One-class approach: models for virtual screening of non-nucleoside HIV-1 reverse transcriptase inhibitors based on the concept of continuous molecular fields

One-class models for virtual screening of potent non-nucleoside HIV reverse transcriptase inhibitors were built for the first time in terms of the one-class approach using the support vector machine method. The training set included 786 structures of 2-substituted pyrimidinones and their inhibitory activity against the enzyme of wild-type and mutant (K103, IRLL98, Y188L) HIV-1 strains. The representation of molecular structures of organic ligands based on continuous molecular fields can be used to build classification models of higher quality compared to conventional approaches using Carhart fragment-based descriptors, molecular fingerprints, and spectrophores.