In this paper, I discuss some examples revealing that genetic diversity of humans (Homo sapiens) has been affected by non-biological, i.e., social and cultural, factors that involve demographic events caused by subsistent styles in both male and female lineages, social structure, dietary culture, and so on. Population genetics considers a “population” as a gene pool: evolution is a fluctuation of allele (types of genes) frequency in the gene pool, which is caused by mutation, natural selection, migration (gene flow), and random genetic drift1. Though mutation is purely a cellular biological process, the other causes are associated with the history of demography in the population. This is the main reason why social and cultural factors in human activities are reflected in human genetic diversity, because these factors are associated with demography in human populations.
How is demography associated with genetic diversity? Random genetic drift is a fluctuation of allele frequency by chance, whereas natural selection is an inevitable change of allele frequency. The effect of migration in the allele frequency change can be a spatial aspect of random genetic drift. An effective population size (Ne) is defined as a population size that relates to reproduction. When an Ne is finite, the contribution of random genetic drift becomes larger than that in the case of natural selection. When an Ne is infinite, then the contribution of natural selection in allele frequency changes becomes larger, i.e., more diverse than that in a finite Ne. Because the Ne of humans is quite small, the allele frequencies of polymorphic sites in human genomes have been influenced mainly by random genetic drift and migration causing admixture between populations. When natural selection rarely occurs, if an allele has the advantage to survive, the genomic region including that allele increases in a population; and when that occurs faster than the shuffling by recombination, the genomic region including the advantageous allele becomes homogeneous in the population, which is called, a “selective sweep”2.
Thus, genetic diversity is a result of random genetic drift, natural selection, and gene flow. Therefore, it is nonsense that only genetic factors play a role in everything that supports genetic determinism.
Recently, human genome data have been exponentially increasing. Because the next generation sequencer was generalized, kits with reagents have become less expensive, and human genome sequencing has become faster. At the outset, the human genome project took approximately 15 years; but now, human genome sequencing can be done in approximately 1 week. Therefore, “population genetics” gave birth to the field of “population genomics.” In social practice, commercially direct-to-consumer (DTC) genetic testing became popular because of the availability of low-cost genome analysis. I am afraid that this prevalence is becoming a trigger of new genetic determinism, which often leads to discrimination and social exclusion. I think it is extremely dangerous and notice that is not well known that human genetic diversity has been affected by social and cultural factors such as life style and dietary behavior.
In this paper, however, I will not mention studies mainly based on the whole genome data. I will, however, intentionally introduce studies of haploid loci, like mitochondrial DNA (mtDNA) and the Y chromosome, or a single locus (or a particular genomic region) of diploid (autosomal) loci, because the inheritance system of mtDNA and the Y chromosome are much simpler than that of the autosomal genome. The data of mtDNA and the Y chromosome intelligibly fit the theory, as I briefly explained above, so that readers would be expected to easily understand the examples suggesting that human genetic diversity is affected by social and cultural factors. For examples of these, there are other papers reporting similar observations. However, to simplify this particular paper, I will only refer to representative papers, mainly those involving our own studies.
1. Within-population genetic diversity and subsistent styles
Genetic diversity within a population is affected by demography as a result of subsistent styles in humans. Pairwise mismatch distribution (PMD) is shown in the graph in Figure 1 as the numbers of differences of bases in the horizontal axis, and the ratios of the people who have the pairwise difference in the vertical axis. In a PMD analysis on a haploid, if a population has experienced demographic expansion (i.e., rapid increase in population size), it is theoretically presumed that the distribution shows a bell-shaped pattern (i.e., a normal distribution). Meanwhile, if a population has a constant population size (i.e., without past demographic expansion), the distribution should show a non-bell-shaped pattern
Figure 1 : Pairwise mismatch distribution (PMD).
The horizontal axis represents the numbers of differences of bases, and the vertical axis represents the ratios of the people who have the pairwise difference.
Such a relationship between subsistent styles and demographic history based on genetic data was initially described in African populations. Watson et al. (1996)3 examined the PMDs of nucleotide sequences of mtDNA and compared them between those of hunter-gatherers and food producers. In the groups who have hunting-gathering traditions in their ethnography, the distributions showed non-bell-shaped patterns, indicating the groups had not experienced rapid demographic growth. Contrarily, for the groups who were food producers, the distributions showed bell-shaped patterns, suggesting past demographic expansions. Thus, subsistent styles are reflected in their demography and result in differences of within-population genetic diversity in human populations.
The PMD patterns are, however, not always fixed in hunter-gatherers and food producers. Using the mtDNA data published in a study by Oota et al. (2001)4, I investigated the PMDs of four slash-and-burn populations who lived in the hilly jungle of northern Thailand (Fig. 1). The Akha people showed a relative bell-shaped distribution, while the Black Lahu, Red Karen, and White Karen people showed collapsed bell-shaped distribution patterns, suggesting their constant population sizes. In the definition of subsistent styles, people who live by slash-and-burn agriculture are classified as food producers. However, the demographic patterns of the slash-and-burn populations were close to those of hunter-gatherer populations. Therefore, a simple typological definition of subsistent styles does not indicate an adequate estimation of demographic history.
2. Patrilocality v. matrilocality in migration patterns and genetic diversity
Mating systems, marital systems in humans, also influence genetic diversity, particularly in mtDNA and the Y chromosome, because mtDNA and the Y chromosome show maternal and paternal inheritance patterns. This was initially described by Seielstad et al. (1998)5, who looked into correlations between genetic distances (between-population diversity) based on mtDNA, and the Y chromosome, and geographic distances in Europe (Fig. 2).
Figure 2 : Correlations between genetic distances (Fst values) based on mtDNA or Y chromosome data and geographic distances (km) in Europe.
Circles represent mtDNA and squares represent the Y chromosome.
They found larger genetic distances based on the Y chromosome between populations far from each other than those close to one another, but the differences were much smaller in genetic distances based on mtDNA. The average of genetic distances based on the Y chromosome was much larger than those based on mtDNA. It is likely that males have not moved as frequently as have females between populations; as a result, frequencies of Y chromosome types in the populations have been changed by random genetic drift, while frequencies of mtDNA types have been homogenized by frequent migrations between populations. Seielstad et al. proposed a hypothesis that a patrilocal marital-system is reflected in mtDNA and Y chromosome diversity in Europe.
To test this hypothesis, Oota et al. (2001) focused on minority groups in northern Thailand because there were both groups who were classified into patrilocal and matrilocal residential patterns in a relatively small region (Fig. 3a).
Figure 3a : Map of northern Thailand and minority groups.
They looked into three matrilocal and three patrilocal groups, using mtDNA and the Y chromosome. The results were clear. When examining within-population genetic diversity, mtDNA diversity was higher in patrilocal groups than that in matrilocal groups, whereas Y chromosome diversity was higher in matrilocal groups than that in patrilocal groups (Fig. 3b).
Figure 3b : Within-population genetic diversity based on mtDNA or Y chromosome data in matrilocal and patrilocal groups.
Figure 3c : Between-population genetic diversity (genetic distances) based on mtDNA or Y chromosome data in matrilocal and patrilocal groups.
Examining between-population genetic diversity, the average of pairwise genetic distances based on Y chromosome haplotypes was higher in patrilocal groups than that in matrilocal groups, whereas the average of pairwise genetic distances based on mtDNA nucleotide sequences was higher in matrilocal groups than that in patrilocal groups (Fig. 3c). That is, the matrilocal and patrilocal residential patterns related to marital-systems were reflected in mtDNA and Y chromosome diversity. These results strongly supported the hypothesis of Seielstad et al. (1998).
Oota et al. (2002)6 investigated mtDNA and the Y chromosome in entire East Asian populations. In mtDNA, pairwise genetic distances between populations were significantly correlated to geographic distances between the populations (Fig. 4).
Figure 4 : Correlations between genetic distances (Fst values) based on mtDNA or Y chromosome data and geographic distances (km) in East Asia.
Circles represent mtDNA, and squares represent the Y chromosome.
Meanwhile, in the Y chromosome, pairwise genetic distances between populations were not correlated to geographic distances between the populations (Fig. 4). This pattern seems to be different from that in Europe. However, the average pairwise genetic distance based on the Y chromosome was approximately 2-fold higher than that based on mtDNA, suggesting the pattern in East Asia was essentially the same as that in Europe. This would explain the results of stronger patriarchal societies among East Asian populations than those among European populations. Therefore, the phenomena of “sex-biased migration/admixtures” have been designated and would be plausibly related to a patriarchal social system.
3. The debate over Genghis Khan’s Y chromosome
Another study, however, showed another interpretation for Y chromosomal characteristics in East Asia that higher genetic distances between populations and no correlation between genetic distances and geographic distances were due to polygamy, which is related to much stronger patriarchal societies. Zerjal et al. (2003) examined the Y chromosome in human populations widely from the eastern part of the Eurasian continent and found that a particular haplotype, which is the center of a star-like structure, showed high frequencies in East and Central Asia. That star-like structure in the phylogenetic network theoretically means the same as the bell-shaped pattern in a PMD, representing a demographic expansion. The demographic expansion was estimated to 1,000 – 2,000 years ago, corresponding to the timing of foundation of the Mongolian empire7. In addition, the Y haplotype was found in the geographical region the Mongolian empire occupied. Therefore, Zerjal et al. interpreted that this observation could be explained by what particular male lineage successfully expanded into vast geographical regions, suggesting that the lineage is plausibly linked to the family of Genghis Khan. The patterns observed in the Y chromosome were similar to selective sweep, so that is typically observed in a genome region where positive natural selection has been taking place. But there is no evidence of biological advantage in the Y haplotype. Only the reproductive success of the Y lineage was observed, indicating a particular marital system, probably polygamy. Zerjal et al. (2003) called the phenomena, “social selection,” instead of “biological selection.”
However, Ilumae et al. (2016)8 examined whole Y chromosome sequences in populations of the eastern Eurasian continent and estimated the timing of the expansion of the Y haplotype that Zerjal et al. (2003) 7 thought as Genghis Khan’s Y chromosome to be older than 5,000 years, suggesting no relationship between the demographic expansion and the Mongolian empire. Therefore, it is still controversial why higher genetic distances are observed in the Y chromosome than in mtDNA and no correlations are observed between genetic distances and geographic distances in the Y chromosome in East Asia. It would be certain, however, that the difference observed between Europe and East Asia is explained by population migration and demographic events, which are related to the transition of subsistent styles.
4. Agropastoralism and effective population size
Recent studies involving larger data sets showed sex-biased demographic patterns based on mtDNA and Y chromosomal diversities in worldwide populations. Karmin et al. (2015)9 determined whole Y chromosome sequences in the global populations and compared them with mtDNA sequences from global populations. They found a recent bottleneck, a reduction of the Ne, occurred in West Asia and Europe, after the Out-of-Africa bottleneck. The timing of that reduction was estimated to have occurred around the period of the Neolithic revolution, which interestingly corresponds to the transition from hunting-gathering to food producing societies. The Ne of mtDNA has continued to increase from 50,000 years ago and still continues to increase even to this day, while the Ne of the Y chromosome stopped increasing approximately 10,000 years ago and had remarkably decreased once some 8,000 to 5,000 years ago.
Because the extreme decrease of Ne had occurred earlier in West Asia than in Europe, and the timings approximately corresponds to the period of the Neolithic revolution in West Asia and Europe, respectively, Karmin et al. (2015) explain that social and cultural changes accompanying practices of agriculture and pastoralism (agropastoralism) had promoted a disproportion of males’ reproduction chances in those times 9. This might have occurred because agropastoralism made a difference between rich and poor people. Rich males reproductively succeeded more often than did poor males, which was reflected in the Ne of the entire society; while no deviation was evidenced in the reproductive success of females, which led to no change occurring in the Ne of the entire society (Fig. 5).
Figure 5 : Deviation of reproductive success in males but not in females.
The process of the development of agropastoralism would have been different between the western and eastern parts of the Eurasian continent. Local regions would have also had different processes. Some regions might have accepted agropastoralism very easily, whereas other regions might have not accepted it until more recently. Such differences produced different migration patterns and demographic events, which in turn had an influence on characteristics of the genome structure in individual populations.
5. Pastoralism and agriculture influence on gene evolution
Development of agropastoralism also reflects on the evolution in genomic regions (which also reflects that of genes as well) related to phenotypic differences among human populations. A well-known example would be the mutation in the lactase enzyme that cancels lactose intolerance. Babies grow up with mothers’ milk that includes lactose. Babies have an enzyme that digests lactose into glucose and galactose. This enzyme, lactase, is expressed mainly in the small intestines of babies during the lactation period and stops at delactation. That is the reason some adults cannot digest lactose, and are lactose intolerant. However, a recent genome-wide investigation found there are mutants who do not stop expression of lactase after delactation10. This is due to a mutation that occurred in the promotor region of LCT, the lactase gene, which evades the stopping expression of lactase. The mutation shows signals of selective sweep (positive selection) on the chromosomal region. The frequencies of the mutation are high in some European and African populations, both of which traditionally do dairy farming. This is widely considered as a typical example that dietary culture (in this case, one in which the people thereof often drink milk from domestic animals) influenced the genome evolution in humans.
Another example concerns mutations of genes that relate to alcohol digestion. When drinking alcohol, in the first step, ethanol is digested into acetaldehyde; and, in the second step, acetaldehyde is digested into acetic acid in the liver. The enzyme of the first step is alcohol dehydrogenase (ADH), and that of the second step is aldehyde dehydrogenase (ALDH). In the human genome, there are 7 ADH genes and 16 ALDH genes. Three ADH genes, and the ALDH2 gene are mainly expressed in the liver when humans’ drink alcohol. One of the ADH genes (ADH1B) has genetic polymorphism: the single nucleotide polymorphism (SNP) changes arginine (Arg) into histidine (His) at the 48th amino acid position, which have low and high activities, respectively11. The ALDH2 gene also has genetic polymorphism: the SNP changes glutamic acid (Glu) into lysine (Lys) at the 504th amino acid position, which have enzymatic activity and deficiency, respectively. People who have an ALDH2 deficiency exhibit phenotypic characteristics: flushing, headache, and nausea after drinking too much alcohol. If a person has a high-activity allele (His) in ADH1B, then ethanol is rapidly metabolized by oxidization into acetaldehyde; and if a person also has the deficiency allele (Lys) in ALDH2, then acetaldehyde is not oxidized to acetic acid but accumulated.
Acetaldehyde is toxic for humans. Nevertheless, a high-activity allele of ADH1B has a high frequency in West and East Asia12, and to date the deficiency allele of ALDH2 has been found only in East Asia13. Oota et al. (2004) found signals implying positive selection on the deficiency allele of ALDH214, and Han et al. (2007) found strong evidence showing selective sweep on the ADH gene cluster15.
The results of population genetic studies on the ADH gene cluster and the ADLH2 gene suggest that weakness against alcohol intake is advantageous in natural selection. This strange advantage can be explained by the toxic effect of acetaldehyde. A high concentration of acetaldehyde in blood is toxic not only for humans but also for pathogens, especially for blood parasites. Namely, alcohol intake might be related to prevention against infectious disease, which might have been contributing to selective pressure concerning these genes involved in the alcohol metabolic system of humans. But why did it work only in East Asian populations? We do not know the exact reason yet. A hint is that the geographical region with high allele frequencies in both the high-activity allele of ADH1B and the deficiency allele of ALDH2 corresponds to the area where rice cultivation began in East Asia and where many infectious cases of blood parasites, malaria, amebiasis, and so on, have been reported. Large-scale rice fields are reclaimed by the changes in nature, which could be a hospitable environment for such pathogens. Thus, environmental adaptation sometimes leaves a trace in the human genome. For humans, culture is also part of their environment. Therefore, cultural adaptation also leaves a trace in the human genome.
6. Spatial and temporal changes of population genetic structure
Subsistent styles influence demography, which leaves a trace in the human genome. Such demographic changes can be traced by looking at genome data from modern human populations. Ancient DNA analyses, however, tell us more details of changes in population genetic structure owing to demography as well as to migration events.
The Jomon and Yayoi were prehistoric cultures in the Japanese archipelago, and roughly correspond to hunting-gathering and agricultural societies, respectively. In anthropological studies, the Jomon people have been considered as indigenous hunter-gatherers since approximately 16,000 years ago, whereas the Yayoi people were farmers who migrated from the eastern part of the Eurasian continent16. Because the dating of carbonized rice on earthenware excavated from the Kyushu island was estimated to be from 3,000 years ago, the Yayoi period is thought to have begun some 3,000 years ago, and the arrival place of the migrants who brought the technology of rice cultivation is thought to be the northern part of Kyushu17.
Oota et al. (1995)18 examined mtDNA of human remains excavated from an archaeological site located in northern Kyushu. Two burial styles coexisted in that site from the Yayoi period. One was “dokobo,” a typical style used in the Jomon period that buried human remains directly in the earth, and the other was “kamekan,” an earthenware burial jar, i.e., a jar-coffin, used particularly in the Yayoi period. The results of ancient DNA analysis showed that mtDNA of individuals buried in “kamekan” had lower diversity in nucleotide sequences that did those buried in the “dokobo” style and occupied a major sequence type (Fig. 6).
Figure 6 : Phylogenetic network based on mtDNA nucleotide sequences from 26 human skeletal remains from the Takuta-Nishibun shell-mound site in the northern part of Kyushu, Japan.
Circles represent sequence type of mtDNA. The numbers on the branches are the nucleotide positions of the mtDNA D-loop region where the nucleotide positions changed. The size of the circles is proportional to the number of the individuals who have the sequence types. The ID numbers with K represent the individuals excavated from “kamekan,” while those with D represent the individuals excavated from “dokobo.”
Archaeological evidence showed there were significant time differences between “dokobo” and “kamekan” in this site: “dokobo” were older than “kamekan.” Therefore, Oota et al. (1995) concluded that the difference of mtDNA diversity between the two burial styles could be explained by population replacement that occurred at this site. It is likely that the individuals buried in the “dokobo” style were indigenous people from the Jomon lineage, whereas the individuals buried in “kamekan” were plausibly migrants from the eastern part of the Eurasian continent and perhaps had consanguineal kinship ties with each other in the maternal lineage.
Wang et al. (2000) examined mtDNA of human remains from two periods (2,500 and 2,000 years ago) excavated from an archaeological site in Linzi, China, and compared mtDNA of present-day people in the area of the dig19. Because ancient DNA was damaged and fragmented, the mtDNA sequence length that could be examined was short, and only five major sequence types were found. However, the frequencies of mtDNA sequences were very different from each other in those three periods. Therefore, the current genetic structure of a population is not the same as that of a past population even in the same geographical area, because genetic drift and migration easily change the frequency patterns of the sequence types in a short time.
More recently, genome sequencing has often been done for many ancient samples and has revealed population replacements. McColl et al. (2018) examined ancient nuclear genomes from Southeast Asians and the Japanese Ikawazu Jomon and reported the peopling history of Southeast Asia20. In the pre-Neolithic period, the Hòabìnhian hunter-gatherers occupied Southeast Asia until agriculture developed and expanded. In the “one-layer” hypothesis, the Hòabìnhian hunter-gatherers adopted agriculture and became farmers without substantial external gene flow. In the “two-layer” hypothesis, the farmers migrated from the North and replaced the Hòabìnhian hunter-gatherers approximately 4,000 years ago. Ancient genome results showed that the 8,000 – 4,000-year-old specimens (the oldest was of the Hòabìnhian culture) were closely related to present-day hunter-gatherers in Southeast Asia, while the 4,000 – 500-year-old specimens were so different from pre-Neolithic/present-day hunter-gatherers and close to present-day East Asians (Austroasians, Tai-Kadai, Austronesians, Southern Chinese, and so on). These suggest replacement from pre-Neolithic hunter-gatherers to Neolithic farmers.
Meanwhile, the ancient genome analyses also found a trace of admixture between pre-Neolithic hunter-gatherers and Neolithic farmers, after the divergence between pre-Neolithic hunter-gatherers and ancient East Asians. The 2,500-year-old Ikawazu Jomon from Japan had much higher affinity with the Hòabìnhian hunter-gatherers than to the present-day East Asians. Thus, the ancient genome data supports the “multi-layer” hypothesis, which is much more complicated than the “one-layer” or “two-layer” hypotheses.
7. Conclusions
In this paper, I have introduced examples that are related to non-biological factors and gene frequencies (allele frequencies) in humans. Though genetic determinism is sometimes considered a principle of biology, that is not always true: the examples mentioned here suggest that human genetic diversity has been considerably affected by non-biological factors, such as subsistent styles, marital systems, social selection, and natural selection. This might be the opposite of an ordinary idea but should be understood by people affected by non-biological factors. As I have mentioned, population geneticists start their studies by regarding a “population” as a gene pool. A “population” is not a priori but a posteriori that population geneticists define in a study, so that a “human population” is always tentative.
If a certain “race” exists as a subspecies among Homo sapiens, we would find a substantial genetic distance between various human populations. However, we do NOT find one. Within-population diversity in a human population is always fluctuated by random genetic drift, natural selection, and migration events. Therefore, between-population diversity, which is the same as genetic distance between populations, is also changed temporally. These are fundamental reasons why human population geneticists no longer use the word “race”, and no “race” is a subspecies in humans.
I would emphasize that researchers become aware of the “way of thinking” about population genetics, as I have mentioned, if you want to understand human genome data. In this context, most of “ancestry discovery” provided by DTC genetic testing is misleading, because there is no concept involving thinking way of population genetics. Some DTC companies provide DNA testing for ancestry discovery using mtDNA haplogroups. But, mtDNA haplogroups only give us a 1/2n ancestor theoretically from n generation(s) in the past. For example, if you want to discover your ancestors from 3,000 years ago, that is 100 generations ago if one generation is assumed to be 30 years; therefore, you theoretically have 2100 ancestors. The DNA ancestry test using mtDNA, however, only tells you one of 2100 ancestors.
The DNA ancestry test gives you information about a population that has a haplogroup of mtDNA at a high frequency, but the population is a present-day population because the frequency data of mtDNA haplogroups are generated from present-day populations. You know where the present-day population is but do not know where the population was. As I have mentioned above, there have been many migration events in human history. Only thing that the mtDNA haplogroup data tell you is that you and the population have a common ancestor somewhere in the past.
This issue is the same in the DNA ancestry test using autosomal genome data. Some DNA ancestry tests give you percentages of “your ancestral populations.” But these are also based on genome data from present-day populations and can be carried out under the assumption that no population migrated. For example, the DNA test might reveal that you are 50% European, 30% Chinese, and 20% Indian. But these are present-day “Europeans,” “Chinese,” and “Indians.” Genetic components of these populations have always been changed by random genetic drift, migration events, and natural selection. You cannot say who are standard Europeans, Chinese, and Indians. Therefore, the percentages in DNA ancestry tests are quite misleading, even if those are based on huge data from autosomal DNA.
Human populations certainly have genetic differences from each other because of random genetic drift, migration events, and natural selection, but the genetic differences between populations are always very small and fluctuating. Among humans, the average difference between genome is less than 0.2%. Without the correct understanding of human population genetics, the current huge genome data of human populations lead to general misunderstandings. It is imperative to understand population genetics, to avoid the rising tide of racism.
Notes
1
D. L. Hartl, A Primer of Population Genetics, 3rd ed., Sunderland, Sinauer, 2000.
2
D. L. Hartl, A. G. Clark, Principles of Population Genetics, 4th ed., Sunderland, Sinauer, 2007.
3
E. Watson et al., « mtDNA sequence diversity in Africa », Am J Hum Genet., 59(2), 1996, p. 437-44.
4
H. Oota et al., « Human mtDNA and Y-chromosome variation is correlated with matrilocal versus patrilocal residence », Nat Genet., 29(1), 2001, p. 20-1.
5
M. T. Seielstad, E. Minch, L. L. Cavalli-Sforza, « Genetic evidence for a higher female migration rate in humans », Nat Genet., 20(3), 1998, p. 278-80.
6
H. Oota et al., « Extreme mtDNA homogeneity in continental Asian populations », Am J Phys Anthropol., 118(2), 2002, p. 146-53.
7
T. Zerjal et al., « The Genetic Legacy of the Mongols », Am J Hum Genet., 72(3), 2003, p. 717-21.
8
A. M. Ilumäe et al., « Human Y Chromosome Haplogroup N : A Non-trivial Time-Resolved Phylogeography that Cuts across Language Families », Am J Hum Genet., 99(1), 2016, p. 163-73.
9
M. Karmin et al., « A Recent bottleneck of Y chromosome diversity coincides with a global change in culture », Genome Res., 25(4), 2015, p. 459-66.
10
S. A. Tishkoff et al., « Convergent adaptation of human lactase persistence in Africa and Europe », Nat Genet., 39(1), 2007, p. 31-40.
11
M. V. Osier et al., « A global perspective on genetic variation at the ADH genes reveals unusual patterns of linkage disequilibrium and diversity », Ann Hum Genet., 71(1), 2002, p. 84-99.
12
H. Oota et al., « The evolution and population genetics of the ALDH2 locus: random genetic drift, selection, and low levels of recombination », Ann Hum Genet., 68(Pt 2), 2004, p. 93-109.
13
H. Li et al., « Ethnic related selection for an ADH Class I variant within East Asia », PLoS One, 3(4), 2008, e1881.
14
H. Li et al., « Refined geographic distribution of the oriental ALDH2*504Lys (nee 487Lys) variant », Ann Hum Genet., 73(Pt 3), 2009, p. 335-45.
15
Y. Han et al., « Evidence of positive selection on a class I ADH locus, Am J Hum Genet., 80(3), 2007, p. 441-56.
16
K. Imamura, Prehistoric Japan : new perspective on insular East Asia, London, UCL Press Ltd, 1996.
17
J. Habu, Ancient Jomon of Japan, Berkeley, Cambridge University Press, 1996.
18
H. Oota et al., « A genetic study of 2,000-year-old human remains from Japan using mitochondrial DNA sequences », Am J Phys Anthropol, 98(2), 1995, p. 133-45.
19
L. Wang et al., « Genetic structure of a 2,500-year-old human population in China and its spatiotemporal changes », Mol Biol Evol., 17(9), 2000, p. 1396-400.
20
H. McColl et al., « The prehistoric peopling of Southeast Asia », Science, 361(6397), 2000, p. 88-92.