Abstract
The Structural Classification of Proteins database (SCOP) defines a protein superfamily as a group of evolutionarily related distinct protein structural units or domains. In some cases the superfamily classification spans billions of years of evolutionary time by relating protein domains at the structural level, which has remained conserved even after the domains have diverged beyond the point where sequence similarity is detectable. The PDB contains around 10,000 multi domain protein structures, allowing us to create a Structurally derived Protein superfamily Interaction network or MAP (PSIMAP). PSIMAP represents a condensed view of the whole universe of protein structural interactions at the superfamily level, and thus gives us a robust snapshot of the interactome at this evolutionary level. Here we have applied several general graph theoretical measures to identify key features of this network. Several key features are described, and subsequently analysed in terms of the wealth of structural information available. Generally we uncover hubs, clusters and sub-networks of the whole map, which are composed of superfamilies involved in critical, ancient life processes.