Hello, doing atac-seq and getting results that show very high signal wide diffuse bands of red that seem to be skewing the rest of the graph. Is there a way to stop this from happening? Why is the background so high and can I reduce the background somehow? I used default parameters in the counting function and summits = 75. Should I use bUseSummarizeOverlaps = TRUE or bRemoveDuplicates = TRUE? Can anything be changed there to help with this or in the normalization function? I was wondering how I could and, more importantly, if I even should filter out these peaks (are they just noise that can safely be ignored?) or if that would mess up the analysis. The picture below is pre normalization. I did a profile plot using background normalization and got the same types of signals at the top of the plot. I was also wondering what a good normalization method is for atac-seq. Very new to atac-seq analysis so any help would be much appreciated.
