How to compute protein residue contacts more accurately?

Computing contacts in proteins is important to several types of studies from Bioinformatics to Structural Biology. An accurate computation of contacts is essential to the correctness and reliability of applications involving folding prediction, protein structure prediction, quality assessment of protein structures, network contacts analysis, thermodynamic stability prediction, protein-protein and protein-ligand interactions, docking and so forth. In this work, we built an extensive database of contacts using about 45,000 structures from PDB to compare three paradigms for contact prospection at the atomic level: distance-based only, distance-geometry-based and distance-angulation-based. The main contribution of this paper is a critical evaluation of three different paradigms that can be used to compute contacts between protein atoms. We focused on protein-protein interfaces and analyzed four types of contacts, namely hydrogen bonds, aromatic stackings, hydrophobic and ionic (attractive) interactions. We scanned for possible contacts in the range from 0 to 7 Å. Our database with all computed contacts as well as the source code used to populate this database is freely available at bioinfo.dcc.ufmg.br/capri Our data showed the importance of a geometric approach to filter out spurious occluded contacts after about 3.5 Å for aromatic stackings, hydrophobic and ionic interactions. For hydrogen bonds, to filter out spurious contacts, we need to consider the angles involved in the interactions.