One over PAR or one under PAR: vive la différence

The concept of race based on physical appearance, social grouping, or self-identification has been challenged for decades. With the recent sequencing of thousands of human genomes, this concept has been further redefined. In this issue of Blood, Heenkenda et al show that the A allele frequency of the F2RL3 rs773902 A/G dimorphism associated with increased platelet reactivity is significantly lower in Somalians than what has been previously described for blacks in the United States, underscoring the importance of genetics and geographic ancestry for determining disease risk.¹ 

Most old and many new medical textbooks describe diseases by using a racial predilection, often without defining the variable “race.” More recently, epidemiologic research has begun using self-identified race and ethnicity as a convenient and probably useful definition. Regardless of the definition, it has been claimed for decades that black individuals in the United States have an increased risk of myocardial infarction and stroke when compared with white individuals, even after accounting for demographic and socioeconomic risk factors. Because occlusive platelet thrombi play a central role in ischemic vascular events, it was of interest to learn that protease-activated receptor 4 (PAR4) activation was significantly higher in blacks.²  Furthermore, the single nucleotide polymorphism rs773902 in F2RL3 (which encodes PAR4) was responsible for ∼50% of the racial variation in platelet aggregation.³  rs773902 determines an alanine or threonine at position 120 in PAR4, and the hyperreactive Thr120 allele had an allele frequency 3 times higher in blacks than in whites. The report by Heenkenda et al determined that the allele frequency of the rs773902 A allele in the Somali population is 38%, much lower than the 63% reported for blacks. This finding underscores the complex and diverse genetic structure of populations in Africa and the limitations of using race or ethnicity in genetic associations studies.

Accumulating fossil and genetic evidence show that modern humans are descended from ancestors that lived in Africa approximately 250 000 years ago and started migrating out of the continent between 50 000 and 80 000 years ago, thus populating the rest of the planet and giving origin to the genetic subsets of populations that comprise the modern world.⁴  Time and human migration out of and back into Africa have resulted in substantially greater genetic diversity, significant population substructure, and shorter shared segments of DNA represented by lower linkage disequilibrium. Recent genome-wide association studies have elucidated the genetic basis of the Gaussian variation in skin pigmentation within populations in Africa providing evidence not only for the selective pressure for darker pigmentation that allows for adaptive protection against UV radiation but also for a 30 000-year-old gene variant likely introduced via migration back to Africa, underscoring the complicated genetic architecture of this complex trait.⁵  The report by Heenkenda et al illustrates another example of African genetic diversity by studies of allele frequencies of F2RL3.

There are more than 50 countries in Africa and many more ethno-linguistic groups. African slaves brought to North America in the sixteenth through the nineteenth centuries largely came from the west coast, including the modern-day countries of Ivory Coast, Ghana, Togo, Benin, Nigeria, Cameroon, Republic of Congo, Democratic Republic of Congo, Gabon, and Angola. As pointed out by Heenkenda et al, these countries dominate the African continental genomic data in the 1000 Genomes Project, which may better represent genomic variation in North Americans of African ancestry. Indeed, the F2RL3 allele frequencies in the cities of Houston and Philadelphia are similar to those observed in Nigerians.³  Because Somalia is on the east coast of Africa, it would not be the obvious region to examine when seeking the source of genetic diversity in North Americans with African ancestry. Nevertheless, Somalia is sub-Saharan, and neighboring Chad has been shown to be a target of Eurasian backflow, which may explain the lower F2RL3 allele frequency in this region of Africa.

The individuals brought from Africa as slaves eventually were named African Americans or blacks, emphasizing the continental origin of their ancestors and the color of their skin. These rather imperfect definitions, sometimes characterized as social constructs, have been adopted by the government and used in decennial censuses, administrative forms and, more importantly, in medical and clinical research. Because most self-identified African Americans carry a genome that is an average of 80% African,⁶  some epidemiologic studies attempting to identify risk factors for diseases have been successful⁷ ; however, this has also led to misdiagnoses.⁸ 

The Heenkenda et al report raises the important issue of the appropriate use of the concept of race in the medical field. It also challenges the scientific community with a simple example of allele frequencies to consider genomics and geographic ancestry as the pillars that define population structure in medical trials. We believe the concept of use of racial predilections for disease is a severely limited approach. The use of race as a variable in association studies is controversial, hard to define, has the potential to be discriminatory, and is often inaccurate for any biologic trait. Furthermore, there are now countless examples of genetic variants that are established as phenotype altering or disease causing when the allele frequencies substantially vary among groups of differing geographical ancestry, such as MYH9 and end-stage renal disease and the chr8q24 locus for prostate cancer.^9,10  In other words, association studies that use a genetic variant as the dependent variable may improve the many problems generated by using race. Heenkenda et al suggest using the term “African American of West African ancestry.” Although cumbersome, this is certainly more accurate than black, African American, or sub-Saharan African. From a genetic epidemiologic point of view, in 10 years even this term would almost certainly lack the precision that will hopefully better predict disease associations.

In the age of home genetics testing and ancestry kits and the efforts by scientists to develop precision medicine approaches, the findings by Heenkenda et al represent a timely reminder that we are more diverse than we previously thought and that this diversity has significant biological implications (see figure).