Typing tumors using pathways selected by somatic evolution

• DOI: 10.1038/s41467-018-06464-y
• Publication Date: 2018-10-02T15:34:17Z
• Authors: Sheng Wang; Jianzhu Ma; Wei Zhang; John Paul Shen; Justin Huang; Jian Peng; Trey Ideker
• Keywords: Epistasis; STK11

Many recent efforts to analyze cancer genomes involve aggregation of mutations within reference maps molecular pathways and protein networks. Here, we find these pathway studies are impeded by interactions that functionally irrelevant or the patient's tumor type, as diminish contrast driver relative individual frequently mutated genes. This problem can be addressed creating stringent tumor-specific networks biophysical interactions, identified signatures epistatic selection during evolution. Using such an evolutionarily selected (ESP) map, major genome atlases derive a hierarchical classification subtypes linked characteristic pathways. These clinically prognostic predictive, including TP53-AXIN-ARHGEF17 combination in liver CYLC2-STK11-STK11IP lung cancer, which validate independent cohorts. ESP framework substantially improves definition from data.