MACS2 for K27me3 peak calling

Question

[ac]

I'm trying to run MACS2 on K27me3 in a non-model organism. I know the peaks can be extremely broad and are generally weak compared to other ChIP signals but I'm wondering if you have input for what parameters work best at calling K27me3 besides the obvious --broad.

I ran it with default callpeaks --broad, using an input and was largely unsuccessful.

Thanks!

Answer 1

[Ian]

Hi. Despite the general pattern that H3K27me3 is a broad, region-like mark I have often found it useful to run both the default (narrow) and --broad versions of MACS2 as the resulting profiles (seen by bigWig coverage) are often mixed. The trouble with --broad is that the enrichment over background is deflated. Otherwise I would go with the default setting as the results for histone mark analysis by MACS2 is usually good. Looking at the results on a genome browser will indicate which way to go.

[msimmer92]

Very good idea to run it in both modes and then merge the results. Two questions: 1) would you recommend this for other histone marks as well (e.g. H4K5, H4K12)? 2) This post is from almost four years ago... does the new MACS2 have a mode to deal with mixed-type peaks data so that we don´t have to do that manually?