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 -- -------- January Weiner --------------------------------------

Hi, On Tue, Jun 25, 2013 at 7:18 AM, January Weiner <january.weiner at="" gmail.com=""> 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? You might try the limma::voom + limma::camera option suggested in this thread: https://stat.ethz.ch/pipermail/bioconductor/2012-May/045420.html Or perhaps the GSVA package: http://bioconductor.org/packages/release/bioc/html/GSVA.html -- Steve Lianoglou Computational Biologist Bioinformatics and Computational Biology Genentech

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. Best, Jim 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. Biostatistician University of Washington Environmental and Occupational Health Sciences 4225 Roosevelt Way NE, # 100 Seattle WA 98105-6099

Le 25/06/2013 18:52, Sylvain Broh?e a ?crit : > Le 25/06/2013 18:52, Sylvain Broh?e a ?crit : >> Hi January, >> >> I recently discovered the nice package "piano" for bioconductor. It >> is very easy to use and displays the results in a very nice way. >> >> http://bioconductor.org/packages/release/bioc/html/piano.html >> >> I also developped my own GSEA tool (web). It works in C++ (multicore) >> and R. That can maybe interest you. If you wish, I can provide you >> with the code. >> >> http://rsat.bigre.ulb.ac.be/~sylvain/new_gsea/php/ >> >> Cheers, >> >> Sylvain >> >> >> Le 25/06/2013 16:18, January Weiner a ?crit : >>> 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 >>> >> >