microrna-seq normalization in differential expression

Question

[kanwarjag]

I have raw counts of miroRNA seq. Do we normalize read count before we apply any t test/ DESEq? I believe we dont need length normlaization

Thanks

Answer 1

[Devon Ryan]

DESeq2/edgeR/limma (via voom) have their own normalization mechanisms. There's no need to do length normalization, the length of a given miRNA is the same between your groups.

[kanwarjag]

So Devon, In other words if I am not using DEseq/EdgeR or Limma we work with read counts without nay further normalization?

[Devon Ryan]

To be frank, if you're asking this question then you don't know enough to correctly analyze this type of data without a pre-made R package like DESeq2/edgeR/limma. If you continue this with some method you're coming up with then the results will very likely be wrong.