Summary¹

Genetics is the study of heredity, specifically the mechanisms by which traits or characteristics are transmitted from one generation to the next. Because it is applicable to many areas of human life, genetics garners wide interest. Researchers use human genetic information to address a variety of questions about human history and evolution, human biology, diseases, and heritable traits (e.g., height or serum cholesterol). Researchers have frequently used population descriptors as a shorthand for capturing the continuous and complex patterns of human genetic variation resulting from history, migration, and evolution. Of particular concern is the long-standing use of race, and more recently ethnicity, as this shorthand. In humans, race is a socially constructed designation, a misleading and harmful surrogate for population genetic differences, and has a long history of being incorrectly identified as the major genetic reason for phenotypic differences between groups. Rather, human genetic variation is the result of many forces—historical, social, biological—and no single variable fully represents this complexity (see Chapter 1). The structure of genetic variation results from repeated human population mixing and movements across time, yet the misconception that human beings can be naturally divided into biologically distinguishable races has been extremely resilient and has become embedded in scientific research, medical practice and technologies, and formal education. Many elements of racial thinking, including essentialism and biological determinism, have influenced modern thinking around human genetics, to the marginalization of some peoples and the benefit of others.

___________________

¹ References are not included in this report summary. Citations appear in subsequent report chapters.

Derived from a person’s complete set of DNA, genomic information is increasingly easy and inexpensive to produce, and tools to analyze genetic information are widely available. Accordingly, genetic and genomic information has become far more accessible, and research using human genetic data has grown exponentially over the last decade. The use of genetic information is now widespread across biomedical research, and genetics and genomics research is now conducted by a range of investigators across different disciplines, creating a need for clarity and providing an important opportunity to implement substantive changes to the ways population descriptors are used. Clear guidance about the use of population descriptors is needed before mistakes of the past are integrated into this new era of genomics research.

Race and racism have recently gained renewed attention from the U.S. scientific community. Recognition by the U.S. biomedical research community of the need to address the complex issue of population descriptors in genetics research has never been greater. Although the history of prior attempts to address population descriptors in genetics and genomics research—and the lack of notable change—may create some skepticism about the usefulness of another report aiming to create best practices for this complex area, this is a crucial moment to offer concrete guidance to the research community.

STUDY CHARGE

The study sponsor, the National Institutes of Health (NIH), asked the National Academies to conduct a study to review and assess existing methodologies, benefits, and challenges in using race, ethnicity, ancestry, and other population descriptors in genomics research.² The statement of task emphasizes the use of appropriate and valid population descriptors in genomics research, and focuses on understanding the current use of population descriptors in genomics research; examining best practices for researchers in the use of race, ethnicity, and ancestry as population descriptors; and identifying how best practices in the use of population descriptors could be widely adopted within the biomedical and scientific communities to strengthen genetics and genomics research.³ To accomplish this task, the

___________________

² The full statement of task is presented in Chapter 1 along with a discussion of what was in and out of scope.

³ The statement of task also identified four areas that are beyond the scope of this committee’s recommendations: examining the use of race and ethnicity in clinical care; examining racism in science and genomics; examining the use of race and ethnicity in biomedical research generally (e.g., beyond genetics and genomics research); and providing policy recommendations to NIH and government agencies. See the section “What Is the Goal of This Report?” in Chapter 1 for more detail.

National Academies empaneled a committee of 17 members with expertise in population genetics, human and clinical genetics, genetic epidemiology, statistical and computational genetics and genomics, anthropology, sociology, social epidemiology, demography and population statistics, as well as historical, ethical, legal, and social implications research. Given the charge, researchers who use genetic and genomic data are the primary audience for the report, especially the more technical recommendations and best practices (Chapter 5). However, much of the report is intended for a broader audience (see Chapters 3, 4, and 6).

GUIDING PRINCIPLES FOR THE USE OF POPULATION DESCRIPTORS

Human populations can be described according to countless characteristics: urban versus rural, for example, or smokers versus nonsmokers. The use of such descriptors as race, ethnicity, or ancestry, however, focuses on “descent-associated” groups—sets of individuals whose members are thought to share some characteristic that derives from their common origin (see Box S-1 for key terms). Importantly, the inclusion of population descriptors in genomics, and which specific ones to include, must be a deliberate decision because their use has high stakes for ensuring that research benefits society and mitigates against potential harm, such as race-based health inequities.

The committee considered a range of population descriptors, each revolving around a somewhat distinct feature of human difference and thus offering researchers a specific tool that is more appropriate for some uses than for others. Over the course of the committee’s work, the following population descriptors emerged as most relevant to the committee’s charge: ancestry, geography, ethnicity, indigeneity, and race/racialized groups. To support researchers making reasoned, deliberate choices in their selection of population descriptors, the description and discussion of each demonstrate what these concepts of human difference can—or cannot—capture in genetics studies.

Although some genetics studies have used descriptors like race as proxies for genetic variation, some genetic epidemiologic studies rely on descriptors like race as proxies for cultural beliefs and practices or for shared environments, in the absence of direct measurements of these latter contextual factors. The environment is the complex of physical, social, chemical, and biotic factors that act upon a person or a community and also shape its form and survival. Social context, an attribute of environment, influences behavior and interacts dynamically with biology, including genetics, throughout the life course to affect human health. Given that human genotypes are not randomly distributed across environmental conditions,

leading to correlations of genetic and environmental effects, and given the extensive evidence for gene-by-environment interactions in experimental organisms, the committee concluded that environmental factors should often be considered alongside population descriptors in genomics research.

Emerging from the mistakes of past and current use of population descriptors is an imperative to transform not only the use of these descriptors but also the field of genomics research. For the recommendations that follow to successfully advance the appropriate use of population descriptors in genetics and genomics research, they must be grounded in ethical and empirical principles that engender trust and drive trustworthiness of research. Accordingly, the committee developed a set of guiding principles that mutually reinforce one another and undergird the recommendations (see Figure S-1). The guiding principles are respect, beneficence, equity and justice, validity and reproducibility, and transparency and replicability (Chapter 3).

In short, respect for individual and community preferences, norms, and values should inform approaches when determining what population descriptors to use in research. The principle of beneficence calls on researchers to assess how the selection of population descriptors may not only generate potential good but also potential harm and requires consideration of the effect of population descriptors on health equity. A commitment to equity and justice requires determining whether and how the selection and use of population descriptors will produce equitable benefit to avoid reinforcing existing inequities or introducing new ones. Upholding validity and reproducibility requires judicious evaluation of research objectives and assessment of the appropriateness and purpose of including population descriptors. Transparency and replicability include the obligation to provide a clear rationale for the selection and/or use of population descriptors and to explain decision-making processes in an open and accessible manner to both other researchers and research participants, thus enhancing replicability.

Given the dynamic nature of research and the limitations of this report to fully capture the range of possible use cases in future genetics and genomics research, the guiding principles also provide a foundation and common vocabulary for researchers and other relevant parties to engage in future decision making for contexts that may not be addressed directly in this report.

RECOMMENDATIONS FOR THE USE OF POPULATION DESCRIPTORS IN GENETICS AND GENOMICS RESEARCH

Researchers in human genetics and genomics have often struggled with a lack of clear, specific guidance concerning the use of population descriptors. The committee’s recommendations are intended to operationalize the

guiding principles with specific practices and procedures to aid researchers. These principles and recommendations offer a starting point for greater harmonization of the uses of population descriptors in genomics research worldwide, without, however, calling for a rigidly standardized approach. The committee does not recommend a standardized nomenclature or typology of population groups, either globally or in the United States alone. The principles and recommendations are intended to provide the basis for a shared approach to grappling with the myriad potential uses of population descriptors in human genomics research worldwide. The recommendations focus on areas that the committee identified as necessary for achieving change, including employing strategies to improve research study design, promoting transparency, tailoring the use of population descriptors for the purpose of a study, and ensuring that researchers have the support needed to implement the recommended best practices.

Requisites for Sustained Change

The committee identified three overarching approaches that are paramount to the long-term success of any effort to resolve the challenging problems surrounding the use of population descriptors in genetics and genomics research: avoiding typological thinking, including environmental factors in study design, and engaging communities. Although not new, these topics warrant increased attention because confronting these challenges could serve as the necessary foundation to catalyze progress.

Avoiding Typological Thinking

Erroneous categorical assumptions can be scientifically and ethically detrimental, particularly when applied to studies of human history, identity, variation, and traits or diseases. There is a pervasive misconception that humans can be grouped into discrete, innate biological categories. The committee cautions against the use of typological categories, such as the racial and ethnic categories established by the U.S. Office of Management and Budget in Statistical Directive 15, for most purposes in human genomics research. While the use of these categories may be required of researchers under certain circumstances (for example, in describing participants in studies receiving federal funding), the fundamentally sociopolitical origins of these categories make them a poor fit for capturing human biological diversity and as analytical tools in human genomics research. Furthermore, use of these categories reinforces misconceptions about differences caused by social inequities. Current practices in human genetics, including the use of descriptors such as continental ancestry, also reinforce these views.

Although these recommendations help lay the essential groundwork for changing the use of population descriptors, specific guidance for appropriate use is also needed and is provided through subsequent recommendations and best practices.

Including Environmental Factors in Study Design

Genetic effects cannot be adequately explained without nongenetic contexts. In the broadest sense, nongenetic or environment in gene–environment research refers to everything outside of DNA that influences a person’s traits. These factors include physical, chemical, and biological exposures; behavioral patterns, such as sexual practices or physical activity; and social context, such as neighborhoods and income, throughout the life span. Epidemiologic and genetics studies sometimes use race and/or ethnicity as a proxy for cultural beliefs or shared exposures without directly measuring them, even though descent-associated descriptors are not reliable proxies for most environmental factors. Whenever possible, researchers should use variables that more precisely capture the information that is needed to answer the question at hand. Moreover, researchers should not attribute unexplained variance to racial or ethnic differences.

___________________

⁴ Johann Friedrich Blumenbach (1752–1840) named Europeans Caucasian because he felt the most beautiful skull in his collection came from the Caucasus region and was thus a fitting symbol for a superior race (Marks, 1995; Painter, 2010).

⁵ The committee chose not to capitalize “black” and “white” throughout the report to recognize and emphasize that they do not signify biological or ethnic groups.

As it may not always be possible to directly measure all relevant environmental variables, the committee provides guidance for navigating the associated nuances in later best practices.

Engaging Communities

Communities vary in how individuals and groups self-identify and in their preferences for involvement in a research study. The ways that communities define themselves are dynamic and change over time. Evolving ethical guidelines and frameworks underscore the importance of engaging communities in the research process, in particular by demonstrating trustworthiness and cultivating trust. Effective community engagement improves communication, study coordination, and long-term collaborations between researchers and communities. Conversely, failing to engage and understand communities and relevant parties can undermine trust and trustworthiness of research, diminish public acceptance of the veracity of research results, and importantly, fail to deliver the research outcomes effectively to the communities whom researchers are trying to serve. Effectively engaging communities requires multidisciplinary approaches that draw on expertise in history, sociology, demography, anthropology, communication, and other areas. Research teams should include members with community engagement expertise to better understand how communities identify themselves and discuss the rationale for descriptors or group labels researchers decide to use. Research teams can develop ongoing partnerships with communities by drawing on emerging models and guidelines of community-engaged research.

Guidance for the Selection and Use of Population Descriptors in Genetics and Genomics Research

Because research conducted using genomic data is broad and varied, the committee concluded that there is no one-size-fits-all solution to the challenge of using population descriptors; rather, the appropriate popula-

tion descriptor depends on the scientific question that is being addressed. Consideration of the different purposes of genetics research gave rise to seven major types of genomics studies, covering both disease and non-disease traits,⁶ to serve as a basis for the development of recommended best practices:

1. Gene discovery for Mendelian traits: studies aimed at identifying the genetic basis (e.g., pathogenic variant) underlying Mendelian disorders or traits.
2. Prediction for Mendelian traits: approaches that rely on the presence of a specific genotype to predict risk for or incidence of a Mendelian disease or specific outcome.
3. Gene discovery for complex and polygenic traits: studies aiming to identify genetic variants associated with quantitative traits or complex disease risk, as done in genome-wide association studies (GWAS).
4. Prediction for complex and polygenic traits: studies that aim to make probabilistic predictions about individual disease risk or traits based on genomic data.
5. Elucidation of molecular, cellular, or physiological mechanisms: studies using related or unrelated participants or cell lines derived from their biological tissues to understand molecular, cellular, or physiological mechanisms.
6. Studies of health disparities with genomic data: elucidation of the role of genetic and environmental effects in how social disadvantage leads to health disparities.
7. Studies of human evolutionary history: inferences about human evolutionary history using samples of related or unrelated participants.

Responsive approaches are needed both to address the varied types of genomics studies and to accommodate community preferences and evolving conceptions of best practices for grouping individuals and naming those groups.

Transparency in methodology is a scientific norm and the bedrock of replicability, yet the challenge of transparency is one of communicating specifically how and why particular decisions were made—that is, stating the rationale behind the classification scheme and group labels applied when using population descriptors. The lack of both specific practices and transparent reporting can lead to confusion and a lack of comparability among data sets. This lack of transparent reporting could ultimately dimin-

___________________

⁶ Examples are provided in “Classification of Genomics Study Types” in Chapter 1.

ish trust in researchers by those participating in research. Researchers tend to rely on commonly used population descriptors without a clear justification for why they used them. When communicating their research methods, findings, and conclusions, researchers should be as transparent as possible about the specific procedures used to name groups within their data sets. To enhance transparency in reporting, the committee’s focus was the conceptual approaches and language that enable appropriate and accurate use of population descriptors in genetics and genomics research. The guidance that follows is intended to provide researchers with feasible best practices and rationales for decision making, in alignment with the guiding principles presented, and is an effort to support the goal of promoting trustworthy research.

To better equip researchers with the information to follow these recommendations, the committee developed best practices for different types of genomics studies as well as decision-making tools, including Table S-1 and a decision tree.⁷ The research context, including the study type and research questions, gives rise to specific best practices and helps the researcher determine which descriptors apply. Table S-1 suggests which population descriptors are most appropriate as analytical tools for each of the genomics study types outlined in this report. Note that each descriptor represents

___________________

⁷ A full discussion of best practices for each study type is presented in Chapter 5. The decision tree can be found in Appendix D.

a particular concept of difference across human populations. The tree and table are not intended to recommend or proscribe specific words. Instead, the recommendations in the decision tree and table focus on the conceptual building blocks that researchers should use in study design and data analysis. The objective is to encourage researchers using genomic data to consider, define, and delineate very carefully the concepts of human difference with which they are working.

A nuanced understanding of key terms and concepts is necessary to approach the table and best practices. In this context, population descriptors refer to conceptual classification schemes used to group people based on specific characteristics. The appropriate application of descent-associated population descriptors in particular study contexts is the primary focus of the recommendations to follow. Group labels are names given to groupings of individuals. There is the tacit assumption that despite the relative similarity of the individuals within a group the individuals may show variation in other dimensions, including their genetic background.

Many of the best practices recommended by the committee rely on distinctions between genetic ancestry, genetic ancestry group, and genetic similarity (Chapter 2). Ancestry is a concept encompassing a person’s origin or descent, lineage, “roots,” or heritage, including kinship. People have an intuitive understanding of ancestry from their family tree, consisting of their biological ancestors (e.g., parents, grandparents, and so forth). The genetic ancestry of a person refers to the paths through their family tree by which they have inherited DNA from specific ancestors. Genetic similarity between individuals is a quantitative measure of their genetic resemblance, reflective of the extent of their shared genetic ancestry. An analogy may be helpful for elucidating this distinction between a concept (genetic ancestry) and the measures or indicators of that concept (genetic similarity). The concept of wealth is generally understood, but the way it is measured can vary from the amount of money in a person’s bank account, to the car they drive, home they own, or sneakers they wear.

Genetic ancestry groups are discrete groups delimited based on one or more measures of genetic similarity. Once demarcated, these groups are typically given a label derived from nongenetic characteristics, including ethnicity, geography, or race. In many contexts, grouping individuals in a study based on genetic similarity alone, without additional labeling, may often be sufficient for the purposes of the study. When choosing to use genetic ancestry or similarity as a population descriptor, careful attention to the intended application is paramount (see best practices in Chapter 5).

As shown in Table S-1, best practices in the use of population descriptors vary by study type. Careful consideration should be given to whether descent-associated population descriptors are needed at all, beyond basic

descriptions of sampling strategy (e.g., sample collection site and study inclusion criteria). Once researchers identify the appropriate population descriptor for the context of their study, they should then apply group labels consistent with that concept to all study participants. More than one descriptor may be appropriate, and studies may benefit from using multiple descriptors. Finally, population descriptors are sometimes used as proxies for environmental effects (Chapter 2). It should be explicitly noted when population descriptors are used as proxies in this way and the rationale provided. The reader is advised to consult the text in Chapter 5 describing best practices in conjunction with viewing the table. In addition, Table S-1 provides only a broad overview and summary of the best practices; additional considerations for decision making are outlined in a decision tree (see Figure D-1 in Appendix D).

Implementation and Accountability

Despite many previous efforts to provide recommendations, guidelines, and strategies promoting culturally sensitive and valid use of population descriptors, there has been relatively little change in how any entities within the genetics research ecosystem use them. Many aspects of the current systems that fund, support, evaluate, and reward genomics research must change to better facilitate implementation of these recommendations. The genomics research ecosystem has many players, including funders of genetics and genomics research, professional societies, research journals, and research institutions, who all share responsibility for making these changes across an interdisciplinary research community (Chapter 6). Individual researchers bear this responsibility too.

___________________

⁸ Often called “human subjects” research training. See also https://www.hhs.gov/ohrp/education-and-outreach/human-research-protection-training/index.html

TABLE S-1 Recommended Approaches for the Use of Population Descriptors by Genomics Study Type

This table should be read and interpreted in conjunction with the report text. Consult the decision tree in Appendix D for more information and Chapter 5 text for best practices for each study type. See also the terminology box preceding the table and descriptions of each study type in Chapter 1 section “Classification of Genomics Study Types.” For any given study, the use of multiple descriptors may be preferable.
LEGEND
Preferred population descriptor(s)				Should not be used
In some cases; refer to Ch. 5 text and the decision tree in Appendix D				Descriptors could be used if appropriate proxies for environmental, not genetic, effects
GENOMICS STUDY TYPE	Race	Ethnicity/Indigeneity	Geography	Genetic Ancestry	Genetic Similarity	Notes
1: Gene Discovery - Mendelian Traits						Similarity suffices as a genetic measure; at fine-scale, other variables may be useful
2: Trait Prediction - Mendelian Traits						No population descriptors may be necessary for analysis
3: Gene Discovery - Complex Traits						Similarity suffices as a genetic measure
4: Trait Prediction - Complex Traits						Similarity suffices as a genetic measure
5: Cellular and Physiological Mechanisms						No population descriptors may be necessary for analysis
6: Health Disparities with Genomic Data						Not all health disparities studies rely on descent-associated population groupings, so none may be necessary for analysis
7: Human Evolutionary History						Reconstructing genetic ancestry may be of central interest

The ability of this report to effect durable change also depends on accountability, which could be enhanced by two mechanisms. First, since both research and social norms will continue to evolve in the future, population descriptors and their use must necessarily change as well. Thus, the committee concluded that it would be valuable to establish multidisciplinary advisory bodies to periodically evaluate current population descriptors and recommend changes based on trusted sociological and scientific data, current cultural norms, and ethical and empirical principles. Second, there is a need for groups with broader powers to monitor and facilitate the implementation of these recommendations.

It will take a concerted effort by all relevant parties, patience, and a good bit of time to reach a place where the proper use and reporting of population descriptors is routine and consistent. The recommendations in this report will need to be implemented broadly and consistently, by all the relevant parties, to generate lasting change.