My experiment has 2 types of sample; one containing my FLAG-tagged protein of interest and the other that is lacking the FLAG tag. Samples were taken at 2 timepoints, from 2 biological replicates and I also have a total DNA control.

I first used MACS2 to normalize every ChIP sample against the total DNA and call peaks. This identified approx 200 peaks varying in length from 200bp-2kb and this corroborates what I see when I view the BAM files in IGV.

I then used DiffBind to look for peaks that are significantly different between my FLAG-tagged samples and those that lack the tag. This identified approx 170 peaks significantly different between the two samples, however each peak is now just 2-3bp wide.

I would like to know how to get the full peak width, so that I can proceed with motif analysis etc. I have e-mailed the DBA files separately and include the session info below.

Thanks,

Emma

session info:
R version 3.1.2 (2014-10-31) -- "Pumpkin Helmet"
Copyright (C) 2014 The R Foundation for Statistical Computing
Platform: x86_64-apple-darwin13.4.0 (64-bit)

Hi Emma-

Thanks for sending a copy of the objects via email. I can see that in your sample sheet, you are using the MACS2 _summits.bed files as the peak files. These contain only the 1bp peak positions for each peak, and not the peaks themselves., You should use either the _peaks.xls files (with PeakCaller="macs") or the _peaks.bed files (with PeakCaller="bed") instead.

Cheers-

Rory