A mutt is more than the sum of its breeds: Fauna Bio collaborates on large-scale citizen science study in dogs

Fauna Bio was founded on the idea that understanding the extraordinary genomes of animals will revolutionize therapeutic discovery in humans. From the extreme physiology of hibernators to the uniquely selected traits of individual dog breeds, many useful genomic secrets are being revealed with the help of animals. Dogs are a particularly remarkable study system for genomics. Their physical and behavioral traits are the most diverse among mammal species, and the combination of purebred and mixed-ancestry “mutts” makes it easier to discover connections between mutations and characteristics. Genetic associations to traits in dogs can typically be made with far fewer individuals than would be required for a similar study in humans, and the ability to compare across breeds and mutts allows for more precision in identifying mutations responsible for various traits.  

Laying the groundwork for this unique model system, a new study published in Science by researchers from UMass Chan Medical School, the Broad Institute, Fauna Bio, and others, investigates the relationship between breed ancestry and traits in dogs. Most people believe that knowing a dog’s breed can tell you a lot about their personality and appearance - “Golden Retrievers are friendly” and “Chows have purple tongues.” Likewise, people believe that they are able to identify the ancestry of a mutt through its appearance and behavior, but how much truth is actually behind breed stereotypes?

I found my own dog, Skyler, at an animal shelter below a sign that read “Labrador-Pit bull mix.” A staff member mentioned that she was previously owned by a local person experiencing homelessness, but nothing else was known about her past. As she was one of the dogs fully genome-sequenced for this study, I was excited to help out with the analysis to learn more about her origins and how mutations in her genome may influence her unique personality and appearance. With her black back and white belly and paws, I thought she looked more like a Staffordshire Terrier than a Labrador-Pit bull mix, but the genomics would decide.

The study was a monumental effort of citizen science that included 2,155 purebred and mutt dogs combined with 18,385 owner surveys to glean information about each dog’s appearance and behavioral traits and connect these back to their DNA. I have fond memories of capturing videos of Skyler's appearance and gait for future analysis, and answering detailed survey questions about her behaviors. – Does Skyler seek companionship from people? Strongly Agree! Does Skyler have difficulty finding food dropped on the floor? Strongly Disagree! – On average 100 questions were answered by each dog owner with the goal of connecting these traits to genetics.

While dogs began to differentiate from wolves 10-15,000 years ago, most dog breeds emerged in the Victorian era about 150 years ago. Could this relatively short period of artificial selection by humans - for ability to hunt, guard, or herd - really dictate personality traits of individual contemporary dogs? The short answer: not really. This study revealed that modern breeds are primarily distinguished by physical characteristics rather than behavior. Dr. Elinor Karlsson, the senior author on the work, explains: “The majority of behaviors that we think of as characteristics of specific modern dog breeds have most likely come about from thousands of years of evolution from wolf to wild canine to domesticated dog, and finally to modern breeds.” Analyzing 78 breeds for 73 behavioral traits revealed that breed explained only a small portion of behavioral differences between dogs (9 percent). Human sociability was the most heritable factor in the study, and while it’s true that a given Golden Retriever was more likely than a random dog to rank in the highest quartile for sociability (1.6-fold), there was a very high degree of variability within breeds, and some Golden Retrievers are in fact less social with humans than the average dog.

In addition to revealing that breed is a poor predictor of an individual dog’s temperament, this study also demonstrated that humans are poor predictors of a mutt’s breed ancestry. When adopting a mixed-breed dog, you might think you know something about their ancestry, but among 10,000 dog lovers, the average participant was only able to detect 20% of a dog’s ancestry (only 0.83 breeds guessed correctly on average per dog). Responses from professionals who work with dogs (e.g. veterinarians or animal shelter staff) only fared slightly better, correctly naming 0.93 breeds correctly per dog! Despite the guesses of Labrador, Pit bull, and Amstaff Terrier, Skyler turned out to be a mix of at least 10 different breeds, with Dalmatian being the highest percentage. I hadn’t guessed that she was part Dalmatian, though in retrospect she does have a few Dalmatian-like spots on her chest and feet.

Beyond this fun exploration into dog ancestry and personality, the study also helps disentangle behavior from breed stereotypes, further establishing dogs as a model system for the study of human diseases. Understanding the genetics behind human behavioral and psychiatric disorders has proven challenging, but the ability to identify genes that impact behavior in dogs will help inform human disease genomics. For instance, the study found that the trait “Gets stuck behind objects” maps to a region including the SNX29 gene, which is associated with human cognitive performance. Also, the “Dog howls” trait maps to a region with the SCN3A gene, which is involved in human language development. Dr. Karlsson explains, “While breed may be a poor predictor of behavior for individual dogs, there are subtle but real behavioral differences in some breeds reflecting thousands of years of evolution as dogs adapted to live alongside people. These differences make dogs an exceptional model for investigating how genetics affects behavior. By understanding how natural behavioral variation evolved in dogs, and its connection to canine behavioral disorders, including compulsive behavior and anxiety, we will gain new insight in the underlying causes of psychiatric diseases like OCD in people.”

While Fauna Bio typically studies animals with extreme physiologic superpowers, perhaps the greatest superpower of the dog as a disease model is their close proximity to humans - they live alongside us in our unnatural environment. I was evolved to chase prey on an African savanna, but instead I spend most days inside hunched over my laptop. Likewise Skyler spent most of her time plopped down on the couch beside me, a far cry from the daily life of her ancestors. The environmental factors that contribute to their diseases mirror our own, and like humans, dogs can develop neuropsychiatric diseases and can be successfully treated with human medications like selective serotonin reuptake inhibitors (SSRIs). There is great potential for man’s best friend to both inform and benefit from human medical treatments, and we’ve only begun to scratch the surface.