QSAR study on melanocortin-4 receptors by support vector machine

Abstract

Quantitative structure activity relationship (QSAR) of the melanocortin-4 receptor (MC4R) binding affinities (K(i)) of trans-4-(4-chlorophenyl) pyrrolidine-3-carboxamides of piperazinecyclohexanes was studied. A suitable set of molecular descriptors was calculated and the genetic algorithm (GA) was employed to select those descriptors that resulted in the best-fit models. The multiple linear regression (MLR), and the support vector machine (SVM) were utilized to construct the linear and nonlinear QSAR models. The models were validated using Leave-One-Out (LOO) and Leave-Group-Out (LGO) cross-validation, external test set, and chance correlation. The SVM model generalizes better than the MLR model. The SVM model, with high statistical significance (R(2)(train)=0.908, Q(2)(LOO)=0.781, Q(2)(LGO)=0.872), could be used to predict melanocortin-4 receptor binding affinities of piperazinecyclohexanes.