A genome-wide approach to link genotype to clinical outcome by simultaneously utilizing next generation sequencing and gene chip data of 6,697 breast cancer patients.

Abstract

e12074 Background: Cancer somatic mutations may results in an altered protein product impinging on a set of additional genes capable to “leverage” the effects of the mutation. To evaluate this inception, one has to capture the signature of all genes displaying altered expression in relation to a given mutation and asses the prognostic potential for this multigene signature. We combined available genotype data generated by next generation sequencing (NGS) with microarray based gene expression data to establish a framework to assess this leveraged effect on clinical outcome. Methods: NGS data generated by the TCGA and gene chip data obtained from GEO were utilized. Transcriptomic fingerprint for mutation status is carried out by running ROC analysis on mutation and RNA-seq data. The average expression (gene chip data) of the significant genes is designated as a genotype‘s metagene. Correlation to survival is assessed by computing Cox regression for both up- and down-regulated signatures. An online interface enables running the analysis for any selected gene (http://www.g-2-o.com). Results: The database contains 763 NGS samples containing mutational status for 22,938 genes and RNA-seq data for 10,987 genes. The gene chip database contains 5,934 patients with 10,987 genes plus clinical characteristics. The system was validated to reassure correlation between RNA-seq and microarray data (r2= 0.73, p < 1E-16). A set of established genes known to influence survival were used as test set (TP53: p < 1E-16 and PIK3CA: p = 8.1E-07). Conclusions: By connecting genotype to gene expression signature and employing this signature for survival analysis we established a pipeline enabling the rapid functional validation of a discovered mutation for any gene in a large breast cancer cohort.