Human immunodeficiency virus infection and acquired immune deficiency syndrome (HIV/AIDS) has been one of the deadliest diseases for the past three decades. Treatment is challenging even today.

Developing a computational method to predict drug resistance will enable the best and most efficient treatment. Several researchers have recently developed algorithms to predict HIV drug resistance, for example, applying machine learning algorithms such as support vector machines (SVMs), random forests, and statistical methods.

In this paper, the authors propose a method for predicting drug resistance by combining sequence and structural features. They applied SVMs and random forests as classifiers, and conducted experiments using the Stanford HIV database. The imbalanced dataset was balanced using a synthetic minority oversampling technique filter. Features were extracted from openly available web servers, FlexPred and PSIPRED, and estimated using a freely available tool (WESA). Experimental results are presented in terms of metrics such as area under the curve, sensitivity, and specificity. The results are well analyzed and compared with other existing methods.

This well-written paper was developed using freely available resources. It will be useful for researchers working in bioinformatics and machine learning.