[BioC] GSEA for R

James W. MacDonald jmacdon at uw.edu
Tue Jun 25 16:32:44 CEST 2013

Hi January,

The limma package has three variants of GSEA, roast, romer, and camera, 
which are intended to do different types of GSEA. The first two are 
likely more similar to what you are talking about (what Goeman calls 
non-competetive gene set testing), and use rotations of the residuals 
for permutation, which greatly increases the number of combinations you 
can use.

The help pages are quite helpful for these functions.



On 6/25/2013 10:18 AM, January Weiner wrote:
> GSEA is an algorithm for testing gene set enrichment. It is
> distributed as a Java standalone program or an R script; however the R
> script is (a) crippled and (b) not straightforward to use (not really
> written "the R way").
> However, the algorithm itself seems interesting, since it (a) does not
> require arbitrary thresholds for selecting the enriched gene set and
> (b) uses a permutation approach to assess the significant changes for
> GOs, and so directly takes into account the actual variability of the
> data.
> What alternative presents itself for R?
> kind regards,
> January

James W. MacDonald, M.S.
University of Washington
Environmental and Occupational Health Sciences
4225 Roosevelt Way NE, # 100
Seattle WA 98105-6099

More information about the Bioconductor mailing list