DiffBind ATAC-Seq Profile Plot Question About Identifying Sites

Question

[B.N.]

Hello, I was wondering if someone could help me out here. My ATAC-Seq profile plot has a lot of noise. The first picture is post analysis between two of the four groups in the second picture. I even tried plotting samples separately and they look pretty similarly to this. Does anyone know of a cause for this? Is there a way to reduce this noise. I also tried plotting just the counts without normalization (2nd picture) with normalize = FALSE and directly making profiles after calling the counts. I was also wondering how I can identify these sites in the plot that have too much noise and how I can take them out of the plot to maybe see how that is. Is that even advisable?