bamCoverage taking four nucleotides instead of three when --MNase is used

[feraltp]

Hello,

I think there may be an error in bamCoverage when using the --MNase option. It seems that four nucleotides are reported (one left to the center, the center, and two right two the center) when I would expect three for odd fragment sizes; when the fragment sizes are even, the two nucleotides at the center are reported as expected. I think it would be good to point out in the documentation on the website that two or three nucleotides are considered (in reality two or four) depending on the fragment size parity.

I attach here the small bam file that has even and odd fragment sizes together with the bedgraph produced by bamCoverage (the issue is present also for the bigwigs).

BamBedgraph.zip

The deeptools version that I used is not the latest but I didn't see there was any change reported on CHANGES.txt for that particular thing.

deeptools --version
deeptools 3.4.1

python --version
Python 3.9.5

This is the command that I used (the original one had more parameters but I don't think they are interfering with the problem):
bamCoverage --MNase -b AlignmentSimulation.bam -o mnase_AlignmentSimulation.bedgraph -bs 1 -of bedgraph

Here is just one of the examples in IGV. The fragment is 449 bps long and the middle part is four bps long:

Thank you,
Luis