Analysis of clustered cancer mutations

Mutations in cancer-associated proteins sometimes cluster within 3D structures, such as substrate-binding pockets of cancer-causing enzymes.

Researchers developed a statistical method called clustering of mutations in protein structures (CLUMPS) to detect novel cancer genes and mutational hotspots in known genes through analysis of clustered protein mutations.

Using CLUMPS, the authors analyzed DNA mutations in 4,742 tumors of 21 types and compared the mutations with structural and protein interaction data for more than 4,000 human proteins in Protein Data Bank, a longstanding repository of protein structures.

Mutation clusters were detected in four known on coproteins, five known tumor suppressor proteins, and a protein called NUF2, required for proper chromosome separation during cell division, but in which mutations had not been previously linked to cancer.

CLUMPS helped confirm tissue and tumor type-specific clustering of mutations in the cancer-associated proteins EGFR and SPOP, respectively. Mutations were also enriched at ion binding sites of cancer-related proteins and at interfaces with their interacting partners, including DNA and RNA.

Because clustered rather than scattered mutations are more likely to reflect selection for functions boosting cell growth, 3D mutation cluster analysis might help narrow the search for the drivers of cancer, according to the authors.

http://www.pnas.org/content/early/2015/09/16/1516373112