DE analysis of low-expression transcripts

Question

[Antonio Camargo]

I'm working with lncRNAs, which have very low expression levels, and I'm not sure if I can conduct a standard DE analysis with them. Does packages like DESeq2, edgeR or sleuth work well with low expression transcripts? If not, how should I proceed?

Answer 1

[Devon Ryan]

The only issue with lowly expressed genes is whether they get filtered out during something like independent filtering. In DESeq2, either disable that or use edgeR instead, since it doesn't do it. Of course if you're just looking at the lncRNAs and nothing else then they won't be "lowly expressed" relative to each other and you won't have any issues.

Do note that statistical power is always at least partly a function of expression strength and that no package or method can get around that (only additional replicates can help).