Multi-trait genomic prediction in pigs using single and multistep methods based on the absorption of ungenotyped animals.

Many quantitative traits measured in breeding programs are genetically correlated. The genetic correlations between the traits indicate that the measurement of one trait carries information on others. To benefit from this information, multi-trait genomic prediction (MTGP) is preferable to use. However, MTGP is more difficult to implement compared to single-trait genomic prediction (STGP), and even more challenging for the goal to exploit not only the information on other traits but also the information on ungenotyped animals. This could be accomplished using both single and multistep methods. The single-step method was achieved by implementing a single-step genomic best linear unbiased prediction (ssGBLUP) approach using a multi-trait model. Here, we examined a multistep analysis based on an approach called "Absorption" to achieve this goal. The Absorption approach absorbed all available information including the phenotypic information on ungenotyped animals as well as the information on other traits if applicable, into mixed model equations of genotyped animals. The multistep analysis included (1) to apply the Absorption approach that exploits all available information and (2) to implement genomic BLUP (GBLUP) prediction on the absorbed dataset. In this study, the ssGBLUP and multistep analysis were applied to 5 traits in Duroc pigs, which were slaughter percentage, feed consumption from 40 to 120 kg (FC40_120), days of growth from 40 to 120 kg (D40_120), age at 40 kg (A40) and lean meat percentage. The results showed that MTGP yielded higher accuracy than STGP, which on average was 0.057 higher for the multistep method and 0.045 higher for ssGBLUP. The multistep method achieved similar prediction accuracy as ssGBLUP. However, the prediction bias of the multistep method was in general lower than that of ssGBLUP.

Across-country genomic predictions in Norwegian and New Zealand Composite sheep populations with similar development history.

The goal of this study was to assess the feasibility of across-country genomic predictions in Norwegian White Sheep (NWS) and New Zealand Composite (NZC) sheep populations with similar development history. Different training populations were evaluated (i.e., including only NWS or NZC, or combining both populations). Predictions were performed using the actual phenotypes (normalized) and the single-step GBLUP via Bayesian inference. Genotyped NWS animals born in 2016 (N = 267) were used to assess the accuracy and bias of genomic estimated breeding values (GEBVs) predicted for birth weight (BW), weaning weight (WW), carcass weight (CW), EUROP carcass classification (EUC), and EUROP fat grading (EUF). The accuracy and bias of GEBVs differed across traits and training population used. For instance, the GEBV accuracies ranged from 0.13 (BW) to 0.44 (EUC) for GEBVs predicted including only NWS, from 0.06 (BW) to 0.15 (CW) when including only NZC, and from 0.10 (BW) to 0.41 (EUC) when including both NWS and NZC animals in the training population. The regression coefficients used to assess the spread of GEBVs (bias) ranged from 0.26 (BW) to 0.64 (EUF) for only NWS, 0.10 (EUC) to 0.52 (CW) for only NZC, and from 0.42 (WW) to 2.23 (EUC) for both NWS and NZC in the training population. Our findings suggest that across-country genomic predictions based on ssGBLUP might be possible for NWS and NZC, especially for novel traits.

Genotype calling of triploid offspring from diploid parents.

BACKGROUND: Polyploidy is widespread in animals and especially in plants. Different kinds of ploidies exist, for example, hexaploidy in wheat, octaploidy in strawberries, and diploidy, triploidy, tetraploidy, and pseudo-tetraploidy (partly tetraploid) in fish. Triploid offspring from diploid parents occur frequently in the wild in Atlantic salmon (Salmo salar) and, as with triploidy in general, the triploid individuals are sterile. Induced triploidy in Atlantic salmon is common practice to produce sterile fish. In Norwegian aquaculture, production of sterile triploid fish is an attempt by government and industry to limit genetic introgression between wild and farmed fish. However, triploid fish may have traits and properties that differ from those of diploids. Investigating the genetics behind traits in triploids has proved challenging because genotype calling of genetic markers in triploids is not supported by standard software. Our aim was to develop a method that can be used for genotype calling of genetic markers in triploid individuals. RESULTS: Allele signals were produced for 381 triploid Atlantic salmon offspring using a 56 K Thermo Fisher GeneTitan genotyping platform. Genotypes were successfully called by applying finite normal mixture models to the (transformed) allele signals. Subsets of markers were filtered by quality control statistics for use with downstream analyses. The quality of the called genotypes was sufficient to allow for assignment of diploid parents to the triploid offspring and to discriminate between maternal and paternal parents from autosomal inheritance patterns. In addition, as the maternal inheritance in triploid offspring is identical to gynogenetic inheritance, the maternal recombination pattern for each chromosome could be mapped by using a similar approach as that used in gene-centromere mapping. CONCLUSIONS: We show that calling of dense marker genotypes for triploid individuals is feasible. The resulting genotypes can be used in parentage assignment of triploid offspring to diploid parents, to discriminate between maternal and paternal parents using autosomal inheritance patterns, and to map the maternal recombination pattern using an approach similar to gene-centromere mapping. Genotyping of triploid individuals is important both for selective breeding programs and unravelling the underlying genetics of phenotypes recorded in triploids. In principle, the developed method can be used for genotype calling of other polyploid organisms.

Genomic dissection of maternal, additive and non-additive genetic effects for growth and carcass traits in Nile tilapia.

BACKGROUND: The availability of both pedigree and genomic sources of information for animal breeding and genetics has created new challenges in understanding how they can be best used and interpreted. This study estimated genetic variance components based on genomic information and compared these to the variance components estimated from pedigree alone in a population generated to estimate non-additive genetic variance. Furthermore, the study examined the impact of the assumptions of Hardy-Weinberg equilibrium (HWE) on estimates of genetic variance components. For the first time, the magnitude of inbreeding depression for important commercial traits in Nile tilapia was estimated by using genomic data. RESULTS: The study estimated the non-additive genetic variance in a Nile tilapia population of full-sib families and, when present, it was almost entirely represented by additive-by-additive epistatic variance, although in pedigree studies this non-additive variance is commonly assumed to arise from dominance. For body depth (BD) and body weight at harvest (BWH), the proportion of additive-by-additive epistatic to phenotypic variance was estimated to be 0.15 and 0.17 using genomic data (P < 0.05). In addition, with genomic data, the maternal variance (P < 0.05) for BD, BWH, body length (BL) and fillet weight (FW) explained approximately 10% of the phenotypic variances, which was comparable to pedigree-based estimates. The study also showed the detrimental effects of inbreeding on commercial traits of tilapia, which was estimated to reduce trait values by 1.1, 0.9, 0.4 and 0.3% per 1% increase in the individual homozygosity for FW, BWH, BD and BL, respectively. The presence of inbreeding depression but lack of dominance variance was consistent with an infinitesimal dominance model for the traits. CONCLUSIONS: The benefit of including non-additive genetic effects for genetic evaluations in tilapia breeding schemes is not evident from these findings, but the observed inbreeding depression points to a role for reciprocal recurrent selection. Commercially, this conclusion will depend on the scheme's operational costs and resources. The creation of maternal lines in Tilapia breeding schemes may be a possibility if the variation associated with maternal effects is heritable.

Genomic selection strategies to improve maternal traits in Norwegian White Sheep.

This study tested and compared different implementation strategies for genomic selection for Norwegian White Sheep, aiming to increase genetic gain for maternal traits. These strategies were evaluated for their genetic gain ingrowth, carcass and maternal traits, total genetic gain, a weighted sum of the gain in each trait and rates of inbreeding through a full-scale stochastic simulation. Results showed genomic selection schemes to increase genetic gain for maternal traits but reduced genetic gain for other traits. This could also be obtained by selecting rams for artificial selection at a higher age. Implementation of genomic selection in the current breeding structure increased genetic gain for maternal traits up to 57%, outcompeted by reducing the generation interval for artificial insemination rams from current 3 to 2 years. Then, total genetic gain for maternal traits increased by 65%-77% and total genetic gain by18%-20%, but at increased rates of inbreeding.

Strong selection pressures maintain divergence on genomic islands in Atlantic cod (Gadus morhua L.) populations.

BACKGROUND: Two distinct populations have been extensively studied in Atlantic cod (Gadus morhua L.): the Northeast Arctic cod (NEAC) population and the coastal cod (CC) population. The objectives of the current study were to identify genomic islands of divergence and to propose an approach to quantify the strength of selection pressures using whole-genome single nucleotide polymorphism (SNP) data. After applying filtering criteria, information on 93 animals (9 CC individuals, 50 NEAC animals and 34 CC × NEAC crossbred individuals) and 3,123,434 autosomal SNPs were used. RESULTS: Four genomic islands of divergence were identified on chromosomes 1, 2, 7 and 12, which were mapped accurately based on SNP data and which extended in size from 11 to 18 Mb. These regions differed considerably between the two populations although the differences in the rest of the genome were small due to considerable gene flow between the populations. The estimates of selection pressures showed that natural selection was substantially more important than genetic drift in shaping these genomic islands. Our data confirmed results from earlier publications that suggested that genomic islands are due to chromosomal rearrangements that are under strong selection and reduce recombination between rearranged and non-rearranged segments. CONCLUSIONS: Our findings further support the hypothesis that selection and reduced recombination in genomic islands may promote speciation between these two populations although their habitats overlap considerably and migrations occur between them.

Genetics of resistance to photobacteriosis in gilthead sea bream (Sparus aurata) using 2b-RAD sequencing.

BACKGROUND: Photobacteriosis is an infectious disease developed by a Gram-negative bacterium Photobacterium damselae subsp. piscicida (Phdp), which may cause high mortalities (90-100%) in sea bream. Selection and breeding for resistance against infectious diseases is a highly valuable tool to help prevent or diminish disease outbreaks, and currently available advanced selection methods with the application of genomic information could improve the response to selection. An experimental group of sea bream juveniles was derived from a Ferme Marine de Douhet (FMD, Oléron Island, France) selected line using ~ 109 parents (~ 25 females and 84 males). This group of 1187 individuals represented 177 full-sib families with 1-49 sibs per family, which were challenged with virulent Phdp for a duration of 18 days, and mortalities were recorded within this duration. Tissue samples were collected from the parents and the recorded offspring for DNA extraction, library preparation using 2b-RAD and genotyping by sequencing. Genotypic data was used to develop a linkage map, genome wide association analysis and for the estimation of breeding values. RESULTS: The analysis of genetic variation for resistance against Phdp revealed moderate genomic heritability with estimates of ~ 0.32. A genome-wide association analysis revealed a quantitative trait locus (QTL) including 11 SNPs at linkage group 17 presenting significant association to the trait with p-value crossing genome-wide Bonferroni corrected threshold P ≤ 2.22e-06. The proportion total genetic variance explained by the single top most significant SNP was ranging from 13.28-16.14% depending on the method used to compute the variance. The accuracies of predicting breeding values obtained using genomic vs. pedigree information displayed 19-24% increase when using genomic information. CONCLUSION: The current study demonstrates that SNPs-based genotyping of a sea bream population with 2b-RAD approach is effective at capturing the genetic variation for resistance against Phdp. Prediction accuracies obtained using genomic information were significantly higher than the accuracies obtained using pedigree information which highlights the importance and potential of genomic selection in commercial breeding programs.

Large-scale genomic prediction using singular value decomposition of the genotype matrix.

BACKGROUND: For marker effect models and genomic animal models, computational requirements increase with the number of loci and the number of genotyped individuals, respectively. In the latter case, the inverse genomic relationship matrix (GRM) is typically needed, which is computationally demanding to compute for large datasets. Thus, there is a great need for dimensionality-reduction methods that can analyze massive genomic data. For this purpose, we developed reduced-dimension singular value decomposition (SVD) based models for genomic prediction. METHODS: Fast SVD is performed by analyzing different chromosomes/genome segments in parallel and/or by restricting SVD to a limited core of genotyped individuals, producing chromosome- or segment-specific principal components (PC). Given a limited effective population size, nearly all the genetic variation can be effectively captured by a limited number of PC. Genomic prediction can then be performed either by PC ridge regression (PCRR) or by genomic animal models using an inverse GRM computed from the chosen PC (PCIG). In the latter case, computation of the inverse GRM will be feasible for any number of genotyped individuals and can be readily produced row- or element-wise. RESULTS: Using simulated data, we show that PCRR and PCIG models, using chromosome-wise SVD of a core sample of individuals, are appropriate for genomic prediction in a larger population, and results in virtually identical predicted breeding values as the original full-dimension genomic model (r = 1.000). Compared with other algorithms (e.g. algorithm for proven and young animals, APY), the (chromosome-wise SVD-based) PCRR and PCIG models were more robust to size of the core sample, giving nearly identical results even down to 500 core individuals. The method was also successfully tested on a large multi-breed dataset. CONCLUSIONS: SVD can be used for dimensionality reduction of large genomic datasets. After SVD, genomic prediction using dense genomic data and many genotyped individuals can be done in a computationally efficient manner. Using this method, the resulting genomic estimated breeding values were virtually identical to those computed from a full-dimension genomic model.

Using genomic relationship likelihood for parentage assignment.

BACKGROUND: Parentage assignment is usually based on a limited number of unlinked, independent genomic markers (microsatellites, low-density single nucleotide polymorphisms (SNPs), etc.). Classical methods for parentage assignment are exclusion-based (i.e. based on loci that violate Mendelian inheritance) or likelihood-based, assuming independent inheritance of loci. For true parent-offspring relations, genotyping errors cause apparent violations of Mendelian inheritance. Thus, the maximum proportion of such violations must be determined, which is complicated by variable call- and genotype error rates among loci and individuals. Recently, genotyping using high-density SNP chips has become available at lower cost and is increasingly used in genetics research and breeding programs. However, dense SNPs are not independently inherited, violating the assumptions of the likelihood-based methods. Hence, parentage assignment usually assumes a maximum proportion of exclusions, or applies likelihood-based methods on a smaller subset of independent markers. Our aim was to develop a fast and accurate trio parentage assignment method for dense SNP data without prior genotyping error- or call rate knowledge among loci and individuals. This genomic relationship likelihood (GRL) method infers parentage by using genomic relationships, which are typically used in genomic prediction models. RESULTS: Using 50 simulated datasets with 53,427 to 55,517 SNPs, genotyping error rates of 1-3% and call rates of ~ 80 to 98%, GRL was found to be fast and highly (~ 99%) accurate for parentage assignment. An iterative approach was developed for training using the evaluation data, giving similar accuracy. For comparison, we used the Colony2 software that assigns parentage and sibship simultaneously to increase the power of the likelihood-based method and found that it has considerably lower accuracy than GRL. We also compared GRL with an exclusion-based method in which one of the parameters was estimated using GRL assignments.This method was slightly more accurate than GRL. CONCLUSIONS: We show that GRL is a fast and accurate method of parentage assignment that can use dense, non-independent SNPs, with variable call rates and unknown genotyping error rates. By offering an alternative way of assigning parents, GRL is also suitable for estimating the expected proportion of inconsistent parent-offspring genotypes for exclusion-based models.

Genetic effects of fatty acid composition in muscle of Atlantic salmon.

BACKGROUND: The replacement of fish oil (FO) and fishmeal with plant ingredients in the diet of farmed Atlantic salmon has resulted in reduced levels of the health-promoting long-chain polyunsaturated omega-3 fatty acids (n-3 LC-PUFA) eicosapentaenoic (EPA; 20:5n-3) and docosahexaenoic acid (DHA; 22:6n-3) in their filets. Previous studies showed the potential of selective breeding to increase n-3 LC-PUFA levels in salmon tissues, but knowledge on the genetic parameters for individual muscle fatty acids (FA) and their relationships with other traits is still lacking. Thus, we estimated genetic parameters for muscle content of individual FA, and their relationships with lipid deposition traits, muscle pigmentation, sea lice and pancreas disease in slaughter-sized Atlantic salmon. Our aim was to evaluate the selection potential for increased n-3 LC-PUFA content and provide insight into FA metabolism in Atlantic salmon muscle. RESULTS: Among the n-3 PUFA, proportional contents of alpha-linolenic acid (ALA; 18:3n-3) and DHA had the highest heritability (0.26) and EPA the lowest (0.09). Genetic correlations of EPA and DHA proportions with muscle fat differed considerably, 0.60 and 0.01, respectively. The genetic correlation of DHA proportion with visceral fat was positive and high (0.61), whereas that of EPA proportion with lice density was negative. FA that are in close proximity along the bioconversion pathway showed positive correlations with each other, whereas the start (ALA) and end-point (DHA) of the pathway were negatively correlated (- 0.28), indicating active bioconversion of ALA to DHA in the muscle of fish fed high FO-diet. CONCLUSIONS: Since contents of individual FA in salmon muscle show additive genetic variation, changing FA composition by selective breeding is possible. Taken together, our results show that the heritabilities of individual n-3 LC-PUFA and their genetic correlations with other traits vary, which indicates that they play different roles in muscle lipid metabolism, and that proportional muscle contents of EPA and DHA are linked to body fat deposition. Thus, different selection strategies can be applied in order to increase the content of healthy omega-3 FAin the salmon muscle. We recommend selection for the proportion of EPA + DHA in the muscle because they are both essential FA and because such selection has no clear detrimental effects on other traits.

Variable selection models for genomic selection using whole-genome sequence data and singular value decomposition.

BACKGROUND: Non-linear Bayesian genomic prediction models such as BayesA/B/C/R involve iteration and mostly Markov chain Monte Carlo (MCMC) algorithms, which are computationally expensive, especially when whole-genome sequence (WGS) data are analyzed. Singular value decomposition (SVD) of the genotype matrix can facilitate genomic prediction in large datasets, and can be used to estimate marker effects and their prediction error variances (PEV) in a computationally efficient manner. Here, we developed, implemented, and evaluated a direct, non-iterative method for the estimation of marker effects for the BayesC genomic prediction model. METHODS: The BayesC model assumes a priori that markers have normally distributed effects with probability [Formula: see text] and no effect with probability (1 - [Formula: see text]). Marker effects and their PEV are estimated by using SVD and the posterior probability of the marker having a non-zero effect is calculated. These posterior probabilities are used to obtain marker-specific effect variances, which are subsequently used to approximate BayesC estimates of marker effects in a linear model. A computer simulation study was conducted to compare alternative genomic prediction methods, where a single reference generation was used to estimate marker effects, which were subsequently used for 10 generations of forward prediction, for which accuracies were evaluated. RESULTS: SVD-based posterior probabilities of markers having non-zero effects were generally lower than MCMC-based posterior probabilities, but for some regions the opposite occurred, resulting in clear signals for QTL-rich regions. The accuracies of breeding values estimated using SVD- and MCMC-based BayesC analyses were similar across the 10 generations of forward prediction. For an intermediate number of generations (2 to 5) of forward prediction, accuracies obtained with the BayesC model tended to be slightly higher than accuracies obtained using the best linear unbiased prediction of SNP effects (SNP-BLUP model). When reducing marker density from WGS data to 30 K, SNP-BLUP tended to yield the highest accuracies, at least in the short term. CONCLUSIONS: Based on SVD of the genotype matrix, we developed a direct method for the calculation of BayesC estimates of marker effects. Although SVD- and MCMC-based marker effects differed slightly, their prediction accuracies were similar. Assuming that the SVD of the marker genotype matrix is already performed for other reasons (e.g. for SNP-BLUP), computation times for the BayesC predictions were comparable to those of SNP-BLUP.

Use and optimization of different sources of information for genomic prediction.

BACKGROUND: Molecular data is now commonly used to predict breeding values (BV). Various methods to calculate genomic relationship matrices (GRM) have been developed, with some studies proposing regression of coefficients back to the reference matrix of pedigree-based relationship coefficients (A). The objective was to compare the utility of two GRM: a matrix based on linkage analysis (LA) and anchored to the pedigree, i.e. [Formula: see text] and a matrix based on linkage disequilibrium (LD), i.e. [Formula: see text], using genomic and phenotypic data collected on 5416 broiler chickens. Furthermore, the effects of regressing the coefficients of [Formula: see text] back to A (LDA) and to [Formula: see text] (LDLA) were evaluated, using a range of weighting factors. The performance of the matrices and their composite products was assessed by the fit of the models to the data, and the empirical accuracy and bias of the BV that they predicted. The sensitivity to marker choice was examined by using two chips of equal density but including different single nucleotide polymorphisms (SNPs). RESULTS: The likelihood of models using GRM and composite matrices exceeded the likelihood of models based on pedigree alone and was highest with intermediate weighting factors for both the LDA and LDLA approaches. For these data, empirical accuracies were not strongly affected by the weighting factors, although they were highest when different sources of information were combined. The optimum weighting factors depended on the type of matrices used, as well as on the choice of SNPs from which the GRM were constructed. Prediction bias was strongly affected by the chip used and less by the form of the GRM. CONCLUSIONS: Our findings provide an empirical comparison of the efficacy of pedigree and genomic predictions in broiler chickens and examine the effects of fitting GRM with coefficients regressed back to a reference anchored to the pedigree, either A or [Formula: see text]. For the analysed dataset, the best results were obtained when [Formula: see text] was combined with relationships in A or [Formula: see text], with optimum weighting factors that depended on the choice of SNPs used. The optimum weighting factor for broiler body weight differed from weighting factors that were based on the density of SNPs and theoretically derived using generalised assumptions.

Application of a Bayesian dominance model improves power in quantitative trait genome-wide association analysis.

BACKGROUND: Multi-marker methods, which fit all markers simultaneously, were originally tailored for genomic selection purposes, but have proven to be useful also in association analyses, especially the so-called BayesC Bayesian methods. In a recent study, BayesD extended BayesC towards accounting for dominance effects and improved prediction accuracy and persistence in genomic selection. The current study investigated the power and precision of BayesC and BayesD in genome-wide association studies by means of stochastic simulations and applied these methods to a dairy cattle dataset. METHODS: The simulation protocol was designed to mimic the genetic architecture of quantitative traits as realistically as possible. Special emphasis was put on the joint distribution of the additive and dominance effects of causative mutations. Additive marker effects were estimated by BayesC and additive and dominance effects by BayesD. The dependencies between additive and dominance effects were modelled in BayesD by choosing appropriate priors. A sliding-window approach was used. For each window, the R. Fernando window posterior probability of association was calculated and this was used for inference purpose. The power to map segregating causal effects and the mapping precision were assessed for various marker densities up to full sequence information and various window sizes. RESULTS: Power to map a QTL increased with higher marker densities and larger window sizes. This held true for both methods. Method BayesD had improved power compared to BayesC. The increase in power was between -2 and 8% for causative genes that explained more than 2.5% of the genetic variance. In addition, inspection of the estimates of genomic window dominance variance allowed for inference about the magnitude of dominance at significant associations, which remains hidden in BayesC analysis. Mapping precision was not substantially improved by BayesD. CONCLUSIONS: BayesD improved power, but precision only slightly. Application of BayesD needs large datasets with genotypes and own performance records as phenotypes. Given the current efforts to establish cow reference populations in dairy cattle genomic selection schemes, such datasets are expected to be soon available, which will enable the application of BayesD for association mapping and genomic prediction purposes.

Simultaneous fitting of genomic-BLUP and Bayes-C components in a genomic prediction model.

BACKGROUND: The rapid adoption of genomic selection is due to two key factors: availability of both high-throughput dense genotyping and statistical methods to estimate and predict breeding values. The development of such methods is still ongoing and, so far, there is no consensus on the best approach. Currently, the linear and non-linear methods for genomic prediction (GP) are treated as distinct approaches. The aim of this study was to evaluate the implementation of an iterative method (called GBC) that incorporates aspects of both linear [genomic-best linear unbiased prediction (G-BLUP)] and non-linear (Bayes-C) methods for GP. The iterative nature of GBC makes it less computationally demanding similar to other non-Markov chain Monte Carlo (MCMC) approaches. However, as a Bayesian method, GBC differs from both MCMC- and non-MCMC-based methods by combining some aspects of G-BLUP and Bayes-C methods for GP. Its relative performance was compared to those of G-BLUP and Bayes-C. METHODS: We used an imputed 50 K single-nucleotide polymorphism (SNP) dataset based on the Illumina Bovine50K BeadChip, which included 48,249 SNPs and 3244 records. Daughter yield deviations for somatic cell count, fat yield, milk yield, and protein yield were used as response variables. RESULTS: GBC was frequently (marginally) superior to G-BLUP and Bayes-C in terms of prediction accuracy and was significantly better than G-BLUP only for fat yield. On average across the four traits, GBC yielded a 0.009 and 0.006 increase in prediction accuracy over G-BLUP and Bayes-C, respectively. Computationally, GBC was very much faster than Bayes-C and similar to G-BLUP. CONCLUSIONS: Our results show that incorporating some aspects of G-BLUP and Bayes-C in a single model can improve accuracy of GP over the commonly used method: G-BLUP. Generally, GBC did not statistically perform better than G-BLUP and Bayes-C, probably due to the close relationships between reference and validation individuals. Nevertheless, it is a flexible tool, in the sense, that it simultaneously incorporates some aspects of linear and non-linear models for GP, thereby exploiting family relationships while also accounting for linkage disequilibrium between SNPs and genes with large effects. The application of GBC in GP merits further exploration.

Response to selection while maximizing genetic variance in small populations.

BACKGROUND: Rare breeds represent a valuable resource for future market demands. These populations are usually well-adapted, but their low census compromises the genetic diversity and future of these breeds. Since improvement of a breed for commercial traits may also confer higher probabilities of survival for the breed, it is important to achieve good responses to artificial selection. Therefore, efficient genetic management of these populations is essential to ensure that they respond adequately to genetic selection in possible future artificial selection scenarios. Scenarios that maximize the maximum genetic variance in a unique population could be a valuable option. The aim of this work was to study the effect of the maximization of genetic variance to increase selection response and improve the capacity of a population to adapt to a new environment/production system. RESULTS: We simulated a random scenario (A), a full-sib scenario (B), a scenario applying the maximum variance total (MVT) method (C), a MVT scenario with a restriction on increases in average inbreeding (D), a MVT scenario with a restriction on average individual increases in inbreeding (E), and a minimum coancestry scenario (F). Twenty replicates of each scenario were simulated for 100 generations, followed by 10 generations of selection. Effective population size was used to monitor the outcomes of these scenarios. Although the best response to selection was achieved in scenarios B and C, they were discarded because they are unpractical. Scenario A was also discarded because of its low response to selection. Scenario D yielded less response to selection and a smaller effective population size than scenario E, for which response to selection was higher during early generations because of the moderately structured population. In scenario F, response to selection was slightly higher than in Scenario E in the last generations. CONCLUSIONS: Application of MVT with a restriction on individual increases in inbreeding resulted in the largest response to selection during early generations, but if inbreeding depression is a concern, a minimum coancestry scenario is then a valuable alternative, in particular for a long-term response to selection.

A fast Newton-Raphson based iterative algorithm for large scale optimal contribution selection.

BACKGROUND: The management of genetic variation in a breeding scheme relies very much on the control of the average relationship between selected parents. Optimum contribution selection is a method that seeks the optimum way to select for genetic improvement while controlling the rate of inbreeding. METHODS: A novel iterative algorithm, Gencont2, for calculating optimum genetic contributions was developed. It was validated by comparing it with a previous program, Gencont, on three datasets that were obtained from practical breeding programs in three species (cattle, pig and sheep). The number of selection candidates was 2929, 3907 and 6875 for the pig, cattle and sheep datasets, respectively. RESULTS: In most cases, both algorithms selected the same candidates and led to very similar results with respect to genetic gain for the cattle and pig datasets. In cases, where the number of animals to select varied, the contributions of the additional selected candidates ranged from 0.006 to 0.08 %. The correlations between assigned contributions were very close to 1 in all cases; however, the iterative algorithm decreased the computation time considerably by 90 to 93 % (13 to 22 times faster) compared to Gencont. For the sheep dataset, only results from the iterative algorithm are reported because Gencont could not handle a large number of selection candidates. CONCLUSIONS: Thus, the new iterative algorithm provides an interesting alternative for the practical implementation of optimal contribution selection on a large scale in order to manage inbreeding and increase the sustainability of animal breeding programs.

Within- and across-breed genomic prediction using whole-genome sequence and single nucleotide polymorphism panels.

BACKGROUND: Currently, genomic prediction in cattle is largely based on panels of about 54k single nucleotide polymorphisms (SNPs). However with the decreasing costs of and current advances in next-generation sequencing technologies, whole-genome sequence (WGS) data on large numbers of individuals is within reach. Availability of such data provides new opportunities for genomic selection, which need to be explored. METHODS: This simulation study investigated how much predictive ability is gained by using WGS data under scenarios with QTL (quantitative trait loci) densities ranging from 45 to 132 QTL/Morgan and heritabilities ranging from 0.07 to 0.30, compared to different SNP densities, with emphasis on divergent dairy cattle breeds with small populations. The relative performances of best linear unbiased prediction (SNP-BLUP) and of a variable selection method with a mixture of two normal distributions (MixP) were also evaluated. Genomic predictions were based on within-population, across-population, and multi-breed reference populations. RESULTS: The use of WGS data for within-population predictions resulted in small to large increases in accuracy for low to moderately heritable traits. Depending on heritability of the trait, and on SNP and QTL densities, accuracy increased by up to 31 %. The advantage of WGS data was more pronounced (7 to 92 % increase in accuracy depending on trait heritability, SNP and QTL densities, and time of divergence between populations) with a combined reference population and when using MixP. While MixP outperformed SNP-BLUP at 45 QTL/Morgan, SNP-BLUP was as good as MixP when QTL density increased to 132 QTL/Morgan. CONCLUSIONS: Our results show that, genomic predictions in numerically small cattle populations would benefit from a combination of WGS data, a multi-breed reference population, and a variable selection method.

Identity-by-descent genomic selection using selective and sparse genotyping for binary traits.

BACKGROUND: Genomic selection (GS) allows estimation of the breeding value of individuals, even for non-phenotyped animals. The aim of the study was to examine the potential of identity-by-descent genomic selection (IBD-GS) in genomic selection for a binary, sib-evaluated trait, using different strategies of selective genotyping. This low-cost GS approach is based on linkage analysis of sparse genome-wide marker loci. FINDINGS: Lowly to highly heritable (h(2) = 0.15, 0.30 or 0.60) binary traits with varying incidences (10 to 90%) were simulated for an aquaculture-like population. Genotyping was restricted to the 30% best families according to phenotype, using three genotyping strategies for training sibs. IBD-GS increased genetic gain compared to classical pedigree-based selection; the differences were largest at incidences of 10 to 50% of the desired category (i.e. a relative increase in genetic gain greater than 20%). Furthermore, the relative advantage of IBD-GS increased as the heritability of the trait increased. Differences were small between genotyping strategies, and most of the improvement was achieved by restricting genotyping to sibs with the least common binary phenotype. Genetic gains of IBD-GS relative to pedigree-based models were highest at low to moderate (10 to 50%) incidences of the category selected for, but decreased substantially at higher incidences (80 to 90%). CONCLUSIONS: The IBD-GS approach, combined with sparse and selective genotyping, is well suited for genetic evaluation of binary traits. Genetic gain increased considerably compared with classical pedigree-based selection. Most of the improvement was achieved by selective genotyping of the sibs with the least common (minor) binary category phenotype. Furthermore, IBD-GS had greater advantage over classical pedigree-based models at low to moderate incidences of the category selected for.

Genomic predictions based on animal models using genotype imputation on a national scale in Norwegian Red cattle.

BACKGROUND: In dairy cattle, current genomic predictions are largely based on sire models that analyze daughter yield deviations of bulls, which are derived from pedigree-based animal model evaluations (in a two-step approach). Extension to animal model genomic predictions (AMGP) is not straightforward, because most of the animals that are involved in the genetic evaluation are not genotyped. In single-step genomic best linear unbiased prediction (SSGBLUP), the pedigree-based relationship matrix A and the genomic relationship matrix G are combined in a matrix H, which allows for AMGP. However, as the number of genotyped animals increases, imputation of the genotypes for all animals in the pedigree may be considered. Our aim was to impute genotypes for all animals in the pedigree, construct alternative relationship matrices based on the imputation results, and evaluate the accuracy of the resulting AMGP by cross-validation in the national Norwegian Red dairy cattle population. RESULTS: A large-scale national dataset was effectively handled by splitting it into two sets: (1) genotyped animals and their ancestors (i.e. GA set with 20,918 animals) and (2) the descendants of the genotyped animals (i.e. D set with 4,022,179 animals). This allowed restricting genomic computations to a relatively small set of animals (GA set), whereas the majority of the animals (D set) were added to the animal model equations using Henderson's rules, in order to make optimal use of the D set information. Genotypes were imputed by segregation analysis of a large pedigree with relatively few genotyped animals (3285 out of 20,918). Among the AMGP models, the linkage and linkage disequilibrium based G matrix (G LDLA0 ) yielded the highest accuracy, which on average was 0.06 higher than with SSGBLUP and 0.07 higher than with two-step sire genomic evaluations. CONCLUSIONS: AMGP methods based on genotype imputation on a national scale were developed, and the most accurate method, GLDLA0BLUP, combined linkage and linkage disequilibrium information. The advantage of AMGP over a sire model based on two-step genomic predictions is expected to increase as the number of genotyped cows increases and for species, with smaller sire families and more dam relationships.

Upweighting rare favourable alleles increases long-term genetic gain in genomic selection programs.

BACKGROUND: The short-term impact of using different genomic prediction (GP) models in genomic selection has been intensively studied, but their long-term impact is poorly understood. Furthermore, long-term genetic gain of genomic selection is expected to improve by using Jannink's weighting (JW) method, in which rare favourable marker alleles are upweighted in the selection criterion. In this paper, we extend the JW method by including an additional parameter to decrease the emphasis on rare favourable alleles over the time horizon, with the purpose of further improving the long-term genetic gain. We call this new method dynamic weighting (DW). The paper explores the long-term impact of different GP models with or without weighting methods. METHODS: Different selection criteria were tested by simulating a population of 500 animals with truncation selection of five males and 50 females. Selection criteria included unweighted and weighted genomic estimated breeding values using the JW or DW methods, for which ridge regression (RR) and Bayesian lasso (BL) were used to estimate marker effects. The impacts of these selection criteria were compared under three genetic architectures, i.e. varying numbers of QTL for the trait and for two time horizons of 15 (TH15) or 40 (TH40) generations. RESULTS: For unweighted GP, BL resulted in up to 21.4% higher long-term genetic gain and 23.5% lower rate of inbreeding under TH40 than RR. For weighted GP, DW resulted in 1.3 to 5.5% higher long-term gain compared to unweighted GP. JW, however, showed a 6.8% lower long-term genetic gain relative to unweighted GP when BL was used to estimate the marker effects. Under TH40, both DW and JW obtained significantly higher genetic gain than unweighted GP. With DW, the long-term genetic gain was increased by up to 30.8% relative to unweighted GP, and also increased by 8% relative to JW, although at the expense of a lower short-term gain. CONCLUSIONS: Irrespective of the number of QTL simulated, BL is superior to RR in maintaining genetic variance and therefore results in higher long-term genetic gain. Moreover, DW is a promising method with which high long-term genetic gain can be expected within a fixed time frame.