What did Denisovans look like? Xujiayao gives them a face

For fifteen years, the Denisovans have been the most frustrating human group in recent prehistoryPrehistoryThe span of human history before the invention of writing, from the Palaeolithic to the Metal Ages, known mainly through material remains.→. Their genome was known in exquisite detail, their genetic legacy ran through the blood of hundreds of millions of living people, yet you would have been unable to recognise one of them in the street. Of these vanished cousins, all that survived was a finger bone, a few teeth and scattered bone fragments spread across an entire continent, far too meagre to sketch a face. In October 2026, a team led by palaeogeneticist Qiaomei FuPalaeogeneticsThe study of ancient DNADNAThe molecule carrying genetic information, used to reconstruct kinship between species.→ extracted from remains (bones, teeth, sediments, walls) to reconstruct the past of populations.→, of the Institute of Vertebrate Paleontology and Paleoanthropology (IVPP) in Beijing, published in Nature a study that changes everything: around twenty fossils from the site of Xujiayao, in northern China, forgotten since the 1970s, have just been tied to the DenisovansDenisovanAn extinct human population, cousin of the Neanderthals, identified in 2010 from the DNA of remains in Denisova Cave (Siberia).→ through the analysis of their ancient proteins. For the first time, or very nearly, we can put a real face on this shadowy people.¹

A face emerging from the shadows

What exactly do you see when you look a Denisovan in the eye? The answer delivered by the Xujiayao study is strikingly precise for humans of whom, only yesterday, we had nothing but crumbs. The skull is marked by a low, sloping forehead, receding backward instead of bulging like ours, topping a massive brow ridge, a thick supraorbital torussupraorbital torusA continuous bony thickening above the eye sockets, forming a massive brow ridge, a pronounced trait in Denisovans as in Neanderthals.→ curved toward the rear, forming a sort of continuous bony visor above the eyes. Beneath that ridge open curiously rectangular eye sockets, sharply defined, contrasting with the rounder orbits of Homo sapiensHomo sapiensThe present-day human species, which emerged in AfricaAfricaThe cradle of humankind: the continent where the first hominins appeared, then Homo sapiens around 300,000 years ago, before the expansion to the rest of the world.→ around 300,000 years ago, the only surviving human lineage after the extinction of Neanderthals and Denisovans.→. The cheeks are flat, the face broad and robust, and the nasal bridge stands out in a moderately projecting way. Nothing in this ensemble recalls the slender silhouette of modern humans.¹

These features do not come out of nowhere. They confirm and refine what other fossils, already tied to the group, had hinted at: a massive humanity, built to last, sitting in our imagination somewhere between modern humans and NeanderthalsNeanderthalsA fossil humanity of Eurasia, robust and cold-adapted, extinct around 40,000 years before present.→, yet entirely original. Still, caution is in order. Proteins do not directly give a portrait; what they do is allow bones to be confidently linked to a lineage. Once that identification is secured, palaeoanthropologists can read the morphology of those bones, the thickness of the brow, the shape of the orbits, the curve of the forehead, and reconstruct the broad lines of the anatomy. The Denisovan face emerging from Xujiayao is therefore the product of two distinct operations: a biochemical identification, then a classic reading of the skulls. It is the combination of both that finally licenses a description this confident.¹²

Xujiayao, twenty fossils forgotten since the 1970s

The Xujiayao site is anything but a recent discovery. Nestled in the Nihewan Basin, a vast region of northern China famous for its density of prehistoric deposits, it was excavated as early as the 1970s. Archaeologists there unearthed a collection of around twenty human fossils, skull fragments, isolated teeth and pieces of jaw, associated with a rich Middle PleistoceneMiddle PleistoceneGeological subperiod (about 770,000 to 126,000 years ago) marked by growing behavioural complexity among hominins.→ fauna and with stone tools. The assemblage is dated to at least 130,000 years, with some estimates pushing the age of the relevant layers back considerably further. From the very start, these remains puzzled researchers: too robust to be simple Homo sapiens, too late and too peculiar to slot easily into the Homo erectusHomo erectusAncient human species, the first hominin to leave Africa and the maker of the earliest Acheulean handaxes.→ box.⁶

For decades, then, these fossils floated in a kind of taxonomic purgatory. Researchers noted that they blended archaicArchaicRefers to an ancient, now-extinct human population or form (Neanderthals, Denisovans, ghost lineages), as opposed to anatomically modern humans.→ characters with more modern traits, without truly resembling any known group. There was talk of a late Homo erectus, of an undetermined Asian branch, and then, more recently, they were filed under the label of a proposed East Asian species, Homo juluensis, characterised notably by a very large braincase. It is precisely this uncertainty that made Xujiayao an ideal candidate for the new molecular methods: well-documented, well-dated bones whose affiliation no one could settle by anatomy alone. A key was missing, one able to open the lock. That key, Qiaomei Fu's team found not in DNA, too degraded in the warm, humid sediments of China, but in proteins.¹⁸

Proteins rather than DNA

Why proteins? Because DNA, the reigning molecule of palaeogenetics, suffers from one major flaw: it is fragile. In cold temperate regions, such as Denisova Cave in Siberia, it can survive tens of thousands of years; but under the warm, humid climates of East and Southeast Asia, it decays fast and often leaves nothing usable. Proteins, by contrast, are far more robust. Some, such as bone collagenCollagenThe main protein of bone, used as the medium for isotopic analysis of ancient diet.→ or tooth enameltooth enamelThe hardest tissue in the body, laid down in rhythmic layers during tooth formation that preserve a precise memory of time.→ proteins, can persist for several hundred thousand years where DNA has long since vanished. This is the whole point of palaeoproteomicsPalaeoproteomicsThe study of ancient proteins preserved in fossils (bone, tooth enamel); can reveal species or sex when DNA is gone.→palaeoproteomicsThe study of ancient proteins preserved in fossils; more resistant than DNA, they allow remains to be identified where the genome has vanished.→, the young discipline that reads the sequence of fossil proteins to draw evolutionary information from them.¹⁴

In practice, the team extracted the proteins preserved in the Xujiayao fossils, then sequenced the amino acid chains that compose them. Within these peptides, researchers hunt for tiny variations: the replacement of one amino acid by another at a precise position in a protein. These changes, inherited and passed from generation to generation, work as lineage markers. At Xujiayao, the analysis isolated a set of endogenous, genuinely ancient proteins, and above all a characteristic amino acid change, carried by collagen, also found in other East Asian fossils already identified as Denisovan. Crucially, this marker is absent in both Homo sapiens and Neanderthals. Its presence signs membership of the Denisovan lineage almost as surely as a genetic test.¹²

Palaeoproteomics is no magic wand, however. It has its limits, and researchers are the first to acknowledge them. Where a genome yields millions of pieces of information, a proteome from bone or tooth provides only a handful: a few proteins, sometimes one or two truly informative amino acid changes. That is enough to say "this fossil belongs to such-and-such a major lineage," but not to reconstruct fine family trees, to date population splits, or to measure the internal diversity of a group. DNA therefore remains richer, more talkative, when it is available. The two approaches are complementary: palaeogenetics for fine detail where preservation allows, palaeoproteomics for reach where DNA has disappeared. It is this second card that Asia, kingdom of climates hostile to ancient DNAAncient DNAGenetic material preserved in old remains, often degraded, sequenced with cutting-edge techniques.→, was condemned to play.⁴

From Homo juluensis to the Denisovans

The outcome of the Xujiayao study, then, is a taxonomic move. Fossils long classed under the name Homo juluensis, or left in cautious limbo, tip over to the Denisovan side. This is no mere game of labels. By tying Xujiayao to the Denisovans, the researchers give that group something it badly lacked: cranial remains numerous and varied enough to finally describe its facial morphology. The Denisovan skeleton, until now reduced to fragments, suddenly gains substance.¹

This reassignment illustrates a deep trend in Asian palaeoanthropologyPalaeoanthropologyThe science that studies human evolution from the fossil remains of hominins (bones, teeth, footprints) and their context, to reconstruct our biological origins.→. In recent years, a growing number of enigmatic fossils from China and its surroundings, long filed under various names, are being potentially drawn into the Denisovan orbit as molecular analyses advance. Some researchers, such as palaeoanthropologist John Hawks, go so far as to suggest that much of the Chinese fossil record postdating 600,000 years may eventually turn out to be Denisovan, in the same way that most European fossils of the same period proved to be Neanderthal once their DNA was analysed. The image of an Asian continent populated by multiple, compartmentalised species is gradually giving way to that of a single great Denisovan lineage, diverse but united, whose fossils were simply given too many names.⁴

The hypothetical element must nonetheless be measured. Not all specialists subscribe to this unifying vision, and some recall that morphology alone, in the absence of DNA, cannot always settle the matter. On a continent where genetic preservation is poor, there is a real risk of seeing each new ambiguous fossil poured a little too quickly into the convenient Denisovan category. Xujiayao, precisely, escapes that criticism, since its identification rests not on the shape of the bones but on a molecular marker. That is what gives the study its strength: it anchors in biochemistry what anatomy alone left hanging.⁸

Denisova, the Siberian starting point

To grasp the importance of Xujiayao, we must go back to the Denisovans' birth certificate, in 2010. That year, an international team announced it had sequenced DNA extracted from a tiny little-finger bone, found in Denisova Cave, deep in the Altai Mountains of Siberia. The surprise: the genome matched neither Homo sapiens nor Neanderthals. It revealed an entirely new human lineage, promptly named after the cave. For the first time in the history of the discipline, a vanished human population was discovered not through its bones, but through its DNA. The Denisovans were born as a genome in search of a body.⁶

This starting point lastingly shaped the problem. From Denisova Cave, only minute remains were available at first: the famous finger bone, a few isolated teeth, later some other fragments. Nothing that allowed a face, a stature, a bearing to be reconstructed. The Denisovans were defined by a genetic signature, not by an anatomy. For years, textbooks had to settle for representing them with a question mark, or with cautiously blurred silhouettes. We knew they had existed, that they had interbred with our ancestors, that they had occupied a huge swath of Asia, but we literally did not know what they looked like. It is this gap, this absence of flesh around the genome, that the discoveries of the following decade would begin to fill, fossil by fossil.⁶

Harbin, the Dragon ManDragon ManNickname of the Harbin skull, the type of the Homo longi group, dated to at least 146,000 years.→ who had already given a face

Before Xujiayao, another centrepiece had already lifted the Denisovans out of abstraction: the Harbin skull, in northeastern China. This near-complete cranium, hidden for decades at the bottom of a well where a worker is said to have concealed it in the 1930s before revealing it late in life, is one of the most spectacular human fossils ever found. Massive, with a braincase comparable to ours, topped by a powerful brow ridge, it was first described in 2021 as a new species, Homo longiHomo longiA hominin group defined in 2021 from the Harbin skull (Dragon Man), often linked to the Denisovans.→, nicknamed "Dragon Man" in reference to the Black Dragon province (Heilongjiang) where it was found.⁷

In 2025, the molecular inquiry reshuffled the deck. Work carried out, again, with the participation of Qiaomei Fu managed the feat of extracting mitochondrial DNA, not from the dense bone of the skull where it had vanished, but from the dental calculus clinging to the single surviving molar, as well as ancient proteins. The results, published in leading journals, tie the Harbin skull unambiguously to the Denisovan lineage: several protein variants specific to Denisovans, and a mitochondrial DNA that falls within their known variation. In other words, "Dragon Man" is a Denisovan. And because it is a nearly complete skull, it allowed palaeoartists to propose a detailed portrait: a broad face, flat cheekbones, a wide mouth, a large nose, a projecting brow, a voluminous brain. Harbin had given the first real face; Xujiayao now confirms and broadens its features across a series of individuals.⁷¹

An Asian mosaic: Baishiya, Cobra Cave, Penghu

Harbin and Xujiayao are not isolated cases. Over the past decade or so, the Denisovans have gradually materialised on an ever-wider map of Asia, and palaeoproteomics has played a decisive role each time. The first major milestone outside Siberia was a half-jaw discovered on the Tibetan Plateau, in Baishiya Cave, at more than 3,000 metres of altitude. For lack of usable DNA, it was the analysis of collagen proteins that revealed, in 2019, its Denisovan nature. This jaw, at least 160,000 years old, proved that the Denisovans had conquered the highlands well before the arrival of our species, a detail that takes on its full meaning with the question of altitude adaptation.⁶

Further south, an isolated tooth discovered in 2018 in Cobra Cave (Tam Ngu Hao 2), in the Annamite Chain of northern Laos, pushed the frontier further still. Dated to roughly 164,000 to 131,000 years, this molar, probably belonging to a young child, was tied to the Denisovans through the study of its internal structure and its proteins, DNA having failed to survive in the tropical heat. For Fabrice Demeter, a researcher at the University of Copenhagen and co-author of the study, this fossil represented the first discovery of Denisovans in Southeast Asia, a region where their presence had until then only been inferred from the genetics of present-day populations. To these milestones one must add the Penghu jaw, dredged up off Taiwan, also identified as Denisovan by the analysis of its proteins. And very recently, new remains from southwestern China, in Yunnan, have filled a geographic gap, documenting for the first time the way of life of these humans: a forested, conifer-dominated environment, specialised hunting of medium- to large-bodied game, and simple bone tools recalling those of Neanderthals.⁵³

Denisovan blood still runs in our veins

If the Denisovans fascinate so much, it is not only for their recovered appearance: it is because they are not truly extinct. Like the Neanderthals, they interbred with Homo sapiens during encounters between populations, and part of their genetic heritage was passed down to us. InterbreedingInterbreedingGenetic mixing between human populations or species; between Neanderthals and Homo sapiens it left 1 to 2% of Neanderthal DNA in non-Africans.→ is not a hypothesis: it is written in black and white in present-day genomes. The populations of island Southeast Asia and Oceania, in particular Papuans and Aboriginal Australians, carry the highest proportions of Denisovan DNA, sometimes several percent of their genome. Fainter traces are found widely across continental Asia.⁶

The most famous case of this heritage is also one of the finest known examples of adaptive introgressionIntrogressionThe lasting transfer of DNA segments from one population or species into another through repeated interbreeding, detectable in genomes long afterwards.→ in humans: the EPAS1 geneEPAS1 geneA gene regulating the response to low oxygen; a variant inherited from Denisovans helps Tibetans live at high altitude.→ in Tibetans. This particular version of the gene, inherited from the Denisovans, helps the body cope with the oxygen scarcity of very high altitudes by limiting an excessive reaction of haemoglobin. Present-day inhabitants of the Tibetan Plateau carry it in high proportion, even though their overall Denisovan ancestry remains modest: natural selection favoured this precise fragment of DNA because it conferred a vital advantage at altitude. The loop closes elegantly with the Baishiya jaw: the Denisovans already lived on the roof of the world, adapted to altitude, and it may be from contact with them that our ancestors inherited this precious version of the gene, a genetic gift from the depths of time that today allows millions of people to live where the air grows thin.⁶

One population, how many names?

One underlying question remains, the one that most stirs the scientific community: how many humanities, exactly, lie behind all these fossils? The taxonomic debates are heated. On one side, the names have multiplied with each discovery: Homo juluensis for Xujiayao and neighbouring fossils, Homo longi for Harbin and the Yunxian skull, "Denisovans" for the lineage defined genetically from the Siberian cave. On the other, an increasingly defended hypothesis holds that all of this ultimately designates only a single great Asian population, diverse in space and time but genetically coherent: the Denisovans.⁴

The heart of the problem lies in a quirk of history: the Denisovans were defined by their DNA before they had a body, whereas the other groups were named after bones, according to the classic rules of zoological nomenclature. Reconciling these two systems is a headache. If Harbin, defined as Homo longi, really is a Denisovan, which of the two names should prevail? And if Xujiayao, described as Homo juluensis, is one too? Some researchers argue for simply abandoning these species in favour of a single Denisovan lineage; others, on the contrary, defend a real diversity, with several branches having coexisted in Asia. Specialists in ancient genetics, such as Sally Wasef, of Queensland University of Technology (QUT), stress that the rise of ancient proteins changes the game by allowing previously mute fossils to be identified, but also recall that these data remain poor compared with a complete genome, and that one must guard against conclusions that are too firm.¹⁸

Whatever becomes of these quarrels over names, one thing is now settled: the Denisovans are no longer a faceless ghost. From icy Siberia to the Tibetan Plateau, from the forests of Yunnan to the tropical caves of Laos, by way of northern China and Taiwan, they appear as a vast, ancient, diverse people, perfectly settled in the human mosaic of the PleistocenePleistoceneThe geological epoch of the great ice ages (c. 2.6 Ma–11,700 BP), spanning most of human prehistory.→. Qiaomei Fu and her colleagues, by reassigning Xujiayao, did far more than settle the fate of a score of old fossils: they gave flesh to a genome, a forehead, eye sockets and cheekbones to a people once thought forever invisible. It remains to be seen how many anonymous fossils, still sleeping in the drawers of Asia's museums, are in turn waiting for proteins to give them back their true identity.¹⁴