While I was poking around for stories about ACTN3, I came upon this Slate piece from the usually reliable Will Saletan about the ACTN3 test and race.
The story has a generally accurate discussion about ACTN3 allele frequencies and race, and I agree with most of what he has to say. But Saletan makes a factually incorrect statement about the relative value of race and genetics in predicting athletic performance:
Race is a less, not more, reliable gauge of physical characteristics than genes are. In fact, that’s one of the chief consolations of nontherapeutic genetic testing: No matter how inaccurate genes are as a predictor of this or that ability, they’re more accurate than predictions based on race. And the sooner we get past judging by race, the better.
Really? Saletan needs to be careful here. Let’s take the cast of ACTN3. Research shows that elite power athletes very rarely have the homozygous null X-X genotype at ACTN3, which is found in approximately 20% of the global population (it varies geographically, but this is not relevant to my point). Now consider (as Saletan does) the Olympic sprint finals, which consisted entirely of men of West African descent. And let’s assume (although they haven’t been tested) that none of these men had the X-X genotype. If you were trying to predict the probability that an individual was going to be in the Olympic sprint finals in 2012, what would you rather know – their genotype or their race? Genotype will narrow it down to around 4.5 billion people. I don’t know how many people there are of West African descent there are on the planet, but it’s probably a few hundred million. Now, of course, both of these methods will do an awful job of picking Olympic athletes out of the general population, but you’ll do 10x better making the prediction based on race alone than you would on ACTN3 genotype alone.
Now Saletan may be right (he very likely is) that there are other undiscovered genetic variants that collectively explain variation in sprintng ability. And once we know all of these factors we’ll probably be able to do a better job of identifying great sprinters than we would by just picking random West Africans. But as things stand race is a better predicter, Saletan is simply wrong when he says, “No matter how inaccurate genes are as a predictor of this or that ability, they’re more accurate than predictions based on race.”
I'm a biologist at UC Berkeley and an Investigator of the Howard Hughes Medical Institute. I work primarily on flies, and my research encompases evolution, development, genetics, genomics, chemical ecology and behavior. I am a strong proponent of open science, and a co-founder of the 
2 Comments
I disagree.
Didn’t the hapmap tell us that it’s very easy to predict race based
on genotypes? yes, you need more than one, but a collection of
(i don’t know how many) SNPs to say that a person is of west african descent,
but these SNPs, together with the ACTN3 variant, already give you a better predictive power than race alone
I am curious to know what happens to the arguement when you have an admixture population. When talking about race, is it the broadly categorized races of Caucasians, Africans, Asians and others or do you go to the contributions of different races and their admixtures? If so, isnt the genotype (or a pool of genotypes) better predictor than just the vaguely defined term “race”?