Fig. 2: Identification of noncoding regulatory elements with primate-specific constraint.
From: Identification of constrained sequence elements across 239 primate genomes
a, Example of a primate-specific constrained DHS element in the GRIA4 locus (hg38; chromosome (chr.) 11:105608279–105612792). Assay for transposase-accessible chromatin with sequencing (ATAC-seq) insertions from human, chimpanzee and mouse iPS cells and phyloP constraint in primates and mammals. A putative TEAD4 binding motif that better matches primate sequences than non-primate mammal sequences is indicated. b, Proportion of constrained DHS elements across clades. c, Number of primate-specific constrained footprints (TFBSs) in DHS elements, stratified by constraint across the entire DHS. Error bars represent 95% confidence intervals. d, Average chromatin accessibility and the number of accessible cell types is higher at more constrained DHS elements. Colours indicate constraint categories from b. Error bars represent 95% confidence intervals. CPM, counts per million. e, A saturation mutagenesis experiment (MPRA) of a DHS element at chr. 2:191049304–191045304 (hg38). Average effects of substitutions at each nucleotide on transcriptional activity are correlated with phyloP scores from primates but not from mammals. f, Heat map of normalized correlation values (Spearman’s ρ) between primate phyloP and sequence-based Enformer predictions of regulatory activity across 438 ENCODE cell types. Categories of similar cell types corresponding to specific tissues are indicated. g, Normalized luciferase reporter activity in human iPS cells for three selected sets of primate-specific constrained DHS elements at orthologous primate and mouse sequences. Colours indicate constraint categories from b. Bars represent mean and error bars represent 95% confidence intervals; n = 36 across 3 elements. P values: 1.4 × 10−5 (left), 2.8 × 10−4 (middle) and 0.54 (right). Raw data are provided in Supplementary Data 6. h, Average chromatin accessibility in fibroblasts for five primate species at orthologous sequence elements stratified by sequence constraint. Colours indicate constraint categories from b. Error bars represent 95% confidence intervals; n = 90,827 DHS elements. i, Average Spearman ρ of H3K27ac levels at orthologous CREs for three pairs of species. Colours indicate constraint categories from b. Error bars represent 95% confidence intervals. n = 12 for human versus mouse, n = 10 for all other comparisons. ***P < 0.001; NS, not significant.
