Re: [gsea-help] phenotype continuous

[Anthony Castanza]

Hi Tabita,

 

This error is caused by the ranking metric selected for the dataset being incompatible with continuous metrics. If you expand the "Basic fields" section in the GSEA application, you'll see a parameter "Metric for ranking genes", by default this is set to "Signal2Noise" which is what GSEA uses to rank categorical (two phenotype) datasets, however this doesn't work for a continuous metric, you'll probably want to change this to "Pearson" which will internally rank genes by computing their Pearson correlation with the designated continuous phenotype.

 

Let me know if you have other questions

 

-Anthony

 

Anthony S. Castanza, PhD

Curator, Molecular Signatures Database

Mesirov Lab, Department of Medicine

University of California, San Diego

 

From: gsea...@googlegroups.com <gsea...@googlegroups.com> on behalf of Tabita Haessner <tabbi...@gmail.com>
Date: Wednesday, September 15, 2021 at 10:01 AM
To: gsea-help <gsea...@googlegroups.com>
Subject: [gsea-help] phenotype continuous

Hi, 

I try the continuous phenotype GSEA, but somehow i always get this error. Even if my data only contains numbers.  

 

As the phenotype choosen was continuous, only continuous class metrics are allowed

 

Is it just me or what else is to consider when doing this? I tried already a lot. But i am pretty new with GSEA, so maybe it is just my fault somewhere...

 

Best wishes, 

Tabita

--
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/a4cf6f5f-bacf-4305-9f3b-db0211167b48n%40googlegroups.com.

[Ezequiel Dantas]

Hey I came across this thread, at found it quite useful. In my case is the opposite I want to do an analysis using a continuous phenotype label, but i dont know how to select it. I might be preparing the .cls file wrong. Would you happen to have one as an example? I would really appreciate your input. Thanks a lot.

Best,

ED

[Anthony Castanza]

Hi ED,

The guidelines for the Continuous format CLS file are given here: https://software.broadinstitute.org/cancer/software/gsea/wiki/index.php/Data_formats#CLS:_Continuous_.28e.g_time-series_or_gene_profile.29_file_format_.28.2A.cls.29

 

In the future, we'd ask that a new thread be opened to address new issues.

 

-Anthony

 

Anthony S. Castanza, PhD

Curator, Molecular Signatures Database

Mesirov Lab, Department of Medicine

University of California, San Diego

 

To view this discussion on the web visit https://groups.google.com/d/msgid/gsea-help/2ad8cb3f-fa58-4d75-8a54-328ec4645034n%40googlegroups.com.