Towards an integrated understanding of the structural characteristics of protein residue networks.

A protein residue network or PRN is a network induced by spatial contacts between amino acid residues of a protein. Studies of the structure of PRNs have revealed a list of network characteristics common to a diverse class of proteins. Explanations for the observed network characteristics for protein folding have been suggested but not tested in an integrated way. In this article, in silico experiments are performed to understand how structural characteristics of PRNs influence protein folding as modeled by a search problem. We find that the blend of structural characteristics PRNs possess help to place them in a sweet spot within the space of all network configurations tested. PRNs are plausible 3D structures and yield competitive search performances. Hence, it appears that PRNs are in a form suited to the function they evolved into. However, this conclusion is partially contingent upon the fitness function preferentially satisfying short-range links but also allowing short- and long-range interactions to cooperate towards the satisfaction of all links. We close with a discussion on the rather intricate interplay among the three main structural characteristics of PRNs, i.e., clustering, average path length, and assortativity, and their impact on search performance and 3D structure plausibility.