DNA testing can bring families together, but gives mixed answers on ethnicity

Ethnicity estimates vary widely depending on which company is doing the testing

Michael Douglas

FINDING FAMILY DNA testing helped Michael Douglas find his biological family in southern Maryland and his Irish roots.

Matthew Rakola

Michael Douglas, a new resident of southern Maryland, credits genetic testing for helping him find his heritage — and a family he knew very little about.

Douglas, 43, is adopted. He knew his birth mother’s name and had seen a birth certificate stating his birth name: Thomas Michael McCarthy. Over the years, Douglas had tried off and on to find his birth family, mostly by looking for his mother’s name, Deborah Ann McCarthy, in phone books and calling the numbers. “I think I must have broken up a lot of marriages,” he laughs.

His search gained urgency in the last five years as he battled a life-threatening illness. “We planned my funeral three times,” he says. Douglas has a genetic disease called Ehlers-Danlos syndrome, caused by a variant in a gene that helps build the body’s connective tissue. His stretchy skin and hyperflexible joints are characteristic of the disease.

“As a kid, I was always dislocating something,” he says. His blood vessels don’t constrict properly to maintain his blood pressure, so Douglas sometimes faints when he stands up. For five years, he has had a constant migraine. Headaches are typical of about a third of people with Ehlers-Danlos. On top of that, he has B cell lymphoma. “I feel like I have the flu every day,” he says. It was time, he decided, to track down his birth family and learn more about his medical history.

In June 2017, Douglas flew to Ireland on what he calls his “death trip.” He wanted to see the land of his McCarthy ancestors. He chose Fethard, because the walled medieval town has a pub called McCarthy’s. (Douglas learned later that he and the pub owner are related.) His health improved during the visit, which he attributes to Ireland’s cool weather. When he returned to Phoenix, where he and his adopted family lived, he had new resolve to find his birth family.

“That’s it,” he decided. “I need my DNA run to find out who I am.” He sent his DNA to three testing companies: Family Tree DNA, AncestryDNA and MyHeritage. With his results plus sleuthing of genealogical records by some helpful strangers, Douglas found his biological family last November and dove headfirst into a new life.

In February, he moved from Phoenix to Maryland to help care for his biological mother as she recovers from a stroke. The new family dynamic hasn’t been easy, but Douglas has bonded with one of his two biological brothers. “And I have a relationship with my ancestors that I did not know before.” He is pleased to find that he resembles his great-grandfather Thomas Rodda, a bicycle maker. Douglas himself is a Star Wars costume maker.

FAMILY HISTORY With this picture, Douglas learned he resembles his great-grandfather Thomas Rodda (center, holding the bicycle frame).  Matthew Rakola

Adoptees like Douglas and birth parents looking for children they gave up often use commercial DNA tests in hopes of reconnecting, says Drew Smith, a genealogical librarian at the University of South Florida in Tampa. Many states make it difficult for adoptees to get birth certificates or other documents that could help them track down birth families. DNA tests are “an end run around the documentation problem,” Smith says.

But the pool of people looking for their genetic roots is much larger. AncestryDNA, the ancestry testing service with the biggest customer base, has persuaded about 10 million people to take its DNA test. 23andMe, Living DNA, Family Tree DNA, MyHeritage, National Geographic’s Geno 2.0 and others also offer customers a chance to use genetics to connect with living relatives and with families’ pasts. A few companies even give hints about ties that go back to Neandertals (SN: 11/11/17, p. 10). But such testing services may not be able to tell you as much about who you are and where your family came from as they claim.

False precision

I got my DNA tested for this multipart reporting project. My assignment was to investigate the science behind DNA testing (SN: 6/9/18, p. 20), but it was also a welcome excuse to learn more about my family’s history.

I already knew a lot about three branches of my family tree. Based on birth and death records, plus census and other documents, most of my family stems from England and Germany. But I dreamed of connecting to relatives on the Hungarian branch, which I knew less about. So I sent saliva or cheek swabs to a handful of testing companies.

My ethnicity estimates were all over the European map. Generally, estimates are most accurate on the broad continental scale. All of the companies agree that my heritage is overwhelmingly European. But that’s where the consensus ends. Even the companies that limit their estimates to broad swaths of the continent told different stories. National Geographic’s Geno 2.0 says that I am 45 percent Southwestern European. Veritas Genetics puts my Southwestern European heritage at only 4 percent and tells me I’m mostly (91.1 percent) north-central European.

The companies that try to dig down to the country level see their confidence in the results go down, but that doesn’t stop them from making very specific estimates. In most reports, the main results given are at the lower end of the confidence scale. 23andMe, for instance, says it has 50 percent statistical confidence in the ethnicity results.

Along with the wide variations between companies, the estimates often didn’t match what I know about my family tree. 23andMe says I’m 16.6 percent Scandinavian. When I sent raw data from 23andMe to MyHeritage to do its own analysis, that company reported no Scandinavian ancestry in my background; it said I’m 16.9 percent Italian. As far as I know, I have no ancestors from Italy or Scandinavia.

Only 23andMe called out my German heritage, though the company lumped it in with French for a total of 18.8 percent. Hungarian is not specifically identified in any company’s estimates. I can only guess that 23andMe’s 3.9 percent Eastern European and 0.3 percent Balkan findings cover that part of my ancestry. Both 23andMe and AncestryDNA say that I have Ashkenazi Jewish heritage. News to me.

Multiple companies agree that a sizable chunk of my heritage is from the British Isles. But even in that, estimates run from 23andMe’s 26.6 percent British and Irish, to Living DNA’s calculation that 60.3 percent of my DNA comes from Great Britain and Ireland, to MyHeritage’s even higher 78.7 percent.

When I shared these inconsistencies with Deborah Bolnick, an anthropological geneticist at the University of Texas at Austin, I could practically hear her shaking her head over the phone.

“They present these very specific, precise numbers down to the decimal point. But it’s a false precision,” Bolnick says. “The tests that are available may not be as nuanced, sensitive and fine scaled as they are presented.”

Checking references

Ethnicity estimates come from comparing patterns of genetic variants — often called single nucleotide polymorphisms, or SNPs — in your DNA with the SNP patterns of pools of people from particular geographic locations. As a way to confirm that a pool solidly represents a place, companies generally require that the people in these pools, known as the reference populations, have four grandparents who were also born in that location. Many of the companies draw reference population DNA samples from people in large public databases compiled by the 1000 Genomes Project, a catalog of genetic variation of thousands of people around the world, and from other studies. Some companies supplement their databases by testing more people in particular parts of the world. So the mixes in reference populations differ across companies.

Who the companies say you are depends in large part on those reference populations, Bolnick says. For instance, you may carry a pattern of SNPs found in people in both southern France and in Italy. If, by chance, the French people a company sampled had that SNP pattern but none of the Italians in the company’s database did, “they may infer that you have French ancestors and not Italian because of who they do and do not have in their database,” Bolnick explains.

Drilling down to tell customers which country or which part of a country their ancestors called home requires sampling many people in those countries, together with more sophisticated math to detect slight differences in the patterns. By looking at more than SNP patterns, Living DNA provides ethnicity estimates down to subregions of the United Kingdom and Ireland. The company analyzes how different stretches of DNA are connected to each other, says David Nicholson, the company’s cofounder and managing director.

It’s a bit like regional differences in the way people in southwest England assemble scones, cream and jam for cream teas. “In Devon you have a scone, cream and then you have jam,” Nicholson says. “In Cornwall you have a scone, jam, cream, so you have them in a different order. Most DNA tests just tell you that you have a scone, jam and cream so you’re from the U.K.” But because his company looks at the order of the DNA ingredients, Nicholson claims his results can tell customers what part of the British Isles was their ancestral home.

REFERENCE CHECK Testing companies estimate ethnicity by comparing customers’ DNA with the DNA of people in reference populations around the world. But companies have different reference populations and divide the world differently, as seen in this comparison of reference population maps from AncestryDNA, MyHeritage and 23andMe.  From left: AncestryDNA, MyHeritage, 23andMe
REFERENCE CHECK Testing companies estimate ethnicity by comparing customers’ DNA with the DNA of people in reference populations around the world. But companies have different reference populations and divide the world differently, as seen in this comparison of AncestryDNA’s and MyHeritage’s reference population maps.  From left: AncestryDNA, MyHeritage

Dividing lines

In reality, what the companies can say with certainty is that you share common DNA patterns with people living in those places today. But your ancestors may not always have lived where their descendants do now, Bolnick says. People move around, which muddies the waters.

For many Americans, some branches of their families may be recent immigrants, while other branches may have deep roots in American soil. Two branches of my family came to Massachusetts and Maryland from England in the 1600s. One branch moved from Germany to Nebraska in the late 1800s, and my Hungarian great-grandparents arrived in 1905.

Most Americans who get tested want to know about family from before the big move to the United States, says human geneticist Joe Pickrell, chief executive of DNA testing company Gencove. But the answer isn’t simple. DNA is a record of thousands of ancestors stretching back deep in time, each from a slightly different place. How companies sort out time and place may produce different ancestry estimates, Pickrell says.

Take a stretch of DNA containing a particular SNP pattern. “Today it may be found in you in the United States and in relatives in England and Germany, but it could be that 500 years ago your shared ancestor lived in Italy,” Bolnick explains. Going further back in time, that stretch of DNA may look like it came from Romania, Mongolia and Siberia. “As people move and the genes that they have move with them, it’s going to change what those geographic ancestries look like,” she says.

Given the timing of my family’s migrations, I would have expected a much bigger percentage of my ethnicity to come from the newer immigrants. I thought my British ancestry would have been diluted after hundreds of years in America, but I guess not.

Further complicating matters, most people think of their ancestry as coming from particular countries, but genetics cuts across and transcends national borders, Bolnick says. In reality, those categories are not genetic, they’re sociopolitical and historic.

Smith, in South Florida, agrees: “From a DNA perspective, it’s hard to tell a French person from a German person.”

Missing groups

And some groups, including aboriginal populations in Australia and big parts of Africa and Asia, are mostly absent from companies’ databases. The same goes for Native Americans, whose samples in public databases are small, and in some cases, were collected by questionable means, says Krystal Tsosie, a geneticist at Vanderbilt University in Nashville.

 Courtesy of Vanderbilt University Medical Center

She’s talking about “vampire projects,” in which geneticists swooped in to draw blood from native people, then disappeared. Some scientists have misused DNA samples taken from members of several indigenous nations, conducting studies the DNA donors didn’t consent to and doing studies that contradicted the groups’ cultural and religious beliefs.

In 2002, the Navajo (Diné) Nation — Tsosie’s tribe — declared a moratorium on genetic research. Recently, tribal members have discussed lifting the moratorium, but for now it remains in place, Tsosie says. “We’ve been, for so long, used as research subjects and not really equitable partners in research,” she says. “We’re still waiting for the conversation to change to allow us to have our interests protected.”

As a result of this mistrust of genetic research, there are not enough people from the 566 federally recognized tribes in the genetic databases to enable customers to learn about their tribal heritage from DNA tests. And even if a DNA test could establish that a person carries DNA inherited from a Native American ancestor, that doesn’t make that person a member of the tribe, Tsosie says. Tribal memberships are based on family and community ties, not DNA.

As a volunteer for the Native American Indian Association of Tennessee, Tsosie gets a lot of questions. People get Native American results and want to know if they can share in gaming profits. “It’s not enough to just call yourself a Native American,” she says. “I tell them, you have to go through the genealogy” and document your ancestry. “Typically, the response is, ‘Oh, that sounds like too much work.’ ”

That response baffles her. “If knowing this Native American past — this part of you — is so important, then undergoing the legwork and documentation should be important,” she says. Equally puzzling is why people base their identities on randomly inherited SNP patterns, she says. “Our character, who we are, who we come from is a complex story of a variety of nonbiological factors. To reduce that to a test kit is actually going to ignore the beauty and complexity that is us.”

WANNA SWAP? DNA recombination can be a confusing concept. We explain with Legos.

Making connections

Some ads for testing companies reinforce the link between DNA and identity. An AncestryDNA ad features Kyle Merker, a real person, who says that he grew up thinking he was of German descent. He even danced in German folk groups and wore lederhosen. Merker’s DNA suggests he’s not German at all, but predominantly Scottish and Irish. He’s swapped his lederhosen for a kilt.

The commercial makes it sound like Merker changed his entire culture because of a DNA test. Dig deeper, though, and you’ll find that he researched his family through newspaper articles and government records. These traditional genealogical resources really told Merker the story of his family, Smith says.

“DNA by itself is rarely of any value,” Smith says. “If you’re really interested in researching your family, there’s much more work to be done.” He likens it to ads from Home Depot or Lowe’s: “They make it look like, ‘Oh my gosh, redoing a room is easy.’ ”

Similarly, to really confirm heritage, people have to follow paper trails composed of birth and death certificates, military forms, immigration records, census rolls, church baptism and marriage records, and more. “DNA is just one more type of record,” Smith says. “You’ve got to pull it all together to build your case.”

Michael Douglas found his Irish roots, but it took more than DNA to untangle his heritage. Douglas learned from a McCarthy lineage group on Family Tree DNA that his Y chromosome suggests he’s a descendent of Donal Gott McCarthy, a 13th century Irish king. “Oh, my god, I’m royalty!” he says. The group helped him trace the McCarthy lineage from the 1200s to 1830s Cork County, Ireland.

FOLLOW THE TRAIL Through consumer DNA testing and other research, Michael Douglas found out that his Y chromosome connects him to medieval Irish royalty. Matthew Rakola

AncestryDNA’s and MyHeritage’s DNA and genealogical records allowed Douglas and four people he calls his “ancestry angels” to connect him with his biological family. The angels were four strangers who friended Douglas on Facebook and helped him with his family research, using genetic connections Douglas had rejected because they didn’t have the McCarthy last name. The helpers disappeared once he tracked down his mother.

Not all endings are happy. Smith has seen DNA testing split families. “You may discover things that are surprising or disturbing,” he says. You could find out that your father isn’t your father. Or matching to other relatives could uncover family secrets, such as an aunt who never told her family that she gave up a child for adoption or an uncle who knowingly or unknowingly fathered a child.

“It’s fun to learn more about our ancestors and what our ethnicity is,” Smith says. But, he warns, keep in mind that what you learn “may upend your personal life or the personal lives of members of your family.” Don’t do it if you’re not prepared for the repercussions.

Tina Hesman Saey is the senior staff writer and reports on molecular biology. She has a Ph.D. in molecular genetics from Washington University in St. Louis and a master’s degree in science journalism from Boston University.

More Stories from Science News on Genetics

From the Nature Index

Paid Content