Within each continental race there are gradations of skin color, from the light-skinned Khoisan speakers of southern Africa to the darker-skinned Bantu speakers of western and central Africa, from the lighter-skinned Scandinavians to the darker-skinned peoples of southern India. Skin color is therefore an ambiguous indicator of continental race.
At the boundaries of these continental divisions are several groups formed by intermarriage between the two neighboring races, a condition for which geneticists use the term “admixture.” Ethiopians and Somalis, for instance, are an admixture of Caucasians and Africans. “The existence of such intermediate groups should not, however, overshadow the fact that the greatest genetic structure that exists in the human population occurs at the racial level,” Risch says.
In the United States there are several populations formed by intermarriage between members of two racial groups. African Americans, largely as a result of slavery, have a share of Caucasian genes that ranges from 12% to 23% in various populations, with an average of about 17%. “Despite the admixture, African Americans remain a largely African group, reflecting primarily their African origins from a genetic perspective,” Risch says.
Another group of admixed populations is counted by the U.S. Census Bureau as Hispanic although Hispanic is a linguistic, not a racial, category. Hispanics vary in their admixture in different parts of the country. In the southwestern United States, Hispanics are mostly Mexican Americans, whose ancestry is 39% Native American, 58% Caucasian, and 3% African, according to one recent estimate. East coast Hispanics come mostly from the Caribbean and have a larger proportion of African genes.
The United States is often referred to as a melting pot of races but the rate of mixing is slower than might be assumed. Figures from the 2000 U.S. census indicate that U.S. citizens do not marry each other at random. Racial endogamy (marrying within the racial group) is the rule: 97.6% of respondents reported themselves to be of one race; only 2.4% said they were of more than one race, presumably having parents of different races. Some 75% of Americans declared themselves to be white, that is, Caucasian; 12.3% said they were black or African American; 3.6% were Asian, 1% Native American, and 5.5% of other races.
Differentiation of the Ancestral Human Population
These continental groups reflect the leading roles of geography and endogamy in shaping human races. As long as everyone intermarries, as would doubtless have been the case in the ancestral human population, there is a single genetic pool. New diversity—that is, new alternative versions of genes—accumulates through mutation, and old diversity is eliminated by drift, but these changes occur within a common pool. Any substantial bar to intermarriage, however, whether a mountain range or a religious ban on marrying outsiders, will set up two genetic pools. Since mutation and drift are both random processes, the changes in the two pools will now take place independently. From that point on, the two populations may follow different evolutionary paths. Migration between the two will sharply reduce genetic difference; time and distance will increase it.
The starting point for the emergence of human races would have been the dispersal, within Africa, from the ancestral homeland some 50,000 years ago. Before people left for the world beyond, the human population in Africa had apparently fragmented, doubtless by geographical distance, into several different populations. As already noted, those who left Africa belonged to just one of these populations, those descended from the L3 branch of the mitochondrial DNA tree. They carried away in their genes only a subset of the African genetic diversity, meaning only some of the alleles of each gene. That fact alone set them on a potentially different evolutionary path.
The emigrants eventually spread out over the rest of the globe and themselves fragmented into many even smaller populations. The smaller a population, the greater is the force of genetic drift, which reduces the number of available alleles. Without interbreeding to keep the human gene pool mixed, the populations of each continent or region would over time have become more distinct and less like the others.
The importance of drift in differentiating a static population has recently been recognized in the population of Iceland.229 As mentioned earlier, even though the island has been settled for only 1,000 years, the people in each region have become sufficiently different genetically that by sampling Icelanders’ genome in just 40 different places it is possible to tell which of 11 regions of the island they come from. In the rest of the world, with some 50 times longer for genetic forces to act, and many severe impediments to movement, a much greater degree of differentiation would be expected.
Besides drift, another differentiating force on the world’s separate human populations would have been natural selection. Selection may have pressed particularly hard on the people who left the African homeland, since they would have had to adapt to radically new diet, terrain and climates. A particularly striking example of selection is a recently discovered gene variant that causes pale skin in Caucasians. Almost all African and Asians have the same, ancient form of the gene, which is known at present as SLC24A5. Some 99% or more of Europeans have a new version, that must have arisen after Caucasians and East Asians had become separate populations. The new version presumably became almost universal among Caucasians because the pale skin it conferred was of overwhelming advantage, whether for reasons of health or sexual attractiveness or both. A different gene, yet to be discovered, must give East Asians their pale skin.230
As Darwin suggested, sexual selection, the partly capricious taste of women and men for partners of a certain type, as well as competition between men, may have been a strong selective force, and one that acted somewhat independently in each human population. Disease has certainly influenced the human genome as people in different regions responded to local diseases like malaria. Warfare, an unremitting pressure, surely played a major role in shaping populations. And another powerful molder of human populations would have been climate, especially the adaptations necessary for living in northern latitudes and the violent climatic swings of the late Pleistocene.
Given all these evolutionary forces at work, it is not so surprising that the widely dispersed human populations in various continents acquired their own distinctive variations on the general human theme. This genetic-geographical difference is reflected in the familiar trees drawn on the basis of mitochondrial DNA or the Y chromosome, and on several other kinds of genetic elements. Risch cited some of these studies as proof of the division of the human population into continent based races.
A few months after Risch’s article of 2002, a more comprehensive study by Marcus Feldman of Stanford University reached a very similar conclusion. Instead of examining just a few markers, or sites on the DNA, as many previous studies had done, Feldman and his colleagues looked at 377 sites throughout the genome, a larger and more representative sample. This was done for each of 1,000 people from 52 populations around the world. A computer was then instructed to group the individuals, based on their DNA differences at the 377 sites, into clusters. They fell naturally into 5 clusters, corresponding to their five continents of origin—Africa, western Eurasia (Europe, the Middle East, the Indian subcontinent), East Asia, Oceania and the Americas.231
Feldman and his colleagues did not use the word “race” in their article, referring instead to “structure” and “self-reported population ancestry” (meaning a person’s own identification of their race), but he acknowledged in an interview that the finding essentially confirmed the popular conception of race. “Neil’s article was theoretical and this is the data that backs up what he said,” Feldman commented in reference to Risch’s study.232
Identifying Race by DNA
A consequence of the Risch and Feldman studies is that they provide, for the first time, an objective way of ascertaining an individual’s race. Most previous systems of race classification, with the principal exception of modern craniometry, have been based on characteristics like skin color, which vary in an unsystematic way, and were often designed
with a malign agenda such as demonstrating one race’s alleged superiority to others. Not only does the genetic definition of race have no such agenda, but it has nothing directly to do with any physical attribute.
The reason is that the genetic markers used to identify race are not part of the genes or their control regions, so far as is known, and therefore play no part in the physical appearance or behavior of an individual. Presumably they are indirectly correlated with genes that do control the body’s physical makeup, but the connection is indirect and at present unknown.
The DNA markers analyzed by the Feldman team are of the same type as is used in the DNA fingerprinting of forensic cases. At various sites on the human genome the sequence of DNA units goes into a sort of stutter, known as a short tandem repeat because a few units of DNA are repeated several times over, as in AC-AC-AC-AC-AC. For some reason, these stutters tend to confuse the cell’s DNA copying apparatus, which every dozen or so generations may accidentally either add or delete a repeat. The exact number of repeats at a given site is therefore quite variable from one person to another, and so can be made the basis of systems for identifying populations or individuals.233
Only some 3% of the DNA in the genome is devoted to genes; the rest of the DNA is mostly yards of filler material. The short tandem repeats are part of the filler material so do not affect a person’s physical makeup. But some repeats lie close to genes, some of which have evolved in different ways in the various races. By selecting the right repeats, geneticists can find ones that are quite diagnostic of race, even though at present they have little idea which genes it is that give people of different races their different appearance.
Risch calculated that if the sites with short tandem repeats were chosen entirely at random, analysis of about 100 sites should suffice to say which of the five major races a person comes from. But as few as 30 sites would be enough if the sites were specially chosen so as to be diagnostic of race. Many hundreds of markers would be needed to distinguish, within a race, between two populations or ethnicities, Risch estimated.
Sets of these sites, known as Ancestry Informative Markers, can be used to identify not just an individual’s race but the racial origin of individual sections of a person’s genome. A company called DNAPrint Genomics has already started offering a test to assess people’s continent of origin and, if of mixed race, the proportions of ancestry due to different races.234 The test is based on a set of markers identified by Mark Shriver, a geneticist at Pennsylvania State University. It has already proved useful in police inquiries by identifying the race of a suspected serial killer from tissue collected at a crime scene. In June 2003 police believed that a serial killer in Louisiana was white, but were informed otherwise by DNAPrint Genomics, whose test showed the killer’s ancestry was 85% African and 15% American Indian; they then arrested a suspect who was black.235 The reliability of the test has not yet been established, but if it helps police identify a suspect, the suspect’s DNA can then be compared with the crime scene DNA in the usual way.
Feldman and his colleagues say they needed varying numbers of markers—in this case sites with tandem repeats—to identify a person’s continent of origin, depending on the genetic variability of the race in question. Native Americans could be assigned to their continent of origin with just 100 markers, whereas almost all 377 markers were required to identify Middle Easterners. This is because Native Americans are all descended from their Siberian founders whereas Middle Easterners are a more complex genetic blend; they are mostly Caucasian but some, like the Bedouin, have an African contribution.
Feldman’s method gives a glimpse of how deeply genetic markers may be able to reach into population history. The computer program used to sort the genome samples into continental clusters could also split an individual’s genome into different parts if the person was of mixed ancestry. People from the Hazara and Uighur of Central Asia, long a crossroads between east and west, emerged with genomes roughly half Caucasian and half East Asian in origin. The Surui, a fairly isolated people of Brazil, have genomes that are entirely American (in terms of the computer program’s 5 racial clusters), whereas Mayan genomes are American with a strong dash of European and East Asian admixture.
With extra markers, and ones chosen to be more diagnostic of geographical origins, it should be possible to explore a population’s ancestry and history in a much more detailed way. For geneticists, the essence of race is not politics but history: race defines through which branch of the human family tree people trace their descent.
Scientific Attitudes to Race
Researchers’ attitude to race has swung through a wide arc in the last century and the new view developed by the work of Risch, Feldman and others has probably not yet become the consensus view.
In the nineteenth century, as European explorers became acquainted with the peoples of other continents who seemed so different from themselves, a serious debate arose as to whether these strange foreigners should be considered as belonging to separate species. Darwin, with his usual unerring insight, rejected the idea in his 1871 book The Descent of Man, arguing there was only one human species, though divided into subspecies or races. The one species must have had a single origin, which Darwin pre sciently placed in Africa, the continent with the greatest diversity of great apes. That human population, in his view, was later fragmented into different races by geographical isolation, followed by a differentiation that in Darwin’s view was principally driven by sexual selection, the preference by women for men of a certain type.
Since Darwin’s time, greater awareness has developed of the dangers of race in light of the many harms and injustices committed by people of one race against those of another. Many academic researchers, including geneticists, have sought to minimize the extent of biological variation within the human family. What is still one of the most influential positions on race is a statement made in 1972 by the geneticist Richard Lewontin.
Lewontin measured a property of proteins (DNA sequencing was not then available) taken from people of different races, and computed a standard measure of variation known as Wright’s fixation index or FST. The idea is to measure some character that varies in members of a population and assess how much of the variation arises because two subpopulations differ from each other in that character. The index, in other words, reflects how much of the variation is general and how much is specific to the subpopulations.
Lewontin’s value for FST came out at 6.3%, meaning that of all the variability in the human population, at least as reflected in the 17 proteins he had measured, only 6.3% lay between races, while a further 8.3% was found to lie between the ethnic groups within races. “Of all human variation, 85% is between individual people within a nation or tribe,” Lewontin concluded. 236
This finding is perfectly in line with the expectation that most of the genetic variation in each race would be the same as that of the ancestral human gene pool from which it was drawn. But the question then arose as to whether the extent of the difference between races was large or small. Lewontin argued that the difference was so trivial that racial classification had no genetic significance or justification.
Many biologists have chosen to go along with his interpretation, and this position has been followed, even taken to extremes, by the major social science organizations in the United States. According to the American Sociological Association, race apparently does not even have a biological foundation, since it is a “social construct.” The association’s official statement on race warns that “Although racial categories are legitimate subjects of empirical sociological investigation, it is important to recognize the danger of contributing to the popular conception of race as biological.”237
The American Anthropological Association also dismisses the idea that biological differences can be recognized between races: “In the United States both scholars and the general public have been conditioned to viewing human races as natural and separate divisions within the human species based on v
isible physical differences,” the AAA statement says. But since physical traits vary smoothly across the globe, and are not correlated with one another, “these facts render any attempt to establish lines of division among biological populations both arbitrary and subjective.”238
But people can now be objectively assigned to their continent of origin, in other words to their race, by genetic markers such as those used by Feldman. And Lewontin’s characterization of the differences he had found as trivial was as much a political as a scientific opinion. The degree of differentiation he had measured in the human population was similar to other estimates that put the value of global FST as between 10 and 15%. Sewall Wright, one of the three founders of population genetics and the inventor of the FST measure, commented that “if racial differences this large were seen in another species, they would be called subspecies.”239 Wright specified that an FST of 5 to 15% in any population of organisms constituted “moderate” genetic differentiation, and 15 to 25% should be considered “great” genetic differentiation.240
But whether an FST of 10 to 15% represents a large or small degree of differentiation is probably not very relevant to the question of whether or not human races can be identified. The reason is that such measures of FST have captured something called neutral variation, which is probably not the kind that underlies most racial differences.
Neutral variation refers to mutations that don’t affect the organism one way or another. Evolution doesn’t care about such changes, and the frequencies of such alleles in a population will vary randomly under genetic drift. Most common variation is neutral, and most measures of FST are likely to sample common—that is, neutral—variation.
Before the Dawn: Recovering the Lost History of Our Ancestors Page 23