Abstract Prostate cancer (PC) is the most common malignancy, and the second most common cause of cancer related death amongst men in the Western world. The PC biomarkers currently available are unable to distinguish between aggressive and indolent PC, causing significant over-diagnosis and over-treatment. Here, by genomewide profiling of the DNA methylome, we aimed to identify new molecular markers that can improve the accuracy of diagnosis and prognosis of PC. In total, 21 PC, 9 normal (N), and 12 adjacent normal (AN) prostate tissue samples, as well as 15 prostate (cancer) cell lines, were run on the Illumina Infinium HumanMethylation450 BeadChip, which investigates <485,000 CpG sites. The methylation data was analyzed using the Mann-Whitney rank-sum test. Identification of differential methylation between N and AN samples could potentially aid correct diagnosis of patients with false negative biopsies. However, only 16 probes showed significant differential methylation in N vs. AN samples (p-value <0.05, Benjamini-Hochberg-adjusted, Δβ≥0.2), none of which corresponded to the same genomic locus. Also, multi-dimensional scaling, using the 10,000 most variable CpG sites, showed that N and AN samples clustered very tightly together, whereas PC samples showed great heterogeneity. Due to this similarity in methylation between N and AN samples, these were pooled into one control group for further analyses. We applied a Δβ cutoff value of 0.2 and a p-value of 0.05 (adjusted), which revealed 26,286 significantly differentially methylated CpG sites between PC and controls (3076 hypomethylated and 23,210 hypermethylated in PC). Global analysis showed that PC samples were hypermethylated in CpG islands (CGIs), a trend which was even more pronounced in CGI shores and shelves. Although less common overall, hypomethylation was also more prominent in shores and shelves compared to CGIs. Applying a strict Δβ cutoff value of 0.6, only 69 significantly differentially methylated CpG sites remained. Of these, 55 sites were gene associated (41 different genes), all were hypermethylated in PC samples, and all but one were located in CGIs. Out of these 41 genes, we selected 5 candidates for validation. In-house and publicly available gene expression datasets showed that at least three of the genes were downregulated in PC, consistent with silencing caused by aberrant promoter methylation. Initial validation was carried out in PC cell lines by bisulphite sequencing, which confirmed the array results for all 5 genes. We are currently validating these findings by methylation-specific qPCR analysis of a large independent patient cohort with long clinical follow-up, in order to assess the diagnostic/prognostic biomarker potential of these 5 novel candidate methylation markers. These results will be presented at the conference. Citation Format: Siri H. Strand, Michal Switnicki, Philippe Lamy, Soeren Hoeyer, Michael Borre, Jakob S. Pedersen, Torben Oerntoft, Karina D. Soerensen. Genome-wide profiling of the prostate cancer methylome for biomarker discovery. [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 659. doi:10.1158/1538-7445.AM2013-659