Estimation of changes in genetic parameters in selected lines of mice using REML with an animal model. 1. Lean mass

The long-term effects of genomic selection: 1. Response to selection, additive genetic variance, and genetic architecture

Background

Genomic selection has revolutionized genetic improvement in animals and plants, but little is known about its long-term effects. Here, we investigated the long-term effects of genomic selection on response to selection, genetic variance, and the genetic architecture of traits using stochastic simulations. We defined the genetic architecture as the set of causal loci underlying each trait, their allele frequencies, and their statistical additive effects. We simulated a livestock population under 50 generations of phenotypic, pedigree, or genomic selection for a single trait, controlled by either only additive, additive and dominance, or additive, dominance, and epistatic effects. The simulated epistasis was based on yeast data.

Results

Short-term response was always greatest with genomic selection, while response after 50 generations was greater with phenotypic selection than with genomic selection when epistasis was present, and was always greater than with pedigree selection. This was mainly because loss of genetic variance and of segregating loci was much greater with genomic and pedigree selection than with phenotypic selection. Compared to pedigree selection, selection response was always greater with genomic selection. Pedigree and genomic selection lost a similar amount of genetic variance after 50 generations of selection, but genomic selection maintained more segregating loci, which on average had lower minor allele frequencies than with pedigree selection. Based on this result, genomic selection is expected to better maintain genetic gain after 50 generations than pedigree selection. The amount of change in the genetic architecture of traits was considerable across generations and was similar for genomic and pedigree selection, but slightly less for phenotypic selection. Presence of epistasis resulted in smaller changes in allele frequencies and less fixation of causal loci, but resulted in substantial changes in statistical additive effects across generations.

Conclusions

Our results show that genomic selection outperforms pedigree selection in terms of long-term genetic gain, but results in a similar reduction of genetic variance. The genetic architecture of traits changed considerably across generations, especially under selection and when non-additive effects were present. In conclusion, non-additive effects had a substantial impact on the accuracy of selection and long-term response to selection, especially when selection was accurate.

Natural selection stops the evolution of male attractiveness

Sexual selection in natural populations acts on highly heritable traits and tends to be relatively strong, implicating sexual selection as a causal agent in many phenotypic radiations. Sexual selection appears to be ineffectual in promoting phenotypic divergence among contemporary natural populations, however, and there is little evidence from artificial selection experiments that sexual fitness can evolve. Here, we demonstrate that a multivariate male trait preferred by Drosophila serrata females can respond to selection and results in the maintenance of male mating success. The response to selection was associated with a gene of major effect increasing in frequency from 12 to 35% in seven generations. No further response to selection, or increase in frequency of the major gene, was observed between generations 7 and 11, indicating an evolutionary limit had been reached. Genetic analyses excluded both depletion of genetic variation and overdominance as causes of the evolutionary limit. Relaxing artificial selection resulted in the loss of 52% of the selection response after a further five generations, demonstrating that the response under artificial sexual selection was opposed by antagonistic natural selection. We conclude that male D. serrata sexually selected traits, and attractiveness to D. serrata females conferred by these traits, were held at an evolutionary limit by the lack of genetic variation that would allow an increase in sexual fitness while simultaneously maintaining nonsexual fitness. Our results suggest that sexual selection is unlikely to cause divergence among natural populations without a concomitant change in natural selection, a conclusion consistent with observational evidence from natural populations.

Estimating genetic architectures from artificial-selection responses: a random-effect framework

Artificial-selection experiments on plants and animals generate large datasets reporting phenotypic changes in the course of time. The dynamics of the changes reflect the underlying genetic architecture, but only simple statistical tools have so far been available to analyze such time series. This manuscript describes a general statistical framework based on random-effect models aiming at estimating key parameters of genetic architectures from artificial-selection responses. We derive explicit Mendelian models (in which the genetic architecture relies on one or two large-effect loci), and compare them with classical polygenic models. With simulations, we show that the models are accurate and powerful enough to provide useful estimates from realistic experimental designs, and we demonstrate that model selection is effective in picking few-locus vs. polygenic genetic architectures even from medium-quality artificial-selection data. The method is illustrated by the analysis of a historical selection experiment, carried on color pattern in rats by Castle et al.

Characterizing the evolution of genetic variance using genetic covariance tensors

Determining how genetic variance changes under selection in natural populations has proved to be a very resilient problem in evolutionary genetics. In the same way that understanding the availability of genetic variance within populations requires the simultaneous consideration of genetic variance in sets of functionally related traits, determining how genetic variance changes under selection in natural populations will require ascertaining how genetic variance-covariance (G) matrices evolve. Here, we develop a geometric framework using higher-order tensors, which enables the empirical characterization of how G matrices have diverged among populations. We then show how divergence among populations in genetic covariance structure can then be associated with divergence in selection acting on those traits using key equations from evolutionary theory. Using estimates of G matrices of eight male sexually selected traits from nine geographical populations of Drosophila serrata, we show that much of the divergence in genetic variance occurred in a single trait combination, a conclusion that could not have been reached by examining variation among the individual elements of the nine G matrices. Divergence in G was primarily in the direction of the major axes of genetic variance within populations, suggesting that genetic drift may be a major cause of divergence in genetic variance among these populations.

Heritability in the genomics era--concepts and misconceptions

Heritability allows a comparison of the relative importance of genes and environment to the variation of traits within and across populations. The concept of heritability and its definition as an estimable, dimensionless population parameter was introduced by Sewall Wright and Ronald Fisher nearly a century ago. Despite continuous misunderstandings and controversies over its use and application, heritability remains key to the response to selection in evolutionary biology and agriculture, and to the prediction of disease risk in medicine. Recent reports of substantial heritability for gene expression and new estimation methods using marker data highlight the relevance of heritability in the genomics era.

Long-term responses, changes in genetic variances and inbreeding depression from 122 generations of selection on increased litter size in mice

Data on mice selected for litter size over 122 generations have been analysed in order to reveal the effect of long-term selection on responses and changes in variances over a long selection period. Originally, three lines were established from the same base population, namely an H line selected for large litter size, an L line selected for small litter size and a K line without selection. In generation 122, the mean number of pups born alive (NBA) was 22 for the H line and 11 for the K line. Phenotypic response to selection is reduced over generations, but crossing of plateaued lines increased responses and realized heritabilities. Both realized heritabilities and heritabilities from residual maximal likelihood (REML) analyses were, in general, calculated from generation (-1)-44 (period 1), 45-70 (period 2) and 71-122 (period 3) separately. Realized heritabilities were in general smaller than heritabilities estimated from mixed model analysis. An overall estimate of heritability for NBA was found to be 0.19 (+/- 0.01) by REML analysis. Additive variance is constant over all periods in the high line and the control line, but is reduced over periods in the low line. The reduction of additive variance in the low line could probably be explained by changes in gene frequencies. In all lines, environmental variances increased over periods. Inbreeding reduced the mean litter size by 0.72 (+/- 0.10) pups per 10% increase in inbreeding, with substantial variance between periods and lines.

Genetic analysis on the direct response to divergent selection for phytate phosphorus bioavailability in a randombred chicken population

The current study was undertaken to evaluate the direct response to 3 generations of divergent selection for phytate P bioavailability (PBA) in the Athens-Canadian randombred chicken population. Cumulated divergent response (R(C)) was measured as the line difference in PBA at a given generation after adjusting for hatch and sex effects. Results showed a significant response at generation (G)1. The R(C) was unchanged from G1 to G2 and increased (1.62%) from G2 to G3 (P < 0.01) due to the application of best linear unbiased prediction (BLUP) selection in the line selected for high PBA at G2. The average BLUP estimated breeding values were used to estimate the genetic trend for the selected trait across generations. The results showed that the genetic trend was symmetric at G1 and G2 but asymmetric at G3. The application of mixed model methodology was effective in separating the environmental component from phenotypic change. When the data of the high (H) line or the low (L) line in the selected generations (G1 to G3) were combined with the data from the base population (G0), the heritability estimates for PBA were 0.07 +/- 0.02 and 0.09 +/- 0.02, respectively. The line selected for high PBA showed gain, and the line selected for low PBA showed a decrease in estimated breeding values across the generations. The results demonstrated that modest progress could be obtained by incorporating PBA into selection programs. However, other correlated traits of economic importance need to be evaluated before any decision to incorporate selection of PBA into breeding schemes be initiated.

Persistence of changes in the genetic covariance matrix after a bottleneck

Genetic variance, phenotypic variance, and the genetic covariance matrix (G) can change as a result of genetic drift. These changes will persist over time to some extent and will continue if population size remains relatively small. Nine populations founded by a single pair of Drosophila melanogaster were measured for a series of six morphological characteristics for a large number of parent-offspring families at both the third generation after the bottlenecks and after 20 generations. From these data, the phenotypic variance, additive genetic variance, and G were estimated for each line at each generation. Phenotypic and genetic variances were highly correlated over time, so that the measurements made at the third generation were predictive of the state of the population 17 generations later. Genetic covariances were also somewhat stable over time; however, the G matrices of some lines changed significantly over the intervening generations. This change did not return the populations toward their original state before the population bottlenecks. We conclude that the genetic covariance matrix can change as a result of mild genetic drift over a short span of time.

Dwarfism in insular sloths: biogeography, selection, and evolutionary rate

The islands of Bocas del Toro, Panama, were sequentially separated from the adjacent mainland by rising sea levels during the past 10,000 years. Three-toed sloths (Bradypus) from five islands are smaller than their mainland counterparts, and the insular populations themselves vary in mean body size. We first examine relationships between body size and physical characteristics of the islands, testing hypotheses regarding optimal body size, evolutionary equilibria, and the presence of dispersal in this system. To do so, we conduct linear regressions of body size onto island area, distance from the mainland, and island age. Second, we retroactively calculate two measures of the evolutionary rate of change in body size (haldanes and darwins) and the standardized linear selection differential, or selection intensity (i). We also test the observed morphological changes against models of evolution by genetic drift. The results indicate that mean body size decreases linearly with island age, explaining up to 97% of the variation among population means. Neither island area nor distance from the mainland is significant in multiple regressions that include island age. Thus, we find no evidence for differential optimal body size among islands, or for dispersal in the system. In contrast, the dependence of body size on island age suggests uniform directional selection for small body size in the insular populations. Although genetic drift cannot be discounted as the cause for this evolution in body size, the probability is small given the consistent direction of evolution (repeated dwarfism). The insular sloths show a sustained rate of evolution similar to those measured in haldanes over tens of generations, appearing to unite micro- and macroevolutionary time scales. Furthermore, the magnitude and rate of this example of rapid differentiation fall within predictions of theoretical models from population genetics. However, the linearity of the relationship between body size and island age is not predicted, suggesting that either more factors are involved than those considered here, or that theoretical advances are necessary to explain constant evolutionary rates over long time spans in new selective environments.

Rates of change of genetic parameters of body weight in selected mouse lines

A method based on the animal model is described which allows the estimation of continuous changes in variance components over time using restricted maximum likelihood (REML). The method was applied to the analysis of a selection experiment in which a foundation population formed from a cross between two inbred strains of mice (C57BL/6J and DBA/2J) was divergently selected for 6 week body weight over 20 generations. The analysis suggested that there was an increase in phenotypic variance of about 50% in the low selected lines over the course of the experiment which was attributed to increases in the environmental and additive variance components. Variance changes in the High selected lines were generally smaller than in the Low lines, although there was an estimated 20% increase in the environmental variance. Simple models to explain these effects involving dominance, linkage and epistasis were explored. Testing which of these was responsible for the variance changes noted in this experiment (if any) is difficult, although the epistasis and dominance models require less stringent conditions than the linkage model, and the dominance model is supported by evidence of heterosis in the F1.

Behavioural changes as a correlated response to selection

Lines of mice have been selected for up to 50 generations on the following traits: high body weight, low body weight, high fat content or low fat content. The lines selected for high or low body weight differ by a factor of 2.5 and those selected for high or low fat content differ by a factor of five, both traits measured in 10 week old males. A set of behavioural traits was measured to ascertain whether this selection had caused correlated responses in behaviour: studies included feeding behaviour, open field behaviour, ultrasound calling rates of pups, and the response to the introduction of a novel physical object. Alterations in behavioural patterns which were expected a priori were observed but there appeared to be no changes in behaviour associated with any one selection criterion. Estimates of the genetic correlations between selected and behavioural traits were, with one exception, generally less than 0.1 in magnitude and not significantly different from zero (the exception was food intake in lines selected on body weight). Assuming that mice are accurate models for commercial species, then these results have important implications for animal welfare: they demonstrate that large scale behavioural changes do not arise as an inevitable consequence of intense long-term selection on traits of economic importance in commercial species.

The genetic basis of response in mouse lines divergently selected for body weight or fat content. I. The relative contributions of autosomal and sex-linked genes

Lines of mice have been divergently selected for over forty generations on either body weight or fat content. Reciprocal crosses were made between the divergent lines and the offspring backcrossed to the parental lines. The resulting data allowed us to investigate the genetic basis of response, including two features of particular interest: (i) the relative contribution of autosomal and sex-linked genes and whether any significant Y chromosome or cytoplasmic effects were present (ii) the mechanism of gene action, whether predominantly additive or whether significant dominance effects were present. A large additive sex-linked effect was observed in lines selected on body weight which accounted for approximately 25% of the divergence. The remaining 75% of the divergence appeared to be autosomal. There was no apparent sex-linked effect in lines selected on fat content and the response appeared to be entirely autosomal and additive.

The genetic basis of response in mouse lines divergently selected for body weight or fat content. II. The contribution of genes with a large effect

Gene action underlying selection responses has been studied using crossbreeding. Maximum likelihood based segregation analysis has been presented for analysing backcross data for the presence of genes with a large effect. Two sets of divergently selected lines (P-lines for body weight and F-lines for fat content) were reciprocally crossed and the F1s were crossed to the high and low lines to produce all possible backcrosses. Earlier analysis had shown that the difference in body weight at 10 weeks (n = 595) between the high and low P-lines was largely (75-80%) explained by autosomal, additive genes with the remainder explained by additive genes on the X chromosome. Maximum likelihood segregation analysis suggested the presence of a major effect on the X chromosome, but as there was only one round of recombination between the X chromosomes in the forming of the backcrosses, linked genes on the X chromosome could have acted together to give the appearance of a single major gene. The difference in fat content between the F-lines (n = 578) could be explained by autosomal genes of largely additive effect. Segregation analysis suggested the presence of a major gene with complete dominance, but this was attributed to a relationship between the mean and the variance: transformation of the data resulted in only polygenic additive genes being of importance. This study concluded that maximum likelihood based analysis and crosses between selected lines provide a powerful means for studying the gene action underlying responses to selection.

The role of growth hormone in lines of mice divergently selected on body weight

An understanding of the physiological and genetic changes which determine the response to selection is critical for both evolutionary theory and to assess the application of new molecular techniques to commercial animal breeding. We investigated an aspect of physiology, growth hormone (GH) metabolism, which might a priori have been expected to play a large part in the response of mouse lines selected for high or low body weight. Disruption of endogenous GH or addition of exogenous GH had similar proportionate effects on body weight in both lines of mice (although differences in body composition arose) suggesting that neither the production of GH nor receptor sensitivity to GH had been altered as a result of selection. This supports a 'pleiotropic model' of the response to selection: that many genes with diverse metabolic roles all contribute to the divergent phenotype. This result has significant commercial implications as it suggests that artificial selection, transgenic technology and environmental manipulation may be synergistic rather than antagonistic strategies.

Estimation of changes in genetic parameters in selected lines of mice using REML with an animal model. 2. Body weight, body composition and litter size

Restricted Maximum Likelihood (REML) with an animal model was used to estimate genetic parameters of body weight, body consumption and litter size of lines of mice selected for 20 generations on an index of lean mass at 10 weeks in males, highly correlated with body weight, and for a further 18 generations on body weight at 10 weeks in males and females. Univariate and multivariate estimates of heritability were about 0.5 and those of common environment correlations were about 0.25 for both body weight and composition. Body weight and fat pad weight had genetic and phenotypic correlations of about 0.5. The heritability estimate of litter size was about 0.15 from univariate analysis, rather lower from multivariate, and the estimate of its genetic correlation with body weight was about 0.25. There were reductions in heritability of both body weight and litter size in later generations, even though full pedigrees were fitted and inferences made to the base population, but a plateau in response to selection for increased body weight could not be explained by a complete attenuation of genetic variance.