Prediction of interacting sites in protein complexes

Most of the cellular functions are the result of the concerted action of complexes of proteins forming pathways and networks. For this reason, much effort has been devoted to the study of protein-protein interactions. Large-scale experiments on whole genomes allowed the identification of interacting protein pairs but the residues involved in the interaction are generally not known and the majority of the interactions still lack a structural characterization. A crucial step towards the deciphering of the interaction mechanism of proteins is the recognition of their interacting surfaces, particularly in those structures for which also the most recent interaction network resources do not contain information. Furthermore, the prediction of interaction sites in protein structures is extremely valuable in the view of targeting specific disease-related interactions. To this purpose, we developed a neural network-based method that is able to characterize protein complexes, through the prediction of the amino acid residues that mediate the interactions, starting from the protein structure. All the Protein Data Bank (PDB) chains, both in the unbound and in the complexed form, are predicted and the results are stored in a database of interaction surfaces (http://gpcr.biocomp.unibo.it/zenpatches). We updated the database both in order to cover the whole PDB and to further validate the method by casting the predictions in the networks of known and experimentally verified protein interactions, such as those available in the STRING database. This data integration can also help to investigate whether the predictions classified as false positives by computational methods are really non-interacting sites or if they give rise to transient interactions. Finally, we performed a survey on the different computational methods for proteinprotein interaction prediction and on their training/testing sets in order to highlight the most informative properties of protein interfaces and to detect the specificity and the sensitivity for the different types of complexes (homo and hetero-complexes, obligate and non-obligate ones). This comparative analysis allows to discuss the adopted definitions of protein, surface and interface. The information enclosed in the method supports the understanding of protein interaction networks and can help both in the prediction of novel protein interactions and in the design of complexes with specific functions.