When local authorities in Hildesheim, Germany, didn’t believe an asylum seeker who claimed to be under 18 years old — and thus eligible for privileged treatment — police turned to a blood-based age test sold by a California company.
The test, offered by Zymo Research in Irvine, uses chemical modifications to DNA that accrue over a lifetime, called an ‘epigenetic clock’, to determine a person’s age. Forensic scientists — aware of its potential benefits but also of its current lack of precision — sounded an alarm.
In a paper published online1 in May, a team led by forensic-medicine specialist Stefanie Ritz-Timme of the University of Dusseldorf in Germany said that these tests were not ready for use in sensitive forensic evaluations.
But now, in the charged political atmosphere that has accompanied the arrival of millions of refugees to Europe, forensic scientists across the continent are joining forces to improve epigenetic-clock-based tests — with a focus on whether they might be used to help determine the age of refugees whose claims to be under 18 are disputed. They hope that, with time, such tests could replace existing methods, which assess the maturity of bones or teeth to determine an individual’s age but are imprecise, and can be controversial.
“The race is now on to develop a more accurate clock that would be more predictive than the anatomical tests — and also more practical for use in forensic science,” says cell biologist Wolfgang Wagner at the University of Aachen, Germany. He has teamed up with other German forensic scientists to overcome these challenges, not only for the test’s use on refugees, but also for other forensic uses.
A call for sensitivity
The development of scientific methods that could feed into decisions about who is granted asylum and how refugees are treated are likely to elicit criticism, says Denise Syndercombe-Court, a forensic geneticist at King’s College London. She says that some scientists, herself included, are wary of these efforts.
But Niels Morling, a forensic geneticist at the University of Copenhagen who is running a national epigenetic-clock programme, defends the development of tests that are more accurate. Given that the law treats those under 18 very differently from adults, he says, then “you have a duty to make sure that it can be implemented fairly”.
Philosopher Thomas Pogge, who specializes in global justice at Yale University in New Haven, Connecticut, says that to keep rising anti-immigration sentiment in check, it is important for authorities to show that they can detect any refugees who pretend to be younger than they are.
Since 2014, Europe has received around 4 million refugees, many of whom arrived without identity documents. More than 1.5 million refugees have sought asylum in Germany, around one-third of whom are registered as minors. In Europe, minor status usually leads to better care and an increased chance of being granted asylum. Minors also have a higher chance than adults of gaining permission for family members to join them. And, if convicted of a crime, people under 18 usually receive shorter sentences — served in special prisons for juveniles.
Authorities say that, because of the benefits, some unaccompanied refugees claim to be younger than they are. But the anatomical tests that are currently used to assess age have an error range of up to 3–4 years, and rely on X-rays and magnetic resonance imaging. In countries such as Norway, refugees have sued the authorities for being forced to submit to these medical tests.
The publication of the first reasonably accurate epigenetic clock in 2013 presented a simpler way of verifying age, because it could be carried out using blood samples2. Developed by biostatistician Steve Horvath at the University of California, Los Angeles, this clock measures an epigenetic mark called methylation at 353 DNA sites across the genome.
Horvath’s paper was based on data from thousands of individuals. He analysed the genomes of different cell types — from blood to heart and brain cells — and identified the modifications that both correlated best with age and were least subject to environmental influences such as diet, which can alter epigenetic modifications. The clock’s accuracy varies between cell types, and diseases such as cancer can throw it off kilter.
But in blood, the median error was 2.7 years, meaning that it could predict the age of half the donors to within 32 months — a similar accuracy to that of the X-ray-based methods. And Horvath’s clock proved particularly accurate in children and adolescents — including those around the legally crucial age of 18.
Horvath has since refined the method. In July, he published a new clock that measured epigenetic marks at 391 DNA sites in different tissues3. It was particularly accurate in buccal cells scraped from the inside of the cheek, which are easier to collect than blood samples. Testing these buccal cells from 53 people aged between 3.5 and 18 years, he found a median error of just 1.03 years. However, there were many outliers — people whose age could not be accurately predicted — and the most extreme result was out by 5 years and 8 months. Horvarth anticipates that epigenetic-clocks, once refined, will help refugees by corroborating their age claims. “At the same time, the tests may help identify individuals who break the law,” he adds.
In the Hildesheim case, Zymo, which bought an exclusive licence to Horvath’s test in 2016, compared the refugee’s sample with those of five others who had similar ethnic backgrounds and whose ages were known. Keith Booher, project manager for Zymo’s epigenetic services, told Nature that the test determined the most likely age of the person to be between 26 and 29 years old. Hildesheim authorities have declined to comment on the case.
Zymo charges around US$300 for a blood sample, and each test requires multiple samples. Forensic scientists around Europe are now working to make an epigenetic clock that would be more accurate and less expensive than what is available – and is also applicable to the easily-collected buccal cells. That means working out the minimum number of DNA sites needed to ensure the greatest accuracy, says forensic geneticist Peter Schneider of the University of Cologne in Germany. He predicts that just 10 or 20 markers will suffice.
Studies are also under way to gauge how the diverse ethnic backgrounds of Europe’s refugees might influence the epigenetic clock. Schneider’s team conducted a preliminary study4 into the relationship, which suggests the effect could prove to be limited. The researchers found that the epigenetic modifications for which the clock tested were broadly similar across populations hailing from Central Europe, the Middle East and West Africa.
Another challenge of using the epigenetic clock is dealing with the statistical outliers. A recent study5 of more than 1,000 British 17-year-olds born in the 1990s found that the clock measured a pleasing median age of 17.2 years — yet it judged some to be under 4 years old and others over 30.
Such outliers may in some cases be explained by undiagnosed illnesses that distort the clock, or by technical errors. Wagner thinks that some of these people, for whom the method simply won’t work, could be identified using the right combinations of epigenetic markers, allowing scientists to discount the test’s results. And with better mathematical models for setting the epigenetic clock, along with improved sequencing technologies for identifying epigenetic marks, it might be possible to get the test’s margin of error well down, he says.
But even if the accuracy of the epigenetic clock cannot be improved to the point where it can replace anatomical tests, it would still be attractive, says Morling. “Every time you add a new method, you improve overall precision of age estimation.”
Nature 561, 15 (2018)