She Has Her Mother's Laugh: The Powers, Perversions, and Potential of Heredity (29 page)

These results gave Lewontin the confidence to broaden his research and take on the great question of race. He embarked on a new study to see how well racial groups aligned with the actual genetic diversity of humans. If races were indeed biologically significant, Lewontin argued, each race should have a starkly distinctive combination of genetic variants. Most of the genetic diversity should exist between the races rather than between individuals of the same race.

Lewontin gathered measurements of seventeen different proteins in a wide range of human populations, from the Chippewa to the Zulu, from the Dutch to the people of Easter Island. When he sorted people according to their race, he found that the genetic differences between races accounted for only 6.3 percent of the total genetic diversity in humans. The genetic diversity
within
populations, such as the Zulu or the Dutch, contained a staggering 85.4 percent.

In 1972, Lewontin published these results in a profoundly influential paper entitled “The Apportionment of Human Diversity.” He concluded that racial classifications had become entrenched in Western society thanks to optical illusions. People defined races based on features “
to which human perceptions are most finely tuned (nose, lip and eye shapes, skin color, hair form and quantity).” But these features were influenced by only a small
number of genes. It was wrong to assume that all the other genes people carried followed the same patterns.

Given his findings—and given all the suffering that had been justified by racial classifications—Lewontin urged that society set them aside. “Human racial classification is of no social value and is positively destructive of social and human relations,” he declared. “Since such racial classification is now seen to be of virtually no genetic or taxonomic significance either, no justification can be offered for its continuance.”

It was a sweeping statement to make based on fairly little data. But in the years since, younger generations of scientists have revisited Lewontin's question with better tools. Instead of proteins, they've examined DNA. They've surveyed more people, from more populations.
In 2015, for example, three scientists—Keith Hunley and Jeffrey Long of the University of New Mexico and Graciela Cabana of the University of Tennessee—studied DNA from 1,037 people belonging to fifty-two different populations around the world. In each person, they sequenced the same 645 segments of DNA. They looked for the differences in these segments from person to person, calculating their genetic diversity.

Hunley and his colleagues confirmed, like others had before them, that most human genetic diversity can be found within populations rather than between the so-called races. And thanks to the huge scale of their study, they could measure human diversity with far greater precision. The people who live in African populations tend to be more genetically diverse from one another than people who live on other continents, for example. The population with the lowest genetic diversity was a small Amazon tribe called the Suruí. Yet even the Suruí—who number only about 1,120 people—possess about 59 percent of all the genetic diversity in our entire species. If you wiped out everyone on Earth except the Suruí, in other words, nearly two-thirds of humanity's genetic variation would survive.

“In sum,” Hunley and his colleagues said, “we concur with Lewontin's conclusion that Western-based racial classifications have no taxonomic significance.”

The Venn diagram that Sushant Kumar made for me—showing me all
the SNPs that are sprinkled over me, a Nigerian, and a Chinese person—felt like a personal emblem of how badly the concept of race explains human genetic diversity. I'd call myself white, and yet 83 percent of my 3.5 million single-nucleotide polymorphisms are shared by either an African or an East Asian. We may inherit some of those shared variants from common ancestors who lived hundreds of thousands of years ago. Some variants may have arisen later, thanks to a new mutation. They then spread from population to population as people mixed their genes the way people always do. All three of us—me and my pair of anonymous far-flung cousins—got showered in the same genealogical glitter.

Race may not be a meaningful biological concept, but it does exist: It has a powerful existence as a tradition of putting people in social categories. Those categories, then, had profound influences on people's lives. Racial categories served as a legal justification to enslave groups of people and declare their children slaves from birth. Race helped turn other people into scapegoats for economic disasters, justifying their slaughter by the millions. Other people were classified into races judged incompetent to make use of their own land, justifying pushing them off it. And racial categories also gave some people the luxury of enjoying those lands and the profits of slave-based economies without having to learn much about their history. Even after racist institutions and laws were abandoned, their effects have endured for generations, extending race's power.

Because race is a shared experience, it can join people together who aren't closely related. American blacks gained their collective identity only when they came together as cargo on slave ships bound for the colonies. Slave traders roamed up and down the coasts of Africa to capture people separated by thousands of years of history, in
Senegal, Nigeria, Angola, even Madagascar. Richard Simson, a surgeon who traveled to South America in 1689 on an English privateer, observed that throwing strangers together was a crucial step in making slavery a profitable business.

The way “
to keep Negros quiet,” Simson wrote, “is to choose them from
several parts of the Country, of different Languages, so that they find they cannot act jointly.”

Leaning on the biological concept of race like a crutch has led doctors into some
embarrassing blunders in their studies of diseases. “
There is no race which is so subject to diabetes as the Jews,” declared W. H. Thomas, a New York doctor, in 1904. As late as the early 1900s, Jews were considered a distinct race, with its own diseases. To guide their immigration policies, the United States Congress compiled a book called
Dictionary of Races or Peoples.
The book treated the evidence of the Jewish race as plain to see. “
The ‘Jewish nose,' and to a less degree other facial characteristics, are found well-nigh everywhere throughout the race,” the report declared. Such racial classifications led doctors to look for diseases that were characteristic of each race.
Jews, doctors came to agree, had diabetes.

The seed of this notion sprouted in 1870, when a doctor in Vienna named Joseph Seegen observed that a quarter of his patients were diabetic. Other physicians later concluded that Jews died from diabetes at a far higher rate than other groups. German doctors started referring to diabetes as the
Judenkrankheit
: the Jewish disease.

Between 1889 and 1910, New York saw its rate of diabetes triple. To J. G. Wilson, a physician with the US Public Health Service, the cause was clear: the influx of Jewish immigrants. Jews had “
some hereditary defect,” Wilson said, that made them vulnerable.
William Osler, the most important clinical doctor of the early 1900s, blamed the vulnerability of Jews to diabetes on their “neurotic temperament,” along with “their racial tendency to corpulence.”

And then, in the middle of the twentieth century, the universally recognized fact that diabetes was a disease of the Jewish race simply disappeared. Historians can't definitively say why. It's true that a few scientists questioned the statistical evidence behind the Jewish disease. But no one ever published a definitive takedown. Maybe after Nazis peddled myths that Jews were a naturally disease-ridden race, American doctors quietly decided to retire their own misconceptions.

Myths like Jewish diabetes do not detract from the fact that some
people who identify themselves with certain labels—black, Hispanic, Irish, Jewish—have relatively high rates of certain diseases. Ashkenazi Jews have a higher rate of Tay-Sachs disease than other groups, for example. African Americans have a higher rate of sickle cell anemia than European Americans.
Hispanics are 60 percent more likely to visit the hospital for asthma than non-Hispanic whites. Researchers have also found significant associations between the race of patients and how their bodies respond to drugs. Chinese people tend to be
more sensitive to the blood-thinning drug warfarin than whites, indicating they should get a lower dose.

In some cases, these patterns are the result of the genes people inherited from their ancestors. But sometimes they aren't.

When
Richard Cooper went to medical school at the University of Arkansas in the late 1960s, he was stunned at how many of his black patients were suffering from high blood pressure. He would encounter people in their forties and fifties felled by strokes that left them institutionalized. When Cooper did some research on the problem, he learned that American doctors had first noted the high rate of hypertension in American blacks decades earlier. Cardiologists concluded it must be the result of genetic differences between blacks and whites. Paul Dudley White, the preeminent American cardiologist of the early 1900s, called it a “racial predisposition,” speculating that the relatives of American blacks in West Africa must suffer from high blood pressure as well.

Cooper went on to become a cardiologist himself, conducting a series of epidemiological studies on heart disease. In the 1990s, he finally got the opportunity to put the racial predisposition hypothesis to the test. Collaborating with an international network of doctors, Cooper measured the blood pressure of eleven thousand people. Paul Dudley White, it turned out, was wrong.

Farmers in rural Nigeria and Cameroon actually had substantially lower blood pressure than American blacks, Cooper found. In fact, they had lower blood pressure than white Americans, too. Most surprisingly of all, Cooper found that people in Finland, Germany, and Spain had higher blood pressure than American blacks.

Cooper's findings don't challenge the fact that genetic variants can
increase people's risk of developing high blood pressure. In fact, Cooper himself has helped run studies that have revealed
some variants in African Americans and Nigerians that can raise that risk. But this genetic inheritance does not, on its own, explain the experiences of African and European Americans. To understand their differences, doctors need to examine the experiences of blacks and whites in the United States—the stress of life in high-crime neighborhoods and the difficulty of getting good health care, for example. These are powerful inheritances, too, but they're not inscribed in DNA. For scientists carrying out the hard work of disentangling these influences, an outmoded biological concept of race offers no help. In the words of the geneticists Noah Rosenberg and Michael Edge, it has become “
a sideshow and a distraction.”

To many people, Rosenberg and Edge may sound as if they're ignoring the evidence staring them in the face. While I may share millions of single-nucleotide polymorphisms with a Nigerian, no one would mistake me for someone whose family goes back centuries in Lagos. I once went to Beijing, and never on my trip did someone walk up to me and ask for directions in Mandarin. It is true that humans have physical differences, and some of those differences are spread geographically across the planet. But clinging to old notions about race won't help us understand the nature of those differences—both the ones we can see and the ones we can't.

What matters is ancestry. A small band of hominins in Africa evolved into
Homo sapiens
around 300,000 years ago, after which they expanded across that continent and then across the world. Those journeys shaped the genomes that people inherited from their ancestors. And today, if we look at our own genomes, we can reconstruct some of that history, even back to ancestors who weren't exactly human.

T
HE
T
A
ITA
THR
USH
is cloaked in black feathers and tipped by a vermilion beak. It can be found only in the cloud forests of the Taita Hills of southern Kenya. Some species of birds fly far across wide ranges, but the Taita thrush is a homebody. It limits its movements to a small territory of the forest floor, where it hops about in search of fruit and insects. This way of life left the bird exquisitely vulnerable to modern change. Most of the Taita forests were cleared from the hills for farming and pine tree plantations, leaving behind just a few islands of trees at the summits. By the end of the twentieth century, only three populations of Taita thrushes survived. Each numbered just a few hundred.

The isolation of the birds left them especially threatened by extinction. Before the deforestation, their genes flowed across the landscape as the birds mated with their neighbors. Now the genes of the Taita thrush were trapped on hilltop islands. As the years passed, each new generation ran a greater risk of inheriting two recessive alleles and developing a genetic disorder—one that might cut a bird's life short or make it infertile.

Hoping to save the species, conservation biologists climbed the hills and captured 155 thrushes from all three forests. They drew blood from the birds, and later isolated short segments of DNA from them. They studied this genetic material to gauge how much diversity was left.

In 1998, a geneticist at the University of Oxford named Jonathan
Pritchard asked the scientists if he could look at the sequences. Pritchard sorted them into three groups, based on their genetic similarities alone. He then asked the conservation biologists where each bird lived. Each of the groups he created perfectly matched each forest.

To sort the Taita thrushes, Pritchard had used a computer program he had recently written with his advisor, Peter Donnelly, and a fellow postdoctoral researcher, Matthew Stephens.
They had named the program STRUCTURE.

Sorting 155 birds by DNA alone was a daunting task. At many positions, their genes were identical. Many of the variants shared by only some birds could be found in more than one forest. But Pritchard and his colleagues recognized that certain combinations were more common in each group than others—a signature of their origins. There was a signal buried in all the genetic noise.

When the three forests became isolated, their gene pools got cut off from each other, too. In each pool, some variants were common and some were rare. Without birds traveling between the forests, each generation passed down those variants to their descendants. After many generations of isolation, this pattern still held true. The birds in each forest tended to have some common variants, and it was unusual for them to have rare variants.

Pritchard used STRUCTURE to take advantage of these patterns to sort the birds into groups. He found that three groups worked best. The birds in each of the three groups had a clearer genetic connection to each other than if he had tried sorting them into two groups, or four, or five. STRUCTURE was so good at this sorting that Pritchard could pick out a single thrush, look at its DNA, guess which forest it came from, and almost always get the right answer.

What made this success even more impressive was the similarity of the Taita thrushes. The birds had become isolated from one another only a century beforehand. These were not distinctive subspecies, in other words. From forest to forest, the birds look pretty much identical. They eat the same food. In every forest, males and females form monogamous bonds. The subtle genetic differences that Pritchard used to trace the birds to their homes meant little to the birds themselves.

Pritchard did not invent STRUCTURE only to identify the homes of Taita thrushes. He wanted to build a program that could automatically sort individuals from any species into meaningful groups. He especially wanted to apply it to
Homo sapiens
. In the 1990s, it had become clear that mapping the genetic structure of humanity would be crucial to finding genes associated with diseases.

Scientists had begun searching for these genes by looking for variants that were unusually common in people with a particular disease. But they could end up with misleading results if they didn't take into account people's ancestry. This danger came to be known as the
chopstick effect, after a fable spun in 1994 by the geneticists Eric Lander and Nicholas Schork.

Imagine, Lander and Schork said, that a team of researchers in San Francisco decided they would find the genetic cause for why some people in the city ate with chopsticks and others did not. They took blood samples from a random selection of people and scanned their DNA. Lo and behold, the scientists discovered an allele for an immune system gene that was far more common among chopstick users than among people who did not use them. Therefore, the geneticists concluded, inheriting that allele caused people to be more likely to use chopsticks.

They were wrong. The allele was more common in chopstick users for an entirely different reason: because it was more common in Asian Americans than people of European descent. Asian Americans were also more likely to use chopsticks than European Americans. The immune system, in other words, has nothing to do with chopsticks.

A real example of the chopstick effect came to light in the 1980s among the Pima Indians of the southwestern United States. They suffer from Type II diabetes at a catastrophic rate: About half of all adults in the community develop the disease. Diabetes began to wreak havoc on the Pima only in the 1900s, after they lost their land and their sophisticated farming system. Suddenly they had to survive on carbohydrate-rich government-supplied food. That diet could put anyone at greater risk of diabetes, but the Pima proved to be especially vulnerable. Geneticists suspected that their higher risk was due to genetic variants they shared.

William Knowler, a researcher with the National Institute of Diabetes and Digestive and Kidney Diseases, led one of the first studies on Pima DNA. He studied 4,920 subjects on the Pima reservation in Arizona. He discovered that about 6 out of every 100 Pima carried a variant in a gene called Gm, which encodes a type of antibody. The Gm variant seemed to protect the Pima against diabetes. Among those who carried it, only 8 percent developed the disease. Among the Pima who lacked the Gm variant, 29 percent developed diabetes.

Knowler might have stopped there and declared victory. But he was well aware that the Pima he studied did not have a simple history. Native Americans arrived in the Western Hemisphere some fifteen thousand years ago. The Pima probably settled in the Southwest by two thousand years ago, and five centuries ago they came into contact with people of European ancestry: first Spanish explorers, and then Mexican farmers. By the mid-1900s, Pima Indians and Mexican migrant laborers were working together on Arizona cotton farms. Some Pima started
families with people outside the tribe. As a result, some of the Pima whom Knowler studied had a fair bit of European ancestry.

To take ancestry into account, Knowler split his subjects into two groups: those with some European background and those with none. When he looked at the Gm variant within each group, the evidence for its defense against diabetes disappeared. Among people with 100 percent Pima ancestry, having the Gm variant didn't lower the risk of diabetes. It also didn't make a difference when Knowler compared the Pimas with some European ancestry to each other.

Knowler had been initially fooled by the Gm variant, he realized, because it was much more common among Pima with some European ancestry. It served as a genetic marker, in other words, rather than as a direct defense against diabetes. Knowler concluded that European versions of certain genes might lower the odds of developing diabetes on a diet high in simple carbohydrates. But he couldn't say from his data which genes those might be. What he did know was that the Gm variant merely came along for the ride.

Knowler managed to overcome the chopstick effect by asking the Pima about their ancestors. Their European forebears had lived recently enough that the Pima could give Knowler a reliable genealogy. He was also fortunate to be studying a small, relatively isolated community. Other scientists who study broader populations of people with mixtures of ancestries and fuzzy family memories do not enjoy Knowler's advantages.

Pritchard and his colleagues, collaborating with Noah Rosenberg at Stanford University, found that they could use STRUCTURE to overcome the chopstick effect, even when they didn't have any information about people's family trees. The geneticists could identify clusters of people based on their DNA alone. To adapt STRUCTURE to the task, the scientists had to reckon with the fact that
people are not Taita thrushes. They do not live in a few forests in a small patch of Africa; they span the globe. And rather than living in isolation, humans have migrated over thousands of years, mixing their DNA in their living descendants.

The scientists created a version of STRUCTURE that let them scan the genetic variation in people and assign each individual's DNA to one or more groups of ancestors. Pritchard and his colleagues could then look at how well they could account for the genetic variation in people with different numbers of groups.

In 2002, Pritchard and his colleagues tried STRUCTURE out on people. They looked at
genetic variations in 1,056 people from around the planet. Just as in other studies of human diversity, they found that the overwhelming amount of genetic diversity was between individuals. The genetic differences between major groups accounted for only 3 to 5 percent. And yet, with the help of STRUCTURE, the researchers used some of those variants to sort people into genetic clusters. When the scientists allowed people to descend from five different groups, for example, they clustered mostly according to the continents they lived on. People in Africa could trace much of their ancestry to one group, while people in Eurasia were linked to a second one. East Asians traced much of their ancestry to a third, Pacific Islanders to a fourth, and people in the Americas to a fifth.

Much to the chagrin of Pritchard and his colleagues, some people mistakenly took these results as evidence for a biological concept of race. But any resemblance between genetic clusters of people and racial categories concocted before genetics existed can have no deep meaning. It would make just as little sense to say that Aristotle's classifications of animals have been vindicated by comparing the DNA of different species. Aristotle put species into categories based on whether they had blood, whether they had hair, and so on. The genes of animals with hair—mammals—show that they do indeed belong to a group. But
Aristotle also threw together species into other categories that have no close evolutionary link. It would be a disaster for biology if scientists cast off two thousand years of progress and followed Aristotle's example. The same is true for race.

Those who claim that STRUCTURE proves the existence of human races also have to ignore how Pritchard and his colleagues actually used it to study human variation. The clusters that the five ancestral groups produced didn't have sharp boundaries. Where two clusters met on a map of the world, the researchers found people who had some DNA that linked them to one group, and some that linked them to the other. What's more, STRUCTURE allows scientists to try out different numbers of ancestral groups to see what sort of clusters emerge. After trying out five ancestral groups, Pritchard and his colleagues decided to see what would happen if they ran their program with six. The results were pretty much the same, with one telling exception: A single population broke off from the Eurasian cluster and formed a cluster of its own.

That population is known as the Kalash, a few thousand people who live in the Hindu Kush mountains of Pakistan. Their separation in Pritchard's study may tell us something important about the history of the Kalash—perhaps a long isolation from other tribes in Pakistan, allowing them to accumulate a small number of genetic variations that set them off from much larger clusters of people. But it doesn't mean that the Kalash are biologically a race of their own.

Pritchard and his colleagues were also able to use STRUCTURE to search for clusters within clusters. For their study, the researchers had picked out five populations of people in the Americas, including the Pima
in Arizona and the Suruí of Brazil. When they created a model of those people based on five ancestral groups, they were able to identify people's tribes based on their DNA alone.

In the years since the 2002 paper came out, scientists have been improving on STRUCTURE, developing more powerful statistical tools for tracing people's ancestry. They've also been accumulating more DNA from more parts of the world, to get a more accurate map of the human genetic landscape. Before long, it became possible for genealogy companies to analyze customers' DNA and produce a rough breakdown of their ancestry. It was this approach that allowed LeVar Burton to learn that three-quarters of his ancestry came from sub-Saharan Africa, for example.

One of Pritchard's students, Joe Pickrell, ended up at the New York Genome Center. He and his colleagues there used his own update on STRUCTURE to compare people's DNA and estimate their ancestry. When Pickrell ran my DNA through his computational pipeline, he quickly discovered—not surprisingly—that my ancestry is entirely European. He and his colleagues then examined stretches of my DNA to see if they could trace them to smaller populations within Europe. To find variants that pointed to my northwestern European ancestors, for example, Pickrell and his colleagues looked at the DNA of people from Iceland, Scotland, England, the Orkney Islands, and Norway.

The one group they looked at that didn't have a clear geographical location was the Ashkenazi Jews. While Ashkenazi Jews lived for generations across much of eastern Europe, they remained a culturally closed group, mainly sharing their variants among themselves. They thus became recognizably distinct from neighboring Christians.