Creative Commons License
This blog by Tommy Tang is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.

My github papge

Thursday, October 3, 2013

compare ChIP-seq data for different conditions and different transcription factors?

I have done a lot of ChIP-seq data analysis, but most of them are just mapping to reference genome and then calling peaks (use MACS from Shirley liu's lab at Harvard). A little bit more is to intersect different peak bed files with bedtools and to make heatmaps.

The reads mapping is much simpler and faster than RNA-seq data. The data are smaller (~5Gb compare to RNA-seq ~25Gb). I recently asked myself how to compare different ChIP-seq data for different conditions ( control vs knockdown, different developmental stages etc).

I quick google search:

The available packages are:
"- ChIPDiff, which you already mentioned but which I haven't tried, but it was actually developed for histone marks so I'd be surprised if it didn't work for those?! I recall someone saying that there was some other issue with it (sorry I can't be more specific)

- DiffBind, which you also mentioned (; it uses DESeq internally

- DBChIP (, which appears to use edgeR

- You can also use edgeR and DESeq directly. The DESeq paper shows you how to re-analyze differential TF binding data in the Kasowski et al Science paper ( "

After reading a little bit, I decided to give diffReps a try.  The author says it is suitable for both histone modifications and transcription factors.
The paper was published in PLOSONE, it compares the results  with  that from DESeq and others.
The author Shen Li developed the other package ngsplot.

Another interesting R package is here
mentioned by me before
I will have to give it a try also.

At the same time, I am reading the Nature protocol:
 2011 Dec 15;7(1):45-61. doi: 10.1038/nprot.2011.420.

A computational pipeline for comparative ChIP-seq analyses.

it gives you a much better understanding of how to make the comparisons.

update 11/19/13

A table from the newly published paper

Practical Guidelines for the Comprehensive Analysis of ChIP-seq Data

Table S3.  Software packages for the analysis of differential binding in ChIP-seq.
Software tool
ChIPDiff [36]
Differential histone modification sites using a hidden Markov model
Comparative ChIP-seq [25]
Fold change ratio between normalized peak heights
DBChIP [33]
Assigns uncertainty measures in a test of non-differential binding (uses edgeR)
DESeq§ [31]
Test based on a model using the negative binomial distribution
Differential binding affinity analysis (uses edgeR and DESeq)
DIME [35]
Differential identiļ¬cation using mixtures ensemble
edgeR§ [32]
Empirical Bayes estimation and exact tests based on the negative binomial distribution
MACS [17] (version 2)
Differential peak detection based on paired four bedGraph files
MAnorm [34]
Robust regression to derive a linear model
Differences in shape using Kernel methods
Shape-based analysis of variation using functional PCA
Non-linear normalization on RNA Pol II profiling

§ Originally developed for gene expression count data.

updated on 02/06/2014

Two more useful links:

No comments:

Post a Comment