GSEA 1006 Error

[Shin Chan]

Dear GSEA help team,

 
I have downloaded and installed GSEA v4.2.3 (for Windows 10). Although the application runs successfully, whenever I try to run the analysis the status of GSEA reports is "Error" almost as soon as I hit the button (shown in supplementary figure). 

I have changed all the settings to try and figure the issue however, my colleague has the GSEA v4.1.0 version to run the application with same condition and files and the status of GSEA reports is "sucess" always.

I saw that a person before me also had this issue and although I'm still not sure why this issue is occurring I would be grateful if I could have the older version (4.1.0) of your software.

Thank you.

Sincerely,

Shin Young Chang.

[David Eby]

Hi Shin Young,

Would you mind please replying to us over on our private email address, gsea...@broadinstitute.org?  We don't send out these older versions through the public forum.  It would be most helpful if you could provide a copy of the data causing the issue so we can investigate further.  We will keep the data private and use it only for testing.

 I should point out a few important things that might make you reconsider using GSEA 4.1.0:

• These older releases use Log4J which had severe security vulnerabilities reported.  You are taking a risk installing this software and, if you are part of an institution, your IT department might require it to be removed.
• There are significant improvements coming soon and these will be based on the current code.  By choosing to stick with an older version you will not have access to these improvements.
• This message indicates an actual problem with your data combined with the chosen analysis parameters.  You would be better served trying to solve this problem than to work around it with an old version.

I should expand on this third point.  We've actually fixed a number of bugs in this release which could be making the difference in this case, where an error condition spotted in 4.2.3 was somehow missed in 4.1.0.  We fixed a number of issues related to metrics in this release and this is almost certainly one of them.

Regardless, this error message is real and points to a problem.  Fixing that might involve changing your analysis parameters or doing some data preprocessing.

Thanks,

David

David Eby
www.gsea-msigdb.org
igv.org
genepattern.org

--
You received this message because you are subscribed to the Google Groups "gsea-help" group.
To unsubscribe from this group and stop receiving emails from it, send an email to gsea-help+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/gsea-help/ca35b34c-a8d4-4f43-85de-a05861988eebn%40googlegroups.com.

[David Eby]

Hi Shin Young,

Just a quick follow-up.  This error should have popped up a dialog mentioning the error and giving a link to an explanation of how to deal with it.  Here's the link: https://software.broadinstitute.org/cancer/software/gsea/wiki/index.php/1006

The summary is that certain metrics are not valid for use with a limited number of samples.  The correct way to deal with this issue is to change which metric is used (under Basic Fields).  Which choice to make depends on your data, but it should be either "Ratio of Classes" (fold change) or "log2 Ratio of Classes" for natural scale data.  For log scale data, use "Diff of Classes" instead.  There is a fuller discussion of the metrics at this link http://www.broadinstitute.org/gsea/doc/GSEAUserGuideFrame.html?Run_GSEA_Page (under Metrics for Ranking Genes).

I hope this helps.  We would still appreciate a copy of your data if you don't mind, and it would be good if you could mention this to your colleague(s) who are running the old version of the software.

Regards,

David

David Eby
www.gsea-msigdb.org
igv.org
genepattern.org