Fig. 4

Bipotent promoter chromatin accessibility is established within basal ductal cells during puberty. a Principal component analysis (PCA) of ATAC-seq profiles (represented by log2-RPKM intensities for merged peaks) for TEB and duct basal and luminal populations. b Number of peaks identified in each cell population classified by distance to the nearest transcription start site (TSS). Overlapping peaks between cell populations have been merged to make consistent peak boundaries across all populations. Black bars show total number of peaks after merging those called in each of the four cell populations (colored bars). c Differential accessibility between TEBs and ducts for the basal and luminal populations respectively (FDR < 0.05). Peak genomic regions are categorized based on the distance to their nearest TSS as in b. d Chromatin accessibility averaged over all genes from 5 kb upstream of TSS to 1 kb downstream of transcription termination site (TTS). e Average intensities of DA peaks in the different cell populations. Panels from left to right show peaks that are significantly more accessible in ductal vs TEB basal cells, TEB vs ductal basal cells, ductal vs TEB luminal cells, and TEB vs ductal luminal cells respectively. Within each panel, boxes show average log2-RPKM of the same peak regions in basal ducts (blue), basal TEBs (red), luminal TEBs (yellow), and luminal ducts (light blue). Boxes show median, quartiles and 10th and 90th percentiles. f Average RNA expression and chromatin accessibility of lineage signature genes. Top panels show results for positive signature genes for the basal, LP, and ML epithelial cell populations respectively and lower panels show results for negative signature genes [19]. Left panels show average log-expression (log2 counts per million) from bulk RNA-seq for the signature genes in basal and luminal cells from ducts and TEBs. Right panel shows average ATAC-seq coverage within 10 kb of the TSS of the same genes in the same cell populations. ATAC-seq coverage is represented as log2(RPKM+1) for the 20 kb TSS neighborhoods, ignoring ATAC-seq reads not associated with a MACS peak. g Scatterplots of genewise expression changes in TEBs vs ducts vs accessibility changes in the same genes. Expression and accessibility are both represented as log2-fold-changes (log2 FC). Upper panels show results for basal cells, lower panels for luminal cells. Panels left to right show results for ATAC-seq peaks within 500 bp of a TSS, 0.5–2.5 kb from a TSS, 2.5–10 kb from a TSS or > 10 kb from a TSS. Only peaks that are DA for the relevant comparison are plotted. Each DA peak is matched with the expression fold-change for the nearest gene. For each plot, the number of genes in each quadrant is shown, along with the linear regression slope and R value. h Coverage plots of chromatin accessibility in the different cell populations for key lineage-specific genes