Add notebook demonstrating how to use fgsea with EM

[risserlin]

add notebook to cytoscape workflow github
modify output files to be gsea or generic compliant
https://bioconductor.org/packages/release/bioc/html/fgsea.html

[aaz398]

I used fgsea to perform my GSEA analysis. I want to use EnrichmentMap to visualize my results but I am seeing this error:

Could you please help? I'm uncertain what the exact parameters are for the UP and DOWN files to ensure they are EnrichmentMap compliant.
Thank you

[risserlin]

Can you send me a sample of the output files for fgsea (just the top two or three lines of the file will be sufficient).

[aaz398]

I sent the files to your email. Thank you

…

On Wed, Jun 30, 2021 at 11:46 AM Ruth Isserlin ***@***.***> wrote: Can you send me a sample of the output files for fgsea (just the top two or three lines of the file will be sufficient). — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#412 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AUVLI5P4ILHD3CFFZUAZPCTTVM36DANCNFSM4M6AHR5Q> .

[risserlin]

Initial method - the format of the files that are coming out of fGSEA are not the same as the regular GSEA files. I would recommend converting them to generic enrichment files as opposed to changing them to the GSEA files. (There is a lot of extra info in GSEA format that can be annoying to add although it would be nice to have the NES values in your analysis) A detailed description of the file formats can be found here. https://enrichmentmap.readthedocs.io/en/latest/FileFormats.html#enrichment-results-files

[yj2024]

I'm using fgsea results as well. I separated the up- and down-regulated pathway list based on NES value and used gmt file downloaded from MSigDB. I got this error below-

I'm using Cytoscape Version: 3.10 and enrichmentmap v3.3.6

My fgsea file column are like this:

<style> </style>

	pathway	pval	padj	log2err	ES	NES	size	leadingEdge
1	KEGG_MEDICUS_REFERENCE_TRANSLATION_INITIATION	0.0001036	0.00043012	0.5384341	0.45868976	1.92531548	63	RPL18A,RPL18,RPL4,RPL21,RPS10,RPS8,RPS7,RPL9,RPL24,RPS23,RPL7A,RPL35A,RPL13,RPL19,RPS17,RPS11,RPS4Y1,RPS12,RPL34,RPL28,RPL35,RPL31,RPS14,RPL36,RPL10A,RPL8,RPS3,RPS20,RPL23,RPL23A,RPL32,RPL11,RPL12,RPS28,RPS24,RPS9,RPS27,RPS25,RPL6,RPS19,RPS6,RPL37A,RPL15,RPL30

Could you please help me with this? I'm not sure where the error is rooted..

[risserlin]

for your fgsea results to mimic GSEA results they need to have the following columns. Your current output format is not a recognized format for EM

NAME
description
GS DETAILS
SIZE
ES
NES
NOM p-val
FDR q-val
FWER p-val
RANK AT MAX
LEADING EDGE

Your above columns would need to be mapped as follows -
NAME --> Name
description --> Name
GS DETAILS --> you can put anything here
SIZE --> size
ES --> ES
NES --> NES
NOM p-val --> pval
FDR q-val --> padj
FWER p-val --> not used, just set to 0
RANK AT MAX --> this is used by EM but you need to calculate from your ranks file* and the leading edge in the results file. If you don't want to you can just set this to a random number but just know that the leading edge feature in EM won't be showing you the right answers.
LEADING EDGE --> leading edge

• to calculate the rank at max in R you could do something like this - where current_fgsea_results are the dataframe with your gsea results and current_ranks is the dataframe with your gene to rank mapping.
  /calculate the rank at max fgsea returns the leading edge.
  just need to extract the highest rank from
  set to get the rank at max/
  calculated_rank_at_max <- apply(current_fgsea_results,1,
  FUN=function(x){ max(which(names(current_ranks) %in% unlist(x[8])))})

One last thing, it looks like the file that you outputted has the rows numbered. Make sure they you are exporting the results from R that you set rownames = FALSE