betaFeedback

WorksPhysics and Astronomy

Limited knowledge exchange across biological subfields in agent-based modelling: an LLM-assisted bibliometric analysis

Abstract

Agent-based models (ABMs) are widely used across the biological sciences. While applications differ between biological subfields, the mathematical and statistical methods underlying ABMs largely do not, raising the question of how much knowledge is exchanged between subfields. We address this question by developing a bibliometric pipeline that combines a database keyword search with coarse- and fine-grained classification by a large language model (LLM). We apply the pipeline to the Web of Science Core Collection, classifying 22,936 ABM articles into ecology, epidemiology, oncology, other-biology, or not-biology. Using citation patterns and author overlap as proxies for knowledge exchange, we find that 83%, 81%, and 69% of citations from ecology, epidemiology, and oncology articles, respectively, go to articles in the same subfield. Furthermore, 91% of authors with at least two articles across these subfields publish exclusively within one subfield. Both proportions far exceed those expected under random assignment of subfields (permutation tests, p≤0.001p \leq 0.001). Terminology also differs, with "individual-based model" used mainly in ecology and "agent-based model" in epidemiology, oncology, and outside biology. Taken together, our results suggest limited knowledge exchange between biological subfields in ABM research, leading us to discuss opportunities for increased exchange through conferences, journals, and training.

main.pdf

Introduction

Agent-based models (ABMs), also known as individual-based models (IBMs), are a mathematical and computational framework used to simulate autonomous agents that interact with each other and with their environment. Each agent follows deterministic or stochastic rules and has its own attributes (e.g., type, position, and age), allowing for heterogeneous populations of agents [1]. ABMs can therefore be used to study how population-level dynamics emerge from individual-level behaviours, even in systems that are difficult to study analytically [2, 3, 4]. The flexibility of ABMs has made them a widely used tool across the biological sciences, including in ecology [5, 6], epidemiology [7, 8], and oncology [9, 10], where agents may represent, e.g., cells, plants, animals, or humans [4].

Modern ABMs have their origins in rule-based systems such as cellular automata (CA), which were developed in the 1940s by pioneering mathematicians including John von Neumann and Stanislaw Ulam, partly to investigate the principles underlying biological self-reproduction [11, 12, 13]. CA reached a wider audience in 1970, when Martin Gardner popularised John Conway’s Game of Life, in which simple rules governing the birth, death, and survival of organisms on a grid generate complex patterns [14, 15]. Around the same time, Thomas Schelling brought rule-based modelling into the social sciences with his model of residential segregation [16, 17]. Stephen Wolfram later demonstrated that even simple CA can generate highly complex behaviour [18, 19], and with the spread of personal computers, ABMs became common across many disciplines in the 1990s [20].

Although the research questions differ between scientific disciplines, the mathematical and statistical methods underlying ABMs are largely shared, such that many methodological advances developed in one discipline should be transferable to others. However, communities of ABM researchers have to a large extent emerged in parallel: individual-based modelling in ecology developed independently of agent-based modelling in the social and computational sciences [21], and the two terms remain associated with different research communities today [22, 23]. This parallel development raises the question of how much knowledge is exchanged between ABM communities in different disciplines or, as is the focus of this study, between biological subfields.

Using citation patterns and author overlap across subfields as example proxies, knowledge exchange can be studied through bibliometric analysis, which uses statistical methods to examine patterns in publication data [24, 25]. These data can be obtained from citation databases such as the Web of Science [26], Scopus [27], and OpenAlex [28]. A search of the Web of Science Core Collection (WoSCC) [29] identified 47 bibliometric studies of ABMs (Supplementary Material, SM1), including studies of ABMs in epidemiology [30], transportation [31], and finance [32]. Moreover, literature reviews of biology-focused ABMs that include bibliometric features have been conducted within individual subfields, such as ecology [33], epidemiology [34], and oncology [35]. In contrast, Vincenot [21] analysed citation patterns of ABM research across disciplines using journal-level subject categories that group biological research into pre-defined areas such as “agricultural and biological sciences” and “medicine”. Separating articles into finer biological subfields is challenging using traditional bibliometric approaches, as keyword searches rely on explicit terminology [36] and database subject categories can be too broad [37]. However, recent advances in large language models (LLMs) offer an alternative approach, allowing users to classify articles based on the semantic content of their text [38, 39, 40].

In this work, we develop an LLM-assisted bibliometric pipeline to study ABM research across biological subfields. We focus on ecology, epidemiology, and oncology: three subfields in which ABMs are a well-established modelling framework [41]. We compare terminology across subfields and use cross-subfield citation patterns and author overlap to assess knowledge exchange.

Methods

Data collection and cleaning

We used the Web of Science (WoS) application programming interface (API) [42] to retrieve metadata for all documents in the WoSCC containing “agent-based model” or “individual-based model” in the title, abstract, or author keywords (Figure 1a). For each of the 23,902 documents retrieved on 9 January 2026, we obtained the following metadata: title, abstract, author keywords, publication year, author ORCID iDs, and references. To ensure sufficient data per year, we excluded the 12 documents published before 1993, the first year with more than 10 publications. We also removed 755 documents with missing titles or abstracts and 199 non-English documents (Figure 1b), resulting in a dataset of 22,936 ABM articles published between 1993 and 2025.

Article selection and classification workflow and stacked area chart

Figure 1. Article selection and classification. (a) Search query used to retrieve documents from the WoSCC. TI, AB, and AK denote the article title, abstract, and author keywords, respectively. The asterisk denotes a wildcard that matches any word ending. (b) Data collection, cleaning, and classification workflow. (c) Number of articles per year in each biological subfield, shown as a stacked area chart. nn denotes the total number of articles in each subfield.

Article classification

To classify the articles, we used the LLM Qwen3:8b [43], 4-bit quantised and run with Ollama (v0.40.1), with temperature 0. Two zero-shot prompts [44, 45], written in American English [46], were used. Prompt 1 separated biology from not-biology articles (Section 2.2.1, Appendix A1), and Prompt 2 categorised the biology articles into four prescribed biological subfields: epidemiology, oncology, ecology, or other-biology (§ 1.2.2.2, Appendix A2). Each prompt was refined iteratively [47, 48, 49] on a human-annotated development set sampled from the dataset it was applied to, and was then evaluated on separate human-annotated data (§ 1.2.2.3).

Coarse-grained classification: biology versus not-biology

Prompt 1 (Appendix A1) instructed the LLM to classify each article abstract as biology or not-biology based on its primary scientific focus. Accordingly, biology was defined as research on living organisms or biological processes, in line with the common definition of biology as the scientific study of life [50], and not-biology as research on physical systems, engineering, modelling, social sciences, or environmental processes without biological mechanisms. The prompt was refined on a development set of 100 randomly sampled articles (without stratification), which one author annotated as 36 biology and 64 not-biology articles. Refinement stopped when the agreement between the human

annotations and the LLM classifications reached an accuracy of 94%. The WoS unique identifiers, DOIs (where available), and annotations are listed in Supplementary Table SM1. Applied to the full dataset (including the development set), the prompt identified 7,285 biology and 15,651 not-biology articles.

Fine-grained classification: biological subfields

Prompt 2 (Appendix A2) instructed the LLM to categorise each biology-classified article into exactly one biological subfield based on the primary focus of its abstract. We specified three main subfields: (1) oncology, if the article focuses on cancer, (2) epidemiology, if it focuses on disease spread or population-level health outcomes in humans or animals, and (3) ecology, if it focuses on interactions among living organisms and their environment; (4) other-biology is a residual category for articles that do not fit these. The prompt applied these in the listed order, prioritising the three main subfields in an effort to minimise the number of incorrectly classified articles in them. The prompt was refined on a development set of 100 randomly sampled biology articles (without stratification), which one author annotated as 54 ecology, 20 epidemiology, 5 oncology, and 21 other-biology articles. Refinement stopped once agreement with the human annotations plateaued at an overall accuracy of approximately 90% (and a precision of 95% across ecology, epidemiology, and oncology). The WoS unique identifiers, DOIs (where available), and annotations are listed in Supplementary Table SM2. Applied to all 7,285 biology articles (including the development set), the prompt identified 3,654 ecology, 1,309 epidemiology, 347 oncology, and 1,975 other-biology articles. Figure 1c shows that ecology was the subfield with the largest number of ABM-related articles published throughout our dataset. Before the year 2000, over 80% of yearly ABM-related biology articles were ecology-based, with this proportion remaining above 50% until 2008, and never falling below 37% throughout the study period.

LLM evaluation

To evaluate Prompt 1, the three authors collectively annotated a total of 150 abstracts predicted to be biology and 150 predicted to be not-biology. The abstracts were randomly sampled within each predicted category, and the annotators were blind to the LLM predictions. Each author independently annotated 100 abstracts from each category, with 50 shared between each pair of authors. Disagreements were first reconsidered by the two authors involved after rereading the abstract, and any remaining ties were broken by the third author. The author pairs initially agreed on 96% of the abstracts. Of the 12 disagreements, 11 were resolved by rereading and 1 by the third author. All LLM classifications and human annotations (initial, after rereading, and tie-break) are listed in Supplementary Table SM3. As abstracts were sampled by predicted category, we evaluated the prompt using two metrics: precision (the percentage of articles predicted as biology that were annotated as biology) and the negative predictive value (NPV; the percentage of articles predicted as not-biology that were annotated as not-biology). The confusion matrix (Table 1) gives a precision of 95% and an NPV of 89%. We considered precision more important, since a false positive contaminates the subfield analysis, whereas a false negative mainly reduces the sample size. This comes at a cost in coverage: based on the NPV, we estimated that approximately 11% (about 1,700) of the 15,651 articles classified as not-biology are in fact biology.

PredictedBiologyNot-biology
Biology142 (94.67%)8 (5.33%)
Not-biology16 (10.67%)134 (89.33%)

Table 1. Confusion matrix for Prompt 1. Percentages are calculated relative to the total number of articles in each predicted category (n=150n = 150). The top-left percentage is the precision and the bottom-right is the negative predictive value.

Prompt 2 was evaluated similarly. Each author independently annotated 50 abstracts from each predicted subfield, with 25 shared between each pair of authors, totalling 75 distinct abstracts per subfield. The author pairs initially agreed on 86% of the abstracts. Of the 43 disagreements, 24 were resolved by rereading and 19 by the third author, 18 of which involved other-biology. All LLM classifications and human annotations are listed in Supplementary Table SM4, and the evaluation results are summarised in the confusion matrix (Table 2). We evaluated Prompt 2 using the precision of each predicted subfield, which is high for ecology (93%), epidemiology (95%), and oncology (100%), and lower for other-biology (72%), where 21 of the 75 abstracts were annotated as ecology (13), epidemiology (7), or oncology (1), consistent with its role as the residual category. Although the stringent, priority-based assignment placed some main-subfield articles in other-biology, the high precision meant that each main subfield contained few articles that belong elsewhere, yielding reliable datasets for investigating cross-subfield knowledge exchange.

Annotated
EcologyEpidemiologyOncologyOther
Predicted Ecology70 (93.33%)1 (1.33%)0 (0.00%)4 (5.33%)
Predicted Epidemiology0 (0.00%)71 (94.67%)0 (0.00%)4 (5.33%)
Predicted Oncology0 (0.00%)0 (0.00%)75 (100.00%)0 (0.00%)
Predicted Other13 (17.33%)7 (9.33%)1 (1.33%)54 (72.00%)

Table 2. Multiclass confusion matrix for Prompt 2. Percentages are calculated relative to the total number of articles in each predicted subfield (n=75n = 75). The diagonal percentages are the precision of each predicted subfield.

Bibliometric analyses

Terminology

To compare terminology across subfields, we used case- and hyphen-insensitive matching to classify each article according to the search terms in its title, abstract, or author keywords: only “agent-based model*”, only “individual-based model*”, or both. Here, * denotes a wildcard matching any word ending, such as “models” or “modelling”. We then calculated the number of articles in each category per year for each of the four biological subfields, and for not-biology articles as a reference.

Citation patterns

To study knowledge exchange through citations, we analysed citations between articles within the biology dataset, excluding any articles outside it. We included all four fine-grained categories, as excluding the residual other-biology category would inflate within-subfield citation proportions. For each subfield pair (including self-pairs), we calculated the number of citations given by articles in the first subfield and received by articles in the second.

To test whether within-subfield citation proportions were higher than expected from subfield sizes alone, that is, under random assignment of subfields, we performed a permutation test [51]. In each permutation, subfield labels were randomly shuffled among articles published in the same year, while citation links between articles remained unchanged. This preserved annual subfield sizes and the citation network, but removed the association between individual articles and their subfields. Within-subfield citation proportions were recalculated for B=1000B = 1000 permutations, giving a null distribution for each subfield. For the one-sided alternative hypothesis that the observed within-subfield citation proportion exceeds that under random assignment of subfield labels, the pp-value was approximated by the Monte Carlo pp-value p=(b+1)/(B+1)p = (b + 1)/(B + 1) [52]. Here, bb is the number of permutations in which the within-subfield citation proportion was at least as high as the observed proportion, and adding one to both the numerator and the denominator, which counts the observed data as one permutation, ensures a valid pp-value that is never zero [53, 52].

Author overlap

To study knowledge exchange through authors who publish in multiple subfields, we identified authors by their ORCID iDs. We restricted this analysis to articles in the biology dataset published from 2010 onwards, when annual ORCID coverage was at least 50% in all subfields (Supplementary Material SM3.1). Unlike in the citation analysis, we also excluded the residual other-biology category, which is heterogeneous by design and contains some misclassified articles from the three main subfields (Section 2.2.3). Thus, author overlap with the other-biology category would overstate knowledge exchange between research communities. To test the robustness of our results to this exclusion, we repeated the analysis with other-biology included (Supplementary Material SM3.2). Furthermore, only authors with at least two articles across the three main subfields (ecology, epidemiology, and oncology) were included, since authors with a single article cannot publish in more than one subfield. For each subfield, we calculated author overlap as the percentage of its authors who also published in at least one of the other two subfields.

To test whether author overlap was lower than expected from authors’ publication counts alone, that is, under random assignment of subfields, we used a permutation test analogous to that for citations (Section 2.3.2). In each permutation, subfield labels were randomly shuffled among articles published in the same year, while the links between authors and articles remained unchanged. This preserved annual subfield sizes and the number and timing of each author’s articles, but removed the association between individual articles and their subfields. Author overlap was recalculated for B=1000B = 1000 permutations, giving a null distribution for each subfield. For the one-sided alternative hypothesis that the observed author overlap is lower than under random assignment of subfield labels, the pp-value was approximated by the Monte Carlo pp-value p=(b+1)/(B+1)p = (b + 1)/(B + 1) [52], where bb is the number of permutations in which author overlap was at least as low as observed.

Results

ABM or IBM? Terminology differs between subfields

Figure 2 shows the annual number of articles using “agent-based model”, “individual-based model”, or both terms, for each biological subfield and for not-biology articles (raw counts in Supplementary Table SM5). In not-biology articles, “agent-based model” dominated throughout the study period and “individual-based model” was rarely used. Within biology, “individual-based model” was the most frequently used terminology in ecology-classified articles between 1993 and 2025. However, “agent-based model” dominated in both epidemiology and oncology. In absolute terms, the use of “agent-based model” increased in all biological subfields across the study period, whereas the use of “individual-based model” grew markedly only in ecology. In relative terms, epidemiology shifted from preferring one term to the other: “individual-based model” was more frequently used until around 2010, after which “agent-based model” became predominant, reaching 87% of articles (82 of 94) in 2024. Articles using both terms together were rare (ranging from 0 to 12%) in all subfields and years.

Terminology used across biological subfields and not-biology articles

Figure 2. Terminology used across biological subfields and not-biology articles. Annual number of articles containing “agent-based model” only, “individual-based model” only, or both terms together, in the title, abstract, or author keywords, across biological subfields and not-biology. nn denotes the total number of articles in each category.

Subfields mainly cite themselves

Figure 3 shows citation flows between biological subfields. In absolute terms, ecology gave and received the most citations, whereas oncology exchanged few, consistent with their respective sizes (Figure 3a). Relative to subfield size, citations were concentrated within subfields: 83%, 81%, 69%, and 63% of citations given by ecology, epidemiology, oncology, and other-biology, respectively, went to articles in the same subfield (Figure 3b). Where cross-subfield citations occurred, exchange was largely reciprocal (Supplementary Material SM2). For all four subfields, the observed within-subfield citation proportion lay far above the null distribution, whose central 95% ranges were [57.77%, 64.90%], [11.88%, 16.42%], [2.34%, 5.45%], and [22.50%, 29.36%] for ecology, epidemiology, oncology, and other-biology, respectively (Figure 3c). No permutation reached the observed values (p≤0.001p \le0.001 for all subfields). The concentration of citations within subfields therefore cannot be explained by differences in subfield size and publication year alone.

Citation flows within and between biological subfields, showing citation flows between ecology, epidemiology, oncology, and other biology, proportions of citations given, and null distributions of within-subfield citation proportions

Figure 3. Citation flows within and between biological subfields. Subfields are ecology, epidemiology, oncology, and the residual other-biology category, and only citations between articles in the biology dataset are included. (a) Citation flows between subfields. Node heights are proportional to the total number of citations received (left) or given (right) by each subfield, and ribbon heights are proportional to the number of citations between subfields. (b) Proportion of citations given to each cited subfield, with colours as in (a). (c) Null distributions of within-subfield citation proportions from 1000 permutations, in which subfield labels were shuffled among articles within each publication year. Vertical lines mark the observed values.

Authors mainly publish within one subfield

Figure 4 shows author overlap between ecology, epidemiology, and oncology for the 2,195 authors with at least two articles across these subfields from 2010 onwards. Of these, 91% published in only one subfield (Figure 4a,Figure 4b). The percentage of authors who also published in at least one of the other two subfields was 12%, 21%, and 28% for ecology, epidemiology, and oncology, respectively (Figure 4c). Relative overlap was lowest for ecology, the largest subfield, and highest for oncology, the smallest.

Figure 4. Authors publishing within and across biological subfields. Subfields are ecology, epidemiology, and oncology, and only authors with an ORCID iD and at least two articles across these subfields published between 2010 and 2025 are included. (a) Number of unique authors publishing in each combination of subfields. (b) For authors who published in ecology, epidemiology, and oncology (bars, left to right), the proportion who published in one, two, or three subfields. Colours as in (a). (c) Null distributions of author overlap from 1000 permutations, in which subfield labels were shuffled among the included articles within each publication year. Vertical lines mark the observed values.

For all three subfields, observed author overlap lay far below the null distribution, whose central 95% ranges were [64.02%, 69.22%], [88.52%, 92.72%], and [96.67%, 99.57%], respectively (Figure 4c). In none of the 1000 permutations was author overlap as low as observed, giving p≤0.001p \le0.001 for all three subfields. When the analysis was repeated to include the residual other-biology category, overlap increased to 34%, 36%, and 66% for ecology, epidemiology, and oncology, respectively, mainly through authors who also published in other-biology (Supplementary Material SM3.2). Nevertheless, overlap remained far below the null distribution for all subfields (p≤0.001p \le0.001). Within our dataset, authors therefore published across subfields considerably less often than they would have if their articles had been randomly assigned to subfields.

Discussion

In this study, we developed an LLM-assisted bibliometric pipeline to classify ABM research into biological subfields. We compared terminology across these subfields and used citation patterns and author overlap as proxies for knowledge exchange between them. Three main findings stand out. Firstly, terminology differed between subfields: “individual-based model” was used mainly in ecology, whereas “agent-based model” was more common in epidemiology and oncology (§ 1.3.1). Secondly, between 69% and 83% of citations given by ecology, epidemiology, and oncology articles went to articles in the same subfield (§ 1.3.2). Thirdly, 91% of authors with at least two articles in these subfields published in only one of them (§ 1.3.3). In both the citation and author analyses, the separation between subfields was greater than under random assignment of subfields. Together, these patterns indicate that biological ABM research is organised into largely distinct subfield communities.

The observed separation between subfields is particularly notable because ABMs share core mathematical, statistical, and computational foundations across applications [4, 22], so methods developed in one subfield could reasonably be expected to be useful in others. Although our bibliometric analyses cannot explain why the separation exists, the differences in terminology indicate distinct modelling traditions in different subfields, which may contribute to it. Terminology differences can make relevant literature harder to find across subfields [54] and may thereby reinforce the separation between communities. Separation between research communities has also been observed beyond ABMs, with fields often becoming increasingly isolated as they mature [55, 56], although our previous bibliometric study suggests that mathematical oncology has become more interconnected over time [57].

The results presented in this study should be interpreted in light of several design choices and their associated limitations. The keyword search used for article retrieval relied on the terms “agent-based model” and “individual-based model”, so relevant articles using other terminology, such as early rule-based models [15, 17], were not included. Keyword bias is a well-documented limitation of bibliometric studies, affecting which articles are included in the dataset [58, 59, 60], and the dataset itself was limited to documents indexed in the WoSCC [29]. Other bibliographic databases, such as Scopus [27] and OpenAlex [28], differ in coverage, particularly of conference proceedings and non-English-language journals [61]. Similarly, the author analysis was limited to researchers with associated ORCID iDs.

The article classification also involved several simplifications. Articles were assigned to exactly one subfield, with overlaps resolved by a fixed priority order (oncology, epidemiology, ecology) chosen in increasing order of subfield size, as estimated from keyword frequencies (Appendix A2). Articles spanning several subfields, such as models of cancer epidemiology, were therefore counted in only one, so knowledge exchange within individual articles was not captured. Classification was based only on abstracts, which generally capture an article’s main focus but may omit details that would clarify its biological application. Moreover, the decision to prioritise precision in the three main subfields over complete coverage came at the cost of lower precision for the residual other-biology category (72%), which therefore contained articles that human annotators assigned to the main subfields. Since any article misclassified in this way turns its within-subfield citations into cross-subfield citations, and may inflate author overlap in the analysis including other-biology (Supplementary Material SM3.2), our findings of limited knowledge exchange can be interpreted as conservative. However, random assignment of subfields provides only a minimal baseline, and without comparison to other modelling approaches or fields, we cannot assess whether the observed separation in biological ABM research is unusually strong.

Furthermore, using citations and author overlap within our dataset as proxies for knowledge exchange is itself a simplification. Our analysis did not distinguish whether a citation built on, critiqued, or only briefly mentioned prior work, and citations to articles outside the dataset were not counted. Bibliometric approaches that could investigate knowledge exchange in more depth include bibliographic coupling [62, 63], citation context analysis [64, 65], textual similarity of abstracts [66], models of topic diffusion between disciplines [67], and analyses of software mentions in articles [68]. As a final limitation, our analysis identifies limited knowledge exchange but does not measure whether this has affected research quality or impact, which for interdisciplinary research can take longer to emerge [69].

Future work could address the discussed limitations, for example by combining several bibliographic databases or by integrating multiple bibliometric approaches [70]. Moreover, studying how citations and author overlap change over time could test whether knowledge exchange between subfields is increasing or decreasing. Journal- or conference-level analyses could furthermore identify where cross-subfield communication occurs and, equally importantly, where it does not. Comparing ABMs with other modelling approaches, such as differential equation models, could also show whether the observed separation is specific to ABMs or characteristic of quantitative biology in general. The funnel-shaped article classification pipeline developed in this study could be applied to other research methods and disciplines, especially since the computationally expensive LLM steps are applied only to progressively smaller subsets of articles, allowing it to scale to large datasets. Any such reuse should include transparent reporting of methodological choices and evaluation against human annotations, in line with recommended practices for LLM-based annotation [71]. Beyond bibliometrics and the quantitative proxies used for knowledge exchange in this work, qualitative approaches such as surveys [72] and interviews [73] could clarify how researchers discover work across subfields and which barriers they face. Such approaches could also help identify examples of knowledge exchange that have proved useful in practice.

While our work identifies, and largely focuses on, limited knowledge exchange between subfields, we stress that subfield-focused research has many benefits, such as deep domain expertise, shared vocabulary, and specialist peer review [74]. Interdisciplinary research, on the other hand, has been noted to carry costs, including lower funding success [75] and fewer publications, albeit with higher visibility [76]. Several studies have investigated whether interdisciplinary research receives more citations or produces better science, often with contradictory results [77, 78, 79, 80]. Specialised research communities also benefit from communication in dedicated modelling journals: mathematical ecologists may publish in Ecological Modelling and Theoretical Ecology, and mathematical epidemiologists in Epidemics and Infectious Disease Modelling. Mathematical oncologists, however, mainly publish in broader journals, as no dedicated mathematical oncology journal has yet become established. Specialisation is also reflected in ABM software. Although general-purpose platforms such as NetLogo [81], Mesa [82], and Repast Simphony [83] can be applied across biological subfields, more specialised platforms have been developed in parallel for specific applications. For example, HexSim [84] and RangeShifter [85] are widely used in ecology, Covasim [86] and Epidemilogical Modelling software (EMOD) [87] in epidemiology, and PhysiCell [88], CompuCell3D [89], Chaste [90], and the Hybrid Automata Library (HAL) [91] in oncology.

We argue that knowledge exchange between subfields can be valuable, and that there are many opportunities to increase such exchange while retaining subfield-specific expertise. Interdisciplinary workshops, methods-focused training, and special issues in scientific journals could make methods and terminology visible across subfields. Notably, the Ten Simple Rules collection in PLOS Computational Biology offers practical guidance on interdisciplinary collaboration [92, 93, 94, 95]. Furthermore, broad mathematical biology meetings, such as those organised by the Society for Mathematical Biology and the SIAM Life Sciences community, bring together researchers working on ecological, epidemiological, and oncological models while also providing subfield-focused symposia and subgroups activities. By connecting researchers across biological subfields, such initiatives can help turn shared methodological foundations into shared scientific progress.

Data accessibility

The code to reproduce the data cleaning, article classification, and results in this study is available on the public GitHub repository https://github.com/KiraPugh/ABM_knowledge_exchange. Data included in this study are derived from the WoS (Clarivate) and cannot be redistributed under its licence terms. The search query and retrieval date are provided in Section 2.1. The WoS unique identifiers and the DOIs (where available) of all articles making up the dataset are available in Supplementary tbl:6.

Declaration of AI use

The authors used AI to edit prose and code. All conceptual, methodological, theoretical, and empirical contributions are the authors’ own.

Conflict of interest declaration

The authors declare no competing interests.

Funding

KP was funded by the Wenner-Gren Stiftelserna/the Wenner-Gren Foundations through a grant to SH (WUP2024-0006). RMC was supported by an Engineering and Physical Sciences Research Council grant (EP/Z534870/1). SH was funded by Wenner-Gren Stiftelserna/the Wenner-Gren Foundations (WGF2022-0044) and the Kjell och Märta Beijer Foundation.

Appendix

Prompt 1

Prompt 1 (biology versus not-biology classification)

"""
You are an expert scientific document classifier. Determine whether a
scientific abstract is primarily about **Biology** or **Not Biology**.

Definitions:
- Biology = Research on living organisms or biological processes.
- Not Biology = Research primarily on physical systems, engineering,
modeling, social sciences, or environmental processes without biological
mechanisms.

Instructions:
1. Only classify abstracts based on their primary scientific focus.
2. Output ONLY JSON in this format:

[
    {"id": 1, "Category": "Biology"},
    {"id": 2, "Category": "Not Biology"}
]

Do NOT add commentary or explanations.
"""

Prompt 2

Prompt 2 (biology subfield classification)

"""
You are an expert biological sciences classifier.

Classify each abstract into exactly ONE PRIMARY category.

You MUST determine the PRIMARY scientific focus of the study, not
secondary mentions or background context.

--------------------------------------------------
DECISION ORDER (follow strictly):

1. oncology
   Use ONLY if cancer or oncology is the central research subject.
   Cancer must be explicitly studied (e.g. tumors, carcinogenesis, cancer
   treatment, cancer-specific therapies).
   Do NOT use if cancer is only mentioned as motivation or comparison.

2. epidemiology
   Use if the PRIMARY focus is disease or health outcomes analyzed across
   human or animal populations, i.e., epidemiology.
   Do NOT use if disease is a minor component of ecological or
   environmental dynamics.

3. ecology
   Use if the study focuses on how living organisms interact with each
   other and their environment, i.e., ecology.
   Includes wildlife disease ONLY when ecological dynamics (not public
   health) are the primary outcome.

4. other-biology
   Use if the study does NOT clearly fit oncology, epidemiology, or ecology.
   Includes all other biology subfields, review papers, or methods papers
   without a clear focus on health-related event in a population,
   ecosystem, or cancer.

--------------------------------------------------

Rules:
- Choose the MOST SPECIFIC applicable category.
- Do NOT infer intent beyond what is stated.
- Never guess.
- Respond ONLY in valid JSON as a list of objects:
  [{"id": 1, "KA": "oncology"}, ...]
"""

Supplementary material to Limited knowledge exchange across biological subfields in agent-based modelling: an LLM-assisted bibliometric analysis

Kira Pugh; Rebecca M Crossley; Sara Hamis

supplementary_material.pdf

⊛ Corresponding author: [email protected].

Identifying previous bibliometric studies of ABMs

To identify previous bibliometric analyses of ABM research, we conducted the following search in the WoS Core Collection (WoSCC):

(TI=(bibliometric*) OR AB=(bibliometric*) OR AK=(bibliometric*))
AND (TI=("agent-based model*" OR "individual-based model*")
OR AB=("agent-based model*" OR "individual-based model*") OR
AK=("agent-based model*" OR "individual-based model*"))

This search query was conducted on 28 October 2026 and retrieved articles containing the term “bibliometric*”, together with “agent-based model*” or “individual-based model*” (or both) in the title (TI), abstract (AB), or author keywords (AK), identifying 47 articles published between 2011 and 2025. Note that the asterisk denotes a wildcard that matches any word ending.

Supporting information on citation patterns

Figure SM1 presents the reverse citation perspective of Figure 3b,Figure 3c in the main manuscript, showing the proportion of citations received from each subfield rather than those given to each subfield. Relative to subfield size, citations were concentrated within subfields: 86%, 84%, 69%, and 57% of citations received from ecology, epidemiology, oncology, and other-biology, respectively, went to articles in the same subfield (Figure SM1a). In the permutation test (Figure SM1b), no permutation reached the observed values (p≤0.001p \le0.001 for all subfields). The similarity between the two citation perspectives suggests that citation exchange between subfields is relatively reciprocal, rather than unidirectional.

Citation flows within and between biological subfields

Figure SM1. Citation flows within and between biological subfields. Subfields are ecology, epidemiology, oncology, and the residual other-biology category, and only citations between articles in the biology dataset are included. (a) Proportion of citations received from each citing subfield. (b) Null distributions of within-subfield citation proportions from 1000 permutations, in which subfield labels were shuffled among articles within each publication year. Vertical lines mark the observed values.

Supporting information on author overlap

Annual ORCID iD coverage

Figure SM2 shows a substantial increase in ORCID iD coverage across all four subfields between 1993 and 2025. From 2010 onward, annual ORCID iD coverage exceeds 50% in all four subfields. We therefore restrict the author overlap analysis in § 1.3.3 to the period 2010–2025.

Annual ORCID iD coverage across the four biological subfields

Figure SM2. Annual ORCID iD coverage across the four biological subfields. The red dashed vertical line indicates the first year from which annual ORCID iD coverage has since remained at least 50% in all four subfields.

Author overlap including other-biology

We repeated the author overlap analysis described in § 1.3.3 of the main manuscript, this time including the residual other-biology subfield in the analysis. Out of the 2,954 authors, 72% published in only one subfield (Figure SM3a,Figure SM3b), compared with 91% when other-biology was excluded. Including other-biology increased the proportion of authors publishing across multiple subfields to 34%, 36%, and 66% for ecology, epidemiology, and oncology, respectively (Figure SM3c). In the permutation test, overlap remained far below the null distribution for all subfields (p≤0.001p \le0.001).

Figure SM3. Authors publishing within and across biological subfields. Subfields are ecology, epidemiology, oncology, and other-biology, and only authors with an ORCID iD and at least two articles across these subfields published between 2010 and 2025 are included. (a) Number of unique authors publishing in each combination of subfields. (b) For authors who published in ecology, epidemiology, oncology, and other-biology (bars, left to right), the proportion who published in one, two, three, or four subfields. Colours as in (a). (c) Null distributions of author overlap from 1000 permutations, in which subfield labels were shuffled among the included articles within each publication year. Vertical lines mark the observed values.

References

  1. [1]Crooks, A. T. and Heppenstall, A. J. “Introduction to agent-based modelling”, Agent-Based Models of Geographical Systems. Springer, 2011, pp. 85–105. DOI: 10.1007/978-90-481-8927-4_5.DOI
  2. [2]Bonabeau, E. “Agent-based modeling: Methods and techniques for simulating human systems”, Proceedings of the National Academy of Sciences 99.suppl_3 (2002), pp. 7280–7287. DOI: 10.1073/pnas.082080899.DOI
  3. [3]Macal, C. M. and North, M. J. “Agent-based modeling and simulation”, Proceedings of the 2009 Winter Simulation Conference (WSC). IEEE. 2009, pp. 86–98. DOI: 10.1109/WSC.2009.5429318.DOI
  4. [4]Macal, C. and North, M. “Tutorial on agent-based modelling and simulation”, Journal of Simulation 4 (2010), pp. 151–162. DOI: 10.1057/jos.2010.3.DOI
  5. [5]DeAngelis, D. L. and Mooij, W. M. “Individual-based modeling of ecological and evolutionary processes”, Annual Review of Ecological and Evolutionary Processes 36 (2005), pp. 147–168. DOI: 10.1146/annurev.ecolsys.36.102003.152644.DOI
  6. [6]Zhang, B. and DeAngelis, D. L. “An overview of agent-based models in plant biology and ecology”, Annals of Botany 126.4 (2020), pp. 539–557. DOI: 10.1093/aob/mcaa043.DOI
  7. [7]Tracy, M., Cerdá, M., and Keyes, K. M. “Agent-based modeling in public health: Current applications and future directions”, Annual Review of Public Health 39 (2018), pp. 77–94. DOI: 10.1146/annurev-publhealth-040617-014317.DOI
  8. [8]Auchincloss, A. H. and Diez Roux, A. V. “A new tool for epidemiology: The usefulness of dynamic-agent models in understanding complex effects on health”, American Journal of Epidemiology 168.1 (2008), pp. 1–8. DOI: 10.1093/aje/kwn118.DOI
  9. [9]Metzcar, J., Wang, Y., Heiland, R., and Macklin, P. “A review of cell-based computational modeling in cancer biology”, JCO Clinical Cancer Informatics 3 (2019), pp. 1–13. DOI: 10.1200/CCI.18.00069.DOI
  10. [10]Norton, K.-A., Gong, C., Jamalian, S., and Popel, A. S. “Multiscale agent-based and hybrid modeling of the tumor immune microenvironment”, Processes 7.1 (2019), p. 37. DOI: 10.3390/pr7010037.DOI
  11. [11]Ilachinski, A. Cellular automata: A discrete universe. World Scientific, 2001. DOI: 10.1142/4702.DOI
  12. [12]Sarkar, P. “A brief history of cellular automata”, ACM Computing Surveys (CSUR) 32.1 (2000), pp. 80–107. DOI: 10.1145/349194.349202.DOI
  13. [13]Neumann, J. von. Theory of Self-Reproducing Automata. Ed. by Burks, A. W. University of Illinois Press, 1966.DOI
  14. [14]Payette, N. “Agent-based models of science”, Models of Science Dynamics: Understanding Complex Systems. Springer, 2012, pp. 127–157. DOI: 10.1007/978-3-642-23068-4_4.DOI
  15. [15]Gardner, M. “Mathematical games: The fantastic combinations of John Conway’s new solitaire game ”life””, Scientific American 223.4 (1970), pp. 120–123. DOI: 10.1038/scientificamerican1070-120.DOI
  16. [16]Schelling, T. C. “Models of Segregation”, The American Economic Review 59.2 (1969), pp. 488–493. ISSN: 00028282.
  17. [17]Schelling, T. C. “Dynamic models of segregation”, Journal of Mathematical Sociology 1.2 (1971), pp. 143–186. DOI: 10.1080/0022250X.1971.9989794.DOI
  18. [18]Wolfram, S. Cellular automata as simple self-organizing systems. Tech. rep. CALT-68-938. California Institute of Technology, July 1982. URL: https://lib-extopc.nek.jp/preprints/PDF/1982/8210/8210261.pdf.
  19. [19]Wolfram, S. “Universality and complexity in cellular automata”, Physica D: Nonlinear Phenomena 10.1-2 (1984), pp. 1–35.
  20. [20]Gilbert, N. and Bankes, S. “Platforms and methods for agent-based modeling”, Proceedings of the National Academy of Sciences 99.suppl-3 (2002), pp. 7197–7198. DOI: 10.1073/pnas.072079499.DOI
  21. [21]Vincent, C. E. “How new concepts become universal scientific approaches: Insights from citation network analysis of agent-based complex systems science”, Proceedings of the Royal Society B: Biological Sciences 285.1874, 20172360 (2018). DOI: 10.1098/rspb.2017.2360.
  22. [22]Grimm, V. and Railsback, S. F. Individual-based Modeling and Ecology. Princeton University Press, 2005. ISBN: 9780691096650. DOI: 10.2307/j.ctt5hhnk8.
  23. [23]An, L. et al. “Challenges, tasks, and opportunities in modeling agent-based complex systems”, Ecological Modelling 457, 109685 (2021). DOI: 10.1016/j.ecolmodel.2021.109685.DOI
  24. [24]Ding, S. “Insights into bibliometric analyses”, JBI Evidence Synthesis 23.4 (2025), pp. 564–565. DOI: 10.11124/JBIES-24-00522.DOI
  25. [25]Donthu, N., Kumar, S., Mukherjee, D., Pandey, N., and Lim, W. M. “How to conduct a bibliometric analysis: An overview and guidelines”, Journal of Business Research 133 (2021), pp. 285–296. DOI: 10.1016/j.jbusres.2021.04.070.DOI
  26. [26]Clarivate. Web of Science platform. 2025. URL: https://clarivate.com/academia-government/scientific-and-academic-research/research-discovery-and-referencing/web-of-science/ (visited on 09/28/2026).
  27. [27]Elsevier. Scopus. 2025. URL: https://www.scopus.com/ (visited on 09/28/2026).
  28. [28]OurResearch. OpenAlex. 2022. URL: https://openalex.org (visited on 09/28/2026).
  29. [29]Clarivate. Web of Science Core Collection. 2025. URL: https://clarivate.com/academia-government/scientific-and-academic-research/research-discovery-and-referencing/web-of-science/web-of-science-core-collection/ (visited on 10/10/2026).
  30. [30]Tang, J., Vinayavekhin, S., Weeramongkonkul, M., Suksanon, C., Pattarapremcharoen, K., Thiwathittayanuphap, S., and Leelawat, N. “Agent-based simulation and modeling of COVID-19 pandemic: A Bibliometric analysis”, Journal of Disaster Research 17.1 (2022), pp. 93–102. DOI: 10.20965/jdr.2022.p0093.DOI
  31. [31]Domenteanu, A., Delcea, C., Chiriță, N., and Ioanăș, C. “From data to insights: A bibliometric assessment of agent-based modeling applications in transportation”, Applied Sciences 13.23 (2023), p. 12693. DOI: 10.3390/app132312693.DOI
  32. [32]Trinidad Segovia, J. E., Di Sciorio, F., Mattera, R., and Spano, M. “A bibliometric analysis on agent-based models in finance: Identification of community clusters and future research trends”, Complexity 2022.1 (2022). DOI: 10.1155/2022/4741566.DOI
  33. [33]Grimm, V. “Ten years of individual-based modelling in ecology: What have we learned and what could we learn in the future?”, Ecological Modelling 115.2-3 (1999), pp. 129–148. DOI: 10.1016/S0304-3800(98)00188-4.DOI
  34. [34]Willem, L., Verelst, F., Bilcke, J., Hens, N., and Beutels, P. “Lessons from a decade of individual-based models for infectious disease transmission: A systematic review (2006–2015)”, BMC Infectious Diseases 17 (2017), p. 612. DOI: 10.1186/s12879-017-2699-8.DOI
  35. [35]Stephan, S., Galland, S., Narsis, O. L., Shoji, K., Vachenc, S., Gerart, S., and Nicolle, C. “Agent-based approaches for biological modeling in oncology: A literature review”, Artificial Intelligence in Medicine 152, 102884 (2024). DOI: 10.1016/j.artmed.2024.102884.DOI
  36. [36]Xu, H., Li, X., Tupayachi, J., Lian, J. J., and Omitaomu, O. A. “Automating bibliometric analysis with sentence transformers and retrieval-augmented generation (RAG): A pilot study in semantic and contextual search for customized literature characterization for high-impact urban research”, Proceedings of the 2nd ACM SIGSPATIAL International Workshop on Advances in Urban-AI. 2024, pp. 43–49. DOI: 10.1145/3681780.3697252.DOI
  37. [37]Waltman, L. and Van Eck, N. J. “A new methodology for constructing a publication-level classification system of science”, Journal of the American Society for Information Science and Technology 63.12 (2012), pp. 2378–2392. DOI: 10.1002/asi.22748.DOI
  38. [38]Santos, D. P. D., Mendonca, F. L., Lustosa, J. P., Serrano, A., Torres, J. A. S., and Silva, D. A. da. “Large language models for text classification: A new era of accuracy and efficiency”, International Conference on Computational Science and Computational Intelligence. Springer. 2024, pp. 3–15. DOI: 10.1007/978–3–031–90341–0_1.DOI
  39. [39]Sun, X., Li, X., Li, J., Wu, F., Guo, S., Zhang, T., and Wang, G. “Text classification via large language models”, Findings of the Association for Computational Linguistics: EMNLP 2023. 2023, pp. 8990–9005. DOI: 10.18653/v1/2023.findings–emnlp.603.arxiv.org/abs/2305.08377
  40. [40]Kostina, A., Dikaiakos, M. D., Stefanidis, D., and Pallis, G. “Large language models for text classification: Case study and comprehensive review”, arXiv preprint arXiv:2501.08457 (2025). DOI: 10.48550/arXiv.2501.08457.DOI
  41. [41]White, K. A., McEntire, K. D., Buan, N. R., Robinson, L., and Barbar, E. “Charting a new frontier: Integrating mathematical modeling in complex biological systems from molecules to ecosystems”, Integrative and Comparative Biology 61.6 (2021), pp. 2255–2266. DOI: 10.1093/icb/icab165.DOI
  42. [42]Clarivate. Web of Science API Expanded. Accessed: 2025-06-21. 2025. URL: https://developer.clarivate.com/apis/wos.
  43. [43]Yang, A. et al. “Qwen3 technical report”, arXiv preprint arXiv:2505.09388 (2025). DOI: 10.48550/arXiv.2505.09388.DOI
  44. [44]Brown, T. et al. “Language models are few-shot learners”, Proceedings of the 34th International Conference on Neural Information Processing Systems. Vol. 33. 2020, pp. 1877–1901.arxiv.org/abs/2005.14165
  45. [45]Wang, Z., Pang, Y., and Lin, Y. “Large language models are zero-shot text classifiers”, arXiv preprint arXiv:2312.01044 (2023). DOI: 10.48550/arXiv.2312.01044.DOI
  46. [46]Nayeem, M. T. and Rafiei, D. “Which English do LLMs prefer? Triangulating structural bias towards American English in foundation models”, arXiv preprint arXiv:2604.04204 (2026). DOI: 10.48550/arXiv.2604.04204.
  47. [47]Schulhoff, S. et al. “The prompt report: A systematic survey of prompt engineering techniques”, arXiv preprint arXiv:2406.06608 (2024). DOI: 10.48550/arXiv.2406.06608.DOI
  48. [48]Liu, P., Yuan, W., Fu, J., Jiang, Z., Hayashi, H., and Neubig, G. “Pre-train, prompt, and predict: A systematic survey of prompting methods in natural language processing”, ACM Computing Surveys 55.9 (2023), pp. 1–35. DOI: 10.1145/3560815.DOI
  49. [49]Mu, Y., Wu, B. P., Thorne, W., Robinson, A., Aletras, N., Scarton, C., Bontcheva, K., and Song, X. “Navigating prompt complexity for zero-shot classification: A study of large language models in computational social science”, Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024). 2024, pp. 12074–12086.
  50. [50]Taylor, M. R., Simon, E. J., Dickey, J. L., Hogan, K. A., and Reece, J. B. Campbell biology: Concepts & connections tenth edition. 10th ed. Pearson, 2021. ISBN: 9781292401454.
  51. [51]Nichols, T. E. and Holmes, A. P. “Nonparametric permutation tests for functional neuroimaging: A primer with examples”, Human Brain Mapping 15.1 (2002), pp. 1–25. DOI: 10.1002/hbm.1058.DOI
  52. [52]Phipson, B. and Smyth, G. K. “Permutation P-values Should Never Be Zero: Calculating Exact P-values When Permutations Are Randomly Drawn”, Statistical Applications in Genetics and Molecular Biology 9.1 (2010). DOI: 10.2202/1544–6115.1585. URL: https://doi.org/10.2202/1544-6115.1585.
  53. [53]North, B. V., Curtis, D., and Sham, P. C. “A note on the calculation of empirical P values from Monte Carlo procedures”, The American Journal of Human Genetics 71.2 (2002), pp. 439–441. DOI: 10.1086/341527.DOI
  54. [54]Bracken, L. J. and Oughton, E. A. “‘What do you mean?’ The importance of language in developing interdisciplinary research”, Transactions of the Institute of British Geographers 31 (3 2006). ISSN: 00202754. DOI: 10.1111/j.1475–5661.2006.00218.x.DOI
  55. [55]Petterson, M. B., Longhurst, C., and John-Paul, J. Y. “Measuring interdisciplinarity of biomedical research, medical specialty performance, and implications for radiology: A retrospective review of 2.6 million citations”, Clinical Imaging 80 (2021), pp. 322–328. DOI: 10.1016/j.clinimag.2021.08.012.DOI
  56. [56]Singh, C. K., Barme, E., Ward, R., Tupikina, L., and Santolini, M. “Quantifying the rise and fall of scientific fields”, PLOS One 17.6 (2022), e0270131. DOI: 10.1371/journal.pone.0270131.DOI
  57. [57]Pugh, K., Gyllingberg, L., Stratiev, S., and Hamis, S. “A bibliometric study on mathematical oncology: Interdisciplinarity, internationality, collaboration and trending topics”, Bulletin of Mathematical Biology 87.12 (2025), p. 174. DOI: 10.1007/s11538–025–01544–9.arxiv.org/abs/2506.20684
  58. [58]Öztürk, O., Kocaman, R., and Kanbach, D. K. “How to design bibliometric research: An overview and a framework proposal”, Review of Managerial Science 18 (2024), pp. 3333–3361. DOI: 10.1007/s11846-024-00738-0.DOI
  59. [59]Grames, E. M., Stillman, A. N., Tingley, M. W., and Elphick, C. S. “An automated approach to identifying search terms for systematic reviews using keyword co-occurrence networks”, Methods in Ecology and Evolution 10 (2019), pp. 1645–1654. DOI: 10.1111/2041-210X.13268.DOI
  60. [60]Chen, G. and Xiao, L. “Selecting publication keywords for domain analysis in bibliometrics: A comparison of three methods”, Journal of Informetrics 10 (1 2016), pp. 212–223. ISSN: 18755879. DOI: 10.1016/j.joi.2016.01.006.DOI
  61. [61]Visser, M., Eck, N. J. van, and Waltman, L. “Large-scale comparison of bibliographic data sources: Scopus, Web of Science, Dimensions, Crossref, and Microsoft Academic”, Quantitative Science Studies 2 (1 2021), pp. 20–41. ISSN: 26413337. DOI: 10.1162/qss_a_00112.DOI
  62. [62]Kessler, M. M. “Bibliographic coupling between scientific papers”, American Documentation 14 (1963), pp. 10–25. ISSN: 0096-946X. DOI: 10.1002/asi.5090140103.DOI
  63. [63]Boyack, K. W. and Klavans, R. “Co-citation analysis, bibliographic coupling, and direct citation: Which citation approach represents the research front most accurately?”, Journal of the American Society for Information Science and Technology 61 (12 2010), pp. 2389–2404. ISSN: 15322882. DOI: 10.1002/asi.21419.DOI
  64. [64]Bornmann, L. and Daniel, H.-D. “What do citation counts measure? A review of studies on citing behavior”, Journal of Documentation 64.1 (2008), pp. 45–80. DOI: 10.1108/00220410810844150.DOI
  65. [65]Tahamtan, I. and Bornmann, L. “What do citation counts measure? An updated review of studies on citations in scientific documents published between 2006 and 2018”, Scientometrics 121 (2019), pp. 1635–1684. ISSN: 15882861. DOI: 10.1007/s11192-019-03243-4.DOI
  66. [66]Boyack, K. W., Newman, D., Duhon, R. J., Klavans, R., Patek, M., Biberstine, J. R., Schijvenaars, B., Skupin, A., Ma, N., and Börner, K. “Clustering more than two million biomedical publications: Comparing the accuracies of nine text-based similarity approaches”, PLOS One 6 (3 2011). ISSN: 19326203. DOI: 10.1371/journal.pone.0018029.DOI
  67. [67]Kiss, I. Z., Broom, M., Craze, P. G., and Rafols, I. “Can epidemic models describe the diffusion of topics across disciplines?”, Journal of Informetrics 4.1 (2010), pp. 74–82. DOI: 10.1016/j.joi.2009.08.002.DOI
  68. [68]Howison, J. and Bullard, J. “Software in the scientific literature: Problems with seeing, finding, and using software mentioned in the biology literature”, Journal of the Association for Information Science and Technology 67 (9 2016), pp. 2137–2155. ISSN: 23301643. DOI: 10.1002/asi.23538.DOI
  69. [69]Zhang, Y., Wang, Y., Du, H., and Havlin, S. “Delayed citation impact of interdisciplinary research”, Journal of Informetrics 18.1 (2024), p. 101468. DOI: 10.1016/j.joi.2023.101468.DOI
  70. [70]Lim, W. M., Kumar, S., and Donthu, N. “How to combine and clean bibliometric data and use bibliometric tools synergistically: Guidelines using metaverse research”, Journal of Business Research 182 (2024), p. 114760. ISSN: 01482963. DOI: 10.1016/j.jbusres.2024.114760.DOI
  71. [71]Törnberg, P. “Best practices for text annotation with large language models”, Sociologica 18 (2 2024), pp. 67–85. ISSN: 19718853. DOI: 10.6092/issn.1971-8853/19461.arxiv.org/abs/2402.05129
  72. [72]Rijnsoever, F. J. V. and Hessels, L. K. “Factors associated with disciplinary and interdisciplinary research collaboration”, Research Policy 40 (3 2011), pp. 463–472. ISSN: 00487333. DOI: 10.1016/j.respol.2010.11.001.DOI
  73. [73]MacLeod, M. “What makes interdisciplinarity difficult? Some consequences of domain specificity in interdisciplinary practice”, Synthese 195 (2018), pp. 697–720. ISSN: 15730964. DOI: 10.1007/s11229-016-1236-4.
  74. [74]Jacobs, J. A. In defense of disciplines: Interdisciplinarity and specialization in the research university. Chicago, IL: University of Chicago Press, 2013. DOI: 10.7208/chicago/9780226069463.001.0001.DOI
  75. [75]Bromham, L., Dinnage, R., and Hua, X. “Interdisciplinary research has consistently lower funding success”, Nature 534 (7609 2016), pp. 684–687. ISSN: 14764687. DOI: 10.1038/nature18315.DOI
  76. [76]Leahey, E., Beckman, C. M., and Stanko, T. L. “Prominent but less productive: The impact of interdisciplinarity on scientists’ research”, Administrative Science Quarterly 62 (1 2017). ISSN: 19303815. DOI: 10.1177/0001839216665364.DOI
  77. [77]Levitt, J. M. and Thelwall, M. “Is multidisciplinary research more highly cited? A macrolevel study”, Journal of the American Society for Information Science and Technology 59 (2008), pp. 1973–1984. DOI: 10.1002/asi.20914.DOI
  78. [78]Larivière, V., Haustein, S., and Börner, K. “Long-distance interdisciplinarity leads to higher scientific impact”, PLOS One 10.3 (2015), e0122565. DOI: 10.1371/journal.pone.0122565.DOI
  79. [79]Okamura, K. “Interdisciplinarity revisited: Evidence for research impact and dynamism”, Palgrave Communications 5 (2019), p. 141. DOI: 10.1057/s41599-019-0352-4.DOI
  80. [80]Wang, J., Thijs, B., and Glänzel, W. “Interdisciplinarity and impact: Distinct effects of variety, balance, and disparity”, PLOS One 10.5 (2015), e0127298. DOI: 10.1371/journal.pone.0127298.DOI
  81. [81]Wilensky, U. NetLogo. Center for Connected Learning and Computer-Based Modeling, Northwestern University. Evanston, IL, 1999. URL: http://ccl.northwestern.edu/netlogo/.
  82. [82]Hoeven, E. ter, Kwakkel, J., Hess, V., Pike, T., Wang, B., Rht, and Kazil, J. “Mesa 3: Agent-based modeling with Python in 2025”, Journal of Open Source Software 10.107 (2025), p. 7668. DOI: 10.21105/joss.07668.
  83. [83]North, M. J., Collier, N. T., Ozik, J., Tatara, E. R., Macal, C. M., Bragen, M., and Sydelko, P. “Complex adaptive systems modeling with Repast Simphony”, Complex Adaptive Systems Modeling 1.3 (2013). DOI: 10.1186/2194-3206-1-3.DOI
  84. [84]Schumaker, N. H. and Brookes, A. “HexSim: A modeling environment for ecology and conservation”, Landscape Ecology 33 (2018), pp. 197–211. DOI: 10.1007/s10980-017-0605-9.DOI
  85. [85]Bocedi, G., Palmer, S. C. F., Pe’er, G., Heikkinen, R. K., Matsinos, Y. G., Watts, K., and Travis, J. M. J. “RangeShifter: A platform for modelling spatial eco-evolutionary dynamics and species’ responses to environmental changes”, Methods in Ecology and Evolution 5 (2014), pp. 388–396. DOI: 10.1111/2041-210X.12162.DOI
  86. [86]Kerr, C. C., Stuart, R. M., Mistry, D., Abeysuriya, R. G., Rosenfeld, K., Hart, G. R., et al. “Covasim: An agent-based model of COVID-19 dynamics and interventions”, PLOS Computational Biology 17.7 (2021), e1009149. DOI: 10.1371/journal.pcbi.1009149.DOI
  87. [87]Bershteyn, A., Gerardin, J., Bridenbecker, D., Lorton, C. W., Bloedow, J., Baker, R. S., Chabot-Couture, G., Chen, Y., Fischle, T., Frey, K., et al. “Implementation and applications of EMOD, an individual-based multi-disease modeling platform”, Pathogens and Disease 76.5 (2018), fty059. DOI: 10.1093/femspd/fty059.DOI
  88. [88]Ghaffarizadeh, A., Heiland, R., Friedman, S. H., Mumenthaler, S. M., and Macklin, P. “PhysiCell: An open source physics-based cell simulator for 3-D multicellular systems”, PLOS Computational Biology 14.2 (2018), e1005991. DOI: 10.1371/journal.pcbi.1005991.DOI
  89. [89]Swat, M. H., Thomas, G. L., Belmonte, J. M., Shirinifard, A., Hmeljak, D., and Glazier, J. A. “Multi-scale modeling of tissues using CompuCell3D”, Methods in Cell Biology. Ed. by Asthagiri, A. R. and Arkin, A. P. Vol. 110. Academic Press, 2012. Chap. 13, pp. 325–366. ISBN: 9780123884039. DOI: 10.1016/B978-0-12-388403-9.00013-8.DOI
  90. [90]Mirams, G. R., Arthurs, C. J., Bernabeu, M. O., Bordas, R., Cooper, J., Corrias, A., Davit, Y., Dunn, S.-J., Fletcher, A. G., Harvey, D. G., et al. “Chaste: An open source C++ library for computational physiology and biology”, PLOS Computational Biology 9.3 (2013), e1002970. DOI: 10.1371/journal.pcbi.1002970.DOI
  91. [91]Bravo, R. R., Baratchart, E., West, J., Schenck, R. O., Miller, A. K., Gallaher, J., Gatenbee, C. D., Basanta, D., Robertson-Tessi, M., and Anderson, A. R. “Hybrid Automata Library: A flexible platform for hybrid modeling with real-time visualization”, PLOS Computational Biology 16.3 (2020), e1007635. DOI: 10.1371/journal.pcbi.1007635.DOI
  92. [92]Knapp, B., Bardenet, R., Bernabeu, M. O., Bordas, R., Bruna, M., Calderhead, B., Cooper, J., Fletcher, A. G., Groen, D., Kuijper, B., et al. “Ten simple rules for a successful cross-disciplinary collaboration”, PLOS Computational Biology 11.4 (2015), e1004214. DOI: 10.1371/journal.pcbi.1004214.DOI
  93. [93]Carey, M. A. and Papin, J. A. “Ten simple rules for biologists learning to program”, PLOS Computational Biology 14.1 (2018), e1005871. DOI: 10.1371/journal.pcbi.1005871.DOI
  94. [94]Stawarczyk, B. and Roos, M. “Establishing effective cross-disciplinary collaboration: Combining simple rules for reproducible computational research, a good data management plan, and good research practice”, PLOS Computational Biology 19.4 (2023), e1011052. DOI: 10.1371/journal.pcbi.1011052.DOI
  95. [95]Robinson, M. D., Cai, P., Emons, M., Gerber, R., Germain, P.-L., Gunz, S., Luo, S., Moro, G., Sonder, E., Sonrel, A., et al. “Ten simple rules for computational biologists collaborating with wet lab researchers”, PLOS Computational Biology 20.6 (2024), e1012174. DOI: 10.1371/journal.pcbi.1012174.DOI

Paper details

vlnc:W_5ke5tr2n
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

AI

View statement
Looking for related papers…
Contents