Fig. 5
Relationships of genetic-variant density with recombination, natural selection and replication timing. a Strength of positive selection indicated by percentage of SNPs displaying DAF > 0.95 in the African (AFR), American (AMR), South Asian (SAS), European (EUR), and East Asian (EAS) populations within non-hotspot regions, total hotspots, total clusters, and Rec-Hs. b Enhancement of positive selection in clusters over non-hotspot regions in terms of |ΔDAF| between ethnic population pairs expressed by negative log10-transformed p values from unpaired one-tailed t tests (see “Statistical analysis” in “Methods” section). c Enhanced positive selection in clusters and hotspots relative to the non-hotspot regions indicated by elevated |nSL| values in five populations. d Excess of nucleotide diversity (left panel) or reduction of phyloP score (right panel) in hotspots and clusters. In (a, c, d), bars share the same color-codes, and p values (from unpaired one-tailed t tests) above the bars are shown in red or in blue indicating significantly higher or lower selection in hotspots/clusters relative to the non-hotspot regions. Since no population-specific phyloP data is available, right panel in (d) compares the non-hotspot, hotspot, and cluster regions of the human reference genome hg19. e Variations of recombination intensity, positive selection (in terms of average |nSL|), and purifying selection (in terms of phyloP score) in 1-kb window groups with low (3.1% of total, blue region), medium (92.8%, green region) and high (4.1%, pink region) SNP densities. f Variation of recombination intensity (orange curve) and positive selection in terms of DAF (red curve) among ten groups of SNP hotspots ranked by SNP density from low (rank 10) to high (rank 1). g Distribution of the low-, medium-, and high-density windows from part (e) in the different DNA replication phases. Error bars in (a, c, d) and shaded bands around the curves in (f) indicate 95% confidence intervals