Hi All,

Is there a package / program that finds the peaks with the most differential pileup between the condition and input? For example, I am currently calling peaks with HOMER and MACS2; these programs automatically determine if the amount of reads in a certain region in the condition exceed the amount of reads in the input, and then labels that region as a peak.

I assume there is some threshold for labeling a region as a peak, but surely not all chip-seq peaks are the same; for example, one peak might have low binding affinity for that protein vs another peak being representative of a strong affinity binding site? I assume that the affinity of the site to the protein corresponds to the amount of reads found within that peak; would this biologically make sense? I am aware that annotation graphs around the TSS / TTS can be generated for the relative read count, but what I am looking for is the strongest peak signal between all peaks.

Furthermore, I did a knockout experiment and then differential analysis with Conditon1/control1 and Condition2/control2 using macs2 bdgdiff. If for example, both condition 1 and condition 2 had the same peak, but condition 2 had 50% less reads / 50% lower binding affinity to that site, how could I tell? Is there some program that can calculate this and tell me?

Basically, is there some way to display the variance of TF binding affinity between all the found peaks?