s2RNAi Determine reagent bias.

[jfear]

Story

Lee noticed that RNAi reagents are being sequenced when running RNA-seq. This causes an increased read count in the region of the reagent. We want to exclude these regions to minimize bias.

Questions and Tasks

• [ ] How big of a problem is reagent bias?

  This is a huge big problem. It does not matter if you are looking at raw counts or per-base counts.

  • [ ] Plot Avg Read Count w/in reagent vs Avg Read count w/o reagent

Definition of done

• [ ] Produce a counts table that accounts for reagent bias.

[HJosephLee]

yeah, and you need to exclude the region from controls samples when performing DE.

[jfear]

@hangnoh I am having an internal debate weather excluding the region would cause a bias. Wondering if scaling the counts based on the rest of the genome would be more appropriate. May not really matter as long as everything like RPKM are calculated on the adjusted gene length.

[HJosephLee]

Oh, I was wondering about it, too. Please let me know when you get the answer. But for me, excluding the region for both control and exp is somewhat similar to removing one exon. So I guess that gene level counts, which we do, could be relatively safe compared to exon level counts?