DNA and Breed ID

Can a dog be assigned to a particular breed by looking at its DNA?  Can DNA be used to reconstruct the evolutionary history of domestic dog breeds? Can DNA settle issues of pure breeding? Until now, the answer was no.

That’s because previous analyses had to rely on mitochondrial DNA, a slowly evolving type of DNA that, unlike nuclear DNA, is found outside the cell nucleus and is inherited only from mother to offspring. Mitochondrial DNA is good for looking at evolution over long time periods, but it evolves too slowly to look at evolution over the relatively short time that most breeds have been in existence.

Instead, some preliminary research projects had shown that microsatellite markers ― segments of nuclear DNA that vary greatly among different dogs but tend to be the same in closely related individuals ― could be used to identify a few breeds and make inferences about their relationships.  Extensive use of microsatellite markers had to wait until enough markers were mapped. Progress in mapping the canine genome has enabled the first wide-scale comparison using microsatellite markers. 

Researchers at Fred Hutchinson Cancer Institute in Seattle, WA, compared 96 DNA microsatellite sequences from 4 to 5 unrelated dogs of 85 different breeds. They found that dogs from the same breed had microsatellite “signatures” that were so distinct that 99% of the dogs could be assigned to their correct breeds based on their DNA. Only four breeds ― Chihuahua, Presa Canario, German Shorthaired Pointer, and Australian Shepherd ― failed to form distinct breed DNA groups.

Several breeds, while retaining breed specific microsatellite signatures, had very similar signatures to other breeds. In most cases this came as no surprise, since the breeds were ones already assumed to be closely related.  Close pairs included Alaskan Malamute and Siberian Husky; Collie and Shetland Sheepdog; Greyhound and Whippet; Bernese Mountain Dog and Greater Swiss Mountain Dog; Mastiff and Bullmastiff; and West Highland White Terrier and Cairn Terrier; and so close as to be indistinguishable, Belgian Tervuren and Belgian Sheepdog. 

The degree of similarity can be used a gauge of degree of relatedness, not only among breeds but in relation to wolves. Eight wolves, one each from China, Oman, Iran, Sweden, Italy, Mexico, Canada, and the United States, were used for comparison. Using these wolves as the root of a phylogenetic tree, a cluster of breeds consisting of four smaller branches were clearly separate from the remaining breeds. These branches and their breeds were 1): the Shar-Pei, Shiba Inu, Akita, and Chow Chow; 2) the Basenji; 3) the Alaskan Malamute and Siberian Husky; and 4) the Afghan Hound and Saluki. By changing the statistical criteria and using these breeds as the root of a phylogentic tree, the Tibetan Terrier, Lhasa Apso, Pekingese, Shih Tzu, and Samoyed were also grouped in this cluster. The fact that this progenitor cluster springs from a wide geographic expanse, suggests that the first dogs originated in Asia and migrated with humans throughout Asia and as far north as the Arctic and south as Africa.

The remaining breeds (those not in this progenitor cluster) probably originated more recently from European founders. Applying more sensitive analyses allows these breeds to be subdivided into three groups. The first is made up of Mastiff-like breeds and includes the Boxer, Bulldog, Newfoundland, and other guarding breeds. The German Shepherd also falls into this group, perhaps reflecting an influx of genes from military dogs with Mastiff forebears. A second subdivision is made up of herding type breeds such as the Collie, Belgian Sheepdog, and Australian Shepherd.  This group also surprisingly contained the Saint Bernard, Greyhound, Borzoi, and Irish Wolfhound. These nonherding breeds may have been either herding breed progenitors or descendents.  The remaining subdivision represents a wide variety of dogs of mostly recent European descent and includes terriers, scent hounds, spaniels, pointers, and retrievers.

These findings are important because they will allow researchers to pool samples from closely related breeds in order to better map disease and complex trait genes in dogs and possibly, humans. The research also has significance for understanding the evolutionary history of dogs, for forensics, and for breed and mixed breed identification.

Being able to assign a dog to a particular breed has many implications. At present, AKC issues ILP numbers based upon a dog’s physical similarity to a particular breed. Many such dogs are fortuitous imposters that are in reality mixed breeds. DNA identification may eventually prove to be more reliable than the current method.

Breeders who produce unexpected phenotypes in their supposedly purebred litters can at present turn to DNA parentage testing, but this only proves the sire and dam are who they claim to be. It’s possible that unusual traits could instead result from the influx of genes from a dog of another breed several generations ago, and that through inbreeding these recessives have been expressed. DNA microsatellite comparisons may or may not differentiate between mutation and cross breeding.

An area in which the implications may be mixed is that of breed identification for dogs in which breed specific legislation is involved. Owners could use DNA to prove their dog is not a member of a target breed. However, outside entities might also use DNA to prove a dog was in fact a member of a target breed.

Evolutionary findings may prove troublesome to breeds targeted for breed specific legislation, as it could conceivably be argued that progenitor breeds are “wolf-like.”  However, that fact that breeds such as the Pekingese are in this progenitor group, and that they are not exactly considered a menace to society, should help in any debate based on DNA sharing with wolves.

The complete report, “Genetic Structure of the Purebred Domestic Dog” originally appeared in the journal Science.