Why cannot call wanted peaks by using IgG as control

Question

[JochenL]

Hi, I am new on Chip-seq analysis. We designed the experiment by compare one wild-type cancer cell line with IgG and one target gene knock-out cell line. I used MACS2 to call peak from two conditions: 1) WT VS IgG 2) WT VS Knockout. I got the logLR signal track and it showed a peak in both of these two conditions. When I look into .narrowpeaks file and the generated excel files, WT VS Knockout has a peak here but WT VS IgG has not. And the peaks from IgG are far less than the peaks call from Knockout. (The DNA concentration from IgG were very low when we finished the experiment) Can anyone give me some suggestion? Thanks

Answer 1

[dariober]

I would annotated the screenshot you posted to mark what tracks are there (WT, KO, IgG) and where are the peaks called by macs. Anyway, at a glance the number of reads is pretty low so it's not surprising that a peak is called in one condition but not in another.

It's not clear to me what suggestion you are looking for... KO and IgG control for different factors so it's expected to have different peak sets. IgG should control for aspecific binding of the antibody and library prep (it's normal to have very little recovery, ideally you should recover nothing). KO controls for bindings that are not affected by the knocked out gene.

In general I find the peak calling exercise very unstable, it's not at all unusual to see peaks present in one case and absent in another even if you would say the two cases look very similar.