Safety Data Mining Based on Chemical Structures

DESCRIPTION (provided by applicant): Predicting and understanding drug toxicity in humans remains a major obstacle to new drug development programs. Pharmacovigilance databases of spontaneous adverse event reports from marketed drugs represent an important source of human toxicity information for addressing this challenge, especially if the associations identified in these databases between various adverse events (types of toxicity) and specific drugs can be understood in terms of the relationship of the adverse events to the underlying chemical structure of the drugs. In Phase I, the focus is on the integration of structural chemistry information into the applicant's existing "'WEBVDME" safety data mining system to support initial investigation of relationships between drug toxicity Profiles (as expressed in spontaneous adverse event data) and underlying chemical structures and substructures (2-D fragments). Phase II will also investigate the feasibility and value of incorporating other types of information (3-D molecular descriptors, pharmacokinetic and metabolic parameters) for further elucidation of patterns and issues in drug toxicity. The key benefit of the project to public health is an improved ability to predict drug toxicity, leading to significant reductions in the financial and human costs of drug development and to improved safety profiles for the drugs that reach the market.