Modular organization of gene-tumor association network allows identification of key molecular players in cancer

The role played by the topological structure of biological networks in their dynamics and function is receiving increasing attention over the last decade as large-throughput experiments have provided large volumes of highly resolved data on the interactions between the components of such networks. This has provided new perspectives on systems diseases: for example, there has been a gradual shift in cancer research away from the study of individual molecules and of single gene mutations to the emerging consensus that it is a complex disease involving large-scale disruptions in the intracellular signaling network. One of the drawbacks of a systems- or network-based approach is the large number of cellular agents whose interactions need to be investigated. We tried to solve this problem by taking a mesoscopic view of the cancer diseases-genes network, whose modular organization we studied after projecting it onto two networks, one comprising only disease types and the other consisting of only genes related to one or more categories of cancer. Using community partitioning, we identified several modules in these networks. Projecting cancer gene clusters onto an abstract 'modular space' allows us to infer the relations between different tumor types. By classifying the functional role of particular genes in terms of their inter- and intra-modular connectivity, we identified a number of genes that play the key role of 'connector hubs' in the network. Using data from the human protein-protein interaction network we showed that genes that are 'connector hubs' or 'global hubs' are, in fact, much more likely to be related to cancer than other genes. More important from a therapeutic point of view, we showed that the connector hubs in the cancer gene network are involved in a significantly larger number of human signaling pathways associated with cancer than other types of cancer genes. Furthermore, the types of cancer linked to connector hub genes have significantly reduced survival rates compared with other types of cancer, thereby enhancing their importance in the search for potential therapeutic targets.

Automated summary

The topological structure of biological networks plays a crucial role in understanding diseases such as cancer. By studying the modular organization and connectivity of cancer disease-gene networks, we can infer relationships between different tumor types and identify potential therapeutic targets.