- Artificial selection is the deliberate choice of individuals to become parents of the next generation based on desirable characteristics. It is one of the fundamental principles of animal breeding and has been used for centuries to improve production, reproduction, health, behaviour, adaptation, and other economically and biologically important traits. Unlike natural selection, which results from differences in survival and reproductive success under environmental conditions, artificial selection is directed by breeders according to defined breeding objectives.
- The effectiveness of artificial selection depends on the presence of genetic variation, the ability to identify animals with superior genetic potential, the accuracy of selection decisions, the intensity and timing of selection, and the management of genetic diversity. The central principle is that animals should be selected primarily according to their genetic merit, rather than simply their observed phenotype. A superior phenotype may result from both genetic and environmental factors, so successful breeding programs attempt to separate genetic effects from environmental effects.
- A useful starting point is the relationship between phenotype and its underlying causes: P = G + E
- where P is the observed phenotype, G is the genetic component, and E represents environmental effects. In practice, the genetic component can include additive genetic effects, dominance effects, and epistatic effects. For most long-term selection programs, additive genetic variation is particularly important because additive effects are transmitted predictably from parents to offspring and form the basis of the breeding value.
- The first principle of artificial selection is therefore the existence of heritable genetic variation. If individuals differ genetically for a trait, and some of that genetic variation can be passed to their offspring, selection can change the population mean over generations. Without genetic variation, selection cannot produce a sustained genetic response. Genetic variation may involve differences in growth, milk production, meat quality, egg production, fertility, disease resistance, temperament, heat tolerance, feed efficiency, survival, or other traits.
- The second principle is that selection should target the breeding objective rather than isolated phenotypic performance. A breeding objective defines the combination of traits that should improve in the population. For example, a dairy breeding program may consider milk yield, milk composition, fertility, udder health, longevity, disease resistance, feed efficiency, and welfare simultaneously. Selecting only for maximum production could unintentionally reduce fertility, health, longevity, or other important traits if these traits are genetically unfavourably correlated.
- The third principle is the distinction between phenotypic value and breeding value. Phenotypic value describes what can be observed or measured in an individual, whereas breeding value represents the individual’s expected additive genetic contribution to its offspring. Two animals with similar phenotypes may therefore have different breeding values because they differ in family background, genetic relationships, environmental circumstances, or other sources of information.
- The breeding value is central to modern artificial selection because the goal is not simply to identify animals that perform well, but to identify animals that are genetically capable of transmitting desirable characteristics to future generations. Estimated breeding values (EBVs) provide statistical predictions of genetic merit using information from the animal itself, relatives, progeny, and other relevant sources.
- A fourth principle is the use of selection criteria that provide reliable information about genetic merit. Selection can be based on an individual’s own performance, family information, progeny performance, or genomic information. Individual selection uses the phenotype of the candidate animal. This can be effective when the trait has moderate or high heritability and can be measured accurately before reproduction.
- Family selection uses information from relatives to improve selection accuracy, particularly when the trait is difficult, expensive, sex-limited, or impossible to measure directly in the candidate. Information from full-sibs, half-sibs, parents, or other relatives can contribute to the evaluation of genetic merit.
- Progeny testing evaluates an individual’s genetic merit using the performance of its offspring. It can be especially valuable for traits that are difficult to measure directly in breeding candidates. However, progeny testing may require many offspring, substantial time, and considerable resources, which increases the generation interval.
- Modern breeding programs increasingly use genomic selection, in which genome-wide DNA marker information is used to predict genetic merit. Genomic information can increase the accuracy of selection, particularly for young animals that have not yet produced offspring. This can allow breeders to reduce the generation interval while maintaining or increasing selection accuracy.
- Another important principle is selection differential. The selection differential is the difference between the mean phenotype of the selected parents and the mean phenotype of the population from which they were selected. It describes how strongly selection has favoured a particular trait in one generation.
- The expected response to selection is commonly expressed as: R = h² × S
- where R is the response to selection, h² is narrow-sense heritability, and S is the selection differential. This relationship illustrates why both genetic variation and selection intensity matter. A large selection differential does not necessarily produce a large response if the trait has little additive genetic variation.
- The principle of heritability is therefore central to artificial selection. Narrow-sense heritability describes the proportion of phenotypic variance attributable to additive genetic variance under a particular population and environment: h² = σ²_A / σ²_P
- where σ²_A is additive genetic variance and σ²_P is phenotypic variance.
- High heritability generally means that phenotypic differences provide more information about additive genetic differences within the population and environment being studied. However, low heritability does not mean that a trait cannot respond to selection. Accurate information from relatives, repeated measurements, progeny, genomic data, and appropriate statistical models can improve selection decisions for traits with relatively low heritability.
- The principle of selection intensity concerns how strongly animals are selected. If only a small proportion of animals are chosen as parents, selection intensity is high. If a large proportion is retained, selection intensity is lower. Stronger selection can increase short-term genetic gain, but excessive selection can also reduce genetic diversity and increase the risk of inbreeding or genetic concentration.
- The principle of selection accuracy is equally important. Selection accuracy describes how closely the selection criterion reflects the animal’s true genetic merit. Accuracy can be improved by combining multiple sources of information, including individual performance, pedigree relationships, repeated records, progeny data, genomic information, and appropriate statistical evaluation.
- Modern BLUP (Best Linear Unbiased Prediction) methods and animal models allow breeders to combine information from large numbers of animals while accounting for environmental effects and genetic relationships. These methods can produce estimated breeding values that are more informative than simple phenotypic rankings.
- Artificial selection also depends on the principle of generation interval. The generation interval is the average age of parents when their selected offspring are born. Shortening the generation interval can increase the annual rate of genetic improvement, provided selection accuracy and other factors remain adequate.
- A simplified expression for annual genetic improvement is: ΔG/year = i × r × σ_A / L
- where i is selection intensity, r is selection accuracy, σ_A is the additive genetic standard deviation, and L is the generation interval. This relationship shows why breeding programs must consider more than simply selecting the most extreme animals. Increasing accuracy and reducing generation interval can be as important as increasing selection intensity.
- Another fundamental principle is multi-trait selection. Most breeding programs have several objectives rather than a single trait. Selection for one trait can cause correlated changes in other traits because traits may share genetic causes.
- The relationship between two traits can be described using the genetic correlation. A favourable genetic correlation can allow improvement in one trait to contribute to improvement in another, whereas an unfavourable correlation can create a trade-off. For example, strong selection for production may have undesirable consequences for fertility, health, longevity, or welfare if the traits are genetically unfavourably associated.
- For this reason, selection index methods are widely used in modern animal breeding. A selection index combines information from multiple traits and sources to rank animals according to an overall breeding objective. This allows breeders to balance production, reproduction, health, welfare, adaptation, and economic value rather than focusing on one characteristic.
- The principle of balanced selection is particularly important for long-term breeding. A breeding program should not maximize a single trait at the expense of the overall biological and economic performance of animals. Excessive emphasis on one production characteristic can produce undesirable changes in fertility, disease resistance, survival, behaviour, structural soundness, or welfare.
- Artificial selection must also account for environmental effects. An animal may perform exceptionally well because it received superior nutrition, housing, management, or health care rather than because it possesses superior genetics. Statistical genetic evaluation attempts to account for systematic environmental differences so that animals can be compared more fairly.
- The interaction between genotype and environment is known as genotype–environment interaction (G×E). Genetic differences may be expressed differently under different environments. For example, animals selected for production under high-input conditions may not have the same ranking under heat stress, limited feed availability, disease challenge, or extensive management systems. Sustainable selection should therefore consider the environments in which animals are expected to perform.
- Another principle is repeatability, particularly for traits measured repeatedly during an animal’s lifetime. Repeated records can provide additional information about an individual’s permanent genetic and environmental effects. This can improve selection decisions for traits such as milk production, egg production, growth, behavioural consistency, or repeated health measurements.
- For traits that are expressed as categories or binary outcomes, such as disease occurrence, pregnancy success, survival, or certain behavioural classifications, breeders may use threshold models or other appropriate statistical approaches. Treating every trait as a simple continuous measurement can produce inappropriate genetic evaluations when the underlying biological process is categorical.
- Artificial selection also operates through differential reproductive contribution. Animals selected as parents contribute genes to the next generation, while animals that are not selected contribute fewer or no genes. Therefore, selection changes allele frequencies over generations. The magnitude and direction of these changes depend on the intensity of selection, genetic architecture, population size, reproductive structure, and mating system.
- A major principle of sustainable artificial selection is maintaining genetic diversity while achieving genetic improvement. Strong selection can increase the representation of particular families or genetic lines. When a small number of animals contribute disproportionately to the next generation, genetic concentration can increase.
- The popular sire effect is a common example. An exceptionally successful male may be used extensively through natural mating, artificial insemination, or other reproductive technologies. Although this can spread desirable genetics rapidly, excessive use can increase relatedness among future animals, reduce the effective population size (Ne), increase the rate of inbreeding, and increase the probability that deleterious recessive variants become homozygous.
- A simplified relationship between effective population size and the rate of increase in inbreeding is: ΔF ≈ 1 / (2Ne)
- This illustrates an important principle: the genetic consequences of selection depend not only on which animals are selected, but also on how their genes are distributed throughout the population.
- For this reason, artificial selection should be combined with mate allocation, mean kinship, and optimal contribution selection when appropriate. Mate allocation can help avoid matings that produce excessive expected inbreeding, while optimal contribution methods attempt to balance genetic gain against the preservation of genetic diversity.
- The expected inbreeding coefficient of offspring can be related to parental coancestry: E(F_offspring) = φ(sire, dam)
- or, when relatedness is expressed on the conventional relationship scale: E(F_offspring) = r(sire, dam) / 2
- These principles are particularly important in small populations, closed breeding populations, endangered populations, and highly selected commercial populations where the number of effective breeding animals may be much smaller than the census population.
- Another principle is the management of genetic load. Populations contain genetic variants that may reduce survival, fertility, health, or performance under particular circumstances. Artificial selection can reduce some harmful variants when they are identifiable and subject to selection, but intense selection and population concentration can also unintentionally increase the frequency of undesirable variants.
- Genetic testing can help identify known deleterious variants, while genomic information can provide broader information about genetic relationships, homozygosity, and population structure. However, genetic testing should be integrated with overall breeding objectives rather than used as the sole basis for selection.
- Runs of homozygosity (ROH) provide another genomic indicator of autozygosity. Long ROH can indicate recent common ancestry and may help identify genomic inbreeding. A common measure is: F_ROH = Total length of ROH / Total autosomal genome length
- This information can complement pedigree-based estimates of inbreeding and help breeders monitor genetic diversity more precisely.
- The principle of selection across generations is also important. Artificial selection produces cumulative genetic change. A breeding program should therefore be evaluated over multiple generations rather than judged solely by short-term performance. Short-term improvement can sometimes hide long-term reductions in diversity, fertility, health, adaptation, or resilience.
- The principle of sustainable genetic improvement requires breeders to consider the entire life cycle and biological performance of animals. Traits such as survival, fertility, disease resistance, longevity, welfare, temperament, heat tolerance, and adaptation may become increasingly important when production environments change.
- Animal welfare should therefore be incorporated into breeding objectives. Genetic selection can influence behaviour, structural traits, disease resistance, reproductive performance, and susceptibility to health problems. However, genetics is only one component of welfare. Nutrition, housing, management, veterinary care, handling, stocking density, environmental conditions, and other factors remain essential.
- The principle of economic weighting is also relevant in commercial breeding. Different traits have different economic consequences. A selection index can assign appropriate economic weights to traits while considering their genetic relationships and expected long-term effects. The objective should be to improve overall profitability and sustainability rather than maximize one biological measurement.
- Artificial selection can also interact with crossbreeding. Within-breed selection improves additive genetic merit, whereas crossbreeding can exploit heterosis, or hybrid vigour, which may improve traits such as fertility, survival, disease resistance, and growth in appropriate mating systems. Therefore, selection and mating strategies should be considered together.
- Another principle is the distinction between selection and mating. Selection determines which animals contribute genetically to the next generation, while mating determines which selected animals are paired. A population can therefore have excellent selection decisions but poor mating management. Combining accurate selection with appropriate mating plans can produce better long-term outcomes.
- The principle of continuous evaluation is essential because breeding objectives should not remain fixed indefinitely. Consumer expectations, production systems, disease challenges, climate conditions, welfare standards, market values, and available technology can change. Breeding objectives should therefore be reviewed periodically.
- Modern artificial selection increasingly integrates phenotypic records, pedigree information, genomic data, environmental information, statistical genetic evaluation, and reproductive technologies. The objective is not simply to select the highest-ranking animals, but to make accurate decisions that produce cumulative genetic improvement while controlling unintended consequences.
- The most effective breeding programs therefore follow several interconnected principles: maintain sufficient genetic variation, define clear breeding objectives, measure relevant traits accurately, estimate breeding values, maximize appropriate selection accuracy, apply suitable selection intensity, manage the generation interval, account for genetic correlations, consider genotype–environment interaction, maintain genetic diversity, control inbreeding, and use appropriate mating strategies.
- Ultimately, artificial selection is a long-term process of changing the genetic composition of a population in a desired direction. Its success should be measured not only by immediate increases in production, but also by improvements in reproduction, health, survival, welfare, adaptation, efficiency, and overall population fitness. Sustainable artificial selection combines genetic gain with responsible management of inbreeding, genetic diversity, and population structure.
- The central principle is therefore not simply to select the best animals, but to select the animals that are most likely to contribute to the best future population. This requires combining genetic information, accurate evaluation, balanced breeding objectives, appropriate mating strategies, and long-term population management.
