Effect of gametic disequilibrium on selection in an autotetraploid population

Summary The effects of a gametic disequilibrium (DSE) in an autotetraploid population on response to selection as measured by the covariance of selection were investigated. The theoretical responses were calculated for mass selection [Mass (1)] and half-sib progeny test selection (HSPT) in a two-allele (B and b), single locus, autotetraploid population. The complexity of calculations precluded analytical expressions for the covariances so numerical analysis was used assuming the following genetic models: monoplex dominance, partial monoplex dominance, duplex dominance, partial duplex dominance, and additive gene action.The results indicated the DSE could greatly affect the covariance of selection. For a constant allele frequency the DSE might double the covariance expected with selection in a population at random mating equilibrium (RME) of gametes, but in other instances approach zero. For all genetic models and the two breeding methods investigated the covariance of selection was always increased when the frequency of BB gamete exceeded p² (where p is frequency of allele B) and decreased when the frequency of BB gamete was less than p². The possible incorporation of this information into a long term breeding program and some other ramifications were briefly discussed.With the DSE the covariances of selection with HSPT and Mass (1) had a proportionality of 1:2, respectively, with the additive genetic model, but this relationship rarely occurred for other genetic models. The deviations from this ratio were not large in comparison to differences between selection in populations in DSE and RME.Cooperative investigations of the Alfalfa Production Research Unit, United State Department of Agriculture, Agricultural Research Service, and the Nevada Agricultural Experiment Station, Reno, Nevada. Paper No. 512. Scientific Journal Series, Nevada Agricultural Experiment Station     

Selection, load and inbreeding depression in a large metapopulation

The subdivision of a species into local populations causes its response to selection to change, even if selection is uniform across space. Population structure increases the frequency of homozygotes and therefore makes selection on homozygous effects more effective. However, population subdivision can increase the probability of competition among relatives, which may reduce the efficacy of selection. As a result, the response to selection can be either increased or decreased in a subdivided population relative to an undivided one, depending on the dominance coefficient F(ST) and whether selection is hard or soft. Realistic levels of population structure tend to reduce the mean frequency of deleterious alleles. The mutation load tends to be decreased in a subdivided population for recessive alleles, as does the expected inbreeding depression. The magnitude of the effects of population subdivision tends to be greatest in species with hard selection rather than soft selection. Population structure can play an important role in determining the mean fitness of populations at equilibrium between mutation and selection.     

Effect of gametic disequilibrium on means and on genetic variances of autotetraploid synthetic varieties

Summary The occurrence and effects of a gametic disequilibrium (DSE) in the first generation of a theoretical two-population synthetic variety were investigated. Theoretical development was limited to the genetics at a single locus with two alleles in an autotetraploid species with random chromosome inheritance. Algebraic expressions were developed for the differences between the mean genotypic values of the two-population synthetic variety at generation one and in random mating equilibrium (RME). For the situation where both parents of the synthetic were in RME, a numerical analysis was performed for all possible allele frequencies assuming the following types of genic action: monoplex dominance, partial monoplex dominance, duplex dominance, partial duplex dominance, and additive. The result indicated that with non-additive genic action the DSE could, in some cases, greatly depress or inflate the mean genotypic value of the first generation (Syn-1_(RME)). Thus, any change of means over advancing generations with loss of DSE could be positive or negative. When additive genic action was assumed, there was no effect associated with DSE and when both parents had the same allele frequencies there was no DSE. The DSE, with only a minor exception, decreased the genetic variance and in numerous cases forced it near zero. Expressions were developed for mean genotypic values of a first generation synthetic with DSE in one parent (Syn-1_(DSE/RME)) or both parents (Syn-1_(DSE)). The deviation of these means from those of Syn-1_(RME) was a function of digenic and quadragenic population effects. An inspection of the response equations for Syn-1_(RME) indicated that in a series of crosses with one common parent the rankings of first generation means would be the same as the ranking of populations at equilibrium though the individual means would be biased. More importantly with DSE of one or both parents there are situations when a ranking of first generation mean genotypic values would not reflect relative frequency of desirable alleles in the populations. These results indicate that statistical analyses and selections based on means of the Syn-1 generation can have an error which is not avoidable by improvement in precision of evaluation.Cooperative investigations of the Alfalfa Production Research Unit, United States Department of Agriculture, Agricultural Research Service, and the Nevada Agricultural Experiment Station, Reno, Nevada, USAPaper No. 590 Scientific Journal Series, Nevada Agricultural Experiment Station     

PREY ADAPTATION AS A CAUSE OF PREDATOR-PREY CYCLES

We analyze simple models of predator-prey systems in which there is adaptive change in a trait of the prey that determines the rate at which it is captured by searching predators. Two models of adaptive change are explored: (1) change within a single reproducing prey population that has genetic variation for vulnerability to capture by the predator; and (2) direct competition between two independently reproducing prey populations that differ in their vulnerability. When an individual predator's consumption increases at a decreasing rate with prey availability, prey adaptation via either of these mechanisms may produce sustained cycles in both species' population densities and in the prey's mean trait value. Sufficiently rapid adaptive change (e.g., behavioral adaptation or evolution of traits with a large additive genetic variance), or sufficiently low predator birth and death rates will produce sustained cycles or chaos, even when the predator-prey dynamics with fixed prey capture rates would have been stable. Adaptive dynamics can also stabilize a system that would exhibit limit cycles if traits were fixed at their equilibrium values. When evolution fails to stabilize inherently unstable population interactions, selection decreases the prey's escape ability, which further destabilizes population dynamics. When the predator has a linear functional response, evolution of prey vulnerability always promotes stability. The relevance of these results to observed predator-prey cycles is discussed.     

Infection before pregnancy affects immunity and response to social challenge in the next generation

Natural selection should favour parents that are able to adjust their offspring's life-history strategy and resource allocation in response to changing environmental and social conditions. Pathogens impose particularly strong and variable selective pressure on host life histories, and parental genes will benefit if offspring are appropriately primed to meet the immunological challenges ahead. Here, we investigated transgenerational immune priming by examining reproductive resource allocation by female mice in response to direct infection with Babesia microti prior to pregnancy. Female mice previously infected with B. microti gained more weight over pregnancy, and spent more time nursing their offspring. These offspring generated an accelerated response to B. microti as adults, clearing the infection sooner and losing less weight as a result of infection. They also showed an altered hormonal response to novel social environments, decreasing instead of increasing testosterone production upon social housing. These results suggest that a dominance-resistance trade-off can be mediated by cues from the previous generation. We suggest that strategic maternal investment in response to an infection leads to increased disease resistance in the following generation. Offspring from previously infected mothers downregulate investment in acquisition of social dominance, which in natural systems would reduce access to mating opportunities. In doing so, however, they avoid the reduced disease resistance associated with increased testosterone and dominance. The benefits of accelerated clearance of infection and reduced weight loss during infection may outweigh costs associated with reduced social dominance in an environment where the risk of disease is high.     

An exact form of the breeder's equation for the evolution of a quantitative trait under natural selection

Starting with the Price equation, I show that the total evolutionary change in mean phenotype that occurs in the presence of fitness variation can be partitioned exactly into five components representing logically distinct processes. One component is the linear response to selection, as represented by the breeder's equation of quantitative genetics, but with heritability defined as the linear regression coefficient of mean offspring phenotype on parent phenotype. The other components are identified as constitutive transmission bias, two types of induced transmission bias, and a spurious response to selection caused by a covariance between parental fitness and offspring phenotype that cannot be predicted from parental phenotypes. The partitioning can be accomplished in two ways, one with heritability measured before (in the absence of) selection, and the other with heritability measured after (in the presence of) selection. Measuring heritability after selection, though unconventional, yields a representation for the linear response to selection that is most consistent with Darwinian evolution by natural selection because the response to selection is determined by the reproductive features of the selected group, not of the parent population as a whole. The analysis of an explicitly Mendelian model shows that the relative contributions of the five terms to the total evolutionary change depends on the level of organization (gene, individual, or mated pair) at which the parent population is divided into phenotypes, with each frame of reference providing unique insight. It is shown that all five components of phenotypic evolution will generally have nonzero values as a result of various combinations of the normal features of Mendelian populations, including biparental sex, allelic dominance, inbreeding, epistasis, linkage disequilibrium, and environmental covariances between traits. Additive genetic variance can be a poor predictor of the adaptive response to selection in these models. The narrow-sense heritability sigma2A/sigma2P should be viewed as an approximation to the offspring-parent linear regression rather than the other way around.     

Effects of random and non-random errors on phenotypic selection in autotetraploids

Summary A theoretical investigation was made to ascertain the effects of random and non-random deviations, called errors, of phenotypic from genotypic values on population means and on the response to phenotypic recurrent selection. The study was motivated as a selection experiment for disease resistance where there was either variability in the inoculation or environment (the random errors) or where the inoculation was above or below the the optimum rate where genetic differences in resistance are maximized (the non-random errors). The study was limited to the genetics at a diallelic locus (alleles B and b) in an autotetraploid population in random mating equilibrium. The response to selection was measured as the covariance of selection and compared to the exact covariance which was the covariance of selection without errors in phenotype. The random errors were modeled by assuming that a given percentage () of the population was uniformly distributed among the five possible genotype classes independent of their true genotypes. This model was analyzed numerically for a theoretical population with the frequency of the B allele (p) ranging from 0.0 to 1.0 and assumed errors of=0.1 and 0.5 for the following six types of genic action of the B allele: additive, monoplex dominance, partial monoplex dominance, duplex dominance, partial duplex dominance, and recessive. The effect of random error was to consistently reduce the response to selection by a percentage independent of the type of genic action at the locus. The effect on the population mean was an upward bias when p was low and a downward bias when p approached unity. In the non-random error model below optimum inoculations altered the phenotypes by systematically including percentage of susceptible genotypes into one or more other genotype classes with more genetic resistance (a positive shift). With above optimum inoculations, some resistant genotypes are classed with the non-resistant genotypes (a negative shift). The effects on the covariance of selection were found by numerical analysis for the same types of genic action and's as investigated for random error. With a negative shift and a low p, the covariance of selection was always reduced, but for an increasing p the covariance approached and exceeded the exact covariance for all types of genic action except additive. With a positive shift and a low p, response to selection was greatly improved for three types of genic action: duplex dominance, partial duplex dominance, and recessive. The effect of a non-random error on population means was to greatly bias the means upwards for a low p and positive shift, but with increasing p the bias decreased. A relatively slight decrease in the mean occurred with a negative shift. This study indicated check varieties commonly used to monitor selection pressures in screening programs are very responsive to positive non-random shifts, but are relatively unresponsive to negative shifts. The interaction of selection pressure, types of genic action, and genotypes in the class shift models was suggested as a partial explanation for the lack of response to increasing selection pressures observed in some breeding programs.Cooperative investigations of the Alfalfa Production Research Unit, United States Department of Agriculture, Agricultural Research Service, and the Nevada Agricultural Experiment Station, Reno, Nevada. Paper No. 404 Scientific Journal Series. Nevada Agricultural Experiment Station     

Evolutionary dynamics of escape from biomedical intervention

Viruses, bacteria, eukaryotic parasites, cancer cells, agricultural pests and other inconvenient animates have an unfortunate tendency to escape from selection pressures that are meant to control them. Chemotherapy, anti-viral drugs or antibiotics fail because their targets do not hold still, but evolve resistance. A major problem in developing vaccines is that microbes evolve and escape from immune responses. The fundamental question is the following: if a genetically diverse population of replicating organisms is challenged with a selection pressure that has the potential to eradicate it, what is the probability that this population will produce escape mutants? Here, we use multi-type branching processes to describe the accumulation of mutants in independent lineages. We calculate escape dynamics for arbitrary mutation networks and fitness landscapes. Our theory shows how to estimate the probability of success or failure of biomedical intervention, such as drug treatment and vaccination, against rapidly evolving organisms.     

ANTIBODY SELECTION AND AMINO ACID REVERSIONS

Pathogens adapt to antibody surveillance through amino acid replacements in targeted protein regions, or epitopes, that interfere with antibody binding. However, such escape mutations may exact a fitness cost due to impaired protein function. Here, it is hypothesized that the recurring generation of specific neutralizing antibodies to an epitope region as it evolves in response to antibody selection will cause amino acid reversions by releasing early escape mutations from immune selection. The plausibility of this hypothesis was tested with stochastic simulation of adaptation at the molecular sequence level in finite populations. Under the conditions of strong selection and weak mutation, the rates of allele fixation and amino acid reversion increased with population size and selection coefficients. These rates decreased with population size, however, if mutation became strong, because clonal interference reduced the rate of adaptation. The model successfully predicts the rate of reversion per allele fixation for an important human immunodeficiency virus type 1 (HIV‐1) antibody epitope region. Therefore, antibody selection may generate complex adaptive dynamics.     

EVOLUTION AND EXTINCTION IN A CHANGING ENVIRONMENT: A QUANTITATIVE-GENETIC ANALYSIS

Because of the ubiquity of genetic variation for quantitative traits, virtually all populations have some capacity to respond evolutionarily to selective challenges. However, natural selection imposes demographic costs on a population, and if these costs are sufficiently large, the likelihood of extinction will be high. We consider how the mean time to extinction depends on selective pressures (rate and stochasticity of environmental change, and strength of selection), population parameters (carrying capacity, and reproductive capacity), and genetics (rate of polygenic mutation). We assume that in a randomly mating, finite population subject to density-dependent population growth, individual fitness is determined by a single quantitative-genetic character under Gaussian stabilizing selection with the optimum phenotype exhibiting directional change, or random fluctuations, or both. The quantitative trait is determined by a finite number of freely recombining, mutationally equivalent, additive loci. The dynamics of evolution and extinction are investigated, assuming that the population is initially under mutation-selection-drift balance. Under this model, in a directionally changing environment, the mean phenotype lags behind the optimum, but on the average evolves parallel to it. The magnitude of the lag determines the vulnerability to extinction. In finite populations, stochastic variation in the genetic variance can be quite pronounced, and bottlenecks in the genetic variance temporarily can impair the population's adaptive capacity enough to cause extinction when it would otherwise be unlikely in an effectively infinite population. We find that maximum sustainable rates of evolution or, equivalently, critical rates of environmental change, may be considerably less than 10% of a phenotypic standard deviation per generation.     

Selection, subdivision and extinction and recolonization

In a subdivided population, the interaction between natural selection and stochastic change in allele frequency is affected by the occurrence of local extinction and subsequent recolonization. The relative importance of selection can be diminished by this additional source of stochastic change in allele frequency. Results are presented for subdivided populations with extinction and recolonization where there is more than one founding allele after extinction, where these may tend to come from the same source deme, where the number of founding alleles is variable or the founders make unequal contributions, and where there is dominance for fitness or local frequency dependence. The behavior of a selected allele in a subdivided population is in all these situations approximately the same as that of an allele with different selection parameters in an unstructured population with a different size. The magnitude of the quantity N(e)s(e), which determines fixation probability in the case of genic selection, is always decreased by extinction and recolonization, so that deleterious alleles are more likely to fix and advantageous alleles less likely to do so. The importance of dominance or frequency dependence is also altered by extinction and recolonization. Computer simulations confirm that the theoretical predictions of both fixation probabilities and mean times to fixation are good approximations.     

Theoretical Modifications of Reciprocal Recurrent Selection

Reciprocal recurrent selection (RRS), which assumes overdominant loci to be important, alters two genetically different populations to improve their crossbred mean. Individual plants from two populations (A and B) are selfed and also crossed with plants from the reciprocal female tester population (B and A, respectively). Selection is based on the mean of crossbred families, and the selected individuals are randomly mated within A and B to form new populations.—We propose two alternatives to RRS. The first (RRS-I) uses, as the tester of population A, a population (LB) that is derived from population B by family selection for low yield. The second (RRS-II) is similar to RRS-I, but also uses, as the tester of B, a population (LA) that is derived from population A by family selection for low yield.—The expected crossbred means of RRS, RRS-I, and RRS-II were compared, assuming equal σP, at several cycles of selection for incomplete and complete dominance, and for several cases of overdominance (depending on the gene frequencies in A and B, and on the equilibrium gene frequency).—The choice of selection method depends on the importance of the effects of overdominant loci compared to loci exhibiting incomplete or complete dominance. If overdominance is unimportant, RRS-II is the best selection method, followed by RRS-I and RRS. If overdominance is important, both RRS and RRS-I are superior to RRS-II; RRS is preferred to RRS-I if the effects of overdominant loci are sufficiently important. If the genetic model is a mixture of levels of dominance at different loci, a combination of selection systems is suggested.     

The conflict between natural and artificial selection in finite populations

Summary A single locus model of the interaction between natural selection and artificial selection for a quantitative character in a finite population, assuming heterozygote superiority in natural fitness but additive action on the character, has been studied using transition probability matrices.If natural selection is strong enough to create a selection plateau in which genetic variance declines relatively slowly, then the total response to artificial selection prior to the plateau will be much less than that expected in the absence of natural selection, and the half-life of response will be shorter. Such a plateau is likely to have a large proportion, if not all, of the original genetic variance still present. In selection programmes using laboratory animals, it seems likely that the homozygote favoured by artificial selection must be very unfit before such a plateau will occur. A significant decrease in population fitness as a result of artificial selection does not necessarily imply that the metric character is an important adaptive character.These implications of this model of natural selection are very similar to those derived by James (1962) for the optimum model of natural selection. In fact, there seems to be no aspect of the observable response to artificial selection that would enable anyone to distinguish between these two models of natural selection.     

The potential impacts of climate change on inshore squid: biology, ecology and fisheries

Squid are important components of many marine ecosystems from the poles to the equator, serving as both important predators and prey. Novel aspects of their growth and reproduction mean that they are likely to play an important role in the changing oceans due to climate change. Virtually every facet of squid life-history examined thus far has revealed an incredible capacity in this group for life-history plasticity. The extremely fast growth rates of individuals and rapid rates of turnover at the population level mean that squid can respond quickly to environmental or ecosystem change. Their ‘life-in-the-fast-lane’ life-style allows them to rapidly exploit ‘vacuums’ created in the ecosystem when predators or competitors are removed. In this way, they function as ‘weeds of the sea’. Elevated temperatures accelerate the life-histories of squid, increasing their growth rates and shortening their life-spans. At first glance, it would be logical to suggest that rising water temperatures associated with climate change (if food supply remains adequate) would be beneficial to inshore squid populations and fisheries—growth rates would increase, life spans would shorten and population turnover would accelerate. However, the response of inshore squid populations to climate change is likely to be extremely complex. The size of hatchlings emerging from the eggs becomes smaller as temperatures increase and hatchling size may have a critical influence on the size-at-age that may be achieved as adults and subsequently, population structure. The influence of higher temperatures on the egg and adult stages may thus be opposing forces on the life-history. The process of climate change will likely result in squids that hatch out smaller and earlier, undergo faster growth over shorter life-spans and mature younger and at a smaller size. Individual squid will require more food per unit body size, require more oxygen for faster metabolisms and have a reduced capacity to cope without food. It is therefore likely that biological, physiological and behavioural changes in squid due to climate change will have far reaching effects.     

A general approach for the study of a population of testcross progenies and consequences for recurrent selection

Summary A model to study genetic effects at the level of a population of testcross progenies is presented. As there is no dominance for the testcross value, with the restriction of epistasis to pairs of loci, only additive x additive epistasis can contribute to the variance among progenies. To estimate the variance among progenies due to epistasis, it is necessary to have the population structured in families of full sibs, half sibs or S₁, with only a few plants per family tested in combination with the tester. Using a two-way mating design to produce the families, it is possible to estimate the variance due to additive x additive epistasis. The consequence of the presence of epistasis is studied at the level of recurrent selection for combining ability with the tester. It seems that epistasis itself does not change the efficiency of the breeding methods considered. However, when the population from intercrossing is structured in families, it could be efficient to use a combined selection when the heritability is very low. In this case it would be efficient to produce full-sib families (by single-pair matings) at the level of intercrossing. The best procedure is to produce such families at the same time as crossing with the tester. In comparison to the classical scheme of selection for combining ability with a tester, such a modification increases the efficiency of selection 41.1% if an off-season generation can be used.     

Selection with Partial Selfing. I. Mass Selection

The expected responses to mass selection carried out before or after reproduction in a population whose members all have a fixed probability of self pollination (s) are formulated using covariances of relatives and their component quadratic functions for a model with arbitrary additive and dominance effects. The response measured in the first generation offspring after selection (immediate gain) can differ from that retained when the population has regained equilibrium (permanent gain). The population mean behaves in a predictable manner during the return to equilibrium, and its value at any time can be predicted from earlier generations. The permanent gain from selection after reproduction is always (1 + s)/2 times as large as that from selection before reproduction, but the relationship of the immediate gains depends on the genetic model assumed. Numerical analysis applied to a model with two alleles per locus and varying allele frequencies, dominance ratios and numbers of loci showed that the proportion of the immediate gain retained at equilibrium was reduced with the large inbreeding depression associated with increasing dominance levels and numbers of loci and was generally lower for selection after reproduction than before. In the absence of information as to the magnitude of genetic variances and inbreeding depression in species reproducing by partial selfing, the importance of this phenomenon is unknown.     

HIV evolution: CTL escape mutation and reversion after transmission

Within-patient HIV evolution reflects the strong selection pressure driving viral escape from cytotoxic T-lymphocyte (CTL) recognition. Whether this intrapatient accumulation of escape mutations translates into HIV evolution at the population level has not been evaluated. We studied over 300 patients drawn from the B- and C-clade epidemics, focusing on human leukocyte antigen (HLA) alleles HLA-B57 and HLA-B5801, which are associated with long-term HIV control and are therefore likely to exert strong selection pressure on the virus. The CTL response dominating acute infection in HLA-B57/5801-positive subjects drove positive selection of an escape mutation that reverted to wild-type after transmission to HLA-B57/5801-negative individuals. A second escape mutation within the epitope, by contrast, was maintained after transmission. These data show that the process of accumulation of escape mutations within HIV is not inevitable. Complex epitope- and residue-specific selection forces, including CTL-mediated positive selection pressure and virus-mediated purifying selection, operate in tandem to shape HIV evolution at the population level.     

Population bottlenecks increase additive genetic variance but do not break a selection limit in rain forest Drosophila

According to neutral quantitative genetic theory, population bottlenecks are expected to decrease standing levels of additive genetic variance of quantitative traits. However, some empirical and theoretical results suggest that, if nonadditive genetic effects influence the trait, bottlenecks may actually increase additive genetic variance. This has been an important issue in conservation genetics where it has been suggested that small population size might actually experience an increase rather than a decrease in the rate of adaptation. Here we test if bottlenecks can break a selection limit for desiccation resistance in the rain forest-restricted fly Drosophila bunnanda. After one generation of single-pair mating, additive genetic variance for desiccation resistance increased to a significant level, on average higher than for the control lines. Line crosses revealed that both dominance and epistatic effects were responsible for the divergence in desiccation resistance between the original control and a bottlenecked line exhibiting increased additive genetic variance for desiccation resistance. However, when bottlenecked lines were selected for increased desiccation resistance, there was only a small shift in resistance, much less than predicted by the released additive genetic variance. The small selection response in the bottlenecked lines was no greater than that observed in the control lines. Thus bottlenecks might produce a statistically detectable change in additive genetic variance but this change has no impact on the response to selection.     

No evidence for competition between cytotoxic T-lymphocyte responses in HIV-1 infection

Strong competition between cytotoxic T-lymphocytes (CTLs) specific for different epitopes in human immunodeficiency virus (HIV) infection would have important implications for the design of an HIV vaccine. To investigate evidence for this type of competition, we analysed CTL response data from 97 patients with chronic HIV infection who were frequently sampled for up to 96 weeks. For each sample, CTL responses directed against a range of known epitopes in gag, pol and nef were measured using an enzyme-linked immunospot assay. The Lotka–Volterra model of competition was used to predict patterns that would be expected from these data if competitive interactions materially affect CTL numbers. In this application, the model predicts that when hosts make responses to a larger number of epitopes, they would have diminished responses to each epitope and that if one epitope-specific response becomes dramatically smaller, others would increase in size to compensate; conversely if one response grows, others would shrink. Analysis of the experimental data reveals results that are wholly inconsistent with these predictions. In hosts who respond to more epitopes, the average epitope-specific response tends to be larger, not smaller. Furthermore, responses to different epitopes almost always increase in unison or decrease in unison. Our findings are therefore inconsistent with the hypothesis that there is competition between CTL responses directed against different epitopes in HIV infection. This suggests that vaccines that elicit broad responses would be favourable because they would direct a larger total response against the virus, in addition to being more robust to the effects of CTL escape.     

Inferences about the distribution of dominance drawn from yeast gene knockout data

Data from several thousand knockout mutations in yeast (Saccharomyces cerevisiae) were used to estimate the distribution of dominance coefficients. We propose a new unbiased likelihood approach to measuring dominance coefficients. On average, deleterious mutations are partially recessive, with a mean dominance coefficient ~0.2. Alleles with large homozygous effects are more likely to be more recessive than are alleles of weaker effect. Our approach allows us to quantify, for the first time, the substantial variance and skew in the distribution of dominance coefficients. This heterogeneity is so great that many population genetic processes analyses based on the mean dominance coefficient alone will be in substantial error. These results are applied to the debate about various mechanisms for the evolution of dominance, and we conclude that they are most consistent with models that depend on indirect selection on homeostatic gene expression or on the ability to perform well under periods of high demand for a protein.