AUTOMATIC CREATION OF DRUG CANDIDATE STRUCTURES BASED ON RECEPTOR STRUCTURE - STARTING POINT FOR ARTIFICIAL LEAD GENERATION

We have developed a new method for automatic generation of drug candidate structures based on a known receptor structure. In our method, various structures which fit well to the receptor cavity are generated, by adding atoms one by one using a force field and random numbers. The usefulness of the program was exemplified by application to the E. coli dihydrofolate reductase system. From dozens of generated structures, we could obtain several promising new structures with considerable internal stability and having favorable interactions with the receptor cavity. It is expected that this method will become an essential starting point for artificial lead generation, which has been impossible so far.