- Threshold Traits are traits that appear in distinct categories or states at the phenotypic level but are influenced by an underlying continuous liability that is determined by genetic and environmental factors. They are an important concept in quantitative genetics because they connect apparently discrete traits with the continuous variation of complex genetic and environmental influences.
- A classic example is a disease that is recorded as either affected or unaffected. Although the observed phenotype has only two categories, individuals can differ continuously in their underlying genetic and environmental susceptibility. Individuals whose liability exceeds a particular threshold express the condition, while individuals below the threshold do not. The observed categorical phenotype therefore results from an underlying continuous distribution.
- The threshold model provides a framework for understanding traits that appear qualitative but have a complex quantitative genetic basis. Instead of assuming that a binary phenotype is controlled by a single gene, the model proposes that many genetic and environmental factors contribute to an underlying liability. The phenotype is then determined when this liability crosses a threshold.
- The underlying liability can be conceptualized as a continuous variable influenced by genetic and environmental components. In a simplified model, liability can be represented as L = G + E, where G represents genetic liability and E represents environmental effects. A threshold determines whether the observable phenotype falls into one category or another. More complex models can include additive genetic effects, dominance, epistasis, maternal effects, permanent environmental effects, and other sources of variation.
- The key feature of a threshold trait is therefore the distinction between the underlying liability and the observed phenotype. Liability varies continuously among individuals, but the phenotype is recorded categorically. This explains why a trait can show a binary or discrete phenotype while still behaving as a quantitative genetic trait.
- Threshold traits can have two categories, such as affected and unaffected, or several ordered categories. Binary threshold traits include disease susceptibility, survival, reproductive success, and certain developmental abnormalities. Ordinal threshold traits may include disease severity scores, fertility categories, behavioral classifications, or other traits recorded using ordered classes.
- The threshold model is closely related to polygenic inheritance. Many threshold traits are influenced by multiple genes, with each locus contributing a small or moderate effect to the underlying liability. Environmental conditions can also influence liability. As a result, individuals with similar genotypes may differ in phenotype because of environmental variation, while genetically different individuals may display the same observed category.
- A threshold trait can therefore be understood as a special type of quantitative trait. The underlying liability is continuous, but the observed phenotype is discontinuous. This provides a bridge between qualitative traits and quantitative traits and helps explain why some apparently Mendelian categories can actually arise from complex genetic architecture.
- The location of the threshold is important. If the threshold is low, many individuals may exceed it and express the trait. If the threshold is high, relatively few individuals may cross it. The frequency of the phenotype therefore depends on the distribution of liability and the position of the threshold.
- For a binary trait, the observed proportion of affected individuals can be described as the incidence or prevalence in the population, depending on the biological context. The proportion of individuals exceeding the threshold is determined by the distribution of underlying liability. Changes in genetic or environmental conditions can shift this distribution and consequently change the frequency of the observed phenotype.
- Heritability of threshold traits requires special consideration because the observed phenotype is categorical rather than continuous. Heritability can be estimated on an underlying liability scale, where the continuous genetic variance of liability is modeled. This is often called liability-scale heritability. The observed-scale proportion of phenotypic variation is not directly equivalent to the heritability of the underlying liability.
- For example, suppose a disease has an affected/unaffected phenotype. The observed phenotype provides only two values, but genetic differences may influence a continuous susceptibility score. Individuals near the threshold may be especially sensitive to environmental influences, while individuals far below or above the threshold may be more resistant or susceptible. Quantitative genetic analysis can model the latent liability rather than treating the binary phenotype as if it were a normally distributed quantitative measurement.
- Threshold traits are often analyzed using threshold models or liability-threshold models. These statistical approaches assume an unobserved continuous liability underlying the categorical phenotype. The observed category is determined by whether the liability crosses one or more thresholds.
- For a binary trait, there is usually one threshold separating the two categories. For an ordinal trait with several categories, multiple thresholds can divide the liability scale into different phenotypic classes. This allows researchers to model ordered categories while retaining the underlying continuous interpretation.
- Threshold models are closely related to generalized linear mixed models, particularly models for binary or ordinal outcomes. In animal breeding and quantitative genetics, threshold models can incorporate pedigree relationships, genomic information, maternal effects, common environmental effects, and other random effects.
- Covariance among relatives provides important information for estimating the genetic basis of threshold traits. If relatives are more likely to share the same phenotype than expected from their environmental circumstances alone, this can provide evidence for genetic variation in liability. However, shared environments and maternal effects must be considered because relatives may resemble each other for non-genetic reasons.
- The genetic contribution to liability can be modeled using an additive genetic variance component. The resulting breeding value represents the individual’s expected additive genetic contribution to liability. This is different from simply predicting the individual’s observed categorical phenotype because the breeding value describes genetic propensity rather than the final threshold outcome.
- This distinction is particularly important in animal breeding. Many economically important traits are threshold traits, including disease resistance, calving difficulty, fertility, survival, reproductive success, and susceptibility to particular disorders. An animal may have a genetic tendency toward greater resistance even if its observed phenotype is simply recorded as healthy or diseased.
- Threshold models can therefore be used to estimate breeding values for categorical traits. These genetic evaluations can help identify individuals with lower genetic liability to disease or higher genetic probability of successful reproduction. When genomic information is incorporated, genomic selection can further improve prediction of genetic liability.
- Threshold traits are also important in human genetics. Many diseases and disorders have complex genetic architectures in which numerous variants contribute to susceptibility together with environmental and lifestyle factors. Although the clinical phenotype may be recorded as affected or unaffected, underlying liability can vary continuously among individuals.
- Examples can include susceptibility to multifactorial diseases, developmental disorders, autoimmune conditions, and other complex phenotypes. The threshold framework does not imply that every disease follows a single universal threshold mechanism; rather, it provides a useful quantitative model for situations in which a continuous liability produces a categorical outcome.
- Threshold traits are also relevant to natural selection. If crossing a threshold affects survival or reproduction, selection can act on the underlying liability even though the phenotype is observed categorically. Individuals with genetic variants that reduce liability below the threshold may have greater fitness under particular environmental conditions.
- Environmental effects are particularly important for threshold traits because the same underlying genetic liability can result in different phenotypes under different conditions. Nutrition, temperature, pathogens, stress, management, developmental conditions, and other environmental factors can shift an individual’s liability and determine whether the threshold is crossed.
- This relationship makes threshold traits closely connected to genotype–environment interaction (G×E). Different genotypes may respond differently to environmental conditions, causing their liabilities to change at different rates. As a result, the probability of crossing a threshold can vary substantially across environments.
- Threshold traits can also involve maternal effects. Maternal nutrition, prenatal conditions, maternal antibodies, maternal behavior, and other maternal influences may alter offspring liability. If these effects are not accounted for, genetic differences in susceptibility can be confused with differences caused by maternal environments.
- Common environmental effects can also influence threshold traits. Relatives or individuals sharing the same environment may have similar risks because they experience the same exposure, nutrition, housing, management, or disease pressure. This shared environmental covariance can contribute to phenotypic similarity independently of genetic relatedness.
- Similarly, permanent environmental effects may alter an individual’s liability over a long period. An early-life disease or persistent developmental condition could increase susceptibility to a later phenotype. Such effects need to be distinguished from inherited genetic liability when estimating genetic parameters.
- Threshold traits can also be associated with maternal genetic effects. A mother’s genotype may influence the environment provided to her offspring and thereby affect the offspring’s liability. This creates a distinction between the offspring’s direct genetic liability and the maternal genetic contribution to that liability.
- The relationship between threshold traits and phenotypic variance differs from that of continuously measured traits. On the observed scale, the phenotype may have only two or a few categories, so standard variance calculations do not fully represent the continuous underlying biological variation. The liability scale provides a more informative framework for genetic analysis.
- The frequency of an observed category can nevertheless change substantially when the underlying liability distribution changes. For example, selection against disease susceptibility can shift the genetic liability distribution toward lower risk, reducing the proportion of individuals crossing the disease threshold in subsequent generations.
- The expected response to selection depends on the genetic variation in liability, the accuracy of genetic evaluation, and the selection strategy. Because the observed phenotype is categorical, predicting response may require models that operate on the underlying liability scale rather than directly applying continuous-trait formulas to the observed categories.
- Threshold traits are therefore closely connected to the breeder’s equation, but application requires care. The traditional breeder’s equation, R = h²S, assumes a continuously measured trait under a particular quantitative-genetic framework. For threshold traits, selection response is often modeled through the underlying liability and the genetic parameters associated with that latent scale.
- Threshold models can also incorporate genetic correlation between multiple threshold or continuous traits. For example, susceptibility to two diseases may share genetic influences, or a disease liability may be genetically correlated with growth, fertility, or production traits. Understanding these relationships is important for avoiding unfavorable correlated responses during selection.
- The genetic covariance between traits can therefore influence selection decisions involving threshold phenotypes. A breeding program selecting strongly for disease resistance could unintentionally alter another trait if the underlying genetic liabilities are correlated. Multivariate genetic models can account for these relationships.
- Threshold traits are also relevant to genetic architecture. A categorical phenotype does not necessarily imply a simple genetic architecture. A threshold outcome may result from numerous loci, small genetic effects, dominance, epistasis, environmental effects, and interactions among these factors. The discrete phenotype therefore should not be mistaken for evidence of single-gene inheritance.
- Modern genomic methods can investigate the genetic basis of threshold traits using GWAS, whole-genome sequencing, and genomic prediction. GWAS can identify genetic variants associated with differences in disease susceptibility or other categorical outcomes. However, individual variants usually explain only part of the variation in complex threshold phenotypes.
- Polygenic scores can sometimes be used to summarize the combined contribution of many genetic variants to disease liability or another threshold phenotype. Their predictive performance depends on the genetic architecture, population, environment, phenotype definition, and population used to develop the score.
- Threshold traits also illustrate the importance of distinguishing phenotype from genotype. An affected individual may have high genetic liability, an unfavorable environment, or a combination of both. Conversely, an individual with substantial genetic susceptibility may remain unaffected if environmental conditions do not push its liability beyond the threshold.
- This makes threshold traits especially useful for understanding the interaction between genes and environments. The observed phenotype represents the final outcome of underlying genetic liability, environmental exposure, and the threshold mechanism. The same genetic background can therefore produce different observed outcomes under different environmental circumstances.
- Threshold traits should also be distinguished from genuinely qualitative traits controlled by a small number of genes with discrete Mendelian inheritance. Some categorical traits are indeed controlled primarily by one or a few genes, while others reflect an underlying continuous liability. Determining which model is appropriate requires biological knowledge and statistical evidence.
- The threshold concept is also useful for understanding why disease susceptibility can appear to run in families. Family clustering may reflect inherited genetic liability, shared environmental exposures, maternal effects, or combinations of these factors. Quantitative genetic and epidemiological models can help separate these sources.
- In conservation genetics, threshold traits can be relevant to survival, reproductive success, disease susceptibility, and environmental tolerance. Small populations may experience changes in genetic variation, inbreeding, and environmental conditions that alter the distribution of liability and the probability that individuals cross fitness-related thresholds.
- Threshold traits can also be important under climate change. Environmental stressors may shift liability distributions or move individuals closer to thresholds associated with mortality, reproductive failure, disease susceptibility, or developmental abnormalities. Genetic variation in environmental tolerance can influence the capacity of populations to respond to changing conditions.
- Overall, threshold traits provide a powerful framework for understanding categorical phenotypes that arise from continuous underlying genetic and environmental variation. They demonstrate that a trait can appear discrete at the phenotypic level while being controlled by complex polygenic and environmental processes. By modeling the underlying liability, researchers can estimate genetic variance, heritability, breeding values, genetic correlations, and expected responses to selection more appropriately.
- Understanding threshold traits is therefore essential in quantitative genetics, animal and plant breeding, human genetics, epidemiology, evolutionary biology, and conservation. The threshold model connects qualitative traits, quantitative traits, polygenic inheritance, environmental variation, heritability, and genetic architecture, providing a unified way to study complex traits that are observed as distinct phenotypic categories.
