Goal and description
Over 3,000 successful Genome-Wide Association Studies (GWAS) have identified and validated a plethora of genetic variants associated with hundreds of phenotypes. These have improved our biological understanding of a multitude of traits and diseases ranging from hair color to diabetes, and the results provide the basis for implementing genome-based medicine.
Despite these success stories, the current approaches have been criticized on multiple fronts. A key concern is that an overwhelming majority of study participants have been recruited among individuals of European ancestry (1,2). For example, 96% of samples in global GWAS up to 2011 were from individuals of European origin for reports1. This has marginally improved by 2016 to include 81% of Europeans and 15% Asians, with only 4% for all other populations(2). The issue of representation is further compounded by uneven sampling within continental groups. It is unclear how association knowledge gained in European panels is informative about individuals with African, Asian, and Native American ancestry(3), particularly when rare genetic variants are considered(4). By contrast, it is perfectly clear that focusing on a single group reduces our ability to discover important aspects of human biology and to make medical application of the results, since we are exploring only a tiny subset of functional variation present in the human population.
The overrepresentation of individuals of European ancestry near major medical centres, where much of the research is conducted, plays a role in creating bias, but this cannot fully account for it. Even within Europe and North America, most research institutions are located in cosmopolitan centres that have much more diverse populations: among the largest 5 cities in Canada, the proportion of visible minority populations range from 24% in Ottawa to 49% in Toronto, but the strong bias towards European ancestry remains. A second key factor is statistical convenience-- homogeneous samples are easier to model and interpret. This drive towards uniformity sharply biases samples towards majority populations.
However, statistical convenience is not the best guiding principle for designing genomic and other medical studies, and there has been a growing realization that studying more diverse populations is beneficial on scientific, medical, and ethical fronts. Broadening ethnic representation in medical genomics research is a key mechanism for redressing this issue and accelerating the discovery of genetic variants associated with chronic, heritable diseases across diverse human populations, an essential prelude for implementing personalized or precision medicine. An increasing number of cohorts of individuals of diverse ancestries are being assembled and genotyped or sequenced. However, important barriers remain in recruitment and in the statistical analysis of more diverse cohorts.
The goal of this symposium is to bring together geneticists, statisticians, cohort leaders, and leaders in underrepresented communities to help overcome these barriers and facilitate the inclusion of more diverse populations in medical research as scientists and as study participants.
The symposium leadership has wide previous experience in the organization of international conferences to draw upon here. For example, Dr. Lathrop organizes the International Workshop on Genomic Epidemiology, which has occurred biannually since 2002, with the most recent versions in Paris (2013), London (2015) and Barcelona (2017).
(1) Bustamante, Carlos D., M. Francisco, and Esteban G. Burchard. "Genomics for the world." Nature 475.7355 (2011): 163-165.
(2) Popejoy, Alice B., and Stephanie M. Fullerton. "Genomics is failing on diversity." Nature 538.7624 (2016): 161.
(3) De Bakker, Paul IW, et al. "Transferability of tag SNPs in genetic association studies in multiple populations." Nature genetics 38.11 (2006): 1298
(4) Gravel, Simon, et al. "Demographic history and rare allele sharing among human populations." Proceedings of the National Academy of Sciences 108.29 (2011): 11983-11988.