- Covariance among relatives is a fundamental concept in quantitative genetics that describes the statistical similarity between relatives for a trait. Relatives tend to resemble one another because they share genes inherited from common ancestors. By measuring how phenotypic values covary among relatives, researchers can estimate components of genetic variance, particularly additive genetic variance, and investigate the inheritance of complex traits.
- Covariance measures the degree to which two variables vary together. In studies of relatives, it describes whether individuals with higher or lower values for a trait tend to have relatives with similarly higher or lower values. Positive covariance indicates that relatives tend to resemble one another, whereas covariance close to zero indicates little systematic association under the conditions of the study.
- The covariance between two individuals can be expressed as: Cov(X,Y) = E[(X − μX)(Y − μY)]
- where X and Y are measurements for two individuals and μX and μY are their respective population means. In quantitative genetics, the important question is how much of this covariance is caused by shared genetic effects rather than shared environmental conditions or other factors.
- Relatives share different expected proportions of their genes depending on their relationship. For example, full siblings and parents and offspring share, on average, approximately one-half of their segregating genetic material, while more distant relatives share smaller expected proportions. These relationships allow researchers to use family resemblance to infer genetic contributions to phenotypic variation.
- The expected genetic covariance between relatives depends strongly on their coefficient of relationship. For additive genetic effects, the covariance between two relatives is commonly expressed as: Cov(Aᵢ,Aⱼ) = 2φᵢⱼVᴬ
- where Aᵢ and Aⱼ are the additive genetic values of individuals i and j, φᵢⱼ is their coefficient of coancestry, and Vᴬ is additive genetic variance. Because the additive relationship coefficient is often written as twice the coefficient of coancestry, the relationship can also be expressed using the additive relationship coefficient.
- For parent–offspring pairs and full siblings, the expected additive genetic covariance is approximately one-half of the additive genetic variance under standard assumptions. For more distant relatives, the expected covariance is smaller because they share fewer genes inherited from common ancestors.
- This relationship makes covariance among relatives particularly useful for estimating additive genetic variance. If closely related individuals show greater resemblance than unrelated individuals, part of their phenotypic similarity may be attributed to shared genetic effects.
- However, relatives may also share environmental conditions. For example, siblings may grow up in the same household, receive similar nutrition, attend the same school, or experience similar socioeconomic conditions. Animals from the same family may share housing, feeding, management, or maternal environments. Therefore, observed covariance among relatives can contain both genetic and environmental components.
- This is an important limitation of simple family-based estimates. Shared environmental effects can make relatives appear more genetically similar than they actually are. Conversely, environmental differences can reduce the observed resemblance between genetically related individuals.
- Heritability is closely connected to covariance among relatives. Narrow-sense heritability is defined as: h² = Vᴬ / Vₚ
- where h² is narrow-sense heritability, Vᴬ is additive genetic variance, and Vₚ is phenotypic variance. Covariance among relatives provides one way of estimating Vᴬ and therefore contributes to the estimation of narrow-sense heritability.
- Different types of relatives provide different information about genetic variance. Parent–offspring covariance, sibling covariance, half-sibling covariance, and other family relationships can be used in quantitative genetic analyses. Each design has different assumptions and sensitivities to environmental confounding.
- In a simple parent–offspring regression, offspring phenotype can be regressed against the phenotype of one parent or the average phenotype of both parents. Under appropriate assumptions, the slope of the regression provides information about heritability and additive genetic transmission.
- For example, when offspring phenotype is regressed on the average phenotype of both parents, the expected regression coefficient is related to narrow-sense heritability under a standard additive model. This approach relies on assumptions about environmental effects, mating patterns, population structure, and the absence of systematic confounding.
- Full-sib covariance provides another source of information. Full siblings share, on average, one-half of their additive genetic effects, but they may also share dominance effects and environmental conditions. Consequently, sibling resemblance can reflect more than additive genetic variance alone.
- Half-sib designs are particularly useful in quantitative genetics because paternal or maternal half-siblings share one parent but usually do not share the same other parent. The covariance among paternal half-sibs, for example, can provide information about additive genetic variance while reducing some forms of shared environmental confounding.
- The classic paternal half-sib design has been widely used in animal breeding and quantitative genetics. When offspring from different females are grouped according to their common sire, the covariance among paternal half-sibs can be used to estimate components of genetic variance under an appropriate statistical model.
- Covariance among relatives can also be studied using pedigrees. A pedigree records relationships among individuals across generations and allows researchers to construct a relationship matrix. The relationship matrix describes the expected genetic relationships among individuals based on their known ancestry.
- The additive relationship matrix, often called the numerator relationship matrix, represents expected additive genetic relationships among individuals. It is widely used in quantitative genetic models and animal breeding to account for genetic relatedness when estimating breeding values.
- Modern genomic approaches can extend this framework beyond recorded pedigrees. A genomic relationship matrix estimates realized genetic similarity using genome-wide genetic markers. Unlike a pedigree relationship, which describes expected relatedness, genomic relationships can capture the actual amount of genetic material shared between individuals.
- Genomic relationship information is particularly useful because individuals with the same pedigree relationship do not necessarily share exactly the same proportion of their genomes. Mendelian sampling causes offspring to receive different random combinations of parental alleles, producing variation in realized genetic relationships among relatives.
- Covariance among relatives can therefore be analyzed using either pedigree information or genomic information. Genomic relationship matrices are especially important in modern breeding programs and genomic prediction because they allow genetic covariance to be estimated using large numbers of DNA markers.
- Covariance among relatives is also related to breeding value. Breeding value represents the additive genetic contribution that an individual is expected to transmit to its offspring. Because relatives share additive genetic effects, information about relatives can be used to improve estimates of an individual’s breeding value.
- This principle is central to Best Linear Unbiased Prediction (BLUP). BLUP combines phenotypic records, pedigree relationships, fixed effects, and other information to predict genetic values. Information from relatives can improve predictions, especially when an individual has limited phenotypic information of its own.
- In animal breeding, covariance among relatives has historically been used to estimate breeding values for traits such as growth rate, milk production, fertility, disease resistance, and meat quality. Modern systems combine these family relationships with genomic data to improve the accuracy of genetic evaluation.
- Covariance among relatives is also important for studying genetic correlations between traits. When the same genetic factors influence two traits, relatives can show correlated differences in both traits. Genetic covariance between traits is therefore an important component of multivariate quantitative genetics.
- For two traits, genetic covariance can be represented as: Cov(A₁,A₂)
- where A₁ and A₂ represent additive genetic values for the two traits. Positive genetic covariance means that genetic factors increasing one trait tend to be associated with increases in the other, whereas negative covariance indicates an opposing genetic relationship.
- Genetic covariance contributes to the genetic correlation between traits. Genetic correlations are important because selection for one trait can cause correlated changes in another trait. Understanding covariance among relatives therefore helps breeders predict both direct and correlated responses to selection.
- Covariance among relatives also provides information about genetic architecture. Differences in covariance patterns among different classes of relatives can help researchers investigate the relative contributions of additive effects, dominance, epistasis, and environmental effects.
- Dominance variance can contribute to covariance among certain relatives. Full siblings, for example, have a greater expected similarity in their dominance relationships than parent–offspring pairs. Carefully designed family studies can therefore help distinguish additive and non-additive genetic components.
- Epistatic variance can also contribute to resemblance among relatives, although estimating epistatic effects is generally more difficult. Interactions among loci depend on the particular combinations of alleles present in individuals and can be difficult to separate from additive and environmental effects.
- Covariance among relatives is influenced by inbreeding and population structure. Inbreeding changes the distribution of genotypes and increases homozygosity, while population structure can create correlations among individuals because of ancestry rather than direct biological relationships. Quantitative genetic models therefore need to account for these factors when appropriate.
- Mating patterns can also influence covariance. Assortative mating, in which individuals with similar phenotypes or genotypes are more likely to mate, can alter the genetic relationships among individuals and affect estimates of genetic parameters.
- Covariance among relatives has applications beyond traditional breeding. In human genetics, family resemblance can be used to study the genetic contribution to complex traits and diseases. Twin studies, adoption studies, sibling studies, and large pedigree analyses have historically been used to investigate genetic and environmental contributions to phenotypic variation.
- However, interpreting human family resemblance requires particular caution because relatives often share environments as well as genes. Genetic and environmental effects can be separated more effectively when researchers use study designs that provide different patterns of genetic relatedness and environmental exposure.
- Covariance among relatives is also relevant to evolutionary studies. If relatives resemble one another for traits associated with fitness, and this resemblance is caused by heritable genetic variation, natural selection can produce evolutionary change. The amount of additive genetic variance available in a population influences its potential response to selection.
- The breeder’s equation, R = h²S, provides a simple connection between heritability and response to selection. Because narrow-sense heritability depends on additive genetic variance, information from relatives can help predict the expected response to selection when appropriate assumptions are satisfied.
- An important distinction is that covariance among relatives does not automatically prove that a trait is genetically inherited. Family members can resemble one another because of shared environments, cultural transmission, maternal effects, developmental conditions, or other non-genetic influences. Proper experimental and statistical designs are therefore essential.
- Maternal effects are particularly important in some species. Offspring may resemble their mothers because of genes, but also because mothers influence offspring through the prenatal environment, egg composition, milk, nutrition, behavior, or other developmental effects. These effects can contribute to covariance between relatives without representing direct additive genetic transmission.
- Environmental covariance can also occur among relatives raised in the same location. When family members experience similar environments, their phenotypes may covary even if the environment rather than genetics is responsible for the resemblance.
- Statistical models can account for these complexities by incorporating multiple sources of variation. Linear mixed models, pedigree-based models, genomic models, and multivariate models can separate genetic effects from environmental and other random effects when sufficient data and appropriate experimental designs are available.
- The strength of covariance among relatives also depends on the trait being studied. Traits with substantial additive genetic variance may show stronger predictable resemblance among relatives, while traits strongly influenced by environmental conditions may show weaker or more inconsistent familial resemblance.
- Covariance among relatives is therefore a bridge between observed family resemblance and the underlying genetic architecture of quantitative traits. It provides a statistical foundation for estimating additive genetic variance, heritability, breeding values, and genetic correlations.
- Overall, covariance among relatives is one of the foundational tools of quantitative genetics. By comparing the phenotypic similarity of individuals with known genetic relationships, researchers can investigate how genetic and environmental factors contribute to complex traits. Combined with pedigrees, relationship matrices, genomic data, and statistical models such as BLUP, covariance among relatives remains essential for modern genetic analysis, breeding, and evolutionary research.
