Identification of small molecule binding sites using CmDock

Abstract

Identification of binding sites for small molecules is critical in light of modern in silico and experimental methods for structure elucidation such as AlfaFold and CryoEM, respectively. The total number of entries in the PDB database is 208,702 through July 2023, and there are a large number of structural data without similar experimental entries and mechanistic or small molecule data. Herein, we report a simple, rapid, and efficient protocol for the identification of drug-like binding sites of small molecules using extended sampling with the self-developed molecular docking software CmDock. The protocol consists of preparing a docking receptor using the RbtProteinMapper method, which uses a full protein surface as a reference. Then, a series of drug-like interaction sampling probes are docked in an extended sampling of 1000 or more runs. The binding conformations calculated by the probes are analyzed using PacMAP reduction and DBSCAN density clustering to identify binding sites. The protocol is capable of identifying known binding sites of small molecules in a very short time frame. In addition, we demonstrate the application of the method to a known DNA Gyrase as well as to a STAT3 NTD domain system.