Recurrent gene fusions involving E26 transformation-specific (ETS) transcription factors ERG, ETV1, ETV4, or ETV5 have been identified in 40% to 70% of prostate cancers. Here, we used a comprehensive fluorescence in situ hybridization (FISH) split probe strategy interrogating all 27 ETS family members and their five known 5' fusion partners in a cohort of 110 clinically localized prostate cancer patients. Gene rearrangements were only identified in ETS genes that were previously implicated in prostate cancer gene fusions including ERG, ETV1, and ETV4 (43%, 5%, and 5%, respectively), suggesting that a substantial fraction of prostate cancers (estimated at 30-60%) cannot be attributed to an ETS gene fusion. Among the known 5' gene fusion partners, TMPRSS2 was rearranged in 47% of cases followed by SLC45A3, HNRPA2B1, and C15ORF21 in 2%, 1%, and 1% of cases, respectively. Based on this comprehensive FISH screen, we have made four noteworthy observations. First, by screening the entire ETS transcription factor family for rearrangements, we found that a large fraction of prostate cancers (44%) cannot be ascribed to an ETS gene fusion, an observation which will stimulate research into identifying recurrent non-ETS aberrations in prostate cancers. Second, we identified SLC45A3 as a novel 5' fusion partner of ERG; previously, TMPRSS2 was the only described 5' partner of ERG. Third, we identified two prostate-specific, androgen-induced genes, FLJ35294 and CANT1, as 5' partners to ETV1 and ETV4. Fourth, we identified a ubiquitously expressed, androgen-insensitive gene, DDX5, fused in frame with ETV4, leading to the expression of a DDX5-ETV4 fusion protein.