Hello, 

I am wondering what the correct way to define the design in the EdgeR workflow is for a pairwise analysis, while also accounting for batch effects, in terms of the design <- model.matrix(~x+y) term. 

The setup of our experiment is quite simple: we have 18 cases who had blood RNA-seq before and after treatment, but these samples were run in different batches. Therefore, we expect batch effects and would like to correct for them. We would like to find the main effect of treatment without the effects of batch. 

Any help would be greatly appreciated. 

Thank you!

Anna

Paired analyses are automatically batch corrected, because you are already correcting for the baseline level of each case. This serves to adjust for batch effects and indeed for any other differences between the cases. There no need to do anything extra in the analysis.