Exploiting statistical methodologies and controlled vocabularies for prioritized functional analysis of genomic experiments: the StRAnGER web application

StRAnGER is a web application for the automated statistical analysis of annotated gene profiling experiments, exploiting controlled biological vocabularies, like the Gene Ontology or the KEGG pathways terms. Starting from annotated lists of differentially expressed genes and gene enrichment scores, regarding the terms of each vocabulary, StRAnGER repartitions and reorders the initial distribution of terms to define a new distribution of elements. Each element pools terms holding the same enrichment score. The new distribution thus derived, is reordered in a decreasing order to the right, according to the observation score of the elements, while elements with the same score, are sorted again in a decreasing order of their enrichment scores. By applying bootstrapping techniques, a corrected measure of the statistical significance of these elements is derived, which enables the selection of terms mapped to these elements, unambiguously associated with respective significant gene sets. The selected terms are immunized against the bias infiltrating statistical enrichment analyses, producing technically very high statistical scores, due to the finite nature of the data population. Besides their high statistical score, another selection criterion for the terms is the number of their members, something that incurs a biological prioritization in line with a Systems Biology context. The output derived, represents a detailed ranked list of significant terms, which constitute a starting point for further functional analysis.