Protein Function Prediction from Protein Interaction Network Using Physico-chemical Properties of Amino Acids

When function of protein cannot be experimentally determined, it can often be inferred from sequence similarity. Analysis of the protein structure can provide functional clues or confirm tentative functional assignments inferred from the sequence. Many structure based approaches exist (e.g. fold similarity, three-dimensional templates), but as no single method can be expected to be successful in all cases, a more prudent approach involves combining multiple methods. In this work, we present a new approach to predict protein function that combines sequential, structural information into protein-protein interaction network. Our PPI network, derivable from protein sequence and structure only, is competitive with other function prediction methods that require additional protein information, such as the size of surface pockets. If we include this extra information about structure and sequence into our protein network, our method yields significantly higher accuracy levels than the others.