Getting a distinct set of Gene Ontology terms

Question

[Seigfried]

Dear all

I have a set of GO's with their corresponding NCBI-Entrez genes. A subset is as follows:

GO:0005982  starch metabolic process    2875020:2875019:2868278:2874098
GO:0006559  L-phenylalanine catabolic process   2873317:2874818:2874819:2874820
GO:0006572  tyrosine catabolic process  2874818:2874819:2874820
GO:0071472  cellular response to salt stress    2871848:2875595:2874155

I got these GO's from DAVID and I already know GO levels BP 1-5, BP_Direct etc

My question is, I want to raise the GO level (i.e. from 5 to 1) such that at the end, I will be left with GO's with distinct genes. i.e. for example for GO terms A and B, (A ∩ B) will be a null set.

Are there any tools to do this?

I tried ReviGO but it didn't give me a new/higher GO term, it just tried to remove 'redundant' GO terms.

Thanks

Answer 1

[Matteo Schiavinato]

Have a look at GO.db.

In this R package, you can easily get PARENT, OFFSPRING and CHILDREN terms from a specified GO term.

[Seigfried]

Thanks. I downloaded the available GO's from GO.db before this to remove some redundant GO's from DAVID. Found like 3 David GOs non-existent in GO.db

[Matteo Schiavinato]

If you found my answer useful, upvote + mark it as answer, so the thread gets the "answered" status. (not because it's mine, just because it works like this here :D)

Answer 2

[Seigfried]

I tried using Simplify from ClusterProfiler and it seems to work although I need to explore the package to check how well it works. For now the results appear to be quite nice!