Australian Shepherd Health & Genetics Institute

Australian Shepherd Health & Genetics Institute

Generic selectors
Exact matches only
Search in title
Search in content

In the Mode: How traits pass in dogs, lines and breeds

by C.A. Sharp

First published in Fall 2002 Double Helix Network News  Rev. March 2013

What dog breeders do is not breeding dogs; normal, healthy dogs can do that without any assistance from us.  Breeders manipulate genes, encouraging some to pass on from generation to generation while at the same time trying to prevent others from doing so.  With somewhere around 25-30 thousand canine genes to work with and, for most of them, no way to know for sure exactly which versions, called alleles, a particular dog carries we are not doing much more than rolling dice unless we develop a thorough understanding of modes of inheritance:  How genes flow from a dog to it’s offspring, as well as down through generations of a line or breed.

Single-gene modes of inheritance

Inheritance from parent to offspring is the most basic and easiest to understand form of gene transmission.  Every dog has two copies of each autosomal gene.  (Autosomal genes are those that are not on the sex chromosomes.)  One of these copies came from its father and the other from its mother.  What combination of alleles it has is its genotype.  How the alleles interact with each other, other genes and the environment will determine what traits you will see in the dog, referred to as phenotype.

The most basic mode of inheritance is simple dominance.  Black vs. liver color is a classic example.  The allele for black is dominant; the allele for liver is recessive.  If a dog has at least one copy of the dominant black allele, it will be black.  For a dog to be liver, a color produced by the recessive allele, it must have two copies.  A black dog might produce liver puppies if it carried a recessive allele, but a liver dog cannot produce black puppies unless bred to a black.

You cannot tell from appearance whether a dog exhibiting a dominant phenotype like black is also carrying a recessive allele.  However, knowing the phenotypes of dogs in its pedigree can give you an indication of whether it might carry the recessive.  If a black dog has a liver parent you know that black dog is heterozygous, meaning it has two different alleles.  Such a dog will produce liver if bred to another dog with at least one copy of the liver allele.  It has a 50/50 chance of giving a liver allele to each of its pups.

When looking at pedigrees and thinking about autosomal dominant or recessive traits, the breeder should follow the pedigree back step by step along each path of ancestry and note where he first encounters a dog he knows was either one or two copies of the recessive allele.  In most cases 4 or 5 generations will be sufficient.  The closer up an ancestor with the recessive is, the more likely it will have been inherited.   Dogs which exhibit the trait have two copies of the recessive allele and will always pass the trait but a carrier which has only one may or may not. If you don’t know the genotype of the dominant phenotype individuals that lie between that ancestor and your dog, you can’t know for sure if the recessive was passed along or not.  The farther back the carrier is, the less likely the gene will have passed on.

By knowing how many and how far back are the ancestors that you know carried a recessive trait, it is possible to precisely calculate the probability that a dog has inherited the recessive allele.  Even the math phobic can have a good idea of what could happen just by studying the pedigree.  However there is one factor those who don’t want to mess with math need to keep in mind.  We tend to think of probability being halved with each generation:  Half the genes come from each parent, a quarter from each grandparent, an eighth from each great-grandparent, and so on.  This often leads people to the erroneous conclusion that the offspring of two carriers of a recessive trait all have a 50-50 chance of carrying the recessive allele.  This is not the case.

Matings of carriers can produce four allele combinations:  Homozygous (two copies) dominant, homozygous recessive, paternally inherited dominant heterozygote (one copy) and maternally inherited dominant heterozygote.  Three quarters of these are phenotypically dominant.  In our black/liver example, that would be three black puppies for every liver.  Among the black puppies, two out of three will be carrying the liver allele.  Therefore, the odds for carrying liver in any black pup out of such a cross are not 50/50 but 2 out of 3.

If a recessive trait is something you want, you can use this process to determine how likely you are to be able to produce it in a litter.  You can increase the likelihood that it will happen through your mating choices.  Conversely, if you do not want to produce the trait you can eliminate the risk of producing it by breeding known or possible carriers to dogs you know are homozygous dominant.

Some genes have an incomplete dominant mode of inheritance.  In this case each genotype will have a distinct phenotype, with the heterozygote being intermediate to the dominant and recessive phenotypes.  The merle color pattern is an excellent example of this.  Dogs with two recessive alleles are not merle, heterozygotes will be merle patterned, and those with two dominant alleles not only have merle patterning but frequently have considerable amounts of white markings and almost always have serious eye defects and deafness.  Since the phenotype always indicates the genotype, the breeder will know what alleles the dog has by looking at it.

If alleles are co-dominant, their traits will both be expressed in the heterozygote.  The genes in the Major Histocompatibility Complex, which governs important aspects of the immune system, are co-dominant.  Both maternal and paternal alleles will be active. In the case of MHC genes, it means the dog has a bigger arsenal to protect itself against disease than a dog whose MHC genes are mostly homozygous.

Some genes have more than two alleles.  Dominance between them may be simple, co-dominant or incomplete.  The gene that produces golden/yellow coat color and black masks has a clear dominance hierarchy among its alleles.  The most dominant allele will not produce mask or yellow, so dogs that have one copy will have coat color determined by other genes.  The most recessive allele, when homozygous, results in hair that is yellowish, as in Golden Retrievers and yellow Labradors.   The middle allele in the series is for a mask.  Any dog that has at least one copy of his allele and does not have a copy of the most dominant allele will have yellow hair with a darker mask on the face.  The color of the mask will depend on what other color genes the dog has.

If the multiple alleles are incompletely dominant, heterozygotes will be intermediate in phenotype to the two alleles resulting in a continuum of phenotype expression with the phenotype of a specific individual dependent on which pair of alleles it had.  (White markings were once thought to be due to this type of inheritance, but recent genomics research has found that there are at least two major and distinct white marking genes.)

Not all breeds will have all alleles possible for a particular gene.  Knowing which ones your breed has can be important.  For years Australian Shepherd breeders have been selecting against yellow, dilute, and sable because the colors are disqualifying though they did occur early in breed history.  As a result sable is absent or nearly so because it is dominant to tan trim and most Aussies have tan trim.  Both yellow and dilute (blue or Isabella) do occur but are rare.

There are two other proposed single-gene autosomal modes of inheritance:  Dominant with incomplete penetrance and dominant with variable expressivity.  With the former, a dog can have the genotype but sometimes will not.  In the latter, when or how trait presents can vary considerably.  As more and more is learned about how genes interact with each other and the environment, as well as how specific genes are structured and function, it is apparent that neither of these modes of inheritance actually arises solely from a single gene.

Inheritance of genes on the sex chromosomes differs from that of autosomal genes because the sex chromosomes come in two different forms:  X and Y.  Female mammals have two X chromosomes while males have an X and a Y.  The Y chromosome contains only a very few genes, all of them are related to specifically male traits.  The X chromosome contains a normal number of genes that produce a wide variety of traits not related to the sex of the individual.  However, only one copy of the X can work in any given cell, so females are a “mosaic.”  Which X operates in each cell is randomly determined during development.  This can be most clearly seen in calico cats, in which black and orange are phenotypes produced by different alleles of the same X chromosome gene.  Whether a calico cat has a black patch or an orange one at some particular spot on her body will depend on which of her X chromosomes was turned off in an embryologic ancestor cell.  The exact pattern will have no bearing on what she might produce beyond being an indicator that she has the potential to pass on both black and orange alleles.

If a particular X allele produces a disease, like hemophilia, it will occur most frequently in male offspring who have only one X.  Females’ mosaicism provides them with sufficient normal cells that they will be healthy.  (In the unlikely case of a female homozygote for hemophilia, whose father would have to be a hemophiliac himself, she would in most cases die during development.  If she did make it to birth she would hemorrhage to death no later then her first heat cycle. )

If a male has an X-linked disease, this is the only time (other than with traits that are clearly autosomal dominant) that a stud owner can truthfully and accurately insist that her stud was not responsible for the problem.  These diseases are inherited from mother to son.  Each daughter of such a mother has a 50/50 chance of herself being a carrier.   The mother of a carrier is probably also a carrier and the health status of her sons should be examined.  However, genes for hemophilia and some other X-linked diseases mutate with unusual frequency so one cannot assume that all the bitches on the direct female line were carriers.

Another form of inheritance in which sex plays a role is imprinting.  With imprinted genes, the phenotype will be determined by which parent the gene was inherited from.   This mode of inheritance is not common and all imprinted genes discovered thus far are involved with development or reproduction.  There is also evidence that epigenetics, a form of gene regulation that can be influenced by environment, may play a role in these traits.

 Multiple genes

Unfortunately, most traits are not inherited in a simple, single-gene fashion.  Many are polygenic, resulting from the action of multiple genes.  Often environment can influence these traits to some degree.  At the present time, there is no way to know the genotype of any particular dog for any polygenic trait.  The best the breeder can do is make an educated guess.  Phenotypes in polygenic traits represent a continuum, rather than a series of similar but more or less distinct types.  Canine hip dysplasia (HD) is a prime example.  Dogs can have hip joint conformation that ranges from superior to abysmal.  Two sound dogs can produce dysplastic offspring and dysplastics can produce sound pups.

With polygenic traits the parental contribution can be unequal.  A parent with just one or a few genes that produce the trait may have offspring that exhibit it if mated to a dog that has all the rest.    Or the trait may show up after many generations of absence because the right combination of genes finally happened to fall together.  With polygenic traits a breeder must consider the history of the trait in the family, rather than in the pedigree.    Dogs that have a family history of the HD (affected siblings, cousins, aunts/uncles or nephews/nieces) are more likely to produce HD than dogs which do not.  The more affected relatives there are, the greater the risk.

This kind of family analysis can be useful for producing desired traits as well as avoiding those not wanted.    For example, if a dog has an excellent front and comes from a family of excellent fronts, it is less likely to produce incorrect fronts than a similar quality dog that has unusually good front conformation for its family.

Sometimes genes that do not interact with each other produce traits that are nearly always found together.  Such genes are linked, occurring close together on the same chromosome.  Chromosomal near-neighbors are unlikely to become separated as the genes are shuffled prior to formation of sperm and eggs.   If a breeder observes that she cannot find a dog that has a trait she likes without it also having some other thing that she does not like, it may be that the traits are linked.  She may have to live with the one if she wants to have the other.

The genetics of the immune system are both polygenic and linked in an extreme degree.   The Major Histocompatibility Complex (MHC) is a set of linked genes that inherited as a unit called a haplotype.  The higher a dog’s level of inbreeding and the more recently that inbreeding has occurred, the greater the probability that the MHC haplotypes will be the same or very similar.  This can result in an impaired immune system, autoimmune diseases, and reproductive problems.  Risk of producing affected offspring is greatly reduced if the breeder makes an effort to produce heterozygous haplotypes by monitoring the degree of inbreeding through the use of coefficient of inbreeding (COI) calculations on proposed litters and opting for suitable mates that will produce lower COIs.

 Environmental effects

Genes do not act in a vacuum.  The environment a dog experiences in the womb and throughout its life impacts the action of its genes.  Even things experienced by parents may have epigenetic effects on the offspring.  Dogs are born with a certain genetic potential.  Whether and how much that inheritance comes to fruition depends on where it and its parents live and what it experiences, both mentally and physically.    The genetic contribution is often described as the “heritability” of a trait.

Heritability is a measure of how much phenotypic variation in a trait results from genes, rather than environmental effects.  Heritability estimates for hip dysplasia vary by the type of exam used, ranging from 54-76% depending on the focus of the exam (PennHip’s distraction index was 61% and OFA’s extended hip joint radiograph was 76%)1 meaning most of what you see in your dogs’ hip joint conformation is the result of genes rather than diet or exercise, the two most important environmental factors.   The higher the heritability, the more control the breeder has over the trait.

Some inherited traits, notably chronic autoimmune diseases, require an environmental trigger.    The dog must have the genes before it will have the disease; however it is possible that a dog will never develop disease if it never encounters something that triggers the immune system to start attacking its own body.  Such conditions are said to be genetically predisposed.   As with traits of high heritability, the genes must be there in order to produce the trait, no matter what the environmental conditions.

 Lines and Breeds

Understanding inheritance in individuals is only the first step a breeder needs to make.   Each individual dog is part of a larger population from which its mates will be selected and of which its offspring will become a part.    Not every breed will have every allele possible for each gene.   MHC haplotypes are an example of this.  The genes tend to have far more alleles than do other types of genes.  Pure breeds have fewer haplotypes than do mongrels because the breeds are a closed subset of the species.  How few haplotypes a breed has depends on its history and how much it has been subject to the effects of popular sires and prominent kennels.

Selection criteria need to be sufficiently broad, encompassing not just physical attributes, but health, behavior and temperament.  Strong selection for or against a particular trait or a few traits can skew a gene pool and inadvertently result in the lowered frequency or elimination of some alleles while at the same time increasing or “fixing” others.  (An allele is fixed in a population if it is the only one present; color gene alleles causing solid black body color are fixed in the Schipperke.)  Fixed genes may be good or bad, depending on what those alleles happen to do.

Breeding according to the current fashion via selecting for this year’s winning “look,” or the excessive use of a popular sire or the output of a prominent kennel can likewise skew a breed gene pool and result in unintended consequences.   The smaller the breed population, the greater the effect narrow selection criteria and breeding for fashion will have.

A line is an extended family of dogs.  It is developed by some degree of inbreeding and thus will necessarily lead to a sub-set of the alleles present in the breed.   The composition of this sub-set can be altered through the same things that will alter allele frequency in the breed.  Since a line is necessarily a smaller population the effects can be more drastic.   Desired traits can be made fairly uniform in a relatively few generations, particularly those that are easily observed and not much influenced by environment.  However, unwanted traits may become intractable features:  Epilepsy became just such a problem in show line and some working line Australian Shepherds in little over a decade.

New (or lost) alleles can be brought into a line by the simple expedient of outcrossing.  On a breed-wide scale, however, this can be difficult in our current system of closed registries.   For breeds with wide geographic distribution, imports may provide sources of fresh genetic material provided the exporting country’s registry is considered acceptable, the populations are not already substantially related, and the populations have not diverged in type to the point that breeders in each country consider the dogs in the other unsuitable.  In a few cases the American Kennel Club has allowed significant additions of fresh stock at the request of member clubs, most notably the admission of a few African tribal Basenjis and the acceptance, via the Field Dog Studbook, of some Salukis of recent desert origin from a small US registry.  But for some breeds, there is nowhere to go without cross-breeding with another breed of similar type.  This was done with several European breeds after they were pushed to the brink of extinction by one or both World Wars.

The breeder’s task is to effectively utilize what is known about modes of inheritance for breed traits, both positive and negative, in order to produce quality dogs that not only meet his competitive or performance goals but which are also physically and mentally healthy.  He must at all times remember that he does not act in isolation.  Whatever he does will have an impact on breeders that follow.  The greater his success, the greater his impact for good or for ill will be.


1.  Z. Zhang, L. Zhu, J. Sandler, et al, “Estimations of heritabilities, genetic correlations, and breeding values for four traits that collectively define hip dysplasia in dogs”, American Journal of Veterinary Radiology, Vol. 70, No. 4, April 2009, pp. 483-492