Extended Data Fig. 4: Sensitivity analysis of constraint in DHS elements.

(a) Distribution of non-primate mammalian scaling factors for DHS elements stratified by clade-specificity of constraint. The dashed gray line denotes where the mammal-constrained and primate-specific constrained distributions intersect. (b) Distribution of primate scaling factors for DHS elements stratified by clade-specificity of constraint. (c) Proportion of DHS with primate-specific constraint for variable FDR cutoffs in mammals excluding primates. Primate FDR is fixed at 5%. (d) Proportion of constrained DHS elements across clades when modeling substitution rates at a 1 Mb scale, compare to Fig. 2b. The estimated proportions are robust to differences between neutral substitution rates modeled in a regional 1 Mb context and a genome-wide averaged model. (e) Normalized Robinson–Foulds distance between 1 Mb scale phylogeny and canonical phylogeny along human chromosome 1. (f) Venn diagram intersecting DHS elements on chr1 classified as constrained in primates using regional substitution rate models and a fixed, canonical topology, or regional substitution rate models and a variable, regional topology. Models that accounting for regional differences in topology due to e.g. incomplete lineage sorting are highly concordant to those that use a single genome-wide topology (OR = 806.5, P ≈ 0, two-sided Fisher’s Exact Test).