So this screenshot is too far out. However, since the scale is 0-1 it has only counted 1 count per region.

If you converted from bed peak file to bedgraph, it essentially just counts 1 over each peak so that could be why you see this.

This is because a bed peak file is just a list of non-overlapping regions, i.e. the called peaks. genomecov breaks up the genome into windows then looks at each window and counts the number of regions in the bed file that overlap that region. Since the peak file is non-overlapping, it counts one read per region.

There could be other reasons, but I think this is most likely. If you just want to visualize the peak regions, then you could load your bed file directly. If you want to visualize the signal strength over these peaks, then you need the reads. Usually these are in a bam file, but could take the form of a bed file, but in that bed file each line would need to be the aligned location of a read. With the reads, you can then accurately count the number of reads over a peak.

If there was a type and you created a bedgraph from bam file, then maybe you can share how to created that file.