Prediction of Human Immunodeficiency Virus Drug Resistance Using Contact Energies

The HIV-1 protease drug susceptibility data sets from the Stanford HIV-1 drug resistance database were utilized to determine drug susceptibility to seven protease inhibitors using viral genotype. Using the drug-specific resistance-fold values associated with each sample, the dataset of phenotypes were grouped into two classes. The contact energies where used to represent the protease sequence of HIV. Two methods were use to predict de drug resistance: multi layer perceptron (MLP) and support vector machine (SMV). SVMs were use with different types of kernel function. Both MLP and SVM were compared with previously published methods to find a relationship between phenotype and classification models. We found prediction percent between genotype. Numerous authors have worked in order to solve 80-92.3 for MLP and prediction percent between 75.2-91.8 SVM