DNA USA

by Bryan Sykes

  The San Bushmen of southern Africa have a remarkable history. They number about ninety thousand and now find themselves in three modern countries: Botswana, Namibia, and Zambia. Among the last true hunter-gatherers, they live in spiritual union with their surroundings and with their prey, mainly the antelope of that parched region. Their language, Khoisan, is like no other, with a repertoire of sharp clicks that is almost impossible for outsiders to imitate. The San are on the edge, having been displaced from much of their former territory by diamond mining and tourism interests. In a reminder of the American government’s treatment of Native Americans in the nineteenth century, the San have been forcibly removed from their traditional lands into newly created “settlements.” They have naturally objected to this and, helped by supporters from around the world, have several court cases pending. Also, the United Nations Human Rights Council has criticized the government of Botswana for preventing their return.

  The majority of the San are in the two clusters of L1D and L1K, which are among the most diverse in Africa. This alone gives an indication of the great antiquity of the San, but the presence of a small percentage of L1D among the Bantu speakers of southeastern Africa is a reminder that the San were once far more widespread and have been absorbed into the great flow of the Bantu agricultural expansion. This is one of the most intense and sustained intracontinental migrations not only in Africa but anywhere in the world. However, unlike in Europe, where the diffusion of farming from its origins in the Near East was largely a cultural transmission of agricultural skills, both linguists and geneticists agree that in Africa there was indeed an overwhelming migration of people.

  In Africa as elsewhere in the world, agriculture allows a given piece of fertile land to sustain a higher population than does the hunter-gatherer lifestyle of the San and others. Farming began in earnest in Cameroon and Nigeria about five thousand years ago, and this is where the Bantu expansion began. It is a familiar story as pressure on space stimulated an expansion and, about 3,500 years ago, the farmers began to spread east and south into the lands of the Bushmen. Then as now the dense equatorial forest of the Congo was a formidable barrier, and the expansion slowly skirted its northern and western fringes. Agricultural settlements appeared around the African great lakes of Victoria, Turkana, and Malawi to the east of the forest, while the western route was largely coastal, probably by boat as the offshore island of Bioko (formerly Fernando Póo), in the Gulf of Guinea, was one of the first to be settled.

  Bantu, the language of the farmers, is very different from that of the Khoisan-speaking hunter-gatherers, and it belongs to a completely different family. Where there was open conflict between the Bantu and the San, and with the Biaka and Mbuti Pygmies of the central forest, the iron weapons of the expanding agriculturalists proved superior to the stone of the indigenous hunters, just as European rifles were a crucial factor in the progress of the Indian wars in America. Those San who were not killed were either displaced to their current range, the Kalahari Desert in Southwest Africa, or absorbed into the Bantu settlements. They took their mitochondrial DNA with them, but in a tellingly familiar reflection of the way of the world, not often their Y chromosomes. As usual, it was the women who were “absorbed,” not the men. While the San group L1D, with L1K, makes up nearly three-quarters of the Bushmen’s mitochondrial DNA, only about 5 percent of Bantu speakers are in that clan, so the “absorption” was numerically relatively minor.

  I am well aware that this kind of description, of phalanxes of Africans on the move with all the trappings of a grand plan, is far removed from the reality. There was no plan and no way of implementing one. All I have given here is a crude factual summary, lacking in any emotion. The expansion of the Bantu speakers happened at a time way before the days of organized armies or command from any central authority. Small bands of people moved slowly, taking with them the seeds and tools of agriculture. Plots of land were cleared and planted. Encounters with the indigenous hunters may sometimes have been violent, although the experience in Europe was relatively peaceful, and many of the hunter-gatherer native Europeans picked up the ways of cultivation from the incoming farmers but also carried on hunting.

  At this stage they had no domesticated animals, but by the time the Bantu reached the great lakes of the East African Rift Valley, they had learned the ways of the pastoral people living in what is now Kenya. As the expansion turned toward southern Africa, it looks from the genetics as if they were joined by East Africans. You can see this by the way that another of the L1 clusters is geographically distributed. The cluster L1A is found in both East and Southeast Africa, and if we use the accumulated diversity as an indicator of its age, it probably started in East Africa about forty thousand years ago, well before the Bantu expansion. Two branches then evolved, one of which, L1A2, moved into the equatorial forest, where it is still found among the Mbuti of the Democratic Republic of Congo and the Biaka of the Central African Republic. The other branch, L1A1, is found throughout Kenya and Ethiopia, where it is also very diverse. However, only a limited number of lineages reached Tanzania, Malawi, and Mozambique, farther south. The most straightforward explanation is that they found their way there much more recently during the later phases of the Bantu expansion. There are some members of the clan in West Africa, but that has a separate explanation that we will come to in a moment.

  The other major branches within the L1 superclan have a completely different geographical distribution within Africa. Cluster L1B is concentrated in West Africa, being virtually absent from the east and southeast. There is some L1B in equatorial Central Africa, where about 5 percent of inhabitants belong to the clan. They are far more varied than in West Africa, which suggests that the clan began in Central Africa and then expanded into West Africa about thirty thousand years ago, judging by its age there. The same goes for the sister cluster, L1C, which is both older and, at 23 percent, more frequent in Central Africa than farther west.

  While Mitochondrial Eve is the matriarch of the L1 superclan, making it the world’s oldest, the two other African superclans L2 and L3, are still very old, and much older than any equivalents in other parts of the world. Of the four clusters within L2, by far the most frequent is L2A, to which a quarter of all Africans belong. Branching off are two rarer granddaughter clans, L2A1a and L2A1b, both of which have all the attributes of a recent origin in West Africa and are probably associated with the very beginning of the Bantu expansions on the Cameroon plateau. These mitochondrial lineages were then carried along to East and Southeast Africa as part of that mass migration. These two clans are at the very tips of the family tree, whereas the age of the main L2A cluster is very much older, at about 55,000 years. The age is similar in both East and West Africa, making it difficult to choose one as the origin of L2A. The eastern and western branches seem to have separated about 14,000 years ago, around the end of the last ice age, when everyone was on the move. The date of this separation is judged on the accumulation of mutations from shared founder lineages, but where it occurred is hard to decide. I am inclined to think a Western origin is the more likely given that the sister clusters L2B–D are found predominantly in West and Central Africa and not in the east. But we are dealing with events of the very deep past, where some degree of uncertainty is both inevitable and forgivable.

  The sister clusters L2C, L2B, and L2D are more or less confined to West and Central Africa, with traces carried along with the Bantu expansion. All four clusters in the L2 superclan are very old, with ages ranging from 120,000 years (L2D) through 55,000 (L2A), to around 30,000 for L2B and L2C. Taken together, the L2 superclan makes up a third of all Bantu mDNA lineages in Southeast Africa, bringing the combined total with a West African origin to 44 percent.

  The superclans L1 and L2 are found only among indigenous Africans, or those like African Americans with comparatively recent roots in Africa. Only descendants of the third superclan, L3, are found among indigenous people outside Africa, whether that is Europe, Asia, Australasia, o
r America. But before we look at these lineages, how is L3 distributed within Africa itself? This superclan is found at its highest frequency in East Africa, but is certainly not confined to that region of the continent. Even so, it makes sense, as this is the closest region to the Red Sea, the Indian Ocean, and the Near East from where our species ventured out of Africa about sixty thousand years ago on its way to settle the rest of the world. Those adventurous souls were all members of cluster L3A.

  Outside of East Africa, clusters L3B, L3D, and L3F are predominantly West African in both origin and geographical distribution. Like other West African lineages that we have already introduced, many were carried into southern Africa by the Bantu expansion. The small sister cluster, L3G, is largely restricted to the Hadza foragers who live along the shores of Lake Eyasi in Tanzania, who share the characteristic Khoisan click language of the San Bushmen. Though L3G does not seem to have moved south with the eastern stream of the Bantu expansion, as we shall see it is found in African Americans, though not frequently.

  The oldest, most complicated, and most widespread of all the L3 clusters is L3E. You may be glad to read that it is the last one we need to look at in any detail. L3E appears to have originated in Central Africa, where it is still common, around 45,000 years ago. The cluster is well represented in the Bantu of Southeast Africa, and there are Kenyan Kikuyu who are also members. L3E is rare in West Africa but surprisingly common among African Americans. The reason for this was a mystery until it was discovered that a third of Brazilians with African roots are members of the clan. This suggested the former Portuguese colony of Angola as the origin, and sure enough, recent research has found that L3E is very frequent there.

  It has taken researchers a long time to unravel the complexities of the genetics of Africa, a complexity forged by the vast size of the continent and its long history of human occupation. But now we have a good idea how the ebb and flow of humanity have sculpted the genetics of today’s Africans, and as we shall now see, how this knowledge has helped so many Americans reestablish their connections to the continent of their ancestors—connections so often severed by the cruelty of slavery over the sea.

  10

  “I Am a Zulu”

  A village in Mozambique.

  “I am a Zulu,” famously declared the best-known African American woman alive today, Oprah Winfrey. Being Oprah, her words were flashed around the world. Literally, of course, she is not a Zulu but an extremely gifted television host, author, businesswoman, philanthropist, and campaigner. But her choice of words was very telling: She used “I am,” not “One of my ancestors was.” Oprah’s declaration of her African genetic roots followed a genetic test of her mitochondrial DNA, and though it was not the first time an African American had made this connection, it was certainly the most celebrated.

  Tracing their African roots through mitochondrial DNA has been enormously helpful to many African Americans. While European genealogists concentrate on their patrilineal genetic roots through the Y chromosome, often flowing side by side with a surname, the obliteration of genealogically meaningful names for African slaves means this avenue is closed off. There are other reasons, too, why the Y chromosome is perhaps the worst DNA for African Americans to use to connect with their African roots. Fortunately the quest for matrilineal ancestors is helped by the fact that more than 97 percent of African Americans have inherited an African mitochondrial lineage from their mothers, their grandmothers, and back through the generations to the slave ancestor who crossed the Atlantic. Of course African Americans usually have many more African ancestors than this, but the precision of mitochondrial DNA and the special place that maternal ancestors hold in all our memories make this small piece of DNA both the most reliable and the most precious.

  As I saw at firsthand with Jendayi Serwah, finding a match between her and a Kenyan Kikuyu was a transforming experience for both of us. The match was an exact one in that the dozen mutations that defined her sequence were exactly the same as the anonymous Kikuyu. Anonymous because, like all academic research projects, the volunteers are protected by confidentiality clauses in the ethical approval that all such studies must obtain before they begin. These anonymous studies have been essential for revealing the indigenous patterns of African mitochondrial DNA, as we have seen, although they will never directly identify an individual as a relative. Nonetheless, close or exact matches between African Americans and indigenous research samples can give clues to an ancestral homeland in Africa. But it is not completely straightforward, for reasons that we have touched on in the previous section. For example, although Jendayi is an exact match with a Kenyan Kikuyu, that does not necessarily mean that the woman from whom they both inherited their mitochondrial DNA was herself a Kikuyu. They are both in the clan of Leisha (L2A). Leisha lived about fifty-five thousand years ago, well before the subsequent migrations within Africa that scattered her descendants far and wide across the continent. The precise sequence match means that their common ancestor lived far more recently than Leisha herself and probably within the last few thousand years. That is an average figure based on the mitochondrial DNA mutation rate, but the range of possibilities is very wide. We also know that the clan of Leisha is found not only in West Africa, but also in southeast Africa as a result of the Bantu expansions, which is what probably carried the lineage from West Africa to Kenya. We know from the history of slavery that it is far more likely that Jendayi’s ancestor was taken from West Africa than from Kenya, but we cannot be absolutely sure. But it is so much better than nothing.

  Comparing the mitochondrial DNA of African Americans with their indigenous African cousins is also very revealing about the geography of slavery. At first the absence of matches was perplexing. Take the clan of Lalamika (L1C). Many African Americans are members of this clan, but among indigenous Africans whose DNA has been tested it is comparatively rare. For example, among all the indigenous Africans sampled until 2006, there were only nine. Seven were from Mozambique in southeastern Africa, another from the Cameroon mainland, and the last from the island of São Tomé a hundred miles offshore. But, among a much smaller sample of African Americans, there were sixteen who belonged to this branch of the clan. Here I am using “African American” in its widest sense, in that this number included two Mexicans, four African Brazilians, and three “White” Brazilians.

  At the opposite extreme, in the sister clan of Latasha (L1D) there are sixty-one Africans among the indigenous volunteers and not, so far, any African Americans at all. This mismatch is also explained by the geography of slavery. The indigenous African representatives of the Latasha clan are made up of forty San Bushmen from the Botswana Kalahari, twenty Bantu speakers from Malawi and Mozambique, and one solitary member of the Turkana tribe from northern Kenya. There are none at all among the almost five hundred African Americans included in this study. The reason is simply that slaves were not taken from Botswana’s Kalahari Desert or from Malawi, both of which are far from the coast. They were, however, taken from Mozambique, so I think it is only a matter of time before we do see an African American in the clan of Latasha.

  These two very different portraits are partially explained by the mechanics of the slave trade. While most ships left Africa bound for the Americas from the west coast, this was not always the original source of their human cargo. The Portuguese colony of Mozambique, facing the Indian Ocean, exported over a million slaves, and African Americans in the clan of Lalamika probably started their journey from there, via the offshore island of Zanzibar that had been an active Arab slave port for centuries. But before they crossed the Atlantic, slaves from Mozambique were often “stored” on the Portuguese island of São Tomé, awaiting shipment. Most likely the ancestors of the two clan members from São Tomé and Cameroon were also from Mozambique but somehow managed to avoid being taken across the Atlantic.

  The relatively few absolutely exact genetic matches between African American and indigenous Africans in this example is partially a reflection of the ex
treme intrinsic diversity of African mitochondrial DNA that has had so very long to accumulate its huge range of mutations. Outside the family, almost everybody else is different. But it is also due to one other important but often-neglected reason. Put simply, if a region has not been sampled, there will be no matches within it. The research results we have been examining are distilled from about twenty different projects. By the nature of academic research, these projects are limited in scope. They are constrained by the amount of funding available, by the opportunities and permissions to study particular tribes or particular regions and by the time pressure to publish the conclusions. Fortunately the results from these separate projects can be combined because the mDNA sequences are all directly comparable. But there are always regions where no research has been done. Either it is too dangerous, or permissions are hard to get, or stretched funding agencies are not eager to pay for yet another project that looks awfully similar to the last one—there are a host of reasons. Angola, for example, was not widely sampled until recently due to the long-running civil war from 1975 to 2002. After the war was over and genetics results were published, mDNA matches between indigenous Angolans and African Americans began to appear.