GeneSets should have unique names

[hoor.a...@googlemail.com]

Hello,

I'm running GSEA on windows machine and I already tested it with some of the data available online (P53 dataset). Therefore I resumed to analyse my own data. However this error is occurring over and over again? my data has unique gene IDs and I'm analysing the data using the full list of available gene sets. When using the online version of the tool, GSEA produces results, therefore my assumption was the downloaded version should have the same capability?

the error message is too, attached the first 20 lines.

Thank you for your help

best

-----------------------------------------------------------------------------------

<Error Details>

---- Full Error Message ----
There were errors: ERROR(S) #:23746
GeneSets should have unique names. The looku ...

---- Stack Trace ----
# of exceptions: 1
------There were errors: ERROR(S) #:23746
GeneSets should have unique names. The lookup is case INsensitive. Found duplicate name: GLI1_UP.V1_DN
GeneSets should have unique names. The lookup is case INsensitive. Found duplicate name: GLI1_UP.V1_UP
GeneSets should have unique names. The lookup is case INsensitive. Found duplicate name: E2F1_UP.V1_DN
GeneSets should have unique names. The lookup is case INsensitive. Found duplicate name: E2F1_UP.V1_UP
GeneSets should have unique names. The lookup is case INsensitive. Found duplicate name: EGFR_UP.V1_DN
GeneSets should have unique names. The lookup is case INsensitive. Found duplicate name: EGFR_UP.V1_UP
GeneSets should have unique names. The lookup is case INsensitive. Found duplicate name: ERB2_UP.V1_DN
GeneSets should have unique names. The lookup is case INsensitive. Found duplicate name: ERB2_UP.V1_UP
GeneSets should have unique names. The lookup is case INsensitive. Found duplicate name: GCNP_SHH_UP_EARLY.V1_DN
GeneSets should have unique names. The lookup is case INsensitive. Found duplicate name: GCNP_SHH_UP_EARLY.V1_UP
GeneSets should have unique names. The lookup is case INsensitive. Found duplicate name: GCNP_SHH_UP_LATE.V1_DN

[Anthony Castanza]

Hello,

This error message implies that there are duplicate gene sets in the Gene Sets Database (.gmt) file. I was not able to replicate your issue with the P53 data set using the msigdb.v7.0.symbols.gmt. Which file are you using to run this analysis so that I can take a closer look at it.

Thanks,

-Anthony

Anthony S. Castanza

Curator, Molecular Signatures Database

Mesirov Lab, Department of Medicine

University of California, San Diego

[hoor.a...@googlemail.com]

thanks,

I'm using my own data now (175 genes) but with all available geneSet (h, c1, 7*c2, 3*c3, 3*c4, 4*c5, c6 & c7). I even reduced the sets to only h, c1, 7*c2, c6,c7, but I got the same error.

[acas...@gmail.com]

What do you mean by "7*c2", "3*c3", etc?

If your intent is to run GSEA against the entire contents of the molecular signatures database (which we don't really recommend, due to the presence of significant redundancy between some of the collections, but can be fine if you're running in phenotype permutation mode), you should be using the msigdb.v7.0.symbols.gmt which contains all of the MSigDB gene sets.

Also, GSEA is designed to be run on entire gene expression data sets, not highly filtered lists of selected genes.

-Anthony

Anthony S. Castanza, PhD

Curator, Molecular Signatures Database

Mesirov Lab, Department of Medicine

University of California, San Diego

[hoor.a...@googlemail.com]

thank you very much,

with the gmt file you suggested the error disappear and so it did when I chosed only the sets named with "all".

Totally agree with you, giving GSEA a pre-filtered list is not the right way giving its underlying algorithm. My confusion relies between giving GSEA the full GXP data (~20000 genes) despite knowing in advanse which subset are related to my phynotype, or reduse the noise and use the subset directly (e.g. ~2000 genes). Is there a way to use GSEA to continue where my analysis stopped instead of starting aparallel line? Please note that I'm not interested in generating hypothesis, but in interpreting the results.

thanks again