In addition to the previously published data, we generated three new replicates of single-cell K562s (ATCC; validated using STR genotyping (Genetica DNA laboratories)) using the previously published protocol4,8. Bulk ATAC-seq and scATAC-seq data was aligned and filtered as described previously4,8. Uniformly processed DNase data was downloaded from the Roadmap Epigenomics Project Portal18. ATAC-seq data from Lavin et al. (2014) was obtained from GSE63341 and processed as follows: adapters were trimmed using Cutadapt20, reads were aligned using Bowtie221, and filtered for mapping quality (mapq > 30). For the scATAC-seq data from the GM12878 and HEK293T mixture from the combinatorial indexing approach, a count matrix was obtained from GSM1647122.
