Supplementary Figure 4: Cis-regulatory elements in hematopoiesis and co-accessibility validation.

(a) Genome tracks of aggregate scATAC-seq data, clustered as indicated in Figure 2b. Arrows indicate the position and distance (in kb) of intragenic or distal enhancers in each gene locus. (b) MetaV4C plot of H3K27ac HiChIP data demonstrating HiChIP signal at Cicero-identified co-accessible cis-elements (linked to promoter elements). Each plot shows the aggregate HiChIP signal (from 3 T cell types, n=2 biologically independent HiChIP profiles per cell type) between linked cis-elements identified in scATAC-seq data. Each link is scaled so that the 0% position indicates the promoter site and the 100% position indicates the linked cis-element site. The peak indicates an enrichment of HiChIP signal at the linked peaks compared to surrounding genomic regions. Biased links are identified by differential peak analysis in the indicated scATAC-seq clusters. The center line represents the Loess fit, and shaded regions indicate 95% confidence interval. (c) Support for Cicero-identified co-accessible cis-elements by GTEX eQTL data. Shown is the mean enrichment of eQTL signal (determined in the indicated tissue type, bars indicate standard deviation from n=250 simulations) in co-accessible sites linked to promoter elements described in GTEX vs 250 permutations of ATAC-seq peak to genes. Greater enrichment is observed in immune tissues because scATAC-seq data profiled the relevant cell types. (d) Heatmap of log-normalized gene scores for the indicated genes. (e) UMAP projection colored by log normalized gene scores demonstrating the accessibility of cis-regulatory elements linked (using Cicero) to the indicated gene.