Figure 2.
Transcription factor binding sites (TFBSs) and their disruption at active and regulatory variants
(A) Comparison of the total numbers of TFBSs between active variants according to their significance (active in none, CMs, VSMCs, or both). Number of observations per group is written below each boxplot, and asterisks denote significance determined by two-sided Mann-Whitney tests (∗∗p = 6.135 × 10−33). Box limits represent upper and lower quartiles. Central boxplot line represents the median and whiskers represent 1.5× interquartile range (IQR). Points represent outliers.
(B) Colored dots denote significant enrichment of predicted TF motifs across active and regulatory variants (hypergeometric testing, FDR calculated using the multicomp.multipletests function in Python with the Benjamini and Hochberg procedure, significance at FDR < 0.05).
(C) Enrichment analysis of TFs predicted to bind regulatory variants revealed cardiac and BP terms (top 10 terms are shown).
(D) Comparison of the total numbers of different TFBSs (absolute [number of TFBSs in reference allele] − [number of TFBSs in alternative allele]) between regulatory variants according to their significance (regulatory in none, CMs [∗p = 0.014], VSMCs, or both [∗p = 0.00073]).
(E) Counts of TFBS disruptions at regulatory variants for the top TFs in CMs and VSMCs.
(F) Enrichment analysis revealed important cardiac and BP terms of TFs that are predicted to be mostly disrupted at sites of regulatory variants.
(G) Examples of TFBSs disrupted by regulatory variants on either reference or alternative alleles. Horizontal black lines represent means.
(H) Bar plot depicting number of regulatory variants bound by expressed TFs in the cardiovascular system (only TFs bound to >2 regulatory variants in CMs are shown).
(I) Overlap between TFs bound (ChIP-seq in cardiovascular cell-types) at regulatory variants in CMs and VSMCs. See also Figure S3.