Differential testing after qsmooth normalization

[tycho-kirchner]

Hi, and thank you for sharing this package.
Currently, I analyze multiple RNAseq datasets where a virus likely triggered a transcriptional shutdown, so global differences between samples are to be expected. I would like to kindly ask for recommendations/best practices on how to proceed with differential testing using e.g. limma/edgeR or Deseq2. Specifically, using limma-voom, currently I use a script similar to this but without calcNormFactors on the already quantile normalized data. Similarly, with Deseq2 I set sizeFactors(dds) <- 1.0 before calling DESeq(dds,. Would you consider this approach correct?

Thanks
Tycho