- Animal models are among the most important statistical tools used in modern animal breeding and quantitative genetics. They help researchers estimate the genetic merit of individual animals by combining performance records, pedigree information, and environmental factors. These models are widely used to improve livestock traits such as milk production, growth rate, fertility, feed efficiency, disease resistance, meat quality, and longevity. By separating genetic differences from environmental influences, animal models allow breeders to make more accurate selection decisions and achieve long-term genetic improvement in livestock populations.
- An animal model is a statistical model in which the genetic effect of each animal is represented directly in the analysis. Instead of evaluating an animal only through its own performance or through the average performance of its relatives, the model uses information from the entire connected population when suitable data are available. This includes the animal’s own records, records of its parents and offspring, and information from other genetically related animals. The model accounts for the fact that relatives share genes and that their observations can provide useful information about one another’s genetic merit.
- Animal models are commonly used within the framework of mixed linear models. A basic animal model can be written as y=Xb+Za+e, where y represents observed performance records, X is the design matrix for fixed effects, b represents fixed-effect solutions, Z is the design matrix linking records to animals, a represents the animals’ additive genetic effects, and e represents residual errors. Fixed effects may include herd, year, season, sex, age, and management group. The additive genetic effects represent the inherited component of variation that is passed from parents to offspring, while residual errors account for unexplained variation.
- The additive genetic effect of an animal is particularly important because it forms the basis of its estimated breeding value. A breeding value predicts the animal’s additive genetic contribution to the next generation for a particular trait. For example, a dairy bull may have a breeding value for milk yield, while a beef bull may have breeding values for growth rate, carcass quality, and feed efficiency. Animals with favorable breeding values can be selected as parents to increase the frequency of desirable genetic characteristics in future generations.
- A major advantage of animal models is their ability to use pedigree relationships. Pedigree records identify the parents and ancestors of animals and help establish the expected genetic relationships among individuals. Full siblings, for example, share both parents, while half siblings share one parent. These relationships provide information about how much genetic material animals are expected to share. The animal model uses this information to combine observations from related individuals and improve the estimation of genetic merit.
- The numerator relationship matrix is a central component of a traditional pedigree-based animal model. It represents the expected additive genetic relationships among animals based on their recorded ancestry. The diagonal elements describe each animal’s expected additive genetic variance relative to the base population, while off-diagonal elements describe the expected additive genetic relationships between different animals. This matrix allows the statistical model to account for the fact that the genetic effects of relatives are correlated rather than independent.
- In practice, this means that an animal can receive a useful genetic evaluation even when it has limited performance information of its own. A young bull, for instance, may not yet have offspring with recorded production data, but the records of its parents, siblings, and other relatives can help predict its genetic merit. Similarly, a young female may be evaluated using its own early performance and information from relatives. As additional records become available, the breeding value can be recalculated, allowing the genetic evaluation to become more informative over time.
- Animal models are closely associated with Best Linear Unbiased Prediction, commonly known as BLUP. BLUP is a statistical method used to predict random effects, including the additive genetic effects of animals. In an animal breeding application, BLUP combines performance records and pedigree relationships while accounting for fixed environmental effects. It also accommodates animals with unequal numbers of records and different amounts of information. This makes it particularly suitable for large livestock populations in which animals are raised in different herds and management systems.
- One of the most important applications of animal models is the estimation of breeding values for economically important traits. In dairy cattle, animal models are used to evaluate milk yield, fat percentage, protein percentage, fertility, udder health, and longevity. In beef cattle, they can be applied to growth rate, birth weight, weaning weight, carcass characteristics, and maternal ability. In sheep and goats, they help evaluate wool production, milk production, reproductive performance, growth, and resistance to parasites. In pigs and poultry, animal models can be used to improve feed efficiency, growth, fertility, egg production, and other traits relevant to production systems.
- Environmental adjustment is another essential feature of animal models. The performance of livestock is influenced by feed quality, housing, disease exposure, weather, management, and other conditions. If these influences are ignored, animals raised in favorable environments may appear genetically superior even when their inherited merit is similar to that of animals raised under less favorable conditions. By including appropriate fixed effects, such as herd, year, season, and management group, the model helps account for systematic environmental differences. This improves the fairness of genetic comparisons, particularly when animals are evaluated across multiple herds or production environments.
- For example, consider a group of beef calves raised on several farms. Some farms provide high-quality feed and effective health management, while others operate under more challenging conditions. Calves on the better-managed farms may have higher weaning weights because of environmental advantages. An animal model can account for relevant management and environmental effects while using pedigree relationships to estimate the calves’ additive genetic merit. When the data contain sufficient genetic connections between farms, the resulting breeding values can support more meaningful comparisons across herds.
- Animal models are also useful for traits measured repeatedly throughout an animal’s life. Dairy cows may have milk production recorded over several lactations, while growing animals may be weighed at multiple ages. Repeated records from the same animal are not usually independent because they share genetic and environmental influences. Models can include permanent environmental effects to account for lasting non-genetic influences on an animal’s repeated performance. More advanced animal models may also incorporate random regression techniques to describe genetic and environmental changes over time, such as changes in milk yield throughout lactation.
- Another important application is the estimation of genetic parameters, including additive genetic variance, heritability, and genetic correlations. Additive genetic variance measures the amount of phenotypic variation associated with additive genetic differences in a population. Heritability describes the proportion of phenotypic variance attributable to additive genetic variance under a particular population and set of environmental conditions. Genetic correlations describe how genetic effects on two traits are associated. These parameters help researchers understand the inheritance of traits and predict how selection for one trait may affect another.
- For instance, a breeding program may aim to increase growth rate while maintaining fertility and disease resistance. Animal models can be used to estimate genetic relationships among these traits. If faster growth is genetically associated with an unfavorable change in another trait, breeders can account for that relationship when designing selection strategies. This is important because selecting animals for a single trait may produce unintended consequences for other economically or biologically important characteristics. Multiple-trait animal models can use information from several traits simultaneously and improve evaluations when traits are genetically related.
- Animal models can be extended to analyze more than one trait at a time. In a single-trait animal model, the genetic evaluation focuses on one characteristic. In a multiple-trait animal model, two or more traits are analyzed together, allowing the model to account for genetic and residual correlations among them. For example, milk yield, fat percentage, and protein percentage may be evaluated jointly. Information from one trait can help improve the genetic evaluation of another when the traits are genetically correlated and the data provide useful information. Multiple-trait models are particularly valuable when some traits are recorded less frequently or are more difficult and expensive to measure.
- The distinction between direct and maternal genetic effects is also important in animal models. Some traits, such as birth weight and weaning weight, may be influenced by the animal’s own genes and by the genetic characteristics of its mother. Maternal genetic effects can influence the environment a dam provides to her offspring, including milk production, maternal behavior, and other aspects of rearing. A suitable model can separate direct additive genetic effects from maternal genetic effects and may also include maternal permanent environmental effects. This helps breeders understand the different genetic pathways affecting early-life performance.
- The use of genomic information has expanded the capabilities of animal models. Traditional animal models rely mainly on pedigree relationships, but genomic data provide information about genetic similarity based on DNA markers. Genomic relationships can differ from expected pedigree relationships because animals inherit different combinations of genetic material from their parents. Incorporating genomic information can improve breeding value prediction, particularly for young animals that have not yet produced their own offspring or accumulated extensive performance records.
- Genomic animal models are often implemented through methods such as genomic BLUP and single-step genomic BLUP. Genomic BLUP uses a genomic relationship matrix derived from marker information to represent genetic relationships among genotyped animals. Single-step genomic BLUP combines pedigree, genomic, and phenotypic information in a unified evaluation framework, allowing genotyped and non-genotyped animals to be analyzed together. These approaches are widely relevant to modern genomic selection programs, in which animals can be selected at a younger age and the generation interval may be reduced.
- Animal models are also used to investigate genotype-by-environment interaction. This occurs when genetic differences between animals vary across environments. For example, animals that perform well under intensive feeding conditions may not perform equally well in environments with limited feed or high temperatures. Researchers can use suitable animal models to investigate whether the same genetic rankings apply across production systems or whether animals need to be selected for specific environments. This is increasingly important as livestock producers face climate variability, changing disease pressures, and differences in available resources.
- Despite their advantages, animal models require reliable data and appropriate statistical assumptions. Incorrect pedigree information, incomplete records, inconsistent measurement methods, and unrecorded environmental differences can reduce the quality of genetic evaluations. The choice of fixed effects, random effects, covariance structures, and variance components must reflect the biological characteristics of the trait and the structure of the population. If important environmental effects are omitted, genetic estimates may be biased. If the model is unnecessarily complex, it may also become difficult to estimate or interpret.
- It is important to recognize that animal models do not automatically identify the specific genes responsible for a trait. They estimate genetic effects from observed records and the relationships specified in the statistical model. The quality of these estimates depends on the amount and quality of available information, the accuracy of the pedigree or genomic relationships, and the suitability of the model. Breeding values should therefore be interpreted alongside their reliability, the population in which they were estimated, and the objectives of the breeding program.
- Modern animal breeding software enables researchers to analyze large datasets using animal models and related statistical methods. These tools can estimate variance components, predict breeding values, analyze multiple traits, and incorporate genomic relationships. However, reliable genetic evaluation still requires careful data management, model selection, validation, and interpretation. Researchers must ensure that records are accurate and that the model reflects the relevant biological and environmental processes. Breeders must also consider the economic importance of traits, animal welfare, genetic diversity, and long-term sustainability when using the resulting predictions.
- Animal models have become a foundation of modern quantitative genetics because they combine information from individual performance, relatives, and environmental conditions in a consistent statistical framework. They improve the estimation of breeding values, help researchers understand genetic variation, and support the design of effective livestock selection programs. Their integration with BLUP, multiple-trait analysis, repeated measurements, and genomic information has made genetic evaluation more comprehensive and useful across livestock species. As breeding goals increasingly include productivity, fertility, health, adaptability, and sustainability, animal models will continue to play a central role in the genetic improvement of animal populations.
