DNA testing is a phrase that does not communicate very well because it means different things to different people. In dogs it can mean (1) determining the likelihood that specifically identified dogs are related, and perhaps how they are related (2) looking for the markers for a specific trait such as color or genetic condition such as PRA (3) attempting to discover what breeds are in a mixed breed dog. This last meaning has only come into being in the last couple years and is proving to be very popular, but is it accurate? That is the question asked here.
It appears that there is some way to go still on the level of accuracy for these tests. The sellers of these tests start with an advantage over their target client base. The target clients don't know the answer and so are in no position to refute the response provided. Some people, however, have submitted samples of dogs with known heritage and received results sufficient to question the accuracy of these tests.
One person produced a video of the process of submitting samples of her AKC registered American Staffordshire Terrier. The dog came back identified as Border Collie, Boston Terrier and Bulldog. The service used was BioPet Vet Labs DNA Breed Test. The site says "We believe that 92% of the mixed breed dog population in the USA is covered by our 58 validated breeds.".
Note that the American Staffordshire Terrier is not one of those 58 validated breeds. It is, perhaps, good news for those whose dogs are constantly faced with ill conceived breed specific legislation. However, while this test was inaccurate, failing to find the breeds typically covered in BSL, I think the point is well taken that it could just as easily be inaccurate in the other direction. Such a test might be used by authorities to "prove" that a particular dog is a mix of a restricted breed even when the dog's background contains none of that restricted breed. Genetic testing has already been used to successfully counter an official's opinion that a particular dog is of a mix subject to a breed ban. If these tests are proven to be unreliable then are we back to the often very inaccurate "expert" opinions of officials?
Another person reported the results of submitting samples for a Boston Terrier and a Keeshond. Canine Heritage identified the Boston Terrier as a mix of Cairn Terrier, Miniature Pinscher and Whippet. Canine Heritage states that they are able to successfully assign the correct breed over 99% of the time when testing purebred dogs on their breed list. Their breed list includes the Boston Terrier and so should have been able to identify it. DOG DNA identified the Keeshond as Golden Retriever, Akita, and Samoyed. They are, apparently, using the same test and breed list as Canine Heritage. (This was reported on a MySpace Group for Mix Breeds and Mutts "So I got my dna test kits this week. so excited").
The dna for one dog, Dakota, a dog of unknown heritage was submitted. The dog, a recessive yellow (ee) weighing 55 pounds, was identifed as 51% Boston terrier, 25% to 49% Basset hound, and less than 25% Bulldog. The dog owner recognized that this was extremely unlikely and contacted the lab. The lab admitted an error had been made but has not yet sent the replacement results. The dogs pictured in an article published in the Houghton Lake Resorter also seem unlikely to have the breed inherirtance reported. An article by Linda Dahlstrom on MSNBC similarly questioned the results ... or rather lack of results but concedes that the problem might have been the very limited number of breeds identifiable at that time (38). The comments following the article provide additional consideration. I'm somewhat puzzled by the number of reports that include field spaniel, usually in trace amounts. The breed is very rare.
Deteriming that the test results on an unknown mixed breed are blatently incorrect is challenging but can sometimes be done by someone sufficiently familiar with typical inheritance patterns. While there are plenty of happy customers and properly identified purebreds and mixed breeds the number of reports of unlikely mixes suggests there is some development yet to go before there is a reasonable level of reliability.
In March 2010 someone requested that I link to http://www.wisdompanel.com/faq.html The interesting thing is that it didn't in any way refute my discussion above. It includes plenty of lovely marketing speak but also some honesty about the real level of accuracy. The bottom line is that if you know the background of your dog they have ready explanations why they couldn't determine your dog's background accurately, and if you don't know your dog's background you won't be able to dispute the accuracy.
I've thought about doing the testing on my dogs, just for the heck of it. Then I think better of it. Nothing can be proved either way. Kind of expensive entertainment. When it comes right down to it you still have to deal with the dog as it is, rather than as it is supposed to be. That is true even with purebred dogs.Back to the top
Copyright © 2008,
Created:June 1, 2008
Updated June 29, 2010