[BioC] brief quesion on DESeq2

daniel.aguirre daniel.aguirre at cbm.uam.es
Wed Apr 10 11:52:32 CEST 2013


 On Wed, 10 Apr 2013 11:26:34 +0200, Michael Love wrote:
> hi,
>
>>> I was expecting that (having no replicates, for several conditions
>>> of a single factor/variable) using a dispersion model built from
>>> all my samples will increase the number of DE genes found with
>>> respect to a single pairwise analysis.
>>>  
>
> Adding C and D to the analysis can either decrease or increase the
> dispersion estimates for a given gene, for example if the counts for 
> C
> and D are close to or far from the counts for A and B.
>
> Mike

 thanks guys!
 D



More information about the Bioconductor mailing list