A different kind of heart: Tibetans’ genetic uniqueness and enduring cultural sway
High-altitude adaptations and a millennium of religious relevance
Note: part 2 of 2. Read part 1.
Fourteen years ago, a UC Berkeley team studying altitude adaptation in Tibetans developed a method to scan the genome of various modern populations, looking for outlier gene frequencies compared to related populations. First, they compared Tibetans, Han Chinese and Northern Europeans, looking for genomic regions where Tibetans were the exception, the outlier. This approach identified EPAS1, a gene implicated in high-altitude adaptation, as a likely target of positive selection in Tibetans years before it was discovered that this gene introgressed from Denisovans.
This sort of technique obviously requires a good grasp of the genome’s broader phylogenetic patterns. You need some sense of the history of the populations to infer the peregrinations of specific genes. Next, calculating the timescale of divergences between Tibetans and Han Chinese, the authors estimated the two populations split 2,750 years ago. The problem with this estimate is that it places the proto-Tibetan stream’s separation from the proto-Han one at 750 BC, at least a millennium after we know the Han already existed as a people, and centuries into their long written history. Of course, this estimate, as estimates generally do, has multiple simplifying assumptions baked in that might be incorrect. For example, Han Chinese and Tibetans were both modeled as homogeneous groups at the tips of a bifurcating tree, rather than mixed populations. If Tibetans were a mix of two populations, each with distinct historic relatedness to the Han Chinese, then the average divergence estimate may have been misleading.
And in fact the 2,750-year estimate makes more sense if we do imagine that the Tibetans as a mixed population, with one of their ancestral groups closely related to the proto-Han. If you further consider the likelihood of subsequent gene flow from Han Chinese into Tibetans over the last few millennia, the estimate fits even better. Ethnolinguistic data supports the proposition that Tibetans and Han Chinese share a close connection despite vast cultural differences and proto-Han records that memorialize proto-Tibetans as alien barbarians; scholars still support the construct of a Sino-Tibetan linguistic family. If such a linguistic connection is observed, it is almost certain that proto-Tibetans and proto-Han diverged relatively late in the Holocene. But the Denisovan provenance of EPAS1 points to the strong likelihood of a separate modern human population occupying the Tibetan uplands by the Pleistocene (since Denisovans proper disappeared about 40,000 years ago); no doubt that Tibet is hostile to human flourishing, but contrary to past scholarly skepticism, some number of our species have clearly managed to eke out an existence in its crags and valleys for millennia.
Tibetan Y-chromosomal lineages hint at an ancient and very exotic ancestral element. About 50% of Tibetan men carry Y-chromosomal haplogroup D. Worldwide, this lineage is very rare, save for in two other populations, both notably inhabiting isolated archipelagos: the Japanese and the Andamanese. In Japan, D is almost certainly a legacy of the Jomon people, whose roots in the archipelago go back more than 20,000 years. D’s three sublineages across East Asia, seem to have split by at least 45,000 years ago, suggesting descendents are part of a widespread diversification of modern humans right after our species first arrived in the region. D is very possibly a marker for a broad set of deeply divergent East Eurasian populations whose former geographical range has since been largely overwritten by the rise and dominance of agricultural peoples from the Yellow River basin and southern China.
Interestingly, when we examine their whole genome, and compare them to other populations, we see that Tibetans tend to skew more northern among East-Asian populations, despite their southernmost range grazing South Asia. The principal component analysis (PCA) above files Tibetans between Siberian Yakuts and northern Han Chinese, while the admixture analysis above also finds greater similarity between Tibetans and northern East Asians than southern ones (and unlike in the Han, we see a low but consistent West-Asian admixture into Tibetans, pointing to connections spanning western Tibet and Central Asia).
Before attempting to understand the historical antecedents of these connections, we must recall that the modern Tibetan Autonomous Region in China, successor to the nation-state of Tibet (1912-1949), is actually just one component of broader historical Tibet, Ü-Tsang. The Tibetan Empire and its successor states had wider suzerainty, extending into Amdo in the northeast and Kham in the southeast. While Ü-Tsang is over 90% Tibetan, Amdo and Kham are multiethnic, and today Tibetans are a minority in these regions. But out of China's six million Tibetans, only about half live in the Tibetan Autonomous Region, with most of the rest in Amdo (today’s Qinghai) and Kham’s eastern districts, (today’s Sichuan). Historically, Amdo and Kham were just as much Tibetan homelands as Ü-Tsang was, and their history presents a plausible path of ancient migration and admixture across the plateau.
A model to explain Tibetan affinities with northern East Asian populations like Yellow River Han Chinese and Mongolians is that the proto-Tibetans, those tribes who spoke Tibeto-Burman languages, migrated westward from the Yellow River basin and began expanding south and west from Amdo, only settling Ü-Tsang and Kham to the south later; and indeed, related Burman populations’ incongruous (given their geographic position) genetic affinity with northern East Asians, points to the migration continuing southward beyond Kham. Once in the Tibetan highlands, it is evident that the Neolithic proto-Tibetans mixed with indigenous foragers, whose roots in the region date back to the Pleistocene, and who mixed with the Denisovans more than 40,000 years ago (thereby acquiring that crucial altitude-adapted variant of EPAS1).
To this case built on modern genomic inference and a few archaic Denisovan genomes, we can now add the smoking gun of ancient DNA from large numbers of anatomically modern humans in the region. A 2023 paper analyzed 89 samples from the Tibetan plateau dating from across the past 5,100 years; it nearly closes our case. For one thing, most of the ancient samples are genetically similar enough to modern Tibetans to confirm that they were at least partly ancestral to them. Populations of the contemporary highlands have occupied the region for much of the Holocene. Their roots go back at least 5,100 years ago, the date for an individual from the Zongri site in Amdo in the northeast whose connections to modern Tibetans are clear. But in relation to non-Tibetans, the Zongri individual genetically most resembles peoples who lived in the Yellow River basin to the east well before 5,000 years ago, indicating a migration long prior to his burial. Unsurprisingly, we see no strong affinity to populations from ancient southern China, confirming proto-Tibetans’ ultimate northern provenance. But that’s not all. All the ancient and modern Tibetans are best modeled with an additional ancestry component from a “ghost” population not yet discerned in its pure form in the ancient DNA. This group is deeply diverged from other East Asians, and is almost certainly a signal from indigenous Initial Upper Paleolithic-descended foragers still inhabiting the highlands when the lowlanders intruded upon the plateau.
What dynamic explains the demographic replacement of the indigenous foragers? Today, barley is Tibet’s principal crop, but this grain arrived from the west, through Indian Kashmir, just some 2000-3000 years ago. Previously, Tibetans cultivated millet, with sites confirmed as early as 2800 BC. Though this post-dates the Tibetan-like individual from the Zongri site by three centuries, the timing is close enough to boost confidence in the hypothesis that his people’s ethnogenesis was in the westward migration of a group of millet farmers before 3000 BC. This was just one in a succession of migrations from the mouth of the Yellow River that kicked off 8,000 years ago, and progressively spread agriculture across northern China. These movements long predated the Han’s emergence, who arose via the assimilation of migrants from southern China, marking them as quite distinct from early proto-Tibetans.
But Tibetan populations were not isolated in the plateau after 3000 BC. Ancient DNA samples from individuals who lived from 3000-1000 BC reflect sporadic and geographically heterogeneous pulses of admixture from adjacent lowland populations into the plateau. At the same time that the eastern populations in Amdo and Kham maintained contact with lowland East Asians, Tibetan migration into the south and west continued until their genetic profile was present in the highlands of Nepal, and overflowed west into Central Asia, attested as far afield as Ladakh, by 1000 BC. Over the next few millennia, Central and South Asian genetic signatures also appeared in Ü-Tsang, in line with the historically attested interactions between Tibetans in this region and both Indian and West Asian peoples.
Consistent with those diverse ancient contacts millennia ago, the ancient samples show clear genetic clustering within Amdo, Ü-Tsang and Kham respectively; the three regions had distinct profiles. But modern Tibetans no longer show such striking population structure. This implies that gene flow between the regions homogenized the Tibetan genetic profile just in the last 1000 years. The Tibetan Empire’s 7th-century AD emergence almost certainly kick-started that process, and then a unified Tibetan Buddhist culture on the plateau after 1000 AD further accelerated it. It was politics and culture that overcame tribal differences and the plateau’s forbiddingly vast distances, accelerating the natural evolution of a new people with a shared identity, written in both their genes and their memes.