Including more diverse populations in gene discovery efforts benefits all individuals, and may be especially beneficial for underserved populations.
Perhaps the most notable example of this is in the improvements to ‘fine-mapping.’ Fine-mapping is using statistics to attempt to identify which variant is responsible for an observed association in a GWAS. Because we inherit entire segments of chromosomes together, we jointly inherit all the variants contained on those segments. As a result, associations that are identified by GWAS may correspond to the effect of a variant on another part of the same chromosome. Due to differences between ancestry groups in the size and location of these segments, studying different groups allows us to narrow the search for the truly associated variant to a smaller number of genetic variants. Since there are fewer variants that need to be examined in laboratory studies, fine-mapping accelerates the process of linking genetic associations to potential biological pathways and the development of potential therapies. One example of leveraging multi-ethnic populations to better determine causal gene sets is in a publication on inflammatory bowel disease (Huang, 2017). By using a combination of European and East Asian cohorts, researchers were able to refine their GWAS signals to single variants with high certainty. This allowed them to link inflammatory bowel disease to specific immune cells and gut mucosa. Including diverse populations thus allows improved determination of strong genetic determinants and pathways for disease that can be targeted therapeutically, both for the minority populations themselves, but also for people of all ancestries.
Another benefit from including diverse populations in GWAS scans is that we have the potential to identify variants that are unique to different populations. Some genetic variants are only found in groups with certain ancestries. If a variant has an effect on a phenotype but is not found in samples with European ancestries, it would be impossible to discover the association between the variant and phenotype in people with only European ancestry. In contrast, studying diverse populations will discover these kinds of variants. Identifying more variants will help map out the biological causes of disease, which will benefit people of all ancestries.
Last but not least, GWAS have become large enough that polygenic scores are now very widely used in research settings and are even being considered for applications in clinical settings. However, given very large Eurocentric study biases in genetic studies, polygenic scores are currently far more accurate in populations with European ancestries. This means we can predict several traits and diseases in European ancestry populations rather well--similarly in fact to how accurately LDL cholesterol predicts heart disease. In contrast, these study biases mean polygenic scores are currently several-fold less accurate for example in East Asian and African ancestry populations, sometimes not much more predictive than a coin flip. One of our driving motivations to do genetic studies in more diverse participants and make these results widely accessible is to ensure that genetics can deliver on its mission to improve healthcare for all.
This study includes results for all phenotypes which had sufficient sample size to run a well-powered GWAS, including many behavioral phenotypes that were collected by the UK Biobank.
Several researchers, including some involved in this project, have also conducted GWAS of behavioral phenotypes in populations with only European ancestries. Behavioral research on these topics is particularly sensitive. We recommend that interested readers should also read the FAQs for those papers, that go into much greater detail on the interpretation and social and ethical implications of studying the genetics of behavior. Those FAQ may be found here: https://www.thessgac.org/faqs.
Genes do not determine our choices or who we become.
If they did, identical twins would make all of the same decisions, have the same interests, etc. Years of twin studies have shown that, while identical twins tend to be more similar than fraternal twins, they also tend to be different in a lot of ways. This suggests that environmental factors--such as culture, institutions, and policy--also play a large role in our phenotypes. And even for highly heritable phenotypes where many genetic variants are strongly associated with the phenotype, the associations that are identified by GWAS may not represent causal mechanisms. Furthermore, even when associations represent causal relationships, these causal pathways are complicated and interact with the environment. For example, imagine that there was a major pandemic that caused countries to shut down the public school systems. In such a scenario, we might imagine that the genetic influences that are related to academic achievement in the pandemic regime may be different than the influences that are related to academic achievement for those in formal public schooling.
That said, despite the limitations in interpreting genetic associations with behavior, this research is still valuable. For example, socioeconomic status (SES) is among the most important risk factors for many diseases and health outcomes. This means that any genetic variant that is associated with a disease may just reflect the association between the disease and SES. By analyzing GWAS of SES-related phenotypes alongside GWAS of disease phenotypes, researchers can focus on genetic variants that are associated with disease but not SES. These variants are more likely to represent strong biological risk factors of disease that can be tested rigorously in follow-up research.
No. Making policy or clinical decisions based on the results of this study would be incredibly premature.
GWAS look for associations between genetic variants and phenotypes in the world as it is today. Policy and clinical questions ask what the world would look like if we did things differently. GWAS cannot answer those questions. We are hopeful that the results of this study will facilitate future policy and clinical research for the global population, but those who use our results to make broad policy claims are overinterpreting the results.
Genetics research unfortunately has a legacy of racist research that has harmed minority populations.
Indeed, the term “eugenics” was coined in the late 1800s by one of most prominent early researchers in heredity, Francis Galton, and perhaps the most influential geneticist and statistician in history, Ronald Fisher, was an active proponent of the belief that socio-economic disparities in society were primarily driven by genetic factors in the early 1900s. These racist attitudes among several in the scientific community laid the groundwork for the “forced sterilizations, anti-miscegenation laws, and immigration restrictions of the 20th century.” These policies overwhelmingly targeted minorities and people of color.
In addition to racist policies in the name of science, a lack of community involvement in research also has led to harms to certain minority groups. For example, in the 1990s, members of the Havasupai tribe, a small Native American group based in the Grand Canyon, approached researchers at Arizona State University (ASU) asking if there was anything that could be done to treat the high rates of diabetes in their community. Blood samples taken from several tribe members, who were told that it would be used to “study the causes of behavioral/medical disorders.” Over the subsequent years, these samples were used in a variety of research projects beyond diabetes research. Some of these studies were about the tribe’s geographic ancestral origins, suggesting narratives that were in direct conflict with the tribe’s traditional stories of its origins in the Grand Canyon. Some worried that this research could be used to threaten their sovereign rights to their land. It was clear that many of the individuals whose blood samples were used in that research would not have consented if they had understood the scope of the projects that would be done. Ultimately, the Havasupai tribe sued ASU. As part of the settlement for the lawsuit, ASU returned the samples, some of the research that had been done was withdrawn, and the tribe received a large payment and other conconsessions from ASU. This story highlights how carefully executed research on topics that may seem benign to researchers can harm disadvantaged groups and sow distrust between researchers and the community.
Acknowledging the harm done to certain groups in the past in the name of science reminds us of the importance of careful communication of the implication of scientific research and intense vigilance that disadvantaged groups are not further harmed by this and other related work.
Yes, but excluding these groups from genetics research is also harmful. For this reason it is important to be aware of and mitigate potential harms.
As described above, genetics has a long history of being used to stigmatize certain groups. Although our results do not imply that phenotypic differences between groups are due to genetic or biological differences, we do anticipate that some racist individuals may mistakenly or willfully misinterpret our study to advance their ideological agendas. However, the exclusion of diverse groups from genetics research directly harms minority populations. When research is based only on one group, subsequent treatments and policies that are tested and implemented will most greatly benefit that population, exacerbating disparities. Remember that human populations are more genetically alike than we are different. Including a broader and more inclusive set of individuals helps support biological understanding for all groups, and does not imply a meaningful difference between them.
We have carried out a number of activities in an effort to maximize benefits and minimize risks from this research. See our response in “What has been done to reduce the potential harms of this research?” for more information.
We have adopted several strategies in an effort to reduce the potential harms of this research.
First, we have written this FAQ so that interested laypeople and the media can understand the value and the limitations of this work. We will treat this as a living document and welcome feedback from the community if any aspect of our analyses or their interpretation remain unclear.
Second, we also discussed this project and this FAQ with groups and individuals with diverse expertise and perspectives. For example, we met on several occasions with members of Shades@Broad, an identity-based affinity group whose mission is to advocate for and support the recruitment, development, and success of ethnic minorities at the Broad. We sought their comments, advice, and perspectives throughout the analysis, interpretation, and dissemination of results. We also obtained feedback on our research and this FAQ from researchers and clinicians who work with diverse communities across the US.
Third, Shawneequa Callier, a bioethics professor who specializes in the ethical, legal, and social implications of genetics research and racial categories reviewed and provided feedback on this FAQ. Professor Callier advised us on ethical considerations surrounding this research and provided comments on the draft of the text of this FAQ. She also helped with the drafting of related manuscripts.
Even with these efforts, it is still likely that some will misinterpret this work. As such occasions arise, we will attempt to correct the public record with firmness where appropriate.