Project description:DNA sequencing studies have identified specific recurrent somatic mutations that drive the aggressiveness of localized prostate cancers. Surprisingly, though, it is poorly understood how the prostate cancer proteome is shaped by genomic, epigenomic and transcriptomic dysregulation. To fill this gap, we profiled the whole genomes, methylomes, epigenomes, transcriptomes and proteomes of 55 localized, intermediate-risk, prostate cancers. This multi-modal dataset revealed that the genomic subtypes of prostate cancer converge on four proteomic subtypes, which are associated with distinct clinical trajectories. ETS fusion genes, the most common mutation in prostate tumours, perturb the proteome and transcriptome in divergent ways – with different genes and pathways affected at each level. Indeed, mRNA abundance changes explain only ~10% of variability in protein levels. Perhaps as a direct result, prognostic biomarkers that combine genomic or epigenomic features with proteomic ones significantly outperform those comprised of either molecular feature alone. These data suggest that the proteome of prostate cancer is shaped by a complex interplay of genomic, epigenomic, transcriptomic and post-transcriptional dysregulation.