Virtual Screening of Molecular Databases Using a Support Vector Machine.

Abstract

The Support Vector Machine (SVM) is an algorithm that derives a model used for the classification of data\ninto two categories and which has good generalization properties. This study applies the SVM algorithm to\nthe problem of virtual screening for molecules with a desired activity. In contrast to typical applications of\nthe SVM, we emphasize not classification but enrichment of actives by using a modified version of the\nstandard SVM function to rank molecules. The method employs a simple and novel criterion for picking\nmolecular descriptors and uses cross-validation to select SVM parameters. The resulting method is more\neffective at enriching for active compounds with novel chemistries than binary fingerprint-based methods\nsuch as binary kernel discrimination.