In doing so, it helped show why dogs might be a better target for direct-to-consumer genetic testing than humans.
The research is the first of its kind to be conducted in nonhumans. It drew on data from more than 6,000 customer dogs in an effort to identify the genetic mutation responsible for blue eyes, a striking trait that’s relatively common in Siberian huskies but rare among other breeds in which it sometimes appears, like border collies and corgis. This meant sifting through the genomes of the dogs who did have blue eyes versus those who didn’t to see if any mutations were common to the former group and (mostly) absent from the latter. Owners conducted DNA tests from Embark and completed online surveys detailing their dogs’ breed and appearance, which included uploading “profile photos” for their pups; the scientists, from Embark and Cornell University, took care of the rest.
The data allowed them to identify a novel association: An allele on chromosome 18, carried by just 10 percent of dogs in the data set overall, was present in 100 percent of blue-eyed Siberian huskies and may be responsible for blue eyes in the breed. It seems likely that a duplication upstream of the gene ALX4, involved in mammalian eye development, is responsible—if so, breeders who can check for the variant in their dogs’ DNA will be better able to select for the trait.
Prospective blue-eyed puppies aside, the success of this first study, now in preprint, speaks to the approach’s potential: Being able to crowdsource genotypic and phenotypic information can lead to key discoveries regarding not just eye color but also more complex traits, behaviors, and overall health. Artificial selection in the form of careful breeding has also left dogs particularly well suited to this kind of analysis—from German shepherds to Chihuahuas, there’s a huge diversity of phenotypes on display, but genomewide divergence is pretty moderate except at the alleles underlying those differences, which is what this kind of testing can help identify.
There’s another reason this small study is exciting. Giving up our own data for such studies is something people are understandably wary of, but the stakes are considerably lower for our pets. And direct-to-consumer DNA studies in other animals could yield worthwhile results—for both them and us. A 2005 paper published in Briefings in Functional Genomics described dogs as “an unrivalled model for the study of human disease,” and regions of the canine genome have already been causally linked to more than 70 Mendelian diseases—heritable disorders caused by a single mutation as opposed to a more complex combination of genes—many of which have human analogues. And since there’s less sequence divergence between humans and dogs than humans and mice (and they’ve cohabitated with us since the hunter-gatherer days), their genomes may be able to reveal things that the murine model can’t. A recent study found that mice first colonized human settlements about 15,000 years ago, but our relationship with dogs may go back more than twice as far. Having shared our environment for so long—and seen us through some key transitions in the process—might put dogs in a unique position to tell us about ourselves.
That’s not to say you should take a genetic testing company’s claims about what your dog’s genes mean for his health without a grain (or more than a grain) of salt. Even if a region of the genome is associated with heritable diseases, that doesn’t necessarily mean a dog or its offspring are guaranteed to have it—we just don’t know enough yet to justify the sweeping, context-less claims many kits are notorious for making. If anything, the benefit of individual testing is that it provides the data for the larger-scale studies that can begin to lay the groundwork for meaningful genome analysis. If you want to sequence your dog’s genome, just remember the result will probably be more meaningful to science than to your ability to care for your pet.
