Cancer-mutation network and the number and specificity of driver mutations

Significance Cancer genomics yields a wealth of information on cancer-associated mutations in various cancer types, but current understanding of the number and tissue specificity of the driver mutations remains limited. We applied mathematical methods for network analysis to identify distinct modules linking tumors to driver mutations. About 27% of the tumors belong to such modules, whereas the rest form a diffuse component of the gene–tumor network. The cancers from the diffuse component show an onset later in life than those in the modules and have fewer associated known drivers, implying the existence of multiple unidentified and/or interchangeable drivers in the former.