[BioC] how to do it with biomaRt

Michal Okoniewski michal at fgcz.ethz.ch
Thu Nov 5 08:51:39 CET 2009


>
>> Btw, the job of translating rat exon probesets into genes and 
>> transcripts is done most quickly with exonmap,
>> assuming that you install a local copy of X:MAP database...
>
> The point is that, what we often do, is to use Gene Array or Exon 
> chips to study gene expression -not alternative splicing- so what I am 
> looking for is a flexible way to get the annotations for these chips a 
> the transcript cluster level.
>
> Thanks for the help
>
> Alex
>
>
Then - as an approximation I use Brainarray CDFs for the Entrez or 
Ensembl gene level - however it comes at a price of loosing
many genes as false negatives (same for transcript clusters, I suppose). 
Brainarray is updated quite often, so should be more precise
than transcript clusters, I suppose.

In the more precise version - get all the significant
probesets in the full set (1M for rat ) and check with exonmap to which 
gene they belong.
If there are several exons having the same direction and roughly similar 
magnitude of fold change - then such a gene is OK differentially expressed,
although you might have missed it with Brainarray mapping or transcript 
cluster approach in Partek .

Cheers,
Michal



More information about the Bioconductor mailing list