Hi All, I am analyzing Chip Seq data generated with 6 different antibodies: H3K4me3 H3K27ac H3K4me1 H3K27me3 H3K9me3 H3K36me3
I used macs2 for peak calling and DiffBind for differential bound sites between the subtypes of a cancer. Now I want to see the status for these histones around my genes of interest. Since many of these are enriched for enhancer and promoter regions, I am not sure which genomic coordinates I should consider in order to check their status around the genes of my interest.
I have read somewhere to consider +-10 Kb window around the TSS. I guess this could be fine for the peaks that have diffuse signal (like H3K9me1 and H3K27me3) but what is best for the narrow peaks?
I know H3K36me3 is highly enriched in intragenic regions, does that mean I should check for this in the CDS?
Any recommendations will be highly appreciated.
Thanks
Which question do you want to answer?
Hi @ATpoint, I want to see the enrichment for all the above mentioned histone marks but confused about which genomic coordinate for each mark I should focus around my gene of interest. Can you please help?
That comes down again to what question you want to answer: Enrichment over what? For enrichments one needs a comparison. What was the purpose of the experiment? If you simply plot signals over a couple of promoter or distal regions you will always find some enrichments over random regions, the question is what the underlying scientific question is. As for window sizes in general, 5kb typically give visually appealing profile plots for narrower marks. For diffuse one, try 10, 20 and 50kb.
To answer you question, I want to elaborate my scientific question.
I am comparing two subtypes of a cancer and found a list of differentially expressed genes predominantly from a specific biological pathway. Now I want to correlate this observation using the chipseq data from the same dataset. I hope to see the enrichment of activation/repressive marks in one of the subtype compare to the other. Since I already have a list of genes, I want to ask which genomic coordinates I should specifically focus on?
See if this make sense and please correct me if I am missing something. Thanks
Thanks for your reply. Can you please tell, how should I deal with the histone mark from intragenic regions like H3K36me3?
HK36me3 is very diffuse and challenging to compare quantitatively. Regardless, I would be focusing on gene promoters and gene bodies rather than intergenic regions for that mark. Your other marks (H3K27ac, H3K4me1, and H3K27me3) are likely to be more useful for intergenic regulatory elements, where a 1-5 kb window is often sufficient.
It is difficult to tie regulatory elements back to a specific gene though, as they can interact with multiple genes and be hundreds of KB away.
Hi @Jared, my question was about diffused peaks for HK36me3 at intragenic regions or at gene body, can you please explain how should I deal with that?
I have few more queries and wondering if you can help.
Fact) As per ENCODE, H3K4me3, H3K4me1 and H3K27ac generate gapped peaks while H3K36me3, H3K27me3, and H3K9me3 generate broad peaks.
a) Are these gapped peaks same as narrow peak? b) If it is so then do I need to check at a window around 1-5kb from TSS for H3K4me3, H3K4me1 and H3K27ac and for H3K36me3, H3K27me3, and H3K9me3 at a window around 10kb from TSS?
Yes, it makes sense to look at H3K36me3 over gene bodies, as it should be relatively spare outside of them.
Yes, peaks for those marks tend to be relatively narrow. There are no hard and fast windows to use for these. Try a few and see which look most appropriate for your use.