New structural concepts for predicting carcinogenicity in rodents: An artificial intelligence approach

Abstract
The Computer Automated Structure Evaluation (CASE) method for studying structure‐activity relationships has been applied to a data base of rodent carcinogens. It has been demonstrated that CASE is able to identify determinants embedded in the molecular structure which, with a high probability, predict rodent carcinogenicity. CASE has also identified determinants associated with the activity of non‐genotoxic carcinogens, thereby suggesting that there is a structural commonality in the activity of these molecules. The present study reveals that there are “universal” as well as species‐specific structural determinants of carcinogenicity. CASE was able to predict the carcinogenicity in rodents of certain endogenous pesticides in edible plants.