HSPPIP: An Online Tool for Prediction of Protein-Protein Interactions in Humans

Recently, protein-protein interaction prediction (PPIP) has been emerging as an appealing question. Although several in silico approaches have been developed to delineate the potential protein-protein interaction (PPI), there are few online tools of human PPIP for further experimental design. Here we present an online service, hsPPIP (Protein-Protein Interaction Predicting of Homo Sapiens), to predict or evaluate the potential PPIs in human. The annotations of functional domain (Interpro) and GO (Gene Ontology) for proteins are employed as two informative features, and are integrated by the naïve Bayesian approach. The prediction accuracy is comparable to the existing tools. Based on the hypothesis that the features correlated with PPIs are conserved in different organisms, the web server hsPPIP is established and could predict the PPIs of human dynamically. hsPPIP is implemented in PHP+MySQL and can be freely accessed at: http://973-proteinweb.ustc.edu.cn/ hsppip/.