Disclaimer: I was one of the authors on a 2003 study reporting a link between ACTN3 and athletic performance, but I have no financial interest in ACTN3 gene testing. The opinions expressed in this post are purely my own.
An article in the NY Times yesterday describes the launch of the grandiosely named Athletic Talent Laboratory Analysis System (ATLAS). The ATLAS test looks at a common genetic variation within the ACTN3 gene, which has been associated in numerous studies with elite athlete status and with variation in muscle strength and sprint ability in the general population. The company claims that this variation “may determine the type of athlete you were born to be”.
In the NY Times article ATLAS explicitly describes its target audience – the parents of young children:
Atlas executives acknowledge that their test has limitations but say that it could provide guidelines for placing youngsters in sports. The company is focused on testing children from infancy to about 8 years old because physical tests to gauge future sports performance at that age are, at best, unreliable.
Setting aside the unsettling ethical issues associated with recreational genetic testing of children, how useful will the test be to parents looking to find out whether or not their kids will be future track superstars? Here are the facts:
ACTN3 is indeed associated with athletic performance
The variant tested by ATLAS (and by a multitude of other companies – see below) disrupts the function of the ACTN3 gene, which normally produces a protein that regulates the function of fast muscle fibres. Fast muscle fibres are the cells that underlie rapid, forceful muscle contraction in activities like sprinting and weight-lifting. My previous lab in Sydney, Australia and several other groups have shown that individuals with two disrupted copies of ACTN3 are rare among elite sprint/power athletes, suggesting that the loss of ACTN3 is deleterious for muscle power production. There is also weaker evidence suggesting that the loss of ACTN3 actually increases endurance performance, although this is much less clear than the association with sprint performance.
Several studies have also found that ACTN3 influences muscle function in the general population, with the ~18% of the population with two disrupted copies of the gene displaying slightly lower muscle strength and poorer sprint performance than individuals with at least one functioning copy of the gene.
ACTN3 is just one of many factors influencing athletic performance
At the highest levels of performance ACTN3 genotype certainly make a big difference: among Olympic-level sprinters the frequency of individuals carrying two disrupted ACTN3 copies is vanishingly low (less than 3%, compared to ~18% in the general population). However, this large effect is due to the exceptionally strong selection that occurs during the slow climb to the Olympic level. The vast majority of athletes who start that climb will never make it to the top; those who do will be the tiny minority who have nearly everything in their favour, including the right genes.
So super-elite athletes need to have the right ACTN3 combination, but they also have to have a whole host of other factors working in their favour – this one gene is just a minor ingredient in a large and complex recipe. In fact, most studies performed so far suggest that ACTN3 explains just 2-3% of the variation in muscle function in the general population. The rest of the variation is determined by a wide range of genetic and environmental factors, most of which (particularly the genetic factors) are very poorly understood.
Parents considering using the ATLAS test on their children need to be very clear about this: the test provides information about only one small component of a much bigger picture. It’s probably not a great idea to base any important life decisions on such limited information.
ACTN3 doesn’t tell you whether or not your child will be a super-athlete
This is obvious when you consider the population frequencies of the three different combinations that the ACTN3 gene comes in: around 30% have two copies of the “sprint” version, about 18% have two copies of the “endurance” version, and just over half the population have one copy of each. Professional athletes, on the other hand, constitute a fraction of 1% of the population. Obviously there is no ACTN3 combination that is unique to super-athletes; instead, whatever combination your child has, he or she will share that with a large chunk of the population, the vast majority of whom will never go on to be international-level athletes.
In fact, as I mentioned above (and as illustrated in this graphic from the NY Times article drawing on data from our 2003 article), the effect of the ACTN3 variant differs markedly between different sports: the effect on sprint performance appears to be balanced somewhat by opposing effects on endurance performance, and there are many sports (e.g. talent-focused sports like shooting and sailing) where it’s unlikely that ACTN3 genotype has any effect at all. That means that when you look at elite athletes as a whole the frequency of the ACTN3 variant is essentially identical to what you see in the general population – the differing effects of the gene in different sports basically cancel each other out.
So while ACTN3 (in combination with other tests of ability and guidance from experienced coaches) might have some limited value in helping talented youngsters decide which sport they might be best suited to, this is not a test that can tell parents whether or not their kid will be able to become an elite athlete in general. No matter what your ACTN3 genotype is there is a wide range of sports that you could theoretically excel in. Whether you do actually excel in these sports will depend on a wide range of factors (particularly psychological factors) that have nothing to do with ACTN3.
There’s currently no evidence that ACTN3 offers improved predictive power over physical talent ID tests
The fact that ACTN3 is associated with muscle strength and sprint performance doesn’t mean that it necessarily provides any useful predictive power. The reason for this is simple: if ACTN3 affects athletic performance purely by effects on muscle strength, then simply testing muscle strength may capture all of the information conveyed by testing ACTN3 – as well as providing information about all of the other genetic and environmental factors that influence this trait. In other words, the ACTN3 test may well be completely redundant in the face of the existing panels of physical tests used by talent selectors to determine the athletic potential of young children.
This is obviously why ATLAS suggests that the test will be most useful for kids between infancy and eight years of age, in whom physical performance tests are hard to perform and give unreliable results. However, it’s hard to see what advantage could possibly be provided by the tiny predictive edge of ACTN3 testing at such a young age. In the NY Times article the ATLAS president argues that, “If ACTN3 suggests a child may be a great athlete [...] parents should take a step back and nurture that potential Olympian or N.F.L. star with careful nutrition, coaching and planning” – but of course bearing in mind the small effects of ACTN3 and the fact that it has different effects on performance in different sports, if you’re interested in an athletic future for your child you should nurture their potential regardless of their result on this genetic test.
To demonstrate that ACTN3 testing actually boosts predictive power beyond that currently provided by physical performance tests will take large prospective studies, and those studies simply haven’t been done yet. Until such studies have been published, parents (and sporting bodies) should treat the results of any genetic test for athletic performance with extreme caution.
ACTN3 testing isn’t new, or exclusive to ATLAS
This is briefly mentioned in the NY Times article, but deserves repeating: this test isn’t the invention of ATLAS. Since soon after we published the association between this gene and athletic performance in 2003, an ACTN3 test has been offered by an Australian company, which subsequently licensed the test to other companies overseas (including a Japanese company, and now ATLAS). The variant is also included in the massive panels of genetic markers tested by personal genomics companies 23andMe and deCODEme.
Kevin Fischer has already noted that from a pure cost-benefit point of view the ATLAS test doesn’t compete with the offerings of personal genomics companies. ATLAS will charge you $150 for testing ACTN3; for just $250 more, you get genetic information pertaining to more than 90 different conditions and traits from 23andMe. Neither test is likely to change your life (the predictive power of most current genetic tests using common markers is extremely low), but if you’re interested enough in recreational genetics to fork out for an ACTN3 test you might as well spend a little extra to get information on a bunch of other traits at the same time.