Plotting ChIP-seq read profiles relative to genomic features

[Gilman85]

Hello all,

I’m fairly new to the business of NGS data analysis. I was wondering whether anybody might have some advice or could suggest some resources for plotting average ChIP-seq profiles relative to genomic features such as transcription start sites and, particularly, enhancers. In my case, I’m looking to use such a plot to compare the H3K4me1 ChIP profiles between control and knockdown cells as they may occur relative to TSSs or enhancers.

Precisely what kind of files are required? A file for the total list of TSSs or gene transcript genomic regions found throughout the genome? A BED file corresponding to the H3K4me1 reads in control and knockdown cells? I’ve also noticed that with such read profiles (at least for certain histone mod reads plotted relative to TSSs) that some people group their reads together in 100bp windows and plot these relative to the TSS. How might I do something like this? Some people use average read density while others use cumulative reads occurring within such windows – Is using either the average read density or cumulative number of reads more advisable?

Also, how is it possible to make a BED file for something like enhancers detailing their genomic positions, or do we have to use surrogates like p300 binding sites and assume that these regions are likely to be associated with enhancers?

Thanks!

[mudshark]

if you are using R and coverage vectors, here is a good starting point

http://bioconductor.org/help/course-materials/2009/EMBLJune09/Practicals/TSS/

[Simon Anders]

If you know at least a little bit of Python (or feel like learning it), give HTSeq a try. I've just added a new chapter to the documentation, which uses the example of making such plots to explain how to use HTSeq.

[jingler]

These are what I am looking for.
Many many thanks

[howi]

If u want to have a quick and easy look at your data simply use SeqMINER. No R no Phyton ;D. It generates a heatmap around e.g. TSS and also shows you a density histogram you were looking for. It has a graphical interface and it will work on Linux/OSX/Windows.

[ParthavJailwala]

I agree with howi...seqMINER works very well to plot heatmaps around TSS.

[sbrohee]

This is an old post ... but I cannot resist to make a small advertising for my (beta-version) tool that might be of interest to some of you.

So I developped a tool which displays high quality pictures of binding profiles (peaks) around features. I presently works for 4 assemblies (2 mice and 2 human) but I would be happy to add more or to give the source code of the tool.

http://rsat.bigre.ulb.ac.be/~sylvain/binding_profiles/binding_profiles_form.php

Thank you for the feed-back.

Sylvain

[asiangg]

Give ngs.plot a try: http://code.google.com/p/ngsplot/

ngs.plot is developed in R but does not require any knowledge in R programming. All you need is just install it and then run like a command line program. It can plot multiple genomic features, such as TSS, TES, genebody, exons, CpG islands. For enhancers, you may have to put the genomic coordinates into a BED file and then give it to ngs.plot.