ToxCastTM : Developing predictive signatures for chemical toxicity

ToxCast, the United States Environmental Protection Agency's chemical prioritization research program, is developing methods for utilizing computational chemistry, bioactivity profiling and toxicogenomic data to predict potential for toxicity and prioritize limited testing resources. The ToxCast research program is designed to forecast toxicity based on bioactivity profiling, and for proof-of-concept is focusing on environmental chemicals with extensive toxicity data to provide an interpretive context for these bioactivity profiles. This set of 320 reference chemicals are largely derived from the active ingredients in food use pesticides and represent numerous structural classes and phenotypic outcomes, including tumorigens, developmental and reproductive toxicants, neurotoxicants and immunotoxicants. Toxicity data from rodent in vivo studies are being captured into a relational database called ToxRefDB, allowing chemical groupings based on toxicities specific to study type, target organ, or effect categories. ToxCast will include a broad spectrum of data on environmental chemicals: physical-chemical properties, computationally-predicted biological activities, biochemical and cell-based phenotypic characterization based on HTS assays, highcontent screening assays, and in vitro gene expression. As the ToxCast proof-of-concept advances, it will expand to a broader range of data types that can be used to develop predictive models for multiple toxicities, useful for the prioritization of chemicals for further toxicological evaluation. This manuscript provides an overview of the ToxCast research program, including description of the supporting informatic infrastructure that will house supporting reference databases.