An Improved FWER-Controlling Method in Gene Ontology Graphs

• Activity Number: 550
• Type: Contributed
• Date/Time: Wednesday, August 7, 2013 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Garrett Saunders*+ and John Stevens and Clay Isom
• Companies: Utah State University and Utah State University and Utah State University
• Keywords: multiple comparison procedures ; GO ; p-value adjustment ; gene set testing

Abstract

The Gene Ontology (GO) provides three separate structured vocabularies by which genes specifying biological functions can be grouped. The structure of each of these vocabularies maps genes to certain biological functions where each function is logically linked as a node in a directed acyclic graph. The number of tests performed to test these gene sets requires a multiplicity adjustment to keep control over the family-wise error rate (FWER). Goeman and Mansmann (2008) proposed a focus level adjustment method which exploits the structure of the GO graph to adjust for multiplicity while controlling the FWER at a pre-specified level. This work proposes an improvement to the focus level method which also controls the FWER at a specified level. The advantage of the new method results in the ability to consider the GO terms individually and not just within their context of the GO graph as in the focus level method. Also, the new method is shown to demonstrate increased power over the focus level method to detect significant GO terms. The resulting method is demonstrated with application to identifying biological process differences between in vivo and in vitro maturated pig embryos.