Abstract
Increasing evidence on the massive biodiversity loss caused by chemical pollution highlights the urgent need to understand the internal exposure and resulting risk of chemicals to aquatic species. Current ecological risk assessment frameworks face challenges in addressing the diversity of physiological and biochemical characteristics of aquatic organisms. Here we pioneer the intelligent high-throughput multi-species physiologically based toxicokinetic (HM-PBTK) model by developing AI-driven models for multi-species physiological and biochemical parameters. The HM-PBTK model precisely quantifies tissue-specific internal exposure and predicts in vivo toxicity for neutral and ionizable chemicals across 151 freshwater and marine fish species. In an oestrogenic-effect case study, 85% of toxicity predictions were within fivefold of the corresponding observations. We further developed a user-friendly platform enabling comprehensive ‘exposure–toxicity’ predictions for chemicals by simply setting up exposure scenarios. Our study helps to explore the risks of chemical pollution to aquatic ecosystem, ultimately contributing to the protection of aquatic biodiversity.
This is a preview of subscription content, access via your institution
Access options
Subscribe to this journal
Receive 12 digital issues and online access to articles
$119.00 per year
only $9.92 per issue
Buy this article
- Purchase on SpringerLink
- Instant access to the full article PDF.
USD 39.95
Prices may be subject to local taxes which are calculated during checkout
Similar content being viewed by others
Curated mode-of-action data and effect concentrations for chemicals relevant for the aquatic environment
Progress and challenges in exploring aquatic microbial communities using non-targeted metabolomics
Mathematical modelling and control approach for sustainable ecosystems in mitigating the impact of pollutants on aquatic species in rivers
Data availability
Data supporting the findings of this study are available within the article and its Supplementary Information. Data are also available via figshare at https://doi.org/10.6084/m9.figshare.28521698 (ref. 59). Source data are provided with this paper.
Code availability
The source code for the HM-PBTK model is available via GitHub at https://github.com/Hanpeiling/HM-PBTK.
References
Ruan, T. et al. Identification and prioritization of environmental organic pollutants: from an analytical and toxicological perspective. Chem. Rev. 123, 10584–10640 (2023).
Google Scholar
Chemical Abstracts Service Home Page. Chemical Abstracts Service https://www.cas.org/cas-data/cas-registry (accessed 5 December 2023).
Jaureguiberry, P. et al. The direct drivers of recent global anthropogenic biodiversity loss. Sci. Adv. 8, 9982 (2022).
Google Scholar
EU biodiversity strategy for 2030: bringing nature back into our lives. European Commission https://environment.ec.europa.eu/strategy/biodiversity-strategy-2030_en (accessed 8 October 2024).
Sigmund, G. et al. Addressing chemical pollution in biodiversity research. Global Change Biol. 29, 3240–3255 (2023).
Google Scholar
Sylvester, F. et al. Better integration of chemical pollution research will further our understanding of biodiversity loss. Nat. Ecol. Evol. 7, 1552–1555 (2023).
Google Scholar
Brodin, T. et al. The urgent need for designing greener drugs. Nat. Sustain. 7, 949–951 (2024).
Google Scholar
Kidd, K. A. et al. Collapse of a fish population after exposure to a synthetic estrogen. Proc. Natl Acad. Sci. USA 104, 8897–8901 (2007).
Google Scholar
Duarte, C. M. et al. Rebuilding marine life. Nature 580, 39–51 (2020).
Google Scholar
Barbarossa, V. et al. Threats of global warming to the world’s freshwater fishes. Nat. Commun. 12, 1701 (2021).
Google Scholar
Zoroufchi, B. K. et al. Environmental occurrence and toxicity of 6PPD quinone, an emerging tire rubber-derived chemical: a review. Environ. Sci. Tech. Lett. 10, 815–823 (2023).
Google Scholar
Butchart, S. H. M. et al. Global biodiversity: indicators of recent declines. Science 328, 1164–1168 (2010).
Google Scholar
Khuong, V. D. Vietnam’s fish kill remains unexamined. Science 365, 333 (2019).
Google Scholar
Tanneberger, K. et al. Predicting fish acute toxicity using a fish gill cell line-based toxicity assay. Environ. Sci. Technol. 47, 1110–1119 (2013).
Google Scholar
Sixth Report on the Statistics on the Number of Animals Used for Experimental and Other Scientific Purposes in the Member States of the European Union https://www.europarl.europa.eu/RegData/docs_autres_institutions/commission_europeenne/sec/2010/1107/COM_SEC(2010)1107_EN.pdf (Commission of the European Communities, accessed 8 October 2024).
Coors, A. et al. Minimizing experimental testing on fish for legacy pharmaceuticals. Environ. Sci. Technol. 57, 1721–1730 (2023).
Google Scholar
Tian, Z. et al. A ubiquitous tire rubber–derived chemical induces acute mortality in Coho salmon. Science 371, 185–189 (2021).
Google Scholar
Mangold-Döring, A. et al. A novel multispecies toxicokinetic modeling approach in support of chemical risk assessment. Environ. Sci. Technol. 55, 9109–9118 (2021).
Google Scholar
Fisher, J. W. et al. Fundamentals of physiologically based pharmacokinetic modeling. In Physiologically Based Pharmacokinetic (PBPK) Modeling: Methods and Applications in Toxicology and Risk Assessment (eds Fisher, J. W. et al.) 57–80 (Academic, 2020).
Zhang, S. et al. Tissue-specific accumulation, biotransformation, and physiologically based toxicokinetic modeling of benzotriazole ultraviolet stabilizers in zebrafish (Danio rerio). Environ. Sci. Technol. 55, 11874–11884 (2021).
Google Scholar
Breen, M. et al. High-throughput PBTK models for in vitro to in vivo extrapolation. Expert Opin.Drug Met. 17, 903–921 (2021).
Google Scholar
Brinkmann, M. et al. Understanding receptor-mediated effects in rainbow trout: in vitro–in vivo extrapolation using physiologically based toxicokinetic models. Environ. Sci. Technol. 48, 3303–3309 (2014).
Google Scholar
Krabbenhoft, T. J. et al. Ficli, the fish and climate change database, informs climate adaptation and management for freshwater fishes. Sci. Data 7, 124 (2020).
Google Scholar
Manel, S. et al. Global determinants of freshwater and marine fish genetic diversity. Nat. Commun. 11, 692 (2020).
Google Scholar
Roberson, L. A. et al. Over 90 endangered fish and invertebrates are caught in industrial fisheries. Nat. Commun. 11, 4764 (2020).
Google Scholar
Grech, A. et al. Generic physiologically-based toxicokinetic modelling for fish: integration of environmental factors and species variability. Sci. Total Environ. 651, 516–531 (2019).
Google Scholar
Brinkmann, M. et al. Cross-species extrapolation of uptake and disposition of neutral organic chemicals in fish using a multispecies physiologically-based toxicokinetic model framework. Environ. Sci. Technol. 50, 1914–1923 (2016).
Google Scholar
Wang, J. et al. A generalized physiologically based kinetic model for fish for environmental risk assessment of pharmaceuticals. Environ. Sci. Technol. 56, 6500–6510 (2022).
Google Scholar
Golmohammadi, H. et al. Quantitative structure–activity relationship prediction of blood-to-brain partitioning behavior using support vector machine. Eur. J. Pharm. Sci. 47, 421–429 (2012).
Google Scholar
Sarigiannis, D. Α et al. Development of QSARs for parameterizing physiology based toxicokinetic models. Food Chem. Toxicol. 106, 114–124 (2017).
Google Scholar
Iwata, H. et al. Prediction of total drug clearance in humans using animal data: proposal of a multimodal learning method based on deep learning. J. Pharm. Sci. 110, 1834–1841 (2021).
Google Scholar
Chou, W.-C. & Lin, Z. Machine learning and artificial intelligence in physiologically based pharmacokinetic modeling. Toxicol. Sci. 191, 1–14 (2023).
Google Scholar
Kosugi, Y. & Hosea, N. Direct comparison of total clearance prediction: computational machine learning model versus bottom-up approach using in vitro assay. Mol. Pharm. 17, 2299–2309 (2020).
Google Scholar
Zhang, S. et al. Multimodal model to predict tissue-to-blood partition coefficients of chemicals in mammals and fish. Environ. Sci. Technol. 58, 1944–1953 (2024).
Google Scholar
Black, S. R. et al. Evaluation and comparison of in vitro intrinsic clearance rates measured using cryopreserved hepatocytes from humans, rats, and rainbow trout. Toxicology 457, 152819 (2021).
Google Scholar
MacRae, C. A. & Peterson, R. T. Zebrafish as tools for drug discovery. Nat. Rev. Drug Discov. 14, 721–731 (2015).
Google Scholar
Matthee, C. et al. Factors determining the susceptibility of fish to effects of human pharmaceuticals. Environ. Sci. Technol. 57, 8845–8862 (2023).
Google Scholar
Louisse, J. et al. Use of physiologically based kinetic modeling-based reverse dosimetry to predict in vivo toxicity from in vitro data. Chem. Res. Toxicol. 30, 114–125 (2017).
Google Scholar
Post-2020 global biodiversity framework. IUCN https://www.iucn.org/sites/default/files/2022-07/post-2020-strategic-initiative_policy-brief_march2022.pdf (accessed 25 September 2024).
Silvestro, D. et al. Improving biodiversity protection through artificial intelligence. Nat. Sustain. 5, 415–424 (2022).
Google Scholar
Arnot, J. A. et al. A quantitative structure–activity relationship for predicting metabolic biotransformation rates for organic chemicals in fish. Environ. Toxicol. 28, 1168–1177 (2009).
Google Scholar
Li, J. et al. High-resolution mass spectrometry screening of emerging organophosphate esters (OPEs) in wild fish: occurrence, species-specific difference, and tissue-specific distribution. Environ. Sci. Technol. 56, 302–312 (2022).
Google Scholar
Harrison, C. 3R centers tap into the human mindset to bolster replacement, reduction and refinement uptake. Lab Anim. 53, 166–169 (2024).
Google Scholar
Hack, C. E. et al. Physiologically based pharmacokinetic modeling software. In Physiologically Based Pharmacokinetic (PBPK) Modeling: Methods and Applications in Toxicology and Risk Assessment (eds Fisher, J. W. et al.) 81–126 (Academic, 2020).
Bell, S. et al. An integrated chemical environment with tools for chemical safety testing. Toxicol. In Vitro 67, 104916 (2020).
Google Scholar
Sustainable Development Goals (SDGs). United Nations https://www.un.org/sustainabledevelopment/ (accessed 2 October 2024).
Erickson, R. J. & McKim, J. M. A model for exchange of organic chemicals at fish gills: flow and diffusion limitations. Aquat. Toxicol. 18, 175–197 (1990).
Google Scholar
Hendriks, A. J. et al. The power of size. 1. Rate constants and equilibrium ratios for accumulation of organic substances related to octanol–water partition ratio and species weight. Environ. Toxicol. 20, 1399–1420 (2001).
Google Scholar
Fitzsimmons, P. N. et al. Branchial elimination of superhydrophobic organic compounds by rainbow trout (Oncorhynchus mykiss). Aquat. Toxicol. 55, 23–34 (2001).
Google Scholar
Bertelsen, S. L. et al. Evaluation of log kow and tissue lipid content as predictors of chemical partitioning to fish tissues. Environ. Toxicol. 17, 1447–1455 (1998).
Google Scholar
Mansouri, K. et al. OPERA models for ADME-related properties. National Institute of Environmental Health Sciences https://ntp.niehs.nih.gov/sites/default/files/iccvam/meetings/sot22/poster/mansouri-sot2022.pdf (accessed 10 February 2026).
Briggs, G. G. Theoretical and experimental relationships between soil adsorption, octanol–water partition coefficients, water solubilities, bioconcentration factors, and the parachor. J. Agric. Food Chem. 29, 1050–1059 (1981).
Google Scholar
Hendriks, A. J. et al. Critical body residues linked to octanol−water partitioning, organism composition, and LC50 QSARS: meta-analysis and model. Environ. Sci. Technol. 39, 3226–3236 (2005).
Google Scholar
Huang, Y. et al. Quantitative structure–activity relationship models for predicting inflammatory potential of metal oxide nanoparticles. Environ. Health Perspect. 128, 067010 (2020).
Google Scholar
Liu, W. et al. Machine learning model for screening thyroid stimulating hormone receptor agonists based on updated datasets and improved applicability domain metrics. Chem. Res. Toxicol. 36, 947–958 (2023).
Google Scholar
Jaworska, J. et al. QSAR applicability domain estimation by projection of the training set in descriptor space: a review. Altern. Lab. Anim. 33, 445–459 (2005).
Google Scholar
Guo, J. et al. Deep mining of reported emerging contaminants in China’s surface water in the past decade: exposure, ecological effects and risk assessment. Water Res. 243, 120318 (2023).
Google Scholar
Scholz, S. et al. Alternatives to in vivo tests to detect endocrine disrupting chemicals (EDCs) in fish and amphibians – screening for estrogen, androgen and thyroid hormone disruption. Crit. Rev. Toxicol. 43, 45–72 (2013).
Google Scholar
Han, P. et al. Supplementary data sets for “Advancing aquatic ecological risk assessment through AI-driven prediction of chemical exposure and toxicity in freshwater and marine fish”. figshare https://doi.org/10.6084/m9.figshare.28521698 (2026).
Funding
This study was supported by the National Key Research and Development Program of China (grant nos. 2022YFC3902104 and 2024YFC3712004), the National Natural Science Foundation of China (grant no. U25A20580) and the Programme of Introducing Talents of Discipline to Universities (B25041).
Author information
Authors and Affiliations
Contributions
X.L. designed the research and provided guidance and supervision. P.H. conducted the research and wrote the paper. P.H., Y.Z. and J.Y. designed the platform’s interface and collected the data necessary to support the platform’s operation. J.C. and W.J.G.M.P. participated in the discussions of the study and revised the paper. All authors provided comments on the paper.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing interests.
Peer review
Peer review information
Nature Water thanks Emilio Benfenati, Yongdong Liu and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
Supplementary Information (download PDF )
Supplementary Methods 1–6, Results 1–4, Tables 1–23 and Figs. 1–34.
Peer Review File (download PDF )
Supplementary Data 1 (download XLSX )
The Excel file lists species information from the multimodal dataset, as well as the mean values of blood flow, tissue volume and tissue content calculated for the species when grouped by species family and environmental medium.
Supplementary Data 2 (download XLSX )
The Excel file contains datasets used in this study to construct the prediction models for multi-species Fcard and VO2, as well as the classification and quantification models for human and multi-species CLin vitro,int.
Supplementary Data 3 (download XLSX )
The Excel file contains 703 internal exposure concentration data collected from extensive literature for validation of the HM-PBTK model.
Source data
Source Data Fig. 3 (download XLSX )
Statistical source data.
Source Data Fig. 4 (download XLSX )
Statistical source data.
Source Data Fig. 5 (download XLSX )
Statistical source data.
Source Data Fig. 6 (download XLSX )
Statistical source data.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
Reprints and permissions
About this article
Cite this article
Han, P., Chen, J., Zhu, Y. et al. Advancing aquatic ecological risk assessment through AI-driven prediction of chemical exposure and toxicity in freshwater and marine fish.
Nat Water (2026). https://doi.org/10.1038/s44221-026-00709-7
Received:
Accepted:
Published:
Version of record:
DOI: https://doi.org/10.1038/s44221-026-00709-7
Source: Ecology - nature.com
