NIH Investigators Discuss Ethics of Ancestry Testing
Photo: Marleen Van Den Neste
Perhaps an old family photo first sparked an interest, or a story passed down through generations. Some wonder about the stories they never heard. Many begin the inquiry for health reasons, searching for clues to an illness that seems to run in the family.
Whatever the impetus, millions of people are inquiring about their roots, trying to get a glimpse into their lineages from at-home ancestry tests. These DNA kits can provide useful information but also have drawbacks. One looming question is their level of accuracy, a concern that’s also prevalent when considering ancestry in clinical research.
At the outset, it’s important to distinguish among ancestry, race and ethnicity, which should not be used interchangeably, advised Dr. Adebowale Adeyemo, deputy director, Center for Research on Genomics and Global Health, NHGRI.
When including ancestry data in research, proper use of these terms is essential to the study’s validity, agreed Dr. Sara Chandros Hull, director of NHGRI’s bioethics core and faculty member in the Clinical Center’s department of bioethics who, along with Adeyemo, spoke at an Oct. 1 NCI ENRICH Forum.
Hull emphasized the need to carefully consider how ancestry is used in both clinical study design and when disclosing and disseminating results.
Were study populations recruited fairly? How should investigators manage expectations in the consent process? Should they return results selectively or at all? Did they explain the value of partnering with the diverse study communities, who are part of, and supposed to benefit from, the research?
Photo: Marleen Van Den Neste
“It’s more than just saying good ethics precedes good science,” Hull said. “It’s really thinking about choosing questions that are important to address that use precise scientific tools correctly before we dive into a lot of the challenging ethical questions that tend to come up around genetics research.”
The terms race, ethnicity and ancestry each capture different but overlapping aspects of identity, said Adeyemo. Race assumes people are homogeneous, but they’re not.
“There are no discrete biological groups or genetic groups of people different from everyone else,” he said. “Human variation exists along a continuum.”
Ethnicity, however, is not a geographic label; it comprises cultural factors such as customs and language. To log race and ethnicity in research studies, clinical staff generally rely on self-reporting. Sometimes they assign a label based on physical appearance, which often correlates poorly with underlying ancestry, Adeyemo said.
Ancestry refers to broad geographic regions or local populations from whence your ancestors came. Someone may be 70 percent European, 20 percent African and 10 percent split among other groups. But people tend to be forced into a single category, which can skew results.
“If you just have people check boxes [for race and ethnicity], then you force them to choose something,” Adeyemo said, “and what they’ve identified [may not be] representative. But if you allow them to choose 3 or 4 [options], you’ll find you have dozens of terms that correlate to any single ancestry.”
Ancestry testing uses genetic markers to infer or estimate lineage. An ancestry panel has limits, said Adeyemo. It often partitions data between races and subgroups under study. It also often targets specific ancestries, e.g., African American vs. European American, or American vs. Native American, he said.
“Sometimes [race and ethnicity] can be inappropriately used as a surrogate for genetic ancestry or as proxies for socio-economic factors,” said Adeyemo. “There are stereotypes based on observations that something is more common in a group. However, this could be due to genetic factors, environmental factors, lifestyle, culture or a combination of these factors. So we get these reductionist explanations that pertain to a racial category.”
An ancestry panel also raises the risk of misclassification.
“If someone is studying a minority population and wants a comparable cell line but is getting something that’s 99 percent something else, then it’s wrong because that was not the intention of the researchers,” Adeyemo said. Such a study may incorrectly fail to confirm findings from other lines of research in that ancestral population.
Photo: Marleen Van Den Neste
If an investigator sets out to enroll 100 African Americans and 100 Native Americans in a study, for example, and some of each group are misclassified as European Americans, this omission can have scientific consequences. “The misclassification would basically stymie our efforts to try to reflect the diversity that we so desperately need,” said Adeyemo.
The most useful testing is a genome-wide SNP array, he added, which isn’t limited to pre-selected ancestries and therefore allows for inferring races and ethnicities across the globe. “It also has enough markers that we can come back later to reanalyze the data in light of new evidence or better techniques.”
In many cases, clinicians are overlooking a proportion of the world’s genetic variation, which can exacerbate existing health disparities.
“There’s a persistent lack of representation across non-European populations,” said Hull, which “fails to ensure that our investment in research will be as relevant as it should be to populations that also experience a disproportionate burden of disease.”
Yet some populations remain reluctant to participate in genomic research due to past stigma, discrimination or lack of informed consent, issues that keep resurfacing in Hull’s work with tribal communities.
“Any time we start talking with tribal communities about genetics research, the question of commercial ancestry testing comes up, how those companies are commodifying and selling genetic ancestry, indigenous ancestry, to the general public,” Hull said. These communities have asked, “Is the science legitimate? Is the science at NIH going to be used in these ways?
“It’s part of the context that research participants are bringing to the table when they sign up for our studies,” she said. “It affects our ability to partner meaningfully with tribal communities because it’s a stated concern.”
Hull urges clinical researchers to be careful not to conflate issues of race, ancestry and biology. Too often she sees definitions of race that are incomplete and fail to account for important political and social determinants.
For example, she said, “naming disease types based on an entire continent on which they were discovered [such as African Kaposi sarcoma] uses ancestry as a disease label that has the potential to be stigmatizing.”
Adeyemo recommended that studies try to capture as many aspects of identity—self-identified race, ethnicity, ancestry and socio-economic variables—as possible. Hull concurred: Aim for precision, good communication and transparency.