An estimate of genetic ancestry is typically included as a control in a genetic association test. This estimate of ancestry dovetails closely with conceptualization of human populations in genetic analysis. Other fields have different ways and different motivations for grouping individuals into populations, such as the OMB categories for race. From a social science point of view the stakes are high; they fear that the way statistical geneticists are starting to use the concept of ancestry will end up reifying race as a biological category. This project aims first to describe, using a schema that we will develop, how the population concept is used across different disciplines relevant to the study of genetic associations. This will help identify points of tension. We will attempt to resolve these through a normative project that refines this schema by identifying how the terms should be used. This will be an interdisciplinary endeavor resulting in guidelines for use of the population and ancestry concepts across these disciplines.
The parent grant to this Administrative Supplement aims to develop methods to analyze genetic data, in order to better detect and interpret associations between genetic variation and traits, with a focus on psychiatric disease. A key aspect of the grant is that the methods be freely available and readily adoptable by the broader genetics community. This Administrative Supplement is focused on the interpretative side of this work, and in particular on the concept of a population and the related concept of ancestry. The process of identifying genetic associations involves averaging over individual genomes. The methodological choices of which genomes to average over, and how, relate to deep seated questions about identity, family history, and social norms, all of which may also influence the traits under study. There are two starting points of the work we propose. The first is that within statistical genetics there is a clear consensus that the methodologies employed are very sensitive to choice of population,1 and that application of current methodologies to current data sets produces very different strengths of association in different populations.2 And moreover, that if the existing results are applied either as tools of public health or as tools within clinical medicine, we run the risk of exacerbating health disparities.3 Within statistical genetics, there is not consistent usage of labeling populations, and this usage has been in flux.4 The second starting point is that numerous disciplines will necessarily be involved in realizing the promise of genomic medicine, and the concepts of population and ancestry are not used consistently across these disciplinary boundaries. This raises far more than academic concerns; social scientists are concerned that the way that these concepts are being adopted in statistical genetics will end up reifying race, a social construct.5 The stakes are hence high, and the timing right, for an interdisciplinary conversation dedicated to clarifying how we do and should use the concepts of population and ancestry in connection with genetic association studies. Many statistical geneticists are eager for guidance both around use of language for these concepts, and for input on how to think through the historical mutual interdependencies of the grouping of individuals and the traits under study. Ben Neale, PI of the parent grant, and Danielle Allen, Director of the E J Safra Center for Ethics at Harvard and the Intellectual Lead for this Administrative Supplement, have been in conversation about these issues since an interdisciplinary workshop organized by Danielle held in October 2018. The research agenda proposed here arose out of that workshop. The interdisciplinary conversation started at that workshop will continue as part of this project; we have a working group of about 20 relevant experts from a broad range of backgrounds who we will be engaging via monthly virtual meetings, and (funding depending) at two 1.5 day workshops. Under the intellectual supervision of Ben and Danielle, Anna Lewis ? who worked in the genetics testing industry for several years and who is retraining in bioethics ? will drive this research agenda forward. A second postdoc with expertise in the relevant empirical social science methodologies will carry out much of the design and implementation of the data gathering and analysis strategies.
