Sequence-based prediction of protein-protein interactions using pseudo substitution matrix representation features and ensemble rotation forest classifier in HIV (human immunodeficiency virus)

Abstract

Human immunodeficiency virus-1 (HIV-1) in acquired immune deficiency syndrome (AIDS) relies on human host cell proteins in virtually every aspect of its life cycle. Knowledge of the set of interacting human and viral proteins would greatly contribute to our understanding of the mechanisms of infection and then to design of new therapeutic approaches. Predicting Protein-Protein Interaction (PPI) is important for making discoveries in the molecular mechanisms within a cell. Sequence-based prediction is the most readily applicable and effective cost method to predict protein-protein interactions. By using computation processes and applying machine learning methods, it is more efficient than conventional method which takes a long time and expensive cost. In this study, we used pseudo-substitution matrix representation as a feature extraction method, after that we used rotation forest ensemble classifier to predict protein-protein interactions class between humans proteins and HIV proteins. The experiment results show that proposed method is more feasible, powerful and can be improved to predict other protein-protein interactions.