13 Background: Whole genome and matching normal tissue analyses in prostate cancer patients have shown that widespread genomic changes occur in multiple distant regions of the genome simultaneously (Baca et al., Cell. 2013). The inherent genomic instability in prostate cancers causes activation of pathways involved in DNA damage, which could be a prognostic and potentially predictive marker for treatment with agents impacting DNA repair (e.g. PARP inhibitors). Using a published dataset, a “DNA damage signature” was generated to test its role as a potential prognostic marker. Methods: A 10-gene signature was generated, using the top 9 genes (MXD3, ZMYND19, BANF1, TOMM40, UBE2S, SNRPF, RPP21, CDK2AP2 and H2AFX) correlated with H2AFX expression, which is known to be unregulated upon DNA damage. TP53 was added due to its significant role in this pathway. This signature was then assessed in The Cancer Genome Atlas dataset (TCGA, unpublished data at NCI) as a discovery set and then validated in the MSK dataset (Taylor et al., 2010) using mutation, copy number, mRNA and proteomic data. Results: Genes in the signature were altered in 31% (77/246) of patients, by amplification and over expression compared to normal. The most frequent alterations were found in TP53 (15%), MXD3 (7%), BANF1 (7%), UBE28 (7%) and CDK2AF2 (7%). Analysis of protein changes associated with alteration in genomic signature showed upregulation of FOXO3, BCL2L1, BAK1 and PCNA (P<0.05). There were significant correlations between the signature and Gleason score (p=0.0015), PTEN (p=0.003) & RAD51 gene expression (p<0.001), but not with ERG,PSA, PIK3CA or BRCA2. Patients with alterations were associated with poor disease free survival (DFS) (p=0.002). When expression was assessed in the validation set, 73%(62/85) samples showed altered gene signatures and those with alterations also had poor DFS (p=0.003). Conclusions: The DNA damage signature can be used to define a group of patients with poor outcome and has the potential to be used as prognostic marker in treatment decisions in early prostate cancer patients.