17 May 2011. Michael Jackson sang, “It don’t matter if you’re black or white”—and that goes for people in the doctor’s office, too, according to a paper in the May PLoS ONE. A joint team from Loyola University in Chicago, Illinois, and Mount Sinai Medical Center in New York analyzed nearly 1,000 people’s genomes to conclude that individual human beings lie all across a broad genetic spectrum, not in narrow racial hues. However, they did find that race matters at the population level, with chunks of DNA inherited in different patterns in peoples of different ancestry. The results suggest that disease-linked single nucleotide polymorphisms (SNPs) from one ethnic group may not translate to another, and that personalized medicine will have to be just that—truly tailored to the person.
In addition to the ongoing interest in disease-causing SNPs, researchers are also seeking variants that indicate how well a person will respond to one medicine or another (Urban, 2010). Armed with such data, doctors could offer people a customized panel of drugs most likely to work. In AD, this may be relevant one day in the future for people who carry certain risk alleles. More generally, that dream will only become reality if physicians have good genetic markers to work with. But will markers based on one ethnic group work for another? First author Bamidele Tayo from Loyola University and senior author Erwin Bottinger of Mount Sinai led the study to analyze how much genomic heterogeneity there is among people and ethnicities. At Mount Sinai, they collected samples from nearly 1,000 people of African, European, or Hispanic descent, and used gene array chips to identify more than 800,000 SNPs in each person. They discovered heterogeneity both within and between ethnic groups, suggesting that many of their subjects were of mixed descent.
The existence of genetically blended races is not surprising, particularly in the bubbling melting pot that is New York, wrote John Hardy of University College London, U.K., in an e-mail to ARF. “The descriptions used, especially in the U.S., to self-identify racial groups are really historical artifacts with little meaning from a purely genetic perspective.” Everyone of mixed descent who has tried to check the right box on a census form will probably agree.
By contrast, when considering the genes of groups of people, ethnicity does matter. It is particularly important in genomewide association studies. In GWAS, marker SNPs stand proxy for the functional mutations the scientists are really after, but may not always be co-inherited with them. For example, the study authors point to the fat mass- and obesity-associated (FTO) gene, which encodes an enzyme that demethylates DNA. A variant in the FTO locus is strongly linked to obesity in those of European descent, but is much more weakly linked in people with African ancestry (Adeyemo et al., 2010). This may sound like tough luck for the obesity field, but lo and behold! This gene has already shown up in an AD study as well (see ARF related news story on Ho et al., 2010).
SNPs only point to the right mutation if the SNP and mutation are inherited together in the same chunk of DNA called a linkage disequilibrium block. In their NYC samples, the researchers found that some genetic loci are passed on in different linkage disequilibrium blocks in different subjects who are ostensibly of the same “race.” For example, at the FTO region there were five such blocks in African Americans, two in European Americans, and four in Hispanics. Thus, the study raises the question of whether SNPs from one ethnic group match up with the mutation locus in another. “Until the ‘causal mutation’ is known for [GWAS] findings, it may not be possible to transfer the information from the original study population, virtually always of European ancestry,” the authors wrote.
The researchers also considered each genotype individually. When they did so, the boundaries between ethnicities were considerably blurred, with Hispanic genomes fitting in between those of European and African Americans. In particular, Hispanic and African American genomes shared more alleles. “If one dispenses with the conventional labels and relies solely on genotype, the Hispanics…can be appropriately merged with the African Americans,” the authors concluded. “In the genomic era, conventional racial/ethnic labels are of little value.”
What does that mean for a doctor faced with a patient? At least in diverse areas such as New York, it may be useless to consider the race that person writes on a form in diagnosing or selecting treatments (Cooper et al., 2003). Similarly, researchers might be using artificial categories if they stratify people by race (Weiss and Lambert, 2010). The study authors are now analyzing DNA samples from 20,000 folks to discover genetic markers that work best for the multiethnic New York populace—data that will bring them one step closer to personalized medicine.—Amber Dance.
Tayo BO, Teil M, Tong L, Qin H, Khitrov G, Zhang W, Song Q, Gottesman O, Zhu X, Pereira AC, Cooper RS, Bottinger EP. Genetic background of patients from a university medical center in Manhattan: Implications for personalized medicine. PLoS One. 2011 May:6(5):e19166. Abstract