How to plot GSEA result?

Question

[Raheleh]

Hello there,

I did standard GSEA for my microarray expression data. As I found out GSEA tools provide GSEA plot just for top 20 of enriched pathways. How can I provide plot for other pathways which are also significantly enriched (adj-pval < 0.05) in my dataset based on GSEA?

I am trying fgsea package in r. Could anyone please bring an example here that shows, how the input should look like for plotEnrichment()? What's wrong with my input?

I want to plot for TGFb gene set and this is my script : plotEnrichment(tgfb_gene, gene_list, gseaParam = 1, ticksSize = 0.2)

so for the first argument, examplePathway I used this list:

> tgfb_gene

$HALLMARK_TGF_BETA_SIGNALING

 [1] "TGFBR1"   "SMAD7"    "TGFB1"    "SMURF2"   "SMURF1"   "BMPR2"    "SKIL"    
 [8] "SKI"      "ACVR1"    "PMEPA1"   "NCOR2"    "SERPINE1" "JUNB"     "SMAD1"   
[15] "SMAD6"    "PPP1R15A" "TGIF1"    "FURIN"    "SMAD3"    "FKBP1A"   "MAP3K7"  
[22] "BMPR1A"   "CTNNB1"   "HIPK2"    "KLF10"    "BMP2"     "ENG"      "APC"     
[29] "PPM1A"    "XIAP"     "CDH1"     "ID1"      "LEFTY2"   "CDKN1C"   "TRIM33"  
[36] "RAB31"    "TJP1"     "SLC20A1"  "CDK9"     "ID3"      "NOG"      "ARID4B"  
[43] "IFNGR2"   "ID2"      "PPP1CA"   "SPTBN1"   "WWTR1"    "BCAR3"    "THBS1"   
[50] "FNTA"     "HDAC1"    "UBE2D3"   "LTBP2"    "RHOA"

and for the second argument, exampleRanks, I used a vector of the logFC (output of differential gene expression analysis), the names of vector is the name of genes. But I got this error and warnings:

Error: Aesthetics must be either length 1 or the same as the data (1): x, xend
In addition: Warning messages:
1: In max(tops) : no non-missing arguments to max; returning -Inf
2: In min(bottoms) : no non-missing arguments to min; returning Inf
3: In max(tops) : no non-missing arguments to max; returning -Inf
4: In min(bottoms) : no non-missing arguments to min; returning Inf
5: In max(tops) : no non-missing arguments to max; returning -Inf
6: In min(bottoms) : no non-missing arguments to min; returning Inf

Is there any other ways to provide or re-plot GSEA results?

I really appreciate any help!!

Answer 1

[tbg]

I have been working on GSEA too and I found two very interesting resources but I am not that proficient with this kind of analysis so I am still not able to interpret the plots.

However, here you can find a very nice tutorial for R that gives you some plotting. Also there is a question and another tutorial about how to interpret the graphical results.

Finally, this is an article about GSEA.

Hope this helps!

[Raheleh]

Many thanks @gabrit. I did it. I am trying to add title for my plot. Do you know how I can do it?

[tbg]

which plot of the many listed in the tutorial are you referring to?

Answer 2

[Barry Digby]

Hi Raheleh,

Here is a link to a RNA-seq analysis where I used GSEA and fgsea. Use the following link : https://github.com/BarryD237/D-O-Connor and scroll down the markdown file until you find "fgsea". This workflow assumed you used DESeq 2.

Regards,

Barry

[Aynur]

Hello Barry, Thanks for this example, but this link no longer exist. Can you please help me with this ? I used it before and it was very helpful.

[Barry Digby]

Some chump deleted my github accounts. I only recovered my primary account so that workflow is no longer available.

Follow this for fgsea - https://stephenturner.github.io/deseq-to-fgsea/#using_the_fgsea_package

And this for GSEA - How to prepare .gct file for GSEA software?

[Barry Digby]

Github repo has been restored, the link should work now. Might I suggest using @KevinBlighe EnhancedVolcano package in place of the (cumbersome) volcano plot code I used.

Answer 3

[guillaume.rbt]

If you are using the GUI version of GSEA you can set the number of plots that will be generated.

You have to go in "advanced fields" where there is a field "Plot graphs for the top sets of each phenotype" which sets the number of plots.