Machine learning can address critical data gaps in the characterization of chemical toxicity
How can machine learning help bridge the gaps in chemical data assessments? That is the subject of a newly published article, written by Mistra SafeChem researchers Kerstin von Borries and Peter Fantke, DTU, Hanna Holmquist, IVL, and others.
The article is the result of Kerstin von Borries' PhD project investigating the potential of machine learning (ML) to address critical data gaps associated with the characterization of chemical toxicity effects.
Machine learning is increasingly used to fill data gaps in assessments to quantify impacts associated with chemical releases and chemicals in products. However, the systematic application of ML-based approaches to fill chemical data gaps is still limited, and their potential to address a wide range of chemicals is unknown.
This study prioritized chemical-related parameters for chemical toxicity characterization to inform ML model development based on two criteria:
- Relevance of each parameter to robustly characterize chemical toxicity as described by the uncertainty in characterization results attributable to each parameter.
- The potential of ML-based approaches to predict parameter values for a wide range of chemicals described by the availability of chemicals with measured parameter data.
The results can systematically inform future ML model development efforts to address data gaps in chemical toxicity characterization.
Read the article Potential for Machine Learning to Address Data Gaps in Human Toxicity and Ecotoxicity Characterization External link, opens in new window., published in Environmental Science & Technology.