Two QSAR approaches have been compared in detail. The DARC/PELCO method is based on the exhaustive generation of all topochromatic sites around the reference structure and the evaluation of their contribution to the property. The OASIS system is an optimized version of the extended Hansch type physicochemical method which makes use of a large set of molecular descriptors (topological, steric, quantum chemical, etc.). The DARC/PELCO models of a series of purine derivatives have a more specific area of reliable prediction and greater accuracy than OASIS models. The latter, in turn, provide more details concerning the possible biological mechanism. In spite of their differences, the two methods produce similar predictions for the properties under study. A conclusion is drawn about the complementarity of these QSAR approaches for optimizing the lead structure in congeneric series of biologically active molecules.