Removal of ERCC spike-in in RNA-Seq

Question

[anu014]

Hello Biostars!

Can anyone tell me if is it necessary to remove spike-in sequences before aligning FASTQs to reference FASTA OR is it necessary to consider spike-in while normalizing in DEseq2 ? And if yes, how to handle spike-in in RNA-Seq? Can STAR or DEseq handle it somehow?

Please suggest me. Thank you :)

Answer 1

[Devon Ryan]

Presuming you mean ERCC spike-ins the best practice is to include them in the reference genome. However in practice not doing so has little if any effect, so we rarely bother. Unless you have a use-case that absolutely requires spike-ins, it's best to simply ignore them (i.e., do not use them in any way in your analysis).

[anu014]

Hi @Devon . What if the data is single cell RNA-seq?

[Devon Ryan]

scRNA-seq can sometimes benefit from spike-ins. Just align to them and use them for computing scaling factors.

[Daniel]

If I am not mistaken this answer is now outdated? It seems that more and more people are starting to only trust in data which has spike-in normalization, for all types of rna-seq.