This value implies that 41% of the variation in spotting of guinea pigs in Wright's population was due to differences in genotype.
Estimating heritability by using this method assumes that the environmental variance of genetically identical individuals is the same as the environmental variance of the genetically variable individuals, which may not be true. Additionally, this approach can be applied only to organisms for which it is possible to create genetically identical individuals.
Heritability by pa rent-offspring regression Another method for estimating heritability is to compare the pheno-types of parents and offspring. When genetic differences are responsible for phenotypic variance, offspring should resemble their parents more than they resemble unrelated individuals, because offspring and parents have some genes in common that help determine their phenotype. Correlation and regression can be used to analyze the association of phe-notypes in different individuals.
To calculate the narrow-sense heritability in this way, we first measure the characteristic on a series of parents and offspring. The data are arranged into families, and the mean parental phenotype is plotted against the mean offspring phenotype (IFigure 22.17). Each data point in the graph represents one family; the value on the x (horizontal) axis is the mean phenotypic value of the parents in a family, and the value on the y (vertical) axis is the mean phenotypic value of the offspring for the family.
Let's assume that there is no narrow-sense heritability for the characteristic (h2 = 0); genetic differences do not contribute to the phenotypic differences among individuals. In this case, offspring will be no more similar to their parents than they are to unrelated individuals, and the data points will be scattered randomly, generating a regression coefficient of zero (I Figure 22.17a). Next, let's assume that all of the phenotypic differences are due to additive genetic differences (h2 = 1.0). In this case, the mean phenotype of the offspring will be equal to the mean phenotype of the parents, and the regression coefficient will be 1 (I Figure 22.17b). If genes and environment both contribute to the differences in phenotype, both heritability and the regression coefficient will lie between 0 and 1 (I Figure 22.17c). The regression coefficient therefore provides information about the magnitude of the heritability.
22.17 The narrow-sense heritability (h2) equals the regression coefficient (b) in a regression of the mean phenotype of the offspring on the mean phenotype of the parents. (a) There is no relation between the parental phenotype and the offspring phenotype.
(b) The offspring phenotype is the same as the parental phenotypes.
(c) Both genes and environment contribute to the differences in phenotype.
A complex mathematical proof (which we will not go into here) demonstrates that, in a regression of the mean phenotype of the offspring against the mean phenotype of the parents, narrow-sense heritability (h2) equals the regression coefficient (b):
(regression of mean offspring against mean of both parents)
An example of calculating heritability by regression of the phenotypes of parents and offspring is illustrated in Figure 22.18. In a regression of the mean offspring phenotype against the phenotype of only one parent, the narrow-sense heritability equals twice the regression coefficient:
Monozygotic (identical) twins have 100% of their genes in common, whereas dizygotic (nonidentical) twins have, on average, 50% of their genes in common. If genes are important in determining variability in a characteristic, then monozygotic twins should be more similar in a particular characteristic than dizygotic twins. By using correlation to compare the phenotypes of monozygotic and dizygotic twins, we can estimate broad-sense heritability. A rough estimate of the broad-sense heritability can be obtained by taking twice the difference of the correlation coefficients for a quantitative characteristic in monozygotic and dizygotic twins:
(regression of mean offspring against mean of one parent)
With only one parent, the heritability is twice the regression coefficient because only half the genes of the offspring come from one parent; thus, we must double the regression coefficient to obtain the full heritability.
Heritability and degrees of relatedness A third method for calculating heritability is to compare the pheno-types of individuals having different degrees of relatedness. This method is based on the concept that, the more closely related two individuals are, the more genes they have in common.
Was this article helpful?