Combined SVM-based and docking-based virtual screening for retrieving novel inhibitors of c-Met

Aberrant c-Met activation has been demonstrated to be implicated in tumorigenesis and anti-cancer drug resistance. Discovery of c-Met inhibitors has attracted much attention in recent years. In this study, a support vector machine (SVM) classification model that discriminates c-Met inhibitors and non-inhibitors was first developed. Evaluation through screening a test set indicates that combined SVM-based and docking-based virtual screening (SB/DB-VS) considerably increases hit rate and enrichment factor compared with the individual methods. Thus the combined SB/DB-VS approach was adopted to screen PubChem, Specs, and Enamine for c-Met inhibitors. 75 compounds were selected for in vitro assays. Eight compounds display a good inhibitory potency against c-Met. Five of them are found to have novel scaffolds, implying a good potential for further chemical modification.

Graphical abstractCombined SVM-based and molecular docking-based virtual screening approach was used to screening several large compounds libraries to retrieve c-Met inhibitors. Some active compounds with novel scaffolds were found.Figure optionsDownload full-size imageDownload as PowerPoint slideHighlights► Combined SVM/docking-based virtual screening was evaluated to be effective. ► This combined approach was used to screen large databases for c-Met inhibitors. ► Eight compounds display a good inhibitory potency against c-Met. ► Five of them are found to have novel scaffolds.