Deconstructing the Gene Pool

Dr. Mark Neff and his team uncover the surprising origin of a potentially deadly mutation
By Jane Brackman PhD, November 2008, Updated June 2021

Until the mid-19th century, any hodge-podge of similar-looking dogs performing similar tasks was awarded the right to be called a breed. However, as inventions (such as guns) mechanized jobs that dogs normally performed, many breeds—like the tumbler, who “tumbled and turned” to mesmerize prey—simply sank back into the ancestral soup, taking their unique traits with them.

One of these ancient breeds, the glacier-climbing Lundehund with its unusual polydactyl triple-jointed toes, survived, but its current population is so small that the breed teeters on extinction’s edge. And a few, like the ubiquitous working collies and spaniels of Great Britain, spawned a number of the breeds created during the prosperous, class-conscious Victorian era. In the age of upward mobility, those on the way up claimed many of the privileges of the upper class, including the luxury of breeding and showing dogs.

More than one-quarter of the world’s estimated 375 breeds were created between 1859, when the first dog show was held in the UK, and 1900, when Westminster and Crufts were well established; even the most subtle differences in weight or color were enough to allow registry of a new breed type. In many cases, the subdivision of farm dogs was an unintended consequence of competitive exhibition in dog shows.

Responding to the shows’ strict criteria for body type, size and color, breeders drew from an increasingly smaller number of founder populations to create dogs who conformed to these standards. Breeding closely related dogs to one another became a popular way to refine a breed, which today means a group of dogs with a common gene pool and characteristic appearance and function.


Sign up and get the answers to your questions.

Email Address:

Unfortunately, the down-side of homozygosity (having two identical genes at a specific location on the DNA strand) is disease and unsoundness. As a consequence of this intense concentration on form, modern dogs suffer from more than 350 genetic illnesses, and today’s breeders bear the burden of restoring their lines to health.

There are no easy answers. Removing affected individuals from breeding populations may decrease the incidence of a particular disease, but smaller gene pools create opportunities for other congenital problems. In cases where an entire breed is afflicted, out-crossing with other breeds means running the risk of losing truly unique traits, such as the Lundehund’s joint flexibility.

Recent research has shown that a single mutated gene, unnoticed for over a century, is responsible for sensitivity to several modern medicines, ranging from ivermectin (a common ingredient in heartworm preventatives) to anticancer agents such as vincristine. These adverse drug responses can cause illness or death in the breeds that harbor the mutation.

A team of researchers led by Professor Mark Neff at UC Davis expanded the results of earlier research by demonstrating that the mutation probably originated in a single generic herding dog who lived in Great Britain in the mid-1800s. This dog must have been a common ancestor of founding dogs for nine different breeds, all of which were found to possess the mutation. Moreover, scientists involved in this study were able to describe the frequency of the mutation in these various breeds, further defining the inherited risk of adverse drug response: Collie (54.6%), Long-haired Whippet (41.6%), Miniature Australian Shepherd (25.9%), Silken Windhound (17.9%), McNab (17.1%), Australian Shepherd (16.6%), Shetland Sheepdog (8.4%), English Shepherd (7.1%) and the Old English Sheepdog (3.6%).

Dr. Neff talks about his research and the implications of genetic testing on the health and well-being of dogs.

Jane Brackman: In addition to helping breeders make informed decisions, your findings provide an opportunity for veterinarians to treat dogs based on their individual genetic profile. What do you mean when you say dogs can be treated with personalized medicines?

Mark Neff: The study of how individuals respond differently to medicines due to their genetic makeup is called pharmacogenetics, and it’s an intense area of investigation in human genetics. Probably all of us are aware of instances where one person responds positively to a medicine and is cured, while another person responds negatively or not at all. These differences are often tied to variation in genes. If we knew the genes that were responsible for side effects, we could identify the individuals at risk and prescribe the medicine that avoids a reaction and still provides relief. The same opportunity exists for dogs. Not all Collies have the mutation; those that don’t can be treated with ivermectin, which is an effective drug for its purpose. The Collies who do have the mutation can be treated with a different medicine.

JB: In dogs with the mutation, what happens?

MN: The normal product of this gene is a protein pump that can eliminate toxic chemicals from the central nervous system, thereby protecting the brain. The mutation causes the pump to be defective. If both copies of the gene are mutated, no functional pump is produced at all, which is the worst scenario. When the dogs are given a drug like ivermectin, which is toxic to neurons in high doses, the drug accumulates in the central nervous system, killing nerve cells. However, dose-sensitivity is an important issue. The smaller dose of ivermectin used to prevent canine heartworm infection, for instance, does not appear to be a problem for these dogs regardless of the mutation, but the higher dose used to treat mange can be fatal in a dog with two defective genes. The dose sensitivity varies by drug as well, so there’s still a lot to be sorted out.

JB: You’ve said that you think there is currently a disconnect between canine genetic research and the application of genetic knowledge. What do you mean?

MN: Breeders and owners are beginning to be inundated with DNA test results, which will only increase in the next few years. There’s an unmet need of genetic counseling that ideally would accompany DNA test results. In addition, we as researchers typically don’t have all the information we need to advise breeders on integrating test results with their breeding strategies. For example, based on our data, we think that both copies of the gene need to be mutated to acquire supersensitivity to ivermectin, but this may pertain to only some of the breeds with the mutation. Sighthounds have very different physiologies from Collies, for instance, and this could alter the effects of the mutation. Science always involves uncertainty, and it’s difficult to convey the ambiguity that remains. Katrina Mealey, our collaborator at Washington State, is continuing the research and adding a lot more detail to this particular story.

JB: In the research article, you advise re-examining how a breed is defined genetically. Would you elaborate?

MN: This is a statement more for academic geneticists than for breeders and owners. There’s a lot of scientific work going on now that neglects the fundamental fact that dog breeds are not natural species, but rather, have evolved through selective breeding and intentional outcrossing to produce new combinations of traits and hence new breeds. Most studies describing breed relationships use statistical and computational tools that were developed to describe relationships between species with distinct lineages. These tools are inappropriate for analyzing breeds of dog. Our paper showed that an identical mutation existed in two very different types of dog, two sighthounds and seven herding breeds. Conventional tools would have almost certainly missed the relatedness of these breeds.

JB: Your research identified seven affected herding breeds, which were mostly developed after the mid-19th century, and provided evidence that these breeds are in fact closely related. How did you draw those conclusions?

MN: The herding breeds, whose origins were in Great Britain, all shared the mutation, indicating that they had common ancestors. This implies that these types of dogs were mixing, presumably before registries were established. If the mutation was shared by these breeds, there probably are many other genes also shared. Great Britain in the 1800s appears to have been a real melting pot for dog genes, and this has been substantiated by subsequent work in our laboratory.

JB: Why is it important to know the history of a genetic mutation?

MN: Because the information may tell us something about the distribution of the mutated gene, and key us into where we might look to find additional affected breeds. For instance, if this particular mutation had been ancient, we would have predicted that many more breeds would have had it. Given all the breeds that exist worldwide, and that only a small fraction of these were affected by the mutation, we can infer that the mutation is relatively young. The mutation probably arose in the mid-1800s, but this is speculative. Reconstructing history, genetic or otherwise, always involves an incomplete data set—you never know what’s been lost.

JB: Should breeders genetically test their breeding stock, and if so, how should they apply the test results?

MN: Breeders have it hard—when they make decisions, they have to consider what’s good for the breed, their bloodline and individual dogs. If the mutation frequency is high in the breed, like it is with Collies and Long-haired Whippets, testing and selective breeding is certainly worthwhile to reduce the frequency and decrease the breed’s risk of drug sensitivity. In a breed where the mutation is relatively rare, such as the Old English Sheepdog, the owner might test only those dogs that are scheduled for treatment with one of the interacting drugs. But it gets complicated when a single breeder’s bloodline is heavily affected; breeders cannot give up their entire breeding stock, so a more gradual approach must be taken. What a breeder should do really depends on the specific situation.

I certainly believe that eliminating mutations from gene pools requires greater debate. Many scientists have suggested that a mutation should be removed from a gene pool gradually so as to preserve genetic diversity in the breed. I think ulterior motives are sometimes at work, as this is most strenuously advocated by service laboratories that offer these kinds of tests. DNA tests should be targeted for obsolescence—ultimately, a mutation should be eliminated from the gene pool, which of course renders the test meaningless. Don’t get me wrong—there is enormous value in preserving genetic diversity across breeds, but most of the diversity in one breed exists in related breeds, so conservation genetics within a breed is not a critical issue. I’m certain that far more genetic diversity is lost from “popular sire” effects and line breeding than from adherence to DNA testing.

JB: When I think of conservation genetics, I think of bringing a species back from the brink of collapse, or more generally, the practice of making a species healthy again. Do you mean that preserving genetic diversity in dog breeds is not critical because the way you create a breed in the first place is by eliminating diversity?

MN: That’s right. The normal rules of conservation genetics don’t apply to artificially selected populations. For instance, there appears to be an advantage in nature for individuals to have a lot of genetic variation in at least one region of the genome—the MHC—that arms the body’s self-defense. It appears that individuals may actually select mates in part based on being genetically different for genes at the MHC. In dogs, we have obliterated this selection because dogs don’t get to choose their mates.

There is an enormous difference between a species that breeds and evolves naturally and one whose breeding and evolution are controlled by humans. The irrefutable fact of closing a gene pool by enlisting dogs to a closed registry to suspend change in a breed is that diversity is being lost, and there’s little or no opportunity to create new diversity. Mutations are incredibly rare, at least those that have effect, and there isn’t any new blood coming into these populations.

JB: So trying to “conserve” diversity, for instance, by not strictly adhering to a DNA test result, doesn’t make much sense to you?

MN: There’s nothing more important to the survival and adaptability of a species than genetic diversity. What I’m saying is that it seems incongruous to hold adherence to DNA testing to a different standard when the gene pools of these breeds have been closed and breeders are striving for conformation to a standard. I’m not sure I’m aware of a single instance where dogs have been brought in to add new “blood” to a breed.

JB: You said that in many cases, even when breeds are crossed with a dissimilar breed, it doesn’t take too long to re-establish the breed’s classic look and behavior. Why is that so?

MN: Most breed-defining traits are shared by multiple related breeds. For example, I would predict that the genes for pointing behavior are common to perhaps a dozen or more pointing breeds. So one could resurrect a breed by bringing together the right combination of genes from related breeds. This won’t always work because there are some characteristics that truly are unique, such as polydactyly in the Lundehund.

What makes breed development rapid is that most selected traits are tied to genes of big effect. So instead of moving a hundred genes of small effect through selective breeding, one only needs to move two or three. I’m speculating on this last part. We don’t entirely know this, but it’s a reasonable guess.


Article first appeared in The Bark, Issue 34: Jan/Feb 2006

Jane Brackman, PhD, is an authority on the cultural history of canine domestication and the author of two books on pets in 19th-century America.