Question: Inconsistency of results GSEA clusterprofiler
gravatar for Stane
9 weeks ago by
Stane40 wrote:


I have been trying to reproduce results from gsea using clusterprofiler and gseKEGG

(for some unknow reason gseGO is not working for me )

simply repeating experiment on the same gene expression data about 13k genes: 

 gseKEGG(geneList = zscores,
                      organism = 'hsa',
                      nPerm = 1000,
                      pvalueCutoff = 0.05,
                      verbose = F)

about 80% of the time I am getting 0 gene set enrichment and the other 20% I am getting approximatively the same 8-9 gene sets. I have to compute gene set for about 400k assays and would like to make sure I can get consistent results. 

I understand that some parts of the process use some random number but I find strange that I am getting gene set enrichment only 20% of the time with the exact same data. 

I could implement a loop to repeat the experiment until I am getting results for each assay but is there any other method which I may not be aware of directly in the package? Or another package that you guys would recommend instead of clusterprofiler? 



ADD COMMENTlink written 9 weeks ago by Stane40

I just bumped the nPerm to 5k and now the results seems more consistent

ADD REPLYlink written 9 weeks ago by Stane40
Please log in to add an answer.


Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.2.0
Traffic: 311 users visited in the last hour