Genetic correlation and response to selection in simulated populations

Summary Effects of truncation selection of a primary trait upon genetic correlation between the primary trait and an unselected secondary trait were observed during 30 generations. Populations were 24 male and 24 female parents per generation randomly mated with replacement, the number of offspring set by intensity of selection. Each trait was controlled by genes with equal effects and complete dominance segregating independently from starting frequencies of 0.5 at each of 48 loci. Three levels each of genetic correlation, selection, and environmental variation were simulated.Genetic correlation decreased faster under more intense selection by lower than by upper truncation but behaved similarly in both by remaining near initial level when as many as one-half of the offspring were saved for parents. Truncation selection decreased genetic correlation in the offspring selected to be parents whether selection was by upper or lower truncation. Estimates of genetic correlation from covariances between phenotypes of parent and offspring were erratic for both directions of selection.Michigan Agricultural Experiment Station Journal Article4841. Part of North Central Regional Project NC-2.     

Effect of selection on genetic parameters of correlated traits

Summary Changes in genetic parameters of correlated traits due to the buildup of linkage (gametic phase) disequilibrium from repeated truncation selection on a single trait are studied. After several generations of selection, an equilibrium is approached where there are no further changes in genetic parameters and limiting values are reached. Formulae are derived under an infinitesimal model for these limiting values of genetic variances and covariances, heritabilities, and genetic correlations between traits directly and indirectly selected. Changes from generation zero to the limit in all these parameters become greater as heritability of the trait under direct selection increases and, to a lesser extent, as intensity of selection increases. Change in heritability of a trait under indirect selection also increases as the absolute value of the correlation between the trait under indirect and the trait under direct selection increases. The change is maximum when the initial value of heritability is close to 0.5 and insignificant when the initital value is close to zero or one. Change in the genetic correlation between the trait under direct selection and the trait under indirect selection is maximum when its initial value is close to ±0.6 and insignificant when its initial value is close to zero or ±1. Heritability of the trait indirectly selected and genetic correlation between that trait and the trait directly selected always decrease in absolute value, whereas genetic correlation between two traits indirectly selected can either decrease or increase in absolute value. It is suggested that use be made of formulae at selection equilibrium in the prediction of correlated responses after several generations of selection.     

A multi-marker test based on family data in genome-wide association study

Background

Requirements for successful implementation of multivariate animal threshold models including phenotypic and genotypic information are not known yet. Here simulated horse data were used to investigate the properties of multivariate estimators of genetic parameters for categorical, continuous and molecular genetic data in the context of important radiological health traits using mixed linear-threshold animal models via Gibbs sampling. The simulated pedigree comprised 7 generations and 40000 animals per generation. Additive genetic values, residuals and fixed effects for one continuous trait and liabilities of four binary traits were simulated, resembling situations encountered in the Warmblood horse. Quantitative trait locus (QTL) effects and genetic marker information were simulated for one of the liabilities. Different scenarios with respect to recombination rate between genetic markers and QTL and polymorphism information content of genetic markers were studied. For each scenario ten replicates were sampled from the simulated population, and within each replicate six different datasets differing in number and distribution of animals with trait records and availability of genetic marker information were generated. (Co)Variance components were estimated using a Bayesian mixed linear-threshold animal model via Gibbs sampling. Residual variances were fixed to zero and a proper prior was used for the genetic covariance matrix.

Results

Effective sample sizes (ESS) and biases of genetic parameters differed significantly between datasets. Bias of heritability estimates was -6% to +6% for the continuous trait, -6% to +10% for the binary traits of moderate heritability, and -21% to +25% for the binary traits of low heritability. Additive genetic correlations were mostly underestimated between the continuous trait and binary traits of low heritability, under- or overestimated between the continuous trait and binary traits of moderate heritability, and overestimated between two binary traits. Use of trait information on two subsequent generations of animals increased ESS and reduced bias of parameter estimates more than mere increase of the number of informative animals from one generation. Consideration of genotype information as a fixed effect in the model resulted in overestimation of polygenic heritability of the QTL trait, but increased accuracy of estimated additive genetic correlations of the QTL trait.

Conclusion

Combined use of phenotype and genotype information on parents and offspring will help to identify agonistic and antagonistic genetic correlations between traits of interests, facilitating design of effective multiple trait selection schemes.     

Efficiency of direct selection on quantitative trait loci for a two-trait breeding objective

Selection on known loci affecting quantitative traits (DSQ) was compared to phenotypic selection index for a single and a two-trait selection objective. Two situations were simulated; a single known quantitative locus, and ten identified loci accounting for all the additive genetic variance. Selection efficiency of DSQ relative to traitbased selection was higher for two-trait selection, than was selection on a single trait with the same heritability. The advantage of DSQ was greater when the traits were negatively correlated. Relative selection efficiency (RSE) for a single locus responsible for 0.1 of the genetic variance was 1.11 with heritabilities of 0.45 and 0.2 and zero genetic and phenotypic correlations between the traits. RSE of DSQ for ten known loci was 1.5 to 1.8 in the first 3 generations of selection, but declined in each subsequent generation. With DSQ most loci reached fixation after 7 generations. Response to trait-based selection continued through generation 15 and approached the response obtained with DSQ after 10 generations. The cumulative genetic response after 10 generations of DSQ was only 93% to 97% of the economically optimum genotype because the less favorable allele reached fixation for some loci, generally those with effects in opposite directions on the two traits.     

The effect of simultaneous selection on the genetic correlation

The theoretical effect of simultaneous selection on the genetic correlations between two traits over 20 generations was examined using simulation. For each generation, a population of 50 male and 50 female diploid gen otypes with 15 loci, each with two alleles, was synthesized. None of the loci exhibited dominance. Five loci affected only trait 1, 5 loci only trait 2 and 5 were pleiotropic (affected both traits). Initial allelic frequencies were equal at each locus. Phenotypes were created by adding a random normal deviation for each trait to the genotype. The size of this deviation for each trait determined its heritability (h²). Index selection with h² combinations of (0.15, 0.15), (0.15,0.45) and (0.45,0.45) and relative economic weights of (1, 1) and (1, 3) for each h² combination was employed. In each generation, the highest ranking 25 genotypes of each sex were used to generate the next generation with single-pair matings, each producing two male and two female offspring. One hundred replicates were run for both negative and positive correlations. With a positive initial value, the genetic correlation tended to decline (toward zero). The rates of change were moderately affected by index weights and h². With a negative initial value, the genetic correlation tended to decrease (towards -1). However, unequal heritabilities and unequal relative economic weights slowed the rate of change with the greatest imbalances tending to hold the correlation constant or move it toward zero. These simulations illustrate that changes in parameters over time can affect the selection practiced. Under some of the conditions simulated, the use of initial genetic parameter values without change could have potentially negative effects on overall genetic gain.     

Effect of various parameter combinations on genetic gain in computer simulated two-character selection

Summary In a simulation study, the effect of various parameter combinations such as linkage, dominance, heritability, and economic weights on the individual trait means was investigated using additive genetic, genotypic and the phenotypic index of Elston (1963). The characters responded differently to these indices under various parameter combinations, indicating favourable and unfavourable effects of the mentioned parameters. Linkage was found to reduce the rate of progress through selection. Depression of genetic gain was greater where the genes governing a character showed dominance and/or heritability coefficients were low. It was, however, noticed that depression of genetic gain due to low heritability of a character could be avoided by assigning higher economic weight to that character. This suggests that desirable changes in the means of characters available for selection can be manipulated by choosing appropriate economic weights. The additive genetic index, where only the additive genetic variances and covariances go into its construction, does not seem to be affected by intra-allelic interactions since they add to variances and covariances due to dominance deviations and these have nothing to do with the additive genetic variances and covariances. It seems that from such studies, if conducted extensively incorporating still more parameters, conclusions may be drawn on the most suitable selection model for simultaneous selection under a given set of parameters available in real biological systems.     

Efficiency of indirect selection at selection equilibrium

Summary Efficiency of indirect selection compared with that of direct selection to increase the mean value of some trait has been usually studied by considering a single generation of indirect and direct responses to selection only. However, under continued selection, genetic variances and covariances, and therefore expected genetic responses, change each generation due to linkage disequilibrium. With directional and truncation selection, genetic parameters asymptote to limiting values after several generations. The efficiency of indirect selection is examined in this limiting situation. The ratio of correlated response to direct response for the trait to improve in the limit is compared with the ratio after the first generation of selection. For all initial parameter values for which indirect selection is more efficient than direct selection, relative efficiency of indirect selection is smaller in the limit than in the first generation. For some parameter values, indirect selection is more efficient than direct selection in the first generation, but less efficient in the limit. Expressions for minimum values of the initial genetic correlation and heritability of the alternative trait required for indirect selection to be preferred in the limit are derived. These values are higher when limiting responses are used instead of single generation responses. The loss in relative efficiency of indirect selection from changes in genetic parameters due to selection should be taken into account when applications of indirect selection are considered.     

Prediction of additive and dominance effects in selected or unselected populations with inbreeding

Summary A genetic model with either 64 or 1,600 unlinked biallelic loci and complete dominance was used to study prediction of additive and dominance effects in selected or unselected populations with inbreeding. For each locus the initial frequency of the favourable allele was 0.2, 0.5, or 0.8 in different alternatives, while the initial narrow-sense heritability was fixed at 0.30. A population of size 40 (20 males and 20 females) was simulated 1,000 times for five generations. In each generation 5 males and 10 or 20 females were mated, with each mating producing four or two offspring, respectively. Breeding individuals were selected randomly, on own phenotypic performance or such yielding increased inbreeding levels in subsequent generations. A statistical model containing individual additive and dominance effects but ignoring changes in mean and genetic covariances associated with dominance due to inbreeding resulted in significantly biased predictions of both effects in generations with inbreeding. Bias, assessed as the average difference between predicted and simulated genetic effects in each generation, increased almost linearly with the inbreeding coefficient. In a second statistical model the average effect of inbreeding on the mean was accounted for by a regression of phenotypic value on the inbreeding coefficient. The total dominance effect of an individual in that case was the sum of the average effect of inbreeding and an individual effect of dominance. Despite a high mean inbreeding coefficient (up to 0.35), predictions of additive and dominance effects obtained with this model were empirically unbiased for each initial frequency in the absence of selection and 64 unlinked loci. With phenotypic selection of 5 males and only 10 females in each generation and 64 loci, however, predictions of additive and dominance effects were significantly biased. Observed biases disappeared with 1,600 loci for allelic frequencies at 0.2 and 0.5. Bias was due to a considerable change in allelic frequency with phenotypic selection. Ignoring both the covariance between additive and dominance effects with inbreeding and the change in dominance variance due to inbreeding did not significantly bias prediction of additive and dominance effects in selected or unselected populations with inbreeding.     

Maximum likelihood techniques for the mapping and analysis of quantitative trait loci with the aid of genetic markers

A method is presented to estimate the biometric parameters of a quantitative trait locus linked to a genetic marker when both loci are segregating in the F-2 generation of a cross between two inbred lines. The method, which assumes underlying normal distributions, is a combination of maximum likelihood and moments methods and uses the statistics of the genetic marker genotype samples for the quantitative trait to estimate the recombination frequency between the two loci and the means and variances of the genotypes of the quantitative trait locus. With this method, the genetic parameters of a locus affecting plant height linked to an electrophoretic marker for esterase were accurately estimated from a sample of 1596 F-2 progeny of a cross between two species of Lycopersicon (tomato). Linkage distance between the two loci was 38 map units and the effect of the quantitative trait locus was 1.6 phenotypic standard deviation units. Accurate estimates of the genetic parameters and linkage distance for populations of 2000 individuals simulated with a segregating codominant locus with an effect of 1.63 standard deviations linked to a genetic marker with .2 recombination were also derived by this method. The method is not effective in distinguishing between complete and partial linkage in samples of only 500 individuals or for quantitative loci with effects less than a phenotypic standard deviation. The method is more effective for codominant than for dominant loci.     

Genetic evaluation by best linear unbiased prediction using marker and trait information in a multibreed population.

Genetic evaluation by best linear unbiased prediction (BLUP) requires modeling genetic means, variances, and covariances. This paper presents theory to model means, variances, and covariances in a multibreed population, given marker and breed information, in the presence of gametic disequilibrium between the marker locus (ML) and linked quantitative trait locus (MQTL). Theory and algorithms are presented to construct the matrix of conditional covariances between relatives (Gv) for the MQTL effects in a multibreed population and to obtain the inverse of Gv efficiently. Theory presented here accounts for heterogeneity of variances among pure breeds and for segregation variances between pure breeds. A numerical example was used to illustrate how the theory and algorithms can be used for genetic evaluation by BLUP using marker and trait information in a multibreed population.     

Natural selection, evolvability and bias due to environmental covariance in the field in an annual plant

Estimates of the form and magnitude of natural selection based on phenotypic relationships between traits and fitness measures can be biased when environmental factors influence both relative fitness and phenotypic trait values. I quantified genetic variances and covariances, and estimated linear and quadratic selection coefficients, for seven traits of an annual plant grown in the field. For replicates of 50 paternal half-sib families, coefficients of selection were calculated both for individual phenotypic values of the traits and for half-sib family mean values. The potential for evolutionary response was supported by significant heritability and phenotypic directional selection for several traits but contradicted by the absence of significant genetic variation for fitness estimates and evidence of bias in phenotypic selection coefficients due to environmental covariance for at least two of the traits analysed. Only studies of a much wider range of organisms and traits will reveal the frequency and extent of such bias.     

Genetic and phenotypic parameters for test day milk yield of Sahiwal cattle in the semi-arid tropics

A total of 19 376 test day (TD) milk yield records from the first three lactations of 1618 cows daughters of 162 sires were used to estimate genetic and phenotypic parameters and determine the relationship between daily milk yield and lactation milk yield in the Sahiwal cattle in Kenya. Variance components were estimated using animal models based on a derivative free restricted maximum likelihood procedure. Variance components were estimated using various univariate and multi-trait fixed regression test day models (TDM) that defined contemporary groups either based on the year-season of calving (YSCV) or on the year-season of TD milk sampling (YSTD). Variance components were influenced by CG which resulted in differences in heritability and repeatability estimates between TDM. Models considering YSTD resulted in higher additive genetic variances and lower residual variances compared with models in which YSCV was considered. Heritability estimates for daily yield ranged from 0.28 to 0.46, 0.38 to 0.52 and 0.33 to 0.52 in the first, second and third lactation, respectively. In the first and second lactation, the heritability estimates were highest between TD 2 and TD 4. Genetic correlations among daily milk yields ranged from 0.41 to 0.93, 0.50 to 0.83 and 0.43 to 86 in the first, second and third lactation, respectively. The phenotypic correlations were correspondingly lower. Genetic correlations were different from unit when fitting multi-trait TDM. Therefore, a multiple trait model would be more ideal in determining the genetic merit of dairy sires and bulls based on daily yield records. Genetic and phenotypic correlations between daily yield and lactation yields were high and positive. Genetic correlations ranged from 0.84 to 0.99, 0.94 to 1.00 and 0.94 to 0.97 in the first, second and third lactations, respectively. The corresponding phenotypic correlation estimates ranged from 0.50 to 0.85, 0.50 to 0.83 and 0.53 to 0.87. The high genetic correlation between daily yield and lactation yield imply that both traits are influenced by similar genes. Therefore daily yields records could be used in genetic evaluation in the Sahiwal cattle breeding programme.     

Accuracy of marker-assisted selection with auxiliary traits

Genetic information on molecular markers is increasingly being used in plant and animal improvement programmes particularly as indirect means to improve a metric trait by selection either on an individual basis or on the basis of an index incorporating such information. This paper examines the utility of an index of selection that not only combines phenotypic and molecular information on the trait under improvement but also combines similar information on one or more auxiliary traits. The accuracy of such a selection procedure has been theoretically studied for sufficiently large populations so that the effects of detected quantitative trait loci can be perfectly estimated. The theory is illustrated numerically by considering one auxiliary trait. It is shown that the use of an auxiliary trait improves the selection accuracy; and, hence, the relative efficiency of index selection compared to individual selection which is based on the same intensity of selection. This is particularly so for higher magnitudes of residual genetic correlation and environmental correlation having opposite signs, lower values of the proportion of genetic variation in the main trait associated with the markers, negligible proportion of genetic variation in the auxiliary trait associated with the markers, and lower values of the heritability of the main trait but higher values of the heritability of the auxiliary trait.     

Bayesian prediction of breeding values for multivariate binary and continuous traits in simulated horse populations using threshold-linear models with Gibbs sampling

Simulated data were used to determine the properties of multivariate prediction of breeding values for categorical and continuous traits using phenotypic, molecular genetic and pedigree information by mixed linear-threshold animal models via Gibbs sampling. Simulation parameters were chosen such that the data resembled situations encountered in Warmblood horse populations. Genetic evaluation was performed in the context of the radiographic findings in the equine limbs. The simulated pedigree comprised seven generations and 40 000 animals per generation. The simulated data included additive genetic values, residuals and fixed effects for one continuous trait and liabilities of four binary traits. For one of the binary traits, quantitative trait locus (QTL) effects and genetic markers were simulated, with three different scenarios with respect to recombination rate (r) between genetic markers and QTL and polymorphism information content (PIC) of genetic markers being studied: r = 0.00 and PIC = 0.90 (r0p9), r = 0.01 and PIC = 0.90 (r1p9), and r = 0.00 and PIC = 0.70 (r0p7). For each scenario, 10 replicates were sampled from the simulated horse population, and six different data sets were generated per replicate. Data sets differed in number and distribution of animals with trait records and the availability of genetic marker information. Breeding values were predicted via Gibbs sampling using a Bayesian mixed linear-threshold animal model with residual covariances fixed to zero and a proper prior for the genetic covariance matrix. Relative breeding values were used to investigate expected response to multi- and single-trait selection. In the sires with 10 or more offspring with trait information, correlations between true and predicted breeding values ranged between 0.89 and 0.94 for the continuous traits and between 0.39 and 0.77 for the binary traits. Proportions of successful identification of sires of average, favourable and unfavourable genetic value were 81% to 86% for the continuous trait and 57% to 74% for the binary traits in these sires. Expected decrease of prevalence of the QTL trait was 3% to 12% after multi-trait selection for all binary traits and 9% to 17% after single-trait selection for the QTL trait. The combined use of phenotype and genotype data was superior to the use of phenotype data alone. It was concluded that information on phenotypes and highly informative genetic markers should be used for prediction of breeding values in mixed linear-threshold animal models via Gibbs sampling to achieve maximum reduction in prevalences of binary traits.     

Genetic variances and covariances of aerobic metabolic rates in laboratory mice

The genetic variances and covariances of traits must be known to predict how they may respond to selection and how covariances among them might affect their evolutionary trajectories. We used the animal model to estimate the genetic variances and covariances of basal metabolic rate (BMR) and maximal metabolic rate (MMR) in a genetically heterogeneous stock of laboratory mice. Narrow-sense heritability (h²) was approximately 0.38 ± 0.08 for body mass, 0.26 ± 0.08 for whole-animal BMR, 0.24 ± 0.07 for whole-animal MMR, 0.19 ± 0.07 for mass-independent BMR, and 0.16 ± 0.06 for mass-independent MMR. All h² estimates were significantly different from zero. The phenotypic correlation of whole animal BMR and MMR was 0.56 ± 0.02, and the corresponding genetic correlation was 0.79 ± 0.12. The phenotypic correlation of mass-independent BMR and MMR was 0.13 ± 0.03, and the corresponding genetic correlation was 0.72 ± 0.03. The genetic correlations of metabolic rates were significantly different from zero, but not significantly different from one. A key assumption of the aerobic capacity model for the evolution of endothermy is that BMR and MMR are linked. The estimated genetic correlation between BMR and MMR is consistent with that assumption, but the genetic correlation is not so high as to preclude independent evolution of BMR and MMR.     

Genetic and phenotypic correlations among size-related traits, and heritability variation between body parts in Drosophila buzzatii

Recent studies have shown that body size is a heritable trait phenotypically correlated with several fitness components in wild populations of the cactophilic fly Drosophila buzzatii. To obtain further information on size-related variation, heritabilities as well as genetic and phenotypic correlations among size-related traits of several body parts (head, thorax and wings) were estimated. The study was carried out on an Argentinean natural population in which size-related selection was previously detected. The genetic parameters were estimated using offspring-parent regressions (105 families) in the laboratory G₂ generation of a sample of wild flies. The traits were also scored in Wild-Caught Flies (WCF). Laboratory-Reared Flies (LRF) were larger and less variable than WCF. Although heritability estimates were significant for all traits, heritabilities were higher for thorax-wing traits than for head traits. Phenotypic and genetic correlations were all positive. The highest genetic correlations were found between traits which are both functionally and developmentally related. Genetic and phenotypic correlations estimated in the lab show similar correlation patterns (r = 0.49; TP = 0.02, Mantel's test). However, phenotypic correlations were found to be typically larger in WCF than in LRF. The genetic correlation matrix estimated in the relatively homogeneous lab environment is not simply a constant multiplicative factor of the phenotypic correlation matrix estimated in WCF. This revised version was published online in July 2006 with corrections to the Cover Date.     

Inheritance of size and shape in a natural population of collared flycatchers,Ficedula albicollis

Four external skeletal and three feather dimensions were measured on adult collared flycatchers (Ficedula albicollis) and their adult offspring. By using mid-offspring-midparent regressions, all traits were found to be heritable with an arithmetic mean heritability of 0.46. Heritability estimates from full-sib analyses were about 1.5 times higher (mean 0.67), indicating that variation in traits was affected by shared nest environment among full-sibs. The overall body size as measured by principal component one (PC1) was found to be heritable (h² = 0.40). However, this multivariate measure of heritability was not significant in offspring-father comparison, while highly so in offspring-mother comparison (h² = 0.60). Low offspring-father resemblance was evident also in univariate estimates of heritability. Possible causes of this (extra-pair copulations, maternal effects, sex-linked variance) are discussed. Genetic correlations among seven traits were estimated to be low (mean 0.22), and of similar magnitude or higher than phenotypic correlations (mean 0.18). All genetic correlations were positive. Genetic and phenotypic correlations as well as covariances were fairly similar to each other (r = 0.85 and r = 0.87, respectively). Environmental correlations did not follow the pattern of genetic correlations (r = 0.11), but were more similar to phenotypic correlations (r = 0.60). Given the low genetic correlations and moderate heritabilities, the overall conclusion is that the external morphology of collared flycatchers is largely under additive genetic control and that there is a strong potential for evolutionary change in morphology even under complex multivariate selection.     

Application of the covariance function approach with an iterative two-stage algorithm to the estimation of parameters of a random regression test day model for dairy production traits

The covariance function approach with an iterative two-stage algorithm of LIU et al. (2000) was applied to estimate parameters for the Polish Black-and-White dairy population based on a sample of 338 808 test day records for milk, fat, and protein yields. A multiple trait sire model was used to estimate covariances of lactation stages. A third-order Legendre polynomial was subsequently fitted to the estimated (co)variances to derive (co)variances of random regression coefficients for both additive genetic and permanent environment effects. Daily and 305-day heritability estimates obtained are consistent with several studies which used both fixed and random regression test day models. Genetic correlations between any two days in milk (DIM) of the same lactation as well as genetic correlations between the same DIM of two lactations were within a biologically acceptable range. It was shown that the applied estimation procedure can utilise very large data sets and give plausible estimates of (co)variance components.     

Genetic response from marker assisted selection in an outbred population for differing marker bracket sizes and with two identified quantitative trait loci.

Effect of flanking quantitative trait loci (QTL)-marker bracket size on genetic response to marker assisted selection in an outbred population was studied by simulation of a nucleus breeding scheme. In addition, genetic response with marker assisted selection (MAS) from two quantitative trait loci on the same and different chromosome(s) was investigated. QTL that explained either 5% or 10% of phenotypic variance were simulated. A polygenic component was simulated in addition to the quantitative trait loci. In total, 35% of the phenotypic variance was due to genetic factors. The trait was measured on females only. Having smaller marker brackets flanking the QTL increased the genetic response from MAS selection. This was due to the greater ability to trace the QTL transmission from one generation to the next with the smaller flanking QTL-marker bracket, which increased the accuracy of estimation of the QTL allelic effects. Greater negative covariance between effects at both QTL was observed when two QTL were located on the same chromosome compared to different chromosomes. Genetic response with MAS was greater when the QTL were on the same chromosome in the early generations and greater when they were on different chromosomes in the later generations of MAS.     

Sexual variation in heritability and genetic correlations of morphological traits in house sparrow (Passer domesticus)

Estimates of genetic components are important for our understanding of how individual characteristics are transferred between generations. We show that the level of heritability varies between 0.12 and 0.68 in six morphological traits in house sparrows (Passer domesticus L.) in northern Norway. Positive and negative genetic correlations were present among traits, suggesting evolutionary constraints on the evolution of some of these characters. A sexual difference in the amount of heritable genetic variation was found in tarsus length, wing length, bill depth and body condition index, with generally higher heritability in females. In addition, the structure of the genetic variance-covariance matrix for the traits differed between the sexes. Genetic correlations between males and females for the morphological traits were however large and not significantly different from one, indicating that sex-specific responses to selection will be influenced by intersexual differences in selection differentials. Despite this, some traits had heritability above 0.1 in females, even after conditioning on the additive genetic covariance between sexes and the additive genetic variances in males. Moreover, a meta-analysis indicated that higher heritability in females than in males may be common in birds. Thus, this indicates sexual differences in the genetic architecture of birds. Consequently, as in house sparrows, the evolutionary responses to selection will often be larger in females than males. Hence, our results suggest that sex-specific additive genetic variances and covariances, although ignored in most studies, should be included when making predictions of evolutionary changes from standard quantitative genetic models.