Does H3K27ac generate narrow or broad peaks?

Question

[jklopp]

Hi,

I have been trying to find out whether the H3K27ac Histone modification generates sharp or broad peaks during a ChiP-Seq experiment and therefore to either use the "--broad"-flag in macs2 or not.

Several studies I found (http://www.nature.com/nrg/journal/v12/n1/full/nrg2905.html for example) mentioned that H3K27me3 "and other" histone modifications generate broad peaks but H3K27ac was never explicitly addressed.

In contrast some studies (https://www.researchgate.net/publication/299653522_Identification_of_Epigenetic_Biomarkers_of_Lung_Adenocarcinoma_through_Multi-Omics_Data_Analysis) even treat H3K27ac as narrow peak creating but unfortunately lack suitable sources.

I really appreciate your help!

Answer 1

[Devon Ryan]

H3K27ac is a sharpish mark, so you can start with the defaults.

@kennethcondon2007: Some marks (e.g., H3K4Me3) are known to produces sharp distinct peaks. Other marks (e.g., H3K9me3) produce wide >10kb regions of enrichment. ChIPseq itself isn't any noisier than ATACseq or anything else, the noise is due to the underlying biology.

[BioinfGuru]

Clarification always help, thank you :)

Answer 2

[BioinfGuru]

All the broad flag does is combine narrow peaks that are very close together into 1 peak. Run MACS2 with 1) broad, 2) narrow and 3) gapped options and then view the 3 results in IGV and you'll see it clearly.

Not so sure about the next bit but my 2 cents is...:

I dont think the mark itself produces the "broad" they are referring to. I think the use of the word "broad" is referring to resolution or noise of Chip-seq data, for example ATAC-seq data would be expected to produce narrower peaks than Chip-seq data.