Hello!

I want to normalise my small RNA seq data. I have allowed for multi-mapping. Should I count each alignment and then normalise with the likes of CPM or use fractional weights for the counts first and then normalise prior to DE analysis ?

Thanks in advance!!

Hey, Michael (the creator of DESeq2) has answered this before, a quick search for "small RNA DESeq2" would have given the answer.

Question 1 (multi mapping): Michael talk about multi-mapping here: Small-RNA analysis using DESeq2 And here: Multiple alignments (multi-mapped reads) and DESeq/edgeR pipeline

Main points: Power of detection (correct multimapping) vs false detection rate (wrong multimapping), is closely-related paralogs important to you? How small are the reads, anything below 20nt starts to become very likely to get hits by chance (non-paralog hits).

Question 2 (normalize counts or not): The DESeq2 vignette (tutorial page) talks about this here: https://bioconductor.org/packages/release/bioc/vignettes/DESeq2/inst/doc/DESeq2.html#why-un-normalized-counts What is most normal is to use un-normalized counts.