Question: What is the best peakcaller for H3 antibody?
0
gravatar for grx0326
8 weeks ago by
grx03260
grx03260 wrote:

Hello there!

I am working on a chip-seq data of general H3 histone antibody. As expected, there is not much difference between treatment group and control group, and I got so few peaks using MACS2 that I can hardly proceed further annotation and GO analysis.

I know it`s kind of weird to use such an antibody, but I am trying to find whether there are still H3 in the nucleus in my mutant.

In this case, what kind of peakcaller would be the best to use? I tried MACS2 and SICER2, but they both output poor number of peaks(less than 500). I assume that if there really are peaks of H3 they should be broad, but adjusting parameters to broad mode in MACS2 would not work.

chip-seq • 197 views
ADD COMMENTlink modified 8 weeks ago by hyperghjhon0 • written 8 weeks ago by grx03260
2

What was the thought behind this experiment? H3 histone is so ubiquitous in the genome that I'd be very surprised if peak calling A) even makes sense and B) has any realistic shot of working given the background is going to be so high.

ADD REPLYlink written 8 weeks ago by jared.andrews077.9k
1

Adding ln this, do you have a control sample? I would not even be surprised if most of these few peaks would either be removed by input control or removed after filtering against the NGS blacklists as published by ENCODE for the reasons Jared mentions above. H3 should be pretty much everywhere so these peaks even be noise if you have no input co trol and do not filter for common artifact regions.

ADD REPLYlink modified 8 weeks ago • written 8 weeks ago by ATpoint41k

I have a control sample ,just the same as normal chip-seq. Actually I am not sure if there`s a difference between my chip-seq and an ATAC-Seq, cause they both seem to have no specific preference to chromatin. If ATAC-Seq can have peaks in macs2, I guess my H3 experiment can also have peaks

ADD REPLYlink written 5 weeks ago by grx03260
1

Actually I am not sure if there`s a difference between my chip-seq and an ATAC-Seq, cause they both seem to have no specific preference to chromatin. If ATAC-Seq can have peaks in macs2, I guess my H3 experiment can also have peaks

This statement raises some concerns. I highly recommend you read up on ATAC-seq and what it actually measures and why ChIP-seq for H3 is not at all comparable to it.

ADD REPLYlink written 5 weeks ago by jared.andrews077.9k

I used DNase to cut the fixed chromatin and precipitate nucleosome with H3 antibody instead of the traditional ultrasonic way, sorry if I said “chip-seq”mislead you. I don`t see any fundamentally difference between this and an atac-seq because they cut and gathered almost the same area in genome--H3 spreads so wide on chromatin that if I precipitate H3, almost every accessible nucleosome is precipitated. If atac-seq can get peaks, my experiment should at least get as many peaks as atac-seq can do. Sorry I am not sure if I have made my points clear.

ADD REPLYlink written 5 weeks ago by grx03260

I am expecting a global decrease of H3 in a very specific mutant. Actually I`m not sure how to demonstrate this,chip-seq is just a naive approach

ADD REPLYlink written 5 weeks ago by grx03260

Is there any reason you aren't just doing a western blot for H3? If you're expecting a global decrease, that's almost certainly the easiest and most reliable way to measure it. It's certainly more reliable, less expensive, and easier than ChIP-seq.

ADD REPLYlink written 5 weeks ago by jared.andrews077.9k

Global decrease is also not measurable by ChIP-seq as you simply get lower amounts if precipitated DNA which is not a reliable measureI think. A WB definitively would be easier here and good to quantity protein amounts.

ADD REPLYlink modified 5 weeks ago • written 5 weeks ago by ATpoint41k

Western blot is exactly what I did, but I am not sure if WB is enough. Just thinking chip-seq might be more accurate?

ADD REPLYlink written 5 weeks ago by grx03260

As I said, sequencing is typically not appropriate when you want to explore global changes because it is a relative measure. If everything changes, then relatively it looks the same. I personally would find WB the most convincing experiment while ChIP-seq would imho be inappropriate and a waste of money both because ChIP-seq only makes sense if you can detect peaks (but H3 is everywhere) and if you can detect changes (which you probably can't given that your hypothesis is true and global differences are the ground truth).

ADD REPLYlink written 5 weeks ago by ATpoint41k

Make sense! I will think about it. Thank you

ADD REPLYlink written 5 weeks ago by grx03260

Maybe you could do ChIP on a couple of loci followed by qPCR and then show (as % against input) that in your treatment condition you get consistently less signal as this is an absolute measure against the input rather than sequencing which is relative (thinking aloud, jared.andrews07 this makes sense?)

ADD REPLYlink written 5 weeks ago by ATpoint41k

It makes sense, but a WB is still more clear and probably less time intensive.

ADD REPLYlink written 5 weeks ago by jared.andrews077.9k
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.3.0
Traffic: 1266 users visited in the last hour