What is the difference between narrow peak, Broad peak and a gapped peak of a chip-seq experiment and is it okay to use all of these regions for downstream analysis

Question

[Saad Khan]

Many publicly available data sets (road map epigenomics data in particular) has three kinds of peaks (based on different parameters used to do peak calls in MACS2) namely "narrow" peaks, "broad" peaks and "gapped" peaks. I was wondering how are these three types of peaks different from each other and when doing downstream analysis using this kinds of peaks data would it be ok to use bedtools to merge these peak regions and use that instead of doing the analysis individually on all the different types of peaks.

Answer 1

[TriS]

You can find more info on the UCSC site and forum.

narrowpeaks can be generally called for TF, since the region bound is pretty much limited. broadpeaks are better for histone modifications or histone modifiers since the regions can be much wider. I'm not totally sure about gapped peaks but looks like you can used them when combining regions that contain gaps (yeah, intuitive...).

And yes, you can use them for downstream analysis.

Why would you want to merge them though? That can give you gigantic peaks with basically no information. If you look at the co-binding of multiple TFs or histone marks, then I'd go for overlaps.

[Saad Khan]

So I am looking at specific regions of Histone Modifications in Human data and these peaks may have overlapping regions in that range. So in your opinion if I am looking at histones I should only look at Broad peaks. Or look at each of those peaks separately.

[sophialovechan]

I have some similar analysis need to do. Have you finished it? How did you cooperate your TF data with your data for histone modification?

[Saad Khan]

I am looking at transcription factor binding sites and the patterns of histone modifications around those sites I was wondering if I should use different peaks(broad/narrow/gapped) for different types of histone modifications??

[TriS]

for histone mods I would use broad peaks, independently from the modification

[Saad Khan]

Then why do the Roadmap data have narrow peaks,broad peaks and gapped peaks for different kinds of modifications?

[msimmer92]

I read that you could also call peaks in two runs, one with MACS2 narrow and another one with MACS2 broad, and then merge the output to have a mixed-type calling. Is this a correct way to do it, or is it better to call the mixed-type factors with other peak callers, such as PeakRanger's BCP, CCAT and Ranger (or even others)? (I would use this for a histone mark that I suspect might be having narrow behavior and broad as well)(PS: I assume with gapped peaks you mean mixed type).

[Dataman]

Hi! The link to the forum is broken. Is it possible to update it?