EUDOC: A computer program for identification of drug interaction sites in macromolecules and drug leads from chemical databases

The completion of the Human Genome Project, the growing effort on proteomic s, and the Structural Genomics Initiative have recently intensified the att ention being paid to reliable computer docking programs able to identify mo lecules that can affect the function of a macromolecule through molecular c omplexation. We report herein an automated computer docking program, EUDOC, for prediction of ligand-receptor complexes from 3D receptor structures, i ncluding metalloproteins, and for identification of a subset enriched in dr ug leads from chemical databases. This program was evaluated from the stand points of force field and sampling issues using 154 experimentally determin ed ligand-receptor complexes and four "real-life" applications of the EUDOC program. The results provide evidence for the reliability and accuracy of the EUDOC program. In addition, key principles underlying molecular recogni tion, and the effects of structural water molecules in the active site and different atomic charge models on docking results are discussed. (C) 2001 J ohn Wiley & Sons, Inc.