Sharon Begley, "Race Is Only Skin Deep, So It Isn't a Solid Basis For Medical Decisions" (2003)

BY Sharon Begley
> <strong><font size="2">Race Is Only Skin Deep, So It Isn't a Solid Basis For Medical Decisions<font>

> WOE BETIDE scientists who find themselves on the &quot;politically correct&quot; side of a debate. All their opponents have to do is slap that label on them and they're toast.<p>

So it has gone for biologists who doubt that there is a deeper, scientific meaning to the popular conception of race, typically defined by skin color, facial features and hair type. They have been met with disdain or worse by opponents. If race is only skin deep, ask these critics, how can forensic anthropologists infer the race of a skeleton? How can blacks respond more poorly to some heart drugs than whites?

The debate over the biological validity of race has been largely abstract, but that's about to change. Earlier this year, the U.S. Food and Drug Administration recommended that scientists running clinical trials of new drugs ask participants their race and ethnicity: American Indian, Asian, black, white, or Pacific Islander; and Hispanic/Latino or not. The premise is that your physiological response to a drug, complications as well as how well it works, reflects your genes, and that those genes can be predicted by which of the 10 categories you fall into. Race couldn't get any more biologically valid than that.

IN THE WAKE of the FDA's "draft guidance," however, prominent geneticists are warning that analyzing drug response by self-reported race is a recipe for medical disaster. For one thing, self-reporting is inherently problematic. Also, six million Americans in the 2000 census picked two or more races to describe themselves. But the problem is more fundamental.

Geneticists, fresh from the triumph of the human-genome project, find that when they add up the tiny genetic variations that make one person different from the next, there are more differences within races than between races. In, December, scientists led by Marcus Feldman of Stanford University in California reported that 93% to 95% of genetic variations reflect within-group differences. Less than 5% reflect differences between races.

Geneticists are finding something even more startling. Obviously, people whose ancestors lived for generations in one region of the world look different from people whose ancestors inhabited a climatically different region. The descendants of Vikings look different from the descendants of Nubians. Skin color, facial features and hair type reflect genes selected by evolution to allow their carriers to thrive in the place they call home.

Ancestry does reflect something genetically real, explaining why forensics can identify a skeleton's race, says anthropologist C. Loring Brace of the University of Michigan, Ann Arbor. Some DNA sequences differ sufficiently between populations so that scientists can tell that a skeleton comes from someone whose ancestors were Bretons and not Maori, or Lapps and not Manchurians. That, of course, also tells you the native continent, and thus race, of those ancestors.

Race in this sense "is too broad," says Prof. Feldman. "Assigning a racial classification to someone hides the biological information that you need to make intelligent medical decisions. An African-American or Hispanic could have any mix of ancestries."

Those who believe that races reflect something biologically important assume that superficial genetic traits "are characteristic of the overall genome," he says, that large differences between races in genes for skin color imply large genetic differences in general between races.

IT TURNS OUT, however, that genes coding J. for skin color, facial features and hair type are atypical. As Prof. Feldman and colleagues write in the journal Nature, "They vary between races much more than genes in general." Thousands of other genetic traits invisible to the naked eye don't neatly sort themselves by one's ancestral continent. If you group people by ABO blood groups, Europeans and Africans are more closely related. If you group people by genetic traits that matter medically-the justification for race-based health data-the utility of race is just about nil.

Response to medication reflects, in part, a patient's genes. But because these genes don't reflect race, race is a lousy proxy for medically important genes, argues Craig Venter of the Center for the Advancement of Genomics, Rockville, Maryland, who led the private effort to sequence the human genome. Analyzing response to drugs based on self-reported race, therefore, "could result in misleading and biologically meaningless conclusions," he and TCAG's Susanne Haga argue in the journal Science.

Sickle-cell anemia, for instance, is more common in parts of sub-Saharan Africa than it is globally. But if you focus sickle-cell prevention efforts on blacks, you'll miss the millions of southern Italians, eastern Mediterraneans, Middle Easterners and southern Indians who also carry the defective gene. Wherever malaria was common, so is the sickle-cell gene. In this sense, the concept of race is too narrow.

Knowing ancestry, in contrast, could well prove useful in understanding variable responses to drugs. (Best of all is a patient's actual genetic information.) Because race alone can mislead, confusing race and ancestry could be devastating for medical practice. Sometimes, politically correct is scientifically correct, too.