Federal Project Grant Award Summary

Duke University's Office of Research Administration received a $444,125 Project Grant from the National Institute of Environmental Health Sciences (NIEHS) under the Environmental Health program (CFDA 93.113), awarded February 17, 2026, with completion targeted for February 16, 2028. The project delivers an integrated computational and experimental workflow to identify endocrine-disrupting plastic additives through machine learning and laboratory validation. Under Aim 1, the awardee develops novel machine learning models trained on publicly available datasets to predict androgen receptor (AR) and estrogen receptor alpha (ERA) modulating activity across thousands of uncharacterized plastic additives, enabling prioritization of candidates for further testing based on predictive uncertainty and novelty. Under Aim 2, Duke University conducts multi-step experimental validation including in-house AR and ERA assays, dose-response studies in responsive cell lines to measure target gene activation and cell proliferation, and in vivo mouse studies to assess systemic hormonal changes for the top three compounds identified in silico.

The project generates deliverables with significant public health application, including a first-in-kind machine learning-assisted predictive toxicological model identifying plastic additives of highest concern for adverse health effects, coupled with a comprehensive dataset characterizing plastic additive effects on endocrine function. These outputs are designed to inform regulatory policy guidance on plastic additives requiring removal or restriction from consumer products, with potential beneficial health outcomes for billions of consumers.