Runx1 binding sites in a GMP-like cell line confirm binding to Runx1 target genes that are downregulated in its absence. (A) Venn diagram showing overlap (±150 bp from peak summit) of binding sites obtained from experiments using either the GMP-like FDC-P1 cells or the HSC/MPP-like HPC7 cells.²⁹  Differences in the binding sites between endogenous and RUNX1-ER may represent alternative binding of the fusion protein and/or higher expression levels. (B) Consensus sequences within 75 bp of the Runx1-peak summits identified by de novo motif discovery. The frequency of the observed motif in randomly selected Runx1 peaks (n = 500) and the calculated E-values are given. (C) Hierarchical clustering of the shared Runx1 peaks identified in FDC-P1 cells based on overlapping occupancy patterns with peaks identified for C/ebpα (GMPs),⁴⁴  Gfi1 (M/E transformed),⁷⁵  and Runx1 and Scl1 (HPC7).²⁹  Peak regions for individual TFs were set at a standard width of 300 bp. Each line corresponds to an individual Runx1-binding region where red/blue coloring indicates the presence/absence of the given TF within 152 bp of the summit. (D) Runx1-bound regions visualized as density blots (green) and peak calling (green bar) for 6 genes, which are upregulated in KOs and downregulated after re-expression of Runx1. Density blots for C/ebpα from GMPs⁴⁴  (blue) and open chromatin structure in GMPs (black),²⁸  as determined by the ATAC methodology, are shown. Called peaks for Runx1 and Scl1 from HPC7 cells²⁵  are depicted with red bars. Exon-intron gene structures are depicted in a 5′-3′ orientation using the UCSC Genome Browser. Asterisk by Gcnt1 denotes a packed view of 3 different transcript variants, each starting at a unique first exon but with shared second and third exons (last exons depicted).