[BioC] clarification on DEXSeq output

[Simon Anders]

Dear Nirmala

On 2012-07-31 17:32, Akula, Nirmala (NIH/NIMH) [C] wrote:
> Since the above mentioned exon belongs to multiple isoforms (in some
> cases to multiple genes), is there a way to identify exactly to which
> isoform (or gene) this exon belongs to?

Yes: It belongs to several isoforms, as you said. And no: DEXSeq cannot 
tell you which of these isoforms are the ones that are differentially 
expressed.

This is by design. The rationale behind DEXSeq is that it is not only 
easier and offers more power to test on the level of exons rather than 
on the level of isoforms but also might give results that are easier to 
interpret. For our arguments, please see the discussion in our paper:

S. Anders, A. Reyes, W. Huber: Detecting differential usage of exons 
from RNA-seq data. Genome Research, 2012.
http://dx.doi.org/10.1101/gr.133744.111

Best regards
   Simon