I've generated a quant seq data that I intend to use to compare male and female gene expresion, with a focus on sexual chromosome.
For my species (three-spined stickleback), it is a classic XY sex determination system, and i have a male and a female reference transcriptome, including or not the Y chromosome.
I am not sure how to handle that with SALMON, and I have 2 questions in particular:
- Given that I use two different transcriptomes depending of the sex, can I still compare the obtained normalized pseudocount of would it affect the estimation too much?
- I want to compare expression levels of genes present in both the X and Y reference (i.e genes that are not recombining but have not been lost or degenerated on the Y). For that, I'd need to add the gene counts of the Y and X copy. Is it Okay to add pseudo count from different gene copies, or will that be issue? I do not understand the pseudo-count statistical model enough to have an idea about that
My other option is to go with a STAR / HTseq pipeline, but my mapping rates seems about 20% lower with this tools.
