Multiple Gene Traits

By Leos Kral

Simple traits such as coat color and the merle pattern are due to the actions of single genes. Individual traits can also be affected by multiple genes. In the preceding article, two forms of genetic heterogeneity were discussed where multiple genes were responsible for slightly different but discrete forms of the same trait or for the same form of a trait.

There are two other forms of multiple gene based inheritance of a single trait. For both of these forms, the genetic effect is additive. For one form the phenotypic effect of the genes is continuously additive while for the other form, the additive effect upon phenotype is only manifested after the genetic component accumulates to a threshold level. For purpose of discussion, I will refer to these types of multigenic traits as quantitative traits and threshold traits.

Quantitative Traits

Quantitative traits are usually measurable. In people, height, weight, and skin color are examples of quantitative traits. Height can be measured as the number of feet and inches, weight can be measured as the number of pounds, and skin color can be measured as the degree of dark pigmentation. Note that these traits can not be grouped into discrete categories but rather these traits represent a continuous gradient of variation within the population. While we may arbitrarily classify people as short, medium or tall, in reality, there are no discrete height measurements into which all people fall. If heights of all individuals were measured and the data plotted, the plot would look like the one in the diagram below. Note the continuous variation in height and the arbitrary categorization.

Quantitative traits are determined by the additive action of many genes. To oversimplify a bit for the sake of explanation, let us assume that height is determined by 5 different genes called A, B, C, D, and E. Each gene has 2 alleles. The "recessive" alleles a, b, c, d, and e make no contribution to height and the "dominant" alleles A, B, C, D, and E make equal contribution to height. An individual who has the genotype aabbccddee would be of the shortest possible height and an individual who has the genotype AABBCCDDEE would be of the tallest possible height. Individuals who have the genotype AaBbCcDdEe would be of average height. Note, that individuals who are AabbCCddEe and individuals who are aaBBccDDEe would also be of average height. Why, because the effect of the "dominant" alleles is additive and all three genotypes that give rise to average height individuals contain exactly 5 "dominant" out of the possible 10 "dominant" alleles. Which of the dominant alleles are present is not important. They all contribute equally to the final phenotype. As another example, note that individuals with Aabbccddee, aabbCcddee, and aabbccddEe genotypes would be all the same height and just a bit taller than aabbccddee individuals.

How are quantitative traits inherited? For example, let's say that dad is of the tallest possible height (AABBCCDDEE) and mom is of the shortest possible height (aabbccddee). Dad's gamete will contain all "dominant" alleles (ABCDE) and mom's gametes will contain all "recessive" alleles. All offspring that result from the fusion of these gametes will have the genotype AaBbCcDdEe and thus will be of average height.

Now, if these average height offspring mate with each other (brother-sister matings [this is illegal in real life, but this is only an example]), the gametes they produce can be of all possible types ranging from all "dominant" (ABCDE) to all "recessive" (abcde) and all combinations in between (AbcDE, aBCDe, abcdE, etc.). Therefore, the offspring of these brother-sister matings will be of all possible sizes, and if a large enough number of offspring were to be conceived, the distribution would look very much like the bell shape curve shown in the diagram earlier in this article.

Ever wonder why mating two perfectly proportioned dogs does not result in all perfectly proportioned offspring? That's because most desirable conformation characteristics are governed by multiple genes with additive effects. The most desirable dogs are those which are the most average for many traits. And as the height example illustrates, the average genotype produces the most variety of gametes with regard to gene combinations and hence the greatest variety of phenotypes in the offspring produced by those gametes.

Threshold Traits

In dogs, an accepted example of a threshold trait is hip dysplasia. Please note that the actual genetics of canine hip dysplasia have not been worked out and the following is a hypothetical example made up to explain the principle of inheritance of this threshold trait.

Let us assume that 5 additive genes are involved in determining this trait. Let's call these genes G, H, I, J, and K. Only the dominant alleles contribute to hip dyspalsia. Individuals who are gghhiijjkk are not dysplastic and have no chance of being dysplastic. Individuals who are GGHHIIJJKK are dysplastic. If dogs of these two genotypes were mated, all offspring would be GgHhIiJjKk and would not be dysplastic. Why would they not be dysplastic? Because they only have 5 dominant alleles and the threshold of expression is 7 dominant alleles. That is, the trait is phenotypically expressed only if 7 or more of the dominant alleles are present in the genotype of an individual. Which dominant alleles are present is not important. For example, both ggHHIIJJKk and GGHHIIJjkk genotypes would produce hip dyspalsia because both genotypes contain 7 dominant alleles.

When trying to minimize diseases such as hip dysplasia that are inherited as threshold traits, the common practice is to make sure that the parents are not affected. However, this strategy does very little to eliminate the condition from the breed. Non-dysplastic dogs can have a very high propensity to produce dysplastic offspring if they have a fairly high number of the dominant alleles. For example, two dogs that each have 6 dominant alleles can produce a sizable proportion of offspring with 7 or more dominant alleles.

So, how do you pick breeding stock that minimizes production of dysplastic offspring? Certainly, you want to check that the sire and dam are non-dysplastic, but you also want to make sure that none of their parents had dysplastic siblings and their own siblings were not dysplastic. The two pedigrees below illustrate the concept.

The pedigree on the left shows no history of hip dysplasia and, therefore, it can be concluded that individuals in this pedigree have a very small number of the dyspalsia producing dominant alleles. The odds are that offspring produced by individuals from this pedigree with individuals from similar pedigrees would have a very low probability of being dyspalstic.

The pedigree on the right shows a fairly high incidence of hip dysplasia. The affected individuals have 7 or more of the dysplasia producing dominant alleles. Any of the phenotypically normal individuals in this pedigree, while not dysplastic themselves, probably have a fairly large number of dysplasia producing dominant alleles. These normal individuals who have dyspalstic siblings, such as the male marked by "x" should probably not be bred. Of course if the male marked "x" is an exceptional specimen that the breeder feels must be bred, then care should taken to find a female from a hip dysplasia free pedigree (such as the one on the left) to minimize the probability of production of dysplastic offspring.

Pedigree analysis is not in itself sufficient to determine if a trait is inherited as a threshold trait. However, a comprehensive database of Australian Shepherd pedigree information should be of tremendous value to better understand the inheritance of those traits shown to be inherited as threshold traits.


Please note that the above examples are hypothetical. They were made up to explain the basic principles of inheritance of additive traits and threshold traits. Also note that the notion of all dominant alleles having an equal effect toward the phenotype is an oversimplification. In actuality, alleles of different genes can have major and minor additive effects. Current research being carried out to identify genes responsible for canine hip dyspalsia is geared toward identifying those genes that have the greatest effect.


Copyright 1998, 1999 Leos Kral.    Australian Shepherd Health and Genetics Information Resource and Health Registry.  All rights reserved.

  Return Home