GSEA assigned na_pos and na_neg

[Tony]

 I am using GSEA (v4.1) linux command line as follows:

./gsea-cli.sh  GSEAPreranked  -rnk tumor_vs_control.rnk -gmx /data/msigdb_v7.1_GMTs/h.all.v7.1.symbols.gmt  -out /data/output

The process runs to completion. 

For the GSEA Report for the dataset. I generate summaries to include statements:

20 / 50 gene sets are upregulated in phenotype na_pos

26 / 50 gene sets are upregulated in phenotype na_neg

The .rnk file has two columns (gene symbol, logFC). The data is from an RNASEQ experiment. In the rnk file, genes have +ve logFC (when the gene expression is high in tumor compared to normal) and genes have a -ve logFC (when the gene expression is low in tumor compared to normal).  

How do I interpret GSEA assigned  "na_pos" and "na_neg"   given the content of the rnk file. Are genesets positively correlated in the tumor when GSEA assigns na_pos? Are genesets negatively correlated in tumor when GSEA assigns na_neg?

Thank you,

[Anthony Castanza]

Hi,

In GSEA's Preranked mode, GSEA uses the positive and negative directions from the RNK file as-is. So, whichever direction you defined as positive when constructing the rnk file will be assigned na_pos, and whichever direction you defined as negative will be assigned na_neg.

-Anthony

Anthony S. Castanza, PhD

Curator, Molecular Signatures Database

Mesirov Lab, Department of Medicine

University of California, San Diego

--
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/50c3101f-f6a1-4f44-b02b-7047d3f509d0n%40googlegroups.com.

[Tony]

Hi Anthony,   

Thank you for clarifications. Regards to a follow up question. Reviewing the GSEA output for the na_neg category and the gene-sets therein.  

if a given gene-set has negative scores for ES and NES  (ES -0.53271891;  NES -1.321645) and further statistical significance (Nominal p-value = 0; FDR q-value = 0.077690321; FWER p-Value = 0.055);

Can I describe the gene-set as being negatively correlated in tumor compared to control or negatively enriched in tumor compared to control?

Thank you

[Anthony Castanza]

Yes, you can describe the gene sets with negative scores as negatively enriched in tumor vs. control, you could also say that the sets are enriched in the control.

This is covered in more detail in the GSEA user guide: https://www.gsea-msigdb.org/gsea/doc/GSEAUserGuideFrame.html. The user guide primarily describes standard GSEA but the underlying logic is the same for GSEA preranked except that you’ll have precomputed the ranking calculations in Preranked mode.

 

-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/ae73fb8e-989e-43df-bc20-a5d6c592efa5n%40googlegroups.com.