Until recently, the evolution of early modern humans in Asia has received less popular attention than their African and European counter parts. The trend is beginning to change, particularly with the blossoming of scientific research in China. Melinda Yang is playing her part in illuminating the story of how modern Asians emerged after the African diaspora. She set aside some time to discuss her research regarding Tianyuan man with SCINQ.
SCIENTIFIC INQUIRER: First, can you provide some background on the Tianyuan man? Where was he discovered? What was the context in which he was found in the cave? Were there any other fossils or tools found?
MELINDA YANG: The Tianyuan skeleton was found in 2003 at Tianyuan Cave, Zhoukoudian, China, not far from Beijing. This area is well-known in paleoanthropology, because the older Peking Man was found six kilometers northeast of the Tianyuan cave site in another cave in the early 1900s. The Tianyuan man was first discovered by local workers at the Tianyuan tree farm in 2001, and further uncovered in 2003 and 2004 by an archaeological team from the Institute of Vertebrate Paleontology and Paleoanthropology of the Chinese Academy of Sciences (where I work now in the Molecular Paleontology Lab founded by Professor Qiaomei Fu). In total, 34 human bone fragments were found, likely from a single individual. The cave deposit has four distinct layers, and his remains were primarily found in layer III, along with several mammalian fossils – primarily different deer species. These mammalian fossils show no evidence of cultural or carnivore related marks, and no stone artifacts or other cultural materials were found. Using samples from both him and other fauna, the site was dated to ~42-39 kBP using radiocarbon dating.
SI: What previous studies have been conducted regarding the Tianyuan man?
MY: There’s been several studies on him prior to our current study, most of them focused on his morphology. The first was “An early modern human from Tianyuan Cave, Zhoukoudian, China”, by Shang et al. (2007)  describing the site and his skeletal morphology, which showed that he looked like an early modern human but with some features more similar to that found in archaic humans. A smaller genetic study was conducted in 2013 (title is “DNA analysis of an early modern human from Tianyuan Cave, China” ) by my postdoctoral advisor Dr. Qiaomei Fu, where mitochondrial DNA, most of chromosome 21 and a few thousand autosomal SNPs were sequenced using bone powder sampled from his femur. This earlier study compared his DNA to the same positions of the genome in present-day humans, and showed that he had a closer relationship to present-day East Asian and Native American than European populations.
However, since 2013, several ancient individuals in West Eurasia have been sequenced, showing that present-day Europeans are composed of more than two populations, and one of those, called ‘Basal Eurasian’ , is likely to have separated from other non-African populations very early. This result could affect the conclusion that the Tianyuan individual is more closely related to Asians than Europeans, as the Basal Eurasian ancestry in the present-day European genome makes present-day Asians and Europeans look more different from each other than present-day Asians and ancient Europeans look from each other. However, the Tianyuan DNA sequenced in 2013 was not enough to effectively compare to the DNA sequenced from the ancient Europeans.
SI: Working with ancient DNA is notoriously tricky. How did you procure the DNA sample from the remains and subsequently analyze it?
MY: So this part is largely thanks to work from Dr. Qiaomei Fu, my postdoctoral advisor, and her fantastic lab technicians at the Molecular Paleontology Lab! The Tianyuan man’s DNA was first extracted from the femoral bone and the ancient DNA samples (called libraries) were stored away in freezers during the first study . They tested the genomic coverage and level of mitochondrial DNA (mtDNA) contamination in these libraries. Looking at genomic coverage tests the amount of human DNA available for analysis in each library. The contamination level refers to how often we see more than one allele at a mitochondrial site (due to its uniparental inheritance, we would only expect to see one allele) and tests for contamination from present-day humans, which may have happened during any point of the excavation or DNA extraction process when the Tianyuan material was handled by different people. Dr. Fu picked the top 15 libraries with the highest genomic coverage and lowest levels of mtDNA contamination to conduct the next step for our study: DNA capture.
DNA capture refers to the lab technique where a subset of the genome are pre-chosen for sequencing. By now, most ancient individuals were sequenced using DNA capture, in part because it makes the process drastically cheaper and more efficient. The bulk of the DNA in libraries are actually bacterial DNA, as the amount of human DNA present in the bones is very small. Previously, everything was sequenced, and methods of post-processing could separate out the human from bacterial DNA. However, usually that meant >99% of the lab resources were sequencing bacterial DNA. DNA capture relies instead on sequencing a standardized set of positions in the genome (called panels), using oligonucleotide probes (tiny strands of DNA sequence) to specifically target those positions. During the lab procedures, DNA fragments that did not have a pattern matching that on one of the probes were washed away, which removed the bulk of the microbial DNA. This was a key part of the process because the Tianyuan sample had high amounts of microbial contamination. A key aspect of this process is that the same panels were being captured in all ancient individuals, allowing enough overlapping data from ancient individuals to easily compare them to each other. Ancient DNA capture was developed in part by Dr. Fu [4, 5], and her and her lab technicians are very skilled at this technique. So, for each of the 15 libraries, they used DNA capture to amplify DNA at the same panels sequenced in other ancient individuals, allowing us to compare variation in the same set of genomic positions.
Dr. Fu and I worked together on the analytical side, after DNA sequencing. The analysis relies on a series of methods developed for comparing populations and/or individuals to each other. The main statistics we used comes from a set called the f-statistics , which compare 2, 3 or 4 populations/individuals to each other, assessing the relative amount of shared alleles between the compared groups. These tests first allow testing for ‘treeness’, where we can determine whether certain sets tend to be more closely related to each other than other sets. They also allow testing for ‘admixture’, where given knowledge of the relationship between sets, we can determine whether some sets violate the expected patterns of the best fitting ‘tree’, requiring another demographic process incorporating DNA sharing to explain the unexpected connections. With the next few questions, I delve more into some of the logic behind the analytical methods.
SI: How did you confirm Tianyuan man’s relationship with modern-day Asians as opposed to European?
MY: We found this in a few different ways. First, we assessed the amount of genetic similarity between the Tianyuan individual and all present-day and ancient humans by using a three population f-statistic comparing the Tianyuan individual, a present-day or ancient individual or population and an outgroup, the Central African Mbuti population. We found the highest amount of genetic similarity between the Tianyuan individual and present-day Asians and lower amounts of genetic similarity relative to both present-day and ancient Europeans. We also used direct, relative comparisons (using four population f-statistics) between Asians, Europeans, the Tianyuan individual and an outgroup (usually a present-day African population) and we found that present-day Asians always shared more alleles with the Tianyuan individual than Europeans, both present-day and ancient. Lastly, we also estimated a best fitting tree  including ancient Europeans, the Tianyuan individual and present-day East Asians and Native Americans, and we found that even allowing for migration between populations, East Asians and Native Americans always shared a common ancestor with the Tianyuan individual sooner than ancient Europeans.
SI: While Tianyuan man’s closest genetic relationship is with modern-day Asians, your research indicates that he shows intriguing similarities to a 35,000-year-old individual from Belgium, Goyet Q116-1. How is this possible?
MY: So, in #4, we established that the main relationship we observed would group Goyet Q116-1 with other ancient Europeans and group the Tianyuan individual with present-day East Asians. And when we ask whether Goyet Q116-1 or Tianyuan is more closely related to an Asian or European, our results are consistent with the above observation. However, one of the f-statistics allows a test that asks whether any of the ancient Europeans are more or less similar to the Tianyuan individual. Without any unusual connections, we would expect that this test would come out to zero, indicating that all ancient Europeans share the same relationship to the Tianyuan individual. Instead, we found that the test came out as significantly different from zero, such that the Tianyuan individual shared more alleles with Goyet Q116-1 than with other ancient Europeans. We did not find this relationship in comparisons to present-day East Asians or using any other ancient Europeans. Thus, there is a genetic similarity unique thus far to Goyet Q116-1 and the Tianyuan individual that needs to be explained.
We did some tests taking into account contamination or DNA damage, but neither can explain the genetic similarity, which makes it much more likely that a demographic explanation is needed. This is where it gets tricky, because we have a single individual from each area of Eurasia with this signal. While one individual’s genome can tell you a lot, modeling the different demographic scenarios that could give rise to this signal currently requires yet more data, hopefully from other individuals who also share the signal. However, we do know that two major types of demographic history can lead to such a signal: admixture or structure.
Admixture would be due to gene flow between populations related to the Goyet Q116-1 individual and the Tianyuan individual or gene flow from a third unsampled population into the populations related to these two individuals. The other scenario is that the separation between Asians and Europeans was complex and there was high ancestral population structure. Here, Asians and Europeans cannot be described as originating from a single ancestral population, but rather, they come from an ancestral population made of several subpopulations that experienced different rates of intermigration. One or more of these ancestral subpopulations had a greater influence on both the populations of Goyet Q116-1 and the Tianyuan individual. In this model, the similarities would not be due to gene flow entering an Asian or European population after they separated, but because prior to the Asian-European separation, the subpopulations they came from already shared more ancestry with each other than other subpopulations that contributed to other Europeans or Asians. While we currently do not have the methods to discern between these two demographic scenarios, we think it is worth investigating the structure model, as the large geographic distance and likely short time scale separating these early Europeans and Asians make it difficult to envision a plausible admixture model.
SI: In addition to similarities to the GoyetQ16-1, Tianyuan man shares some similarities with Native Americans from South America. How do you explain the link between the two?
MY: Answering this question also draws heavily on an f-statistic test. If Native Americans come from the same single founding population, we would expect that they share a similar relationship to the Tianyuan individual, or the f-statistic test would be approximately zero. Instead, we found that some Native Americans from South America share significantly more alleles with the Tianyuan individual than other Native Americans. Again, either admixture or structure has to explain this connection. An interesting previous study, by Dr. Pontus Skoglund and colleagues in 2015 , showed that a similar connection is found when using some Oceanian and Andamanese populations.
They proposed two potential models – one involving direct gene flow across the Pacific and the second mediated by population structure in East Asia that was carried across into the Americas because the founding population actually consisted of at least two subpopulations, one with a closer relationship to the Oceanians and Andamanese Islanders. They did not see large contiguous tracts of DNA related to populations across the Pacific, suggesting that the gene flow occurred very early (recent gene flow would show up as large blocks of DNA that is slowly broken into smaller pieces over time thanks to recombination). They thought that the direct admixture model was unlikely, as the earliest humans are believed to have crossed the Pacific Ocean only 2-3,000 years ago. The other model, however, required population structure in East Asia, and in their study, they found no evidence of any present-day East Asians showing a similar signal. Thus, our finding that the signal appears with the Tianyuan individual is the first found in mainland Asia, supporting the structure model. What is really interesting, however, is that he is 40,000-years-old, while the colonization of the Americas occurred much later, around 20 kBP. Thus, a population with at least partial ancestry coming from a population related to the Tianyuan individual had to persist in Asia around for tens of thousands of years to be carried into the Americas. In Asia, populations carrying that ancestry eventually dies out, leaving no representation in today’s Asian populations!
SI: Can you place the study’s findings in the broader context of what is known about the evolution of early man in East Asia?
MY: There’s a few ways to think about this. First, morphological studies of him and other specimens in East Asia have a long history of what seems like archaic features that are not usually found in modern humans. Our study does not find that he shares any greater similarities to Neanderthals and Denisovans (archaic humans for whom we have genetic data) than a typical early modern human, which suggests that finding archaic features is not enough to indicate archaic admixture. However, that does not mean localized archaic admixture did not occur, as in Europe, the example of Oase 1 from Romania showed mixed morphological features and was found to have high Neanderthal ancestry, suggesting an ancestor of his from 4-6 generations back was a Neanderthal .
That the Tianyuan individual is more closely related to Asians and is 40,000-years-old is also very interesting. This shows that the prehistory of present-day East Asians is localized to Asia over the last 40,000 years. At the same time, however, you get the intriguing connection between Tianyuan and Goyet Q116-1. Other studies have shown some gene flow between Europeans and Asians, but I think most people thought that occurred more recently in time. Thus, to see a connection so early in time, not long after Europeans and Asians separated from each other, that is really interesting. Of course, a structure model could be a better explanation than the admixture model, but even that is fascinating – to have had a very complex Asian-European split. That scenario is also not well discussed yet in the literature.
Lastly, the Tianyuan individual highlights the high genetic diversity that was present in East Asia that is not observed today. His Native American connection indicates more than one East Asian subpopulation must have existed during the period of 40-20,000 years ago. Present-day East Asians just don’t show these signals, which leaves questions such as where did populations carrying Tianyuan-like ancestry live and when did they die out? In some ways, the Tianyuan individual points to how invaluable ancient DNA will be in understanding past East Asian prehistory.
SI: On a more personal-professional level, how did you come to a life in science? Did you always want to be a scientist?
MY: For me, it was always something in the back of my head, though the idea really started to take shape in college. I was always really interested in ancient or pre history, and when I discovered a way to study history using biology, I really started to appreciate the scientific process. Exploring research opportunities as an undergraduate, I discovered computational biology, and studying humans in the genomic age, that was really important. The idea of finding patterns in the data, of matching these patterns to different possible scenarios – it was very exciting from the first time I experienced the feeling and that excitement hasn’t gone away since.
SI: The Sciences seem to be progressing by leaps and bounds in Asia, especially China. Can you discuss the direction paleontology and paleoanthropology are taking locally?
MY: Yes, science funding has really exploded in China, which is fantastic. The biggest shift in paleoanthropology today is that ancient genomics is arriving in China. This field has largely been limited to European (and one American) labs, so it isn’t surprising that the largest advances thus far have been in Europe. However, as I mentioned above, the Tianyuan individual has left a bunch of intriguing hints on the diversity in East Asia and the loss of that diversity in present-day East Asians. There are so many questions related to East Asian prehistory, and Dr. Fu’s Molecular Paleontology Lab is a sign that in the next few years, much ancient DNA will be sequenced in China to help us understand the biological history of humans in China. Her lab is the first in China with the capability and expertise to sequence ancient genome-wide data and being part of that lab allows me to work on the coolest data, so I’m very lucky. There’s also a lot of really exciting discussion and work outside of just humans, using ancient animals and plants and their genetics to study domestication or environmental change, etc, so there is so much to be explored about more recent human prehistory!
Another big area is the relationship with archaic humans. I don’t discuss archaic humans much here, but the Denisovan is an archaic human from Siberia with a connection to present-day Oceanian populations. A big piece of land – Asia – separates these two regions, suggesting that an interesting history between populations related to the Denisovan and modern humans occurred in Asia. The Denisovan is interesting because there’s not much morphological data to go with it – only a couple of teeth and a finger bone have been discovered. Much of the history of the Denisovans comes solely from genetics. There is a huge push to figure out whether a Denisovan individual can be found in China, and that will take a mix of genetic and morphological study to find.
SI: What are the most important traits a researcher needs to have in order to succeed on a day to day basis?
MY: Oh, there’s so many directions I can take this! I think the four I will highlight here are perseverance, patience, effective communication, and enthusiasm. It takes so much time to find samples, generate data, analyze the data, and put it together into something coherent to share that the task can be daunting at times. Thus, it is important not to get frustrated easily, push through difficult parts of the research process and keep in mind both short and long term goals. For the Tianyuan project, I often found myself going in circles around a question, trying to determine if certain patterns answered the question I was asking. I definitely had to push myself sometimes to work out the logical thought process needed. Research is a slow process, so pacing myself appropriately is key.
I also really want to emphasize effective communication. Many people think of researchers as hiding behind piles of books, completely absorbed in investigating a research question. But there’s so much communication needed at many different levels – conveying the importance of your work to funding agencies, talking to researchers both within and outside of your field as well as other relevant parties to obtain samples, writing an article that is both scientifically accurate and engaging to readers (who may or may not have familarity with your field), and especially in this day and age, sharing your work with the public who mostly do not have the vocabulary and skills you might take for granted. Communication is a huge part of science, and I don’t think that’s emphasized enough. Lastly, enthusiasm – because these questions take so much time to answer, it is hard to stay in the business if you do not have enthusiasm for the material. At the hardest moments, reminding myself why I love the work I do is what keeps me going.
1. Shang, H., Tong, H., Zhang, S., Chen, F., and Trinkaus, E. (2007). An early modern human from Tianyuan Cave, Zhoukoudian, China. Proc Natl Acad Sci U S A 104, 6573-6578.
2. Fu, Q., Meyer, M., Gao, X., Stenzel, U., Burbano, H.A., Kelso, J., and Pääbo, S. (2013). DNA analysis of an early modern human from Tianyuan Cave, China. Proceedings of the National Academy of Sciences of the United States of America 110, 2223-2227.
3. Lazaridis, I., Patterson, N., Mittnik, A., Renaud, G., Mallick, S., Kirsanow, K., Sudmant, P.H., Schraiber, J.G., Castellano, S., Lipson, M., et al. (2014). Ancient human genomes suggest three ancestral populations for present-day Europeans. Nature 513, 409-413.
4. Fu, Q., Hajdinjak, M., Moldovan, O.T., Constantin, S., Mallick, S., Skoglund, P., Patterson, N., Rohland, N., Lazaridis, I., Nickel, B., et al. (2015). An early modern human from Romania with a recent Neanderthal ancestor. Nature 524, 216-219.
5. Haak, W., Lazaridis, I., Patterson, N., Rohland, N., Mallick, S., Llamas, B., Brandt, G., Nordenfelt, S., Harney, E., Stewardson, K., et al. (2015). Massive migration from the steppe was a source for Indo-European languages in Europe. Nature 522, 207-211.
6. Patterson, N., Moorjani, P., Luo, Y., Mallick, S., Rohland, N., Zhan, Y., Genschoreck, T., Webster, T., and Reich, D. (2012). Ancient admixture in human history. Genetics 192, 1065-1093.
7. Pickrell, J.K., and Pritchard, J.K. (2012). Inference of population splits and mixtures from genome-wide allele frequency data. PLoS Genet 8, e1002967.
8. Skoglund, P., Mallick, S., Bortolini, M.C., Chennagiri, N., Hünemeier, T., Petzl-Erler, M.L., Salzano, F.M., Patterson, N., and Reich, D. (2015). Genetic evidence for two founding populations of the Americas. Nature 525, 104-108.