Neanderthal adaptive introgression shaped LCT enhancer region diversity without linking to lactase persistence in East Asian populations

Abstract

Positive selection at the 2q21.3 enhancer region for lactase gene (LCT) expression in Europeans and Africans has long been attributed to selection for lactase persistence (LP), the capacity of adults to digest lactose in milk, presumably because of the benefits associated with milk consumption. While considered a classic example of gene-culture coevolution, recently doubts have been raised about the link between selection at 2q21.3 and LP. Analysis of additional populations could shed further light; here, we demonstrate that a haplotype spanning ~467 kb at the 2q21.3 locus has risen to high frequency in East Asians (~25%) but is absent from Africans and Europeans. This haplotype likely derived from Neanderthals and has been under positive selection in East Asians. The East Asian-specific haplotype is associated with alterations in LCT expression and promoter methylation in certain cell types, similar to what is observed with LP-associated haplotypes in Europeans. Moreover, its frequency is comparable to that of LP in East Asians, suggesting a potential association with LP in East Asians. However, it is highly unlikely that selection in East Asians was related to milk-drinking habits. We find that this haplotype impacts the expression of UBXN4, DARS1, and DARS1-AS1 in immune cells and is associated with neutrophil and white blood cell counts. Hence, the selection might be linked to certain aspects of immune function. This implies that selection on 2q21.3 has thus either occurred for different reasons in different populations or the selection observed in other populations is also not due to LP.

Keywords: East Asian; LCT; Neanderthal; adaptive introgression; lactase persistence.

Fig. 1.

Genetic differentiation between East Asians and Europeans at the 2q21.3 locus. (A) Genetic differentiation between East Asian and European populations. The dashed line indicates F_ST = 0.2. The Bottom panel shows the protein-coding and micro-RNA genes in this region. The x axis indicates the GRCh37 coordinate. (B) The genetic differentiation in European populations at the LCT region. The y axis represents the locus-specific branch length of each locus for European populations. (C) The genetic differentiation in East Asian populations at the LCT region. The y axis represents the locus-specific branch length of each locus for East Asian populations. Each point represents one variant. The blue and red colors indicate variants that have undergone large allele frequency changes in European and East Asian populations (as determined by the locus-specific branch lengths analysis), respectively. (D) The allele frequency distribution of rs4988235 in worldwide populations. (E) The allele frequency distribution of rs4988245 in worldwide populations.

Fig. 2.

Neanderthal origin of the East Asian–specific haplotype. (A) Allelic states of haplotypes from East Asian (represented by CHB), European (represented by CEU), and African (represented by YRI) populations, together with haplotypes from Altai Neanderthal and Altai Denisovan. These representative haplotypes at the 2q21.3 locus are defined by highly differentiated SNPs, which are at high frequency in European or East Asian populations and carry an allele absent in African populations. The Top panel shows variants that were used to define haplotypes. The next panel shows the allele state for these haplotypes. (B) The network for haplotypes from present-day modern populations together with Altai Neanderthal, Altai Denisovan, and Chimpanzee. The haplotypes were defined by the highly differentiated SNPs. Each circle represents one unique haplotype, with the size of the circle proportional to the number of chromosomes with that haplotype, colored according to population. (C) The neighbor-joining tree for the inferred sequences of modern human haplotypes and the Neanderthal, Denisovan, and Chimpanzee genome sequences. Branches are colored according to population.

Fig. 3.

The signal of positive selection for the East Asian–specific haplotype. (A) The DIND test result for the 2q21.3 locus. Each point represents one SNP with the derived allele >0.1. The y axis indicates the empirical P-value of this SNP in the DIND test. The blue dashed line indicates the 5% empirical distribution threshold. The red color indicates SNPs with introgressed alleles. (B) The extended haplotype homozygosity (EHH) results for rs4988245 for all haplotypes from the CHB population. (C) The Tajima’s D result for the 2q21.3 locus. Each point represents one 50-kb window. For each window, we calculated Tajima’s D for all the haplotypes, the introgressed haplotypes, and the nonintrogressed haplotypes from the CHB population. (D) The estimate of the selection time for rs3087348 in East Asian populations. The selection time was estimated separately for each population (see Materials and Methods for details).

Fig. 4.

Allele frequency change over time. (A) The proportion of carriers of the East Asian–specific alleles at rs3087348 across time. The number above each bar represents the number of samples carrying the East Asian–specific allele and the total number of samples with known genotypes. (B) The allele frequency changes over time for rs3087348. The solid line indicates the allele frequency change inferred by RELATE and CLUES using all haplotypes from different East Asian populations. The filled region indicates the 95% CI.

Fig. 5.

Expression and methylation change for the LCT gene/region associated with the East Asian–specific haplotype. (A) The relationship between the expression of the LCT gene and the genotype at rs149881318 in pDCs in Japanese. (B) The relationship between the methylation level at cg20242066 at LCT promoter and the genotype at rs149881318 in whole blood in Han Chinese.