Reduce redundancy in a GO enrichment analysis

Question

[dago]

I performed a GO enrichment analysis using topGO in BioC.

The analysis worked perfectly and now I get the following GO enriched (GO_ID) with the respective p-values (Fisher):

> head(cyt.MF.merged.combined[,c("GO_ID","Fisher")] )

 GO_ID  Fisher
1 GO:0000156 0.01314
2 GO:0004322 0.01705
3 GO:0004519 0.04516
4 GO:0004601 0.01917
5 GO:0005315 0.01705
6 GO:0008509 0.00769

However, many terms (e.g. GO:0016722, GO:0016724) are one the child of the other.

Since I do not need such a level of details, I was wondering if there is a tool to combine those terms that are related in a lower level of annotation. Of course, I would need to recalculate another p-value for the obtained parent node.

Has anyone experience with that?

Thanks in advance.

Answer 1

[russhh]

You could use GOstats, which provided a conditional=TRUE argument is provided to hyperGTest, can account for some of the conditional dependencies between GO terms

HTH

[dago]

Thanks for your answer. I know GOstats, and you are right it is a good tool for this kind of analysis. I was just wondering if is available another tool which can reduce redundancy after the enrichment is already performed with topGO.

Answer 2

[dariober]

Have a look at Revigo and/or EnrichmentMap (although revigo seems to be down just now...).

Dario