Interpro Enrichment - Any Organism

Question

[boczniak767]

Among many tools for over-/under-representation analysis there are few which are not constrained to fixed list of species.

For my knowledge it is Ontologizer and ASCA-Functional from SEA suite.

There one have to just provide already available association of genes with functional categories.

I wonder if you know any other "open-species" enrichment analysis tools.

I'm mostly interested in analysis of InterPro enrichment tools, where analysis rely on user submitted population and experimental sets with IDs and file with IDs association with InterPro IDs.

EDIT

I ended with creation of GO-like association and obo files for InterPro and use Ontologizer program. Much work to do but works fine.

[Michael]

Interesting question. Funny enough, after 11 years there exists no proper InterPro enrichment tool, one which can test for enrichment of IPR's directly (not IPRs mapped to GO). At least I haven't found any, not to mention species or model organism independence. The answers below provide no answer to the original question either.

When one googles for "interpro domain enrichment analysis" the top results are all doing something different, have misleading titles or otherwise are only distantly related to the problem. The top hit is this question. This biocnductor documentation is a good example of google's top hits.

Answer 1

[Casual]

I just found this one, WEGO by BGI in China. It can accept interpro scan results(raw ,xml,txt) as input.

Answer 2

[ALchEmiXt]

Not sure if I understand your question completely but maybe DAVID can do that for you. It can show enriched Interpro domains...

[boczniak767]

I saw this program at GO website along with similar software.

The problem with it is that it requires specification of type of IDs used. I have a mixed list so it doesn't work for me.

[boczniak767]

I learned that mgsa package allows what I need to do.

Answer 3

[Michael]

Currently, no out-of-the-box solution can do InterPro (IPR) enrichment analysis for any given organism I know of. The other answers did either never answer the question or are outdated as of today.

IF such an analysis should be conducted, and I think it perfectly makes sense (also indicate by the relatively large number of views of this question), one has to whip it up from scratch. It shouldn't be too difficult:

• Retrieve a background set consisting of tabulated counts for each IPR and all proteins in the organism, e.g. from UniProt or your own annotation with InterProScan
• Retrieve the test set of tabulated IPR counts for the subset of proteins in question from the same source
• Apply Fisher's exact test or hypergeometric test for each IPR with count > 0 using R.
• Optionally adjust p-values for multiple testing.