Search

The increasing availability of high throughput data arising from gene expression studies leads to the necessity of methods for summarizing the available information. As annotation quality improves it is becoming common to rely on the Gene Ontology (GO) to build functional profiles that characterize a set of genes using the frequency of use of each GO term or group of terms in the array. In this work we describe a statistical model for such profiles, provide methods to compare profiles and develop inferential procedures to assess this comparison. An R-package implementing the methods is available., Comment: Submitted to Journal of Statistical Planning and Inference