Who We Are and How We Got Here

by David Reich

  Ancient DNA studies with large numbers of samples also offer the promise of being able to estimate human population sizes at different times in the past, a topic about which we have almost no reliable information from the period earlier than the invention of writing, but which is important for understanding not just human history and evolution but also economics and ecology. In a population of many hundreds of millions (such as the Han Chinese), a pair of randomly chosen people is expected to have few if any shared segments of DNA within the last forty generations because they descend from almost entirely different ancestors over this period. By contrast, in a small population (like the indigenous people of Little Andaman Island, who have a census size of fewer than one hundred), all pairs of individuals are closely related and will show evidence of relatedness through many shared segments of DNA. Measuring how related people are has been used to show, correctly, that the size of the population of England in the last few centuries has averaged many millions.11 In ongoing work, Pier Palamara and I have demonstrated that the same approach can be used to show that early farmers from Anatolia of around eight thousand years ago were part of much larger populations than the hunter-gatherers from southern Sweden who were their contemporaries, as expected based on the higher densities that can be supported by agriculture. I have no doubt that applying this approach to ancient DNA will provide rich insight into how populations changed in size over time.

  Ancient DNA’s Promise for Revealing Human Biology

  Ancient DNA in principle has just as much insight to offer about how human biology has changed over time as it does about human migrations and mixtures. And yet while the power of ancient DNA to reveal population transformations has been a runaway success, so far the insights into human biology have been limited. A key reason is that to track human biological change over time, it is important to be able to study how mutation frequencies change. But this requires hundreds of samples, and to date, the sample sizes of ancient DNA have been relatively small, just a handful from each cultural context. What will happen once we have genome-wide data from a thousand European farmers living shortly after the transition to agriculture? Comparing the results of a scan for recent natural selection in these individuals to the same scan performed in present-day Europeans should make it possible to understand whether the pace and nature of human adaptation has changed between preagricultural times and the time since the transition to agriculture. It might even be possible to determine whether natural selection has slowed down in the last century due to medical advances that allow individuals with genetic conditions that would have prevented them from surviving and having families to live and procreate. Examples of such medical conditions include poor eyesight, which can now be fully corrected with spectacles, or infertility, which can now be corrected by medical interventions, or cognitive challenges, which can now be controlled by medication and psychotherapy. It is possible that this change in natural selection is leading to a buildup of mutations contributing to altering these traits in the population.12

  The power of ancient DNA to track the rate at which the frequencies of biologically important mutations have changed is important not just because it offers the possibility of tracking the evolution of specific traits, but also because it provides a previously unavailable tool that we can use to understand the fundamental principles of how natural selection proceeds. A central question in human evolutionary biology is whether human evolution typically proceeds by large changes in mutation frequencies at relatively small numbers of positions in the genome, as in the case of pigmentation, or by small changes in frequencies at a very large number of mutations, as in the case of height.13 Understanding the relative importance of each type of adaptation is important, but addressing this question is made more challenging when the only tool available is analysis of people all of whom lived in a single window of time. Ancient DNA overcomes this obstacle—the time trap of only being able to study the present.

  Ancient DNA research also reveals pathogen evolution. When grinding up human remains, we sometimes encounter DNA from microorganisms that were in an individual’s bloodstream when he or she died and so were the likely cause of death. This approach proved that the bacterium Yersinia pestis was the cause of the fourteenth-to-seventeenth-century CE Black Death,14 the sixth-to-eighth-century CE Justinianic plague of the Roman Empire,15 and an endemic plague that was responsible for at least about 7 percent of deaths in skeletons from burials across the Eurasian steppe after around five thousand years ago.16 Ancient pathogen studies have also revealed the history and origins of ancient leprosy,17 tuberculosis,18 and, in plants, the Irish potato famine.19 Ancient DNA studies are now regularly obtaining material from the microbes that inhabit us, including from dental plaque and feces, providing information about the food our ancestors ate.20 We are only just beginning to mine this new seam of information.

  Taming the Wild West of the Ancient DNA Revolution

  The speed at which the ancient DNA revolution is moving is exhilarating. The technology is evolving so quickly that many papers being published right now use methods that will be obsolete within a few years. Ancient DNA specialists are multiplying—for example, my own laboratory has already graduated three people who have founded their own ancient DNA laboratories. A major trend is specialization. The pioneers of ancient DNA spent a large portion of their time traveling the world to remote locations, talking with archaeologists and local officials, and bringing back unique remains that they have then analyzed in their molecular biology laboratories. Travel to exotic places and a gold rush to obtain key bones are central to this way of doing science. Some in the second generation of ancient DNA research have adopted this model. But others, including myself, travel far less, and instead spend most of our time developing expertise in improved laboratory techniques or statistical analysis, obtaining the samples we study through increasingly equal partnerships with archaeologists and anthropologists.

  Ancient DNA laboratories will also become more specialized. At present, we who are working on ancient DNA have the privilege of doing research on populations from all over the world and from a wide range of times. We are like Robert Hooke turning his microscope to describe an extraordinary array of tiny objects in his book Micrographia, or like explorers in the late eighteenth century, sailing to every corner of the globe. But we have at best a superficial knowledge of the historical and archaeological and linguistic background of any topic we work on, and as knowledge grows, a deeper understanding of each region and the specific questions associated with it will be needed to make progress. Over the next two decades, I expect that ancient DNA specialists will be hired into every serious department of anthropology and archaeology, even history and biology. The professionals hired into these roles will be specialized in studying particular areas—for example, Southeast Asia or northeastern China—and their research will not flit from China to America to Europe to Africa as mine does today.

  Ancient DNA will also go the way of specialization and even professionalization when it comes to setting up service laboratories, analogous to the service laboratories that exist for radiocarbon dating. Ancient DNA service laboratories will screen samples, generate genome-wide data, and provide reports that are easily interpretable, much like those currently provided by commercial personal ancestry testing companies. The reports will determine species, sex, and family relationships, and reveal how newly studied individuals relate to individuals for whom there is previously reported data. The researchers submitting the samples will receive an electronic copy of the data to use in any way they wish. The whole process shouldn’t cost more than twice what radiocarbon dating does.

  Service laboratories will proliferate, but researchers analyzing the data to study population history will never be entirely replaced. Archaeologists interested in learning about ancient populations using DNA will always need to partner with experts in genomics if they wish to use the technology to address any question that has subtlety. Getting information about sex,
species, family relatedness, and ancestry outliers from ancient DNA will eventually be routine. But deeper scientific questions that can be accessed with ancient DNA data—such as how populations mixed and migrated, and how natural selection occurred over time—are unlikely ever to be addressed adequately through standardized reports.

  The future for ancient DNA laboratories that I find appealing is based on a model that has emerged among radiocarbon dating laboratories. For example, the Oxford Radiocarbon Accelerator Unit processes large numbers of samples for a fee, and uses this income stream to support a factory that churns out routine dates and produces data more cheaply, efficiently, and at higher quality than would be possible if its scientists limited themselves to their own questions. But its scientists then piggyback on the juggernaut of the radiocarbon dating factory they have built to do cutting-edge science, such as the study led by Thomas Higham that clarified the record on the demise of Neanderthals in Europe, showing that they disappeared everywhere within a few thousand years of contact with modern humans.21 This is also the model that I learned when I was a postdoctoral scientist at the Massachusetts Institute of Technology at one of the half dozen sequencing centers that carried out the brute-force work for the Human Genome Project, funded by large data production contracts from the U.S. National Institutes of Health. The center’s leader, my supervisor, Eric Lander, also took advantage of the fact that he could turn the power of his sequencing center to address scientific problems that intrigued him. This is my model too: to build a factory, and then to commandeer it to answer deep questions about the past.

  Out of Respect for Ancient Bones

  I first went to Jerusalem when I was seven years old, taken there by my mother along with my older brother and younger sister. We stayed that summer and the next in an apartment that my grandfather owned in a poor, ultra-Orthodox neighborhood populated by men dressed in long black kaftans and women in layered modest dresses and headscarves. The boys attended morning-to-night religious schools, but on Friday afternoons before the Sabbath they were dismissed early and often joined political demonstrations. During the protests, they sometimes set fire to dumpsters and pelted policemen with stones. I remember watching the boys running, cloths pressed to faces, eyes streaming from the tear gas lobbed at them by the police.

  Some of these protests were in response to excavations in the City of David, a site that spills down the hillside of the Temple Mount south of the Old City of Jerusalem, and covers much of the area that became the capital of Judaea after about three thousand years ago. The protesters were upset that the excavations would disturb ancient Jewish graves, an ever-present possibility when digging in Israel. For the protesters, the opening of graves, whether by accident or for scientific investigation, was desecration.

  What would those protesters think of what my laboratory is doing now, grinding through the bones of hundreds of ancient people every month? Perhaps they would not care much about samples from outside Israel, but I think the issue is more general, and I have found myself reflecting more and more about opening up the graves and sampling the remains of any ancient human. It is likely that many of the people whose bones we sample would not have wanted their remains to be used in this way.

  One argument that some ancient DNA specialists and archaeologists have made is that most of the skeletons we are studying are from cultures so remote in time that they have no traceable connection to peoples of the present. This is the standard encoded in law in the U.S. Native American Graves Protection and Repatriation Act, which states that remains should be returned to Native American tribes when there is evidence of a cultural or biological connection to present-day peoples. However, this standard is now breaking down, as exemplified by the approximately 8,500-year-old Kennewick Man skeleton and the approximately 10,600-year-old Spirit Cave skeleton that are being returned to tribes despite having no clear cultural or genetic connections to specific groups living today.22 As we study skeletons that draw ever closer in time to the present, it is important to think about the implications of modern claims on ancient samples. Ancient remains are the remains of real people whose physical integrity we should perhaps only violate if we have good reasons.

  In 2016, I decided to ask a rabbi, in this case my mother’s brother, for counsel. He is Orthodox, which means that he follows the intricate rules specified in the Jewish Oral Tradition. I had a hope that he might be open to my question, as he has also been an advocate of adapting Orthodox Judaism as much as possible to the modern world while abiding by the constraints of its fixed rules, a movement of inclusivity that has been called “Open Orthodoxy”—most recently, he set up a religious seminary to train women as Orthodox rabbis, a role from which women in that community had previously been excluded. I told him that in my lab we were grinding through the bones of ancient peoples, many of whom might not have wanted their remains to be disturbed, and that I felt I had not thought enough about this. He was obviously troubled, and asked me for some time to think. Afterward he came back with the judgment a rabbi gives to provide guidance when there is no precedent set by earlier decisions or judgments made by other rabbis. He said all human graves are sacrosanct, but there are mitigating circumstances that make it permissible to open graves as long as there is potential to promote understanding, to break down barriers between people.

  The study of human variation has not always been a force for good. In Nazi Germany, someone with my expertise at interpreting genetic data would have been tasked with categorizing people by ancestry had that been possible with the science of the 1930s. But in our time, the findings from ancient DNA leave little solace for racist or nationalistic misinterpretation. In this field, the pursuit of truth for its own sake has overwhelmingly had the effect of exploding stereotypes, undercutting prejudice, and highlighting the connections among peoples not previously known to be related. I am optimistic that the direction of my work and that of my colleagues is to promote understanding, and I welcome our opportunity to do our best by the people, ancient and modern, whom we have been given the privilege to study. I see it as our role to midwife ancient DNA into a field that is not only the domain of geneticists, but also of archaeologists and the public—to realize its extraordinary potential to reveal who we are.

  Notes on the Illustrations

  Map sources. All maps were made with data from Natural Earth (http://www.naturalearthdata.com/​).

  Figure 1. Contours in panel (a) are based on Fig. 2A of L. L. Cavalli-Sforza, P. Menozzi, and A. Piazza, “Demic Expansions and Human Evolution,” Science 259 (1993): 639–46. Contours in panel (b) are based on interpolation of the numbers shown in Fig. 3 of W. Haak et al., “Massive Migration from the Steppe Was a Source for Indo-European Languages in Europe,” Nature 522 (2015): 207–11. The interpolation was performed using the POPSutilities.R software of F. Jay et al., “Forecasting Changes in Population Genetic Structure of Alpine Plants in Response to Global Warming, Molecular Ecology (2012): 2354–68 and the parameter settings recommended in O. François, “Running Structure-like Population Genetic Analyses with R,” June 2016, http://membres-timc.imag.fr/​Olivier.Francois/​tutoRstructure.pdf.

  Figure 2. The plot shows the 3,748 unique individuals in the author’s internal laboratory database as of November 19, 2017, broken down by the year when they became available.

  Figure 4. The number of genealogical ancestors expected to have contributed DNA to a person living today is based on simulation results shared with the author by Graham Coop. The simulations were performed as described in G. Coop, “How Many Genetic Ancestors Do I Have,” gcbias blog, November 11, 2013, https://gcbias.org/​2013/​11/​11/​how-does-your-number-of-genetic-ancestors-grow-back-over-time/.

  Figure 5. The number of mutations in a given segment that separate the genome a person receives from his or her father and the one he or she receives from his or her mother can be used to estimate how much time has elapsed since the common ancestor at that location in the genome. Panel (2), which is based on the
analyses reported in S. Mallick et al., “The Simons Genome Diversity Project: 300 Genomes from 142 Diverse Populations,” Nature 538 (2016): 201–6, shows the estimated times since the most recent shared ancestor averaged across 250 non-African genome pairs (solid line), and 44 sub-Saharan African genome pairs, measured at equally spaced locations in the DNA. Panel (3) shows the maximum estimated time at each location in the genome over 299 genome pairs and is based on analyses from the same study.

  Figure 6. The approximate range of the Neanderthals is adapted from Fig. 1 of J. Krause et al., “Neanderthals in Central Asia and Siberia,” Nature 449 (2007): 902–4.

  Figure 7. The counts of shared mutations are based on the French-San-Neanderthal comparison in Table S48 of the Supplementary Online Materials of R. E. Green et al., “A Draft Sequence of the Neandertal Genome,” Science 328 (2010): 710–22.

  Figure 8. The illustration is based on the data in Fig. 2 of Q. Fu et al., “An Early Modern Human from Romania with a Recent Neanderthal Ancestor,” Nature 524 (2015): 216–19.

  Figure 9. This illustration replots the data shown in Fig. 2. of Q. Fu et al., “The Genetic History of Ice Age Europe,” Nature 534 (2016): 200–5.

  Figure 10. The pie chart data come from columns AJ and AK of Supplementary Table 2 of S. Mallick et al., “The Simons Genome Diversity Project: 300 Genomes from 142 Diverse Populations,” Nature 538 (2016): 201–6. Each population is represented by an average of the individuals in that population. The proportion of archaic ancestry is expressed as a fraction of the maximum seen in any population in the dataset. Numbers less than 0.03 are set to 0 and numbers greater than 0.97 are set to 1. A subset of 47 populations is plotted to highlight the geographic coverage while reducing visual clutter.