Support Vector Machine Applied to Predicting the Activity of HIV-1 IN Inhibitors

HIV integrase(IN) catalyzes the integration process in the viral life cycle.IN helps the viral reverse transcription product cDNA integrating into the host chromosome.As an indispensable enzyme,IN is also an important target for designing and developing the novel anti-HIV drugs.Constructing the QSAR model of the HIV-1 IN pyrimidones inhibitors can help for understanding of the structural factors.In this paper,the CoMFA software was used to calculate topological descriptors,polarizable descriptors and hydrophilic descriptors and other structural parameters for 68 compounds.The structural parameters were selected as inputs of support vector machine(SVM) to establish the non-linear regression model.Resultsshow that the SVM algorithm combined with QSAR can establish the forecasting model for inhibitors,and provide biological information for the design of anti-HIV drugs.