- Realized heritability is an empirical measure of heritability estimated from the observed response to selection in a population. Unlike conventional estimates of heritability that are often obtained from variance components, relatives, pedigrees, or genomic relationships, realized heritability is calculated from the actual change in a trait produced by selection across generations. It therefore provides a practical way to assess how much of the observed response to selection is associated with heritable genetic differences.
- Realized heritability is closely related to the breeder’s equation, which expresses the expected response to selection as R=h2SR = h^2S, where RR is the response to selection, h2h^2 is narrow-sense heritability, and SS is the selection differential. Rearranging this relationship gives the basic expression for realized heritability: hR2=R/Sh^2_R = R/S. Here, hR2h^2_R represents realized heritability estimated from the observed response and selection differential. This approach connects the theory of quantitative genetics with measurable changes in a real breeding or experimental population.
- The selection differential is the difference between the mean phenotype of the individuals selected as parents and the mean phenotype of the population from which they were selected. The response to selection is the difference between the mean phenotype of the offspring generation and the mean phenotype of the original population. When both quantities are measured accurately, their ratio provides an estimate of the realized heritability of the trait under the specific selection experiment and environmental conditions.
- For example, suppose a population has a mean trait value of 50 units and individuals selected as parents have a mean of 60 units. The selection differential is therefore S=10S = 10. If the offspring generation has a mean of 55 units, the response to selection is R=5R = 5. The realized heritability would then be hR2=5/10=0.5h^2_R = 5/10 = 0.5. This means that approximately half of the selection differential was expressed as a response in the next generation under the conditions of the experiment. It does not mean that 50% of an individual’s phenotype is genetic.
- Realized heritability is particularly useful because it measures the outcome of actual selection rather than relying solely on a statistical decomposition of phenotypic variance. A traditional narrow-sense heritability estimate is based on the ratio of additive genetic variance to phenotypic variance, h2=VA/VPh^2 = V_A/V_P. Realized heritability instead uses the relationship between observed response and selection differential. Although the two measures are conceptually related, they are not necessarily identical because realized heritability can be affected by changes in genetic variance, environmental conditions, selection procedures, population structure, and other processes occurring during the experiment.
- The distinction between realized heritability and conventional heritability is important. A variance-component estimate describes the expected contribution of additive genetic variation to phenotypic differences under a particular model and set of conditions. Realized heritability describes the observed effectiveness of selection in producing a response. The latter therefore incorporates the combined consequences of the genetic and experimental conditions that actually occurred during the selection process.
- Realized heritability is often estimated using selection experiments conducted over multiple generations. Researchers measure the trait in the initial population, select individuals according to their phenotypes, breed the selected individuals, and measure the trait in their descendants. Repeating the process over several generations provides a series of selection differentials and responses. The cumulative response can be compared with cumulative selection differential to estimate realized heritability.
- A common graphical approach is to plot cumulative response to selection against cumulative selection differential. Under suitable assumptions, the slope of the regression provides an estimate of realized heritability. If the relationship is approximately linear, the slope represents the average response per unit of cumulative selection differential. This method can be especially useful when selection is applied over multiple generations and individual generation-to-generation estimates are affected by sampling variation.
- Realized heritability is strongly influenced by additive genetic variance because additive genetic effects are transmitted between generations and therefore contribute directly to the response to selection. If a population contains substantial additive genetic variation for a trait, sustained selection can produce a measurable response. If additive genetic variance becomes depleted, the response may decline even when selection continues.
- This depletion of additive genetic variance is one reason why realized heritability may change during a long-term selection experiment. Strong directional selection can alter allele frequencies and reduce genetic variation for the selected trait. As favorable alleles become more common, the remaining population may have less additive genetic variance available for further response. Consequently, the relationship between selection differential and response may not remain constant across generations.
- Other genetic processes can also affect realized heritability. Mutation can introduce new genetic variation, while genetic drift can randomly change allele frequencies, particularly in small populations. Inbreeding can alter genetic variation and increase homozygosity. Genetic correlation can cause selection on one trait to produce correlated changes in another trait. These processes can influence the observed response and therefore the realized heritability calculated from the experiment.
- Genotype–environment interaction (G×E) can also influence realized heritability. Genotypes may respond differently to environmental conditions, so the same population may show different selection responses under different temperatures, diets, locations, management systems, or developmental conditions. Consequently, realized heritability is specific to the population and environmental context in which the selection experiment is conducted.
- Environmental effects can also complicate the interpretation of realized heritability. If environmental conditions change systematically between generations, part of the observed change in the population mean may be caused by environmental differences rather than genetic evolution. For this reason, well-designed selection experiments often use control populations, replicated lines, standardized environments, or appropriate experimental designs to distinguish genetic response from environmental change.
- A control population is particularly valuable in long-term selection experiments. A control line is maintained without directional selection and provides an estimate of changes that occur because of environmental trends, drift, or other non-selection processes. Comparing the selected population with the control population can help determine how much of the observed phenotypic change is associated with the selection treatment.
- Realized heritability can also be affected by maternal effects, paternal effects, shared environmental effects, and other sources of resemblance between generations. For example, if selected mothers provide different prenatal or postnatal environments to their offspring, offspring may resemble their parents even when the resemblance is not entirely caused by inherited additive genetic differences. Such effects can lead to an overestimate of the genetic response if they are not properly controlled.
- The relationship between realized heritability and breeding value is also important. Breeding value represents the additive genetic contribution expected to be transmitted from an individual to its offspring. Selection based on phenotype is effective when phenotypic differences provide useful information about breeding values. Realized heritability provides an empirical indication of how effectively phenotypic selection produced a response in the next generation.
- Realized heritability is widely used in animal breeding, plant breeding, experimental evolution, and laboratory selection studies. In agriculture, breeders may use long-term selection records to determine how strongly traits such as growth rate, yield, body composition, flowering time, or disease resistance respond to selection. In experimental evolution, realized heritability can help demonstrate whether a population contains sufficient heritable variation for a trait to evolve under artificial or natural selection.
- The measure is also closely connected to genetic gain. Genetic gain describes the improvement in a breeding population achieved through selection. A trait with substantial realized heritability can show a strong response when sufficient selection differential is applied. However, genetic gain also depends on factors such as selection intensity, generation interval, population size, genetic variance, accuracy of selection, and genetic correlations with other traits.
- Realized heritability should not be confused with response to selection itself. Response is the observed change in the population mean caused by selection, whereas realized heritability is the response standardized by the selection differential. Two populations may show different responses because they experienced different selection differentials, while their realized heritabilities may be similar. Conversely, similar selection differentials can produce different responses if their underlying genetic variation differs.
- Realized heritability also differs from broad-sense heritability. Broad-sense heritability is H2=VG/VPH^2 = V_G/V_P and includes total genetic variance, including additive, dominance, and epistatic components. Realized heritability is primarily interpreted in relation to the additive genetic response predicted by the breeder’s equation, although empirical estimates can be influenced by other biological and environmental processes. It should therefore not be treated as a direct substitute for every form of heritability estimation.
- The value of realized heritability can theoretically range from zero to one under the simplest interpretation of the breeder’s equation, but empirical estimates can sometimes fall outside this range because of sampling error, measurement error, changing environments, nonrandom selection, statistical estimation, or violations of model assumptions. A negative estimate may occur when the observed response is opposite to the direction expected from the selection differential, while an estimate above one can arise from experimental or statistical complications rather than implying that a biological heritability greater than one exists.
- Long-term selection experiments can reveal changes in realized heritability over time. A declining response may indicate depletion of additive genetic variance, changes in allele frequencies, increasing environmental variance, correlated responses, or other evolutionary processes. Conversely, sustained response may indicate that sufficient genetic variation remains available or that new variation has been introduced through mutation, recombination, migration, or other mechanisms.
- Modern genomic approaches can complement realized heritability estimates. Genomic relationship matrices, genome-wide markers, QTL mapping, and GWAS can help identify genetic variants and genomic regions associated with the trait being selected. Comparing phenotypic response with genomic changes can provide additional evidence about the genetic basis of selection response. Genomic selection can also use genome-wide information to predict breeding values and improve selection decisions.
- Realized heritability is particularly informative when combined with other quantitative-genetic measures. Genetic variance describes the amount of variation associated with genetic differences, additive genetic variance describes the component most directly relevant to predictable inheritance, heritability describes the relationship between genetic and phenotypic variance, and realized heritability measures the empirical relationship between selection and observed response. Together, these concepts provide a more complete understanding of how populations respond to selection.
- In evolutionary studies, realized heritability provides experimental evidence that phenotypic variation can produce evolutionary change when it is associated with heritable genetic differences. However, natural populations are influenced by many additional processes, including mutation, migration, genetic drift, natural selection, genetic correlations, changing environments, and demographic structure. Realized heritability therefore represents one component of the broader framework of evolutionary quantitative genetics.
- Overall, realized heritability is an empirical measure that connects selection intensity with the actual response observed across generations. It is commonly estimated as the ratio of response to selection to selection differential, hR2=R/Sh^2_R = R/S, or from the slope of cumulative response on cumulative selection differential. Its interpretation depends on genetic variation, environmental conditions, population structure, selection design, and the assumptions of the underlying quantitative-genetic model.
- Realized heritability is especially valuable because it demonstrates the practical consequences of additive genetic variation and provides a direct link between heritability, selection response, breeding value, and genetic gain. When interpreted alongside conventional heritability estimates, genetic variance components, covariance among relatives, and genomic information, it provides a powerful framework for understanding and predicting evolutionary and breeding responses to selection.
