© 2005 Lowell Ackerman [No part of this material may be reproduced or distributed without express written consent of author].

Today, as at no time in the past, we can often determine the genotype (the genetic makeup for the trait) for specific diseases in pets, even though the phenotype (what we can see) for the trait may be "normal". Until now, we had to wait for an affected animal to be born to determine if the parents were carriers of the trait. Not any more! For a condition such as progressive retinal atrophy (PRA) in Irish setters, we now have a test that will determine genotype (the genetic makeup for the trait), even though the phenotype (what we can see) is normal for the trait (i.e., they don't have evidence of clinical PRA). So, even though we can see the phenotype for traits (hemophilia, PRA, hip dysplasia, and so forth), in most cases we have to infer the genotype from observing how traits seem to run in families. Only in the past several years has it become possible to determine the actual genotype for several different conditions.

As our knowledge of genetics expands, judging genotype becomes less of a guessing game and more of a science. A new DNA test has allowed that to happen for progressive retinal atrophy in the Irish setter and for several other diseases as well. In addition, genetic information can be used for nonmedical advances. For example, we can remove all the guesswork from predicting coat color genetics in several breeds, including the Labrador retriever, Doberman pinscher, American Cocker Spaniel, Poodle, and Scottish terrier.

To explain what we see (the phenotype), we often make up rules to help us infer which actual genes are involved (the genotype). This works fairly well for some problems (such as hemophilia) and not so well for others (hip dysplasia, for example).

GENETIC COUNSELING

With new methods of identifying genetic diseases, veterinarians may think they will have an easy time advising breeders. With the new technologies, though, come new responsibilities. This is especially true when a trait is highly entrenched in a breed. For example, approximately 75% of Bedlington terriers either are affected with copper toxicosis or are carriers of the trait. Because a DNA linkage test is now available for copper toxicosis, should we advise breeders to breed only animals that are "clear" for the trait? Unfortunately, things aren't that simple.

Even under perfect circumstances, the linkage-based test for copper toxicosis could be wrong 1% to 5% of the time. Still, it is a handy tool until a direct DNA test is developed. A dog that tests as marker type 1/1 is 90% likely to be completely clear of the trait. About 95% of dogs that are 1/2 are carriers of the trait. Those that are 2/2 are mostly affected (about 72%), but some are carriers (about 24%). The results of the test can be formally registered with the Orthopedic Foundation for Animals.

Even with direct DNA tests that detect animals that are clear, affected, or carriers with almost absolute certainty, enforcing a "clear only" policy of breeding may not be possible. It may not even be desirable. In so doing, we may select a very small number of animals to become the foundation for a new breed line. Although this is laudable, this narrow gene base conceivably could concentrate other deleterious genes instead. Accordingly, if mating only clear-to-clear animals is not possible, breeding clear to carrier may be necessary until a large enough population of clear animals is created so the gene pool isn't quite so shallow and will preserve some genetic diversity.

Selection of Breeding Animals

There are many ways to select animals for breeding purposes. Some breeders select animals based on their conformation, their appearance. Perhaps they feel the individual has an exceptional topline, or head shape. However, this does not take into consideration genetic health, so this method is not the best to recommend during genetic counseling. Another method is to set "culling levels" for traits. For example, the breeder may eliminate all dogs that have less than perfect hip conformation, or that test "positive" for carrier status of von Willebrand disease, etc. Although this allows for many characters to be considered, it has the tendency to be too rigid for a flexible breeding program. Perhaps the best option for the conscientious breeder and accommodating veterinarian is a selection index scheme.

The selection index is a numerical score assigned to an animal that provides a relatively objective measure of overall suitability for breeding. The index can be customized for each situation, but to be fair, should be consistent among all dogs evaluated. In my own selection index for show breeders, I generally assign 50 points for health concerns, 30 points for conformation, and 20 points for behavior. For performance dogs, I might assign 50 points for health concerns, 30 points for performance factors, and 20 points for behavior. For pet owners not interested in show or performance, more weight can be given to health and behavioral concerns.

In creating a selection index, it is also possible to attribute more significance to certain traits and less to others, while still remaining constant within groups. For example, with a particular collie breeder who is trying to eliminate collie eye anomaly from her lines, but has a negligible incidence of hip dysplasia, more of the health points may be assigned to CEA evaluation. It is also important to be aware of the heritability of individual traits when assigning significance. The lower the heritability, the more generations will be required before improvement will be seen. Here is an example of a possible selection index for our fictional collie breeder.

Hardy-Weinberg Law

Sometimes the magnitude of the problem is not immediately obvious. We can bring it more sharply into focus with something called the Hardy-Weinberg law. The Hardy-Weinberg law applies as long as there is no mutation (allelic changes), migration ("new blood"), or active selection for a trait. While the Hardy-Weinberg law is very useful in canine genetics, it must be remembered that dog breeding invalidates some of the basic tenets of the law, since matings don't occur randomly or without selection, the way they might in nature. Also, the population is rarely in equilibrium since breeders often bring in breeding animals from outside the local population. As such, gene frequencies predicted by the Hardy-Weinberg law may not be completely accurate. Still, much useful information is provided.

Test Mating

To establish a foundation stud to rid a line of a recessive disorder such as von Willebrand disease, it is helpful to confirm that the animal is not a normal-appearing carrier of the trait. This can be done with a test mating. The number of offspring needed for evaluation is simply a function of how sure we want to be. In a litter of n normal pups, the chance of a carrier remaining undetected is 0.5ⁿ and the chance that the animal is a non-carrier is 100-0.5ⁿ% Accordingly, there is less than a 1% chance that the stud is a carrier if there are seven pups, all normal. If any affected pups are produced, the dog must be a carrier, heterozygous for the von Willebrand disease trait.

For some recessive genetic conditions, breeding affected individuals in a test mating is not possible. For example, in some of the lysosomal storage diseases, affected animals may not survive until maturity. In these cases, a test mating can be done with our prospective stud and a known carrier, such as a dam that previously has produced an affected individual. In this case, in a litter on n normal pups, the chance of the prospective stud being a carrier and remaining undetected is 0.75ⁿ. In this situation, 11 pups, all normal, would be required to be 95% sure, and 16 pups to be at least 99% sure, that the prospective stud was not a carrier. If any affected pups are produced, the prospect must be a carrier, heterozygous for the von Willebrand disease trait.

Although establishing harsh criteria to rid breeds of genetic disorders by completely eliminating affected animals and carriers may seem reasonable, this is neither practical nor realistic. Because each animal carries about 30,000 different genes, ridding lines of all deleterious alleles is not possible. If carriers can be determined, however, breeding phenotypically normal dogs is a real possibility by never breeding two carriers together. If we are aware of heterozygotes, we can safely breed carriers with known normal individuals and we will never see cases of the disorders we are trying to avoid. Isn't that what genetic counseling is all about?

Trying to overcome polygenic traits takes longer and is more troublesome. Obviously, the higher the heritability of a trait, and the more ruthless we are at selecting superior individuals for breeding, the more successful our selection pressure will be. This response to selection (R) also can also be described as an algebraic function, R=h²S, where R is the response to our breeding strategy, h² is the heritability of the condition, and S is the selection differential, the phenotypic superiority of the parents we selected versus the general population from which they came. The result is that the mean improves; it doesn't mean that the pups will be superior to their parents.

Regardless of the mode of inheritance of a trait, genetic counseling is always more effective when veterinarians and breeders work as a team. This includes keeping records of affected animals, as well as those that are clear for a disorder. Registries are wonderful tools for genetic selection, if they are nonjudgmental so as to encourage free exchange of information. Closed registries release only information of animals certified as having desirable phenotype. Open registries provide information on specific animals and those to which they are related, including both desirable and undesirable phenotypes.

APPLICATION FOR VETERINARIANS

Why do veterinary practitioners need knowledge of genetic diseases and selection pressures? Despite the fact that genetics is a critical component of veterinary medicine in general, the discipline holds very real implications for effective medical practice. Issues of legal liability and medical competence also have to be addressed.

Although some veterinarians believe that breeders represent too small a segment of their practice to warrant gaining proficiency in genetics and genetic counseling, this is a poorly conceived argument. In most circumstances, when someone purchases a dog, whether from a pet store or a breeder, the dog must be evaluated by a veterinarian within a fixed time period (usually 2-5 business days) and found to be fit. During that visit, most astute practitioners will discern the most common anomalies, such as luxating patellas, umbilical hernias, congenital cataracts, heart murmurs, and the like. If the veterinarian finds the animal to be unfit, the purchaser often has the option of returning the animal, either for a refund or an exchange, depending on the seller's policies.

Let's investigate what happens when a veterinarian determines the pup to be fit. The purchaser, now the owner of the dog, starts to notice strange behavior in the dog at 4 months of age. The dog is an English springer spaniel, and, unsure of the diagnosis, you refer it to a neurologist. The final diagnosis is fucosidosis, with a grave prognosis. The dog will be dead before a year of age, but not before getting progressively more impaired than it is now. The owner asks the specialist how this could have been avoided and is informed of a simple blood test that should have been performed on the parents of the affected dog, which would have identified them as carriers.

The owner now has a dog that was accepted as a family member that will die shortly, will incur a huge veterinary bill, and has little hope for satisfactorily resolving the problem. Should the owner have expected more from the veterinarian selected to do the post-purchase examination? Should the veterinarian have informed the owner that he or she was examining the animal only for obvious problems the dog had at that point in time, and not for problems that may be prevalent in the breed but not yet evident?

The problem was driven home to the American public in an exposé that appeared in Time magazine on December 12, 1994. The correspondent concluded that up to 25% of the 20 million purebred dogs in the U.S.--1 in 4--were afflicted with a serious genetic problem, with costs of about $1 billion annually in veterinary bills and lost revenues from stillborn pups that couldn't be sold.

At least 13 states now have pet lemon laws that give purchasers of dogs (or cats) some recourse against people who sell dogs (or cats) that develop genetic problems of which the purchasers were not advised. Typically, the liability is limited to the cost of the animal. Once veterinarians become involved, however, the total costs of diagnosis and treatment are often many times the cost of the animal itself. Whether veterinarians share any culpability when they perform post-purchase examinations without advising owners of breed-related problems is something that should be discussed and settled within the profession, and hopefully not in the courts.

Currently, most veterinarians are not involved in the pet selection process until after the fact. This puts them and their clients at a distinct disadvantage. If a client asks, "Where can I get a good cocker spaniel?" few veterinarians take this opportunity to offer preemptive genetic counseling. On their own, few prospective dog owners know where to look for a dog or what questions to ask. All too often, they purchase a dog with "papers," mistakenly believing that this is some certificate of genetic superiority, a "Good Housekeeping Seal of Approval." Most people would be genuinely surprised to learn that this dog could be blind, have epilepsy, be dying of a lysosomal storage disease, and bear little resemblance to the breed it is supposed to represent, and would still qualify to get "papers." They are expecting assurances from "papers" while the registries are only doing what they are mandated to do--to register purebred dogs produced as offspring of registered purebred dogs.

Veterinarians are often at a loss when clients inquire as to where they might buy a problem-free purebred dog. While the standard answer is often "a good breeder," it is important to appreciate that it is almost impossible for prospective dog owners to differentiate a good breeder from a bad one. Since the goal is to match owners with healthy and well-adjusted dogs, veterinarians should probably spend more time on objective criteria by which clients can select dogs, rather than the source of dogs selected by owners. If clients are educated to only buy dogs from sources that focus strongly on health and behavioral issues, these same clients become better and more informed consumers, regardless of where they shop.

Risk Calculations

During the course of genetic counseling, veterinarians are often asked to comment on the "risks" of certain matings. For example, a breeder may be very interested to know the risk of progressive retinal atrophy in her pups from the mating of a specific stud and bitch. If the mode of inheritance is known for the breed, and family history or diagnostic testing can determine carrier status, the assessment is simple. If the condition is inherited as a autosomal recessive trait in the breed and it is known that both potential parents are carriers, it is simple to advise that each pup has a 25% chance of being affected.

If we don't know the carrier status of each potential parent, we can often use statistics to give an approximation of the risk. For example, with an autosomal recessive trait, when an affected animal results, both parents must be obligate carriers of the trait. If a full sibling is affected, but our animal of interest is not, there is a 2/3chance that it is a carrier and a 1/3 chance that it is not. A table can be used to determine carrier risk depending on the relationship to known affected dogs.

The probability of producing a homozygous recessive (affected) puppy can also be calculated from looking at known carriers and affected animals in the pedigree. Depending on how far back in the pedigree you need to look to find affected and/or carrier animals, a rough measure of the likelihood of producing an affected pup can be accomplished. One factor is assigned if an individual of the pedigree is affected, depending of generation, and half that amount is assigned if the individual on the pedigree is a carrier.

Sometimes, the situation is more complicated and simple statistics cannot be used to accurately predict risk. In these cases, risk can still be approximated using Bayesian risk calculations.

References

1.  Ackerman, L: The Genetic Connection: A Guide to Health Problems in Purebred Dogs. AAHA Press, 2000.