Ethnic related selection for an ADH Class I variant within East Asia

Abstract

Background: The alcohol dehydrogenases (ADH) are widely studied enzymes and the evolution of the mammalian gene cluster encoding these enzymes is also well studied. Previous studies have shown that the ADH1B*47His allele at one of the seven genes in humans is associated with a decrease in the risk of alcoholism and the core molecular region with this allele has been selected for in some East Asian populations. As the frequency of ADH1B*47His is highest in East Asia, and very low in most of the rest of the world, we have undertaken more detailed investigation in this geographic region.

Methodology/principal findings: Here we report new data on 30 SNPs in the ADH7 and Class I ADH region in samples of 24 populations from China and Laos. These populations cover a wide geographic region and diverse ethnicities. Combined with our previously published East Asian data for these SNPs in 8 populations, we have typed populations from all of the 6 major linguistic phyla (Altaic including Korean-Japanese and inland Altaic, Sino-Tibetan, Hmong-Mien, Austro-Asiatic, Daic, and Austronesian). The ADH1B genotyping data are strongly related to ethnicity. Only some eastern ethnic phyla or subphyla (Korean-Japanese, Han Chinese, Hmong-Mien, Daic, and Austronesian) have a high frequency of ADH1B*47His. ADH1B haplotype data clustered the populations into linguistic subphyla, and divided the subphyla into eastern and western parts. In the Hmong-Mien and Altaic populations, the extended haplotype homozygosity (EHH) and relative EHH (REHH) tests for the ADH1B core were consistent with selection for the haplotype with derived SNP alleles. In the other ethnic phyla, the core showed only a weak signal of selection at best.

Conclusions/significance: The selection distribution is more significantly correlated with the frequency of the derived ADH1B regulatory region polymorphism than the derived amino-acid altering allele ADH1B*47His. Thus, the real focus of selection may be the regulatory region. The obvious ethnicity-related distributions of ADH1B diversities suggest the existence of some culture-related selective forces that have acted on the ADH1B region.

Figure 1. Locations of the populations and distributions of the ADH1B rs1229984 and rs3811801 derived allele frequencies.

Note: The map of part A showed the ethnic phyla in East Asia, and part B displayed the locations of the populations. Populations marked with stars were cited from literature , . The codes of the star-marked populations are ISO639-3 codes. Populations shown by gray spots are previously published by our team . The colorful spots are the populations collected in this study. Part C is the distribution of the derived allele frequency of rs1229984 (ADH1B*47His). Part D is the distribution of the regulatory region polymorphism rs3811801 derived allele frequency.

Figure 2. Pattern of regions of high LD using HAPLOT and the default r² algorithm.

Note: The codes of the non-East Asian populations are shown as NAS(Nasioi), KTY(Khanty), KMZ(Komi), FIN(Finns), SAM(Samaritans), DRU(Druze), ASH(Ashkenazi Jews). CN: New collected samples from China. LA: Newly collected samples from Laos. Both dbSNP numbers and ALFRED UID numbers are presented for the SNPs in the ADH region we typed.

Figure 3. Haplotype frequencies of the ADH1B gene region including the regulatory region.

Note: the SNPs in the haplotypes are rs2066701-rs2075633-rs4147536-rs1229984-6810842-rs3811801, corresponding to SNPs 6–11 in Figure 2. Phyla, PN: Papuan-New Guinean, AU: Austronesian, TK: Daic, AA: Austro-Asiatic, TB: Tibeto-Burman, HM: Hmong-Mien, SN: Sinitic Han, KJ: Korean-Japanese, AT: Altaic (inland), UR: Uralic, AF: Afro-Asiatic. The patterns of the non-East Asian phyla (UR, AF, PN) are quite different from those of East Asian phyla. The patterns of the phyla in East Asia can be classified into four groups as the colors shown in the left side bar. The frequency data for all haplotypes are in Table S1.

Figure 4. Principal Component Analysis plots.

Notes: The plots show the relationships among populations estimated by PCA. Plots in part A were based on the ADH1B haplotypes frequency data in Figure 3. Plots in part B used the haplotype frequency data of the whole ADH region in Figure 2. In part A, an ethnic related distribution is obvious, while in part B the distribution shows no strong distinct clusters corresponding to ethnicity.

Figure 5. Frequencies and network of the core haplotypes (rs1229984-rs6810842-rs3811801).

Notes: Haplotype codes (1) to (8) are in the same system for both part A and part B of the figure. The sizes of the balls in the network of part B represent the rough relative frequencies of the haplotypes. The arrows are the most likely mutational relationships. The broken lines indicate possible historical recombinations. Haplotype (3) is more derived and presumably younger than haplotypes (1), (2), (4), and (6). Its high frequency in some populations suggests selection may have operated.

Figure 6. Extended Haplotype Homozygosity (EHH) and Relative Extended Haplotype Homozygosity(REHH) of Altaic, Han and Hmong populations.

Note: Colorful lines are data of core haplotype (3)AGA, and gray lines are data of other haplotypes. The data following the population codes are frequencies of the core haplotype in the populations. EHH and REHH of the other populations are in Figure S1.

Figure 7. REHH of observed and simulated populations.

Note: The colored dots are observed REHH data of core haplotype AGA both in chart A and C. In chart A, the observed REHH data shows that most of the REHH values of haplotype AGA are higher than those of the other haplotypes. Part B is the East Asian population history model determined by complicated factors. Six phases were defined with the effective population numbers and the generation numbers to present. Chart C indicates the REHH data simulated by the model in part B along with the data in Chart A. The lines in chart C are comparison borders of the simulated data. The observed REHH of haplotype AGA of all the Altaic and Hmong populations are above the 95% border, which is the evidence of positive selection.

Figure 8. Significance P values of the positive selection on the ADH1B gene in East Asia.

Note: The map of part A displayed the distribution of the significance P value of the positive selection on the ADH1B gene. Populations from most areas of East Asia have been significantly selected for except those in the southwest. In part B, the selection area was compared with the high frequency areas of two ADH1B SNPs. The dotted line encloses the region in which the ADH1B*47His frequency is >40%, and the broken line encloses the region in which the ADH1B rs3811801 frequency is >40%. The distributions of ADH1B*47His and selection area differ from each other, which indicates that they are unrelated. The high frequency area of rs3811801 derived allele is included in the selection area, indicating the possible effect of this allele in the selection on the ADH1B gene.