Evidence of still-ongoing convergence evolution of the lactase persistence T-13910 alleles in humans

Abstract

A single-nucleotide variant, C/T(-13910), located 14 kb upstream of the lactase gene (LCT), has been shown to be completely correlated with lactase persistence (LP) in northern Europeans. Here, we analyzed the background of the alleles carrying the critical variant in 1,611 DNA samples from 37 populations. Our data show that the T(-13910) variant is found on two different, highly divergent haplotype backgrounds in the global populations. The first is the most common LP haplotype (LP H98) present in all populations analyzed, whereas the others (LP H8-H12), which originate from the same ancestral allelic haplotype, are found in geographically restricted populations living west of the Urals and north of the Caucasus. The global distribution pattern of LP T(-13910) H98 supports the Caucasian origin of this allele. Age estimates based on different mathematical models show that the common LP T(-13910) H98 allele (approximately 5,000-12,000 years old) is relatively older than the other geographically restricted LP alleles (approximately 1,400-3,000 years old). Our data about global allelic haplotypes of the lactose-tolerance variant imply that the T(-13910) allele has been independently introduced more than once and that there is a still-ongoing process of convergent evolution of the LP alleles in humans.

Figure 1.

The physical map showing the analyzed genome region flanking the C/T_-13910 and G/A_-22018 variants associated with LP. The distance (in kb) from the first ATG of LCT is shown. A, Genes in the region studied. B, Expanded map of the 30-kb region in the LCT and MCM6 genes, showing SNPs 1–9 analyzed in the population samples.

Figure 2.

Correlation between the frequency of the LP trait, as measured by lactose-tolerance tests and/or disaccharidase activities, and the frequency of LP, as predicted by the frequency of the C/T_-13910 allele with the assumption of Hardy-Weinberg equilibrium in different populations. A perfect correlation is not expected because the phenotype data are collected separately from the genotype data. Data about the frequency of LP were obtained from or referenced in studies reported elsewhere.^,,– Populations are coded as indicated in table 3, except for “Fin,” which includes both FiE and FiW as one group. The r coefficient was calculated using SPSS version 10.0. The two-sided test was performed using a .01 significance level.

Figure 3.

Population frequencies for the T_-13910 allele associated with LP in worldwide populations. For each population, the pie chart denotes the frequency of the T_-13910 allele (green shading). Populations and frequency details are shown in table 3.

Figure 4.

MJ haplotype network for eight SNPs and one indel marker in the 30-kb LCT region among 37 populations, constructed using NETWORK version 4.1.1.2. The analysis includes all haplotypes with an estimated population frequency >4% in at least one population. The arrow denotes the rooted haplotype LNP H1 of the network. LNP haplotypes are shown as yellow circles, and LP haplotypes are shown as green circles. The size of the circles corresponds to the estimated haplotype frequency in the global sample, and the haplotypes are shown inside or above the circles. The population frequencies for some haplotypes relevant to the origin of the LP alleles and discussed in the text are shown. The positions of the C/T_-13910 and G/A_-22018 alleles within the haplotype are shown in the enclosed box. The SNPs have been coded for each site as 1 for the ancestral SNP and 2 for the derived SNP. The haplotype details are shown in tables 2 and 5.