Abstract 105: Integrative analysis of genomic and pharmacologic data from the Cancer Cell Line Encyclopedia

The Cancer Cell Line Encyclopedia (CCLE) represents a collaborative effort to assemble a comprehensive resource of human cancer models for basic and translational research. Thus far, the CCLE contains high-density SNP array data, gene expression microarray data and selected cancer gene mutation data for approximately 1,000 human cancer cell lines spanning many tumor types. Additionally, we are assessing the sensitivity of these same cell lines using a series of pharmacological compounds that represent both conventional cytotoxic and targeted agents. Another goal of the CCLE collaboration involves systematic integration of the genomic and pharmacologic datasets in order to identify putative targets of prevalent genetic alterations as well as predictors and modifiers of pharmacologic sensitivity and resistance. The availability of high-quality data generated by uniform criteria across hundreds of cell lines markedly enhances the statistical power to discover genetic alterations involved in carcinogenesis and molecular predictors of pharmacologic vulnerability. As proof of principle, we have carried out systematic nomination of putative targets of genetic alterations using integrative analyses. Here, significant regions of genomic gains and losses have been linked to expression and mutation data to find significant correlations at both single-gene and pathway levels. We have also begun to assemble systematic algorithms that identify genetic predictors of sensitivity or resistance to particular pharmacological compounds, taking advantage of the fact that the CCLE is a comprehensive resource with extensive genomic characterization. Toward this end, we integrated a preliminary sensitivity dataset for 28 compounds accurately profiled against more than 400 cell lines with all genomic data available in the CCLE. To enhance the robustness of our method, we reduced the number of significant genomic features for each cell line to a number that allows properly determined prediction of sensitivity. Expression data was converted to cell line-specific readouts of gene set expression; and DNA gains and losses are reduced to statistically significant regions using the GISTIC algorithm. These values were combined with critical oncogene mutations as inputs to a multifaceted prediction model for pharmacological sensitivity, the accuracy of which was assessed using cross-validation. Our results suggest that this integrative approach applied to a robust cancer cell line collecti