Scientists used genetic testing in over 900 shelter dogs to identify breed heritage in the largest study of its kind. The researchers found widespread genetic diversity: 125 breeds in the sample and an average of three breed matches per dog. The accuracy of shelter staff in identifying more than one breed in the dog's heritage based just on physical appearance was only 10 percent. How breed labels can impact shelter dogs is discussed.
Imagine meeting a potential roommate for coffee but instead of questions that gauge how compatible you both would be living together, you were asked about the ancestry of your parents' families. Though this situation seems ridiculous, it happens all the time in animal shelters where dogs are assigned breeds that are often just guessed from their physical appearance. These assigned breeds are then used to infer how the dogs might behave and also often impact the length of time a dog waits to be adopted.
The first step to understanding how breed labels might affect shelter dogs is to identify who shelter dogs actually are, and researchers at Arizona State University have done just that. The ASU scientists genotyped shelter dogs in Arizona and California and compared the genetic information to the breed labels assigned in shelters. The findings were published August 23 in PLOS ONE.
Canine cheek swabs
Who are shelter dogs? To answer this question, ASU's Canine Science Collaboratory researchers Lisa Gunter and Clive Wynne collected DNA from over 900 shelter dogs housed at the Arizona Animal Welfare League and Society for the Prevention of Cruelty to Animals (AAWL) in Phoenix, AZ, and the San Diego Humane Society and Society for the Prevention of Cruelty to Animals (SDHS) in San Diego, CA.
For the genetic testing, the researchers used the Wisdom Panel Canine DNA Tests from Mars Veterinary, which is a commercially available product. A small brush was used to collect cells from the dog's cheeks and gums, and the samples were sent to a lab for processing. At the lab, DNA was extracted from the dog's cells and compared to over 300 sites in the canine genome that have been matched to specific breeds.
The three most common breeds at both the AAWL and SDHS were the same: the American Staffordshire Terrier, Chihuahua and Poodle. Yet, these three breeds accounted for less than half of the dogs in the two shelters.
"The level of genetic diversity in the shelter dogs exceeded our expectations: we found 125 distinct breeds," said Gunter, who is a Maddie's Research Fellow in the ASU Department of Psychology. "We also found that just 5% of the shelter dogs were purebred, even though it is commonly assumed that up to a quarter of dogs in shelters are purebred."
The genetic testing gave the researchers information about three generations of ancestors for each dog. On average, most dogs were comprised of three different breeds, with some dogs having up to five breed signatures identified at the great-grandparent level.
"Breed identification has quite an outsize role in people's perceptions of dogs," said Wynne, professor of psychology and head of the Canine Science Collaboratory. "'What breed is he?' is often the first question people ask about a dog, but the answer is often terribly inaccurate."
The accuracy and unintended consequences of a label
The genetic diversity among shelter dogs can make it difficult for shelter staff to assess the breed heritage of dogs. Gunter and Wynne partnered with ASU's Rebecca Barber, assistant clinical professor in the Mary Lou Fulton Teachers College, to compare the breed information from the genetic testing to the labels given to the dogs in the San Diego shelter. At the SDHS, shelter staff used the physical appearance of a dog and breed descriptions from the American Kennel Club Breed Identification Guides to identify a primary and secondary breed.
When the researchers looked at whether either the assigned primary or secondary breed matched the information from the genetic profile of a dog, they found the shelter staff accuracy was 67%. The accuracy fell to 10% when staff identified more than one breed.
Though breed labels are common in animal shelters, they can have unintended consequences. In a previous study also published in PLOS ONE, the researchers found that dogs labeled as pit bulls waited over three times as long to be adopted. In this study, the researchers found that dogs in the San Diego shelter with a pit bull-type ancestry waited more than three times as long as other dog breeds.
How a breed label impacts a shelter dog extends beyond identification. The relationship between behavior and breed is murky except for certain behaviors that are breed-specific, like pointing.
"The genetics of behavior is so complex that a dog who is a cross of two breeds might not behave much like the typical members of either of its parents' families," Wynne said. "Then you have a situation where breed-typing is worse than stereotyping members of our own species. Breed labels would be better dropped altogether."
Beyond breed labels to behavior
The behavioral diversity even within single breeds has led Gunter and Wynne to advocate for the importance of behavioral assessments instead of breed labels. What really matters is a dog's behavior and how it might fit into an adoptive family, Gunter said.
"Shelter dogs are interesting and complex genetically," Gunter said. "They really are individuals, and labeling them with a single breed can minimize their uniqueness."
An important goal of the Canine Science Collaboratory is to design and validate a behavioral assessment that would provide insight into how dog behavior in a shelter translates to behavior in a home. Such an assessment would be more informative than breed labels and might positively impact shelter dogs.
The AAWL in Phoenix recently stopped publicizing breed information on their website and on kennel cards, though the breed labels are available upon request. The shelter, which has a long-standing relationship with the Collaboratory, made the decision based on the findings from the previous study that showed how certain breed labels increased the length of stay in a shelter.
"Everything about the life experience of a dog -- where he was before coming to the shelter or any medical issues he might have -- is what makes him who he is, not who his grandparents might have been," said Michael Morefield, director of marketing and communications for the AAWL. "When you adopt a dog, you are not adopting a bully, a German Shepherd or St. Bernard, you are adopting Jerry or Mo. When you love a dog, you don't love a German Shepherd. You love Jerry."