Computer Prediction of Low Molecular Organic Compounds Interaction with Protein-Targets

Pogodin PV1,2, Lagunin AA1, Ivanov SM1, Konova VI1, Filimonov DA1, Poroykov VV1,2
About authors

1 Department of Bioinformation, Laboratory of Structure-Function Based Drug Design,
Orekhovich Institute of Biomedical Chemistry of RAMS, Moscow, Russia

2 Department of Biochemistry, Biomedical Faculty,
Pirogov Russian National Research Medical University, Moscow, Russia

Correspondence should be addressed: Pavel Pogodin
ul. Pogodinskaya, d. 10, Moscow, Russia, 119121; moc.liamg@vpnidogop

Received: 2013-07-01 Accepted: 2013-10-29 Published online: 2017-01-05

The purpose of the study is development of a specialized version of the computer system PASS — PASS Targets for computer-aided prediction of the spectra of interactions between chemical compounds and protein-targets based on ChEMBLdb database (14 version). Data on the interaction between 348 137 different compounds and protein-targets were extracted from ChEMBLdb and were used for PASS training. The average prediction accuracy calculated by leave-one-out cross-validation procedure is 98.1 %. The developed copputer system PASS Targets and be used to predict interaction with 3257 protein-targets representing eight major classes of protein-targets of 215 species.

Keywords: SAR, PASS, spectrum of ligand-protein interactions, computer-aided prediction, in silico pharmacology, protein-targets, ChEMBLdb