Use RLE normalised RNA-seq data in limma?

Question

[ntelkar]

I've decided to use the RLE (Relative Log Expression) normalisation method found in the DESeq2 package for my small-RNA RNA-seq data. Can I input these normalised counts into limma, or does it have to be logCPM transformed counts?

[ATpoint]

It should be on the log-scale, that is pretty much the only limitation if the underlying normalization is good which RLE is. So convert your norm. values to log and go ahead.

[ntelkar]

Oh, that's great - thank you!

[ATpoint]

For RNA-seq there is limma-voom though, you could simply follow their vignette.