As soon as we start zooming out further, additional levels of precalculated

bins

within the bigWig files start averaging over the underlying (identical) values, which now means that the 0-values start bringing the averages down, whereas the `n/a`s are simply ignored for the average calculation. In essence this means that with

skipNAs

the floor value a bin can have is the scale factor (or value for exactly the coverage of 1 read), whereas for

0-value

bins the minimum value can be as low as scale factor divided by #bp in the bin, so for 50 bp bins scale factor/50. This will eventually change the values for bins with a coverage close to background level, and since this is related to the Zoom level in IGV, it gets rendered differently, appearing as a massively increased background for what used to be identical coverage values: