ChIP-seq: Should we be consistent with the control used (IgG/Input) among different conditions?

Question

[salamandra]

Imagine this situation with ChIP-seq analysis.

We have to conditions: A and B. For each condition there are three samples: one is ChIP, the other two are the controls (IgG and Input). For sample A the control with best quality (with higher number of reads and that passes QC) is IgG and for sample B the control with best quality is Input. If we want to compare condition A with condition B, should we use the same control in both conditions regardless of the quality of that control or can we use different controls for different conditions (IgG for A and Input for B)?

Answer 1

[Devon Ryan]

You should always use the same type of control, which should always be input. In my opinion IgG is a waste of sequence, it's replaceable with a blacklist.

[m.sadman.sakib]

What about H3 ChIP-seq as control? As H3 gives nucleosome occupancy, do you think it might be more appropriate instead of Inputs?

[Devon Ryan]

For some specific questions, sure. With that you're asking less "where's the enrichment of my modification" and more "where's the enrichment of my modification above where one would normally expect it assuming it's evenly distributed among H3 instances". For most cases a standard input would make more sense, but on occasion the aforementioned question is relevant.

[m.sadman.sakib]

Makes sense! Thank you very much!