Discovery Analysis of TCGA Data Reveals Association between Germline Genotype and Survival in Ovarian Cancer Patients Rosemary Braun 1,2 *, Richard Finney 2 , Chunhua Yan 2 , Qing-Rong Chen 2 , Ying Hu 2 , Michael Edmonson 2 , Daoud Meerzaman 2 , Kenneth Buetow 3,2 1 Robert H. Lurie Comprehensive Cancer Center, Northwestern University, Chicago, Illinois, United States of America, 2 Center for Biomedical Informatics and Information Technology, National Cancer Institute, National Institutes of Health, Bethesda, Maryland, United States of America, 3 Computational Science and Informatics Program, Complex Adaptive Systems Initiative, Arizona State University, Phoenix, Arizona, United States of America Abstract Background: Ovarian cancer remains a significant public health burden, with the highest mortality rate of all the gynecological cancers. This is attributable to the late stage at which the majority of ovarian cancers are diagnosed, coupled with the low and variable response of advanced tumors to standard chemotherapies. To date, clinically useful predictors of treatment response remain lacking. Identifying the genetic determinants of ovarian cancer survival and treatment response is crucial to the development of prognostic biomarkers and personalized therapies that may improve outcomes for the late- stage patients who comprise the majority of cases. Methods: To identify constitutional genetic variations contributing to ovarian cancer mortality, we systematically investigated associations between germline polymorphisms and ovarian cancer survival using data from The Cancer Genome Atlas Project (TCGA). Using stage-stratified Cox proportional hazards regression, we examined w650,000 SNP loci for association with survival. We additionally examined whether the association of significant SNPs with survival was modified by somatic alterations. Results: Germline polymorphisms at rs4934282 (AGAP11/C10orf116) and rs1857623 (DNAH14) were associated with stage- adjusted survival (p = 1.12e-07 and 1.80e-07, FDR q = 1.2e-04 and 2.4e-04, respectively). A third SNP, rs4869 (C10orf116), was additionally identified as significant in the exome sequencing data; it is in near-perfect LD with rs4934282. The associations with survival remained significant when somatic alterations. Conclusions: Discovery analysis of TCGA data reveals germline genetic variations that may play a role in ovarian cancer survival even among late-stage cases. The significant loci are located near genes previously reported as having a possible relationship to platinum and taxol response. Because the variant alleles at the significant loci are common (frequencies for rs4934282 A/C alleles = 0.54/0.46, respectively; rs1857623 A/G alleles = 0.55/0.45, respectively) and germline variants can be assayed noninvasively, our findings provide potential targets for further exploration as prognostic biomarkers and individualized therapies. Citation: Braun R, Finney R, Yan C, Chen Q-R, Hu Y, et al. (2013) Discovery Analysis of TCGA Data Reveals Association between Germline Genotype and Survival in Ovarian Cancer Patients. PLoS ONE 8(3): e55037. doi:10.1371/journal.pone.0055037 Editor: Amanda Ewart Toland, Ohio State University Medical Center, United States of America Received November 9, 2011; Accepted December 21, 2012; Published March 21, 2013 This is an open-access article, free of all copyright, and may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose. The work is made available under the Creative Commons CC0 public domain dedication. Funding: The authors are supported by the Intramural Research Program of the National Cancer Institute, United States National Institutes of Health, Bethesda, MD. There were no external funding sources for this study. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. * E-mail: rbraun@northwestern.edu Introduction Ovarian cancer accounts for about three percent of all cancers in women and is the fifth leading cause of cancer-related death among women in the United States, with an age-adjusted incidence rate of 12.8 per 100,000 women per year and death rate of 8.6 per 100,000 women per year (2003–2007) [1]. Of the gynecological cancers, ovarian cancer has the highest mortality, with an overall five-year survival rate of 43.7% for white women and 34.9% for black women [1]. The poor survival statistics are attributable to the late stage at which ovarian cancers are diagnosed due to their asymptomatic nature: while stage I tumors have a 92.4% relative survival rate, they account only for 15% of ovarian cancer diagnoses; by contrast, stage III and IV cancers have survival rates of 34% and 18%, respectively, and together account for 65.4% of diagnoses [1]. Response to standard chemotherapy (platinum plus taxane) is highly variable [2,3], and tends to be poor for advanced cases [2]. Understanding the genetic determinants of ovarian cancer survival and response to treatment may improve these statistics, particularly for stage III and IV patients who comprise the majority of cases. In particular, identifying variations that predict response to chemotherapy PLOS ONE | www.plosone.org 1 March 2013 | Volume 8 | Issue 3 | e55037