DNA ancestry tests compare your DNA to reference populations(not exact ancestry) so results vary by company and over time. Despite this, they feel compelling. Experts like Kostas Kampourakis and Adam Rutherford stress their limits but highlight value in finding relatives.
By Sofia Woodfine
For £100 and a tube of saliva, a DNA test promises something more seductive than information: a clearer story about who you are. Millions of people have mailed off their spit in search of an answer to a deceptively simple question, where am I really from? What comes back is a neat breakdown of percentages, the kind of result that looks more precise than a family story and more personal than a map (Royal et al., 2010).
But that certainty is part of the appeal, not the reality. Geneticists say these tests do not uncover who we are so much as generate statistical estimates, results that can shift over time, differ between companies, and depend on databases few users ever see or question (Ray, 2014). There is no single answer waiting in your saliva: each company is working from its own data, categories and assumptions.
Part of the appeal of DNA testing is that the results look exact. A screen filled with percentages, 20% this, 15% that, gives the impression that ancestry can be measured cleanly, almost like a blood test. But that is not quite what is happening. “What the tests really measure is not ancestry,” says Kostas Kampourakis. “Its similarity with people in the database.”
These companies are comparing snippets of DNA to samples from living people in their databases, grouped into so-called reference populations. The result is not a definitive answer to where someone is “really from”, but an estimate of which groups in that company’s system their DNA most closely resembles.
The percentages depend heavily on how the company has built the test: who is included in its database, how large it is, and how it decides where one population ends and another begins. At broad scales, the tests can be reasonably reliable. At finer scales, they become much shakier. “They are highly inaccurate when they go below the continental level,” Kampourakis says.
Even that result is not fixed. “The results change with time,” he says. “The DNA hasn’t changed, but it is the way the sequence is analysed, and the inferences made, that change.” For some users, that uncertainty becomes obvious quickly, with different companies producing conflicting results (Erlich et al., 2018).
Part of the power of DNA testing lies not just in the science, but in what people want from it. “People have always turned to the authority of science for definitive answers to nebulous questions,” says Adam Rutherford. The idea that a company can reveal something fundamental about you is hard to resist.
“Essentialism is intuitive,” says Kampourakis. People easily think that there is something deep inside that is important. Even when the science is probabilistic, the presentation, percentages, maps, and colour-coded regions, give an impression of clarity. As Rutherford puts it, these tests are “selling cute answers to complex problems.”
Despite their limitations, these tests are not useless. “It’s mostly about finding relatives,” says Kampourakis. When two people share long stretches of DNA, it is strong evidence of a recent common ancestor. “The genetics is a piece of evidence,” Rutherford says, best used as a starting point rather than a conclusion.
DNA tests are not purely personal. When someone sends off a sample, they are also revealing information about their relatives. Companies may store and use this data in ways users do not fully understand (Phillips, 2016).
DNA tests can offer clues. But the desire for a single, definitive account of identity is harder to satisfy. “Everyone wants to know where they are really from,” Rutherford says. “But what does that question even mean?”
References
Erlich, Y., Shor, T., Pe’er, I., Carmi, S., 2018. Identity inference of genomic data using long-range familial searches. Science 362, 690–694. https://doi.org/10.1126/science.aau4832
Phillips, A.M., 2016. “Only a click away - DTC genetics for ancestry, health, love…and more: A view of the business and regulatory landscape.” Appl Transl Genom 8, 16–22. https://doi.org/10.1016/j.atg.2016.01.001
Ray, S., 2014. Human Evolutionary Genetics. Yale J Biol Med 87, 603.
Royal, C.D., Novembre, J., Fullerton, S.M., Goldstein, D.B., Long, J.C., Bamshad, M.J., Clark, A.G., 2010. Inferring genetic ancestry: opportunities, challenges, and implications. Am J Hum Genet 86, 661–673. https://doi.org/10.1016/j.ajhg.2010.03.011
