Dna: The Secret of Life

by Watson, James

  Cann, Wilson, and their colleagues had changed fundamentally the way we understand our human past.

  Subsequent research has confirmed Cann and Wilson's conclusion. Much of the newer work has come out of the Stanford laboratory of Luigi Luca Cavalli-Sforza, who pioneered the application of genetic approaches to anthropological problems. Raised in a distinguished Milanese family, Cavalli-Sforza was fascinated with microscopes. And in 1938, he enrolled as a precocious sixteen-year-old in medical school at the University of Pavia. "It turned out to be a very lucky choice," he notes: the alternative would have been service in Mussolini's army. When I first met him in 1951, Cavalli-Sforza was still an up-and-coming bacterial geneticist. But a chance remark made by a graduate student would inspire a turn away from the genetics of bacteria toward the genetics of humans. The graduate student, who had trained to be a priest, mentioned that the Catholic Church had kept detailed records of marriages over the past three centuries. Realizing that in these records there lurked a wealth of research possibilities, Cavalli-Sforza began to apply himself more and more to human genetics, and he probably remains one of a very few human geneticists who can legitimately claim to have found their vocation via the Church.

  Cavalli-Sforza understood that the most convincing confirmation of Cann and Wilson's assertions about human evolution would ideally come from genes only transmitted from father to son, i.e., some component of the human genome passed down through the male line. If one could arrive at their conclusions tracking the male lineage – taking a patrilineal route as opposed to the matrilineal path Cann and Wilson found through mtDNA analysis – one could be assured of a truly independent corroboration. The male-specific component of the genome is, of course, the Y chromosome. By definition, the possessor of a Y is male (the Y chromosome, remember, is inherited by men from their fathers, whose sperm cells can contain either an X or a Y; upon fusing with the egg cell, which always contains an X, the sperm thus determines our sex, XX combinations producing females and XY males). The Y chromosome, then, holds the key to the genetic history of men. In addition, because recombination occurs only between paired chromosomes, the use of the Y allows us to avoid that dreaded pitfall of evolutionary analysis, recombination: a Y is unique whenever it is present, and so there is never a matching Y with which it might trade material.

  In a blockbuster paper published in 2000, Cavalli-Sforza's colleague Peter Underhill did for the Y chromosome what Cann and Wilson had done for mtDNA. The findings were strikingly similar. Again the family tree was found to be rooted in Africa, and again it was shown to be remarkably shallow: not the ancient mighty oak imagined by anthropologists, but the shrub of Cann and Wilson's analysis, around 150,000 years old.

  The existence of two independent data-sets yielding a similar picture of the human past is extremely compelling. When only one region, say mtDNA, is studied, the results, while suggestive, are still inconclusive; the pattern may simply reflect the peculiarities of the history of that particular region of DNA, rather than the impact of some major historical event on our species as a whole. Critically, the point at which a family tree converges – the most recent common ancestor of all the sequences in the study, that great-great-. . . -grandfather/mother of us all – is not necessarily associated with any particular event in human history. Though it may connote the origin of our species or some other historically significant demographic episode, it may just as likely signify something much more trivial from the point of view of human history – perhaps nothing more, say, than the effect of past natural selection on mtDNA. If, however, the same pattern of change can be observed in more than one region of the genome, the chances are that one has indeed found the genetic footprint of an important past event.

  To better understand how natural selection can affect patterns of genetic variation (and the overall age of a family tree), imagine the following scenario: 150,000 years ago, the tribe of protohumans boasted a plethora of mtDNA sequences, just as our species does today, but then a beneficial mutation – one favored by natural selection – arose on one of those sequences. The mutation would increase in frequency until, after many generations, every member of the species would have it. Because there is no recombination in mitochondria, no exchange between mtDNAs, the selective process would affect the entire sequence in which a favored mutation first appeared, so every member of the species would end up with the same mtDNA sequence. So by the time that natural selection has finished its job and every individual possesses the favored mutation, there would be no mtDNA genetic variation in the species. Gradually over subsequent years, though, mutations would occur and variation would build up again, but all these new mtDNA sequences would ultimately be descended from that single sequence: the family tree's convergence point, the most recent common ancestor of all the sequences. The pattern would be exactly what Cann and Wilson found, but in this case the convergence point represents nothing more than an episode of evolution's fine-tuning of mtDNA.

  This was the ambiguity that dogged Cann and Wilson's result: Was it produced by evolutionary tinkering, or by something much more significant in the overall scheme of human prehistory? But when Underhill observed a similar pattern for the Y chromosome, that ambiguity vanished. The coincidence suggested forcefully that at the moment in question (150,000 years ago), human populations did indeed undergo a radical genetic alteration, one capable of affecting mtDNAs and Y chromosomes simultaneously. The phenomenon involved, to which we shall turn in a moment, is called a "genetic bottleneck."

  How can demographic factors affect a family tree? Any genealogy is the outcome of the waxing and waning of the lineages composing it: over time, some will thrive and others become extinct. Think of surnames. Assume that a thousand years ago on some remote island everyone had one of three surnames: Smith, Brown, and Watson. Assume, too, that small errors of transcription – "mutations" – occasionally occurred when the names of newborns were inscribed in the birth registers. The errors are infrequent and slight, so we can still tell which of the original names the altered forms derived from: "Browne" is clearly a mutation of "Brown." Now let us imagine that in the population today, a thousand years later, we find that everyone is called Brown, Browne, Bowne, Frown, or Broun. Smith and Watson have gone extinct while the Brown line has thrived (and diversified through mutation). What has happened? Pure chance has led to the loss of the Smith and Watson lines. Perhaps, for instance, several Mr. & Mrs. Smiths of one generation managed to produce mainly daughters. Assume (in accordance with tradition, though not the modern alternative convention) that surnames are transmitted along the male line; the bumper crop of daughters would thus have the effect of reducing the representation of Smiths in the next generation. Now say that the new generation of Smiths also overproduced daughters, and the demographic effect was heightened once again – well, you get the picture: eventually, the Smith name disappeared altogether. So did Watson.

  This kind of random extinction is, in fact, statistically inevitable. Usually, however, it happens so slowly that its impact can be felt only over huge periods of time. Sometimes, though, a bottleneck – a period of very much reduced population size – will massively accelerate the process. With only three couples (six individuals) on the island at the beginning of its population history, it was reasonably likely that we would lose Smith and Watson within a single generation, the chances being fairly good that both the Smiths and the Watsons would have only daughters, or fail to procreate at all. In a large population, such abrupt disappearances of lines cannot occur; it is statistically inconceivable, given a population with many Smith couples, that they could all wind up producing only girls or simply fail to have children. Only over the course of many generations would the effects of the dwindling ranks gradually mount up. Indeed a real-life example of this hypothetical name-extinction process actually occurred in the South Pacific, when the six Bounty mutineers colonized Pitcairn Island with their thirteen Tahitian brides. Within seven generations, the number of surnames had dwindled to thr
ee.

  When we look today at the surnames in our theoretical population, Brown, Browne, Bowne, Frown, or Broun, we can infer that they are all descended from just one of the three starting lineages, Brown. And so the implication of the human mtDNA and Y chromosome data should hardly surprise us: 150,000 years ago there were many different mtDNA sequences and many different Y chromosome sequences, but today's sequences are all descended from just one of each. All the others went extinct, most probably disappearing during some ancient bottleneck event – a population crash caused by plague, a change in climate, what have you. But whatever this cataclysmic event in our early history, one thing is clear: some time afterward, groups of our ancestors started to head out of Africa, beginning the epic saga of the human colonization of the planet (see Plate 50).

  Another interesting finding confirmed by both the mtDNA and Y chromosome data is the position on the human family tree of the San of southern Africa.* Theirs is the longest, and therefore the oldest, branch on the tree. This by no means implies that they are more "primitive" than the rest of us: every human is at the same evolutionary and molecular remove from our closest relatives among the great apes. If we trace lineages back to the last common ancestor of both chimpanzees and humans, my lineage is about 5 million years old, and so is a San's. In fact, our two lineages are the same for most of those eons; only 150,000 years ago did the San lineage separate from other human lines.

  * The San are also known as Bushmen (Sanqua, in Dutch), a derogatory term given to them by Dutch settlers in the late seventeenth century.

  It appears, from the genetic evidence, that after an initial migration into southern and eastern Africa, the San remained relatively isolated throughout history. This pattern is borne out by sociolinguistics when we consider the distribution of the San's unusual (at least to my ears) "click" languages. Their current distribution is extremely limited owing to the expansion of Bantu-speaking people from west central Africa starting about 1,500 years ago. The Bantu expansion displaced the San to marginal environments like the Kalahari Desert.

  Given their relatively stable history, do the San provide a snapshot of what the ancestors of all modern humans were like? Possibly, but not necessarily – substantial change may well have occurred along the San lineage over the past 150,000 years. Even inferences from the San about our early ancestors' ways of living are questionable: the San's present lifestyle is an adaptation to the harsh desert environment to which they have been confined since the relatively recent arrival of the Bantu speakers. In 2000 I experienced the unique thrill of living for several days in a San community in the Kalahari. I was struck by their remarkable pragmatism, their efficient no-nonsense way of taking on all tasks before them, even those outside their normal experience, like fixing a flat tire. I found myself wishing that more of my colleagues were likewise adaptable. And if, in genetic terms, these people are as genetically "different" from me as any on the planet, I could not fail to be impressed by just how like-minded we were.

  The genetic and cultural uniqueness of the San will disappear shortly. Young people in the Kalahari show little desire to continue the simple hunter-gatherer lifestyle of their nomadic parents. When, for instance, the group I visited staged a traditional "trance dance," the younger members were visibly embarrassed by their elders' antics. They will move away from their communities and marry into other groups.

  In fact, history has already recorded a trend toward mixture between the San and other groups. Nelson Mandela's Xhosa tribe, for one, represents a biological mix of Bantu and San peoples, as the Xhosa language, though Bantu-based, reveals in its many typically San clicks. In our technologically accelerated day and age, it is unlikely that the genetic and cultural integrity of the San will survive much longer. It is, therefore, fortunate indeed that considerable efforts have been made over the past few decades to understand and document this unique people and their way of life. Philip Tobias of the University of Witwatersrand in Johannesburg both initiated these studies and, for many years, championed the San as an unofficial spokesman during the dark days of apartheid. And Trefor Jenkins, a voluble Welshman who arrived in South Africa after working as a doctor in Zambian copper-mining towns, has long spearheaded genetic studies of the San and other indigenous groups.

  Sadly it currently remains beyond the reach of even the most sophisticated genetic methods to elucidate the origins of human culture. Archaeological evidence shows that our ancestors were up to much the same activities as other hominids, Neanderthals included, during the first phase of their evolution. Indeed, a cave site at Skhul in Israel offers proof that about 100,000 years ago populations of Homo sapiens and Homo neanderthalensis coexisted, neither apparently endangering the other. But, as we have seen, modern humans subsequently wiped out their heavy-browed cousins around 30,000 years ago. It therefore seems likely that in the intervening 70,000-year period modern humans, through technological and/or cultural advances, somehow acquired the edge.

  Independent archaeological information supports this hypothesis. It would appear that, around 50,000 years ago, modern humans suddenly became culturally modern: we see in the remains from this time the first indisputable ornaments, the first routine use of bone, ivory, and shell to produce familiar useful artifacts, and the first of many improvements in hunting and gathering technology. What happened? We shall probably never know. But one is tempted to speculate that it was the invention of language that made all of this – and all we have accomplished since – possible.

  Prehistory by definition refers to the period prior to written records, and yet we find written in every individual's DNA sequences a record of our ancestors' respective journeys. The new science of molecular anthropology uses patterns of genetic variation among different groups to reconstruct this history of human colonization. Human "prehistory" has thus become accessible.

  Studies of the distribution of genetic variation across the continents combined with archaeological information have revealed some details of our ancestors' global expansion. The journey along the fringes of Asia and through the archipelagoes of modern Indonesia to New Guinea and Australia was accomplished by about 60,000 years ago. Getting to Australia required crossing several substantial bodies of water, suggesting that our ancestors were already using boats at that early stage. Modern humans arrived in Europe around 40,000 years ago, and penetrated northern Asia, including Japan, some 10,000 years later.

  Like so many other leaders in this field (including Rebecca Cann and Svante Pääbo), Michael Hammer, at the University of Arizona, received his training in Allan Wilson's Berkeley lab. And though Hammer's initial interest was mice, the publication of Cann and Wilson's mtDNA study diverted him from rodents to the human past. He was among the first to realize that information from the Y chromosome would provide the crucial test of Cann and Wilson's overall hypothesis. But the Y proved reluctant at first to yield its secrets. One study (done in Wally Gilbert's lab) sequenced the same chunk of DNA drawn from multiple individuals, only to find the sequence identical in every instance – a laborious effort that yielded zero information about genetic interrelations. Hammer persisted, however, and eventually he and others turned the Y chromosome into an anthropological gold mine, whose payoff culminated in Underbill's landmark paper.

  A major vein in the Y chromosome mine has enriched our attempts to reconstruct the human colonization of the New World, a relatively late development. The identity of the oldest human settlement in the Americas remains contentious: a site in Clovis, New Mexico, is the traditional titleholder, dating back some 11,200 years; but fans of a site in Monte Verde, Chile, claim it to be at least 12,500 years old. It is also debated whether the first Amerindians crossed a land bridge across the Bering Strait during the last Ice Age or took a more southerly route in boats. What the genetic data make clear, however, is that the founding group was small: with only two major classes of Y chromosome sequences detected, there appear to have been just two distinct arrivals, each perhaps involving no more than
a single family. Among Amerindians mtDNA variation is much more extensive than Y chromosome variation, suggesting that there were more women than men in each founding group. Probably the more common of the two Y chromosome sequences represents the first arrival; the descendant population would then already have been established before the arrival of the second group, which included the ancestors of today's Navajo and Apache. The more common sequence also boasts another distinction: the presence (first noted in 2002) of a mutation that is rarely found elsewhere on the planet. Giving further evidence of its bearers' precedence as pioneers, this mutation is calculated to be about 15,000 years old, not much older than the earliest known archaeological sites.

  Genetic analyses have permitted the reconstruction of more recent phases of prehistory as well. Hammer, for example, has shown that modern Japanese are a mix of the Jomon ancient hunter-gatherers, currently represented by Japan's aboriginal Ainu population, and relatively recent immigrants, the Yayoi, who arrived about 2,500 years ago from the Korean peninsula, bringing with them weaving, metalworking, and rice-based agriculture. In Europe, too, we see evidence of waves of migration, often associated with advances in agricultural technology. Groups like the Basques (who live in the mountainous Pyrenees on the French-Spanish border) and the Celts (who arrived later and are found throughout the northwest margin of Europe, from Brittany in France through Ireland and western Britain) are genetically distinct from the rest of Europe. One explanation is that each of these groups was displaced to relatively far-flung regions by more recent arrivals.