Genetics of Dog Breeding

Any dog lover knows that Labrador retrievers are friendly, Dalmatians are hyper, and Australian shepherds are smart (Scott & Fuller, 1974). Some dog lovers also know that Labradors are susceptible to hip dysplasia, while deafness and kidney stones run in Dalmatians. But why is this the case?

Breeding dogs for particular characteristics, or phenotypes, has been going on for centuries. Dogs are companions and workers, in service to humans, and they have thus been bred to accentuate desired traits. For instance, Dalmatians have long been coach dogs, in part because of their striking looks and their comfort around horses. Bred for endurance, they can run alongside horse-drawn carriages all day. When kept as a housebound family pet, however, a Dalmatian's excess of energy can make the dog seem wired and can lead to less desirable behaviors, such as gnawing on furniture.

Dogs' closest living relatives are wolves. Analysis of the two species' genomes has revealed differences that some scientists believe are a result of dogs being subject to artificial selection imposed by humans. It appears that with domestication, beginning as long as 14,000 years ago, came a relaxation of selective forces typical of nature (forces that continued in earnest on wolves), as well as an increase in variability in the dog genome compared with the genome of their ancestral stock (Björnerfeldt et al., 2006).

One question that tugged at Swedish researcher Carles Vilà is how dogs can have such a wide variety of phenotypes—imagine a tiny Chihuahua standing next to a Great Dane, or a Chinese shar-pei peering from under its skin folds at an Old English sheepdog who peers back through its long hair. In fact, the variation among breeds of dogs is far greater than the variation among other completely distinct species in the family Canidae.

If dogs evolved from wolves, which seems to be the case, then wolves must have had the capacity for this diversity somewhere in their genomes. Thus, Vilà and his colleagues decided to compare the mitochondrial DNA of dogs and wolves in an attempt to understand the genetic consequences of these species' different lifestyles: domesticated versus wild. (Remember, both dogs and wolves evolved from a common ancestral wolf species, so wolves are an ideal control with which to study the consequences of dogs' life with humans.) The mitochondrial genome was used because of earlier work by Vilà that showed the nuclear genomes of dogs and wolves to be too similar to study their molecular evolution. On the other hand, this research indicated that mitochondrial lineages are clearly distinguishable for the two species.

Vilà hypothesized that certain mutations—those that might be deleterious, but not strongly so—accumulated faster in populations in which natural selection had been relaxed, resulting in a decline in fitness. In other words, after dogs started to live with humans, less fit individuals were more likely to survive and reproduce than they were in the wild. In addition, it is highly likely that dogs were strongly selected for certain behavioral traits, such as tameness. "It is therefore possible that this process led to an increase in functional genetic diversity throughout the entire dog genome," wrote Vilà, "including both genes and elements affecting gene expression." Such a relaxation of selective pressures might have led to the wide phenotypic diversity in dogs, as well as the variety of diseases seen in dogs today (Figure 1).

Many researchers have noted that beyond tameness, dogs appear to retain certain traits associated with juvenile wolves, especially behavioral traits such as whining, barking, and submissiveness. Russian geneticist Dmitry Belyaev focused on tamability as a guiding characteristic. His idea was not only that early humans would have selected the tamest animals to live with them, but also that selecting for a single trait could give rise to an entire set of changes in form, physiology, and behavior. Belyaev thus launched an experiment that would last longer than his life, seeking to test whether selecting for tameness would indeed produce a set of domesticated traits similar to those seen in dogs (Trut, 1999).

Belyaev chose the silver fox for his experiment; this species is related to the dog, but it is not domesticated. The initial foxes in Belyaev's experiment were not trained in any way, but simply tested for tameness at an early age. Starting at age one month, a human researcher would try to feed and pet the foxes, either alone or in the company of other foxes. The animals' responses varied from aggressive behaviors (such as biting), to indifference, to seeking interaction with the person more than with the other foxes. The tamest foxes were then selected for breeding the next generation, although fresh genes were supplied through continual outbreeding.

Belyaev and his colleagues did indeed create a population of foxes that differed in temperament and behavior from their wild cousins. The foxes changed physically as well, with alterations in coat color appearing as early as the eighth generation—typically a loss of pigment resulting in white patches. The foxes also developed floppy ears and curved tails, mirroring traits seen in dogs as well as other domesticated species.

One of Belyaev's hypotheses was therefore satisfied: Selecting for one trait (behavior) also changed other traits (here, aspects of the foxes' physical form). A common thread in many of the observed changes across the generations in this experiment was that the timing of key developmental steps had been altered. Belyaev predicted that hormonal and neurochemical differences would be evident, and that such changes would be regulatory in nature and would control early development in a top-down fashion. In particular, two developmental milestones were different in the tamer foxes: their eyes opened several days earlier, and their fear response kicked in about three weeks later than the norm for wild foxes. These two events might have worked together to increase the openness of young foxes to interacting with humans and doing so without fear. At the same time, Belyaev found reduced levels of the stress hormone corticosterone in the domesticated foxes. Even the changes in coat color were linked to changes in the timing of development.

In addition to tameness, another unique trait of dogs is their ability to understand humans. For example, if you point or even shift your gaze toward a certain object (say, a jar that contains dog treats), a dog will likely investigate the object (Hare & Tomasello, 2005). Even our closest animal relatives, chimpanzees, do not have this skill. Brian Hare, a German researcher, posits that this is an example of convergent evolution—the emergence of a trait (referred to by Hare as "social skills") that developed independently in two species. Hare's working hypothesis is that domestication comes first, and after the fear response has been tuned down enough, the development of social skills can take place.

Domestication of dogs has occurred over many millennia. More recently, the advent of controlled breeding practices has segregated genetic variability into distinct phenotypes. In fact, the evolution of the majority of dog breeds is a relatively recent phenomenon, beginning with selective breeding practices during the past 200 years. Today, various breeds demonstrate a huge variability in size and shape, as well as coat characteristics. Behavioral traits have also been bred based on humans' use of dogs for herding, hunting, guarding, and companionship. Phenotypic variation among dogs is currently partitioned into more than 350 distinct breeds worldwide; these breeds are largely closed populations that receive little genetic variation beyond that which existed in the original founders (Ostrander & Wayne, 2005).

"These restrictive breeding practices reduce effective population size and increase overall genetic drift among domestic dogs, resulting in the loss of genetic diversity within breeds and greater divergence among them," writes Ostrander, who participated in a landmark study of the genomic relationship of 85 different dog breeds. "For example, variation among breeds accounts for 27% of total genetic variation, as opposed to 5-10% among human populations" (Parker et al., 2004). Now that the dog genome has been sequenced, the potential to learn more about man's best friend, and perhaps ourselves, has increased (Ellegren, 2005).