Hi,
I would like to run some GO BP enrichment analysis in Cytoscape (not GSEA based on expression or rankings, just hypergeometric tests on shortlist vs. reference). There are two conditions I would like to fulfill when carrying it out:
1-To be able to define my own reference geneset (i.e. the genes that I have detected in my RNA-Seq)
2-To be able to discard GO terms comprising more than a given number of genes.
I have done so using topGO in R, but I wanted to try it on Cytoscape to get some nice visualization, using either BINGO or ClueGO. However, I have not been able to fulfill the conditions above cause:
1-It seems to me that BINGO does not allow to define a custom reference geneset and that enrichment of the query set is calculated vs the whole genome annotation.
2-I am not sure if ClueGO does that when you compare to sets, one of them being the reference set. Even if this is the case, ClueGO does not seem to allow discriminating GOs according to their gene size.
Am I wrong? Any hint on how to proceed will be greatly appreciated.
Thanks!
Best,
David
1 answer
Hi David,
Looks to me like ClueGO is your best bet, although it may not do exactly what you want. If you look under "Statistical Options" you'll see that you can choose a "Custom Reference Set", so that should take care of #1. I don't see a way to do #2 specifically in ClueGO, but you can take advantage of Term Fusion, which will definitely limit the terms that lots of genes have given the term's place in the hierarchy.
Hope this helps!
-- scooter
Log in to answer this question.