GO enrichment between two samples

[Fernas]

Dear all,

I want to study the GO enrichment between two samples (they differ in age) of a particular species.
I have two sets of data:
1) list of differentially expressed genes in sample 1 along with the GOs associated with each gene
1) list of differentially expressed genes in sample 2 along with the GOs associated with each gene

How can I study which enriched GOs between the two sets? do I need to have background genes (i.e. non differentially expressed genes)?
is there a tool that do that? (FYI: this species is not supported by most of GO analysis tools like Gorilla).

[mbblack]

Since you don't mention the species, it is hard to know what available or pre-built tools, like DAVID would be useful.

There are tools in BioConductor to compute a hypergeometric enrichment test for any set of gene lists (ie. library GOstats). So, if you have lists of genes of interest, what you then need is a list of the gene "universe" you wish to compare them to in your test for significant enrichment and a list of GO terms associated with them (an ontology in other words).

Do you have, or can you put together a gene universe and a (custom or otherwise) ontology for this species? If so, there is a slide deck under the courses at the BioConductor web site that lays out how to perform a hypergeometric enrichment analysis of those three things, namely:

1. A species-specific Gene "universe"
2. Your ontology, or GO categories (categorize genes by GO terms).
3. A list of interesting genes (e.g. differentially expressed genes based on your significance testing)

[Fernas]

Thanks Michael for your quick reply!.

My species is a type of honey bee. So, I am wondering if DAVID would be useful in this case? shall I paste the GOs of one sample as target and the other as background? or I put the uniprot IDs instead of GOs?

regarding your suggestion of Bioconductor GOstats tool. Yes. I can get the set gene universe (i.e. background set), but, let me ask what do we mean by the gene universe? is it the set of all genes in the annotation for this species? the set of all expressed genes in this species? or simply it is the background for both sets that I have (i.e. the non differentially expressed genes for each of the two samples?)?

I tried to find this slide deck but still unable to find it. Can you put the link for it?

Thanks very much

[mbblack]

You can certainly try running your list of differentially expressed gene IDs through DAVID and see what you get.

For R, the universe of genes would be all known/annotated coding genes for your species. And you would want the known GO terms associated with them. It could end up being a bit of a chore to put a decent custom ontology together to perform enrichment with, unless the information is somewhat readily available for honeybee somewhere.

You can also use R tools to just take your list of differentially expressed gene IDs and perform enrichment against all of the public GO. The entire public GO database is available in R already.

[chadn737]

Give Blast2GO a try.