Figure 5

Same biological processes in the interactome network neighborhoods. (a) Left panel, strategy used to examine the interactome network; given a seed or benchmark protein encoded by a recognized low-penetrance susceptibility gene and using a shortest path algorithm, we calculated at each step the percentage of nodes annotated with Cell Communication or Cell Adhesion among proteins annotated with any term (excluding non-annotated proteins). Right panel, distribution of all possible short paths in the giant network component. (b) Left panels, results for percentages in short paths of up to seven steps for benchmark proteins. Over-representation in Cell Communication and Cell Adhesion annotations is suggested for CDH1, FGFR2, HMMR and RASSF1 at direct and/or one-hop interactions. Right panel, asymmetrical distribution of CDH1, FGFR2, HMMR and RASSF1 direct and one-hop interactors in the complete WGAS rank. (c) Over-representation of processes in the one-hop neighborhood of CDH1 or FGFR2 (vertical lines) using as controls seed proteins with the same Level 3 annotations (curves). The x-axis represents the cumulative percentage up to 200. The CDH1 and FGFR2 percentiles are shown.