It has been a awhile since I have done a ChIP-Seq experiment. I need to ChIP-Seq a chromatin remodeling subunit BPTF. I was wondering what the recommended samples would be for this technique. Besides your ChIP samples with your Ab of interest (anti-BPTF antibody), would an IGG control be used? The problem I have had in the past with that is that there is very little DNA in that sample and it is hard to make the library, or is total input used. The problem with that is that you need to sequence deeply to get good coverage of the genome. Or are both recommended?
Personally, I also do input only for the reason you mention. I would not be concerned about depth because peak callers like macs3 anyway downsample to the smallest file so sequencing input deeply does not help. After all, what you want from the inout is to reveal those regions in the genome that at a typical ChIP-seq depth show up, be it due to PCR or any other bias, and then like "sequester" this signal from the peak calling. There is to me not more use for inputs beyond protecting peak calling from artificial peaks.
Hi Joe: I recommend using input chromatin rather than an IgG control to avoid the issues you mentioned with IgG. Input controls offer a more reliable reference for assessing IP enrichment while allowing for deeper sequencing coverage of the genome.
For a more rigorous approach, you might consider siQ-ChIP (Dickson et al., 2020; 2023), which represents the cutting edge of ChIP-seq normalization, enabling quantitative comparisons between samples IP'd with the same or different antibodies, provided that reaction conditions are well-controlled (e.g., matched antibody concentration, chromatin concentration, and reaction volumes). This is possible because siQ-ChIP computes IP efficiency. Additionally, using input chromatin facilitates the interpretation of IP enrichment or depletion relative to background chromatin.
If this sounds useful for your work, I recommend checking out the siQ-ChIP papers: link (2020) and link (2023).
Personally, I also do input only for the reason you mention. I would not be concerned about depth because peak callers like macs3 anyway downsample to the smallest file so sequencing input deeply does not help. After all, what you want from the inout is to reveal those regions in the genome that at a typical ChIP-seq depth show up, be it due to PCR or any other bias, and then like "sequester" this signal from the peak calling. There is to me not more use for inputs beyond protecting peak calling from artificial peaks.