I need to draw the shape of peaks from different chip-seq samples.
I want to get something like this picture, which I found on a Nature paper.
Can you help me in doing that?
For this sort of profiles I'm quite happy with deepTools (look at the computeMatrix and profiler tools). Also SeqMonk has and option to draw profiles but is more limited and you are bound to the GUI (maybe a plus?)
I also a fan of deepTools for this. They also have a Galaxy server for those that want to avoid the command-line. Disclaimer: I have never used their Galaxy server.
I've never done this before, but I think you could do something like the following:
1. Get the summit position of all the peaks in your sample and convert them to BED format.
2. Expand the summit position upstream and downstream by 2,500 bp (like in the figure), using bedtools slop.
e.g. bedtools slop -i summits.bed -g my.genome -b 2500 > summits.2500.bed
3. Calculate the per base pair coverage for each peak region using bedtools coverage.
e.g. bedtools coverage -counts -d -a summits.2500.bed -b sample.bam
4. Load the coverage data in R, modify it so that you take the average coverage for each position, then plot a scatter plot of position against coverage.
Login before adding your answer.
Use of this site constitutes acceptance of our User Agreement and Privacy