The overall objective of SBIR Topic 32 is also the major goal of an NIEHS project from which it emanates, to develop predictive toxicology methods that support decision making and provide insight into pertinent mechanisms of toxic action. Fundamentally different methods are needed to assess the extent to which any one of them applies to the multifactorial prediction space of interest; accordingly, their simultaneous development is essential to achievement of significant progress in this research area. The specific objective of Phase I is to demonstrate the feasibility of reducing the representation of specific-organ toxicity datasets using symbolic artificial intelligence (SAI) methodology, so as to enable their more effective utilization in research aimed at predicting the carcinogenic potential of noncongeneric chemicals, other environmental agents, and mixtures. Realistically, little is known about relationships between specific-organ toxicity and carcinogenicity, or about how to utilize the extensive but complicated NTP database of specific-organ toxicity effects to predict carcinogenicity. Therefore, the Phase I research is viewed as a major challenge by the NIEHS, especially in view of the short, six-month period of the contract. The successful product of Phase I is expected to be a report that describes experiments performed, evaluates results obtained, and demonstrates the applicability of the SAI approach chosen for solving the specific-organ toxicity data analysis problem. Any Phase 2 support for longer-range research to develop a particularly promising approach for predicting carcinogenicity will depend on the extent to which the Phase I objective is accomplished.
