We have updated our workflows for per base sequence coverage to use genomeCoverageBed from BAM files. However for pair-end data it seems as though the regions between pair-end reads are not counted.
To be clear I am not talking about using -split for not counting introns in a single read of a paired-end, instead I am looking to count the probable whole insert when the insert size is greater than the combined read length of the paired reads.
We've looked at using iRanges from BioConductor as well but cannot tell if this would do what we want.
Is there is hidden flag in genomeCoverageBed to count the whole insert as coverage, not just the sequenced ends? Is there another program out there what would work on BAM files?
I know I can alter the SAM file before BAM conversion but this seems like something that should be coded somewhere already.