DEG analysys using pseudobulk from single cell RNA-seq

Question

[Diego]

Hi,

I would like to do DEG analysis of a specific cluster that is present in two different species (eg cluster A in species A vs cluster A in species B). I am wondering how I can do this using pseudo-bulk form snRNA-seq while accounting for differences in gene length between the species. I have seen that tools like Deseq2 and edger only accept row counts, so normalization that accounts for gene length such as TPM would not be optimal.

Looking forward to hearing from you, Diego

[Bastien Hervé]

If you are using the snRNA-seq protocol from 10X, you never get the full coverage of a gene but only 3' or 5' of it. So, I would say, genes length would not matter that much in that case. Check your alignment files for some genes to validate.

[Diego]

Hi, Thanks for the input, Bastien. I agree with your comment. The problem is that the length of 3' between sps can be quite different. There is also mapping to intronic regions, which also differ in length. I could try to only keep reads that map in exonic regions, but then I do not know how to preserve the clustering structure of my original single-cell object.