Method and apparatus for computer automated detection of protein and nucleic acid targets of a chemical compound

Abstract

A method and apparatus are disclosed for automated detection of possible protein and nucleic acid targets of a given drug. The advantage of this invention over existing methods is its unique capability of drug target detection. In contrast, existing computer methods are designed for screening multiple chemical compounds to find one or more compounds that can bind to a protein or nucleic acid, and they are not capable of finding drug targets. Potential applications of this method and apparatus include unknown target or secondary target identification for drugs, lead compounds, and natural products. It may also potentially facilitate the prediction of drug side-effect and toxicity based on the analysis of function of identified protein or nucleic acid targets. A ligand-biomolecule inverse-docking algorithm is disclosed to flexibly dock a ligand to multiple entries in a biomolecular cavity database. Docking is accomplished by matching atoms of a ligand in single or multiple conformations to spheres in a sphere cluster representing a cavity in the biomolecule by means of a disclosed vector-vector matching algorithm. The docked structures are subject to conformation optimization for both the ligand and the side-chains of protein or nucleic acid residues around the ligand. A method is also disclosed for computer automated generation of a biomolecular cavity database using entries from protein and nucleic acid 3D structure database. The number of proteins and nucleic acids in this database is comparable to that in Brookhaven Protein Databank (the most popular public domain protein and nucleic acid 3D structure database).