- Selection for multiple traits is an approach to animal breeding in which animals are evaluated and selected simultaneously for two or more characteristics that contribute to the overall breeding objective. Unlike single-trait selection, which focuses primarily on improving one trait, multiple-trait selection recognizes that animal performance is determined by many biological characteristics and that improvement in one trait can influence other traits through genetic correlations, environmental relationships, and correlated responses to selection. Multiple-trait selection is therefore central to modern animal breeding programs seeking balanced improvement in production, reproduction, health, survival, welfare, adaptation, and economic performance.
- The fundamental purpose of multiple-trait selection is to achieve genetic improvement across a defined breeding objective rather than maximizing one measurement in isolation. For example, a dairy breeding program may seek improvement in milk production, milk composition, fertility, udder health, longevity, feed efficiency, and disease resistance. A beef breeding program may simultaneously consider growth rate, carcass quality, feed efficiency, calving ease, fertility, maternal ability, and survival. Poultry breeding programs may combine egg production, egg quality, growth, feed efficiency, fertility, hatchability, health, and welfare. The exact combination of traits depends on the production system, economic objectives, biological constraints, and long-term sustainability goals.
- The genetic foundation of multiple-trait selection is the same quantitative genetic framework used for individual traits. An observed phenotype can be represented as P = G + E, where P is the phenotype, G is the genetic component, and E represents environmental effects. The genetic component includes additive genetic effects, dominance effects, and epistatic effects. For selection decisions, additive genetic variation is particularly important because it determines the component of genetic merit that can be transmitted predictably from parents to offspring.
- When several traits are evaluated simultaneously, breeders must consider not only the genetic variance of each trait but also the genetic covariance among traits. Genetic covariance describes whether genetic differences for two traits tend to occur together. If animals with higher genetic merit for one trait also tend to have higher genetic merit for another, the traits have a positive genetic association. If improvement in one trait tends to be accompanied by deterioration in another, the traits may have an unfavorable genetic relationship.
- The standardized form of genetic covariance is the genetic correlation, expressed as r_A = Cov_A(X,Y) / (σ_A,X × σ_A,Y). Genetic correlations are extremely important in multiple-trait selection because they determine how selection for one trait is expected to influence other traits. A favorable genetic correlation can allow simultaneous improvement, whereas an unfavorable genetic correlation may create a trade-off that requires deliberate balancing.
- For example, suppose growth rate and mature body size are positively genetically correlated. Selecting strongly for rapid growth may also increase mature size. If larger mature size increases maintenance requirements or creates reproductive or welfare concerns, simply maximizing growth may not be desirable. Similarly, production and fertility may sometimes show antagonistic genetic relationships. A breeding program that selects only for production may therefore experience an undesirable correlated response in fertility if the breeding objective does not account for both traits.
- This is one of the major reasons why single-trait selection can be inadequate for long-term breeding. A trait may show substantial genetic improvement while another economically or biologically important trait deteriorates. Multiple-trait selection attempts to account for these relationships before selection decisions are made.
- The first step in multiple-trait selection is defining the breeding objective. The breeding objective describes the aggregate genetic improvement desired in the population. It should include traits that affect profitability, biological efficiency, health, fertility, welfare, adaptation, longevity, and other long-term goals relevant to the production system. The objective should distinguish between traits that are directly measured and the underlying genetic characteristics that breeders ultimately want to improve.
- Not every economically or biologically important trait can be measured directly. Some traits are expensive, difficult, late-expressed, sex-limited, or affected strongly by environmental conditions. For this reason, breeding programs often use indicator traits that provide information about difficult-to-measure traits. For example, an easily recorded measurement may provide information about feed efficiency, health, fertility, or longevity. Multiple-trait genetic evaluation can combine these sources of information to improve the prediction of breeding value.
- Heritability differs among traits and strongly influences their response to selection. Narrow-sense heritability is expressed as h² = σ²_A / σ²_P, where σ²_A is additive genetic variance and σ²_P is phenotypic variance. A highly heritable trait may respond relatively quickly to selection based on individual performance, while a low-heritability trait may require information from relatives, repeated records, progeny, or genomic data. Multiple-trait evaluation can improve the prediction of low-heritability traits when genetically correlated traits provide additional information.
- This is an important advantage of multiple-trait genetic evaluation. Suppose fertility has low heritability but is genetically correlated with another trait that is measured accurately and has higher heritability. Information from the correlated trait can contribute to the estimation of genetic merit for fertility. This is known as indirect selection or correlated information and can be particularly valuable when direct measurement of the target trait is difficult.
- The expected response to selection for a single trait can be represented simply as R = h² × S, where R is selection response and S is selection differential. With multiple traits, however, selection response becomes more complex because selection applied to one trait can produce correlated changes in other traits. The total response therefore depends on the genetic variance-covariance structure, selection criteria, selection intensity, and accuracy of genetic evaluation.
- A useful way to manage multiple traits is through a selection index. A selection index combines information from several traits or information sources into a single numerical value that can be used to rank animals. The index can include estimated breeding values, phenotypic records, family information, progeny information, genomic information, and economic or biological weights.
- Conceptually, a selection index can be written as I = b₁x₁ + b₂x₂ + … + bₙxₙ, where I is the selection index, x values represent available information, and b values are the index weights. The weights are determined using the genetic and phenotypic relationships among the information sources and the desired breeding objective. The exact statistical formulation depends on the breeding program and genetic evaluation system.
- The selection index approach is particularly powerful because it allows breeders to balance traits with different units. Milk yield may be measured in kilograms, fertility as a percentage, somatic cell count on a transformed scale, longevity in years, and disease resistance using a binary or threshold-scale measurement. These traits cannot simply be added together in their raw units. The selection index provides a statistically appropriate way to combine information.
- Economic weights are often used when the breeding objective is primarily economic. An economic weight represents the relative economic value of changing a trait while holding other traits constant. However, breeding objectives should not necessarily be based only on immediate financial return. Health, welfare, environmental impact, resilience, adaptation, and genetic diversity may also need to be incorporated because they influence long-term sustainability.
- Economic weights can change over time. Feed prices, market demands, disease risks, environmental regulations, consumer expectations, climate conditions, and production systems can all change. A breeding objective that was appropriate several decades ago may not be optimal for future production. Multiple-trait breeding programs should therefore periodically review and update their objectives.
- Genetic correlations are not always favorable or unfavorable in an absolute sense. Their importance depends on the breeding objective and production environment. A positive genetic correlation means that genetic improvement in one trait tends to be associated with improvement in another, but this may not always be desirable. For example, increasing mature body size may be positively correlated with growth but may also increase maintenance requirements. The breeder must therefore evaluate the overall biological and economic consequences.
- Multiple-trait selection is also important when traits have antagonistic relationships. If production and fertility have an unfavorable genetic correlation, selecting exclusively for production may reduce fertility. Including fertility directly in the breeding objective creates selection pressure in both directions and can moderate the correlated decline. The same principle applies to production and health, growth and calving ease, muscling and reproductive performance, or other traits with potential trade-offs.
- The relationship between selection criteria and breeding objectives is important. A selection criterion is a measurement or source of information used to make selection decisions, while the breeding objective describes the genetic changes that breeders ultimately want. The two may not be identical. A breeder may use an indicator trait because it is easier or cheaper to measure than the target trait.
- Modern genetic evaluation makes multiple-trait selection much more powerful. BLUP and animal models can simultaneously analyze records from several traits while accounting for pedigree relationships, contemporary groups, environmental effects, maternal effects, repeated records, and other sources of variation. The resulting estimated breeding values (EBVs) can be combined across traits to create a multi-trait selection index.
- The pedigree relationship matrix, commonly represented by the A matrix, allows information from relatives to contribute to genetic evaluation. An animal with limited own records may still receive useful information from parents, full-sibs, half-sibs, offspring, and other relatives. This is particularly valuable for traits that are difficult to measure directly.
- Genomic selection adds another source of information. Genomic data can be used to estimate genomic estimated breeding values (GEBVs) for multiple traits, sometimes before animals have sufficient own performance or progeny records. This can increase selection accuracy and shorten the generation interval. Genomic selection is particularly useful when breeding programs have large reference populations with reliable phenotypes and genotypes.
- Multiple-trait genomic selection can also help improve traits that are difficult or expensive to record. A correlated, well-recorded trait may provide information about a less frequently measured target trait. However, the usefulness of such indirect information depends on the strength and stability of the genetic relationship between the traits and the relevance of the reference population.
- Progeny testing remains valuable for multiple-trait selection because offspring performance provides information about parental genetic merit. This is particularly important for sex-limited traits, low-heritability traits, late-expressed traits, and traits that cannot be measured reliably on the candidate itself. Progeny records can contribute to multi-trait BLUP evaluations and genomic reference populations.
- Family selection and within-family selection can also contribute to multiple-trait decisions. Relatives provide additional information about genetic merit, particularly when individual phenotypes are strongly affected by environmental variation. Combined selection can integrate individual, family, progeny, pedigree, and genomic information into a comprehensive evaluation.
- Multiple-trait selection is especially important for reproductive traits. Fertility, age at sexual maturity, conception rate, litter size, calving or lambing performance, semen quality, maternal ability, reproductive longevity, and offspring survival may all influence overall reproductive efficiency. Selecting for production without considering reproductive performance can create undesirable genetic consequences when production and fertility are genetically antagonistic.
- Health traits are equally important. Disease resistance, disease susceptibility, immune function, disease tolerance, survival, and resilience can be incorporated into multi-trait breeding objectives. These traits may have low heritability or require specialized phenotyping, making family, progeny, and genomic information especially useful.
- Disease resistance and disease tolerance should not be treated as identical traits. Resistance refers to the ability to prevent or limit infection or pathogen burden, whereas tolerance describes the ability to maintain performance despite infection or disease. A breeding objective may need to distinguish these biological mechanisms depending on the production system.
- Multiple-trait selection also has a major role in animal welfare. Traits associated with health, locomotion, temperament, injury risk, reproductive problems, survival, and behavioural responses may be incorporated into breeding objectives. Selecting for welfare-related traits alongside production traits can help avoid genetic improvement strategies that maximize output while compromising animal well-being.
- Adaptation traits are increasingly important in breeding objectives. Heat tolerance, disease resilience, feed efficiency, water-use efficiency, survival under challenging conditions, and performance under variable climates can influence the long-term suitability of animals. Genetic improvement should therefore consider the environment in which animals will be expected to perform.
- This becomes particularly important because of genotype–environment interaction (G×E). Genetic rankings may change across environments. An animal that performs exceptionally well in a high-input production system may not be the best animal under heat stress, low-input management, disease challenge, or restricted feed availability. Multiple-trait selection can incorporate adaptation traits and environmental information into the breeding objective.
- The choice of traits should also consider genetic antagonisms. Selecting simultaneously for too many traits can dilute selection intensity across individual traits. If a breeding program includes dozens of traits without appropriate prioritization, the expected response for important traits may become small. Multiple-trait selection is therefore not simply a matter of adding more traits. It requires careful prioritization and appropriate weighting.
- Selection intensity influences the amount of genetic progress that can be achieved. If a small proportion of animals are selected, selection intensity is high. If many animals are retained, selection intensity is lower. In a multiple-trait program, selection intensity must be balanced against the need to retain animals with acceptable performance across all important traits.
- The generation interval also influences the rate of genetic improvement. The approximate rate of genetic improvement per year can be represented as ΔG/year = i × r × σ_A / L, where i is selection intensity, r is selection accuracy, σ_A is additive genetic standard deviation, and L is generation interval. Genomic selection can increase accuracy at younger ages and reduce generation interval, potentially accelerating improvement across several traits simultaneously.
- However, faster genetic gain is not always the only objective. High selection intensity can increase genetic concentration if a small number of animals contribute disproportionately to future generations. This can reduce effective population size and increase inbreeding. Multiple-trait selection must therefore be combined with responsible reproductive management.
- The approximate relationship ΔF ≈ 1 / (2Ne) illustrates the relationship between the rate of inbreeding and effective population size (Ne). When genetic contributions become concentrated in a small number of animals, effective population size can decline even if the census population remains large. This is why breeding programs should monitor genetic diversity while pursuing multi-trait genetic improvement.
- Popular sire effects are particularly relevant. A sire with excellent performance across several traits may be selected extensively, increasing the representation of his genes in the population. Although this can accelerate genetic improvement, excessive reproductive contribution can increase relatedness and inbreeding and reduce the number of independent genetic lineages contributing to future generations.
- Mate allocation provides a way to manage this problem. Instead of selecting animals solely on their multi-trait genetic merit, breeders can consider the relatedness between potential mates. If φ(sire, dam) represents the coefficient of coancestry between the parents, the expected inbreeding of their offspring can be expressed as E(F_offspring) = φ(sire, dam). When relationship is defined as twice coancestry, the same concept can be expressed as E(F_offspring) = r(sire, dam) / 2.
- Optimal contribution selection goes further by controlling the proportion of genes contributed by individual animals to the next generation. It seeks to balance genetic gain against the rate of increase in relatedness or inbreeding. This is particularly valuable in multi-trait breeding programs because animals with high overall breeding values can be used strategically without allowing genetic contribution to become excessively concentrated.
- Genomic information can strengthen this management. Genomic relatedness can provide information about realized genetic similarity among animals, while runs of homozygosity (ROH) can provide information about segments of the genome that are identical by descent. The proportion of the autosomal genome contained in ROH can be expressed as F_ROH = Total length of ROH / Total autosomal genome length.
- These genomic measures can complement pedigree-based estimates of relatedness and inbreeding. This is particularly useful in populations where pedigrees are incomplete, parentage is uncertain, or actual genomic relationships differ substantially from pedigree expectations.
- Multiple-trait selection must also consider genetic load and deleterious variants. Selecting animals for superior performance across multiple traits does not automatically guarantee that they are free from harmful recessive alleles. Genetic testing can identify known deleterious variants and help breeders avoid high-risk matings. The goal should be to manage harmful alleles while maintaining useful genetic diversity rather than unnecessarily eliminating large portions of the breeding population.
- The concept of inbreeding depression is also relevant. Increasing homozygosity can expose harmful recessive variants and may reduce fertility, survival, growth, disease resistance, and other fitness-related traits. A breeding objective that includes only production traits may fail to detect these longer-term consequences. Including fertility, health, survival, and welfare traits in multiple-trait selection can help create a more balanced genetic response.
- Multiple-trait selection can also be used in conservation breeding. In small populations, breeders may need to maintain genetic diversity while improving health, fertility, adaptation, or other traits. Selection decisions can incorporate genetic merit and relatedness simultaneously. This is especially important when the available breeding population is limited.
- A major practical challenge is data quality. Multi-trait genetic evaluation requires reliable records for all important traits. Measurement errors, missing data, inconsistent definitions, changing recording systems, and environmental confounding can reduce selection accuracy. Standardized phenotyping and accurate animal identification are therefore fundamental components of successful multiple-trait breeding.
- The scale of measurement should also be considered. Some traits are easy and inexpensive to record, whereas others require specialized equipment, laboratory analysis, disease challenges, slaughter measurements, or long-term monitoring. The cost of phenotyping should be compared with the expected genetic and economic value of improving the trait.
- Threshold traits require particular attention. Traits such as conception success, disease status, calving difficulty, survival, or certain genetic disorders may be recorded as categories even though the underlying genetic liability is continuous. Threshold models can be used to analyze such traits within multi-trait genetic evaluations.
- Multiple-trait selection can also incorporate repeatability when traits are recorded repeatedly over an animal’s lifetime. Repeated milk yield, egg production, behavioural measurements, health events, or other records can provide additional information about permanent differences among animals. This can improve the accuracy of genetic evaluation when appropriately modeled.
- Maternal effects and common environmental effects should also be considered. For traits expressed early in life, offspring performance may reflect both direct genetic effects and maternal genetic effects. Animals raised together may also share common environmental influences. Failure to account for these effects can lead to inaccurate estimates of genetic merit and genetic correlations.
- The success of multiple-trait selection therefore depends on a combination of biological knowledge, statistical methodology, accurate phenotyping, and appropriate breeding objectives. It is not enough to rank animals according to a simple average of trait measurements. The genetic relationships among traits, relative economic or biological importance, measurement accuracy, environmental effects, and long-term consequences must all be considered.
- A well-designed multiple-trait breeding program should begin by defining the long-term breeding objective. The next step is to identify the traits or indicator traits that provide information about that objective. Appropriate phenotypic, pedigree, family, progeny, and genomic data are then collected. Genetic parameters, including heritabilities and genetic correlations, are estimated or updated. Genetic evaluation models are used to calculate EBVs or GEBVs, and a selection index or another decision framework is used to rank candidates.
- Selection should then be combined with appropriate mating management. High-ranking animals are not necessarily all mated to one another. Relatedness, genetic diversity, reproductive contribution, genetic defects, and population structure should be considered when designing matings. This allows breeders to achieve genetic progress without unnecessarily increasing inbreeding.
- The breeding objective should also be monitored over time. Genetic trends should be evaluated for each important trait, not only for the overall selection index. If one trait improves rapidly while another deteriorates, the index weights or breeding strategy may need to be adjusted. Genetic correlations may also change as populations, environments, and selection pressures change.
- Multiple-trait selection is therefore fundamentally about balance. The goal is not to maximize every trait simultaneously because biological and genetic constraints make that impossible in many cases. Instead, the objective is to achieve the most desirable overall combination of traits while managing trade-offs and maintaining long-term population health.
- In modern animal breeding, multiple-trait selection represents a major transition from simple phenotype-based selection toward integrated genetic improvement. Individual selection, family information, progeny testing, BLUP, genomic selection, genetic correlations, and selection indexes can all contribute to the process. The breeder is no longer simply asking which animal has the highest value for one trait, but which animals are most valuable for the entire breeding objective.
- Ultimately, selection for multiple traits provides a framework for improving animals in a biologically and economically balanced way. By considering production, growth, fertility, health, disease resistance, survival, welfare, feed efficiency, adaptation, and other relevant characteristics together, breeders can reduce the risk of undesirable correlated responses and create more resilient populations. When combined with accurate genetic evaluation, genomic information, responsible mating strategies, and management of genetic diversity, multiple-trait selection can support sustained genetic gain while protecting the health, welfare, adaptability, and long-term sustainability of animal populations.
