I am using DiffBind and I find it very useful and straightforward to run. However, after running a few trials, I seem to get very different correlation plots (and matrix) when I change the file format of the peaks from the same peak calling. I use MACS2, therefore for each sample I get 3 different files: summit.bed; narrowPeak; .xls. I have tried to input summit.bed and .xls (changing the PeakCaller in the sample sheet in, respectively, bed and macs), and results are very different just running the basic
bed_samples <- dba(sampleSheet = bed_samples_dataframe) plot(bed_samples) corvals_bed <- plot(bed_samples)
xls_samples <- dba(sampleSheet = xls_samples_dataframe) plot(xls_samples) corvals_xls <- plot(xls_samples)
It gets even more inconsistent when I run the respective dba.count (with summits=250), with correlation values going from 0.75 to 0.5 in the two cases for a given comparison, for example. Any idea what's going on?
Many thanks for your help! Anna