A risk in treatments based on race

In the still-young field of personalized medicine, one goal has been to determine if medical treatments could be guided by the concept of race.

But some caution is in order, according to a new study in Proceedings of the National Sciences, codirected by a University of Pennsylvania geneticist.

In a sample of 365 people who identified themselves as African American, their proportion of European ancestry ranged from less than 1 percent to - in a handful of cases - more than 99 percent.

So if a physician diagnoses anemia in an African American patient, for example, he or she should not automatically conclude that it's the sickle-cell variety, more prevalent in African Americans, says Penn's Sarah A. Tishkoff.

Working with Cornell University's Carlos Bustamante, Tishkoff and others analyzed 300,000 genetic markers from each of the 365 African Americans, and did the same with 203 people from western Africa and 400 from Europe.

The work offered some sobering reminders of the slave trade.

For example, the information on the X chromosomes was predominantly of African origin, whereas the information on the Y chromosome revealed more European influence - consistent with plantation-era forced coupling between white men and their female slaves.

Also, the genomes of the African Americans generally did not bear a strong influence from any particular tribe - consistent with people from one village being scattered among different owners upon coming to North America.

Tishkoff says the findings raise questions about some of the commercial ancestry-testing services that purport to reveal tribal origin.

"African Americans," she says, "appear to have ancestry from all over West Africa."

- Tom Avril