In summary, we envision that chromVAR will be broadly applicable to single-cell and bulk epigenomics data to provide an unbiased characterization of cell types and the trans regulators that define them. As such, we applied chromVAR to two bulk chromatin accessibility data sets18,19 down-sampled to 10,000 fragments per sample and data from an alternate scATAC-seq approach and find chromVAR to generalize to these additional data (Supplementary Figs 13–15; Supplementary Note 3). As methods for measuring the epigenome in single-cells and bulk populations continue to improve in throughput and in quality, scalable analytical infrastructure is needed. Analysis workflows for ATAC or DNase-seq data often include the identification of motifs enriched in differentially accessible peaks, but such approaches scale poorly to comparisons across many sample types and require sufficient read depth per-locus to determine differential peak accessibility (Supplementary Note 4). In contrast, chromVAR analysis is highly robust to low sequencing depth and readily scales to hundreds or thousands of cells or samples. Budget-constrained researchers often face a trade-off between the number of samples to sequence and the sequencing depth for each sample;
