A systems biology approach for the identification of significantly perturbed genes

Identifying the genes involved in the mechanisms differentiating two phenotypes is a crucial step in the analysis of high-throughput gene expression experiments. Although in the last decade several approaches have been developed in order to address this challenge, a number of important issues remain open. Even the most widely used approaches fail to incorporate information about known interactions among genes, and they often fail to yield reproducible results across similar experiments, both in terms of the set of genes and in terms of the mechanisms related to those genes. Here we propose a novel systems biology approach able to i) identify the genes that are involved in a biological mechanism relevant to the condition in analysis and ii) yield reproducible results across multiple data sets related to the same condition. This is achieved by using gene expression levels and existing knowledge about the interactions among genes. We apply our method on four data sets describing two conditions, and we compare our results with the results of the classical approach of identifying genes based on their differential expression and p-values. The results show that our approach is better at identifying genes that are involved in the mechanisms relevant to the phenotype in analysis, as well as producing more consistent results across data sets describing the same biological condition.