SAN FRANCISCO – Ancestry.com recently began releasing new estimates for tests that promise to trace people's ethnic origins, resulting in a flurry of anxious social media posts when this dramatically changed their ancestral homelands.
Some users were peeved to suddenly hear they were more French than Italian, or German than French, raising questions about these popular kits – and about a science that's only recently been introduced to consumers.
The explanation behind the overhaul points to how rapidly these genealogical tests are expanding. They've proved remarkably popular with consumers, offering a chance to see a genetic breakdown of a person's ancestry for less than $100 in many cases.
And that means companies such as Ancestry.com and 23andMe have been able to increase the specificity of their results in recent years, allowing for predictions that wouldn’t have been possible when the technologies were first introduced.
Ancestry, for example, has 10 million people in its DNA database. Some of those give it permission to use their information to expand the collection of genomes a customer might be compared with. This allows much greater specificity. A person might be told they have ancestors not simply from Norway, but central Norway. Or not just Native American from Mexico, but from northwestern Michoacán.
“These very large customer databases, with millions of people, enable these companies to sometimes deliver more fine-scale insights into population history” than scientists can, said Alicia Martin, a geneticist at the Broad Institute and the Analytic & Translational Genetics Unit at Massachusetts General Hospital.
When Catherine Ball, Ancestry’s chief scientific officer, started with the company seven years ago, the deepest they could go was to the continent level.
“We were only able to break down your origins between Europe, Africa and Asia. Now we give you the specificity that you have not only had ancestors that lived in Ireland, but that they lived in Cork, Ireland,” she said.
Just a year ago, 23andMe identified 31 populations when it reported a person’s ethnicity. This year that number has expanded to 160. That’s possible because of advances in the technology and the broadening of the population samples used, the company said.
Even customers who did their tests before the databases were expanded have their analysis rerun and sent an update. That’s what happened in the Ancestry case.
How these kits work
Genealogical DNA testing companies look at hundreds of thousands of locations on a person’s genome and compare them with databases of known DNA samples, giving customers information about what population groups their ancestors might have come from.
They do that by taking a sample of their customer’s DNA from their saliva, which is mailed to the company. The companies then compare points on the customer’s DNA with patterns found on those same points in public and proprietary databases of human genomes.
The size of those databases is key to their specificity. The more people you have to compare with, the better match. That’s why the companies are all working to increase the number of people whose genomes they have to compare with. That’s especially an issue for populations outside of Europe because there is less data available about them.
“What’s powered (the increase in specificity) is collecting more of the diversity that exists across the planet. Early on we had a lot of data about Europe. As we’ve grown and expanded, we’ve increased into East Asia, and just a few weeks ago we released an update for African-Americans that was made possible through specific efforts we made to collect samples from people from Africa,” said Robin Smith, a senior ancestry manager at 23andMe.
The collection comes in part by asking customers where their grandparents were born and then asking people who have four grandparents from the same country whether they would allow their information to be part of a reference panel, he said.
Another way to increase genetic diversity in a company’s panel is doing community events with people whose ancestors aren’t well-represented.
The business model is very new. Family Tree DNA, the first company to offer genealogical testing to individuals, only launched in 2000. Today there are five main companies in the United States offering genealogical testing, including 23andMe, AncestryDNA, National Geographic, MyHeritage and Living DNA.
Along with their popularity has come controversy. Some scientists note that because none of them release their reference panel data, it’s impossible to evaluate them.
“For it to be scientifically valid, the methods and the information have to be available to research scholars so the methods can be tested to make sure they’re predictable and can be replicated,” said Sheldon Krimsky, a professor of public health and community medicine at Tufts University in Boston.
Privacy advocates worry that the companies can and do sell anonymized genetic information to drug companies. And scientists have more doubts about their ability to predict diseases, another offering by some of these firms.
“There are very few diseases where you can say for certain that if you have Gene X, you’ll get Disease Y. But the other side, which we all believe in, is the ancestry,” said Dr. Esteban Burchard, a professor of pharmaceutical science at the University of California, San Francisco and a noted researcher in the effects of genetic differences on drug effectiveness.
Reached in Kenya, where she is teaching statistical genetics and bioinformatics, Martin said for her, these tests are incredible resources of genetic data that are paving the way toward widespread public acceptance and regular use of genetic information for learning about population history and improving health outcomes.
“It's hard to think of any other single clinical test that has the potential to cut across so many biomedical domains, such as cardiovascular, psychiatric, and autoimmune disease risk,” she said via email.