The challenge of predicting the toxic potential of chemicals has been met in the last few years by the development of sophisticated structure-activity methodologies. However, most of the endpoints studied dealt with in vitro or whole animal effects. The challenge presented by the RFP is to extend the scope of the methodologies to address the organ specific toxicity of chemicals. This is now made possible because the organ specific toxicity of a variety of chemicals has been assessed over the years at the NIEHS. Rationalization of this data could lead to a better understanding of the factors that govern local toxicity, and possibly yield enough information to predict the organ specific toxicity of untested compounds. We will use the Multiple Computer Automated Structure Evaluation (Multi-CASE) method to analyze the data gathered by NIEHS. Structural databases of increasing size and complexity will be analyzed to identify biophores, the chemical substructural features probably responsible for the endpoints tested. These biophores will be used to create predictive models for each of the endpoints that, in turn, will be validated using standard statistical techniques. Phase I will demonstrate the project's feasibility by completing the analysis of two distinct endpoints selected as representative of the problem.
