The subject of race in America is a lightning rod that many researchers, especially geneticists, prefer to avoid. But a population geneticist at Stanford University has grabbed it with both hands, claiming that 'colour-blind' approaches to medical genetics could actually discriminate against minorities.

The geneticist, Neil Risch, claims in a commentary article that political sensitivities may be leading some geneticists to abandon race as a variable in their studies — an approach that he thinks is dangerously misguided. “We need to value diversity, not to fear it,” Risch and his colleagues argue in the online journal Genome Biology (see http://genomebiology.com/2002/3/7/comment/2007). “Ignoring our differences, even with the best of intentions, will ultimately lead to the disservice of those who are in the minority,” they say.

Geneticists agree that a diverse population such as that of the United States cannot be considered as a homogeneous group. So, when trying to map the genes that confer susceptibility to particular diseases, or studying the performance of new drugs, it is useful to work out if the population can be divided into subgroups with particular genetic characteristics that might affect the results. One subgroup, for instance, might fail to respond to a drug that works well in the rest of the population.

Some geneticists are therefore studying genetic marker sequences to determine how people cluster into subgroups within the overall population. Last October, for instance, a team led by David Goldstein at University College London used some 40 markers to subdivide a genetically diverse sample of people from various populations throughout the world. They found that the resulting subgroups showed differences in the genes for enzymes that metabolize drugs (see J. F. Wilson et al. Nature Genet. 29, 265–269; 2001). An accompanying editorial praised this “race-neutral” approach to population-genetic studies.

But Risch maintains that self-reported labels of racial origin — such as Caucasian, Asian, African American and Hispanic — can do the job just as well as sophisticated gene-clustering studies. “Effectively, you're just going to recreate those racial groups,” he argues. What's more, Risch claims, if the data on genetic clustering are collected without regard to race, there is a tendency to assume that differences between the resulting subgroups are due to genetics — when they might actually be due to socio-economic or cultural factors that just happen to correlate with race.

Risch also argues that minority groups should be selectively recruited into biomedical studies to ensure that they are sufficiently well represented to reveal any important differences from the population as a whole. Take a race-neutral approach, says Risch, and important information about minority health may be lost.

Not surprisingly, some scientists take issue with Risch's arguments. Harold Freeman, a prominent African-American researcher who is director of the National Cancer Institute's Center to Reduce Cancer Health Disparities, based in Rockville, Maryland, agrees that cultural factors should be considered in genetic studies. But he argues that racial labels are too crude to be useful. “Culture is not equal to race,” says Freeman.

Worse, Freeman fears that the use of race as a variable in biomedical studies will perpetuate historical discriminatory attitudes. He adds that statistical findings from population-genetic studies may be wrongly applied to racial groups as a whole. If a drug is found to work less well in African Americans than in Caucasians, for example, it may end up being denied to individuals within the black population who might nevertheless benefit from it.

Goldstein worries that the tone of the Risch paper will bring unwanted political overtones into what ought to be a technical discussion. He accepts that it can be useful to divide populations into subgroups that reflect the geographical origins of their members' ancestors. However, Goldstein says that “race is not a terribly good framework. I believe we can do better. This is a technical question. I think that Neil has politicized it too much.”

Risch argues that his article is merely responding to statements made in the scientific literature by other commentators. For instance, an editorial in The New England Journal of Medicine last year argued that “race is biologically meaningless” (see R. S. Schwartz N. Engl. J. Med. 344, 1392–1393; 2001).

Francis Collins, director of the National Human Genome Research Institute in Bethesda, Maryland, feels that Risch makes some valid points. He points out that the international 'haplotype map' project, an effort to discover disease-susceptibility genes (see Nature 412, 105; 2001), includes DNA samples from the main geographical areas of the world. But Collins thinks that Risch was unwise to frame the issue in terms of racial labels. “It's open to broad misinterpretation,” he says. “I'd be happier if we could get away from the highly charged terminology of race and refer instead to geographical origin of ancestors.”