Superfund Research Program

August 2024

Hot Publication

TAMU SRP Center researchers created a new method to identify points of departure (PODs), or the lowest dose of a chemical that triggers a biological response.

Regulatory agencies use PODs to assess health risks from chemical exposures. However, most chemicals lack PODs due to insufficient toxicity data. To address this challenge, the team created a machine learning model to predict the biological activity of chemicals based on their physical and chemical properties.

The model performs in two stages. The first stage identifies the structural, physical, chemical, and toxicological properties of chemicals, from a variety of databases. The second stage uses those properties and data on health effects to predict chemicals’ biological activity.

The researchers tested their model on over 34,000 chemicals and found that it accurately predicted PODs for chemicals that already had them available. The model also identified PODs for chemicals that previously lacked PODs. The results identified several thousand chemicals of moderate concern and several hundred of high concern for health effects.

This new approach significantly increases the number of chemicals that can be evaluated for health risks, according to the authors.

To learn more, please refer to the following sources:

  • Kvasnicka J, Aurisano N, von Borries K, Lu E, Fantke P, Jolliet O, Wright FA, Chiu WA. 2024. Two-stage machine learning-based approach to predict points of departure for human noncancer and developmental/reproductive effects. Environ Sci Technol doi:10.1021/acs.est.4c00172 PMID:38693844