And agreed, it would interesting to see if peak-calling from bedGraph files is affected by this, as it would then not just be a cosmetic thing that looks less clean….

Update: We think we have understood the reason for the changes in background level. We have confirmed that the values between options

--bg

or

--bga

for bedtools and

--skipNAs

or default for Deeptools bamCoverage are exactly identical, of course with the exception than zero-covered bins have either

0

or

n/a

values. When looking in IGV at bp resolution, both

n/a

and

0

values are displayed in an identical way, as is expected: