Inferring Strengths of Protein-Protein Interactions Using Support Vector Regression

Protein-protein interactions (PPIs) play various important roles in living organisms. Hence, many efforts have been made to investigate and predict PPIs. Analysis of strengths of PPIs is important as well as PPIs because such strengths are involved in functionality of proteins. In this paper, we propose several feature space mappings from protein pairs, which make use of protein domain information, and perform five-fold cross-validation for data obtained from biological experiments. The result of average root mean square error (RMSE) using support vector regression (SVR) with our proposed feature was better than that by the best existing method, APM proposed by Chen et al.