Description and assessment of a model for GSK-3β database virtual screening

Abstract

Molecular docking was used in order to prioritize organic syntheses or experimental evaluations. Different GSK-3β protein models were generated in silico from a known X-ray structure. A set of 42 known inhibitors were then flexibly docked into each rigid model and re-scored with various functions, which led to different rankings. The biological activities of the chemicals were then compared to each set of results and one of the rigid models emerged in combination with two scoring functions as giving the best correlation. This methodology constitutes an easy and accurate way to generate reliable models for virtual database screening. © 2010 Informa UK Ltd.