Collections of simultaneously altered genes as biomarkers of cancer cell drug response
Menée sur 416 lignées cellulaires, cette étude évalue les performances d'un nouvel algorithme pour identifier un ensemble d'anomalies génétiques susceptibles de prédire la réponse à un agent anticancéreux
Computational analysis of cancer pharmacogenomics data has resulted in biomarkers predictive of drug response, but the majority of response is not captured by current methods. Methods typically select single biomarkers or groups of related biomarkers, but do not account for response that is strictly dependent on many simultaneous genetic alterations. This shortcoming reflects the combinatorics and multiple-testing problem associated with many-body biological interactions. We developed a novel approach, MOCA (Multivariate Organization of Combinatorial Alterations), to partially address these challenges. Extending on previous work that accounts for pairwise interactions, the approach rapidly combines many genomic alterations into biomarkers of drug response, using Boolean set operations coupled with optimization; in this framework the union, intersection, and difference Boolean set operations are proxies of molecular redundancy, synergy, and resistance, respectively. The algorithm is fast, broadly applicable to cancer genomics data, is of immediate utility for prioritizing cancer pharmacogenomics experiments, and recovers known clinical findings without bias. Furthermore, the results presented here connect many important, previously isolated observations. Major Findings When applied to 416 pharmacogenomically characterized cancer cell lines, MOCA identifies many known and potential markers of drug response. For instance, correlation with ERBB inhibitor response drastically increased when considering EGFR (ERBB1), ERBB2, ERBB3, ERBB4, and KRAS alterations in a single feature. Similarly, a feature combining IGF1, IGF1R, and RAD51 drastically increased correlation with IGF1R inhibitor response, relative to any of these three genetic markers considered in isolation. This approach is also powerful for determining subsets of site-specific mutations, for a particular gene, that increase correlation with drug response. For example, MOCA captures the differential EGFR inhibitor response conferred by common EGFR mutations. Similarly, we find specific HDAC1 mutations cooperate with HDAC5 overexpression to potentiate cells to the HDAC inhibitor panobinostat. Additionally, considering all pairwise gene-drug interactions, MOCA recovers known and compelling correlations, including: RTK inhibitor resistance via c-MET, EGFR, ERBB2, and PDGFRB kinase switching; mutual exclusivity of TP53 mutation and response to the MDM2 inhibitor nutlin-3; greater nutilin-3 potentiation via MDM4, rather than MDM2, overexpression; MEK and RAF inhibitor response in BRAF mutated cell lines; and, MEK inhibitor potentiating NRAS mutations.
Cancer Research , résumé, 2013