plotCorrelation scatter plot ChIP-seq data

Question

[Marco Pannone]

Hey everybody!

I am performing correlation analysis between replicates of my ChIP-seq dataset and my aim is to generate scatter plots with the Pearson correlation score associated.

In order to do so, I have first generated a correlation table with multiBamSummaryand then used it as input into plotCorrelation. When I generate a scatter plot, I also use --skipZeros and --removeOutliers arguments.

However, I noticed that my scatter plots look a bit unusual and not comparable at all with regular ones, as you can see in the image below. It's my first time producing these specific plots and I wonder if there is something missing or wrong in my procedure.

Thanks!

[ATpoint]

Based on our previous interaction my guess is a far too large binsize.

bamCoverage bin size for narrow and broad histone marks

[Marco Pannone]

Thanks for the reply (and for remembering my previous question) :) I think the two topics are not related in this case since I use .bam files as input into multiBamSummary, not the .bigWig output files of bamCoverage. However, I have noticed that multiBamSummaryalso has a bin size argument and that, for this tool, the default value is 10000. Maybe then this can be the reason?