Benchmarking pathway analysis methods offer novel prognostic cancer biomarkers and therapeutics: application in bladder cancer

The pathogenesis of cancer is typically driven by alterations in multiple cellular pathways that are challenging to identify and target. However, it is unclear how suitable the existing pathway analysis tools are for unbiased discovery and ranking of dysregulated and/or cancer-specific pathways. Here, we created a new platform called Benchmark to evaluate the potential of pathway analysis tools for discovery under experimental conditions. Unexpectedly, we found that despite wide-spread success in confirming hypothesized dysregulated pathways, common pathway analysis tools are less than ideal for unbiased discovery. Nevertheless, our pathway ensemble tool (PET) that combines the rank statistics from the exisiting methods significantly enhanced discovery. We applied PET to transcriptomics data from 12 independent tumor types to identify prognostic pathways. We showed that the genes from prognostic pathways are excellent biomarkers and can define cancer molecular subtypes. Moreover, drug prediction to normalize genes from prognostic pathways identified effective known and novel drugs. Finally, in vitro and/or a xenograft models of bladder cancer treated with the top predicted drug significantly restricted tumor cells. We anticipate that our unbiased approach for pathway discovery will have tangible impacts on cancer management. Comparision of transcriptome of T-24 cells treated with DMSO and cells treated with CCT068127. T-24 (1.5 million) cells were seeded in a 6 well plate. After 12 hrs, cells were treated with either 1 micro molar concentrations of CCT068127 or DMSO for 48hrs. Trypsinized the cells and extracted RNA using Trizol reagent. Experiments were performed in three independent biological replicates.