Find GO terms sharing transcripts

[tboothby]

Hi,
Does anybody know if a tool exists for looking that the distribution of transcripts to GO terms?

Basically I have BLASTed a bunch of transcripts from and RNAseq project (non-model organism) against a C. elegan's refseq protein database. When I look at GO enrichments I notice that a lot of the GO terms are redundant (e.g., term A and term B share all the same transcripts).

I'd like a way to cross reference all my "enriched" GO terms with each other by looking at what transcripts are shared between them.

-Thomas

[JackieBadger]

couldn't you just use excel?

[SNPsaurus]

Thomas, the GO annotation has terms of increasingly fine granularity, so you will expect that at some level the GO annotation will probably have all the transcripts of the parent term. For example, if you heat shock your samples, you might see hsp70, hsp23 and hsp60 as the only terms in GO terms cell function, response to stress, response to heat, unfolded protein response.

If I understand what you want, you are looking for a way to figure out which GO terms are essentially all based on the same set of transcripts. Most GO enrichment programs will output the hierarchy, so if you take your up-regulated genes, add the C. elegans BLAST hit and submit those you should be able to see that. Try http://cbl-gorilla.cs.technion.ac.il/ , or look at http://neurolex.org/wiki/Category:Re...Ontology_Tools for a listing of tools.

[chadn737]

Are those terms really "redundant"? Sharing the same transcripts would not make a GO term redundant. For example, DNA binding and transcription factor activity would be non-redundant terms that share a lot of the same genes.

A better approach would be to visualize your GO in a tree, as a lot of the terms are children of another term. ReviGO is also nice for reducing a set of enriched GO terms to something a little more generalized.