We aim to develop statistical methods for assessing health effects of chemical exposures, focusing on mixtures, interactions, and leveraging similar molecular structure data.
The objectives are to develop new statistical methods for improving our ability to assess human health effects of chemical exposures in the environment. A focus is on mixtures of exposures to different chemical contaminants, and improving statistical models that can detect and accurately characterize not only main effects but also interactions. To this end approaches that isolate synergistic vs antagonistic vs null interactions are developed, including in the longitudinal context in which data on exposures and outcomes are collected over time. In addition, the project develops improved approaches for borrowing of information across chemicals that have similar molecular structure, leveraging on data from ToxCast/Tox21.