Extraction of pharmacophore information from high-throughput screens

Two major advances have been made in the computational perception and utili zation of pharmacophores in compound libraries, both real and virtual. Firs tly, a hierarchical set of filtering calculations has emerged that can be u sed to efficiently partition a library into a trial set of pharmacophores. This sequential filtering permits large libraries to be efficiently process ed, as well as compounds judged as 'hits' to be analyzed in great detail. S econdly, new and extended methods of QSAR (quantitative structure-activity relationship) analysis have evolved to translate pharmacophore information into QSAR models that, in turn, can be used as virtual high-throughput scre ens for activity profiling of a library.