[BioC] Go term memberships.

[Seth Falcon]

"Piero Cantone" <piero.cantone at gmail.com> writes:

> Dear mailinglist,
>
> I am new to bioconductor, and I have a simple task to performs, so I would
> like to ask for yours assistance. I have to analyze an old gene expression
> experiment, performed on 5 microarray chip: mu11kA mu11kB, mgu74B, mgu74C2,
> and an old GEM platforms for with I have accession number as univocal IDs. I
> need to check for some GO terms previously defined by my previously; in
> particular I have to retrive all the member.s of the terms in mus musculus,
> and then check for intersection with differentially regulated genes.
>
> So in contrast to the typical procedure, I am not looking for enriched
> categories, but for some specific terms, regardless of their enrichments (at
> least in the beginning). I looked in manual and Workshop, but I can't make
> it up.
> I tried
>
>> tmp=as.list(GOALLENTREZ)
>> match(tmp, my_entrez, nomatch=0)
>
> but I didn't find matches. Actually i think this is because GOALLENTREZ
> provides entrez from different species. furthermore HyperG function in
> GOstats requires the definition of a chip, but in my case it is a little bit
> more complicated.
>
>
> which strategy would you suggest?

Perhaps the mouseLLMappings package would be a better place to start?

my_matches = eapply(mouseLLMappingsGO2LL,
                    function(egids) intersect(my_entrez, egids))

And the above eapply approach should also work with GOALLENTREZ, but
since this contains data for multiple organisms it will be slower.

+ seth



-- 
Seth Falcon | Computational Biology | Fred Hutchinson Cancer Research Center
http://bioconductor.org