[BioC] edgeR: Contrast selection

[Simon Anders]

Hi

Krys Kelly wrote:
> You could try Tom Hardcastle's baySeq package which allows more complex
> designs.  There isn't an example of a three group design in the vignette,
> but there is one in the draft manuscript which we are about to send off.
> Perhaps Tom might email you a copy of this example.

Actually, I don't have a more complex design. The main limitation of 
edgeR is that you cannot form complex contrasts such as an interaction 
vs the sum of two main effects, or the like.

However, having more than two conditions should be possible, provided 
one always only contrasts one against the other.

Actually, I already got a reply by Mark (which somehow did not make it 
on the list): he is travelling at the moment, but they will look into it 
very soon.

Cheers
   Simon