TABLE 1. Significance of replication time-dependence of evolutionary divergence and human.

Human-chimpanzee and human-macaque divergence and human SNP density were stratified by one of six controlling variables including G+C content; the number of CpGs; exons (RefSeq); gene units (RefSeq); conserved non-coding sequences (CNS); and recombination hotspots (see Supplemental Fig. S1). p-Values were computed by both stratification (generalized Cochran-Mantel-Haenszel) and permutation re-sampling (Monte-Carlo) approaches. The distribution of each compositional feature was produced by sliding a non-overlapping 50kb window across each of the four temporal replication states (S1–S4) and binning composition values obtained into three classes (low, medium and high content), with an equal number of pooled windows under each level. Cochran-Mantel-Haenszel tests for three-way interactions were then performed using each compositional factor as a controlling variable. Monte-Carlo tests were carried out by running 200,000 random permutations of replication timing assignments for a set of generated 50kb windows for each feature, followed by calculation of Cochran-Mantel-Haenszel statistic as described above.

	p-Value	G+C	CpG	Exons	Genes	CNS	Recombination hotspots
Human-Chimpanzee divergence	Stratification	2.8·10−47	1.1·10−81	3.3·10−43	1.5·10−49	7.1·10−44	1.2·10−43
	Permutation	< 5·10−6	< 5·10−6	< 5·10−6	< 5·10−6	< 5·10−6	< 5·10−6
Human SNP density	Stratification	2.0·10−13	1.2·10−22	2.9·10−13	8.1·10−14	3.0·10−12	1.8·10−13
	Permutation	1.5·10−4	1.0·10−5	1.8·10−4	1.1·10−4	2.2·10−4	1.7·10−4
Human-Macaque divergence	Stratification	2.9·10−30	3.7·10−43	8.9·10−29	1.0·10−30	1.4·10−28	1.5·10−28
	Permutation	< 5·10−6	< 5·10−6	< 5·10−6	< 5·10−6	< 5·10−6	< 5·10−6