This project applies a variety of computer-based induction methods to the problem of estimating the biological activity of noncongeneric (structurally dissimilar) chemicals. The long-term goal is to provide chemical and health-care product manufacturers, toxicologists, regulators, managers, and others with a computer-based system that can

1. predict or assist in predicting the toxic effects of chemical compounds;
2. discover relationships that serve as the basis for new hypotheses for mechanistic research; and
3. provide information that supports decision making about appropriate and effective utilization of resources, including a reduction in the reliance on animals in health research.