To avoid false positives, we applied two stringent selection thresholds before we considered a category as consistently enriched. First, only the categories reported to be enriched by several tools in a gene list were selected. From them, only the categories common in at least two of the three gene lists were considered to be consistently enriched. Using these two selection criteria, six general GO Biological Process categories, five GO Molecular Function categories,APS-2-79 hydrochloride and seven KEGG pathways were consistently overrepresented in the GEP studies on prognosis of CRC. The proportion of up and down-regulated genes was similar within each of the consistently enriched GO and KEGG categories, as in the 124 gene list. The ratio of enrichment was higher for the more specific and welldefined KEGG pathways than for the broad GO categories. A high overlap of the individual genes between these 18 categories was also observed. The large number of published microarray studies on prognosis of CRC, showing a very low overlap in results, has provided no generally accepted gene expression profile for prediction of CRC prognosis. Additionally, no genome-wide association studies of outcome in CRC have been published,YU142670 but are now underway. The heterogeneity in the GEP study design regarding the features related to disease progression makes a consistent comparison of results between the single studies very difficult. Here, we report the results of our approach, in which we used the largest collection of GEP studies on CRC prognosis so far, and for the first time applied and compared several enrichment tools to the extracted gene lists. This strategy allowed us to identify the oxidative phosphorylation chain and the extracellular matrix receptor interaction categories, as well as a general category related to cell proliferation and apoptosis, as the only significantly and consistently overrepresented pathways involved in CRC progression. In the first part of the study, we tried to overcome the lack of reproducibility in the GEP studies on CRC prognosis by selecting the genes reported in more than one study, in an attempt to reduce false positive results.