A new study of families from Mexico City has shown how genetic data from relatives can help distinguish inherited genetic effects from social and environmental factors when studying associations between genetic ancestry and complex traits. The study, led by researchers at Oxford Population Health and Universidad Nacional Autónoma de México (UNAM), is published in Nature.
Populations with different genetic ancestries—reflecting patterns of genetic variation inherited from ancestral populations—can differ on average in disease prevalence and other characteristics. But comparing groups only at the population level cannot tell us why these differences occur. Genetic ancestry can be correlated with social and environmental factors, making it difficult to determine whether an observed association reflects inherited genetic differences, environmental circumstances, or a combination of both.
The researchers addressed this problem using a family-based design within a large, genetically admixed population. Because siblings inherit different combinations of their parents' DNA at random, comparing ancestry differences within families can substantially reduce the ancestry-environment confounding that affects comparisons between unrelated people.
The team analysed genetic and health data from more than 92,000 adults in the Mexico City Prospective Study (MCPS), including over 39,000 people from more than 17,000 families. They estimated each person's proportion of Indigenous American, European, African and East Asian genetic ancestry and examined relationships with 15 complex traits, both across the population and within families.
Key findings:
-
Indigenous American ancestry was strongly associated with shorter height and higher type 2 diabetes risk at the population level, and significant ancestry effects were also observed within families. This supports an inherited genetic contribution to these differences in this population, which was further supported by analyses of trait-associated genetic variants;
-
In contrast, a strong population-level association between Indigenous American ancestry and lower educational attainment was not seen within families. This suggests that the population-level association is driven by social or environmental factors correlated with ancestry rather than inherited genetic differences;
-
The researchers also found evidence that natural selection may have contributed to differences in height-associated variants between Indigenous American and European ancestries. The evidence for selection on type 2 diabetes-associated variants was inconclusive.
Siqi Wang, Postdoctoral Researcher at Oxford Population Health and first author of the study, said ‘Genetically admixed participants from MCPS provided a valuable opportunity to examine ancestry effects both within and between families. A major challenge was making the best use of genetic data from thousands of families with different structures. By properly accounting for these family relationships, we could get a much clearer picture of which population-level associations reflected inherited genetic effects.’
Jason Torres, Genetic Epidemiologist Lead at Oxford Population Health, who co-led the study, added:
One of the main lessons from this study is that differences associated with genetic ancestry across a population shouldn’t automatically be assumed to have a genetic explanation. We saw this clearly for educational attainment: there was a large association with ancestry across the cohort, but essentially no corresponding effect within families. For height and type 2 diabetes, however, the family results supported an inherited genetic contribution, which was also supported by differences in trait-associated alleles.
Different traits can therefore give very different answers. We hope this family-based approach provides a principled framework for researchers investigating these questions in other populations, and across other traits and diseases.
Jaime Berumen Campos, co-Principal Investigator of the MCPS and Professor of Molecular Genetics at UNAM, said
A major strength of the Mexico City Prospective Study is the combination of substantial genetic diversity and a large number of participants who are related to one another. Rather than representing a limitation of the cohort, this family structure has proved to be an important advantage relative to many other large population studies, as it has allowed us to carry out family-based analyses that are not possible in most cohorts.
Peter Visscher, Professor of Quantitative Genetics at the University of Oxford’s Big Data Institute, commented
What this study adds is separating ancestry differences into between-family and within-family components while accounting for the genetic similarity between siblings. Because differences in ancestry between siblings arise from random genetic segregation, the within-family comparison provides a robust way to estimate genetic ancestry effects that is much less susceptible to the environmental confounding that affects comparisons between unrelated individuals.
The researchers emphasise that genetic ancestry is not equivalent to race and that human populations are not genetically discrete. The findings relate specifically to participants in the Mexico City Prospective Study and should not be extrapolated to other populations or traits without further evidence.