The Genome Speaks: Origins of the Japanese

In 1991, the physical anthropologist Kazuro Hanihara proposed what would become the long-standing orthodoxy on the origins of the Japanese population: the dual-structure model.

The model posits that modern Japanese populations arose from the admixture of two distinct groups: the Jōmon, who first inhabited the Japanese archipelago, and the Yayoi, continental migrants who arrived via the Korean Peninsula beginning around 900 BCE.

The Jōmon had cultivated a culture of their own for some sixteen thousand years, producing among the world’s oldest known pottery. Their physical traits—deep-set features, double eyelids, wet earwax—persist most visibly today among the Okinawan and Ainu peoples.

The Yayoi migrants, by contrast, introduced wet-rice agriculture and metalworking, fundamentally transforming the culture of the archipelago. Their characteristic traits—elongated faces, single eyelids, dry earwax—are found with particular frequency among modern inhabitants of the Kinki region.

In September 2021, an international research team led by Takashi Gakuhari of Kanazawa University published a paper in Science Advances.

Through genomic analysis of ancient skeletal remains, they demonstrated that the origins of Japanese populations follow not a dual but a tripartite structure.

The first layer is Jōmon: people who separated from a basal continental population some twenty to fifteen thousand years ago and settled across the archipelago. The second layer is Northeast Asian: migrants who arrived during the Yayoi period via the Korean Peninsula, bringing wet-rice agriculture. The third layer is East Asian: a population that arrived during the Kofun period (3rd to 7th century CE) and had been entirely overlooked until this study.

The decisive finding was that the genome of Kofun-period skeletal remains closely matched that of modern Japanese populations. The genetic foundation of the people living in Japan today was laid not during the Yayoi period but during the Kofun period.

This discovery lent genetic confirmation to the historical significance of the Kofun era—an age when colossal keyhole-shaped burial mounds appeared abruptly and the Yamato polity unified the archipelago.

In April 2024, RIKEN, a Japanese national research institute, published a study in Science Advances that validated the tripartite-structure model using large-scale data from living populations.

Whole-genome sequencing of 3,256 Japanese individuals revealed three distinct ancestral components, confirming the tripartite framework.

K1 (Okinawan lineage) retains the strongest Jōmon genetic signature and is highest in Okinawa Prefecture. K2 (Tōhoku lineage) lies closest to Northeast Asian ancestry, peaking in Akita and Aomori. K3 (Kansai lineage) lies closest to East Asian ancestry, reaching its highest proportion in Shiga Prefecture.

The three components vary sharply by prefecture. Okinawa shows a pronounced K1 dominance; Shiga leads in K3. Unexpectedly, the entire Shikoku region also registers a high K3 (continental-migrant) ratio.

When this prefecture-by-prefecture genetic map is overlaid with shrine distribution patterns, the correlations are strong enough to warrant a chapter of their own—taken up in the next chapter.

In October 2024, a team led by Professor Jun Ohashi of the University of Tokyo published an important challenge to the tripartite-structure model.

Their whole-genome analysis of Yayoi-period migrant remains from the Doigahama site in Yamaguchi Prefecture showed that those migrants already carried both Northeast Asian and East Asian genetic components.

The implications are considerable. Where the tripartite-structure model envisions Northeast Asians arriving in the Yayoi period and East Asians arriving separately in the Kofun period, this finding suggests that an already mixed population may have migrated as a single wave during the Yayoi era.

In other words, the three ancestral “waves” may not have been distinct arrivals at all: the migrants themselves may have been a genetically diverse, already-blended population from the start.

This debate remains unresolved. Ancient DNA from further skeletal remains will be the key to a definitive answer.

Professor Jun Ohashi’s team at the University of Tokyo calculated a “Jōmon ancestry index” for each prefecture—a measure of how much Jōmon genetic heritage persists among each region’s inhabitants.

The highest Jōmon ancestry belongs to Okinawa Prefecture, followed by the prefectures of the Tōhoku region. At the opposite end, the highest continental-migrant ratio (lowest Jōmon ancestry) is found in Shiga Prefecture, with the broader Kinki region showing strong continental influence throughout.

Two results run against expectation. Shikoku as a whole shows high continental ancestry. And Hokkaido, despite the presence of the Ainu people, registers values close to the Honshu average, likely a consequence of later Wajin (mainland Japanese) settlement.

Set against shrine distribution data, the same map yields a further pattern: shrines of the Hachiman lineage tend to concentrate in regions with high continental-migrant ratios, while Suwa-lineage shrines cluster in areas of high Jōmon ancestry. The next chapter takes up the correlation between genetics and faith in detail.

Extracting DNA from human remains thousands of years old is extremely difficult. Ancient DNA fragments over time and picks up contamination from microbial and soil DNA. What researchers work with is often a degraded sliver amounting to less than one percent of the full genome.

The breakthrough came with next-generation sequencing (NGS). Where the earlier Sanger method struggled with fragmented ancient DNA, NGS processes vast quantities of short fragments in parallel, assembling them like pieces of a puzzle. From the 2010s onward, this technical advance drove a rapid expansion in ancient genome research.

Japan presents difficulties of its own. The hot, humid climate severely degrades DNA preservation. High-quality genome data was recovered from approximately 3,800-year-old Jōmon remains at the Funadomari site on Rebun Island, Hokkaido, but that was an exceptional success owing to the cold climate. South of Hokkaido, extracting DNA from remains of comparable age is often impossible.

In recent years, however, a technique for efficiently extracting DNA from the petrous portion of the temporal bone—the very dense bone surrounding the inner ear—has been established, and success rates have risen sharply. The era when ancient genome research in Japan was deemed impossible is, at last, drawing to a close.

The first thing researchers noticed in the Jōmon genome was its isolation. The Jōmon are genetically distant from every modern population on Earth. They diverged from mainland East Asians more than twenty thousand years ago—one of the deepest splits among all living human groups.

Yet they were not entirely alone. A degree of genetic affinity has been confirmed with the bearers of the Hoabinhian culture, ancient hunter-gatherers of Southeast Asia. The Hoabinhian was a lithic culture that flourished across the Indochinese Peninsula from roughly 44,000 to 4,000 years ago; like the Jōmon, its people sustained a hunter-gatherer way of life across long stretches of time.

This genetic connection suggests that the ancestors of the Jōmon reached the archipelago not only via a northern route through Siberia but also along a southern route, island-hopping northward from Southeast Asia. During the last glacial period, sea levels were approximately 120 meters lower than today, and the islands of Southeast Asia were joined as a single landmass. From this vast continent known as Sundaland, people may have traveled north.

The genetic singularity of the Jōmon—their deep divergence from continental East Asians, their affinity with ancient Southeast Asian peoples, and over ten thousand years of independent evolution on the archipelago—demonstrates genetically that the population called “Japanese” is not simply a subset of East Asia.

That singularity, blended with later Yayoi and Kofun-period migrants, produced a genetic mosaic unlike any other in the world.

The immune system is a ledger of every pathogen a population’s ancestors fought. That ledger sometimes tells a story quite different from the one whole-genome analysis reveals.

HLA (Human Leukocyte Antigen) genes govern how the immune system distinguishes self from non-self. They constitute the most polymorphic region of the human genome and determine organ transplant compatibility, and they also mirror a population’s evolutionary history in fine detail.

Research by Professor Katsushi Tokunaga of Tokai University and colleagues has identified several HLA haplotypes characteristic of Japanese populations. The most notable is HLA-A24-B52-DR15, carried by approximately 12% of the Japanese population yet exceedingly rare among Chinese and Korean populations, and, worldwide, concentrated almost exclusively on the Japanese archipelago.

One hypothesis holds that this haplotype was selected during the Jōmon people’s long adaptation to pathogens endemic to the archipelago’s hot, humid forests—the viruses and parasites that flourished in dense woodland. More than ten thousand years of genetic isolation produced a path of immunological evolution distinct from that of all other Asian populations.

HLA is particularly revealing because it evolves under balancing selection rather than genetic drift. While most of the genome trends toward homogenization in isolated populations, HLA diversifies in response to each new pathogen threat. The Jōmon’s HLA recorded not only where they came from but what they fought. Where whole-genome analysis narrates the history of human migration, HLA narrates the history of war against disease—two chronicles bound within the same DNA.

Certain HLA haplotypes are also associated with susceptibility to autoimmune diseases. The relatively high incidence of rheumatoid arthritis and Graves’ disease in Japan is not unrelated to this immunogenetic singularity. The weapons those ancestors won in ancient battles now act on the modern body in unexpected ways.

There is one decisive gap in Japanese genomic research: the restriction on investigating imperial tombs.

The Imperial Household Agency administers 896 burial sites across the country. Of these, 188 are classified as “imperial mausolea” (tombs of emperors and empresses) and 555 as “imperial graves” (tombs of other members of the imperial family). Academic investigation of the vast majority is effectively prohibited, placing them beyond the reach of ancient DNA research.

In 2008, sixteen academic organizations including the Japanese Archaeological Association petitioned the Imperial Household Agency for limited access and permission to conduct scholarly surveys. As a result, observational visits to the outer perimeters of select sites were permitted, but excavation and DNA sampling remain forbidden to this day.

The gap matters because, as the tripartite-structure model showed, Kofun-period migrants defined the modern Japanese genome. Yet the genetic roots of the very center of Kofun-period power—the Yamato court, the imperial house—have never been scientifically examined. The tomb attributed to Emperor Nintoku (Daisen Kofun), one of the largest burial mounds on Earth, has never undergone a comprehensive academic excavation.

This “zone of the unexaminable” has become the breeding ground for urban legends claiming foreign origins for the imperial family: theories linking the population of the archipelago to ancient Israel, or tracing the imperial line to Sumer. The more science advances, the more the imagination swells around what remains beyond investigation. That is the structural mechanism by which urban legends form. The academic opening of imperial tombs could provide definitive evidence in the debate over Japanese origins, but the day that door opens remains nowhere in sight.

Y-chromosome haplogroup D has one of the most unusual distribution patterns among all human lineages. It is found at high frequency in only two populations: Japanese populations (approximately 35–40%) and Tibetans (approximately 40–50%). The two are separated by the width of a continent, yet share the same ancient lineage at high frequency.

The prevailing hypothesis runs as follows: the ancestral population carrying haplogroup D left Africa some 60,000 to 70,000 years ago and migrated eastward along the southern rim of Eurasia from South Asia. They once ranged across broad swaths of East Asia but were subsequently displaced by later-expanding groups (notably haplogroup O). The Japanese archipelago, shielded by ocean, and the Tibetan Plateau, shielded by mountains, both limited the influx of later populations and functioned as refugia where the ancient lineage survived.

In genetic terms, the Japanese D lineage is classified as D1a2a (formerly D1b), while the Tibetan lineage is D1a1 (formerly D1a)—branches of the same root that diverged tens of thousands of years ago. The precise formulation, then, is not that “Japanese and Tibetan populations are close relatives” but that “both are descendants of an ancient lineage that branched from a common ancestor.”

Notably, Y chromosomes extracted from Jōmon skeletal remains show a high probability of belonging to the D lineage, while Yayoi-migrant remains yield predominantly O-lineage results. The concentration of the D lineage thus serves as another index of a region’s Jōmon ancestry. The elevated D-lineage frequencies among Okinawan and Ainu populations are fully consistent with whole-genome analyses.

A caveat is in order, however. The Y chromosome reflects only paternal inheritance and therefore shows one dimension of a population’s history. Only when it is combined with maternal mitochondrial DNA and autosomal analysis does the full picture of Japanese origins come into view.

The genome does not only record an ancestral past; it acts directly on the body a person inhabits today. The genetic composition of Japanese populations produces distinctive patterns of susceptibility to specific diseases.

Consider alcohol metabolism. The ALDH2*2 variant—an inactive form of the aldehyde dehydrogenase enzyme common in East Asian populations—is carried by approximately 40% of the Japanese population. This variant, attributed to Yayoi-period migrants, is responsible for the “Asian flush” reaction to alcohol. Among populations with high Jōmon ancestry, such as Okinawans and Ainu, the frequency is notably lower. Alcohol tolerance turns out to mirror the Jōmon-to-migrant ratio in an individual’s ancestry.

The BioBank Japan project, led by RIKEN, has accumulated genetic and clinical data from approximately 270,000 Japanese individuals. Analysis has identified numerous disease-susceptibility variants specific to the Japanese population. Risk variants for type 2 diabetes, rheumatoid arthritis, and atrial fibrillation, among others, operate against a genetic background fundamentally different from that of European populations.

Equally significant are differences in drug response. Reactions to certain anticancer agents and immunosuppressants vary markedly depending on genetic background. Medications developed in the West do not necessarily produce the same effects in Japanese patients, and conversely, developing drugs that work particularly well in this population requires a deep understanding of the Japanese genome.

The ancestral diversity set out by the tripartite-structure model has direct consequences for modern medicine. The three genetic components—Jōmon, Yayoi, and Kofun-period migrant—shape disease risk, drug response, and physical traits. The genome is not only a chronicle of the past; it is a text that inscribes the present and the future.

At every drinking gathering, there are those whose faces turn crimson after a single sip. That reaction is the visible end of an evolutionary puzzle.

Alcohol is first metabolized in the body into acetaldehyde, then converted by the enzyme ALDH2 (aldehyde dehydrogenase type 2) into harmless acetic acid. Individuals carrying a single-nucleotide variant known as rs671 (ALDH2*2) have sharply reduced enzyme activity. Acetaldehyde accumulates, triggering facial flushing, palpitations, and nausea—the so-called Asian flush.

The geographic skew of this variant is pronounced. Approximately 40% of the Japanese population carries it, as do 30–40% of Han Chinese in southern China and 25–30% of Koreans. Among Europeans, Africans, and South Asians, it is virtually absent: a “non-drinker gene” concentrated in East Asia alone.

The spread of an apparently disadvantageous mutation to such high frequencies calls for explanation. The leading hypothesis is co-evolution with rice cultivation. Wet-rice agriculture, which began in the Yangtze River basin roughly ten thousand years ago, created paddy environments ideal for mosquito breeding, spreading malaria and other infections. The ALDH2*2 variant limits alcohol consumption by making drinking unpleasant, reducing deaths from alcohol dependence and intoxication-related accidents. Some studies further suggest that elevated acetaldehyde levels may inhibit replication of the malaria parasite.

Within the Japanese archipelago, the distribution of this variant maps neatly onto the tripartite-structure model. Populations with high Jōmon ancestry—Okinawans and Ainu—show low ALDH2*2 frequency, while the Kinki region, with its high continental-migrant ratio, shows high frequency. Of all the genetic indicators of that ratio, alcohol tolerance is the one most people can observe in themselves.

A powerful counterargument notes that molecular-clock estimates for the emergence of ALDH2*2 range widely, from 2,000–3,000 to 7,000–10,000 years ago. If the variant predates rice cultivation, the co-evolution account weakens. Moreover, India and Southeast Asia, where rice farming flourishes, show virtually no presence of this variant. Yet the hypothesis survives because the variant’s emergence and its selective expansion are separate events: even if the mutation arose before agriculture, its rapid spread may have been driven by survival advantages within rice-farming societies.

That flush at a drinking party may be the remnant of a “survival device” acquired ten thousand years ago, when the ancestors who carried it began cultivating rice in a hazardous environment.

Of the entire human genome, the regions that encode proteins constitute a mere two percent. The remaining ninety-eight percent was long dismissed as “junk DNA,” evolutionary debris that served no purpose—or so it was believed.

In 2012, the international ENCODE project overturned that assumption at its foundations. Over four years, 442 researchers scrutinized all three billion base pairs and found that roughly eighty percent of the so-called junk regions carry biochemical function. They are regulatory switches—determinants of when, where, and which genes are activated.

The implication is that DNA resembles not so much a blueprint as a musical score. The same score, played with different dynamics and emphases, yields entirely different music. The non-coding regions provide those performance instructions. Epigenetics—the mechanism by which gene expression is altered without changing the DNA sequence itself—can be rewritten by environment, diet, stress, and even the experiences of one’s ancestors.

In 2013, a research team at Emory University published an animal study with an unexpected result. Mice were conditioned to associate a particular odor with electric shock. Their offspring—a generation that had never experienced the stimulus directly—showed fear responses to the same odor. The memory of fear had been “inherited” through epigenetic modification.

If this mechanism applies to humans as well, then the Jōmon people’s fifteen-thousand-year accumulation of attunement to the natural world, the continental memories carried by Yayoi migrants, and the statecraft brought by Kofun-period arrivals may not be cultural inheritance alone. They may be inscribed as epigenetic traces in the non-coding regions of the genome. The silence of that ninety-eight percent may be an archive holding tens of thousands of years of ancestral experience.

Sixteen thousand five hundred years ago, as the last ice age loosened its grip on the Japanese archipelago, something happened that had never occurred anywhere on Earth. Someone shaped wet clay into a vessel, placed it in fire, and waited. The pottery fragment recovered from the Odai Yamamoto I site (大平山元I遺跡) in Aomori Prefecture has been radiocarbon-dated to approximately 16,500 years before present—the oldest confirmed pottery in the world. It predates the earliest Mesopotamian ceramics by more than four millennia.

The conventional narrative of civilization links pottery to agriculture: grain needs storing, food needs cooking, therefore vessels follow farming. But the Jōmon were not farmers. They were hunter-gatherers who invented pottery before any known society on the planet had planted a crop. This single fact dismantles the Western-derived model of civilizational progress—the assumption that the sequence must run agriculture, then settlement, then technology (see Chapter 5, “The Western Bias of the Four Great Civilizations”).

The most persuasive explanation is what might be called the “boiling revolution.” Tree nuts abundant in the Jōmon forests—acorns, horse chestnuts, buckeyes—are inedible raw because of their high tannin content. Removing those toxins requires prolonged boiling, and prolonged boiling requires a vessel that can withstand sustained heat. The Jōmon did not invent pottery to store surplus. They invented it to make edible what was otherwise poisonous. Pottery was not a consequence of abundance; it was a technology of survival in the forest.

What confounds explanation, however, is the sheer extravagance of Jōmon pottery. If function were the sole driver, the vessels could have remained plain. Instead, the Jōmon produced flame-style pottery (kaen-doki, 火焔型土器)—soaring, coiling forms of almost hallucinatory complexity. When the artist Taro Okamoto (岡本太郎) encountered Jōmon pottery at the Tokyo National Museum in 1951, he described the experience as “a dialogue with the fourth dimension.” The impulse to pour excessive beauty into utilitarian objects runs unbroken through sixteen millennia of Japanese craft, from Jōmon ceramics to Edo lacquerware to contemporary industrial design (see Chapter 4, “Miso, Soy Sauce, and Natto—The DNA of Fermentation Culture”).

The strongest counterargument cites pottery fragments from Xianren Cave (仙人洞遺跡) in Jiangxi Province, China, claimed to be as old as 20,000 years. But those dates derive from surrounding sediment rather than from the pottery itself, and the methodology remains contested. The Odai Yamamoto I dating was performed directly on carbonized material adhering to the pottery surface, making it considerably more robust. Regardless of which site ultimately claims priority, the fundamental fact stands: one of the world’s oldest ceramic traditions arose not in the cradle of agriculture but at the forested edge of a volcanic archipelago, among people who had no need for farming—and who would continue to have no need for it for another ten thousand years.

Between 266 CE and 413 CE, the record of the Japanese archipelago falls silent for a hundred and fifty years.

The Chinese chronicle Wei Zhi (the “Account of the Wa”) records in meticulous detail the 3rd-century realm of Wa—the rule of Queen Himiko of Yamatai. But after the final tributary entry in 266 CE, all record of Wa breaks off. When Wa next appears in Chinese histories, in 413 CE, it has become an almost unrecognizable state.

Before the void, Wa was a society of fragmented small kingdoms governed by a queen wielding shamanistic authority. After the void, the Yamato polity was building colossal keyhole-shaped burial mounds, fielding iron weapons and horses, and unifying the archipelago from its power base in the Kinki region as a powerful military state. What happened?

The archaeological record only deepens the enigma. With the onset of the Kofun period (late 3rd century), objects appeared across the archipelago abruptly: the distinctive keyhole-shaped burial mound, vast arsenals of iron weaponry, horses and equestrian gear, gilt-bronze ornaments. None of these had any significant presence during the Yayoi period. Whether they evolved gradually from indigenous culture or were introduced all at once from outside remains a subject of active scholarly debate.

In 1948, Namio Egami of the University of Tokyo advanced the “horse-rider conquest theory”—the hypothesis that continental horse-riding peoples conquered the archipelago via the Korean Peninsula and established the Yamato dynasty. Though mainstream scholarship remains skeptical of the theory, it persists as a compelling framework for the abrupt transformations of the Kofun period.

What DNA shows is genetic change across that same interval. As the 2021 tripartite-structure model demonstrated, a third ancestral population arrived during the Kofun period and decisively shaped the modern Japanese genome. During those 150 silent years, the gene pool of the archipelago was rewritten at its foundation. In the century and a half when written records fell mute, DNA carries the evidence that something occurred. Until the Imperial Household Agency permits excavation of the imperial tombs, the full truth remains in the dark.

In September 1872—Meiji 5—someone wrapped an iron knife in washi paper and wrote, in ink, “Nintoku Emperor’s Tomb” (仁徳帝陵). The knife had come from just beyond the moat’s edge, not the mound itself. The antiquarian who wrote the note was Kaichirō Kashiwagi (柏木貨一郎). For the next 150 years, the wrapping paper and the artifact passed through the hands of the Meiji industrialist Takashi Masuda (益田孝) and vanished into a private collection, out of scholarly view.

These escaped grave goods test the limits imposed on the Imperial Household mausolea discussed earlier in this chapter (see Chapter 1, “The Unopened Door—Imperial Tombs and the Forbidden DNA Research”). Daisen-ryō Kofun (大仙陵古墳), attributed by tradition to Emperor Nintoku, is among the largest keyhole-shaped burial mounds on Earth, yet no full-scale academic excavation of its interior has ever been permitted. The single most important monument of the Kofun period, the era in which the modern Japanese genome was finalized, has remained a black box to the present day.

Then in 2024, Kokugakuin University Museum (國學院大學博物館) acquired the long-lost collection. In June 2025 the university published the results of a joint study with Sakai City (堺市) and Nippon Steel Technology: materials analysis found the gilt-bronze-mounted knife, the armor fragments, and the Meiji-5 wrapping paper all consistent with fifth-century Daisen-ryō grave goods. The knife’s construction is especially singular: thin gold-plated copper sheets joined with silver rivets, a combination unrecorded in any other fifth-century kofun artifact. The workshops of the Yamato court were practicing metalworking techniques that cannot be reconstructed from the 5th-century finds alone.

But a boundary must be observed. That these items are “grave goods of fifth-century Daisen-ryō Kofun” is confirmed by materials analysis and the 1872 documentation. That they “belonged to Emperor Nintoku himself” remains hypothesis. The Imperial Household Agency’s designation of the mausoleum is unchanged; no academic evidence attributes an occupant. What 2025 crossed was the question “Are these Daisen-ryō artifacts?” The question “Whose artifacts are they?” remains wide open.

And yet the announcement revealed a quiet paradox. Without opening the tomb, by gathering the fragments that escaped its moat 150 years ago, scholarship can approach the center of Kofun-period power. The real field of inquiry was not inside the black box but among the pieces scattered along its edge. Just as DNA has begun to speak for the silent interval in early Japan’s history (see Chapter 1, “The 150-Year Void—Lost Records and a Fault Line in DNA”), the Daisen announcement shows that paths around forbidden excavation exist—provided, of course, that someone thought to wrap an artifact in paper and keep it safe for a century and a half.