General schema of the approach. First, module and expression data repositories are created. (a) Left: gene modules were obtained from a number of sources such as Gene Ontology and MSigDB as well as from expression datasets listed in Table 2. Right: high-throughput expression profiling experiments of tumor samples with clinical information were normalized by median-centering the expression value of each gene across all the samples and dividing the value by the standard deviation. In the heat map, purple color indicates low expression while yellow means high expression. (b) The first step in the pipeline is sample-level enrichment analysis (SLEA) of the dataset with each of the modules (M) in each of the datasets (D). (c) The second step is survival analysis according to the enrichment pattern (EP) for a module. (d) The results of the enrichment analysis (EA) can be visualized in Gitools as heat maps. Red indicates significant upregulation of the module while blue indicates significant downregulation. Grey is for non-significant results. GOBP, Gene Ontology biological process. (e) Differentially enriched modules are studied for their correlation to some clinical feature, in this case, survival. Shown here are Kaplan-Meier curves of samples with two different enrichment patterns.