1 2 3 4 5 6 7 8 Phenotypic value (units of pigment)

Figure 9.5 Phenotypic distributions for a trait determined by two loci with either complete dominance or epistasis.

(a) Due to dominance, the majority of phenotypes are 9 units of pigment and no individuals display phenotypes with 3 or 7 units of pigment. In this example of two diallelic loci, the frequency of heterozygotes is at a maximum because all allele frequencies are V2. In general, total phenotypic variation in the population can be increased or decreased by dominance.

(b) The bottom phenotypic distribution shows an example of epistasis identical to the two-locus system that determines coat color in Labrador retriever dogs. One dominant locus controls pigment color (BB and Bb genotypes have black coats and bb genotypes have brown coats) while a second completely dominant locus controls the presence or absence of pigment in hairs (AA and Aa genotypes have pigmented hair and aa genotypes have unpigmented hair). In this example of dominance by dominance epistasis, phenotypic expression of the genotype at the coat color locus depends on the genotype at the pigmentation locus. In both graphs, the genotype frequencies are those expected under Hardy-Weinberg assumptions where all allele frequencies are V2. The mean and variance are based on a population of 1000 individuals.

increases VP somewhat compared to additivity (see Fig. 9.2) because more genotypes (heterozygotes and dominant homozygotes) exhibit genotypic values at the upper extreme of the distribution. In general, dominance can either increase or decrease VP depending on allele frequencies in the population. The Mendelian basis of dominance and the dominance variance are explained in detail in Chapter 10.

Dominance genetic variance (VD) The proportion of the total genotypic variance (VG) caused by the deviation of genotypic values from their values under additive gene action caused by the combination of alleles assembled into a single-locus genotype. Epistasis or interaction genetic variance (V) The proportion of the total genotypic variance (VG) due to the deviation of genotypic values from their values under additive gene action caused by interactions between and among loci.

Epistasis literally means "standing on" and denotes interactions between two or more loci that dictate the phenotypic value of a multilocus genotype. When epistasis is absent, the phenotypic value of a multi-locus genotype is the sum of the phenotypic value of all single locus genotypes. With epistasis, the phenotypic contribution of the genotype at one locus depends on the genotypes of other loci that it is paired with. When there is epistasis then the combination of genotypes at two or more loci dictates the phenotypic value of a genotype. In the most extreme case, epistasis produces a phenotypic value for each multilocus genotype that is unique to that combination of genotypes. For example, with epistasis the phenotypic value of the AA genotype paired with the BB genotype (AABB) cannot be predicted from the phenotypic values of the AABb or AAbb genotypes. The interaction genetic variance (Vj) is caused by deviations of genotypic values from the values that would occur if each locus had additive effects.

Since the interactions between two loci can take many forms, a range of terminology has been used to describe the impact of locus interactions on pheno-typic values (see Phillips et al. 2000). For example, the term synergistic epistasis describes the situation where a genotype has a larger effect on the pheno-typic value in the presence of certain other genotypes than would be expected under additivity. In contrast, antagonistic epistasis describes an interaction where a genotype has a smaller effect on the phenotypic value in the presence of certain other genotypes than would be expected under additivity.

A classic example of two-locus epistasis are the coat-color phenotypes of Labrador retriever dogs (see Fig. 9.5b). For the sake of illustration, assume one completely dominant locus controls pigment color, with BB and Bb genotypes having black coats and bb genotypes having brown coats (see Kerns et al. 2007 for a more complete description of the genetic basis of coat color in dogs). A second completely dominant locus controls the presence or absence of pigment in hairs. At the pigmentation locus, AA and Aa genotypes have pigmented hair and therefore exhibit the black or brown coat color determined by the B locus. (The pigmentation locus is often symbolized as the E locus but A is used here for consistency across examples.) However, if the pigmentation locus genotype is aa then the coat color locus has no effect and the coat is yellow because hair pigment is not produced. In a population of Labrador retriever dogs, interaction between the coat color and pigmentation loci will alter the mean and variance of coat color phenotypes relative to two loci that combine additively. The exact impact of the interaction on the phenotypic mean and variance will depend on the genotype frequencies at the two loci.

The distinction between phenotypic variation produced by additive gene action and dominance or epistasis helps clarify a subtle point of terminology in quantitative genetics. The problem is what exactly to call VG. It is sometimes called genetic variation because multiple alleles in a population lead to variation that under additive gene action produces multiple phenotypes. But with dominance and epistasis, genetic variation is a product of multiple genotypes that then produce a range of phenotypes in a population. Calling VG genotypic variation is probably best since it encompasses phenotypic variation due to both alleles and genotypes.

Inheritance of additive (VA), dominance (VD), and epistasis (V,) genotypic variation

Another way to appreciate the differences among the VA, VD, and Vj components of the total genotypic variation is to consider an example of inheritance across a generation. Additive genetic variation (VA) has a distinct pattern of inheritance compared to dominance and epistasis genetic variation (VD and Vj). A critical distinction among the three components of the total genotypic variance is that VA is caused by the average phenotypic effects of alleles while VD and Vj are caused by the average phenotypic effects of genotypes.

Additive genetic variation is the component of the genotypic variance that causes the phenotypic resemblance between relatives. For example, parents and their offspring or siblings (brothers and sisters) have a higher degree of phenotypic resemblance than two randomly sampled individuals in the same population. This average phenotypic resemblance comes about because relatives share alleles that are identical by descent. When alleles combine additively, then shared alleles translate into shared phenotypic values. Only when alleles have additive effects does genetic variation contribute to average resemblance between parents and offspring or among related individuals.

Examples of additive gene action across one generation are shown in Table 9.3. In the top half of the table alleles at one locus are assumed to act additively to determine phenotypic values. When crossing BB x bb parents or Bb x Bb parents, the parental mean phenotype and the progeny mean phenotype are always identical. The equality of mean pheno-types across one generation is remarkable giving that the genotypes of the parents and progeny are not identical. In fact, in the BB x bb cross, none of the progeny share their genotype with the parents since all progeny possess Bb genotypes. What is common between the parents and progeny in each case are allele frequencies. As long as alleles combine additively to determine phenotypic values, identical allele frequencies in two separate populations will produce identical mean phenotypes. This can be seen well in the second cross (B) for additive gene action, where a population of all Bb heterozygotes has the same allele frequencies and same mean phenotype as a population of 1/4 BB, 1/2 Bb, and 1/4 bb individuals.

In contrast to the additive effects of alleles are the genotype effects of dominance and epistasis. Dominance can be thought of as an interaction between the two alleles that make up a single-locus diploid genotype. With dominance, the genotypic value of the heterozygote is not just the sum of the two allelic effects but is some other value depending on how the two alleles interact when packaged into one genotype. While dominance is a continuous variable, complete dominance, where the heterozygote and the dominant homozygote have identical phenotypes, is a useful point of reference. In complete dominance, alleles behave differently in how they contribute to pheno-type depending on whether a genotype is composed

Additive gene action

Genotypes BB Bb bb Phenotypes 3 2 1

0 0

Post a comment