H, Chen L, et al. Individual and interactive effects of socio-ecological factors on dengue fever at fine spatial scale: a geographical detector-based analysis. Int J of Environ Res Public Health. 2017; 14; 795. doi: 10.3390/ijerph14070795. 6. Thomson MC, Muñoz ÁG, Cousin, R, Shumake-Guillemot J. Climate drivers of vector-borne diseases in Africa and their relevance to control programmes. Infect Dis Poverty. 2018; 7: https://doi.org/10.1186/s40249-018-0460-1 7. Parham PE, Waldock J, Christophides GK, Hemming D, Agusto F, Evans KJ et al. Climate, environmental and socio-economic change: weighing up the balance in vector-borne disease transmission. Philos Trans Royal Soc B. 2015; 370 (1665): http://dx.doi.org/10.1098/rstb.2013.0551. 8. Berkes F, Colding J, Folke C. Navigating Social-Ecological Systems: Building Resilience for Complexity and Change. New York: Cambridge University Press. 2003. 8. Berkes F, Colding J, Folke C. Navigating Social-Ecological Systems: Building Resilience for Complexity and Change. New York: Cambridge University Press. 2003. 9. Walker B, Holling CS, Carpenter SR, Kinzig A. Resilience, adaptability and transformability in social– ecological systems. Ecol Soc. 2004; 9:5. 10. References Applying the socio-ecological model to improving fruit and vegetable intake among low- income African Americans. J Comm Health. 2008; 33: 395–406. doi: 10.1007/s10900-008-9109-5. 20. Aguirre AA, Basu N, Kahn LH, Morin XK, Echaubard P, Wilcox BA. et al. Transdisciplinary and social- ecological health frameworks—Novel approaches to emerging parasitic and vector-borne diseases. Parasite Epidemiol Control. 2019; 4: e00084. doi: 1016/j.parepi.2019.e00084. 20. Aguirre AA, Basu N, Kahn LH, Morin XK, Echaubard P, Wilcox BA. et al. Transdisciplinary and social- ecological health frameworks—Novel approaches to emerging parasitic and vector-borne diseases. Parasite Epidemiol Control. 2019; 4: e00084. doi: 1016/j.parepi.2019.e00084. 21. Camara N, Ngasala B, Leyna G, Abade A, Rumisha SF, Oriyo NM. et al. Socio-demographic determinants of dengue infection during an outbreak in Dar es Salaam City, Tanzania. Tanzania J Health Res. 2018; 20(2). doi: http://dx.doi.org/10.4314/thrb.v20i2.3. 21. Camara N, Ngasala B, Leyna G, Abade A, Rumisha SF, Oriyo NM. et al. Socio-demographic determinants of dengue infection during an outbreak in Dar es Salaam City, Tanzania. Tanzania J Health Res. 2018; 20(2). doi: http://dx.doi.org/10.4314/thrb.v20i2.3. 22. Population and Housing Census 2012. United Republic of Tanzania, Dar es Salaam, 2013. 22. Population and Housing Census 2012. United Republic of Tanzania, Dar es Salaam, 2013. 23. Trpis M. Seasonal changes in the larval populations of Aedes aegypti in two biotypes in Dar es Salaam, Tanzania. Bull World Health Org. 1972; 47: 245-255. 23. Trpis M. Seasonal changes in the larval populations of Aedes aegypti in two biotypes in Dar es Salaam, Tanzania. Bull World Health Org. 1972; 47: 245-255. 24. Trpis M. Dry season survival of Aedes aegypti eggs in various breeding sites in the Dar es Salaam area, Tanzania. Bull World Health Org. 1972; 47(3): 433–437. 24. Trpis M. Dry season survival of Aedes aegypti eggs in various breeding sites in the Dar es Salaam area, Tanzania. Bull World Health Org. 1972; 47(3): 433–437. 25. Vairo F, Nicastri E, Yussuf SM, Cannas A. et al. IgG against dengue virus in healthy blood donors, Zanzibar, Tanzania. Emerg Infect Dis, 2014; 20: 465-468. doi: 3201/eid2003.130150. 26. Okada K, Morita R, Egawa K, Hirai Y, Kaida A, Shirano M. Dengue virus type 1 infection in traveller returning from Tanzania to Japan, 2019. Emerg Infect Dis. 2019; 25(9):1782-1784. doi: 3201/eid2509.190814. 27. World Health Organization. Weekly Bulletin on Outbreaks and other emergencies. Week 31-47: 29 July-24 November 2019. 28. Ostrom E. A diagnostic approach for going beyond panaceas. References Liu J, Dietz T, Carpenter SR, Alberti M, Folke C, Moran E et al. Complexity of coupled human and natural systems. Science. 2007; 317:1513–1516. doi:10.1126/science.1144004. 11. Gerdes GH. Rift Valley fever. Rev Sci Tech. 2004; 23 (2):613-623. doi: 10.20506/rst.23.2.1500. 12. Powell JR, Tabachnick WJ. History of domestication and spread of Aedes aegypti - a review. Mem Inst Oswaldo Cruz. 2013; 108:11–17. doi: 1590/0074-0276130395. 13. Smith KR, Ezzati M. How environmental health risks change with development: The epidemiologic and environmental risk transitions revisited. Ann Rev Environ Res. 2005; 30: 291–333. https://doi.org/10.1146/annurev.energy.30.050504.144424. 14. Wilcox BA, Colwell RR. Emerging and re-emerging infectious diseases: biocomplexity as an interdisciplinary paradigm. EcoHealth 2005; 2(4): 244–257. https://doi.org/10.1007/s10393-005- 8961-3. 14. Wilcox BA, Colwell RR. Emerging and re-emerging infectious diseases: biocomplexity as an interdisciplinary paradigm. EcoHealth 2005; 2(4): 244–257. https://doi.org/10.1007/s10393-005- 8961-3. 15. Wilcox BA, Gubler DJ. Disease ecology and the global emergence of zoonotic pathogens. Environ Health Prev Med 2005; 10(5): 263–272. doi: 1007/bf02897701. 15. Wilcox BA, Gubler DJ. Disease ecology and the global emergence of zoonotic pathogens. Environ Health Prev Med 2005; 10(5): 263–272. doi: 1007/bf02897701. 16. McLaren L, Hawe P. (2005) Ecological perspectives in health research. J Epidemiol Comm Health 2005; 59(1):6-14. doi: 1136/jech.2003.018044. 16. McLaren L, Hawe P. (2005) Ecological perspectives in health research. J Epidemiol Comm Health 2005; 59(1):6-14. doi: 1136/jech.2003.018044. 17. McLeroy K, Bibeau D, Steckler A, Glanz K. An ecological perspective on health promotion programs. Health Edu Behav. 1988; 15(4): 351–377. https://doi.org/10.1177/109019818801500401. 17. McLeroy K, Bibeau D, Steckler A, Glanz K. An ecological perspective on health promotion programs. Health Edu Behav. 1988; 15(4): 351–377. https://doi.org/10.1177/109019818801500401. Page 15/21 Page 15/21 18. Gregson J, Foerster S, Orr R, Jones L, Benedict J, Clarke B, et al. (2001) System, environmental, and policy changes: Using the social-ecological model as a framework for evaluating nutrition education and social marketing programs with low-income audiences. J Nutr Educ Behav. 2001; 33: 4–15. doi: 10.1016/s1499-4046(06)60065-1. 18. Gregson J, Foerster S, Orr R, Jones L, Benedict J, Clarke B, et al. (2001) System, environmental, and policy changes: Using the social-ecological model as a framework for evaluating nutrition education and social marketing programs with low-income audiences. J Nutr Educ Behav. 2001; 33: 4–15. doi: 10.1016/s1499-4046(06)60065-1. 19. Robinson T. Applying the socio-ecological model to improving fruit and vegetable intake among low- income African Americans. J Comm Health. 2008; 33: 395–406. doi: 10.1007/s10900-008-9109-5. 19. Robinson T. References Proc Natl Acad Sci USA. 2007; 104(39):15181–15187. http://dx.d org/10.1073/pnas.0702288104. 29. Ostrom E. A general framework for analyzing sustainability of social-ecological systems. Science 2009; 325 (5939): 419–422. http://dx.doi.org/10.1126/science.1172133. 30. Poteete A, Janssen MA, Ostrom E. Working together: collective action, the commons and multiple methods in practice. Princeton University Press, Princeton, New Jersey, USA 2002. http://dx.doi.org/10.1515/9781400835157. 30. Poteete A, Janssen MA, Ostrom E. Working together: collective action, the commons and multiple methods in practice. Princeton University Press, Princeton, New Jersey, USA 2002. http://dx.doi.org/10.1515/9781400835157. 31. McGinnis MD, Ostrom E. Social-ecological system framework: initial changes and continuing challenges. Ecol Soc. 2014; 19(2): 30. http://dx.doi.org/10.5751/ES-06387-190230. 31. McGinnis MD, Ostrom E. Social-ecological system framework: initial changes and continuing challenges. Ecol Soc. 2014; 19(2): 30. http://dx.doi.org/10.5751/ES-06387-190230. 32. Partelow S. Key steps for operationalizing social–ecological system framework research in small- scale fisheries: A heuristic conceptual approach. Mar Policy. 2015; 51(C): 507-511. doi: 10.1016/j.marpol.2014.09.005. 32. Partelow S. Key steps for operationalizing social–ecological system framework research in small- scale fisheries: A heuristic conceptual approach. Mar Policy. 2015; 51(C): 507-511. doi: 10.1016/j.marpol.2014.09.005. Page 16/21 33. Bengtsson, M. How to plan and perform a qualitative study using content analysis. Nurs Open. 2016; 8(2): 8-14. doi: 10.1016/j.npls.2016.01.001. 34. Sharma RC, Thakur A, Sharma M, Mishra S. Teachers’ Perception of Open Educational Resources: Data Collection through Workshops (ROER4D). Paper presented at the Open Education Global Conference 2015. Calgary, AB. 2015. 35. Ørngreen R, Levinsen K. Workshops as a Research Methodology. EJEL. 2017; 15 (1): 70-81. 36. Lain S. Show, Don’t Tell: Reading Workshop Fosters Engagement and Success. TJLE. 2017; 5(2): 160- 167. 37. Knudsen AB, Slooff R. Vector-borne disease problems in rapid urbanization: new approaches to vector control. Bull World Health Organ. 1992; 70:1–6. 38. Kendall C, Hudelson P, Leontsini E, Winch P, Lloyd L. Urbanization, dengue, and the health transition: Anthropological contributions to international health. Med Anthropol Q. 1993; 5:257–268. doi: 1525/maq.1991.5.3.02a00050. 39. Rubio A, Cardo V, Vezzani D. Tire-breeding mosquitoes of public health importance along an urbanisation gradient in Buenos Aires, Argentina. Mem Inst Oswaldo Cruz. 2011; 16(6): 678-684. doi: 1590/s0074-02762011000600006. 39. Rubio A, Cardo V, Vezzani D. Tire-breeding mosquitoes of public health importance along an urbanisation gradient in Buenos Aires, Argentina. Mem Inst Oswaldo Cruz. 2011; 16(6): 678-684. doi: 1590/s0074-02762011000600006. 40. Philbert A, Ijumba JN. Preferred breeding habitats of Aedes aegypti (Diptera- Culicidae) mosquito and its public health implications in Dar es Salaam, Tanzania. J Environ Res Manage. 2013; 4:344– 351. References 40. Philbert A, Ijumba JN. Preferred breeding habitats of Aedes aegypti (Diptera- Culicidae) mosquito and its public health implications in Dar es Salaam, Tanzania. J Environ Res Manage. 2013; 4:344– 351. 41. Mathias L, Baraka V, Philbert A, Innocent E, Francis F, Nkwengulila G, Kweka EJ. Habitat productivity and pyrethroid susceptibility status of Aedes aegyptimosquitoes in Dar es Salaam, Tanzania. Infect Dis Poverty. 2017; 6:1–10. https://doi.org/10.1186/s40249-017-0316-0. 42. Saleh F, Kitau J, Konradsen F, Alifrangis M, Lin C, Juma S et al. Habitat characteristics for immature stages of Aedes aegyptiin Zanzibar City, Tanzania. J Am Mosq Control Assoc. 2018; 18 (34): 190- 200. doi: 2987/17-6709.1. 42. Saleh F, Kitau J, Konradsen F, Alifrangis M, Lin C, Juma S et al. Habitat characteristics for immature stages of Aedes aegyptiin Zanzibar City, Tanzania. J Am Mosq Control Assoc. 2018; 18 (34): 190- 200. doi: 2987/17-6709.1. 43. Maheu-Giroux M, Castro MC. Impact of Community-Based Larviciding on the Prevalence of Malaria Infection in Dar es Salaam, Tanzania. PLoS ONE 2013; 8(8): e71638. https://doi.org/10.1371/journal.pone.0071638. 43. Maheu-Giroux M, Castro MC. Impact of Community-Based Larviciding on the Prevalence of Malaria Infection in Dar es Salaam, Tanzania. PLoS ONE 2013; 8(8): e71638. https://doi.org/10.1371/journal.pone.0071638. 44. World Health Organization. Larval source management: a supplementary measure for malaria vector control. Geneva, Switzerland, 2013. 44. World Health Organization. Larval source management: a supplementary measure for malaria vector control. Geneva, Switzerland, 2013. 45. Hales S, Weinstein P, Souares Y, Woodward A. El Nino and the dynamics of vector-borne disease transmission. Environ Health Perspect. 1999; 107(2): 99–102. doi: 1289/ehp.9910799. 45. Hales S, Weinstein P, Souares Y, Woodward A. El Nino and the dynamics of vector-borne disease transmission. Environ Health Perspect. 1999; 107(2): 99–102. doi: 1289/ehp.9910799. 46. Pham HV,Huong, Doan HTM, Phan TTT, Minh NNT. Ecological factors associated with dengue fever in a central highlands Province, Vietnam. BMC Infect Dis. 2011; 11: 172. https://doi.org/10.1186/1471-2334-11-172. 47. Arunachalam N, Tana S, Espino F, Kittayapong P, Abeyewickreme W, Wai KT. et al. Eco-bio-social determinants of dengue vector breeding: a multicountry study in urban and periurban Asia. Bull World Health Organ. 2010; 88(3):173–184. Page 17/21 Page 17/21 48. Ault SK. Environmental management: a re-emerging vector control strategy. Am J Trop Med Hyg. 1994; 50:35–49. doi: 10.4269/ajtmh.1994.50.35. 49. Anno S, Imaoka K, Tadono T, Igarashi T, Sivaganesh S, Kannathasan S. et al. Space–time clustering characteristics of dengue based on ecological, socio-economic, and demographic factors in northern Sri Lanka. References Geosp Health 2015; 10(376):215-222. doi: 4081/gh.2015.376. 50. Gubler DJ, Clark GG. Community involvement in the control of Aedes aegypti. Acta Trop. 1996; 61:169–179. doi: 10.1016/0001-706x (95)00103-l. 51. Parks W, Lloyd L. Planning Social Mobilization and Communication for Dengue Fever Prevention and Control a Step-by-step Guide. Geneva: World Health Organization, 2004. 52. Grimshaw JM, Russell IT. Effect of clinical guidelines on medical practice: a systematic review of rigorous evaluations. Lancet 1993; 342:1317–1322. doi: 10.1016/0140-6736(93)92244-n. 53. Nathwani D, Rubinstein E, Barlow G, Davey P. Do guidelines for community-acquired pneumonia improve the cost-effectiveness of hospital care? Clin Infect Dis. 2001; 32 (5):728–741. https://doi.org/10.1086/319216. 54. Grol R, Grimshaw J. From best evidence to best practice: effective implementation of change in patients’ care. Lancet 2003; 362(9391):1225–1230. doi: 10.1016/S0140-6736(03)14546-1. 55. Gagliardi AR, Brouwers MC, Palda VA, Lemieux-Charles L, Grimshaw JM. How can we improve guideline use? A conceptual framework of implementability. Implement Sci. 2011; 6:26. https://doi.org/10.1186/1748-5908-6-2.6. 56. Armstrong PW. Do guidelines influence practice? Heart 2003; 89(3): 349–352. 57. Schunemann HJ, Fretheim A, Oxman AD. Improving the use of research evidence in guideline development: 1. Guidelines for guidelines. Health Res Policy Syst. 2006; 4:13. doi: 1186/1478-4505- 4-13. 58. Stanback J, Griffey S, Lynam P, Ruto C, Cummings S. Improving adherence to family planning guidelines in Kenya: an experiment. Int J Qual Health Care 2007; 19(2):68–73. doi: 10.1093/intqhc/mzl072. 59. Francke AL, Smit MC, de Veer AJ, Mistiaen P. Factors influencing the implementation of clinical guidelines for health care professionals: a systematic meta-review. BMC Med Inform Dec. 2008; 8:38. https://doi.org/10.1186/1472-6947-8-38. 60. Karimuribo ED, Mutagahywa E, Sindato C, Mboera L, Mwabukusi M, Kariuki Njenga M. et al. A Smartphone App (AfyaData) for Innovative One Health Disease Surveillance from Community to National Levels in Africa: Intervention in Disease Surveillance. JMIR Public Health Surv. 2017; 3 (4): e94. doi: 10.2196/publichealth.7373. 61. Anno S, Hara T, Kai H, Chang Y, Lee M-A, Oyoshi K. et al. Deep learning applications for predicting Dengue fever outbreak. J Environ Res. 2018; 2: 21. 62. Nelson, C., Lurie, N. & Wasserman, J. Assessing public health emergency preparedness: concepts, tools, and challenges. Annu Rev Public Health 2008; 28: 1–18. pmid:17129174. doi: 10.1146/annurev.publhealth.28.021406.144054. Page 18/21 63. Mghamba JM, Mboera LEG, Krekamoo W, Senkoro KP, Rumisha SF, Shayo E. et al. Challenges of implementing an Integrated Disease Surveillance and Response strategy using the current Health Management Information System in Tanzania. Tanzania Health Res Bull. 2004; 6: 57–63. https://doi.org/10.4314/thrb.v6i2.14243. 64. References Phalkey RK, Yamamoto S, Awate P, Marx Challenges with the implementation of an Integrated Disease Surveillance and Response (IDSR) system: systematic review of the lessons learned. Health Policy Plann. 2015; 30 (1): 131–143. doi: 10.1093/heapol/czt097. 64. Phalkey RK, Yamamoto S, Awate P, Marx Challenges with the implementation of an Integrated Disease Surveillance and Response (IDSR) system: systematic review of the lessons learned. Health Policy Plann. 2015; 30 (1): 131–143. doi: 10.1093/heapol/czt097. 65. Rumisha SF, Mboera LE, Senkoro KP, Gueye D, Mmbuji PK. Monitoring and evaluation of integrated disease surveillance and response in selected districts in Tanzania. Tanzania Health Res Bull. 2007; 9: 1–11. https://doi.org/10.4314/thrb.v9i1.14285. 65. Rumisha SF, Mboera LE, Senkoro KP, Gueye D, Mmbuji PK. Monitoring and evaluation of integrated disease surveillance and response in selected districts in Tanzania. Tanzania Health Res Bull. 2007; 9: 1–11. https://doi.org/10.4314/thrb.v9i1.14285. 66. Alemu T, Gutema H, Legesse S, Nigussie T, Yenew Y, Gahse K. Evaluation of public health surveillance system performance in Dangila district, Northwest Ethiopia: a concurrent embedded mixed quantitative/qualitative facility-based cross-sectional study. BMC Public Health 2019; 19: https://doi.org/10.1186/s12889-019-7724-y. 66. Alemu T, Gutema H, Legesse S, Nigussie T, Yenew Y, Gahse K. Evaluation of public health surveillance system performance in Dangila district, Northwest Ethiopia: a concurrent embedded mixed quantitative/qualitative facility-based cross-sectional study. BMC Public Health 2019; 19: https://doi.org/10.1186/s12889-019-7724-y. 67. Begashaw B, Tesfaye T. Assessment of integrated disease surveillance and response implementation in special health facilities of Dawuro Zone. J Anesthesiol. 2016; 4(3):11–15. doi: 11648/j.ja.20160403.11 67. Begashaw B, Tesfaye T. Assessment of integrated disease surveillance and response implementation in special health facilities of Dawuro Zone. J Anesthesiol. 2016; 4(3):11–15. doi: 11648/j.ja.20160403.11 68. Motilewa O, Akwaowo CD, Ekanem, AM. Assessment of implementation of integrated disease surveillance and response in Akwaibom State Nigeria. Ibom Med J. 2015; 8(1):24-25. 68. Motilewa O, Akwaowo CD, Ekanem, AM. Assessment of implementation of integrated disease surveillance and response in Akwaibom State Nigeria. Ibom Med J. 2015; 8(1):24-25. Figures Page 19/21 District profiles and socio-ecological factors District profiles and socio-ecological factors Page 20/21 Page 20/21 Figure 2 Presence (√) and absence (X) of risks, prevention and control activities for dengu Presence (√) and absence (X) of risks, prevention and control activities for dengue in the study districts Presence (√) and absence (X) of risks, prevention and control activities for dengue in the study districts Page 21/21
https://openalex.org/W2065516946	https://europepmc.org/articles/pmc4004462?pdf=render	English	null	Improved multi-level protein–protein interaction prediction with semantic-based regularization	BMC bioinformatics	2,014	cc-by	14,966	METHODOLOGY ARTICLE Open Access © 2014 Saccà et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. Saccà et al. BMC Bioinformatics 2014, 15:103 http://www.biomedcentral.com/1471-2105/15/103 Saccà et al. BMC Bioinformatics 2014, 15:103 http://www.biomedcentral.com/1471-2105/15/103 Improved multi-level protein–protein interaction prediction with semantic-based regularization Claudio Saccà1, Stefano Teso2, Michelangelo Diligenti1 and Andrea Passerini2* Claudio Saccà1, Stefano Teso2, Michelangelo Diligenti1 and Andrea Passerini2* Abstract Background: Protein–protein interactions can be seen as a hierarchical process occurring at three related levels: proteins bind by means of specific domains, which in turn form interfaces through patches of residues. Detailed knowledge about which domains and residues are involved in a given interaction has extensive applications to biology, including better understanding of the binding process and more efficient drug/enzyme design. Alas, most current interaction prediction methods do not identify which parts of a protein actually instantiate an interaction. Furthermore, they also fail to leverage the hierarchical nature of the problem, ignoring otherwise useful information available at the lower levels; when they do, they do not generate predictions that are guaranteed to be consistent between levels. Results: Inspired by earlier ideas of Yip et al. (BMC Bioinformatics 10:241, 2009), in the present paper we view the problem as a multi-level learning task, with one task per level (proteins, domains and residues), and propose a machine learning method that collectively infers the binding state of all object pairs. Our method is based on Semantic Based Regularization (SBR), a flexible and theoretically sound machine learning framework that uses First Order Logic constraints to tie the learning tasks together. We introduce a set of biologically motivated rules that enforce consistent predictions between the hierarchy levels. Conclusions: We study the empirical performance of our method using a standard validation procedure, and compare its performance against the only other existing multi-level prediction technique. We present results showing that our method substantially outperforms the competitor in several experimental settings, indicating that exploiting the hierarchical nature of the problem can lead to better predictions. In addition, our method is also guaranteed to produce interactions that are consistent with respect to the protein–domain–residue hierarchy. Notwithstanding the increased availability of interac- tion data, the natural question of whether two arbitrary proteins interact, and why, is still open. The growing literature on protein interaction prediction [4-6] is symp- tomatic of the gap separating the amount of available data and the effective size of the interaction network [7]. The present paper is a contribution towards filling this gap. Correspondence: passerini@disi.unitn.it 2Dipartimento di Ingegneria e Scienza dell’Informazione, University of Trento, Trento, Italy Full list of author information is available at the end of the article Saccà et al. BMC Bioinformatics 2014, 15:103 http://www.biomedcentral.com/1471-2105/15/103 Page 2 of 18 Saccà et al. BMC Bioinformatics 2014, 15:103 http://www.biomedcentral.com/1471-2105/15/103 Saccà et al. BMC Bioinformatics 2014, 15:103 http://www.biomedcentral.com/1471-2105/15/103 interaction [9]. At the lowest level, the interaction is instantiated by the binding of a pair of protein inter- faces, patches of solvent accessible residues with compat- ible shapes and chemical properties [10]. The low-level features of the binding sites determine whether the inter- action is transient or permanent, whether two proteins compete for interaction with a third one, etc. Figure 1 illus- trates the multi-level mechanisms with an example taken from the PDB. Finally, residue-level predictions, i.e., interface recogni- tion, enable the detailed study of the principles of protein interactions, and are crucial for tasks such as rational drug design [3], metabolic reconstruction and engineering [19], and identification of hot-spots [20] in the absence of structure information. Given the usefulness of knowing the details of protein– protein interactions at diverse levels of detail, and based on earlier ideas of Yip et al. [21], in this paper we address the problem of collectively predicting the binding state of all proteins, domains, and residues in a network. We call this task the multi-level protein interaction prediction problem (MLPIP for short). Despite the significance of low-level details in elu- cidating the mechanics of protein–protein interactions, most of the current experimental data comes from high- throughput screening techniques, such as yeast two- hybrid (Y2H) assays [11]. These techniques do not provide information on domain- or residue-level interactions, which require solving the three-dimensional structure of each protein-protein complex, an expensive and time con- suming task addressed by X-Ray crystallography, NMR, or electron microscopy techniques [12]. As a consequence, protein–protein interaction data is under-characterized at the domain and residue levels: the current databases are relatively lacking when compared to the magnitude of the existing body of data about protein-level interac- tions [13]. At the time of writing, the PDB hosts 84,418 structures, but merely 4,210 resolved complexes (accord- ing to http://www.rcsb.org/pdb/statistics/ holdings.do, retrieved on 2013/06/20). The latter cover only a tiny fraction of the interactions stored in databases such as BioGRID and MIPS. From a computational point of view, the most important feature of the multi-level prediction problem is its inher- ently relational nature. Proteins, domains and residues are organized in a hierarchy, which dictates constraints on the binding state of pairs of objects at the differ- ent levels, as follows. Saccà et al. BMC Bioinformatics 2014, 15:103 http://www.biomedcentral.com/1471-2105/15/103 On the one hand, whenever two proteins are bound, at least two of their domains must also be bound, and, similarly, there must be residues in the two domains that form an interface. On the other hand, if no residues of the two proteins interact, neither do their domains, nor the proteins themselves. In other words, predictions at different levels must be consistent. In this paper we cast the multi-level prediction problem as a statistical-relational learning task, leveraging the lat- est developments in the field. Our prediction method is based on Semantic Based Regularization [22], an elegant semi-supervised prediction framework that caters both the effectiveness of kernel machines and the expressiv- ity of First Order Logic (FOL). The constraints described above are encoded as FOL rules, which are used to enforce consistent predictions at all levels of the interaction hier- archy. By computing multi-level predictions, our method can not only infer which protein pairs are likely to interact, but also provide details about how the interactions take place. Our empirical evaluation shows the effectiveness of this constraint-based approach in boosting predictive per- formance, achieving substantial improvements over both From a purely biological perspective, predictions at dif- ferent levels have several important applications. The network topology and individual features of protein inter- actions are an essential component of a wide range of biological tasks: inferring protein function [14] and local- ization [15], reconstructing signal and metabolic pathways [16], discovering candidate targets for drug development [2]. Finer granularity predictions at the domain level allow to discover affinities between domain types that can be carried over to other proteins [17,18]; domain– domain networks have also been assessed as being typ- ically more reliable than their protein counterparts [13]. Figure 1 The protein–domain–residue hierarchy. Two bound proteins and their interacting domains and residues, captured in PDB complex 4IOP. The proteins are a Killer cell lectin-like receptor (in violet) and its partner, a C-type lectin domain protein (in blue). (Left) Interaction as visible from the contact surface. (Center) The two C-type lectin domains instantiating the interaction. (Right) Effectively interacting residues in red. Figure 1 The protein–domain–residue hierarchy. Two bound proteins and their interacting domains and residues, captured in PDB complex 4IOP. The proteins are a Killer cell lectin-like receptor (in violet) and its partner, a C-type lectin domain protein (in blue). (Left) Interaction as visible from the contact surface. Background Physical interactions between proteins are the workhorse of cell life and development [1], and play an extremely important role both in the mechanisms of disease [2] and in the design of new drugs [3]. In recent years, there has been enormous interest in reverse engineer- ing the protein–protein interaction (PPI) networks of several species, particularly due to the availability of high- throughput experimental techniques, leading to an abun- dance of large databases on all aspects of PPIs [4]. Our work is based on the observation that physical interactions can be viewed at three levels of detail. At a higher level, two proteins interact to perform some function within a biological pathway (e.g. metabolism, signaling, regulation, etc.) [8]. At a lower level, the same interaction occurs between a pair of specific domains appearing in the proteins; the types of the domains involved characterize the functional semantics of the Correspondence: passerini@disi.unitn.it 2Dipartimento di Ingegneria e Scienza dell’Informazione, University of Trento, Trento, Italy Full list of author information is available at the end of the article Page 2 of 18 Problem definition PPI networks are most naturally formalized as graphs, where nodes represent proteins and edges represent inter- actions. Given a set of features describing the properties of the proteins in the network (e.g. primary structure, local- ization, tertiary structure —when available—, etc.), infer- ring the PPI network topology amounts to determining those pairs of proteins that are likely to interact. This task is often cast as a pairwise classification problem, where a binary classifier takes as input a pair of proteins (or rather their feature-based representations) and predicts whether they interact or not. Standard binary classification meth- ods, such as Support Vector Machines [23], can be used to implement the pairwise classifier. In this setting, the inter- action depends only on the features of the two incident nodes, and is independent of all other nodes. Interactions between domains or residues can be predicted similarly. The most straightforward way to address the MLPIP problem is to cast the three interaction prediction prob- lems, for proteins, domains and residues respectively, as independent pairwise classification tasks. However, as previously discussed, these problems are clearly strongly related: two proteins interact via one or more domains, which in turn contain patches of residues that consti- tute the interaction surface. Ignoring these relationships can lead to heavily suboptimal, inconsistent predictions, where, e.g. two proteins are predicted to interact but none of their domains are predicted to be involved in this interaction. Making these relationships explicit and forc- ing predictors to satisfy consistency constraints is the key contribution of this work. In the machine learning com- munity, this kind of scenario characterized by multiple related prediction tasks is usually cast as a statistical- relational learning problem [24,25], where the goal is to collectively classify the state of all objects of interest, tak- ing into account the relations existing between them. The solution we adopt is grounded in this learning framework. ∀(p,p’)boundp(p,p’) ⇒∃(d,d’) boundd(d,’d) ∧ parentpd(p,d) ∧ parentpd(p’,d’) ∀(p,p’)boundp(p,p’) ⇒∃(d,d’) boundd(d,’d) ∧ parentpd(p,d) ∧ parentpd(p’,d’) Each binary classifier is implemented in the SBR frame- work as a kernel machine [26]. The key component of kernel machines is the kernel function, which measures the similarity between objects in terms of their repre- sentations. A protein, for instance, can be represented as the sequence of its residues, plus additional infor- mation as its subcellular localization and/or its phylo- genetic profile. Saccà et al. BMC Bioinformatics 2014, 15:103 http://www.biomedcentral.com/1471-2105/15/103 Page 3 of 18 Saccà et al. BMC Bioinformatics 2014, 15:103 http://www.biomedcentral.com/1471-2105/15/103 corresponding sets XP, XD and XR respectively. Predicates represent properties of objects or relationships between them. Depending on the scenario, some predicates are always known (called given predicates), some other are known only for a subset of the objects, and their value should be predicted when unknown (query or target pred- icates). The parentpd(p,d) predicate, for instance, specifies that domain d ∈XD is part of protein p ∈XP, i.e. the predicate is true for all (p,d) pairs for which d is a domain of p, and false otherwise. The value of this predi- cate is known for all objects in our domain; note that there indeed are many proteins whose domains are unknown, but in this case there is no corresponding domain object in our data). The boundp(p,p’) predicate specifies whether two proteins p and p’ are interacting. This is one of the target predicates, whose truth value should be pre- dicted for novel protein-protein pairs. Similar predicates are defined for domain and residue level bindings. Target predicates are modelled as binary classifiers, i.e. functions trained to predict the truth value of the predicate. Rela- tionships between predicates can be introduced in order to enforce constraints known to hold in the domain. SBR allows to exploit the full power of First Order Logic in doing this. As a matter of example, the notion that two interacting proteins should have at least one interacting domain can be modelled as (see Methods for details on First Order Logic notation): an unconstrained baseline and the only existing alterna- tive MLPIP method [21]. Problem definition Having the same subcellular localization, for instance, should increase the similarity between two proteins, as having a similar amino acid composition. Designing appropriate kernels is a crucial component of a successful predictor. A kernel machine is a function which predicts a certain property of an object x in terms of a weighted sum of similarities to other objects for which the property is known, i.e.: Saccà et al. BMC Bioinformatics 2014, 15:103 http://www.biomedcentral.com/1471-2105/15/103 (Center) The two C-type lectin domains instantiating the interaction. (Right) Effectively interacting residues in red. Overview of the proposed method In this paper we propose solving the multi-level prediction problem adapting a state-of-the-art statistical-relational learning framework, namely Semantic Based Regulariza- tion (SBR) [22]. SBR ties multiple learning tasks, which are themselves addressed by kernel machines, using con- straints expressing First Order Logic knowledge. In the following we give an overview of the SBR framework, also pictured in Figure 2; see Methods for further details. f (x) = i wiK(x, xi) A kernel machine could for instance predict whether a protein is an enzyme or not (binary classification), in terms of weighted similarity to other proteins. Being sim- ilar to an enzyme xi will drive the prediction towards the positive (enzyme) class (positive weight wi), while Let X be a set of objects. In most scenarios, objects are typed, so that objects of the same type can be con- sidered as belonging to the same group. In our setting, object types are proteins, domains and residues, with Saccà et al. BMC Bioinformatics 2014, 15:103 http://www.biomedcentral.com/1471-2105/15/103 Page 4 of 18 e 2 Visualization of the proposed method. Visualization of the proposed method. (a) Kernel preparation at the three levels. A kernel is ed for each input feature (Left); the resulting matrices are summed up to obtain a per-object kernel (Middle), which is transformed into a ise kernel using Eq 1. Here Np (Nd, Nr) is the number of individual proteins (respectively domains, residues) in the level, while Npp (Ndd, Nrr) is umber of protein (respectively domain, residue) interactions in the dataset. (b) Instantiation of all predicates (Table 1) over a pair of proteins p’ ’ and their parts. Circles represent proteins, domains and residues. Dotted lines indicate a parent-child relationship between objects, senting the parentpd and parentdr predicates. Solid lines link pairs of bound objects, i.e. objects for which the boundp, boundd or ndr predicates are true. (c) Visualization of the experimental pipeline. Given the pairwise kernels, the set of rules (Table 2), a set of example ctions, and a description of the protein-domain-residue hierarchy, SBR finds a prediction for the query predicates consistent with the rules. Figure 2 Visualization of the proposed method. Visualization of the proposed method. (a) Kernel preparation at the three levels. A kernel is derived for each input feature (Left); the resulting matrices are summed up to obtain a per-object kernel (Middle), which is transformed into a pairwise kernel using Eq 1. Overview of the proposed method f (x, x′) = i wiK((x, x′), (xi, x′ i)) f (x, x′) = i wiK((x, x′), (xi, x′ i)) Here the kernel function measures the similarity between two pairs of objects, so that, e.g. two proteins will be pre- dicted as interacting if they are similar to protein pairs which are known to interact, and dissimilar from pairs known to not interact. Here the kernel function measures the similarity between two pairs of objects, so that, e.g. two proteins will be pre- dicted as interacting if they are similar to protein pairs which are known to interact, and dissimilar from pairs known to not interact. Given a kernel between objects K(x, x′), it is possible to construct a pairwise kernel by means of a the following transformation [27]: K((xi, xj), (xk, xl)) = K(xi, xk) · K(xi, xl) + K(xj, xk) · K(xj, xl) (1) (1) (1) This transformation guarantees that, if the input func- tion K is a valid kernel, so is the resulting pairwise function. SBR is a semi-supervised method [28], meaning that the set of target proteins is given beforehand and can be exploited during the learning stage to fine-tune the model. Semi-supervised learning is known to enhance the pre- diction ability when appropriately used [29], and can be applied very naturally to PPI prediction, as the full set of proteins is always known. As already explained, in SBR each target predicate is implemented as a kernel machine, and the state of a predicate for an uncharacterized pair of proteins can be inferred by querying the machine. Positive predictions correspond to true predicates, i.e. bound protein pairs, and negative predictions to false ones. The confidence of the kernel machine, also called margin, embodies the confidence in the state of the predicate, that is, how strongly two proteins are believed to interact (or not). To summarize, at each level the state of an uncharac- terized pair of objects, e.g. proteins p and p′, is mainly inferred by the similarity of the pair (p, p′) to other pairs that are known to interact or not, through the pair- wise kernel function K and the learned weights w. Thus the kernel allows to propagate information horizontally within the same level. At the same time, the FOL con- straints allow to propagate information vertically between the levels, by keeping the interaction pattern along the protein-domain-residue hierarchy consistent. Overview of the proposed method Here Np (Nd, Nr) is the number of individual proteins (respectively domains, residues) in the level, while Npp (Ndd, Nrr) is the number of protein (respectively domain, residue) interactions in the dataset. (b) Instantiation of all predicates (Table 1) over a pair of proteins p’ and p’ and their parts. Circles represent proteins, domains and residues. Dotted lines indicate a parent-child relationship between objects, representing the parentpd and parentdr predicates. Solid lines link pairs of bound objects, i.e. objects for which the boundp, boundd or boundr predicates are true. (c) Visualization of the experimental pipeline. Given the pairwise kernels, the set of rules (Table 2), a set of example interactions, and a description of the protein-domain-residue hierarchy, SBR finds a prediction for the query predicates consistent with the rules. Saccà et al. BMC Bioinformatics 2014, 15:103 http://www.biomedcentral.com/1471-2105/15/103 Saccà et al. BMC Bioinformatics 2014, 15:103 http://www.biomedcentral.com/1471-2105/15/103 Page 5 of 18 of protein, domain and residue pairs, y∗, is a solution to the following optimization problem: being similar to a non-enzyme xj will drive the prediction towards the opposite class (negative weight wj). j In the interaction prediction setting, target predicates actually predict properties of pairs of objects (proteins, domains or residues). We thus employ a pairwise kernel machine classifier to model the target predicate: y∗= arg max y consist(y, f ) + consist(y, KB) where the first term accounts for consistency between inferred truth values and confidence of the individual pre- dictions, and the second incorporates information on the degree of satisfaction of the constraints build from the FOL knowledge. Contrarily to standard kernel methods, this optimization problem is non-convex. This is com- monly the case for complex statistical-relational learning tasks [24], and implies that we are restricted to finding local optima. SBR employs a two-stage learning process to make training effective even in presence of local optima. In particular, the first stage of SBR learning takes into account the fitting of the individual predictions to the supervised data. This learning task is convex and can be efficiently solved. The solution found in the first stage is used as starting point for a second stage, where the FOL knowledge is also considered. This optimization strategy has been experimentally proved to find high-quality solu- tions without adding the computational burden of other non-convex optimization techniques [22]. Overview of the proposed method Given the output of the kernel machines for all target predicates, SBR uses the First Order Logic rules to con- dition the state of the correlated predicates. It does so by first translating the FOL rules into continuous constraints, which we discuss more thoroughly in Methods. The vari- ables coming into play into the continuous constraints are the confidences of all target predicates (and the state of all given predicates) appearing in the equivalent FOL constraint. The amount of violation is reflected by the value of the continuous constraints: if the predicted pred- icates satisfy a FOL rule, the corresponding constraint will have a value equal to 1; on the other hand, the closer the constraint value to zero, the more the FOL rule is violated. Modeling multi-level interactions For a complete list of predicates, see Table 1. For a visualization of the predicates instantiated over a protein pair, see Figure 2-b. ∀(p,p’) hasdom(p)∧ hasdom(p’) ⇒ (boundp(p,p’) ⇒ ∃(d,d’) boundd(d,d’)∧ parentpd(p,d)∧ parentpd(p’,d’)) This is the complete P→D rule. The left-hand side is always false for proteins without domains, making the rule always satisfied in this case (effectively disabling the effect of the rule on the learning process). We define the complementary D→P rule as follows: In what follows we describe how to design inter-level FOL constraints to properly enforce consistency between predictions at different levels. We focus on modeling the constraints tying proteins and domains; it is easy to see that the ones between domains and residues can be mod- elled similarly (with one peculiar exception that will be pointed out later). Table 2 reports the complete list of rules. ∀(p,p’) (∃(d,d’) boundd(d,d’) ∧ parentpd(p,d) ∧ parentpd(p’,d’) ⇒boundp(p,p’)) Inter-level constraints can be seen as propagating infor- mation from the upper layer to the lower one and in the opposite direction. To model this mechanism, we use two distinct constraints: the P→D rule and the D→P rule. A simplified version of the P→D rule is: This rule is applied to all protein pairs, demanding that if there is a pair of bound children domains then the pro- teins must be bound too, and vice versa that if the parent proteins are unbound so are the domains. The P→D and D→P rules could be merged into a single equivalent rule using the double implication (⇔). However, the rules have been considered separately to keep their effects on the results separated and easier to analyze. ∀(p,p’) boundp(p,p’) ⇒∃(d,d’) boundd(d,d’)∧ parentpd(p,d) ∧ parentpd(p’,d’) ∀(p,p’) boundp(p,p’) ⇒∃(d,d’) boundd(d,d’)∧ parentpd(p,d) ∧ parentpd(p’,d’) Intuitively, the rule means that whenever two proteins are bound (and therefore the left-hand side (LHS) of the implication is true) then there must be at least one pair of child domains that are bound (the right-hand side (RHS) is true). In classical First Order Logic the rule would require that, whenever none of the child domains is bound (the To simulate the unidirectional information propaga- tion between levels, as done by Yip et al. [21] (see Related work), we modified how SBR converts logic impli- cations by using the t-norm residuum, which states that a logic implication is true if the RHS is at least as true as the LHS. Modeling multi-level interactions As already explained, we use two distinct kinds of predi- cates: given predicates and target predicates. Given pred- icates encode a priori knowledge about the problem, in our case the structure of the multi-level object hierar- chy. In particular, given a protein p and a domain d, the parentpd(p,d) predicate is true if and only if domain d occurs in protein p; the parentdr predicate is the analo- gous for domains and residues. This simple representation suffices to encode the whole protein–domain–residue hierarchy. To simplify the notation, we also introduce the SBR computes a solution to the inference problem, i.e. deciding the truth value of all target predicates, that maxi- mizes both the confidence of individual predicates and the amount of satisfaction of all constraints. Informally, the optimal assignment to all predicates, i.e. the binding state Page 6 of 18 Saccà et al. BMC Bioinformatics 2014, 15:103 http://www.biomedcentral.com/1471-2105/15/103 hasdom(p) predicate to encode the fact that protein p has at least one domain. More formally: RHS is false), then the parent proteins must not be bound (the LHS is false). ( ) Note that, in the above formulation, the rule is applied indiscriminately to all protein pairs, even to those that have no known child domains in the considered dataset. Therefore, the rule can be reformulated in order to enforce it only for those protein pairs that do in fact have child domains, using the hasdom predicate, as follows: Note that, in the above formulation, the rule is applied indiscriminately to all protein pairs, even to those that have no known child domains in the considered dataset. hasdom(p) := ∃d parentpd(p,d) The hasdom predicate can be computed directly by SBR using the above definition; we instead pre-compute its value for all protein pairs for run-time efficiency. The hasdom predicate can be computed directly by SBR using the above definition; we instead pre-compute its value for all protein pairs for run-time efficiency. Therefore, the rule can be reformulated in order to enforce it only for those protein pairs that do in fact have child domains, using the hasdom predicate, as follows: The boundp(p,p’) target predicate models the bind- ing state of two distinct proteins. Its state is known for certain protein pairs, i.e. those in the training set, and our goal is to predict its state on the remaining ones. The boundd(d,d’) predicate plays the same role for domains. Modeling multi-level interactions This modification also removes a bias in the translation of the implication that was affecting the origi- nal formulation of SBR, whose effect is to often move the LHS toward the false value. See Methods for details. Table 1 Predicates Target predicates boundp(p,p’) true iff the protein pair (p,p’) is bound boundd(d,d’) true iff the domain pair (d,d’) is bound boundr(r,r’) true iff the residue pair (r,r’) is bound Given predicates parentpd(p,d) true iff protein p is parent of domain d parentdr(d,r) true iff domain d is parent of residue r parentpr(p,r) true iff protein p is parent of residue r hasdom(p) true iff protein p has at least one domain hasres(d) true iff domain d has at least one residue List of predicates used by SBR. The constraints for domains and residues can be simi- larly defined with one important exception. The P→D rule described above (correctly) requires at least one domain couple to be bound for each interacting protein pair. However, when two domains are bound, the interaction interface involves more than one residue pair: for instance, binding sites collected in the protein–protein docking benchmark version 3.0 [30] consist of 25 residues on aver- age [31]. We integrate this observation in the D→R rule using the n-existential operator ∃n in place of the regu- lar existential (see Table 2 for the complete formulation), so that whenever two domains are bound, at least n pairs Saccà et al. BMC Bioinformatics 2014, 15:103 http://www.biomedcentral.com/1471-2105/15/103 Page 7 of 18 Saccà et al. BMC Bioinformatics 2014, 15:103 http://www.biomedcentral.com/1471-2105/15/103 Saccà et al. BMC Bioinformatics 2014, 15:103 http://www.biomedcentral.com/1471-2105/15/103 Table 2 Rules Name Definition P→D ∀(p,p’) hasdom(p) ∧hasdom(p’) ⇒ boundp(p,p’) ⇒∃(d,d’) boundd(d,d’) ∧parentpd(p,d) ∧parentpd(p’,d’) D→P ∀(p,p’) ∃(d,d’) boundd(d,d’) ∧parentpd(p,d)∧parentpd(p’,d’) ⇒boundp(p,p’) D→R ∀(d,d’) hasres(d) ∧hasres(d’) ⇒ boundd(d,d’) ⇒∃n (r,r’) boundr(r,r’) ∧parentdr(d,r) ∧parentdr(d’,r’) R→D ∀(d,d’) ∃(r,r’) boundr(r,r’) ∧parentdr(d,r) ∧parentdr(d’,r’) ⇒boundd(d,d’) P→R Same as D→R, with proteins in place of domains R→P Same as R→D, with proteins in place of domains List of FOL constraints used by SBR. ∀(p,p’) hasdom(p) ∧hasdom(p’) ⇒ List of FOL constraints used by SBR. Semi-supervised learning (SSL) techniques [28,29] attempt to solve these issues. In the SSL setting the set of target proteins is known in advance, meaning that the learning algorithm has access to their distribution in fea- ture space. Modeling multi-level interactions This way the inference task can be simplified by introducing unsupervised constraints that assign the same label to proteins that are, e.g., close enough in feature space, or linked in the interaction network, instantiat- ing a form of information propagation. There are several works in the PPI literature that embed the known net- work topology using SSL constraints. Qi et al. [34] employ SSL methods to the special case of viral-host protein inter- actions, where supervised examples are extremely scarce. Using similar methods, You et al. [35] attempt to detect spurious interactions in a known network by projecting it on a low-dimensional manifold. Other studies [36,37] applied SSL techniques to the closely related problems of gene–protein and drug–protein interaction prediction. Despite the ability of SSL to integrate topology informa- tion, no study so far has applied it to highly relational problems such as the MLPIP. of their residues must be bound. Since interfaces in the employed dataset are typically 5 residues long, n = 5 has been used in the experiments. Our results demonstrate that this seemingly small modification has a rather exten- sive impact on the prediction of domain and residue level interactions. Related work In this section we briefly summarize previous PPI interac- tion prediction approaches using methods that are most closely related to the present paper: kernel methods, semi-supervised methods, and logic-based methods. For a broader exposition of interaction prediction methods, please refer to one of the several surveys on the subject [4,6,9,32]. The earliest attempt to employ kernel methods [26] for PPI prediction is the work of Bock et al. [33], which casts interaction prediction as pairwise classification, using amino-acid composition and physico-chemical properties alone. Ben-Hur et al. [27] extended the previous work by applying pairwise kernels and combining multiple data sources (primary sequence, Pfam domains, Gene Ontol- ogy annotations and interactions between orthologues). Successive publications focused primarily on aggregat- ing more diverse sources, including phylogenetic pro- files, genetic interactions, and subcellular localization and function [6]. Kernel machines have also been applied to the prediction of binding sites from sequence, as summa- rized in [10]. The appeal of supervised kernel methods is that they provide a proved and theoretically grounded set of techniques that can easily integrate various information sources, and can naturally handle noise in the data. How- ever, they have two inherent limitations: (i) the binding state of two proteins is inferred independently from the state of all other proteins, and (ii) due to their supervised nature, they do not take advantage of unsupervised data, which is very abundant in the biological network setting. An alternative strategy for interaction prediction is Inductive Logic Programming (ILP) [38], a group of logic- based formalisms that extract rules explaining the likely underlying causes of interactions. ILP methods were stud- ied in the work of Tran et al. [39] using a large number of features: SWISS-PROT keywords and enzyme properties, Gene Ontology functional annotations, gene expression, cell cycle and subcellular localization. Further advances in this direction, with a special focus on using domain information, can be found in [17,18]. The advantage of ILP methods over purely statistical methods is that they are inherently able to deal with relational information, making them ideal candidates for solving the MLPIP problem. Alas, contrary to kernel methods, they tend to be very susceptible to noise, which is a very prominent feature of interaction dataset, and are less effective in Page 8 of 18 Saccà et al. BMC Bioinformatics 2014, 15:103 http://www.biomedcentral.com/1471-2105/15/103 exploiting complex feature representations, e.g. involv- ing highly non-linear interactions between continuous features. Related work information between levels, not to enforce consistency on the predictions. In particular, the downward propagation rule is rather arbitrary: it is not clear why all domains of bound proteins should be themselves bound with the same confidence. Finally, these rules, which are intimately tied to the specific implementation, are not defined using a formal language, and are therefore difficult to extend. For instance, it would be difficult to implement in said framework something similar to an n-existential propa- gation rule, which is extremely useful for dealing with residue interactions. Recently, some works highlighted the importance of the multi-level nature of protein–protein interactions. Gonzalez et al. [40] propose a method to infer the residue contact matrix from a known set of protein interactions using SVMs; on the contrary, our goal is to predict the interactions concurrently at all levels of the hierarchy. Another study [13] highlights the relevance of domain- level interactions, and the unfortunate lack of details thereof, and formulates a method to reinterpret a known PPI network in terms of its constituent domain interac- tions; the present work has a different focus and a more general scope. Semantic Based Regularization seems to have many obvious advantages in this context. A first advantage is that it decouples the implementation of the functions from how consistency among levels is defined. Indeed, consistency is implemented via a set of constraints, which are applied over the output of the predictors. However, there is no limitation to which kind of predictors are used. For example, we used kernel machines as basic machinery for implementing the predictor, where different state-of- the-art kernels can be used at the single levels, while still be able to define a single optimization problem. Most relevant to this paper is the work of Yip et al. [21], where the authors propose a procedure to solve the MLPIP problem based on a mixture of different techniques. The idea is to decompose the problem as a sequence of three prediction tasks, which are solved iter- atively. Given an arbitrary order of the three levels (e.g. proteins first, then domains, then residues), their proce- dure involves computing putative interactions in the first level (in this case proteins), then using the most confident predictions as novel training examples at the following level (i.e., domains). The procedure is repeated until a termination criterion is met. Related work Furthermore, SBR allows to natively propagate the pre- dictions of one level to the other levels. Since the pre- dictions and not the supervisions are propagated, SBR accuracy can get advantage of the abundant unsupervised data. The availability of an efficient implementation of the n-existential quantifier is also a crucial advantage: if two proteins or domains are interacting, a small set of residues must be interacting as well. SBR does not simply propagate a generic prior to all the residues for a protein or domain, which could decrease accuracy of the reduc- tions for the negative supervisions. SBR instead performs a search process in order to select a subset of residue can- didates, where to enforce the interaction. As shown in the experimental results, this greatly improves residue pre- diction accuracy. Finally, the circular dependencies that make learning difficult are dealt in the context of a general and well defined framework, which implements various heuristics to make training effective. Intra-level predictions are obtained with Support Vec- tor Regression (SVR) [41]. In particular, each object has an associated SVR machine that models its propensity to bind any other object in the same level. The extrapolated values act as confidences for the predictions themselves. The mechanism for translating the most confident pre- dictions at one level into training examples for the next level depends on the relative position of the two levels in the hierarchy. Downward propagation (e.g. from proteins to domains) simply associates to each novel example the same confidence as the parent prediction: in other words, if two proteins are predicted as bound with high confi- dence, all their domains will be considered bound with the same confidence. Upward propagation (e.g. from domains to proteins) is a bit more involved: the confidence assigned to the novel example (protein) is a noisy-OR combination of confidences for all the involved child objects (domains). While this method has been shown to work reason- ably well, it is afflicted by several flaws. First of all, while the iterative procedure is grounded in co-training [42], the specific choice of components is not as theoreti- cally sound. For instance, the authors apply regression techniques on a classification task, which may lead to sub-optimal results. Related work The inter-level example propagation mechanisms are ad hoc, do not exploit all the informa- tion at each level (only the most confident predictions are propagated), and are designed to merely propagate Intra-level predictions are obtained with Support Vec- tor Regression (SVR) [41]. In particular, each object has an associated SVR machine that models its propensity to bind any other object in the same level. The extrapolated values act as confidences for the predictions themselves. The mechanism for translating the most confident pre- dictions at one level into training examples for the next level depends on the relative position of the two levels in the hierarchy. Downward propagation (e.g. from proteins to domains) simply associates to each novel example the same confidence as the parent prediction: in other words, if two proteins are predicted as bound with high confi- dence, all their domains will be considered bound with the same confidence. Upward propagation (e.g. from domains to proteins) is a bit more involved: the confidence assigned to the novel example (protein) is a noisy-OR combination of confidences for all the involved child objects (domains). Results and discussion Dataset Kernels computed from the individual features were com- bined additively into a single kernel function for each level, and then transformed into pairwise kernels using Equation (1); the resulting functions were used as inputs to SBR. A visualization of the process can be found in Figure 2. SBR has two scalar hyper-parameters that control the contribution of various parts of the objective function: λc is the weight associated to the constraints (how much the current solution is consistent with respect to the rules) and λr, which controls the model complexity (see the Methods section for more details). The λr parameter was optimized on the first fold by training the model with- out the logic rules and it was then kept fixed for all the folds of the k-fold cross-validation. The resulting value is λr = 0.1. The λc parameter has not been optimized and kept fixed at λc = 1. Please note that further significant gains for the proposed method could be achieved by fine- tuning this meta-parameter. However, since the dataset from Yip et al. does not include a validation split, no sound way to optimize this parameter was possible with- out looking at the test set, or redefining the splits (making it difficult to compare against the results of Yip et al.). Therefore, we decided to not perform any tuning for this meta-parameter. This procedure yields a dataset of 1681 proteins, 2389 domains, and 3035 residues, with a gold standard of 3201 positive (interacting) protein pairs, 422 domain pairs, and 2000 residue pairs. Since interaction experiments can not determine which pairs do not interact, the gold standard of negative pairs is built by randomly sampling, at each level, a number of pairs that are not known to interact (i.e. not positive). This is a common approach to negative labeling in the PPI prediction literature [44]. To keep the dataset balanced, the number of sampled negative pairs is identical to the number of objects in the gold standard of positives. For more details on the dataset prepara- tion, please refer to [21]. We further refined the dataset by running CD-HIT [45] with a 20% sequence similar- ity threshold, identifying 23 redundant proteins. These proteins were not used when comparing the method per- formances. We computed three performance metrics: the Receiver Operating Characteristic (ROC) curve, the area under the ROC (AUCROC, or AUC for short), and the F1 score. Results and discussion Dataset In this work we use the dataset of Yip et al. [21], described here for completeness. The dataset represents proteins, domains and residues using features gathered from a variety of different sources: While this method has been shown to work reason- ably well, it is afflicted by several flaws. First of all, while the iterative procedure is grounded in co-training [42], the specific choice of components is not as theoreti- cally sound. For instance, the authors apply regression techniques on a classification task, which may lead to sub-optimal results. The inter-level example propagation mechanisms are ad hoc, do not exploit all the informa- tion at each level (only the most confident predictions are propagated), and are designed to merely propagate • Protein features include phylogenetic profiles derived from COG, subcellular localization, cell cycle and environmental response gene expression; protein-pair features were extracted from Y2H and TAP-MS data. The gold standard of positive interactions was constructed by aggregating Saccà et al. BMC Bioinformatics 2014, 15:103 http://www.biomedcentral.com/1471-2105/15/103 Page 9 of 18 Saccà et al. BMC Bioinformatics 2014, 15:103 http://www.biomedcentral.com/1471-2105/15/103 Page 9 of 18 the comparison completely fair, we repeated said pro- cedure with SBR, reusing the very same train/test splits. Since correlated objects, e.g. a protein and its domains/residues, share information, the folds were structured as to avoid such information to leak between train and test folds: this was achieved by keeping cor- related objects in the same fold. In order not to bias the performance estimates, all redundant proteins were ignored, along with their domains and residues, when computing the results of both SBR and the method of Yip et al. The full experimental setup and instructions to replicate the experiments can be downloaded at http://sites.google.com/site/semanticbasedregularization/ home/software/protein_interaction. experimentally verified or structurally determined interactions taken from MIPS, DIP, and iPfam. • At the domain level, the dataset includes both features for domain families and for domain instances based on frequencies of domains within one or more species and phylogenetic correlations of Pfam alignments. The gold standard of positive interactions was built from 3D structures of complexed proteins taken from PDB. • Residue features consist of sequence-based properties, namely charge complementarity, Psi-Blast [43] profiles, predicted secondary structure, and predicted solvent accessibility. Results and discussion Dataset The ROC curve represents the relation between the false pos- itive rate (FPR) and the true positive rate (TPR), and can be seen as the proportion of true positives gained by “pay- ing” a given proportion of false positives. By definition, the ROC curve is monotonically non-decreasing; the steeper the curve, the better the predictions. The AUC measures the ability to correctly discriminate between positives and negatives, or alternatively, the ability to rank positives above negatives. It is independent of any classification threshold, and thus particularly fit to evaluate models over the whole spectrum of possible decision thresholds. The F1 score is the harmonic mean of precision and recall. Contrary to the AUC, the F1 takes into account the predicted label, but not its confidence (margin). Turning our attention to the resulting dataset, we note that most of the supervision is located at the protein level: out of all possible interactions between pairs of proteins, which are (1681×1680)/2, 0.226% are known (either pos- itive or negative). On the contrary, the other levels hold much less information: only 0.042% of all possible residue pairs, and 0.014% of all possible domain pairs, are in the dataset. The low number of residue pairs is due to i) dif- ferent requirements for experimentally determining the interactions at the three levels, i.e. whether the structure is available; and ii) sampling choices operated by Yip et al. [21]. Evaluation procedure We computed the average AUC and F1 of our method and those of our competitor over all folds of the cross- validation; the results can be found in Table 3 and Table 4. The ROC curves have been computed by collating the In this work we compare our method to that of Yip et al. [21], where the authors evaluated their method using a 10-fold cross-validation procedure. To keep Saccà et al. BMC Bioinformatics 2014, 15:103 http://www.biomedcentral.com/1471-2105/15/103 Saccà et al. BMC Bioinformatics 2014, 15:103 http://www.biomedcentral.com/1471-2105/15/103 Page 10 of 18 Table 3 Results (AUC) Independent Unidirectional Bidirectional Full Level P→D D→R P→R P↔D D↔R P↔R Results for Yip et al. [21] Proteins 0.723 0.722 0.725 0.724 Domains 0.531 0.619 0.688 0.695 0.673 Residues 0.563 0.542 0.549 0.576 0.659 0.722 Results for SBR Proteins 0.808 0.820 0.819 0.820 Domains 0.605 0.814 0.837 0.896 0.937 Residues 0.591 0.664 0.671 0.675 0.673 0.676 Results for SBR-∃n Proteins 0.808 0.820 0.819 0.821 Domains 0.605 0.814 0.837 0.895 0.956 Residues 0.591 0.745 0.760 0.778 0.772 0.797 Area under the ROC curve values attained by Yip et al. [21], SBR, and SBR-∃n (SBR equipped with the n-existential quantifier). experiments using rules of increasing complexity. This setup follows closely that of Yip et al. [21]. results of all test folds, and can be found in Figure 3. Since the ROC and F1 are not present in [21], and the dataset is slightly smaller because of the redundancy elimination step we introduced, we had to compute their results on a local re-run of their experiment. As a result, the AUC val- ues presented in Table 3 are slightly different from those reported in [21]. However, we note that the results of our analysis would still apply if we had chosen to use the AUC values reported in [21]. Independent levels As a baseline, we estimate the performance of our method when constraints are ignored. This is equivalent to the method of Yip et al. when no information flow between levels is allowed. The results can be found in the “Inde- pendent” column of Tables 3 and 4. In absence of constraints SBR reduces to standard ℓ2- regularized SVM classification: learning and inference become convex problems, and the method computes the globally optimal solution. Thus, the only differences Results To evaluate the effects of the constraints on the per- formances of SBR, we performed three independent To evaluate the effects of the constraints on the per- formances of SBR, we performed three independent Table 4 Results (F1) Independent Unidirectional Bidirectional Full Level P→D D→R P→R P↔D D↔R P↔R Results for Yip et al. [21] Proteins 0.665 0.665 0.666 0.666 Domains 0.518 0.620 0.662 0.659 0.661 Residues 0.522 0.510 0.514 0.602 0.609 0.613 Results for SBR Proteins 0.718 0.722 0.722 0.723 Domains 0.568 0.693 0.696 0.731 0.750 Residues 0.579 0.605 0.605 0.605 0.605 0.602 Results for SBR-∃n Proteins 0.717 0.722 0.722 0.722 Domains 0.568 0.693 0.696 0.729 0.757 Residues 0.579 0.635 0.639 0.641 0.644 0.650 F1 values attained by Yip et al. [21], SBR, and SBR-∃n (SBR equipped with the n-existential quantifier). Saccà et al. BMC Bioinformatics 2014, 15:103 http://www.biomedcentral.com/1471-2105/15/103 Page 11 of 18 Figure 3 ROC curves. ROC curves obtained with the 10-fold cross-validation procedure, for all experimental settings and all levels of the hierarchy. (Left) Results for Yip et al. (Right) Results for SBR-∃n. (Top) ROC curves for protein-level predictions with different sets of constraints, from fully independent to fully connected levels. (Middle) Domain-level predictions. (Bottom) Residue-level predictions. Each plot includes multiple ROC curves, one for each experimental setting; see the legends for more details. Figure 3 ROC curves Figure 3 ROC curves. ROC curves obtained with the 10-fold cross-validation procedure, for all experimental settings and all levels of the hierarchy. (Left) Results for Yip et al. (Right) Results for SBR-∃n. (Top) ROC curves for protein-level predictions with different sets of constraints, from fully independent to fully connected levels. (Middle) Domain-level predictions. (Bottom) Residue-level predictions. Each plot includes multiple ROC curves, one for each experimental setting; see the legends for more details. In the P→D case only the P→D rule is active, meaning that bound protein pairs enforce positive domain pairs and negative domain pairs enforce negative protein pairs. The D→R and P→R cases are defined similarly. In all three cases, the level not appearing in the rule (e.g. the residue level in the P→D case) is predicted independently. This setup makes it easy to study the effects of propa- gating information from one level to the other without interferences. The results can be found in the “Unidirec- tional” column of Tables 3 and 4. In the same column we also show the results for Yip et al. Results for the unidirec- tional flow setting, where examples are propagated from one level to the next but not vice versa. However, since between our method and the competitor are: (i) using classification versus regression, and (ii) using pairwise classification, instead of training a single model for each entity (protein, domain, residues). These differences alone produce a substantial increase in performance: the F1 changes by about +0.05 in all three cases. The AUC of proteins and domains is improved by about +0.09 and +0.07, respectively, while residues are less affected, with a +0.03 difference. between our method and the competitor are: (i) using classification versus regression, and (ii) using pairwise classification, instead of training a single model for each entity (protein, domain, residues). These differences alone produce a substantial increase in performance: the F1 changes by about +0.05 in all three cases. The AUC of proteins and domains is improved by about +0.09 and +0.07, respectively, while residues are less affected, with a +0.03 difference. Bidirectional constraints In the third experiment we study the impact of using bidirectional constraints between pairs of levels; the level not appearing in the rules is predicted independently, as above. In the P↔D case, both the P→D and D→P rules are active, meaning that the protein and domain levels are enforced to be fully consistent; the P↔R and D↔R cases are defined analogously. This experiment is comparable to the bidirectional flow setting of Yip et al.. The results can be found in the “Bidirectional” column of Tables 3 and 4. In this experiment, protein predictions are stable with respect to the previous experiments, confirming the intu- ition that the abundance of supervision at this level makes it less likely to benefit from predictions at the other ones. On the contrary, domains see a large performance upgrade, from 0.896 to 0.937 AUC and from 0.731 to 0.750 F1, when made to interact with both proteins and residues. The change for residues is instead only marginal. We observe that the new constraints have a positive effect on predictions at all three levels: proteins change from 0.808 AUC to 0.820, domains from 0.814 to 0.896 and residues from 0.671 to 0.673. In terms of F1, the changes are from 0.718 to (up to) 0.722 for proteins, from 0.693 to 0.731 for domains, and no change for residues. The change is not as marked as between the independent and unidirectional experiments. In particular, domains see the largest increase in performance (+0.08 AUC, +0.04 F1), in particular thanks to the contribution of residue-level information, which is more abundant. Pro- teins and residues are less affected. The result is unsur- prising for proteins, which hold most of the supervision and are thus (i) more likely to be predicted correctly in the independent setting, and (ii) less likely to be assisted from hints coming from the other, less supervised levels. The results for Yip et al. are mixed, with proteins faring almost identically to the previous experiment, domains showing a slight drop in AUC but an equally slight increase in F1, and residues improving in AUC (+0.08) but not in F1 (unchanged) over the bidirectional P↔R case. The improvement in residue prediction (in terms of AUC) stands in contrast with the results of SBR, and is the only case in which the method of Yip and col- leagues works better than SBR. All constraints h f l In the final experiment we activate the P→D, D→P, D→R and R→D rules, as defined in Table 2, making all levels interact. This is the most complex setting, and produces fully consistent predictions through the hierarchy. It is comparable to the “PDR” bidirectional setting of Yip et al.. The AUC scores can be found in column “Full” of Tables 3 and 4. Compared to SBR, the method of Yip et al. does not benefit as much from unidirectional information flow. Protein-level information allows to improve domain pre- dictions only (+0.1 F1, +0.06 AUC for P→D), while residue predictions are worse than in the independent case (−0.05 and −0.04 AUC, and −0.01 F1, in the D→R and P→R cases, respectively). In this experiment the P→R and R→P constraints are not used. Direct information flow between proteins and residues is not needed, because it would be redundant: from a formal logic point of view, this corresponds to the observation that the logic rule expressing protein to residue consistency is implied by the other consistency rules. Indeed, we have experimentally verified that adding this propagation flow does not significantly affect the results. Unidirectional constraints In the second experiment, we evaluate the effect of intro- ducing unidirectional constraints between pairs of levels. Page 12 of 18 Saccà et al. BMC Bioinformatics 2014, 15:103 http://www.biomedcentral.com/1471-2105/15/103 the competitor’s algorithm is iterative, information about lower levels can indeed affect the upper levels in succes- sive iterations. performance increase, SBR is able to largely outperform the competitor in all configurations except one (F1 of the P↔R case for residues). The results show that introducing unidirectional con- straints in SBR improves the predictions in all cases. In particular, using (predicted and known) protein interac- tions helps inferring correct domain interactions, which improve by about +0.13 F1 and +0.2 AUC (P→D case). Residues improve independently of whether protein or domain-level information is used, with a +0.03 F1 in both cases, and a +0.08/+0.07 AUC difference, respec- tively. Interestingly, proteins tend to help residue predic- tions slightly more than domains, despite the indirection between the two levels; this is likely an effect of the larger percentage of supervised pairs available. We note that the fact that all three cases (P↔D, P↔R and D↔R) improve over both the independent and the unidirectional experiments shows that not only the bidi- rectional constraints are in fact sound, but also that, despite the increased computational complexity, SBR is still able to exploit them appropriately. Bidirectional constraints The issue lies within our formulation of the D→R rule: whenever two domains are bound, the rule is satisfied when at least one residue pair is bound. As already mentioned above, this is not realistic: protein interfaces span more than two residues, typically five or more. We therefore extended SBR to support the n-existential quantifier, which allows to reformulate the D→R rule to take this observation into account (see the Methods section for more details on the n-existential quantifier). The new D→R rule, shown in Table 2, requires for each pair of bound domains at least n = 5 residues As for the method of Yip et al., the bidirectional flow mostly affects the domain and residue levels, whose improvement is +0.07 AUC/+0.04 F1 and +0.11 AUC/+0.09 F1, respectively; the change for protein interactions is negligible. Regardless of the relative Saccà et al. BMC Bioinformatics 2014, 15:103 http://www.biomedcentral.com/1471-2105/15/103 Saccà et al. BMC Bioinformatics 2014, 15:103 http://www.biomedcentral.com/1471-2105/15/103 Page 13 of 18 Page 13 of 18 to be bound. We chose the constant n = 5 to be both realistic and, since the computational cost increases with n, small enough to be easily computable. We applied the same modification to the P→R rule. regression models of Yip et al., a considerable improve- ment was achieved even in the unconstrained experiment. Furthermore, when enforcing consistency among the pro- tein, domain and residue levels and using the n-existential quantifier, the experimental results are significantly better than both the unconstrained baseline and the correspond- ing results of Yip and colleagues, at all levels and in all experimental settings. The complete results for the resulting method, termed SBR-∃n, can be found at the bottom of Tables 3 and 4. When comparing to standard SBR, i.e., without the n-existential, we see that the performance of residues consistently improves in all cases (unidirectional, bidirec- tional, and with all constraints activated), allowing SBR-∃n to always outperform the method of Yip et al. by a sig- nificant margin also on the residue interactions. As a side effect of the better residue predictions, thanks to the D→R and R→D rules domains also improve in the all- constraints experiment. In particular, in the “Full” exper- iment the AUC improvement of SBR-∃n over Yip et al. is +0.1/+0.26/+0.07 AUC and +0.06/+0.1/+0.04 F1 for proteins, domains and residues respectively. Bidirectional constraints We show in Figure 4 an example prediction obtained by SBR-∃n for the VPS25 and VPS36 ESCRT-II complex subunits. The figure shows that, while the unconstrained (baseline) pre- dictions are inconsistent, the addition of the constraints effectively makes the protein- and domain-level predic- tions correct and consistent, and enables SBR to improve the residue-level predictions. It is worth noting that SBR performance improves monotonically with the increase of constraint complexity in the reported experiments. This result is far from obvi- ous, and confirms both that the biologically-motivated knowledge base is useful, and that SBR is able to effec- tively apply it. In contrast, the competitor’s method does not always improve in a similar manner. In general, the performance gain brought forth by inter-level propagation is not homogeneously distributed between the three levels. We register a large improvement for domains and residues, especially when SBR is used in conjunction with the n-existential quantifier. Proteins are less affected by consistency enforcement, most likely due to the availability of more supervised examples. We note that the FOL rules have a twofold effect. Firstly, they propagate information between the levels, enabling predicted interactions at one level to help inferring cor- rect interactions at the other two levels. This is especially clear in the “Full” experiment with the n-existential quan- tifier: in this case, better residue level predictions increase the overall quality of domain predictions as well. Secondly, the rules also guarantee that the predictions are consistent along the object hierarchy. Summing up, these results highlight the ability of SBR to enforce constraints even with highly complex combi- nations of rules, allowing the modeler to fully exploit the flexibility and performance improvement offered by non- standard FOL extensions like the n-existential operator. Discussion Summarizing, SBR is able to largely outperform that of Yip and colleagues, and moreover enforces the predictions to be consistent among levels. As previously mentioned, the data taken from Yip et al. has some peculiarities worth discussing. First, it contains a low number of residue– residue interactions, partially due to design choices taken The results presented in the previous section offer a clear perspective on the advantages of the proposed method. By employing appropriate classification techniques and training a single global pairwise model per level, rather than relying on the less than optimal local (per-object) YJR102C PF05871 00022 00029 00083 YLR417W PF04132 00543 00546 00562 00548 00555 YJR102C PF05871 00022 00029 00083 YLR417W PF04132 00543 00546 00562 00548 00555 INDEPENDENT FULL Figure 4 Example prediction. Prediction for the interaction between two ESCRT-II complex subunits: VPS25 (YJR102C) and VPS36 (YLR417W). The two proteins, their domains, and all the residue pairs in the dataset, are known to interact. Solid black lines indicate a predicted interaction, dotted lines a non-interaction; residue pairs not connected by either a solid or dotted line are not present in the dataset. (Left) SBR-∃n predictions with no constraints: the predictions at the three levels are inconsistent. (Right) SBR-∃n predictions with the full set of constraints: the protein- and domain-level predictions are now both consistent and correct. A further residue pair is now correctly predicted as interacting. Figure 4 Example prediction. Prediction for the interaction between two ESCRT-II complex subunits: VPS25 (YJR102C) and VPS36 (YLR417W). The two proteins, their domains, and all the residue pairs in the dataset, are known to interact. Solid black lines indicate a predicted interaction, dotted lines a non-interaction; residue pairs not connected by either a solid or dotted line are not present in the dataset. (Left) SBR-∃n predictions with no constraints: the predictions at the three levels are inconsistent. (Right) SBR-∃n predictions with the full set of constraints: the protein- and domain-level predictions are now both consistent and correct. A further residue pair is now correctly predicted as interacting. Figure 4 Example prediction. Prediction for the interaction between two ESCRT-II complex subunits: VPS25 (YJR102C) and VPS36 (YLR417W). The two proteins, their domains, and all the residue pairs in the dataset, are known to interact. Solid black lines indicate a predicted interaction, dotted lines a non-interaction; residue pairs not connected by either a solid or dotted line are not present in the dataset. Kernel machines Machine learning and statistical methods are very well defined for the linear case, and statistical learning the- ory can provide optimal solutions in terms of gen- eralization performance. Unfortunately, non-linearity is often required in order to solve most applications, where exploiting complex dependencies is essential to predict some higher level property of the objects. Kernel meth- ods try to combine the potential classification power of non-linear methods and the optimality and computational efficiency of linear methods by mapping the input patterns into a high dimensional feature space, where parameter optimization remains linear. SBR is a flexible framework to inject domain knowledge into kernel machines. In this paper SBR has been used to tie together protein, domain and residue interaction predictions tasks. In particular, the domain knowledge expresses that two proteins interact if and only if there is an interaction between at least one pair of domains of the proteins. Similarly two domains can interact if and only if there are at least some residues interacting. While these tasks could be learned separately, tying them together has multiple advantages. First the predictions will be consis- tent and more accurate, as the predictions at one level will help the predictions at the other levels. Secondly, the domain knowledge can be enforced also on the unsu- pervised data (proteins, domains and residues for which interactions are unknown). Unsupervised data is typically abundant in protein interaction prediction tasks but often neglected. This methodology allows to powerfully lever- age it, significantly improving the prediction accuracy. Note also that, while the resolved complexes are required during the training stage, no structural information is required for performing inference on novel proteins. Kernel methods have a wide range of applications in many fields, and can be used for many different tasks like regression, clustering and classification (the latter being the main interest of this paper). In particular, the repre- senter Theorem [46] shows that a large class of problems admits solutions in terms of kernel expansions having the following form: f (x) = N i=1 wiK(x, xi) (2) (2) where x is the representation of the pattern, K(x, xi) =< (x), (xi) > is a kernel function, where (·) is some mapping from the input space to the feature space. Discussion (Left) SBR-∃n predictions with no constraints: the predictions at the three levels are inconsistent. (Right) SBR-∃n predictions with the full set of constraints: the protein- and domain-level predictions are now both consistent and correct. A further residue pair is now correctly predicted as interacting. Page 14 of 18 Page 14 of 18 Saccà et al. BMC Bioinformatics 2014, 15:103 http://www.biomedcentral.com/1471-2105/15/103 binding state of consecutive residues, which are likely to share the same state. More ambitious goals, requiring a redesign of the experimental dataset, include encoding selected information sources, such as domain types, sub- cellular co-localization and Gene Ontology annotations, as First Order Logic constraints rather than with kernels, to better leverage their relational nature. by [21]. Second, it is artificially balanced by including an appropriate number of non-interactions, while in a real- world case all possible pairs would qualify as candidates. We decided to keep the dataset as-is in order to facili- tate a fair comparison with Yip et al.. We postpone further analysis with other datasets to future work. Methods In this work we tackle the multi-level protein interaction prediction (MLPIP) problem, first introduced by Yip et al. [21], which requires to establish the binding state of all uncharacterized pairs of proteins, domains and residues. Contrary to standard protein–protein interaction predic- tion, the MLPIP problem offers many advantages and opens up new challenges. The primary contribution of this paper is the extension and application to the MLPIP task of a state-of-the-art statistical relational learning tech- nique called Semantic Based Regularization. Kernel machines Intuitively, the kernel function measures the similarity between pairs of instances, and the prediction f (x) for a novel instance is computed as a weighted similarity to training instances xi. There is a large body of literature on kernel machines, see e.g. [26] for an introduction. While other work in the literature has exploited the possibility of tying the predictions at multiple levels, the presented methodology employs a more principled infer- ence process among the levels, where the domain knowl- edge can be exactly represented and precisely enforced. The experimental results confirm the theoretical advan- tages by showing significant improvements in domain and residue interaction prediction accuracy both with respect to approaches performing independent predictions and the only previous approach attempting at linking the pre- diction tasks. The optimization of the weights wi of the Kernel machine can be formulated in different ways. Let us indi- cate yj ∈{−1, +1} the desired output for pattern xi, w = [w1, . . . , wn] is a vector arranging the kernel machine parameters and G is the gram matrix, having its (i, j) ele- ment defined as Gij = K(xi, xj). \|\|f \|\|2 = wtGw, and it can be shown that the following cost function: \|\|f 2\|\| + λ N j=1 L(yj, f (xj)) Given the flexibility offered by SBR, the proposed method can be extended in several ways. The simplest extension involves engineering a more refined rule set, for instance by introducing (soft) constraints between the reduces to the formulation of hard margin ℓ2SVMs [23] if L(·) is the hinge loss and λ →∞. A very similar cost Saccà et al. BMC Bioinformatics 2014, 15:103 http://www.biomedcentral.com/1471-2105/15/103 Page 15 of 18 Saccà et al. BMC Bioinformatics 2014, 15:103 http://www.biomedcentral.com/1471-2105/15/103 function has been employed to solve the protein, domain, residue interaction presented in this paper. function has been employed to solve the protein, domain, residue interaction presented in this paper. Let us consider a multitask learning problem, where each task works on an input domain where labeled and unlabeled examples are sampled from. For example, in the case study presented in this paper, three separate tasks for protein, domain and residue interaction need to be conjunctively learned. Each input pattern is described via some representation that is relevant to solve the tasks at hand. First-order logic The following FOL clause can be used to express that any protein is either an enzyme or it is not: ∀x Protein(x) ⇒ Enzyme(x) ∨NonEnzyme(x). Variables and quantifiers can be combined. For exam- ple, given the predicates Protein(x) holding true if x is a protein and ResidueOf(x, y) holding true if y is a residue of x, the following clause expresses the fact that any protein has a at least one residue: ∀x Protein(x) ⇒ ∃y ResidueOf(x,y). For example, let x be a variable and let Protein(x), Enzyme(x), NonEnzyme(x) indicate three predicates expressing whether, given a grounding x=PDB1a3a, PDB1a3a is a protein, an enzyme, a non-enzyme, respec- tively. The following FOL clause can be used to express that any protein is either an enzyme or it is not: ∀x Protein(x) ⇒ Enzyme(x) ∨NonEnzyme(x). Therefore, using kernel expansions, Equation 3 becomes: Variables and quantifiers can be combined. For exam- ple, given the predicates Protein(x) holding true if x is a protein and ResidueOf(x, y) holding true if y is a residue of x, the following clause expresses the fact that any protein has a at least one residue: ∀x Protein(x) ⇒ ∃y ResidueOf(x,y). λr T k=1 w′ kGkwk + T k=1 L Gkwk, yk + + λc H h=1 φh(G1w1, . . . , GTwT), Kernel machines Let us indicate with T the total number of tasks, where task k is implemented by a function fk, which lives in an appropriate Reproducing Kernel Hilbert Space. In the following, f = [ f1, . . ., fT]′ indicates the vector col- lecting all task functions. The basic assumption of SBR is that the task functions are correlated as they have to meet a set of constraints that can be expressed by the functionals φh : H1 × . . . × HT →[ 0, +∞) such that φh(f ) = 0 h = 1, . . . , H must hold for any valid choice of fk ∈Hk, k = 1, . . . , T. Following the classical penalty approach for constrained optimization, the constraints are embedded by adding a term that penalizes their violation: First-order logic Propositional logic is based on the basic concept of propo- sitions, which can assume either a true or false value. It is possible to perform operations on the propositions by connecting them via the and (∧), or (∨) and not (¬) oper- ators. In particular, given two propositions A, B, it holds that: A ∧B is true iff A = true, B = true, A ∨B is false iff A = false, B = false and ¬A flips the current truth value of A. The operator ⇒can be used to express a conditional statement: A ⇒B expresses the fact that B must hold true if A is true. The sentence A ⇒B is false iff A = true and B = false, and it can be expressed in terms of other operators through the equivalence A ⇒B ≡¬A ∨B. First-order logic (FOL) extends propositional logic to compactly express generic properties for a class of objects, thanks to the use of predicates, variables and quantifiers. A variable can assume as value any object in some con- sidered domain. A variable is said to be grounded once it is assigned a specific object. A predicate is a function that, taking as input some objects (or grounded variables), returns either true or false. Predicates can be connected with other predicates using the same operators defined for propositional logic. The universal quantifier (∀) expresses the fact that some proposition is true for any object, while the existential quantifier (∃) expresses the fact that some proposition is true for at least one object. λr T k=1 \|\|fk\|\|2 + T k=1 xj k,yj k ∈Lk L fk(xj k), yj k + + λc H h=1 φh(S, f ), (3) (3) where L(·) is a loss function measuring the distance of the function output from the desired one and S is the considered sample of data points over which the func- tions are evaluated. In the experimental setting, L(·) has been set to be the hinge function. It is possible to extend the Representer Theorem to show that the best solution for Equation 3 can be expressed as a kernel expansion as showed in Equation 2 [22]. For example, let x be a variable and let Protein(x), Enzyme(x), NonEnzyme(x) indicate three predicates expressing whether, given a grounding x=PDB1a3a, PDB1a3a is a protein, an enzyme, a non-enzyme, respec- tively. T-norms T-norms [47] are commonly used in fuzzy logic [48] to generalize propositional logic expressions to real valued functions of continuos variables. A continuous t-norm is a function t :[0, 1] ×[0, 1] →R, that is continuous, com- mutative, associative, monotonic, and featuring a neutral element 1 (i.e. t(a, 1) = a). A t-norm fuzzy logic is defined by its t-norm t(a1, a2) that models the logic AND, while the negation of a variable ¬a is computed as 1 −a. Once defined the t-norm functions corresponding to the logi- cal AND and NOT, these functions can be composed to convert any arbitrary logic proposition into a continuous function. Many different t-norm logics have been pro- posed in the literature. For example, the product t-norm used in the experimental section: Quantifiers are also converted into real value operators. The universal quantifier corresponds to the sum of the degrees of violation of the continuous expression coming from t-norms over all possible groundings for the quanti- fied variable. Let us consider a universally quantified FOL formula ∀v E(v, P). When considering the real–valued mapping tE(f , x) of the original boolean expression where x is the representation of v, the universal quantifier can be converted measuring the degree of non-satisfaction of the expression over the domain S of x: (a1 ∧a2) mapped −→ t(a1, a2) = a1 · a2 (¬a1) mapped −→ t(a1) = 1 −a1 (a1 ∨a2) mapped −→ t(a1, a2) = a1 + a2 −a1 · a2 (a1 ∧a2) mapped −→ t(a1, a2) = a1 · a2 (¬a1) mapped −→ t(a1) = 1 −a1 (a1 ∨a2) mapped −→ t(a1, a2) = a1 + a2 −a1 · a2 ∀v E(v, P) mapped −→ φ(f , S) = x∈S 1 −tE(f , x) For example, the formula ∀v A(v) ∧B(v) corresponds to: Please note that the t-norm behaves as classical logic when the variable approaches the value 0 (false) or 1 (true). φ(f , S) = x∈S 1 −fA(x)fB(x) φ(f , S) = x∈S 1 −fA(x)fB(x) The equivalence a1 ⇒a2 ≡¬a1 ∨a2 can be used to represent implications (modus ponens) before perform- ing t-norm conversion. However, this process does not powerfully capture the inference process performed in a probabilistic or fuzzy logic context. Any t-norm has a corresponding binary operator ⇒called residuum, which where fA(x)fB(x) is the t-norm generalization of the propositional expression A(v)∧B(v) for a given grounding of v. Translation of first-order logic clauses into real-valued constraints With no loss of generality, we restrict our attention to FOL clauses in the PNF form, where all the quantifiers (∀, ∃) and their associated quantified variables are placed at the beginning of the clause. For example: Quantifiers ∀v1∀v2 Quantifier−free expression A(v1) ∧B(v2) ⇒C(v1) (4) For example, the quantifier-free expression in Equation 4 corresponds to: (4) 1 fA(x1) · fB(x2) ≤fC(x1) fB(x2) else (5) Please note that the quantifier-free part of the expres- sion is equivalent to an assertion in propositional logic for any given grounding of the quantified variables. As stud- ied in the context of fuzzy logic and symbolic AI, different methods can be used for the conversion of a propositional expression into a continuous function with [0, 1] input variables. (5) where the predicates have been substituted by the unknown corresponding functions and x1, x2 are the rep- resentations of the objects identified by the grounded variables v1, v2, respectively. The representation of the object must be compatible with what is accepted as input by the kernels used by the predicate approximations fA, fB. For example, they can be a vector of real valued features when using a linear or Gaussian kernel, or graph or tree representations when using kernel for structures. It is also possible to use the methodology when no explicit repre- sentations are known, but only the kernel values for each pair of input objects. Semantic-based regularization Semantic Based Regularization (SBR) [22] is a general framework for injecting prior knowledge expressed in FOL into kernel machines for semi-supervised learning tasks. The prior knowledge is converted into a set of con- tinuous constraints, which are enforced during training. The SBR framework is very general and allows to employ the full expressiveness of FOL in the definition of the prior knowledge. The SBR framework also allows to perform collective classification on the test set, in order to enforce the output to respect the logic knowledge. where Gk, wk, f k = Gkwk and yk are the gram matrix, the weights, the function values over the data sample and the desired output column vectors for the patterns in the domain of the k-th task. Evidence tasks do not need to be approximated as their are fully known. Optimization of the wk parameters for the cost function in Equation 4 can be done using gradient descent. Con- straint φh are non-linear in most interesting cases like the one presented in this paper. Therefore, the cost function can present multiple local minima, making optimization Saccà et al. BMC Bioinformatics 2014, 15:103 http://www.biomedcentral.com/1471-2105/15/103 Saccà et al. BMC Bioinformatics 2014, 15:103 http://www.biomedcentral.com/1471-2105/15/103 Page 16 of 18 Page 16 of 18 is used in fuzzy logic to generalize implications in case of continuous variables. In particular, for a minimum t-norm, it holds that the residuum converting an implica- tion is defined as: difficult. SBR uses a two-step heuristic to solve this prob- lem: first it computes the theoretically global optimum for all predicates independently (setting λc = 0), which are convex kernel machines. Then, it introduces the con- straints and proceeds to find a good solution using a gradient descent. (a1 ⇒a2) mapped −→ t(a1, a2) = 1 a1 ≤a2 a2 a1 > a2 In the following we show how to express first order logic clauses in terms of constraints φh. The residuum allows to relax the condition of satisfac- tion for the implication: an implication is satisfied if the right end side of the implication is more verified than the pre-condition on the left side. This makes the fuzzy or probabilistic inference process easier and better defined. While the original SBR formulation represents implications using modus ponens, the minimum t-norm residuum has been used in the experimental section of this paper to convert implications. References K k O Cary MP, Bader GD, Sander C: Pathway information for systems biology. FEBS Lett 2005, 579(8):1815–1820. h k h k h φ(f , S) = x1∈S1 x2∈S2 1 −fA(x1)(1 −fB(x2)) The existential quantifier is mapped into the continuos domain as: ∃v E(P, v) mapped −→ φ(f , S) = min x∈S 1 −tE(f , x) This framework also allows a natural definition of the ∃n operator, generalizing the existential operator to n objects. This operator is usually defined in description logic, while it can only indirectly defined in FOL. This operator will be used in the experimental section and its continuous mapping is defined as: This framework also allows a natural definition of the ∃n gy 9. Shoemaker BA, Panchenko AR: Deciphering protein–protein interactions. Part II. Computational methods to predict protein and domain interaction partners. PLoS Comput Biol 2007, 3(4):e43. 10. Ezkurdia I, Bartoli L, Fariselli P, Casadio R, Valencia A, Tress ML: Progress and challenges in predicting protein–protein interaction sites. Brief Bioinform 2009, 10(3):233–246. 11. Fields S, Song O: A novel genetic system to detect protein protein interactions. Nature 1989, 340(6230):245–246. 11. Fields S, Song O: A novel genetic system to detect protein protein interactions. Nature 1989, 340(6230):245–246. ∃nv E(v, P) mapped −→ φ(f , S) = arg maxnx∈Sn 1−tE(f , x) 12. Shoemaker BA, Panchenko AR: Deciphering protein–protein interactions. Part I. experimental techniques and databases. PLoS Comput Biol 2007, 3(3):e42. 12. Shoemaker BA, Panchenko AR: Deciphering protein–protein interactions. Part I. experimental techniques and databases. PLoS Comput Biol 2007, 3(3):e42. arg maxnx∈Sn 13. Memiševiˇc V, Wallqvist A, Reifman J: Reconstituting protein interaction networks using parameter-dependent domain-domain interactions. BMC Bioinformatics 2013, 14:154. 13. Memiševiˇc V, Wallqvist A, Reifman J: Reconstituting protein interaction networks using parameter-dependent domain-domain interactions. BMC Bioinformatics 2013, 14:154. where argmaxnx ∈Sn indicates the n assignments of x that maximize the value of 1 −tE(·) over the set S. The conversion of the ∃n operator consistently reduces to the ∀conversion when n = \|S\|, and to the conversion of the ∃ operator when n = 1. 14. p g The authors declare that they have no competing interests. 18. Nguyen TP, Ho TB: An integrative domain-based approach to predicting protein–protein interactions. J Bioinform Comput Biol 2008, 6(06):1115–1132. 18. Nguyen TP, Ho TB: An integrative domain-based approach to predicting protein–protein interactions. J Bioinform Comput Biol 2008, 6(06):1115–1132. The authors declare that they have no competing interests. References K k O Radivojac P, Clark WT, Oron TR, Schnoes AM, Wittkop T, Sokolov A, Graim K, Funk C, Verspoor K, Ben-Hur A, Pandey G, Yunes JM, Talwalkar AS, Repo S, Souza ML, Piovesan D, Casadio R, Wang Z, Cheng J, Fang H, Gough J, Koskinen P, Törönen P, Nokso-Koivisto J, Holm L, Cozzetto D, Buchan DW, Bryson K, Jones DT, Limaye B, et al.: A large-scale evaluation of computational protein function prediction. Nat Methods 2013, 10(3):221–227. As a final example of the conversion procedure, let’s refer to the FOL clause in Equation 4, which is converted into the real-valued constraint φ(f , S): 15. Jiang JQ, Wu M: Predicting multiplex subcellular localization of proteins using protein-protein interaction network: a comparative study. BMC Bioinformatics 2012, 13(Suppl 10):S20. 15. Jiang JQ, Wu M: Predicting multiplex subcellular localization of proteins using protein-protein interaction network: a comparative study. BMC Bioinformatics 2012, 13(Suppl 10):S20. 16. Zhao XM, Wang RS, Chen L, Aihara K: Uncovering signal transduction networks from high-throughput data by integer linear programming. Nucleic Acids Res 2008, 36(9):e48–e48. 16. Zhao XM, Wang RS, Chen L, Aihara K: Uncovering signal transduction networks from high-throughput data by integer linear programming. Nucleic Acids Res 2008, 36(9):e48–e48. x1∈S1 x2∈S2 0 fA(x1) · fB(x2) ≤fC(x1) 1−fB(x2) else x1∈S1 x2∈S2 0 fA(x1) · fB(x2) ≤fC(x1) 1−fB(x2) else 17. Nguyen TP, Ho TB: Discovering signal transduction networks using signaling domain-domain interactions. Genome Inform 2006, 17(2):35–45. 17. Nguyen TP, Ho TB: Discovering signal transduction networks using signaling domain-domain interactions. Genome Inform 2006, 17(2):35–45. Authors’ contributions 19. Pitkänen E, Rousu J, Ukkonen E: Computational methods for metabolic reconstruction. Curr Opin Biotechnol 2010, 21:70–77. 19. Pitkänen E, Rousu J, Ukkonen E: Computational methods for metabolic reconstruction. Curr Opin Biotechnol 2010, 21:70–77. CS and ST implemented and carried out the experiments. MD and AP provided the original idea and supervised the experiments. All authors wrote and approved the final manuscript. 20. Tuncbag N, Gursoy A, Keskin O: Identification of computational hot spots in protein interfaces: combining solvent accessibility and inter-residue potentials improves the accuracy. Bioinformatics 2009, 25(12):1513–1520. 20. Tuncbag N, Gursoy A, Keskin O: Identification of computational hot spots in protein interfaces: combining solvent accessibility and inter-residue potentials improves the accuracy. Bioinformatics 2009, 25(12):1513–1520. Author details 1 1Dipartimento di Ingegneria dell’Informazione e Scienze Matematiche, University of Siena, Siena, Italy. 2Dipartimento di Ingegneria e Scienza dell’Informazione, University of Trento, Trento, Italy. φ(f , S) = x1∈S1 . . . xn∈Sn 1 −tE(f , x1, . . . , xn) Received: 8 August 2013 Accepted: 3 March 2014 Published: 12 April 2014 Received: 8 August 2013 Accepted: 3 March 2014 Published: 12 April 2014 Received: 8 August 2013 Accepted: 3 March 2014 Published: 12 April 2014 Received: 8 August 2013 Accepted: 3 March 2014 Published: 12 April 2014 T-norms When multiple universally quantified variables are present, the conversion is performed recursively Page 17 of 18 Saccà et al. BMC Bioinformatics 2014, 15:103 http://www.biomedcentral.com/1471-2105/15/103 Saccà et al. BMC Bioinformatics 2014, 15:103 http://www.biomedcentral.com/1471-2105/15/103 from the outer to the inner variable. In particular, ∀v1 . . . ∀vn E(P, v1, . . ., vn) is mapped to the constraint: from the outer to the inner variable. In particular, ∀v1 . . . ∀vn E(P, v1, . . ., vn) is mapped to the constraint: References K k O For example, the formula ∀v1∀v2 A(v1) ∧¬B(v2) corre- sponds to: 1. Keskin O, Gursoy A, Ma B, Nussinov R: Principles of protein-protein interactions: what are the preferred ways for proteins to interact? Chem Rev 2008, 108(4):1225–1244. 1. Keskin O, Gursoy A, Ma B, Nussinov R: Principles of protein-protein interactions: what are the preferred ways for proteins to interact? Chem Rev 2008, 108(4):1225–1244. 1. Keskin O, Gursoy A, Ma B, Nussinov R: Principles of protein-protein interactions: what are the preferred ways for proteins to interact? Chem Rev 2008, 108(4):1225–1244. 2. Hopkins AL: Network pharmacology: the next paradigm in drug discovery. Nat Chem Biol 2008, 4(11):682–690. 3. Csermely P, Korcsmáros T, Kiss HJ, London G, Nussinov R: Structure and dynamics of molecular networks: A novel paradigm of drug discovery. Pharmacol Ther 2013, 138(3):333–408. 4. Tuncbag N, Kar G, Keskin O, Gursoy A, Nussinov R: A survey of available tools and web servers for analysis of protein–protein interactions and interfaces. Brief Bioinform 2009, 10(3):217–232. 5. Lewis AC, Saeed R, Deane CM: Predicting protein–protein interactions in the context of protein evolution. Mol Biosyst 2010, 6:55–64. 6. Skrabanek L, Saini HK, Bader GD, Enright AJ: Computational prediction of protein–protein interactions. Mol Biotechnol 2008, 38:1–17. 7. Wass MN, David A, Sternberg MJ: Challenges for the prediction of macromolecular interactions. Curr Opin Struct Biol 2011, 21(3):382–390. 8. Cary MP, Bader GD, Sander C: Pathway information for systems biology. FEBS Lett 2005, 579(8):1815–1820. 1. Keskin O, Gursoy A, Ma B, Nussinov R: Principles of protein-protein interactions: what are the preferred ways for proteins to interact? Chem Rev 2008, 108(4):1225–1244. 2. Hopkins AL: Network pharmacology: the next paradigm in drug discovery. Nat Chem Biol 2008, 4(11):682–690. 3. Csermely P, Korcsmáros T, Kiss HJ, London G, Nussinov R: Structure and dynamics of molecular networks: A novel paradigm of drug discovery. Pharmacol Ther 2013, 138(3):333–408. 4. Tuncbag N, Kar G, Keskin O, Gursoy A, Nussinov R: A survey of available tools and web servers for analysis of protein–protein interactions and interfaces. Brief Bioinform 2009, 10(3):217–232. 5. Lewis AC, Saeed R, Deane CM: Predicting protein–protein interactions in the context of protein evolution. Mol Biosyst 2010, 6:55–64. 6. Skrabanek L, Saini HK, Bader GD, Enright AJ: Computational prediction of protein–protein interactions. Mol Biotechnol 2008, 38:1–17. 7. Wass MN, David A, Sternberg MJ: Challenges for the prediction of macromolecular interactions. Curr Opin Struct Biol 2011, 21(3):382–390. 8. Acknowledgements g This research was supported by grant PRIN 2009LNP494 (Statistical Relational Learning: Algorithms and Applications) from Italian Ministry of University and Research and by a Google Faculty Research Award (Integrated Prediction of Protein Function, Interactions and Pathways with Statistical Relational Learning). 21. Yip KY, Kim PM, McDermott D, Gerstein M: Multi-level learning: improving the prediction of protein, domain and residue interactions by allowing information flow between levels. BMC Bioinformatics 2009, 10:241. 21. Yip KY, Kim PM, McDermott D, Gerstein M: Multi-level learning: improving the prediction of protein, domain and residue interactions by allowing information flow between levels. BMC Bioinformatics 2009, 10:241. Page 18 of 18 Saccà et al. BMC Bioinformatics 2014, 15:103 http://www.biomedcentral.com/1471-2105/15/103 Saccà et al. BMC Bioinformatics 2014, 15:103 http://www.biomedcentral.com/1471-2105/15/103 22. Diligenti M, Gori M, Maggini M, Rigutini L: Bridging logic and kernel machines. Mach Learn 2012, 86:57–88. 23. Cortes C, Vapnik V: Support-vector networks. Mach Learn 1995, 20(3):273–297. 4. Getoor L, Taskar B: Introduction to Statistical Relational Learning. 2 25. Raedt LD, Frasconi P, Kersting K, Muggleton S (Eds): Probabilistic Inductive Logic Programming - Theory and Applications, Volume 4911 of Lecture Notes in Computer Science. 2008. 26. Hofmann T, Schölkopf B, Smola AJ: Kernel methods in machine learning. Ann Stat 2008, 36(3):1031–1508. 27. Ben-Hur A, Noble WS: Kernel methods for predicting prot interactions. Bioinformatics 2005, 21(suppl 1):i38–i46. 28. Chapelle O, Schölkopf B, Zien A: Semi-Supervised Learning, Volume 2: MIT press; 2006. 29. Zhu X: Semi-supervised learning literature survey. Computer Science, University of Wisconsin-Madison 2006, 2:3. [http://pages.cs.wisc.edu/ jerryzhu/research/ssl/semireview.html] j y 30. Hwang H, Pierce B, Mintseris J, Janin J, Weng Z: Protein–protein docking benchmark version 3.0. Proteins: Struct, Funct, Bioinf 2008, 73(3):705–709. 31. Li B, Kihara D: Protein docking prediction using predicted protein-protein interface. BMC Bioinformatics 2012, 13:7. 32. Qi Y, Noble WS: Protein interaction networks: protein domain interaction and protein function prediction. In Handbook of Statistical Bioinformatics. 2011:427–459. 33. Bock JR, Gough DA: Predicting protein–protein interactions from primary structure. Bioinformatics 2001, 17(5):455–460. 34. Qi Y, Tastan O, Carbonell JG, Klein-Seetharaman J, Weston J: Semi- supervised multi-task learning for predicting interactions between HIV-1 and human proteins. Bioinformatics 2010, 26(18):i645–i652. 35. You ZH, Lei YK, Gui J, Huang DS, Zhou X: Using manifold embedding for assessing and predicting protein interactions from high-throughput experimental data. Bioinformatics 2010, 26(21):2744–2751. 36. Xia Z, Wu LY, Zhou X, Wong ST: Semi-supervised drug-protein interaction prediction from heterogeneous biological spaces. BMC Syst Biol 2010, 4(Suppl 2):S6. 37. Nguyen TP, Ho TB: Detecting disease genes based on semi-supervised learning and protein–protein interaction networks. Artif Intell Med 2012, 54:63–71. 38. De Raedt L: Inductive Logic Programming. 2010. 39. Tran TN, Satou K, Ho TB: Using inductive logic programming for predicting protein-protein interactions from multiple genomic data. In Knowledge Discovery in Databases: PKDD 2005; 2005:321–330. g y 40. González AJ, Liao L, Wu CH: Prediction of contact matrix for protein–protein interaction. Bioinformatics 2013, 29(8):1018–1025. 41. Smola AJ, Schölkopf B: A tutorial on support vector regression. Stat Comput 2004, 14(3):199–222. 42. Blum A, Mitchell T: Combining labeled and unlabeled data with co-training. In Proceedings of the Eleventh Annual Conference on Computational Learning Theory. Saccà et al. BMC Bioinformatics 2014, 15:103 http://www.biomedcentral.com/1471-2105/15/103 ACM; 1998:92–100. 43. Altschul SF, Madden TL, Schäffer AA, Zhang J, Zhang Z, Miller W, Lipman DJ: Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res 1997, 25(17):3389–3402. 43. Altschul SF, Madden TL, Schäffer AA, Zhang J, Zhang Z, Miller W, Lipman DJ: Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res 1997, 25(17):3389–3402. Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color ﬁgure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit Submit your next manuscript to BioMed Central and take full advantage of: 46. Schölkopf B, Smola AJ: Learning with Kernels. 2002. 47. Klement E, Mesiar R, Pap E: Triangular Norms. 2000. 48. Klir G, Yuan B: Fuzzy Sets and Fuzzy Logic: Theory and Applications. 1995. doi:10.1186/1471-2105-15-103 Cite this article as: Saccà et al.: Improved multi-level protein–protein interaction prediction with semantic-based regularization. BMC Bioinfor- matics 2014 15:103.

Subsets and Splits

No community queries yet

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