“We seek out and cultivate identities to fill our need to belong, and it’s through that lens of identity that we see and understand the world,” said Jay Van Bavel, a psychology professor at New York University who researches how group identities, values and beliefs shape the mind and brain. “So when you get information that challenges your identity, many people tune it out, just like we do with headlines and news stories when they counter our politics and belief system.”
When white test takers see results that indicate they have African ancestry, some, especially young people, welcome their newfound multicultural heritage, even when the percentage is small, which raises an interesting question: How much ancestry is enough to give someone the authority to claim that identity? Research also shows that some whites whose reports indicate African lineage conclude that it’s irrelevant, and still others, no matter their race or ethnicity, disbelieve results they didn’t expect. For example, many blacks and whites whose families have long claimed that some of their forebears were Native American dismiss DNA reports that say otherwise. And Asians, like whites, often rebuff results that indicate that their heritage isn’t pure. Some people take that to extremes: White nationalists who use DNA tests to prove their racial purity adamantly reject any non-European results. A professor at the University of California at Los Angeles and another researcher studied comments on the online white-supremacist forum Stormfront. They found that some posters who had taken DNA tests and were upset with their results argued that they were “rigged” to “spread multiculturalism” or that the non-European findings were merely “noise DNA.” Many African-Americans, meanwhile, upon seeing how much of their lineage is European, are not necessarily surprised or doubtful about the results, but they feel gut-punched by the bald reminder that even their genes carry slavery’s legacy. Underlying all these reactions is the question of identity: What do these results mean about who I am? How do these results fit with the stories I’ve long clung to that connected my past, my present and my future?
Ever since Johnson received her disorienting DNA results, she wondered if her saliva sample might have been accidentally mislabeled or she had been sent someone else’s results. But it turns out that the company that analyzed her DNA focuses on forensic genetics and legal paternity tests, which evaluate only a few segments of DNA, not the hundreds of thousands used by most ancestry-testing companies. (Foeman used this company for a minority of her research.) So this summer, when The New York Times offered to buy Johnson ancestry tests from more mainstream companies, AncestryDNA and 23andMe, she eagerly agreed.
Their tests determine ethnicity by analyzing segments of customers’ DNA that give clues to their ancient geographic origins. Five hundred to 1,000 years ago, before large-scale transcontinental migration, people who lived in the same region had similar genetics. Scientists have been able to identify distinct patterns of genetic variation among people whose ancestors hail from the same lands, which is easiest to do with populations that were geographically isolated, like Finns and Filipinos, or were insular, like Ashkenazi Jews. Ideally, ancestry-testing companies would compare customers’ DNA to that of people from premigration days. But given that impossibility, the companies use an imperfect proxy: people alive today who have a deep family tree in a particular geographic area, and sometimes a paper trail to prove it. Those people’s DNA becomes the company’s reference data set for that geographic area. When a segment of your DNA closely matches the data for that location, the company assigns you that ancestry. The more segments on your genome that match that genetic pattern, the larger your estimated percentage will be for that ancestry.
The larger the reference data set for any particular corner of the world, the better the resolution will be: suggesting that your ancestors aren’t, say, just from Europe but from Northwestern Europe, or more specifically from Ireland and Scotland. Each testing company builds its own reference data set, drawn primarily from its own customers, and each company also creates its own algorithm for assigning heritage. In other words, customers’ results are based on inferences and are merely an estimate, often a very rough one — something many test takers don’t realize and testing companies play down.
Still, Johnson, now 65, hoped the new tests would conclude that her genes aligned with who she believed herself to be. In early August, with the kits in hand, she walked around her apartment, trying to work up enough saliva to fill the little collection tubes. Afterward, Johnson was both eager for quick results and hesitant about what they might say. “You know,” she said, “even if the results are the same as they were before, I am still a black woman.”
Weeks later, her AncestryDNA report was posted. It marked more than a third of her ancestry as “low confidence,” meaning it couldn’t establish its ethnicity because her DNA didn’t sufficiently match the company’s reference data sets. She was disappointed. It’s a common experience for customers with non-European ancestry, because Africa and Asia are underrepresented in many companies’ data sets, in part because most of their customers — the building blocks of their reference set — are of European descent. Many companies are trying to remedy that by seeking DNA from people in regions underrepresented in the data set.