Bones Behind the Name: Ancient Proteins Bring Denisovans into Focus

admin · November 2, 2026

A fragment of skull and part of a lower jaw from Xujiayao, northern China, have supplied molecular evidence for an identity that anatomy alone had left unsettled. Excavated between 1976 and 1979, the specimens belong to an assemblage long debated by specialists in human evolution. In a study published in Nature on 21 October 2026, Song Xing and colleagues recovered 18 ancient proteins across the two samples. A particular collagen variant supported their assignment to the Denisovan lineage. The researchers then compared the Xujiayao remains with other Asian fossils, proposing a broader anatomical definition of Denisovans. Their study connects a molecular identification to an archaeological collection whose evolutionary position has resisted agreement for decades.

The distinction between those two steps matters. Identifying sampled bones through proteins and assigning other fossils through resemblance are different kinds of argument. Together they can be productive, but their evidential strength should remain visible. The immediate achievement is a firmer connection between ancient molecules and surviving anatomy; the larger proposal concerns how much of eastern Asia's disputed human fossil record that connection might explain.

A lineage discovered before its anatomy was understood

Denisovans entered scientific discussion through an unusual route. Most fossil human groups were first recognised from conspicuous anatomical differences. Denisovans were identified through genetic material recovered from small remains at Denisova Cave in Siberia's Altai Mountains. Research published in 2010 established a population related to Neanderthals but genetically distinguishable from both them and living humans. It also detected a Denisovan contribution to the ancestry of present-day Melanesians. A fragmentary fossil record had disclosed a population history extending far beyond the place where its bones were found. David Reich and colleagues' original analysis made that geographical mismatch apparent from the beginning.

This created an awkward archaeological problem. Geneticists could recognise Denisovan ancestry, while palaeoanthropologists lacked a substantial reference collection against which to identify unfamiliar skulls and jaws. An Asian fossil could look unlike a modern human or a typical Neanderthal without thereby becoming demonstrably Denisovan. Anatomical difference established a question; it did not supply the answer.

The first major bridge came from the Tibetan Plateau. In 2019, Fahu Chen and colleagues reported that ancient proteins linked the Xiahe mandible from Baishiya Karst Cave to Denisovans. Uranium-series dating of carbonate adhering to the jaw established a minimum age of approximately 160,000 years. Here was a substantial lower jaw, with teeth, giving anatomical substance to a lineage previously known chiefly from Siberian molecular evidence. Its location also demonstrated that Denisovans occupied high-altitude environments outside the Altai. The Xiahe study therefore changed both the anatomical and geographical terms of the discussion.

Harbin brought a much fuller view in 2025. Researchers associated its nearly complete cranium, older than 146,000 years, with Denisovans through protein analysis and mitochondrial DNA. Attempts to retrieve DNA from the tooth and petrous bone had failed, but dental calculus, the mineralised deposit on a tooth, preserved mitochondrial sequences within known Denisovan variation. Qiaomei Fu and colleagues' DNA study illustrates how preservation can differ even between tissues belonging to one individual. The accompanying proteomic investigation provided another molecular route to its identification.

Xujiayao therefore belongs to a developing anatomical record. Describing it as the first revelation of a Denisovan face would obscure the importance of those earlier discoveries.

What a surviving protein can establish

Palaeoproteomics studies ancient proteins and their surviving fragments, called peptides. Proteins consist of amino acid sequences, some positions in which vary between evolutionary lineages. Mass spectrometry measures peptides and their fragments, allowing researchers to infer sequences and compare them with reference data. Certain proteins can persist under conditions in which recoverable DNA has disappeared. Their survival depends on the tissue and its burial history, however; a fossil's age alone does not determine whether analysis will succeed. Christina Warinner and colleagues' methodological review explains both the opportunities and the complications.

Contamination requires particular attention in museum material. Handling, laboratory preparation and conservation can introduce proteins unrelated to the ancient individual. Chemical alteration, including deamidation of particular amino acids, can help assess authenticity, but it is influenced by burial conditions and laboratory treatment as well as time. Consequently, damage patterns must be considered alongside controls and the biological composition of the recovered material. They are supporting evidence, rather than a self-sufficient certificate of antiquity.

At Xujiayao, the diagnostic collagen change is designated COL1A2 R996K: lysine replaces arginine at a particular sequence position. The anatomical comparisons identify a combination including a low braincase, thick cranial bone and features shared with other archaic humans. The authors extend their Denisovan interpretation to fossils including Dali, Jinniushan and Hualongdong. That broader assignment rests on comparative morphology rather than a molecular identification of every specimen.

The appropriate reading is therefore graduated. The sampled bones provide a molecular anchor. Other remains can be evaluated against that anchor, but resemblance still needs interpretation. A shared feature may reflect common ancestry, inheritance from a much older ancestor, or a developmental response that occurred in more than one population. Combinations of characters are generally more useful than an isolated thick bone or prominent brow.

Nor does a protein sequence supply a portrait. Reconstructed bones constrain the form of a head; the appearance of lips, ears, hair and other soft tissues introduces additional assumptions. A virtual reconstruction should also distinguish preserved surfaces from mirrored or otherwise restored portions. These distinctions allow readers to appreciate a reconstruction without mistaking every visible detail for an observation.

The chronology belongs to the argument

The age of Xujiayao deserves attention independent of its molecular identification. A recent dating investigation by Qingfeng Shao and colleagues combined uranium-series work on hominin remains with coupled electron-spin-resonance and uranium-series dating of mammalian teeth. A Bayesian model integrating new and earlier results placed the upper cultural layer containing the hominin fossils between approximately 199,000 and 138,000 years ago, with uncertainties at both boundaries. A lower cultural layer was older. The dating study makes the distinction between archaeological layers explicit.

Such a range should not be converted into a precise birthday for the population. Electron-spin-resonance dating estimates accumulated radiation exposure; uranium uptake and the surrounding radiation environment enter the calculation. Archaeological ages consequently depend on models of physical processes as well as measurements. Agreement between dating evidence and stratigraphy strengthens a chronology, while uncertainty remains part of the result.

Secure context provides a different advantage from molecular survival. A protein may identify the organism from which a bone came, but it cannot by itself establish whether that bone was deposited alongside particular tools, moved after burial or incorporated into a younger deposit. Those questions belong to excavation records, sediment analysis and taphonomy, the study of what happened to remains after death. Molecular archaeology makes those records more consequential because it adds identities to material whose associations must still be assessed.

The contrast with Penghu 1 is instructive. Ancient protein analysis identified this mandible from the Taiwan Strait as a male Denisovan in 2025. Yet its chronology remained broad, with alternative age ranges reported in the research. The Penghu study shows that confidence in biological affiliation need not imply equivalent confidence in age. Questions about population identity, chronology and archaeological association require separate answers.

From individual fossils to inhabited landscapes

The expanding distribution of Denisovan fossils makes a narrowly Siberian account increasingly inadequate. In September 2026, Huiyun Rao and colleagues reported five molecularly identified Denisovan specimens from Bianfu Cave in Yunnan: two teeth, two skull fragments and part of a radius, one of the forearm bones. Their archaeological layers date to approximately 167,000–134,000 years ago. The radius added a new anatomical region to the limited Denisovan record. The Bianfu investigation also demonstrated the value of examining collections of fragmented animal bone for overlooked human remains.

A companion study examined the cave's archaeological record and subsistence evidence. Its consideration of stone artefacts, animal remains and bone use moved discussion towards activities carried out in a particular landscape. Qijun Ruan and colleagues' analysis supplies the kind of contextual evidence required to turn a fossil distribution into an account of human occupation.

Baishiya offers another example. Research published in 2024 identified a Denisovan rib from a layer dated to approximately 48,000–32,000 years ago, while analysing a large animal-bone assemblage. Human modifications to bones supported an interpretation of carcass processing and extensive use of animal resources. The study connected biological identification with evidence for subsistence on the Tibetan Plateau.

These cases suggest a useful direction for Xujiayao research. Establishing who was present makes questions about resource use and occupation more specific. It does not automatically make an associated tool form a Denisovan cultural signature. Technologies can be shared, independently developed or maintained by populations with different ancestries. Conversely, related populations living in different environments need not leave identical archaeological assemblages.

How much variation fits within Denisovan?

Genetic research already cautions against imagining Denisovans as a uniform population. Guy Jacobs and colleagues' 2019 study of present-day genomes identified evidence for deeply divergent Denisovan ancestries in Papuans. The inferred history involved contacts with populations separated over substantial evolutionary time. Their analysis describes population structure that a single Siberian genome could not represent adequately.

That history makes anatomical diversity plausible, but does not prove any particular fossil belongs within it. The danger is circular reasoning: uncertain specimens are assigned to Denisovans because their differences are considered acceptable variation, then used to demonstrate how variable Denisovans were. Further molecular identifications can test the proposed anatomical grouping independently.

Interbreeding complicates simple divisions as well. In 2018, Viviane Slon and colleagues reported the genome of an individual with a Neanderthal mother and a Denisovan father. This direct evidence of mixed parentage establishes contact between distinguishable lineages. It does not mean that every fossil combining different-looking traits was a recent hybrid. That explanation requires its own evidence.

Formal species names raise another issue. Connecting a fossil to a genetic lineage does not, by itself, decide how that lineage should be classified zoologically. Anatomical diagnoses, population relationships and species boundaries address overlapping problems, but they are not interchangeable. The most useful immediate task is to establish which specimens can be connected securely, before expecting a name to settle their evolutionary history.

Xujiayao strengthens the connection between Denisovan molecular ancestry and a recognisable anatomical record. The more ambitious proposition, that numerous disputed East Asian fossils belong within the same lineage, remains testable. Its resolution will depend on additional molecular samples, careful dating and comparisons that preserve rather than conceal anatomical differences. The central question is now how the diversity of these ancient populations was distributed across places and periods, and which parts of that diversity the surviving collections actually preserve.

Selected Sources and Further Reading

  1. Xing, S., et al. (2026). The proteome of Middle Pleistocene Xujiayao and common Denisovan morphology. Nature. Published 21 October 2026.
  2. Shao, Q., et al. New U-series and coupled ESR/U-series dating of Xujiayao (northern China), the type site for Homo juluensis. Quaternary Science Reviews. DOI: 10.1016/j.quascirev.2025.109742.
  3. Warinner, C., et al. (2022). Paleoproteomics. Chemical Reviews 122, 13401–13446.
  4. Reich, D., et al. (2010). Genetic history of an archaic hominin group from Denisova Cave in Siberia. Nature 468, 1053–1060.
  5. Chen, F., et al. (2019). A late Middle Pleistocene Denisovan mandible from the Tibetan Plateau. Nature 569, 409–412.
  6. Fu, Q., et al. (2025). The proteome of the late Middle Pleistocene Harbin individual. Science.
  7. Fu, Q., et al. (2025). Denisovan mitochondrial DNA from dental calculus of the >146,000-year-old Harbin cranium. Cell 188, 3919–3926.e9.
  8. Tsutaya, T., et al. (2025). A male Denisovan mandible from Pleistocene Taiwan. Science.
  9. Rao, H., et al. (2026). Ancient proteins identify various Denisovan remains from Southwest China. Nature.
  10. Ruan, Q., et al. (2026). Denisovans from southwestern China and their subsistence strategies. Nature.
  11. Middle and Late Pleistocene Denisovan subsistence at Baishiya Karst Cave (2024). Nature 632, 108–113.
  12. Jacobs, G. S., et al. (2019). Multiple Deeply Divergent Denisovan Ancestries in Papuans. Cell.
  13. Slon, V., et al. (2018). The genome of the offspring of a Neanderthal mother and a Denisovan father. Nature 561, 113–116.

Comments 0

Log in to leave a comment.
Loading comments...