Tibetan people can live at altitudes above 4,000 metres where many humans struggle with thin air, and one of the genetic tricks that helps them do it did not arise in Tibetans alone: a 2014 Nature study found that a key high-altitude variant near EPAS1 was inherited from Denisovans, an extinct group of archaic humans who interbred with our ancestors.

Tibetans have lived for thousands of years on a plateau that averages above 3,500 metres, with many communities settled higher than 4,000, in air thin enough to leave most lowland visitors short of breath and prone to altitude sickness. Part of how they manage it was not evolved within their own lineage.

It was inherited from another kind of human.

The clearest evidence for this comes from a 2014 paper in Nature, “Altitude adaptation in Tibetans caused by introgression of Denisovan-like DNA,” led by Emilia Huerta-Sánchez. The study found that a key high-altitude variant in the gene EPAS1 carries the genetic signature of Denisovans, an extinct group of archaic humans known to have interbred with the ancestors of some modern populations.

This is no longer a lone or contested result. A decade of follow-up work has reinforced it, and the Tibetan EPAS1 case is now one of the most firmly established examples of what geneticists call adaptive introgression, where a useful stretch of DNA from an archaic human species is retained and spread by natural selection. The strength of the evidence is not in question. What is worth handling carefully is its scope.

What the 2014 paper actually found

EPAS1 is a hypoxia pathway gene. It codes for a transcription factor that helps regulate the body’s response to low oxygen. It had already been flagged as carrying the most extreme signature of natural selection found in Tibetans, and as being linked to differences in haemoglobin concentration at altitude. The 2014 study set out to understand why.

The team re-sequenced the EPAS1 region in 40 Tibetan and 40 Han Chinese individuals at high coverage. They found a haplotype, a block of variants inherited together, with a structure so unusual that the most convincing explanation was the introduction of DNA from Denisovans or a closely related population. Scanning a much larger set of populations worldwide, they found the selected haplotype in only two places: in the Denisovan genome, and in Tibetans, with a trace presence among Han Chinese.

The length of the shared segment, and its near-total absence everywhere else, made ordinary shared ancestry an unlikely explanation. Introgression fit the data better. Interbreeding, rather than coincidence.

The adaptation works by doing less, not more

The mechanism is the part most popular accounts get backwards. The intuitive assumption is that people adapted to thin air must make extra red blood cells to carry more oxygen. The Tibetan variant does close to the opposite.

Carriers of the derived EPAS1 allele tend to have lower haemoglobin concentrations at altitude, not higher. The common short-term response to thin air is to raise haemoglobin, which thickens the blood. Sustained over a lifetime, that thickening carries costs, and in the research literature it is associated with chronic mountain sickness, raised cardiovascular risk, and complications in pregnancy. The Tibetan adaptation appears to sidestep that response, supporting oxygen delivery without the same rise in blood viscosity.

The useful version of the finding is therefore narrower, and more interesting, than “a gene for thriving at altitude.” It is a variant associated with not over-reacting to altitude in a way that, over generations, would do harm.

Why this counts as borrowed, not evolved

What makes the case striking is that Tibetans do not carry much Denisovan ancestry across the rest of their genomes. They hold this particular segment at high frequency while carrying relatively little archaic DNA overall, which is the fingerprint of selection. A borrowed variant that proved useful in one specific environment was kept and amplified, while most of the rest washed out over time.

The contrast with other populations sharpens the point. Some Oceanian groups, including Papuans, carry considerably more Denisovan ancestry genome-wide than Tibetans do, yet they do not carry the EPAS1 haplotype. The most economical reading, set out in a 2021 paper in the Proceedings of the National Academy of Sciences on the history of the Denisovan EPAS1 haplotype, is that without the selective pressure of a high-altitude environment there was nothing to favour keeping it.

The adaptation was not built from scratch on the plateau. It was already present in the gene pool as archaic inheritance, and the plateau is what made it matter.

The Denisovans were there first

There is a further layer the 2014 genetics could only hint at. In 2019, a separate team reported in Nature a Denisovan jawbone, the Xiahe mandible, recovered from Baishiya Karst Cave on the northeastern edge of the Tibetan Plateau at around 3,280 metres. Identified through ancient protein analysis and dated to at least 160,000 years ago, it is the first confirmed Denisovan fossil found outside Denisova Cave in Siberia.

The fossil shows that Denisovans occupied the high-altitude plateau long before modern humans reached it, and the authors describe them as having adapted to that hypoxic environment. It does not, on its own, prove that this individual carried the EPAS1 variant that later passed into Tibetans, since the mandible yielded proteins rather than the nuclear DNA that would settle the point. But it makes the inference cleaner. The variant that helps Tibetans live at altitude may have been carried by a human species that was already living there.

What the case does and does not show

The limits are worth stating plainly. This is one gene, in one population, against one environmental stressor. EPAS1 carries the strongest single signal of altitude adaptation in Tibetans, but it is not the whole story, and altitude tolerance involves several genes alongside a good deal of non-genetic acclimatisation. The finding does not mean humans in general are built for altitude. Most of us are not, which is why the plateau remains hard on visitors.

Nor does archaic inheritance imply design. Interbreeding scattered Denisovan and Neanderthal DNA through modern populations more or less indiscriminately. Most of it was neutral or mildly harmful and faded out. A small fraction happened to be useful in a particular place, and selection did the rest.

The wider point is a modest one. Some of the human capacity to live at the edges of what the body can tolerate did not have to be invented. It was borrowed from other kinds of human who got there first, and kept because it worked.