The problem when all the fossils are male
Protein data shows that most known Denisovan teeth come from male individuals, hamstringing attempts to understand the variation of this group.

Compared to most kinds of mammals, fossils of our own family are pretty numerous. A century and a half of discovery has given us thousands of pieces of ancient humans, ancestors, and relatives. Even so, most of those fossils come from a few times and places, and a few branches of our family tree. Amid this crowded cabinet, a few ancient species can claim only a pittance of fragments.
These cases can be the most misleading.
The fossils known as Denisovans have been like that. The first hint of their existence was a lone finger bone. If not for its exceptional preservation of DNA, we might have spent another decade or more without recognizing the group at all. Fifteen years on, our knowledge of the Denisovans remains mostly genomic. We still have hardly any evidence about their skeletons.
This year has seen the addition of a new fossil to the Denisovan group. This jawbone, designated as Penghu 1, came from an unusual circumstance, dredged from the bottom of the Taiwan Strait nearly twenty years ago. Protein evidence now aligns this jaw with the DNA from Denisova Cave, Russia, making it the furthest-flung yet of Denisovan fossils.
The very scarcity of evidence about Denisovan anatomy frustrates many anthropologists. What can it mean to be part of an ancient group defined entirely from biochemistry?
Protein evidence puts a new twist on this question. Usually dental anatomy provides reliable anatomical evidence of connections between groups. But all Denisovan teeth with proteomic evidence of sex happen to come from male individuals. Are the differences between the so-called Denisovans and other fossils due to species, or sex?
Penghu 1 as Denisovan
From the first finger bone from Denisova Cave, biochemistry and detective work by several research teams have built up a sample tied to this population. The core group includes eight fragments of bone from Denisova Cave, Russia: the finger, three teeth, a piece of skull, shards of limb bone. The dirt floor of Denisova Cave also contains their DNA locked within microscopic pieces of bone, going back as early as 300,000 years ago.
Denisovan dirt DNA was detected at a second site, Baishiya Karst Cave in northwest China, reported in 2020 by Dongju Zhang and coworkers. A rib fragment from that site likewise the DNA signature. The Xiahe 1 fossil, which is roughly half a jawbone estimated to be more than 160,000 years old, also comes from this cave. That jaw preserves no DNA but protein from the bone shares unique amino acid changes with the Denisovan genome. Later, in 2022, Fabrice Demeter and collaborators described a molar tooth from Tam Ngu Hao 2 cave, Laos. While it had neither DNA nor good protein preservation, they observedthat this tooth as very close to the molar form of the Xiahe 1 jaw—possibly another Denisovan.

With each new fossil the geographic and temporal span of fossil Denisovans grows. Baishiya Karst Cave is some 2000 kilometers from Denisova Cave. Tam Ngu Hao 2 is around 2000 km south of Baishiya Karst Cave. The Penghu specimen broadens the southeastern extent of this range.
The history of the Penghu 1 mandible began in 2008 when Kun-Yu Tsai walked into an antiquary and bought the fossil. Chun-Hsiang Chang and colleagues told the story in 2015 when they described the jaw. The fossil had been dredged from the sea bottom and was covered in remains of marine invertebrates, which Tsai scraped off, revealing the jaw’s form.
Penghu 1 looks a lot like Xiahe 1. Both individuals congenitally lacked a third molar, a rare trait among fossil hominins. The similarity of dimensions and appearance of the teeth of the two led many scientists to suspect they belonged to the same population.
Getting proteins from ancient remains takes a mass spectrometer. This sensitive device can detect the differences in amino acid sequences between short protein fragments called peptides. With this approach, Takumi Tsutaya and collaborators recovered fragments of 51 different proteins from the bone and a tooth of Penghu 1. It’s a small trace that connects this fossil with the known Denisovans. One change, affecting one of the many protein chains of collagen, has only been seen so far in four fossils: Denisova 3, Xiahe 1, the Xiahe 2 rib, and Penghu 1. Another is shared by some Denisovans but also a small fraction of living people, including around 22% of people sampled in the Philippines.
Tying the data together enabled Tsutaya and coworkers to draw a tree of protein relationships, with a branch connecting Penghu 1 and Denisova 3. Clear enough.
Still, a couple of the amino acid changes in the Penghu 1 proteome are also found in a fraction of living people but not other fossils. And a handful of distinctive protein changes seen in one or another Denisovan fossils are not in Penghu 1.
These mismatches with the tree reinforce that some branches were interlaced with tendrils of genetic ancestry. Possibly when we are thinking about “Denisovans” we must consider ancestral connections beyond those visible in the few cases with ancient genome and proteome data.
How proteins reveal genetic sex
To analyze such small samples and compare them to other kinds of hominins, researchers have to really think about how such tiny numbers can bias our view.
One of the biggest potential biases is sex. In a small sample, it’s unlikely that the number of male and female individuals will be evenly balanced. There may be a few more fossils of one sex than the other. One sex might even be missing entirely.
When scientists take such a skewed sample and compare it with other groups, they get skewed comparisons. Garbage in, garbage out.
In humans and other kinds of primates, skeletal and physiological sex contribute to variation in body size and shape, lifespan, and pace of maturation. So anyone trying to study those things—which is pretty much everyone in paleoanthropology—has to come to some way of working with the small sample bias.

Today it has become possible to test some fossil teeth directly to discover the genetic sex based on proteins. This is starting to change the way we understand fossil variation.
Determination of genetic sex in fossil teeth is possible because of a protein called amelogenin. This protein exists in two forms, AMELX and AMELY, which are products of two different genes—one on the X chromosome and one on the Y. Since genetic females have two X chromosomes, their enamel contains only AMELX protein, while genetic males with both X and Y chromosomes have enamel with peptides consistent with both AMELX and AMELY.
Even in a very tiny samples of enamel from ancient teeth, it is often possible to detect the specific peptides corresponding to AMELX and AMELY. When both are in the sample, the tooth came from a genetic male individual. If only AMELX is present, the individual was likely a genetic female. The confidence in sex assessment depends on the degree of protein data preserved. Too little data, and the lack of AMELY in a sample might just be due to poor data recovery.
Tsutaya and coworkers found plenty of AMELY in the Penghu 1 tooth—so confident in their assessment that they put “male Denisovan” in the title of their article.
Then they noted something curious that got me thinking. There are not many Denisovan teeth with protein data. The protein data from Tam Ngu Hao 2 did not report enough amelogenin peptides to be confident of sex; no enamel results have been reported for Xiahe 1. But both Denisovan teeth from Denisova Cave itself were also male. Three big-toothed hominins, all male.
That result is not at all unlikely in a sample so small. Far from newsworthy. It’s not even that there are no females: the Denisova 3 finger bone that started it all came from a female individual.
But the tooth bias is important when we start asking the question of what other fossils might be part of this population. Usually teeth are the most numerous fossils, and details of tooth form give important clues about relationships. With Denisovans, the shared aspects of tooth anatomy revolve around the large size of the molars. And all those molars came from male individuals.
What did female Denisovans look like?
I wrote last year about the diverse fossils of China from the time between 350,000 and 130,000 years ago. Over the last few years, researchers have proposed that these fossils represent several different groups or species: well-worn names like Homo erectus and “archaic” Homo sapiens, new ones like Homo longi and the so-called Julurens.
One of the key areas of difference is related to size. The whole concept of the Julurens is centered around the exceptionally large size of these fossil heads.
Xujiayao, in northern China, has the kind of fossils scientists dream of finding—more than a dozen pieces of fossil hominin individuals, probably between 250,000 and 160,000 years old. One had a very big head indeed, around 1700 milliliters. That’s why Xiujie Wu first coined the term Juluren—“big headed people”. But the very big head doesn’t negate that even this site is a small sample of anatomical variation. With little anatomical overlap between different fragments, it’s not possible to say how variable most features were.
Comparing such a sample to other sites, some far away in space and time, requires assumptions. When anatomy or size differ, does that reflect different species, different populations, or different sexes?
Which—if any—of these fossils will prove to have the genetics identified from Denisova Cave? I’d like to say I’m Club Juluren: Xujiayao in many ways seems like an obvious Denisovan-like sample. The smaller individuals from other sites, like Hualongdong, don’t fit the same mold as easily. Yet maybe these sites just happen to have a female or two, with their apparent differences representing sex biased samples.
This problem is one of the oldest in human origins research. Again and again, from early hominins like Australopithecus and Paranthropus up to Neanderthals and early modern people, anthropologists have asked the exact same questions. In the cases where we’ve made progress over the years, it’s because researchers have discovered fairly complete skeletons.
I don’t have an answer to the question of what the teeth, jaw or skull of a female Denisovan would look like. Soon we will. A new horizon is opening. A growing set of genetic and protein data from Neanderthals is giving us a window on the variation attributable to sex in that group. Recent papers have shown that amelogenin sex assessment works in Paranthropus and Australopithecus from South Africa.
Denisovans are coming online. Genetic and protein results never can go fast enough, but almost certainly researchers will keep adding to this sample. The time cannot be long before a fossil skull like the Homo longi skull from Harbin, China, or a fossil skeleton like the one from Jinniushan, China, are added to the list.
From a wider view, we’ll soon know the genetic sex of many fossil hominins. That will go a long way to clarify how the bias of small samples has warped some ideas about human origins. I can’t wait to find out how wrong we’ve been.
References
Chang, C.-H., Kaifu, Y., Takai, M., Kono, R. T., Grün, R., Matsu’ura, S., Kinsley, L., & Lin, L.-K. (2015). The first archaic Homo from Taiwan. Nature Communications, 6(1), 6037. https://doi.org/10.1038/ncomms7037
Chen, F., Welker, F., Shen, C.-C., Bailey, S. E., Bergmann, I., Davis, S., Xia, H., Wang, H., Fischer, R., Freidline, S. E., Yu, T.-L., Skinner, M. M., Stelzer, S., Dong, G., Fu, Q., Dong, G., Wang, J., Zhang, D., & Hublin, J.-J. (2019). A late Middle Pleistocene Denisovan mandible from the Tibetan Plateau. Nature, 569(7756), Article 7756. https://doi.org/10.1038/s41586-019-1139-x
Demeter, F., Zanolli, C., Westaway, K. E., Joannes-Boyau, R., Duringer, P., Morley, M. W., Welker, F., Rüther, P. L., Skinner, M. M., McColl, H., Gaunitz, C., Vinner, L., Dunn, T. E., Olsen, J. V., Sikora, M., Ponche, J.-L., Suzzoni, E., Frangeul, S., Boesch, Q., … Shackelford, L. (2022). A Middle Pleistocene Denisovan molar from the Annamite Chain of northern Laos. Nature Communications, 13(1), Article 1. https://doi.org/10.1038/s41467-022-29923-z
Peyrégne, S., Slon, V., & Kelso, J. (2024). More than a decade of genetic research on the Denisovans. Nature Reviews Genetics, 25(2), 83–103. https://doi.org/10.1038/s41576-023-00643-4
Stewart, N. A., Gerlach, R. F., Gowland, R. L., Gron, K. J., & Montgomery, J. (2017). Sex determination of human remains from peptides in tooth enamel. Proceedings of the National Academy of Sciences, 114(52), 13649–13654. https://doi.org/10.1073/pnas.1714926115
Tsutaya, T., Sawafuji, R., Taurozzi, A. J., Fagernäs, Z., Patramanis, I., Troché, G., Mackie, M., Gakuhari, T., Oota, H., Tsai, C.-H., Olsen, J. V., Kaifu, Y., Chang, C.-H., Cappellini, E., & Welker, F. (2025). A male Denisovan mandible from Pleistocene Taiwan. Science, 388(6743), 176–180. https://doi.org/10.1126/science.ads3888
Wu, Xiujie. (2024). Research progress on human fossils from the Xujiayao site in late Middle Pleistocene. Acta Anthropologica Sinica, 43(01), 5. https://doi.org/10.16359/j.1000-3193/AAS.2023.0044
Wu, Xiujie & Bae, Christopher J. (2024). Xujiayao Homo: A new form of large brained hominin in eastern Asia. PaleoAnthropology, 2024. https://paleoanthropology.org/ojs/index.php/paleo/libraryFiles/downloadPublic/18
Zhang, D., Xia, H., Chen, F., Li, B., Slon, V., Cheng, T., Yang, R., Jacobs, Z., Dai, Q., Massilani, D., Shen, X., Wang, J., Feng, X., Cao, P., Yang, M. A., Yao, J., Yang, J., Madsen, D. B., Han, Y., … Fu, Q. (2020). Denisovan DNA in Late Pleistocene sediments from Baishiya Karst Cave on the Tibetan Plateau. Science, 370(6516), 584–587. https://doi.org/10.1126/science.abb6320