An embedded and charset-unspecified text was scrubbed... Name: not available Url: https://stat.ethz.ch/pipermail/bioconductor/attachments/20070715/ 1cb9fcb5/attachment.pl

Hi Edward, Edward Bolanger <ebolanger at="" yahoo.com=""> writes: > Hi, > > I use a set of 1000 yeast genes as the 'universe' for the > hyperGTest. I then filter the resulting significant GO terms to those > that have a size count greater than 5. > > I now wanted to get the 'universe' of all possible GO nodes/terms for > my set of 1000 yeast genes. I suppose that this would involve mapping > all the 1000 genes onto GO, and then selecting only those nodes that > are associated with more than 5 genes. > > Is there any way that I can use hyperGtest to do this? Or is there any > other method/package you suggest that could help me with this? I think you can get what you want using the geneIdUniverse accessor on the result object. res = hyperGTest(p) GOMap = geneIdUniverse(res) has5 = sapply(goMap, function(x) length(x) >= 5) GOUni = names(GOMap[has5]) All of the accessors for the result object(s) are documented at: help("HyperGResult-accessors") Hope that helps. + seth -- Seth Falcon | Computational Biology | Fred Hutchinson Cancer Research Center http://bioconductor.org