Figure 1.
Signal of Positive Selection around the SCL24A5 Gene in Europeans from the HapMap Data
(A) The EHH decay plot for i, a neutral site outside the selective sweep; and ii, a site in the middle of the SCL24A5 gene, which shows strong evidence of positive selection in Europeans (Eur). Afr, Africans; Chn, Chinese.
(B) iES plotted against physical distance, back to back for Europeans and Africans.
(C) Signals of positive selection shown in the plot of Rsb (European/African) against physical distance.
Figure 2.
Comparisons of iES and Rsb Distributions between the Perlegen and HapMap Datasets
The iES and Rsb values are plotted for a 20-Mb region on Chromosome 15.
(A and B) iES against physical distance in Europeans (Eur) and Africans (Afr), respectively, back to back for the HapMap and Perlegen data.
(C) Plot of Rsb of Europeans over Africans against physical distance, back to back for the HapMap and Perlegen data.
Figure 3.
Comparisons of the ln(Rsb) Distributions between Various Neutral Simulations and between the Full Neutral Model and Empirical Data
(A) shows the full neutral model (model 10) and several other neutral models; (B) shows model 10 and empirical data. The ln(Rsb) distributions are compared by both a QQ plot and superimposing standardized histograms. In every QQ plot, quantile points of model 10 are plotted on the x-axis. For all of the superimposed standardized histogram plots, the blue color designates the full model (model 10) and the red color designates the alternative model.
Figure 4.
KSD Statistic for the Comparisons of Standardized ln(Rsb) Distributions within Neutral Models, and between Neutral Models and Empirical Data
Categories on the x-axis denote the neutral models 1 to 10-asc3, where 10-asc1, 10-asc2, and 10-asc3 indicate the three different ascertainment schemes (see Methods). Each neutral distribution is compared to the corresponding empirical (in blue) or model 10 (in yellow) distribution. A cluster of three bars in the same color denotes the KSD values in the three pairwise comparisons in the order AE, AC, and CE.
Table 1.
Some Examples of Candidate Regions with Strong Signals
Figure 5.
Power of the Two ln(Rsb)-Based Tests for Different Values of f (the Strength of the Bottleneck Experienced by the Derived Population)
(A) The power plot for test I, which uses ratio of tSNPs as the test statistic (see Methods). The critical value is set by independent neutral simulation at a level of 0.01. Curves in red show the power to detect selection that occurs in the central 1 Mb. Curves in blue show the power to detect selection that occurs outside the central 1 Mb, but within 200 kb. Dashed curves in green are the overall power for the entire simulated 2-Mb segment.
(B) The power plot for test II, the ad hoc test. A case is called significant only if the selected site lies within a candidate region.
Table 2.
Pairwise Overlapping Test of Different Candidate Region Lists of Positive Selection
Figure 6.
Comparisons of the Major Allele Spectra between the iHS and Rsb Candidate Regions, and the Genetic Diversity Pattern in the SLC24A5 Region
(A and B) The frequency spectra of the major alleles for SNPs are compared between the iHS candidate regions (in blue) and the 100-kb regions around the centers of the Rsb candidate regions (in orange) in either (A) Europeans or (B) Chinese. Overlapping parts are in dark purple. For comparison, the corresponding major allele–frequency spectra for the whole genome are shown as black curves.
(C) shows, as an example, the genetic diversity pattern for the 200-kb region around the SLC24A5 gene in Africans (Afr; upper section) and in Europeans (Eur; lower section), from the HapMap phase I data. Rows denote individual chromosomes, and columns denote the SNP sites.
Table 3.
Summary of the Gene Ontology Analysis Results