Uncertainty-aware machine learning to predict non-cancer human toxicity for the global chemicals market
Preprint
- 7 March 2025
- preprint
- Published by Springer Nature in Research Square
Abstract
Humans are exposed to many chemicals, yet limited toxicity data hinder effectively managing their impact on human health. High-performing machine learning models hold potential for addressing this gap, but their uncharacterized prediction performance across the wider chemical space undermines confidence in their results. We developed uncertainty-aware models to predict reproductive/developmental and general non-cancer human toxicity points of departure, aimed at supporting chemical risk and impact assessments. Our well-calibrated models provide uncertainty estimates aligned with observed prediction errors and chemical familiarity. We predict toxicity with 95% confidence intervals for >130,000 marketed chemicals and identify toxicity and uncertainty hotspots, including organothiophosphates, PFAS, steroids and dioxin-like compounds. These insights help prioritize efforts to reduce human health risks and improve model performance. By enhancing transparency in prediction uncertainty, our work supports the sound integration of ML-based predictions in chemical assessments and provides key insights for systematically building confidence in digitally derived toxicity information.Keywords
All Related Versions
This publication has 58 references indexed in Scilit:
- Exposure to thiourea during the early stages of development impedes the formation of the swim bladder in zebrafish larvaeJournal of Applied Toxicology, 2024
- Global trends of dioxin and dioxin-like PCBs in animal-origin foods: a systematic review and gap areasEnvironmental Monitoring and Assessment, 2024
- A framework for establishing scientific confidence in new approach methodologiesArchives of Toxicology, 2022
- Hybrid modeling approach for polymer melt index predictionJournal of Applied Polymer Science, 2022
- A framework for chemical safety assessment incorporating new approach methodologies within REACHArchives of Toxicology, 2022
- Exposure and toxicity characterization of chemical emissions and chemicals in products: global recommendations and implementation in USEtoxThe International Journal of Life Cycle Assessment, 2021
- Aleatoric and epistemic uncertainty in machine learning: an introduction to concepts and methodsMachine Learning, 2021
- Dioxin risk assessment: mechanisms of action and possible toxicity in human healthEnvironmental Science and Pollution Research, 2015
- Basic and Clinical Toxicology of Organophosphorus Compounds2014
- Evaluation of developmental toxicity induced by anticholinesterase insecticide, diazinon in female ratsBirth Defects Research Part B: Developmental and Reproductive Toxicology, 2011