Hi,

I have raw counts per exon matrix and want to normalize and do a VST or log2 transform without running any differential expression (for downstream usage separate to DE), so I just need the VST/log2 count matrix.

Is it sufficient to simply run estimateSizeFactors on the DEXseq object and then run vstfrom DESeq2 (or log2 + 1) on the normalized counts before exporting to a csv? are there reasons that VST or log transformations might not be appropriate for exon counts?

thank you

Hello. Short answer is yes, count-level transoformation for variance stabilization also work for exon-count level matrices.