SVM model for Predicting Human Proteins Interacting with HCV Proteins
Chao Fang, Guangyu Cui, Kyungsook Han · 2011
Several computational methods have been developed for predicting protein-protein interactions, but most of these methods are intended for finding the protein-protein interactions within a species rather than for the interactions across different species. Methods for predicting the interactions between homogeneous proteins are not appropriate for predicting the interactions between heterogeneous proteins since they do not distinguish the interactions between proteins of the same species from those of different species. In this paper we present the development of a support vector machine (SVM) model that predicts the interactions between hepatitis C virus (HCV) proteins and human proteins using the sequence data. The average accuracy of the SVM model in predicting the interactions between HCV proteins and human proteins is 81.5%. Using the SVM model and the Gene Ontology (GO) annotations of proteins, we also predicted a total of 456 new interactions between HCV and human proteins.