Portal de Periódicos da CAPES

Integration of Advanced Large Language Models into the Construction of Adverse Outcome Pathways: Opportunities and Challenges

2024; American Chemical Society; Volume: 58; Issue: 35 Linguagem: Inglês

10.1021/acs.est.4c07524

1520-5851

Haochun Shi, Yanbin Zhao,

Environmental and Social Impact Assessments

InfoMetricsFiguresRef. Environmental Science & TechnologyASAPArticle This publication is free to access through this site. Learn More CiteCitationCitation and abstractCitation and referencesMore citation options ShareShare onFacebookX (Twitter)WeChatLinkedInRedditEmailJump toExpandCollapse ViewpointAugust 22, 2024Integration of Advanced Large Language Models into the Construction of Adverse Outcome Pathways: Opportunities and ChallengesClick to copy article linkArticle link copied!Haochun ShiHaochun ShiState Environmental Protection Key Laboratory of Environmental Health Impact Assessment of Emerging Contaminants, School of Environmental Science and Engineering, Shanghai Jiao Tong University, 800 Dongchuan Road, Shanghai 200240, ChinaMore by Haochun ShiYanbin Zhao*Yanbin ZhaoState Environmental Protection Key Laboratory of Environmental Health Impact Assessment of Emerging Contaminants, School of Environmental Science and Engineering, Shanghai Jiao Tong University, 800 Dongchuan Road, Shanghai 200240, China*+86 188 1820 7732, [email protected]More by Yanbin ZhaoView Biographyhttps://orcid.org/0000-0001-5632-5371Open PDFEnvironmental Science & TechnologyCite this: Environ. Sci. Technol. 2024, XXXX, XXX, XXX-XXXClick to copy citationCitation copied!https://pubs.acs.org/doi/10.1021/acs.est.4c07524https://doi.org/10.1021/acs.est.4c07524Published August 22, 2024 Publication History Received 22 July 2024Published online 22 August 2024article-commentary© 2024 American Chemical Society. This publication is available under these Terms of Use. Request reuse permissionsThis publication is licensed for personal use by The American Chemical Society. ACS Publications© 2024 American Chemical SocietySubjectswhat are subjectsArticle subjects are automatically applied from the ACS Subject Taxonomy and describe the scientific concepts and themes of the article.AnatomyDepositionEnvironmental modelingEnvironmental pollutionQuality managementThe adverse outcome pathway (AOP) concept has attracted a significant amount of attention in recent years. It connects a molecular initiating event (MIE) to an adverse outcome (AO) at higher levels of biological organization through key events (KEs) specified by key event relationships (KERs). (1) This approach to interpreting mechanistic toxicological information began in 2010. Since then, more than 400 AOPs have been established, as documented in AOPwiki (https://aopwiki.org/). They enhance the understanding of toxicity mechanisms and assist in predicting the potential impacts of chemical exposures, which is crucial for risk assessment, regulatory decision making, and the development of safer chemicals and pharmaceuticals. (2)However, constructing AOPs is a complex process that still faces numerous challenges. One major challenge is to effectively integrate a vast amount of scientific information dispersed across a wide array of literature. Manual integration from disparate sources, as commonly performed recently, demands substantial effort and expertise, and this approach carries a significant risk of overlooking key insights or introducing errors, often leading to incomplete and inaccurate integration. (3) Moreover, constructing AOPs requires multidisciplinary knowledge and expertise. The process requires proficiency in toxicology, developmental biology, physiology, and even ecology, necessitating a broad knowledge base. However, few researchers possess comprehensive expertise across all of these areas, thereby limiting the depth and quality of AOPs. Furthermore, the standardization and normalization of AOP construction remain insufficient. While several guidelines exist to guide the AOP development and assessment, (4) their interpretation and implementation vary among researchers, leading to inconsistencies in AOP quality. In light of these challenges, there is a clear demand for innovative solutions in AOP construction.Large language models (LLMs), such as GPT-4 and Claude-3, are designed for advanced language understanding and generation. Recent studies have demonstrated their effectiveness in various fields, including biomedicine and healthcare. (5) For instance, GPT-4 exhibited proficient capabilities in cell type annotation within single-cell RNA sequencing analysis, leading to the transition in the annotation process from manual to a semi-automated or even fully automated procedure. This advancement offered cost efficiency and seamless integration into existing single-cell RNA-seq analysis pipelines, eliminating the need for developing additional pipelines or sourcing high-quality reference data sets. (6) Hence, given the attributes of LLMs, including GPT-4, there is a compelling rationale to consider their utilization for assisting in AOP construction.LLMs, with their advanced natural language processing capabilities, have shown potential to accelerate the process of extracting relevant information from scientific literature and databases. (7) They could help to efficiently filter through a wide array of literature sources, identify key insights, and integrate them into AOP construction. This capability would significantly reduce the time and effort required for data identification and synthesis and minimize the risk of overlooking crucial information. LLMs also specialize in knowledge synthesis by integrating information from diverse sources, filling knowledge gaps, and providing a more comprehensive view of biological pathways. Their pattern recognition abilities allow them to identify relationships within complex biological data, essential for mapping out the intricate networks involved in AOPs. LLMs can also bridge the multidisciplinary knowledge gap by serving as both a knowledge repository and an expert assistant. They offer researchers access to a wide range of multidisciplinary information and insights, helping them make informed decisions and fill knowledge gaps. They also assist researchers in integrating and interpreting data from diverse fields, thereby enhancing the depth and quality of AOPs. Moreover, LLMs can contribute to standardization efforts by providing consistent and objective interpretations of the guidelines and recommendations from OECD and other organizations. (2) They can assist researchers in adhering to standardized protocols and best practices, thereby improving the consistency and quality of AOP construction across various research groups.To better understand the effectiveness of LLMs in AOP construction, here we reconstructed five well-documented AOPs described in the AOPwiki using the representative model, GPT-4, for automated annotation. The flowchart for reconstruction and results are depicted in Figure 1. The detailed procedure for the reconstruction and verification process is available on GitHub (https://github.com/ShiHaochun/AOP-project). As shown in panels B and C of Figure 1, the results demonstrate a high degree of structural consistency between the GPT-4-generated AOPs and expert-validated AOPs. GPT-4 accurately identified the MIEs, mapped KEs, and linked these events to AOs with a high degree of accuracy. In the case of the AOP related to liver fibrosis, originating from protein alkylation, GPT-4 successfully delineated the subsequent key events, including the induction of a cellular stress response and inflammation, activation of hepatic stellate cells, and deposition of an extracellular matrix, and correctly linked the final adverse outcome, liver fibrosis. Meanwhile, GPT-4 identified an additional critical step, the excessive deposition of ECM proteins that results in the formation of scar tissue and fibrotic lesions. These alterations disrupt the normal architecture and functions of the liver, impacting crucial processes such as blood flow regulation, bile production, and detoxification and finally leading to liver fibrosis, as described previously. (8) This insight provided by GPT-4 extends beyond the representation found in the graphical network within the AOPwiki, thereby emphasizing the capacity of GPT-4 to facilitate a comprehensive understanding of the complex biological processes associated with liver fibrosis. Another example includes the AOP delineating aromatase inhibition leading to reproductive dysfunction, (9) wherein GPT-4 successfully outlined the key events such as decreased levels of estrogen and vitellogenin synthesis, impaired gonadal development, and effectively linked them to the reduction of cumulative fecundity and spawning. Hence, these results demonstrate that GPT-4 can effectively participate in AOP construction, providing reliable drafts that closely align with expert-validated frameworks.Figure 1Figure 1. Integration of GPT-4 into the construction of adverse outcome pathways (AOPs). (A) Flowchart of GPT-4's annotation in AOP development and assessment, taking liver fibrosis as an example. (B and C) Reconstruction of five well-documented AOPs utilizing GPT-4's automated annotation and comparison with the corresponding expert-validated AOP information documented in the AOPwiki. The identified molecular initiating events (MIEs), mapped key events (KEs), and their linkages to adverse outcomes (AOs) are depicted. Green indicates identical MIEs and KEs. Red indicates mismatched or missed KEs in GPT-4 annotation. Blue indicates missed KEs in the AOP framework documented in the AOPwiki.High Resolution ImageDownload MS PowerPoint SlideWhile promising results were generated, it should be noted that achieving high accuracy in AOP construction using LLMs still poses significant challenges. There are still limitations and areas for improvement that need to be addressed. First, the undisclosed nature of LLMs' training corpus makes verifying the basis of their annotations challenging. (6) While LLMs demonstrate proficiency in natural language processing, if their training data do not adequately cover the breadth of the relevant scientific literature, their understanding of complex biological concepts and mechanisms may be limited. AOP construction requires a profound understanding of intricate biochemical pathways, physiological processes, and toxicological mechanisms, which may not be adequately captured by the training data or algorithms of the LLMs. As a result, there is a risk of oversimplification or misinterpretation of scientific data, leading to inaccuracies or deficiencies in the AOP construction. Other challenges are the quality and completeness of the input data during AOP construction. The outputs of LLMs are only as reliable as the data on which they are trained. Incomplete or low-quality data extracted from diverse sources can lead to gaps or inaccuracies in AOP construction. Moreover, LLMs can easily generate a large number of biological pathways during AOP construction, leading to an abundance of intricate information. Consequently, quality control becomes a critical issue for ensuring the effectiveness of AOP construction.To address these issues, enhancing the training data sets with more specialized and high-quality biological and toxicological data is essential. Collaborating with human experts to curate and refine these data sets can significantly improve the accuracy and reliability of the model. In addition, developing hybrid models that combine LLMs with other AI models or expert systems may also help. By integrating different approaches, hybrid models exhibit the potential to generate more comprehensive and accurate AOPs. For example, integrating machine learning models specialized in bioinformatics with GPT-4's NLP capabilities can enhance the interpretation and mapping of complex biological pathways. (10) Moreover, creating interactive platforms that facilitate seamless collaboration between LLMs and human experts can further improve the efficiency and accuracy of AOP construction. These platforms enable real-time feedback and validation, allowing experts to refine and validate AOPs generated by LLMs iteratively.In conclusion, the integration of advanced large language models into AOP construction presents significant opportunities for addressing existing challenges and improving the efficiency and accuracy of AOP frameworks. Despite current limitations, ongoing advancements in AI technology, coupled with collaborative efforts between AI systems and human experts, hold promise for the future of AOP construction. By leveraging the strengths of LLMs, we can enhance our understanding of adverse effects caused by environmental pollution and better protect public health through more effective risk assessment and regulatory decision making.Author InformationClick to copy section linkSection link copied!Corresponding AuthorYanbin Zhao - State Environmental Protection Key Laboratory of Environmental Health Impact Assessment of Emerging Contaminants, School of Environmental Science and Engineering, Shanghai Jiao Tong University, 800 Dongchuan Road, Shanghai 200240, China; https://orcid.org/0000-0001-5632-5371; Email: [email protected]AuthorHaochun Shi - State Environmental Protection Key Laboratory of Environmental Health Impact Assessment of Emerging Contaminants, School of Environmental Science and Engineering, Shanghai Jiao Tong University, 800 Dongchuan Road, Shanghai 200240, ChinaNotesThe authors declare no competing financial interest.BiographyClick to copy section linkSection link copied!Yanbin ZhaoHigh Resolution ImageDownload MS PowerPoint SlideYanbin Zhao is an Associate Professor at Shanghai Jiao Tong University. He obtained his B.S. from Northwest A&F University and his Ph.D. from Peking University. He serves as an editorial board member of Eco-Environment & Health and ACS ES&T Water and as a member of the Professional Committee of Birth Defect Prevention and Control of the Chinese Environmental Mutagen Society. His research interests focus on the environmental behavior and toxic effects of emerging pollutants by employing molecular biology, bioinformatics, and analytical chemistry technologies. He has led more than 10 research projects, including those funded by the National Natural Science Foundation of China and the National Key Research and Development Program of China, and has published more than 50 research papers in peer-reviewed journals.AcknowledgmentsClick to copy section linkSection link copied!This research was supported by the National Natural Science Foundation of China (22122605) and the National Key Research and Development Program of China (2022YFC3902100).ReferencesClick to copy section linkSection link copied! This article references 10 other publications. 1Ankley, G. T.; Bennett, R. S.; Erickson, R. J.; Hoff, D. J.; Hornung, M. W.; Johnson, R. D.; Mount, D. R.; Nichols, J. W.; Russom, C. L.; Schmieder, P. K.; Serrrano, J. A.; Tietge, J. E.; Villeneuve, D. L. Adverse outcome pathways: a conceptual framework to support ecotoxicology research and risk assessment. Environ. Toxicol. Chem. 2010, 29 (3), 730– 741, DOI: 10.1002/etc.34 Google Scholar1Adverse outcome pathways: a conceptual framework to support ecotoxicology research and risk assessmentAnkley, Gerald T.; Bennett, Richard S.; Erickson, Russell J.; Hoff, Dale J.; Hornung, Michael W.; Johnson, Rodney D.; Mount, David R.; Nichols, John W.; Russom, Christine L.; Schmieder, Patricia K.; Serrrano, Jose A.; Tietge, Joseph E.; Villeneuve, Daniel L.Environmental Toxicology and Chemistry (2010), 29 (3), 730-741CODEN: ETOCDK; ISSN:0730-7268. (John Wiley & Sons Ltd.) A review. Ecol. risk assessors face increasing demands to assess more chems., with greater speed and accuracy, and to do so using fewer resources and exptl. animals. New approaches in biol. and computational sciences may be able to generate mechanistic information that could help in meeting these challenges. However, to use mechanistic data to support chem. assessments, there is a need for effective translation of this information into endpoints meaningful to ecol. risk - effects on survival, development, and reprodn. in individual organisms and, by extension, impacts on populations. Here we discuss a framework designed for this purpose, the adverse outcome pathway (AOP). An AOP is a conceptual construct that portrays existing knowledge concerning the linkage between a direct mol. initiating event and an adverse outcome at a biol. level of organization relevant to risk assessment. The practical utility of AOPs for ecol. risk assessment of chems. is illustrated using five case examples. The examples demonstrate how the AOP concept can focus toxicity testing in terms of species and endpoint selection, enhance across-chem. extrapolation, and support prediction of mixt. effects. The examples also show how AOPs facilitate use of mol. or biochem. endpoints (sometimes referred to as biomarkers) for forecasting chem. impacts on individuals and populations. In the concluding sections of the paper, we discuss how AOPs can help to guide research that supports chem. risk assessments and advocate for the incorporation of this approach into a broader systems biol. framework. >> More from SciFinder ®https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC3cXjt12ju7Y%253D&md5=3b2dfc0652b76d1bed60b8885091dbc22Carusi, A.; Davies, M. R.; De Grandis, G.; Escher, B. I.; Hodges, G.; Leung, K. M. Y.; Whelan, M.; Willett, C.; Ankley, G. T. Harvesting the promise of AOPs: An assessment and recommendations. Sci. Total Environ. 2018, 628–629, 1542– 1556, DOI: 10.1016/j.scitotenv.2018.02.015 Google Scholar2Harvesting the promise of AOPs: An assessment and recommendationsCarusi, Annamaria; Davies, Mark R.; De Grandis, Giovanni; Escher, Beate I.; Hodges, Geoff; Leung, Kenneth M. Y.; Whelan, Maurice; Willett, Catherine; Ankley, Gerald T.Science of the Total Environment (2018), 628-629 (), 1542-1556CODEN: STENDL; ISSN:0048-9697. (Elsevier B.V.) A review. The Adverse Outcome Pathway (AOP) concept is a knowledge assembly and communication tool to facilitate the transparent translation of mechanistic information into outcomes meaningful to the regulatory assessment of chems. The AOP framework and assocd. knowledgebases (KBs) have received significant attention and use in the regulatory toxicol. community. However, it is increasingly apparent that the potential stakeholder community for the AOP concept and AOP KBs is broader than scientists and regulators directly involved in chem. safety assessment. In this paper we identify and describe those stakeholders who currently-or in the future-could benefit from the application of the AOP framework and knowledge to specific problems. We also summarize the challenges faced in implementing pathway-based approaches such as the AOP framework in biol. sciences, and provide a series of recommendations to meet crit. needs to ensure further progression of the framework as a useful, sustainable and dependable tool supporting assessments of both human health and the environment. Although the AOP concept has the potential to significantly impact the organization and interpretation of biol. information in a variety of disciplines/applications, this promise can only be fully realized through the active engagement of, and input from multiple stakeholders, requiring multi-pronged substantive long-term planning and strategies. >> More from SciFinder ®https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC1cXjtFCksLw%253D&md5=6bfe101adb982bf676e8e64692b704813Carvaillo, J. C.; Barouki, R.; Coumoul, X.; Audouze, K. Linking bisphenol S to adverse outcome pathways using a combined text mining and systems biology approach. Environ. Health Perspect. 2019, 127 (4), 47005, DOI: 10.1289/EHP4200 Google Scholar3Linking Bisphenol S to Adverse Outcome Pathways Using a Combined Text Mining and Systems Biology ApproachCarvaillo Jean-Charles; Barouki Robert; Coumoul Xavier; Audouze Karine; Carvaillo Jean-Charles; Barouki Robert; Coumoul Xavier; Audouze Karine; Audouze KarineEnvironmental health perspectives (2019), 127 (4), 47005 ISSN:. BACKGROUND: Available toxicity data can be optimally interpreted if they are integrated using computational approaches such as systems biology modeling. Such approaches are particularly warranted in cases where regulatory decisions have to be made rapidly. OBJECTIVES: The study aims at developing and applying a new integrative computational strategy to identify associations between bisphenol S (BPS), a substitute for bisphenol A (BPA), and components of adverse outcome pathways (AOPs). METHODS: The proposed approach combines a text mining (TM) procedure and integrative systems biology to comprehensively analyze the scientific literature to enrich AOPs related to environmental stressors. First, to identify relevant associations between BPS and different AOP components, a list of abstracts was screened using the developed text-mining tool AOP-helpFinder, which calculates scores based on the graph theory to prioritize the findings. Then, to fill gaps between BPS, biological events, and adverse outcomes (AOs), a systems biology approach was used to integrate information from the AOP-Wiki and ToxCast databases, followed by manual curation of the relevant publications. RESULTS: Links between BPS and 48 AOP key events (KEs) were identified and scored via 31 references. The main outcomes were related to reproductive health, endocrine disruption, impairments of metabolism, and obesity. We then explicitly analyzed co-mention of the terms BPS and obesity by data integration and manual curation of the full text of the publications. Several molecular and cellular pathways were identified, which allowed the proposal of a biological explanation for the association between BPS and obesity. CONCLUSIONS: By analyzing dispersed information from the literature and databases, our novel approach can identify links between stressors and AOP KEs. The findings associating BPS and obesity illustrate the use of computational tools in predictive toxicology and highlight the relevance of the approach to decision makers assessing substituents to toxic chemicals. https://doi.org/10.1289/EHP4200. >> More from SciFinder ®https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A280%3ADC%252BB3M%252FmtlKrsw%253D%253D&md5=8aa032fd31db5bfb878d13589f572b2d4Svingen, T.; Villeneuve, D. L.; Knapen, D.; Panagiotou, E. M.; Draskau, M. K.; Damdimopoulou, P.; O'Brien, J. M. A pragmatic approach to adverse outcome pathway development and evaluation. Toxicol. Sci. 2021, 184 (2), 183– 190, DOI: 10.1093/toxsci/kfab113 Google ScholarThere is no corresponding record for this reference.5Thirunavukarasu, A. J.; Ting, D. S. J.; Elangovan, K.; Gutierrez, L.; Tan, T. F.; Ting, D. S. W. Large language models in medicine. Nat. Med. 2023, 29, 1930– 1940, DOI: 10.1038/s41591-023-02448-8 Google ScholarThere is no corresponding record for this reference.6Hou, W.; Ji, Z. Assessing GPT-4 for cell type annotation in single-cell RNA-seq analysis. Nat. Methods. 2024, 21, 1462, DOI: 10.1038/s41592-024-02235-4 Google ScholarThere is no corresponding record for this reference.7Polak, M. P.; Morgan, D. Extracting accurate materials data from research papers with conversational language models and prompt engineering. Nat. Commun. 2024, 15 (1), 1569, DOI: 10.1038/s41467-024-45914-8 Google ScholarThere is no corresponding record for this reference.8Schuppan, D.; Ashfaq-Khan, M.; Yang, A. T.; Kim, Y. O. Liver fibrosis: Direct antifibrotic agents and targeted therapies. Matrix Biol. 2018, 68–69, 435– 451, DOI: 10.1016/j.matbio.2018.04.006 Google Scholar8Liver fibrosis: Direct antifibrotic agents and targeted therapiesSchuppan, Detlef; Ashfaq-Khan, Muhammad; Yang, Ai Ting; Kim, Yong OokMatrix Biology (2018), 68-69 (), 435-451CODEN: MTBOEC; ISSN:0945-053X. (Elsevier B.V.) A review. Moreover, drugs that would speed up fibrosis reversal are needed for patients with cirrhosis, since even with effective causal therapy reversal is slow or the disease may further progress. Therefore, highly efficient and specific antifibrotic agents are needed that can address advanced fibrosis, i.e., the detrimental downstream result of all chronic liver diseases. This review discusses targeted antifibrotic therapies that address mols. and mechanisms that are central to fibrogenesis or fibrolysis, including strategies that allow targeting of activated hepatic stellate cells and myofibroblasts and other fibrogenic effector cells. Focus is on collagen synthesis, integrins and cells and mechanisms specific including specific downregulation of TGFbeta signaling, major extracellular matrix (ECM) components, ECM-crosslinking, and ECM-receptors such as integrins and discoidin domain receptors, ECM-crosslinking and methods for targeted delivery of small interfering RNA, antisense oligonucleotides and small mols. to increase potency and reduce side effects. With an increased understanding of the biol. of the ECM and liver fibrosis and an improved preclin. validation, the translation of these approaches to the clinic is currently ongoing. Application to patients with liver fibrosis and a personalized treatment is tightly linked to the development of noninvasive biomarkers of fibrosis, fibrogenesis and fibrolysis. >> More from SciFinder ®https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC1cXotlGitLs%253D&md5=b8d068dad20f40cb027a233fb088b5019Conolly, R. B.; Ankley, G. T.; Cheng, W.; Mayo, M. L.; Miller, D. H.; Perkins, E. J.; Villeneuve, D. L.; Watanabe, K. H. Quantitative adverse outcome pathways and their application to predictive toxicology. Environ. Sci. Technol. 2017, 51 (8), 4661– 4672, DOI: 10.1021/acs.est.6b06230 Google Scholar9Quantitative Adverse Outcome Pathways and Their Application to Predictive ToxicologyConolly, Rory B.; Ankley, Gerald T.; Cheng, WanYun; Mayo, Michael L.; Miller, David H.; Perkins, Edward J.; Villeneuve, Daniel L.; Watanabe, Karen H.Environmental Science & Technology (2017), 51 (8), 4661-4672CODEN: ESTHAG; ISSN:0013-936X. (American Chemical Society) A review. A quant. adverse outcome pathway (qAOP) consists of one or more biol. based, computational models describing key event relationships linking a mol. initiating event (MIE) to an adverse outcome. A qAOP provides quant., dose-response and time course predictions that can support regulatory decision-making. Herein the authors describe several facets of qAOPs, including (a) motivation for development, (b) tech. considerations, (c) evaluation of confidence, and (d) potential applications. The qAOP used as an illustrative example for these points describes the linkage between inhibition of cytochrome P 450 19A aromatase (the MIE) and population-level decreases in the fathead minnow (FHM; Pimephales promelas). The qAOP consists of three linked computational models for: (a) the hypothalamic-pituitary-gonadal axis in female FHMs, where aromatase inhibition decreases the conversion of testosterone to 17β estradiol (E2), thereby reducing E2-dependent vitellogenin (VTG; egg yolk protein precursor) synthesis, (b) VTG-dependent egg development and spawning (fecundity), and (c) fecundity-dependent population trajectory. While development of the example qAOP was based on expts. with FHMs exposed to the aromatase inhibitor fadrozole, the authors also show how a toxic equivalence (TEQ) calcn. allows use of the qAOP to predict effects of another, untested aromatase inhibitor, iprodione. While qAOP development can be resource-intensive, the quant. predictions obtained, and TEQ-based application to multiple chems., may be sufficient to justify the cost for some applications in regulatory decision-making. >> More from SciFinder ®https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC2sXltlWhu7Y%253D&md5=a16c89301fbdda3c09055137e9f9321510Jin, Q.; Yang, Y.; Chen, Q.; Lu, Z. GeneGPT: augmenting large language models with domain tools for improved access to biomedical information. Bioinformatics 2024, 40 (2), btae075 DOI: 10.1093/bioinformatics/btae075 Google ScholarThere is no corresponding record for this reference.Cited By Click to copy section linkSection link copied!This article has not yet been cited by other publications.Download PDFFiguresReferencesOpen PDF Get e-AlertsGet e-AlertsEnvironmental Science & TechnologyCite this: Environ. Sci. Technol. 2024, XXXX, XXX, XXX-XXXClick to copy citationCitation copied!https://doi.org/10.1021/acs.est.4c07524Published August 22, 2024 Publication History Received 22 July 2024Published online 22 August 2024© 2024 American Chemical Society. This publication is available under these Terms of Use. Request reuse permissionsArticle Views-Altmetric-Citations-Learn about these metrics closeArticle Views are the COUNTER-compliant sum of full text article downloads since November 2008 (both PDF and HTML) across all institutions and individuals. These metrics are regularly updated to reflect usage leading up to the last few days.Citations are the number of other articles citing this article, calculated by Crossref and updated daily. Find more information about Crossref citation counts.The Altmetric Attention Score is a quantitative measure of the attention that a research article has received online. Clicking on the donut icon will load a page at altmetric.com with additional details about the score and the social media presence for the given article. Find more information on the Altmetric Attention Score and how the score is calculated.Recommended Articles FiguresReferencesAbstractHigh Resolution ImageDownload MS PowerPoint SlideFigure 1Figure 1. Integration of GPT-4 into the construction of adverse outcome pathways (AOPs). (A) Flowchart of GPT-4's annotation in AOP development and assessment, taking liver fibrosis as an example. (B and C) Reconstruction of five well-documented AOPs utilizing GPT-4's automated annotation and comparison with the corresponding expert-validated AOP information documented in the AOPwiki. The identified molecular initiating events (MIEs), mapped key events (KEs), and their linkages to adverse outcomes (AOs) are depicted. Green indicates identical MIEs and KEs. Red indicates mismatched or missed KEs in GPT-4 annotation. Blue indicates missed KEs in the AOP framework documented in the AOPwiki.High Resolution ImageDownload MS PowerPoint SlideYanbin ZhaoHigh Resolution ImageDownload MS PowerPoint SlideYanbin Zhao is an Associate Professor at Shanghai Jiao Tong University. He obtained his B.S. from Northwest A&F University and his Ph.D. from Peking University. He serves as an editorial board member of Eco-Environment & Health and ACS ES&T Water and as a member of the Professional Committee of Birth Defect Prevention and Control of the Chinese Environmental Mutagen Society. His research interests focus on the environmental behavior and toxic effects of emerging pollutants by employing molecular biology, bioinformatics, and analytical chemistry technologies. He has led more than 10 research projects, including those funded by the National Natural Science Foundation of China and the National Key Research and Development Program of China, and has published more than 50 research papers in peer-reviewed journals.References This article references 10 other publications. 1Ankley, G. T.; Bennett, R. S.; Erickson, R. J.; Hoff, D. J.; Hornung, M. W.; Johnson, R. D.; Mount, D. R.; Nichols, J. W.; Russom, C. L.; Schmieder, P. K.; Serrrano, J. A.; Tietge, J. E.; Villeneuve, D. L. Adverse outcome pathways: a conceptual framework to support ecotoxicology research and risk assessment. Environ. Toxicol. Chem. 2010, 29 (3), 730– 741, DOI: 10.1002/etc.34 1Adverse outcome pathways: a conceptual framework to support ecotoxicology research and risk assessmentAnkley, Gerald T.; Bennett, Richard S.; Erickson, Russell J.; Hoff, Dale J.; Hornung, Michael W.; Johnson, Rodney D.; Mount, David R.; Nichols, John W.; Russom, Christine L.; Schmieder, Patricia K.; Serrrano, Jose A.; Tietge, Joseph E.; Villeneuve, Daniel L.Environmental Toxicology and Chemistry (2010), 29 (3), 730-741CODEN: ETOCDK; ISSN:0730-7268. (John Wiley & Sons Ltd.) A review. Ecol. risk assessors face increasing demands to assess more chems., with greater speed and accuracy, and to do so using fewer resources and exptl. animals. New approaches in biol. and computational sciences may be able to generate mechanistic information that could help in meeting these challenges. However, to use mechanistic data to support chem. assessments, there is a need for effective translation of this information into endpoints meaningful to ecol. risk - effects on survival, development, and reprodn. in individual organisms and, by extension, impacts on populations. Here we discuss a framework designed for this purpose, the adverse outcome pathway (AOP). An AOP is a conceptual construct that portrays existing knowledge concerning the linkage between a direct mol. initiating event and an adverse outcome at a biol. level of organization relevant to risk assessment. The practical utility of AOPs for ecol. risk assessment of chems. is illustrated using five case examples. The examples demonstrate how the AOP concept can focus toxicity testing in terms of species and endpoint selection, enhance across-chem. extrapolation, and support prediction of mixt. effects. The examples also show how AOPs facilitate use of mol. or biochem. endpoints (sometimes referred to as biomarkers) for forecasting chem. impacts on individuals and populations. In the concluding sections of the paper, we discuss how AOPs can help to guide research that supports chem. risk assessments and advocate for the incorporation of this approach into a broader systems biol. framework. >> More from SciFinder ®https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC3cXjt12ju7Y%253D&md5=3b2dfc0652b76d1bed60b8885091dbc22Carusi, A.; Davies, M. R.; De Grandis, G.; Escher, B. I.; Hodges, G.; Leung, K. M. Y.; Whelan, M.; Willett, C.; Ankley, G. T. Harvesting the promise of AOPs: An assessment and recommendations. Sci. Total Environ. 2018, 628–629, 1542– 1556, DOI: 10.1016/j.scitotenv.2018.02.015 2Harvesting the promise of AOPs: An assessment and recommendationsCarusi, Annamaria; Davies, Mark R.; De Grandis, Giovanni; Escher, Beate I.; Hodges, Geoff; Leung, Kenneth M. Y.; Whelan, Maurice; Willett, Catherine; Ankley, Gerald T.Science of the Total Environment (2018), 628-629 (), 1542-1556CODEN: STENDL; ISSN:0048-9697. (Elsevier B.V.) A review. The Adverse Outcome Pathway (AOP) concept is a knowledge assembly and communication tool to facilitate the transparent translation of mechanistic information into outcomes meaningful to the regulatory assessment of chems. The AOP framework and assocd. knowledgebases (KBs) have received significant attention and use in the regulatory toxicol. community. However, it is increasingly apparent that the potential stakeholder community for the AOP concept and AOP KBs is broader than scientists and regulators directly involved in chem. safety assessment. In this paper we identify and describe those stakeholders who currently-or in the future-could benefit from the application of the AOP framework and knowledge to specific problems. We also summarize the challenges faced in implementing pathway-based approaches such as the AOP framework in biol. sciences, and provide a series of recommendations to meet crit. needs to ensure further progression of the framework as a useful, sustainable and dependable tool supporting assessments of both human health and the environment. Although the AOP concept has the potential to significantly impact the organization and interpretation of biol. information in a variety of disciplines/applications, this promise can only be fully realized through the active engagement of, and input from multiple stakeholders, requiring multi-pronged substantive long-term planning and strategies. >> More from SciFinder ®https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC1cXjtFCksLw%253D&md5=6bfe101adb982bf676e8e64692b704813Carvaillo, J. C.; Barouki, R.; Coumoul, X.; Audouze, K. Linking bisphenol S to adverse outcome pathways using a combined text mining and systems biology approach. Environ. Health Perspect. 2019, 127 (4), 47005, DOI: 10.1289/EHP4200 3Linking Bisphenol S to Adverse Outcome Pathways Using a Combined Text Mining and Systems Biology ApproachCarvaillo Jean-Charles; Barouki Robert; Coumoul Xavier; Audouze Karine; Carvaillo Jean-Charles; Barouki Robert; Coumoul Xavier; Audouze Karine; Audouze KarineEnvironmental health perspectives (2019), 127 (4), 47005 ISSN:. BACKGROUND: Available toxicity data can be optimally interpreted if they are integrated using computational approaches such as systems biology modeling. Such approaches are particularly warranted in cases where regulatory decisions have to be made rapidly. OBJECTIVES: The study aims at developing and applying a new integrative computational strategy to identify associations between bisphenol S (BPS), a substitute for bisphenol A (BPA), and components of adverse outcome pathways (AOPs). METHODS: The proposed approach combines a text mining (TM) procedure and integrative systems biology to comprehensively analyze the scientific literature to enrich AOPs related to environmental stressors. First, to identify relevant associations between BPS and different AOP components, a list of abstracts was screened using the developed text-mining tool AOP-helpFinder, which calculates scores based on the graph theory to prioritize the findings. Then, to fill gaps between BPS, biological events, and adverse outcomes (AOs), a systems biology approach was used to integrate information from the AOP-Wiki and ToxCast databases, followed by manual curation of the relevant publications. RESULTS: Links between BPS and 48 AOP key events (KEs) were identified and scored via 31 references. The main outcomes were related to reproductive health, endocrine disruption, impairments of metabolism, and obesity. We then explicitly analyzed co-mention of the terms BPS and obesity by data integration and manual curation of the full text of the publications. Several molecular and cellular pathways were identified, which allowed the proposal of a biological explanation for the association between BPS and obesity. CONCLUSIONS: By analyzing dispersed information from the literature and databases, our novel approach can identify links between stressors and AOP KEs. The findings associating BPS and obesity illustrate the use of computational tools in predictive toxicology and highlight the relevance of the approach to decision makers assessing substituents to toxic chemicals. https://doi.org/10.1289/EHP4200. >> More from SciFinder ®https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A280%3ADC%252BB3M%252FmtlKrsw%253D%253D&md5=8aa032fd31db5bfb878d13589f572b2d4Svingen, T.; Villeneuve, D. L.; Knapen, D.; Panagiotou, E. M.; Draskau, M. K.; Damdimopoulou, P.; O'Brien, J. M. A pragmatic approach to adverse outcome pathway development and evaluation. Toxicol. Sci. 2021, 184 (2), 183– 190, DOI: 10.1093/toxsci/kfab113 There is no corresponding record for this reference.5Thirunavukarasu, A. J.; Ting, D. S. J.; Elangovan, K.; Gutierrez, L.; Tan, T. F.; Ting, D. S. W. Large language models in medicine. Nat. Med. 2023, 29, 1930– 1940, DOI: 10.1038/s41591-023-02448-8 There is no corresponding record for this reference.6Hou, W.; Ji, Z. Assessing GPT-4 for cell type annotation in single-cell RNA-seq analysis. Nat. Methods. 2024, 21, 1462, DOI: 10.1038/s41592-024-02235-4 There is no corresponding record for this reference.7Polak, M. P.; Morgan, D. Extracting accurate materials data from research papers with conversational language models and prompt engineering. Nat. Commun. 2024, 15 (1), 1569, DOI: 10.1038/s41467-024-45914-8 There is no corresponding record for this reference.8Schuppan, D.; Ashfaq-Khan, M.; Yang, A. T.; Kim, Y. O. Liver fibrosis: Direct antifibrotic agents and targeted therapies. Matrix Biol. 2018, 68–69, 435– 451, DOI: 10.1016/j.matbio.2018.04.006 8Liver fibrosis: Direct antifibrotic agents and targeted therapiesSchuppan, Detlef; Ashfaq-Khan, Muhammad; Yang, Ai Ting; Kim, Yong OokMatrix Biology (2018), 68-69 (), 435-451CODEN: MTBOEC; ISSN:0945-053X. (Elsevier B.V.) A review. Moreover, drugs that would speed up fibrosis reversal are needed for patients with cirrhosis, since even with effective causal therapy reversal is slow or the disease may further progress. Therefore, highly efficient and specific antifibrotic agents are needed that can address advanced fibrosis, i.e., the detrimental downstream result of all chronic liver diseases. This review discusses targeted antifibrotic therapies that address mols. and mechanisms that are central to fibrogenesis or fibrolysis, including strategies that allow targeting of activated hepatic stellate cells and myofibroblasts and other fibrogenic effector cells. Focus is on collagen synthesis, integrins and cells and mechanisms specific including specific downregulation of TGFbeta signaling, major extracellular matrix (ECM) components, ECM-crosslinking, and ECM-receptors such as integrins and discoidin domain receptors, ECM-crosslinking and methods for targeted delivery of small interfering RNA, antisense oligonucleotides and small mols. to increase potency and reduce side effects. With an increased understanding of the biol. of the ECM and liver fibrosis and an improved preclin. validation, the translation of these approaches to the clinic is currently ongoing. Application to patients with liver fibrosis and a personalized treatment is tightly linked to the development of noninvasive biomarkers of fibrosis, fibrogenesis and fibrolysis. >> More from SciFinder ®https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC1cXotlGitLs%253D&md5=b8d068dad20f40cb027a233fb088b5019Conolly, R. B.; Ankley, G. T.; Cheng, W.; Mayo, M. L.; Miller, D. H.; Perkins, E. J.; Villeneuve, D. L.; Watanabe, K. H. Quantitative adverse outcome pathways and their application to predictive toxicology. Environ. Sci. Technol. 2017, 51 (8), 4661– 4672, DOI: 10.1021/acs.est.6b06230 9Quantitative Adverse Outcome Pathways and Their Application to Predictive ToxicologyConolly, Rory B.; Ankley, Gerald T.; Cheng, WanYun; Mayo, Michael L.; Miller, David H.; Perkins, Edward J.; Villeneuve, Daniel L.; Watanabe, Karen H.Environmental Science & Technology (2017), 51 (8), 4661-4672CODEN: ESTHAG; ISSN:0013-936X. (American Chemical Society) A review. A quant. adverse outcome pathway (qAOP) consists of one or more biol. based, computational models describing key event relationships linking a mol. initiating event (MIE) to an adverse outcome. A qAOP provides quant., dose-response and time course predictions that can support regulatory decision-making. Herein the authors describe several facets of qAOPs, including (a) motivation for development, (b) tech. considerations, (c) evaluation of confidence, and (d) potential applications. The qAOP used as an illustrative example for these points describes the linkage between inhibition of cytochrome P 450 19A aromatase (the MIE) and population-level decreases in the fathead minnow (FHM; Pimephales promelas). The qAOP consists of three linked computational models for: (a) the hypothalamic-pituitary-gonadal axis in female FHMs, where aromatase inhibition decreases the conversion of testosterone to 17β estradiol (E2), thereby reducing E2-dependent vitellogenin (VTG; egg yolk protein precursor) synthesis, (b) VTG-dependent egg development and spawning (fecundity), and (c) fecundity-dependent population trajectory. While development of the example qAOP was based on expts. with FHMs exposed to the aromatase inhibitor fadrozole, the authors also show how a toxic equivalence (TEQ) calcn. allows use of the qAOP to predict effects of another, untested aromatase inhibitor, iprodione. While qAOP development can be resource-intensive, the quant. predictions obtained, and TEQ-based application to multiple chems., may be sufficient to justify the cost for some applications in regulatory decision-making. >> More from SciFinder ®https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC2sXltlWhu7Y%253D&md5=a16c89301fbdda3c09055137e9f9321510Jin, Q.; Yang, Y.; Chen, Q.; Lu, Z. GeneGPT: augmenting large language models with domain tools for improved access to biomedical information. Bioinformatics 2024, 40 (2), btae075 DOI: 10.1093/bioinformatics/btae075 There is no corresponding record for this reference.