Abstract 3649: Correlation of expression data and SNPs associated with aggressiveness of prostate cancer identifies specific associations.

Proceedings: AACR 104th Annual Meeting 2013; Apr 6-10, 2013; Washington, DC Previous genome-wide association studies (GWAS) encompassing 83,975 patients and 96,376 controls have identified more than 40 low-penetrance susceptibility loci for prostate cancer. The majority of the loci are located in regions of unknown function. Expression quantitative trait loci (eQTL) analysis may be used to study the association between gene expression and risk loci and provides a potential way to understand the functional consequences of risk loci. In the current study, we collected 32 prostate cancer samples with biochemical recurrence and 41 prostate cancer samples without biochemical recurrence for an eQTL association study. The Illumina HumanHap 1M platform was used to get SNP data from each patient. We used 559 SNPs previously associated with prostate cancer risk as candidate risk loci in the association analysis. This included the 40+ most susceptible loci reported in GWAS articles and 500+ loci with weaker associations, reported in their supplemental documents. The Affymetrix U133plus2 platform was used to get transcript expression profiling data from a tumor-enriched portion of the prostate in the same patients. We selected 4030 transcripts previously identified in prostate cancer diagnosis and prognosis studies as candidate expression quantitative traits. We developed a Bayesian clustering method to analyze these expressed transcripts and risk variants jointly in a single model. The model assumed that a risk variant can be associated with multiple transcripts and a transcript can be associated with multiple risk variants. An Expectation-Maximization algorithm was used to estimate each variable in the model. From the model, we observed 356 statistical significant linkages between 76 risk variants and 255 transcripts. Risk variants rs10163421 and rs14656182 are of the greatest interest because they were associated with 81 and 53 transcripts respectively. The two identified variants were also reported as common variants associated with aggressiveness for prostate cancer in two separate studies3, 4. We propose to do further analyses in order to verify the associations between the two risk variants and the associated transcripts, and also seek themes among the correlated transcripts, which might facilitate understanding the functions of the two risk variants in the genetic etiology of prostate cancer. 1. Gudmundsson et al. 2007 Nature Genetics, 39(5): 631 - 637 2. Eeles et al. 2009 Nature Genetics, 41(10): 1116-1121 3. Pal et al. 2009 The Prostate, 69(14):1548-1556 4. Catalona et al,, manuscript in preparation Citation Format: Xin Chen, Zhenyu Jia, Michael McClelland, Dan Mercola. Correlation of expression data and SNPs associated with aggressiveness of prostate cancer identifies specific associations. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 3649. doi:10.1158/1538-7445.AM2013-3649