NEW YORK (GenomeWeb News) – In a study appearing online last night in the PLoS ONE, German and Swedish researchers used a variety of bioinformatics tools to find biological pathways that most often contained differentially expressed genes in nearly two dozen published gene expression studies of colorectal cancer outcomes.
The team used a pathway-based enrichment strategy to sift through expression data for almost 1,500 genes identified in 23 past studies that used gene expression profiling to try to find prognostic patterns for colorectal cancer. In the process, they found a handful of pathways that are enriched in these gene expression studies — including genes from cell proliferation, apoptosis, and oxidative phosphorylation pathways.
"[W]e used the largest collection of [gene expression profiling] studies on [colorectal cancer] prognosis so far, and for the first time applied and compared several enrichment tools to the extracted gene lists," corresponding author Jesús Lascorz, a molecular genetic epidemiology researcher at the German Cancer Research Center, and co-authors wrote.
"This strategy allowed use 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 over-represented pathways involved in [colorectal cancer] progression," they added.
Although there have been numerous studies looking at how gene expression relates to colorectal cancer progression and outcomes, the researchers explained, there is still little to no consensus on which genes and pathways are most informative on the prognostic front.
"The large number of published microarray studies on prognosis of [colorectal cancer], showing a very low overlap in results, has provided no generally accepted gene expression profile for prediction of [colorectal cancer] prognosis," they explained. "Additionally no genome-wide association studies of outcome in [colorectal cancer] have been published, but are now underway."
For the current study, the team used 10 bioinformatic enrichment software tools to find over-represented pathways in 23 published gene expression studies of colorectal cancer prognosis, using pathway information found in the Kyoto Encyclopedia of Genes and Genomes databases.
Of the 1,475 genes assessed, just over eight percent — 124 genes — were reported in two or more of the studies, the researchers noted, including eight of 13 genes identified in a past meta-analysis of colorectal cancer-related gene expression profiles. The direction of the gene expression changes was the same for 54 of the 124 genes found in multiple studies.
Among the pathways containing genes whose expression most often corresponded to colorectal cancer outcomes: cell proliferation, apoptosis, oxidative phosphorylation, and extracellular matrix receptor pathways.
"These categories have been functionally clearly related with cancer progression, and deserve further investigation," the researchers concluded. "It would be of special interest if future [gene expression profiling] studies performed in large sample cohorts could validate our results and identify these categories as classifiers for bad prognosis."