The report discusses some important general strategies and issues relative to the application of computational SAR techniques for modeling genotoxicity and carcinogenicity endpoints. Problems particular to the SAR modeling of such endpoints pertain to: the complexity of the carcinogenicity endpoint; the unclear relationship of the multitude of short-term bioassay endpoints to each other and to the carcinogenicity endpoint; and the limitations of available data bases which are most often underrepresented with respect to common endpoint, testing protocol and chemical class. Interrelationships between various elements in an SAR model study, the data base requirements and limitations of such studies, and the composition of data bases available to SAR modelers are considered. Methods for representing, organizing and evaluating such data to maximize its utility for SAR investigation are discussed in terms of biologically-based comparative assessment approaches, i.e. genetic activity profiles, and weight-of-evidence evaluation schemes. A sampling of SAR programs which have been applied to SAR analysis of carcinogenicity and genotocitity endpoints are briefly discussed and contrasted. The CASE fragment-based SAR program and profile matching techniques are considered in more detail in the context of an illustrative example involving analysis of a very general chemical class - the organic halides.