Abstract
It has been commonly accepted that protein-protein interaction network contains valuable information for elucidation of biological processes. Such information, which is difficult to be identified from traditional genetic studies, is important for uncovering the mechanisms of complex diseases such as cancers. In this study, we performed a systematic analysis of global network characteristics for the proteins encoded by cancer genes in the human interactome. We observed that the network topology of the cancer genes was much different from that of non-cancer genes. Overall, compared to non-cancer genes, proteins of cancer genes tend to have higher connectivity, higher betweenness, shorter shortest-path distance, and weaker clustering coefficient in the human interactome. Furthermore, we examined and compared topological features of the recessive and dominant cancer genes. Among the four topological measures we took in this study, recessive cancer genes had significantly different degree distribution or significantly stronger betweenness from dominant cancer genes, while their difference in terms of shortest-path distance and clustering coefficient was not significant. The results suggest that cancer genes’ protein-protein interactions play important roles in the etiology of cancers.