Support Vector Machine-Based Quantitative Structure-Activity Relationship Study of Cholesteryl Ester Transfer Protein Inhibitors

To explore inhibition of cholesteryl ester transfer protein, a support vector machine in quantitative structure-activity relationship was developed for modeling cytotoxicity data for a series of cholesteryl ester transfer protein inhibitors. A large number of descriptors were calculated and genetic algorithm was used to select variables that resulted in the best-fitted models. The data set was randomly divided into 68 molecules of training and 17 molecules of test set. The selected molecular descriptors were used as inputs for support vector machine. The obtained results using support vector machine were compared with those of multiple linear regression which revealed superiority of the support vector machine model over the multiple linear regression. The root mean square errors of the training set and the test set for support vector machine model were calculated to be 3.707, 5.273 and the correlation coefficients (r (2)) were obtained to be 0.947, 0.899, respectively. The obtained statistical parameter of leave-one-out cross-validation test correlation coefficients (q (2)) on support vector machine model was 0.852, which indicates the reliability of the proposed model.