by David Reich
Before diving into the book, I will recount something that happened during a guest lecture I gave at the Massachusetts Institute of Technology in 2009. Mine was one of the last lectures of the term, meant to add spice to a course aimed at introducing students to computer-aided research into genomes with the goal of finding cures for disease. As I addressed Indian population history, an undergraduate sitting at the center of the front row stared me down. When I concluded, she asked me, with a grin, “How do you get funded to do this stuff?”
I mumbled something about how the human past shapes genetic variation, and about how, in order to identify risk factors for disease, it is important to understand that past. I gave the example of how among the thousands of distinctive human populations of India, there are high rates of disease because mutations that happened to be carried by the founders increased in frequency as the groups expanded. I make arguments along these lines in my applications to the U.S. National Institutes of Health, in which I propose to find disease risk factors that occur at different frequencies across populations. Grants of this type have funded much of my work since I started my laboratory in 2003.
True as these arguments are, I wish I had responded differently. We scientists are conditioned by the system of research funding to justify what we do in terms of practical application to health or technology. But shouldn’t intrinsic curiosity be valued for itself? Shouldn’t fundamental inquiry into who we are be the pinnacle of what we as a species hope to achieve? Isn’t an attribute of an enlightened society that it values intellectual activity that may not have immediate economic or other practical impact? The study of the human past—as of art, music, literature, or cosmology—is vital because it makes us aware of aspects of our common condition that are profoundly important and that we heretofore never imagined.
Part I
The Deep History of Our Species
1
How the Genome Explains Who We Are
The Master Chronicle of Human Variation
To understand why genetics is able to shed light on the human past, it is necessary to understand how the genome—defined as the full set of genetic code each of us inherits from our parents—records information. Francis Crick, Rosalind Franklin, James Watson, and Maurice Wilkins showed in 1953 that the genome is written out in twin chains of about three billion chemical building blocks (six billion in all) that can be thought of as the letters of an alphabet: A (adenine), C (cytosine), G (guanine), and T (thymine).1 What we call a “gene” consists of tiny fragments of these chains, typically around one thousand letters long, which are used as templates to assemble the proteins that do most of the work in cells. In between the genes is noncoding DNA, sometimes referred to as “junk” DNA. The order of the letters can be read by machines that perform chemical reactions on fragments of DNA, releasing flashes of light as the reactions pass along the length of the DNA sequence. The reactions emit a different color for each of the letters A, C, G, and T, so that the sequence of letters can be scanned into a computer by a camera.
Although the great majority of scientists are focused on the biological information that is contained within the genes, there are also occasional differences between DNA sequences. These differences are due to random errors in copying of genomes (known as mutations) that occurred at some point in the past. It is these differences, occurring about one every thousand letters or so in both genes and in “junk,” that geneticists study to learn about the past. Over the approximately three billion letters, there are typically around three million differences between unrelated genomes. The higher the density of differences separating two genomes on any segment, the longer it has been since the segments shared a common ancestor as the mutations accumulate at a more or less constant rate over time. So the density of differences provides a biological stopwatch, a record of how long it has been since key events occurred in the past.
Figure 3. The genome contains about three billion nucleotides, which can be thought of as four letters in a biological alphabet: adenine (A), cytosine (C), guanine (G), and thymine (T). Around 99.9 percent of these letters are identical across two lined-up genomes, but in that last ~0.1 percent there are differences, reflecting mutations that accumulate over time. These mutations tell us how closely related two people are and record exquisitely precise information about the past.
The first startling application of genetics to the study of the past involved mitochondrial DNA. This is a tiny portion of the genome—only approximately 1/200,000th of it—which is passed down along the maternal line from mother to daughter to granddaughter. In 1987, Allan Wilson and his colleagues sequenced a few hundred letters of mitochondrial DNA from diverse people around the world. By comparing the mutations that were different among these sequences, he and his colleagues were able to reconstruct a family tree of maternal relationships. What they found is that the deepest branch of the tree—the branch that left the main trunk earliest—is found today only in people of sub-Saharan African ancestry, suggesting that the ancestors of modern humans lived in Africa. In contrast, all non-Africans today descend from a later branch of the tree.2 This finding became an important part of the triumphant synthesis of archaeological and genetic and skeletal evidence that emerged in the 1980s and 1990s for the theory that modern humans descend from ancestors who lived in the last hundred thousand years or so in Africa. Based on the rate at which mutations are known to accumulate, Wilson and his colleagues estimated that the most recent African ancestor of all the branches, “Mitochondrial Eve,” lived sometime after 200,000 years ago.3 The best current estimate is around 160,000 years ago, although it is important to realize that like most genetic dates, this one is imprecise because of uncertainty about the true rate at which human mutations occur.4
The finding of such a recent common ancestor was exciting because it refuted the “multiregional hypothesis,” according to which present-day humans living in many parts of Africa and Eurasia descend substantially from an early dispersal (at least 1.8 million years ago) of Homo erectus, a species that made crude stone tools and had a brain about two-thirds the size of ours. The multiregional hypothesis implied that descendants of Homo erectus evolved in parallel across Africa and Eurasia to give rise to the populations that live in the same places today. The multiregional hypothesis would therefore predict that there would be mitochondrial DNA sequences among present-day people that are separated by close to two million years, the age of the dispersal of Homo erectus. However, the genetic data was impossible to reconcile with this prediction. The fact that all people today share a common mitochondrial DNA ancestor about ten times more recently showed that humans today largely descend from a much later expansion from Africa.
Anthropological evidence pointed to a likely scenario for what occurred. The earliest human skeletons with “anatomically modern” features—defined as falling within the range of variation of all humans today with regard to having a globular brain case and other traits—date up to two hundred to three hundred thousand years ago and are all from Africa.5 Outside of Africa and the Near East, though, there is no convincing evidence of anatomically modern humans older than a hundred thousand years and very limited evidence older than around fifty thousand years.6 Archaeological evidence of stone tool types also points to a great change after around fifty thousand years ago, a period known to archaeologists of West Eurasia as the Upper Paleolithic, and to archaeologists of Africa as the Later Stone Age. After this time, the technology for manufacturing stone tools became very different, and there were changes in style every few thousand years, compared to the glacial earlier pace of change. Humans in this period also began to leave behind far more artifacts that revealed their aesthetic and spiritual lives: beads made of ostrich eggshells, polished stone bracelets, body paint made from red iron oxide, and the world’s first representational art. The world’s earliest known figurine is a roughly forty-thousand-year-old “lion-man” carved from a woolly mammoth tusk, found in Hohlenstein-Stadel in Germany.7 The approximately
thirty-thousand-year-old drawings of pre–ice age beasts, found on the walls of Chauvet Cave in France, even today are recognizable as transcendent art.
The dramatic acceleration of change in the archaeological record after around fifty thousand years ago was also reflected by evidence of population change. The Neanderthals, who had evolved in Europe by around four hundred thousand years ago and are considered “archaic” in the sense that their skeletal shape did not fall within present-day human variation, went extinct in their last holdout of western Europe between about forty-one thousand and thirty-nine thousand years ago, within a few thousand years of the arrival of modern humans.8 Population turnovers also occurred elsewhere in Eurasia, as well as in southern Africa, where there is evidence of abandonment of sites and the sudden appearance of Later Stone Age cultures.9
The natural explanation for all these changes was the spread of an anatomically modern human population whose ancestors included “Mitochondrial Eve,” who practiced a sophisticated new culture, and who largely replaced the people who lived in each place before.
The Siren Call of the Genetic Switch
The finding that genetics could help to distinguish between competing hypotheses of human origins led in the 1980s and 1990s to exuberance about the power of the discipline to provide simple explanations. Some even wondered if genetics might be able to do more than provide a supporting line of evidence for the spread of modern humans from Africa and the Near East after around fifty thousand years ago. Perhaps genes could also be the cause of that spread, offering an explanation as simple and beautiful as the four-letter code written in DNA for the quickening pace of change in the archaeological record.
The anthropologist best known for embracing the idea that a genetic change might explain how we came to be behaviorally distinct from our predecessors was Richard Klein. He put forward the idea that the Later Stone Age revolution of Africa and the Upper Paleolithic revolution of western Eurasia, when recognizably modern human behavior burst into full flower after about fifty thousand years ago, were driven by the rise in frequency of a single mutation of a gene affecting the biology of the brain, which permitted the manufacture of innovative tools and the development of complex behavior.
According to Klein’s theory, the rise in frequency of this mutation primed humans for some enabling trait, such as the ability to use conceptual language. Klein thought that prior to the occurrence of this mutation, humans were incapable of modern behaviors. Supporting his notion are examples among other species of a small number of genetic changes that have effected major adaptations, such as the five changes that are sufficient to turn the tiny ears of the Mexican wild grass teosinte into the huge cobs of corn that we buy in the supermarket today.10
Klein’s hypothesis came under intense criticism almost as soon as he suggested it, most notably from the archaeologists Sally McBrearty and Alison Brooks, who showed that almost every trait that Klein considered to be a hallmark of distinctly modern human behavior was evident in the African and Near Eastern archaeological records tens of thousands of years before the Upper Paleolithic and Later Stone Age transitions.11 But even if no single behavior was new, Klein had put his finger on something important. The intensification of evidence for modern human behavior after fifty thousand years ago is undeniable, and raises the question of whether biological change contributed to it.
One geneticist who came of age at this time of exuberance about the power of genetics to provide simple explanations for great mysteries was Svante Pääbo, who arrived in Allan Wilson’s laboratory just after the “Mitochondrial Eve” discovery, and who would go on to invent much of the toolkit of the ancient DNA revolution and to sequence the Neanderthal genome. In 2002, Pääbo and his colleagues discovered two mutations in the gene FOXP2 that seemed to be candidates for propelling the great changes that occurred after around fifty thousand years ago. The previous year, medical geneticists had identified FOXP2 as a gene that, when mutated, produces an extraordinary syndrome whose sufferers have normal-range cognitive capabilities, but cannot use complex language, including most grammar.12 Pääbo and his colleagues showed that the protein produced by the FOXP2 gene has remained almost identical during the more than hundred million years of evolution separating chimpanzees and mice. However, two changes to the protein occurred on just the human lineage since it branched out of the common ancestral population of humans and chimpanzees, showing that the gene had evolved much more rapidly on the human lineage.13 Later work by Pääbo and his colleagues found that engineered mice with the human versions of FOXP2 are identical to regular mice in most respects, but squeak differently, consistent with the idea that these changes affect the formation of sounds.14 These two mutations at FOXP2 cannot have contributed to the changes after fifty thousand years ago, since Neanderthals shared them,15 but Pääbo and his colleagues later identified a third mutation that is found in almost all present-day humans and that affects when and in what cells FOXP2 gets turned into protein. This change is absent in Neanderthals, and thus is a candidate for contributing to the evolution of modern humans after their separation from Neanderthals hundreds of thousands of years ago.16
Regardless of how important FOXP2 itself is in modern human biology, Pääbo cites the search for the genetic basis for modern human behavior as a justification for sequencing the genomes of archaic humans.17 Between 2010 and 2013, when he led a series of studies that published whole-genome sequences from archaic humans like Neanderthals, Pääbo’s papers highlighted an evolving list of about one hundred thousand places in the genome where nearly all present-day humans carry genetic changes that are absent in Neanderthals.18 There are surely biologically important changes hiding in the list, but we are still only at the very beginning of the process of determining what they are, reflecting a more general problem that we are like kindergartners in our ability to read the genome. While we have learned to decode the individual words—as we know how the sequence of DNA letters gets turned into proteins—we still can’t parse the sentences.
The sad truth is that it is possible to count on the fingers of two hands the examples like FOXP2 of mutations that increased in frequency in human ancestors under the pressure of natural selection and whose functions we partly understand. In each of these cases, the insights only came from years of hand-to-hand combat with life’s secrets by graduate students or postdoctoral scientists making engineered mice or fish, suggesting that it will take an evolutionary Manhattan Project to understand the function of each mutation that we have and that Neanderthals do not. This Manhattan Project of human evolutionary biology is one to which we as a species should commit ourselves. But even when it is carried out, I expect that the findings will be so complicated—with so many individual genetic changes contributing to what makes humans distinctive—that few people will find the answer comprehensible. While the scientific question is profoundly important, I expect that no intellectually elegant and emotionally satisfying molecular explanation for behavioral modernity will ever be found.
But even if studying just a few locations in the genome will not provide a satisfying explanation for how modern human behavior evolved, the great surprise of the genome revolution is the explanations it is starting to provide from another perspective—that of history. By comprehending the entire genome—by going beyond the tiny slice of the past sampled by our mitochondrial DNA and Y chromosome and embracing the story of our past told by the multiplicity of our ancestors that is written in the record of our whole genome—we have already begun to sketch out a new picture of how we got to be the way we are. This explanation based on migrations and population mixture is the subject of this book.
One Hundred Thousand Adams and Eves
When the journalist Roger Lewin in 1987 dubbed the common maternal ancestor of all people living today “Mitochondrial Eve,” he evoked a creation story—that of a woman who was the mother of us all, and whose descendants dispersed throughout the earth.19 The name captured the collective imagination, and is
still used not only by the public but also by many scientists to refer to this common maternal ancestor. But the name has been more misleading than helpful. It has fostered the mistaken impression that all of our DNA comes from precisely two ancestors and that to learn about our history it would be sufficient to simply track the purely maternal line represented by mitochondrial DNA, and the purely paternal line represented by the Y chromosome. Inspired by this possibility, the National Geographic Society’s “Genographic Project,” beginning in 2005, collected mitochondrial DNA and Y-chromosome data from close to a million people of diverse ethnic groups. But the project was outdated even before it began. It has been largely recreational, and has produced few interesting scientific results. From the outset, it was clear that most of the information about the human past present in mitochondrial DNA and Y-chromosome data had already been mined, and that far richer stories were buried in the whole genome.
The truth is that the genome contains the stories of many diverse ancestors—tens of thousands of independent genealogical lineages, not just the two whose stories can be traced with the Y chromosome and mitochondrial DNA. To understand this, one needs to realize that beyond mitochondrial DNA, the genome is not one continuous sequence from a single ancestor but is instead a mosaic. Forty-six of the mosaic tiles, as it were, are chromosomes—long stretches of DNA that are physically separated in the cell. A genome consists of twenty-three chromosomes, and because a person carries two genomes, one from each parent, the total number is forty-six.
But the chromosomes themselves are mosaics of even smaller tiles. For example, the first third of a chromosome a woman passes down to her egg might come from her father and the last two-thirds from her mother, the result of a splicing together of her father’s and mother’s copies of that chromosome in her ovaries. Females create an average of about forty-five new splices when producing eggs, while males create about twenty-six splices when producing sperm, for a total of about seventy-one new splices per generation.20 So it is that as we trace each generation back further into the past, a person’s genome is derived from an ever-increasing number of spliced-together ancestral fragments.
