Hey guys,
I am working on ChIP-Seq analysis using MACS and annotation of peaks using Homer. I have 2 different IgG ChIP-seq control sequences which give me different sets and numbers of peaks using MACS. I was wondering if there is a way to check the quality of reads or sam files to be able to choose 1. I have tried FASTQC for sequence quality scores and samstat for qualilty of sam files and both the files pass the criteria. What would be the best way to pick 1? Thank you for your help!
Most peak callers expect the control sample to represent the actual genomic background (e.g. input). So, though you can use it, the IgG-only control is not recommended for use as a control samples in ChIP-Seq peak call analyses as they don't actually represent the true genomic background for that sample, unlike using an input DNA control. IgG-only controls are useful for assessing whether there was potential non-specific carryover, e.g. DNA binding non-specifically to IgG. Also, the concentrations of DNA from those samples are (if performed correctly) significantly lower than IP or input.
We normally recommend using them in screening samples pre-sequencing as a control, e.g. assessing samples via QPCR, but we don't recommend sequencing them anymore.
EDIT: I should add, some callers do allow for IgG but I would certainly look at the documentation for the tool. You can certainly run the samples w/o any control as well.
I think that it is arguable wether input represents the actual background in a ChIP-seq better than an IgG control, but probably a lot of considerations went into this, so I'll gladly stand corrected :-)
As I see it, then the background might be affected by carry-over as you mention, and more or less systematically distored by the purification method - both locally and in more diffuse ways. Input would not capture this.
On the other hand, the higher DNA yield and number of reads as mentioned will improve statistics, and be a good reason to prefer input over IgG control. However, I think that having both is preferable.
@Chris Fields, @ Mads Lerdrup, Thank you for your comments. I will certainly consider an input control for the future experiments. Right now I am working on analyzing some ChIP Seq results previously done in my lab and only have the IgG control available. My concern is that I have 2 IgG control samples and each of them give different results in MACS analysis outputs and was wondering if there is a way to pick 1 of these 2 IgG control samples for further analysis or should I consider using a combined file?
The problem I have seen from IgG-only controls is that you only get the DNA that is present in the IgG only sample, which is not what most peak callers recommend for a control (they want a non-biased genomic background, with roughly equivalent or more reads than the IP). See What Control For Chip-Seq: Input, Igg Or Untagged Strain? for a pretty detailed set of answers on this.
I think that it is arguable wether input represents the actual background in a ChIP-seq better than an IgG control, but probably a lot of considerations went into this, so I'll gladly stand corrected :-)
As I see it, then the background might be affected by carry-over as you mention, and more or less systematically distored by the purification method - both locally and in more diffuse ways. Input would not capture this.
On the other hand, the higher DNA yield and number of reads as mentioned will improve statistics, and be a good reason to prefer input over IgG control. However, I think that having both is preferable.
@Chris Fields, @ Mads Lerdrup, Thank you for your comments. I will certainly consider an input control for the future experiments. Right now I am working on analyzing some ChIP Seq results previously done in my lab and only have the IgG control available. My concern is that I have 2 IgG control samples and each of them give different results in MACS analysis outputs and was wondering if there is a way to pick 1 of these 2 IgG control samples for further analysis or should I consider using a combined file?
The problem I have seen from IgG-only controls is that you only get the DNA that is present in the IgG only sample, which is not what most peak callers recommend for a control (they want a non-biased genomic background, with roughly equivalent or more reads than the IP). See What Control For Chip-Seq: Input, Igg Or Untagged Strain? for a pretty detailed set of answers on this.
EDIT: added a bit more text to clarify