Self-Reported Race and Genetic Admixture

Self-Reported Race and Genetic Admixtureto the editor: The use of data on self-reported race in health research has been highly debated.1,2 For example, Burchard et al. recently argued that important information on disease susceptibility may be derived from the use of data on selfreported race, 1 whereas Cooper et al. 2 cited Wilson et al., 3 who argued that ethnic labels "are inaccurate representations of the inferred genetic clusters."Cooper et al., however, ignored later work 4 that identified limitations in the analyses of Wilson et al. -specifically, inappropriate classification of groups, the use of a suboptimal model for cluster identification, and reliance on only 39 microsatellite markers for cluster analyses.With larger numbers of markers, it was shown that genetically distinct groups can be almost completely inferred from self-reported race.5 Despite advances in genetic analysis, clinicians and epidemiologists still disagree on the usefulness of race in clinical reporting.For example, sleep apnea is a condition that is more common in black children than in white children and is associated with distinct craniofacial features.Although some of the group differences may relate to environmental factors, it is unlikely that these would substantially influence facial structure and other relevant traits.With support from a U.S. Public Health Service grant, we applied an admixture analysis 6 to a sample population in Cleveland.Participants were clearly separated into unique groups with the use of this genetic approach.Whereas 93 percent of self-reported whites were classified as having predominantly European ancestry, less than 2 percent of blacks were so classified.Only 4 percent who reported their race as black had predominantly African ancestry; yet, the admixture proportions of this group made it possible to separate the population into two groups, in which 94 percent of self-reported blacks and 7 percent of self-reported whites were classified as being of mixed race (Fig. 1).The sharp peak at the left in Figure 1 indicates that there are many persons who have no African ancestry (i.e., the values correspond to those of self-reported whites), and the broad peak at the right indicates that most blacks are of mixed race and do not originate from any single population.Thus, self-