To ensure that each cell was both adequately sequenced and had a high signal-to-background ratio, we filtered cells with less than 1,000 unique fragments and enrichment at TSSs < 8. To calculate TSS enrichment > 2, genome-wide Tn5-corrected insertions were aggregated ± 2,000 bp relative (TSS-strand-corrected) to each unique TSS. This profile was normalized to the mean accessibility ± 1,900–2,000 bp from the TSS, smoothed every 51 bp and the maximum smoothed value was reported as TSS enrichment in R. To construct a counts matrix for each cell by each feature (peaks), we read each fragment.tsv.gz fill into a GenomicRanges object. For each Tn5 insertion, which can be thought of as the “start” and “end” of the ATAC fragments, we used findOverlaps to find all overlaps with the feature by insertions. Then we added a column with the unique id (integer) cell barcode to the overlaps object and fed this into a sparseMatrix in R. To calculate the fraction of reads/insertions in peaks, we used the colSums of the sparseMatrix and divided it by the number of insertions for each cell id barcode using table in R.