IDEOLOGY ALWAYS TRUMPS BIOLOGY III:

Code of Many Colors: Can researchers see race in the genome? (Christen Brownlee, 4/09/05, Science News)

It's difficult to get most scientists even to say the word race when referring to people. That's because in traditional scientific language, races are synonymous with subspecies—organisms in the same species that can interbreed but nevertheless are distinctive genetically.

Many species split into subspecies after being separated geographically for an extended amount of time. During generations of genetic mixing within but not between the isolated groups, some of each group's genes develop slightly different versions, or alleles. Scientists often use a rule called Wright's F statistic to judge whether separate groups are actually subspecies. If 25 percent or more of one group's alleles are different from another's, then by F-statistic standards, the two groups are considered subspecies. A difference of 100 percent would separate them into distinct species.

Subspecies, or races, exist for many animals—for example, the alleles in some populations of grey wolves score up to 70 percent on the F-statistic scale. However, the groups of people considered to be of different races have allelic differences of at most 15 percent, too little to constitute subspecies.

To the nonscientist, however, race clearly is a meaningful term, says Vivian Ota Wang of the Ethical, Legal, and Social Implications Research Program at the National Human Genome Research Institute (NHGRI) in Bethesda, Md. The concept seems to depend on a collection of physical features, "like a checklist," she says, "so that people can categorize each other into groups." Items on the list might include skin tone, hair texture, and the shapes of eyes, noses, or lips.

Most people don't carry a conscious perception of the checklist. Wang says that race has a lot in common with Supreme Court Justice Potter Stewart's famous definition of pornography: We know it when we see it.

About 100,000 years ago, defining race wasn't an issue—all early humans lived in Africa and had similar characteristics. That relatively small population of recently evolved humans carried the majority of alleles present in people today.

But over the next 50,000 years or so, as humans separated into groups, slight differences among populations crept into the genome. First, as waves of emigrants left Africa and spread throughout the world, our ancestors took slightly different groups of alleles with them. Just as each handful of jellybeans scooped out of a jar might have a different mix of flavors, every group of migrant humans carried a slightly different array of alleles.

Later, when roaming humans settled into permanent residences on different continents, new genetic mutations gradually built up within groups as they adapted to their distinct environments. Because people mated most frequently with others from the same region, each population developed its own set of mutational differences, some influencing survival and some being just genetic quirks. [...]

Race and family origin aren't entirely synonymous in modern times, when people can relocate around the globe. However, many researchers have found that the distribution of certain genetic variations can lump people into ancient ancestral groups uncannily similar to what nonscientists call races.

For example, Noah Rosenberg of the University of Southern California in Los Angeles and his colleagues published a study in 2002 that analyzed the number and type of microsatellite variations in the DNA of 1,056 people from 52 populations around the world.

Rosenberg's team masked any information about the study volunteers' ancestral backgrounds and then plugged the microsatellite information into a computer program that clusters people by genetic similarities. Six main clusters emerged.

After restoring individuals' ancestry data to the files, the researchers found that five of the six microsatellite clusters corresponded with geographic regions: Africa, Eurasia (Europe, the Middle East, central and south Asia), east Asia, Oceania (islands of the central and South Pacific), and the Americas (specifically native Americans). The sixth and smallest cluster linked to an isolated group of mountain-dwelling Pakistanis known as the Kalash.

The scientists weren't surprised that people's genetic mutations usually lump them into continental groups. For much of history, people have been land bound and so have mated mostly with people from the same continent.

However, Rosenberg says that he was surprised that he and his colleagues found it impossible to predict with certainty which combination of gene variants any specific person in each cluster had. The computer runs couldn't determine, for example, exact shades of skin color or types of hair texture for individuals.

"In a lot of classical anthropological views of race, race is thought to be a quality predictive of a large variety of traits about a person. We found that for any given person, it's not possible to predict accurately which [variant] they have at any particular site in the genome based on their group membership," Rosenberg says.

Neil Risch of the Stanford University School of Medicine and his colleagues recently used a similar method to come to a very different conclusion. Using microsatellite information from another study that had looked for a genetic link with hypertension in several U.S. populations, Risch's team ran data from 3,636 people through a computer program similar to Rosenberg's. However, instead of searching for clusters based on geography, Risch and colleagues compared clusters from the genetic data with self-described race/ethnicity categories.

The genetic data sorted into four categories—white, African American, east Asian, and Hispanic—which neatly matched what each person had checked on a form at the beginning of the study. Only five people had results inconsistent with their self-described race/ethnicity, giving an error rate of 0.14 percent, the team reports in the February American Journal of Human Genetics.

"This shows that people's self-identified race/ethnicity is a nearly perfect indicator of their genetic background," says Risch.

Yes, when you shade science to fit politics the commonsense becomes uncanny.

Posted by Orrin Judd at April 14, 2005 12:02 AM