Adverse outcome pathway-informed machine learning for predicting vascular toxicity of emerging organic pollutants
Clinical Snapshot
PICO Framework
| P — Population | 312 emerging organic pollutants (EOPs) detected in wastewater |
| I — Intervention | Adverse outcome pathway (AOP)-informed machine learning approach using multilayer perceptron models |
| C — Comparator | ToxCast high-throughput bioassay data and Morgan fingerprints as reference standards |
| O — Outcomes | Vascular toxicity prediction accuracy (70-95%) and chemical prioritisation for risk assessment |
Bottom Line
This computational study presents a novel approach combining adverse outcome pathways with machine learning to predict vascular toxicity of emerging organic pollutants in wastewater. The method achieved high accuracy (70-95%) across multiple toxicity endpoints and successfully identified known vascular toxicants like ketoconazole and sertraline. While the approach shows promise for chemical risk prioritisation, the study lacks detailed validation methodology and external testing. The framework could significantly enhance environmental risk assessment by providing mechanistically-informed predictions for chemicals with limited toxicological data. For Australian practice, this tool could support regulatory decision-making and environmental monitoring programs, particularly for water quality assessment and chemical registration processes. However, regulatory application would require more comprehensive validation and uncertainty quantification before implementation in formal risk assessment frameworks.
Key Findings
P Value: Not reported in abstract
Effect Size: Predictive accuracy ranging from 70-95% across fourteen multilayer perceptron models
Primary Outcome: Vascular toxicity prediction accuracy for 312 emerging organic pollutants
Nnt Or Sensitivity: Model sensitivity and specificity not detailed; accuracy range 70-95% across AOP endpoints
Confidence Interval: Not reported in abstract
Clinical Application
Highly feasible as computational tool requiring no additional laboratory testing; can be integrated into existing risk assessment workflows Relevant for Australian environmental protection agencies, water utilities, and TGA chemical assessment. Could inform National Industrial Chemicals Notification and Assessment Scheme (NICNAS) evaluations and Australian Water Quality Guidelines updates Environmental health professionals, regulatory toxicologists, and public health authorities assessing chemical risks
Abstract
Emerging organic pollutants (EOPs) are increasingly detected in wastewater and pose potential vascular toxicity risks that remain inadequately assessed in current regulatory frameworks. This study developed an adverse outcome pathway (AOP)-informed machine learning approach to evaluate vascular toxicity for 312 EOPs. By integrating ToxCast high-throughput bioassay data with Morgan fingerprints, we trained fourteen multilayer perceptron (MLP) models targeting key events in AOP 509, including Nrf2 inhibition, oxidative stress, mitochondrial dysfunction, apoptosis, endothelial impairment, and angiogenesis disruption. Our optimized models achieved high predictive accuracy (70-95%), enabling activity classification for both tested and untested chemicals. The predicted activation profiles prioritized chemicals such as ketoconazole, sertraline, and miconazole, with literature evidence supporting their vascular toxicity potential. This AOP-guided modelling framework demonstrates how integrating mechanistic pathways with machine learning can inform chemical risk assessment and prioritization, supporting hazard evaluation and environmental decision-making for pollutants with limited toxicological data.
References
- 1.Toukak, A. S., Gao, W., Wan, S., Li, N., & Chen, G. (2026). Adverse outcome pathway-informed machine learning for predicting vascular toxicity of emerging organic pollutants. Environmental Science: Processes & Impacts. https://doi.org/10.1039/d5em00562k
Related Research
Environmental science & technology
Bridging the Gap between External and Internal Exposure: Challenges and a Machine Learning Approach Validated Using Benzene Ring-Containing Pollutants
22 July 2026
PloS one
Air quality index prediction using machine learning regression models: A comparative analysis
10 July 2026
BMJ open
Impact of climatic variables on postoperative outcomes in various surgical disciplines: a systematic review.
5 July 2026
This content is for educational purposes for healthcare professionals only and does not constitute clinical advice. Clinical decisions should be based on individual patient assessment, current guidelines, and appropriate specialist consultation. Editorial Standards · Privacy Policy · Terms of Service