How to combine all three Ontology classes (BP, MF and CC) in a single analysis ?

Question

[User1029725]

Code snippet from topGO Reference Manual (page 4):

sampleGOdata <- new ("topGOdata", description = "Simple session", ontology = "BP", allGenes = geneList, geneSel = topDiffGenes, nodeSize = 10, annot = annFUN.db, affyLib = affyLib)

I tried

ontology=c("BP", "MF", "CC")

But, that is not allowed. Any other syntax to specify all three in one go ?

[Lluís R.]

I think it is not possible, as the topology method is based precisely on the differences on the structure of each sub ontology. But why do you want to do this?

[User1029725]

At the moment, my interest is to filter Genes that are over-represented based on enriched GO terms, not the GO class. After analyzing for BP, MF and CC separately, I could combine the enriched GO terms. My question is to find a way to do all three analysis in one go.

[Lluís R.]

The structure of the GO data, makes such a combination difficult: it is designed to have separate structure. However, I would recommend to run the over-representative analysis and then see which genes map to which GO found over-represented in your analysis. That way you could further filter your genes based on how much they are related to the significant GO terms. Another approach would be to filter by a function, location and process of your interest and see which genes are left. I think that the first approach will be easier and more robust, but depending on your study you might be want to pursue the second one.

Answer 1

[h.mon]

topGO uses the GO hierarchical structure to reduce the redundancy in GO terms, so I don't think it is possible to mix different ontologies in the same topGOdata object. Anyway, it seems more intuitive to me to test and present results of each ontology separately.