BioPerl: Best Gene Ontology API

[BioLion]

Dear Monks,

A bioinformatics question: I have been doing more and more Gene Ontology (GO) work, specifically, I am looking for enriched terms in my various gene sets. I usually go through the web-based programs like DAVID or Babelomics.

However, I now need a high-throughput approach, but was disappointed to only find GO::TermFinder on CPAN.

There are plenty of GO handlers/parsers available, but far fewer analytical tools, so I wondered if anyone had any experience of doing GO term enrichment on many many groups of genes?

1. Any particular API I should use?
2. Any particular problems I should look out for?
3. Is there another web-based program I should think about?
4. Is there a downloadable program that might be a good alternate? (i.e. a commandline based one that I can use via perl)

Any help or advice much appreciated!

NB: I have cross-posted a similar question on seqAnswers in an attempt to keep this one BioPerl/perl related...

[tospo]

I have used Go::TermFinder in the past and it worked fine for me for enrichment analyses, basically as desribed in the documentation for the module.

Might also be worth to post to the BioPerl mailing list: http://www.bioperl.org/wiki/Mailing_lists

[BioLion]

Thanks for the tip - I just wanted to survey opinion before committing to a particular API - I wasn't sure if I had missed another (possibly better!?) option.

Thanks again!

[zentara]

I havn't dabbled in BioPerl, but when you said "high throughput" I googled for "BioPerl PDL" and saw this: simplify the handling of biosequence alignments

[BioLion]

I was more looking for advice on the gene ontology part than the HT part - but thanks for the pointer to PDL, I am sure it will be useful!

[lima1]

You should maybe have a look at the Bioconductor R packages Category and GOstats as well. There are many annotation databases in Bioconductor and they are well maintained. So you can use basically the same code for enrichment tests of other annotation databases, for example KEGG, PFAM, Chromosome Bands etc. And it's easy to map ids (affy, illumina,..) to Entrez for example.

[BioLion]

Yes - this is what the SeqAnswers guys thought too - R/bioConductor seems to be much more on the ball with GO analysis compared to Perl. I still love her though.

[lima1]

Check out the Bioconductor case studies book (By Florian Hahne, Wolfgang Huber, Robert Gentleman, Seth Falcon). There's a whole chapter about this. The code is also available online. It will save you lots of time...

[Anonymous Monk]

You should have a look to the GOminer tool.
This page is a good overview of the available tools : http://www.geneontology.org/GO.tools.shtml

David