I am trying to use diffbind on paired end read data. I have done the following to get our proper peaks of interest ( we have technical replicates:
myDBA=dba.peakset(myDBA, consensus=-DBA_REPLICATE, minOverlap=3)
I then used Rory's instructions to extract the raw read counts (this is what I need for our next step):
myDBA <- dba.count(myDBA, peaks=NULL, score=DBA_SCORE_READS)
counts <- dba.peakset(myDBA, bRetrieve=TRUE)
However, when I look at the bam files (using IGV) there appear to be way more reads than are reported in the count matrix. How exactly are the reads counted? I have been using your raw read matrix to move forward in my analysis and I am concerned that I am using at the wrong counts.
Any advice would be greatly appreciated.