Differential histone modifications

[maryhagen]

Hello,

I am new to this group and to ChIP-Seq. I am looking for software for analyzing ChIP-Seq data from histone modification pull-downs. I've found the SICER program which looks like a good resource for normalizing data sets and finding peaks, but haven't had much luck finding programs that identify differences between samples. Does anyone have advice, suggestions on what to use?

Many thanks,
Mary

[adamdeluca]

There is a thread at the top of the bioinformatics section with a list of software.

http://seqanswers.com/forums/showthread.php?t=43

[kopi-o]

I have used CCAT with one of the samples as "control" sample and the other as "treated". This will give you "differential peaks" (that are there in the "treated" but not the "control")

[maryhagen]

Thanks Adam,

This is an awesome and somewhat daunting list of software. However, it's not clear to me which of these packages are designed to work with histone modification profiles. I like the fact that SICER has been designed to call peaks specifically for histone modification profiles, but I still need a program that will find the differential peaks between samples.

Mary

[Xi Wang]

Check http://cmb.gis.a-star.edu.sg/ChIPSeq/tools.htm and their paper:
An HMM approach to genome-wide identification of differential histone modification sites from ChIP-seq data

[hongbo919]

A professional tool for identification of differential histone modifications

Recently, we developed a tool EpiDiff including a professional tool QDCMR for identification of differential histone modifications.
The software of EpiDiff is available at http://bioinfo.hrbmu.edu.cn/epidiff/. Look forward to you for testing this software and helping us to improve the software.

[NextGenSeb]

Hey Mary,

I have traditionally used MACS peak finder to do the job, seemed to work well for well defined peaks like histone acetylation and H3K4me3, but struggled a bit with more extended peak regions such as H3K27me3. Moreover running treatment vs control in MACS assumes that there is no significant difference (i.e. sharing bias) between the input of treated and untreated samples, so falls down there a bit.

I have recently discovered DiffBind (http://www.bioconductor.org/packages.../DiffBind.html) which although I haven't tested it yet, looks promising and might be worth looking into.

Cheers
Seb