Computer Assisted Studies of Structure-Activity Relations Using Pattern Recognition

Structure-activity relationships (SAR) can be investigated for large sets of biologically active compounds using computer-assisted methods. The combination of chemical structure information handling, molecular mechanics model building, substructure searching, and related methods along with pattern recognition and statistical analysis provided an approach to SAR studies. An interactive computer software system, called ADAPT, has been implemented to automate this approach to SAR studies and to allow its application to large sets of organic compounds. A number of SAR studies have been performed using ADAPT, including studies of pharmaceuticals, anti-tumor compounds, olfactory stimulants, and genotoxic compounds (mutagenic or carcinogenic compounds). Results of two studies are discussed in detail: a study of 9-anilino-acridine anti-tumor compounds, and a study of N-nitroso carcinogens.