Assessing chemical toxicity across Eukaryota using multimodal transformers
- 1Chalmers University of Technology
Abstract
Biodiversity is globally threatened by chemical pollution, yet toxicity data remain unavailable for millions of species and tens of thousands of chemicals, severely limiting our ability to assess ecological impacts. Here we present TRIDENT-2, a multimodal artificial intelligence model for predicting chemical toxicity across evolutionarily diverse eukaryotic species. Trained on 560,780 toxicity assays spanning 82,775 chemicals, 6,793 species, and multiple exposure scenarios, TRIDENT-2 accurately predicts toxicity across Eukaryota with an average median absolute error ranging from 1.76 to 3.80. By jointly learning from chemical, biological, and experimental information, it remains accurate across broad chemical and taxonomic distances, allowing for toxicity assessment for species and chemicals beyond the current experimental evidence. Our findings demonstrate that artificial intelligence can help overcome longstanding data limitations in ecotoxicology, paving the way for improved decision-making and reducing chemical impacts on biodiversity and ecosystems.
Related
- Assessing Low-Mineralized Surface Water via Danio rerio Embryo Toxicity Test: A Case Study of Glacial Lakes in the Northern Tien Shan.
- Assessing financial toxicity with COST-FACIT: experience from a tertiary care cancer center.
- Precision Cardio-Oncology: A Multimodal Approach to Cancer Therapy-Related Cardiovascular Toxicity Prediction and Assessment
- A multimodal system for assessing iatrogenic breast deformities in women (Breast Iatrogenic Deformity System – BIDS)
- Spatially Context-Aware Transformers Facilitate Modeling-Based Anomaly Detection of Subtle Lesions in Brain MRI Images
BibTeX
@misc{kall2026assessing,
title = {{Assessing chemical toxicity across Eukaryota using multimodal transformers}},
author = {S. Käll and M. Gustavsson and P. Svedberg and M. Dalman and J. Henriksson and T. Backhaus and E. Kristiansson},
howpublished = {bioRxiv},
year = {2026},
month = sep,
doi = {10.64898/2026.09.23.753795},
url = {https://www.biorxiv.org/content/10.64898/2026.09.23.753795v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - Assessing chemical toxicity across Eukaryota using multimodal transformers
AU - S. Käll
AU - M. Gustavsson
AU - P. Svedberg
AU - M. Dalman
AU - J. Henriksson
AU - T. Backhaus
AU - E. Kristiansson
PY - 2026
DA - 2026/09/23
PB - bioRxiv
DO - 10.64898/2026.09.23.753795
UR - https://www.biorxiv.org/content/10.64898/2026.09.23.753795v1
AB - Biodiversity is globally threatened by chemical pollution, yet toxicity data remain unavailable for millions of species and tens of thousands of chemicals, severely limiting our ability to assess ecological impacts. Here we present TRIDENT-2, a multimodal artificial intelligence model for predicting chemical toxicity across evolutionarily diverse eukaryotic species. Trained on 560,780 toxicity assays spanning 82,775 chemicals, 6,793 species, and multiple exposure scenarios, TRIDENT-2 accurately predicts toxicity across Eukaryota with an average median absolute error ranging from 1.76 to 3.80. By jointly learning from chemical, biological, and experimental information, it remains accurate across broad chemical and taxonomic distances, allowing for toxicity assessment for species and chemicals beyond the current experimental evidence. Our findings demonstrate that artificial intelligence can help overcome longstanding data limitations in ecotoxicology, paving the way for improved decision-making and reducing chemical impacts on biodiversity and ecosystems.
ER -