To visualize scATAC-seq data, we read the fragments into a GenomicRanges object in R. We then computed sliding windows across each region we wanted to visualize for every 100 bp “slidingWindows (region, 100, 100)”. We computed a counts matrix for Tn5-corrected insertions as described above and then binarized this matrix. We then returned all non-zero indices (binarization) from the matrix (cell × 100-bp intervals) and plotted them in ggplot2 in R with “geom_tile”. For visualizing aggregate scATAC-seq data, the binarized matrix above was summed and normalized. Scale factors were computed by taking the binarized sum in the global peak set and normalizing to 10,000,000. Tracks were then plotted in Loupe Cell Browser, an interactive visualization software that shows scATAC-seq peak profiles for scATAC-seq cell clusters, similar to the analysis done in this manuscript and described at https://support.10xgenomics.com/single-cellatac/software/visualization/latest/what-is-loupe-cell-browser.