Search

BackgroundGenomic evaluation, based on the use of thousands of genetic markers in addition to pedigree and phenotype information, has become the standard evaluation methodology in dairy cattle breeding programmes over the past several years. Despite the many differences between dairy cattle breeding and poultry breeding, genomic selection seems very promising for the avian sector, and studies are currently being conducted to optimize avian selection schemes. In this optimization perspective, one of the key parameters is to properly predict the accuracy of genomic evaluation in pure line layers.ResultsIt was observed that genomic evaluation, whether performed on males or females, always proved more accurate than genetic evaluation. The gain was higher when phenotypic information was narrowed, and an augmentation of the size of the reference population led to an increase in accuracy prediction with regard to genomic evaluation. By taking into account the increase of selection intensity and the decrease of the generation interval induced by genomic selection, the expected annual genetic gain would be higher with ancestry-based genomic evaluation of male candidates than with genetic evaluation based on collaterals. This advantage of genomic selection over genetic selection requires more detailed further study for female candidates.ConclusionsIn conclusion, in the population studied, the genomic evaluation of egg quality traits of breeding birds at birth seems to be a promising strategy, at least for the selection of males.

Use of data from crossbred animals for genomic evaluation

Use of data from crossbred animals for a genomic evaluation of pure line layer. XIth European symposium on poultry genetics (ESPG)

To reduce the cost of genomic selection, low density SNP chip can be used in combination with imputation for genotyping the selection candidates instead of using high density (HD) SNP chip. Concurrently, next-generation sequencing (NGS) techniques has been increasingly used in livestock species but remain expensive to be routinely used in selection. An alternative and costefficient solution is the Genotyping by Genome Reducing and Sequencing (GGRS) techniques to sequence only a fraction of the genome by using restriction enzymes. This approach was simulated from sequences of 1027 individuals in a pure layer line, using four enzymes (EcoRI, TaqI, AvaII and PstI). Imputation accuracy on HD genotypes was assessed as the mean correlation between true and imputed genotypes. Egg weight, egg shell color, egg shell strength and albumen height were evaluated with single-step GBLUP methodology. The impact of imputation errors on the genomic estimated breeding values (GEBV) was also investigated.AvaII or PstI led to the detection of more than 10K SNPs in common with the HD SNP chip resulting in imputation accuracy higher than 0.97. The impact of imputation errors on the ranking of the selection candidates was reduced with Spearman correlation (between GEBV calculated on true and imputed genotypes) higher than 0.97 for AvaII and PstI. Finally, the GGRS approach can be an interesting alternative to low density SNP chip for genomic selection. However, with real data, heterogeneity between individuals with missing data has to be taken into account.

The main goal of selection is to choose breeders of the next generation among a set of selection candidates. In genomic selection, the choice of breeders is based on the use of information on DNA polymorphisms, in particular SNP, in addition of performance measures. Since 2013, a commercial high density genotyping chip (600,000 markers) for chicken allowed the implementation of genomic selection in layer and broiler breeding. However, genotyping costs with this chip still remain high for a routine use on a large number of selection candidates. Consequently, it is interesting to develop, at a lower cost, low density genotyping chips. To do so, a set of SNP markers has to be selected to enable an imputation (prediction) of missing genotypes with high accuracy on a high density chip (HD chip). In this perspective, we conducted various simulation studies to choose the optimal strategy for low density genotyping of two different lines of laying hen. Different low density genotyping chips were designed according to two methodologies: a choice of SNP depending on a clustering of SNP based on linkage disequilibrium threshold or a choice of SNP at regular intervals (kb) along each chromosome. Imputation accuracy was assessed as the mean correlation between true and imputed genotypes. Results showed that correlations were more sensitive to false imputation of SNPs with low Minor Allele Frequency (MAF) with the equidistant methodology. Imputation accuracy improved with SNP density and when a higher LD threshold is used for SNP selection. Given the particular structure of the avian genome with chromosomes of very heterogeneous sizes and extents of LD, imputation accuracy differed according to the type of chromosome. All the simulation studies showed that linkage disequilibrium methodology enabled to get better results of imputation than with equidistant methodology.

The main goal of selection is to choose breeders of the next generation among a set of selection candidates. In genomic selection, the choice of breeders rests on the use of information on DNA polymorphisms, in particular SNP, in addition of performance measures. Since 2013, a commercial high density genotyping chip (600,000 markers) for chicken allowed the implementation of genomic selection in layer and broiler breeding. However, genotyping costs with this chip still remain high for a routine use on a large number of selection candidates. Consequently, it is interesting to develop, at a lower cost, low density genotyping chips. To do so, a set of SNP markers has to be selected to enable an imputation (prediction) of missing genotypes on a high density chip (HD chip). This imputation enables to predict missing genotypes of all selection candidates from high density genotyping of a reference population with phenotypes. In this perspective, according to the reference population, various simulation studies were conducted to choose the best strategy for low density genotyping of laying hen lines. Two different low density genotyping chips of 10K SNP were designed according to two methodologies: a choice of SNP depending on a clustering based on linkage disequilibrium threshold or a choice of SNP at regular intervals (kb) along each chromosome. Imputation accuracy was assessed as the mean correlation between true and imputed genotypes. Focusing on populationnal factors that can influence imputation accuracy, it is shown that imputation accuracy improves with an increase in the size of the reference population. By decreasing the kinship degree between reference and candidate population, it is seen that imputation accuracy decreases. Most importantly, results show that a key point in getting good imputations is to have the direct parents in the reference population. Finally, all different genotyping strategies focused on population factors show that linkage disequilibrium methodology enables to get better results of imputation than with equidistant methodology.

Background[br/] The main goal of selection is to achieve genetic gain for a population by choosing the best breeders among a set of selection candidates. Since 2013, the use of a high density genotyping chip (600K Affymetrix (R) Axiom (R) HD genotyping array) for chicken has enabled the implementation of genomic selection in layer and broiler breeding, but the genotyping costs remain high for a routine use on a large number of selection candidates. It has thus been deemed interesting to develop a low density genotyping chip that would induce lower costs. In this perspective, various simulation studies have been conducted to find the best way to select a set of SNPs for low density genotyping of two laying hen lines.[br/] Results[br/] To design low density SNP chips, two methodologies, based on equidistance (EQ) or on linkage disequilibrium (LD) were compared. Imputation accuracy was assessed as the mean correlation between true and imputed genotypes. The results showed correlations more sensitive to false imputation of SNPs having low Minor Allele Frequency (MAF) when the EQ methodology was used. An increase in imputation accuracy was obtained when SNP density was increased, either through an increase in the number of selected windows on a chromosome or through the rise of the LD threshold. Moreover, the results varied depending on the type of chromosome (macro or micro-chromosome). The LD methodology enabled to optimize the number of SNPs, by reducing the SNP density on macro-chromosomes and by increasing it on micro-chromosomes. Imputation accuracy also increased when the size of the reference population was increased. Conversely, imputation accuracy decreased when the degree of kinship between reference and candidate populations was reduced. Finally, adding selection candidates' dams in the reference population, in addition to their sire, enabled to get better imputation results.[br/] Conclusions[br/] Whichever the SNP chip, the methodology, and the scenario studied, highly accurate imputations were obtained, with mean correlations higher than 0.83. The key point to achieve good imputation results is to take into account chicken lines' LD when designing a low density SNP chip, and to include the candidates' direct parents in the reference population.

Design of a low density SNP chip for genomic selection in layer chicken. 10. European Symposium on Poultry Genetics (ESPG)

La sélection génomique utilise des informations sur les polymorphismes de l’ADN, notamment SNP, en complément des mesures de performances, afin de choisir les futurs reproducteurs parmi un ensemble de candidats à la sélection. Depuis 2013, une puce commerciale de génotypages à haute densité (600 000 marqueurs) de l’espèce poule a permis le développement de la sélection génomique dans les filières ponte et chair. Toutefois, les coûts de génotypages avec cette puce restent élevés pour une utilisation en routine sur un grand nombre de candidats à la sélection. Un des enjeux actuels de la sélection génomique est donc le développement de puces de génotypage de basse densité (dites puces LD) à plus faible coût. Il s’agit pour cela de sélectionner un panel de marqueurs SNP permettant une imputation (prédiction) des génotypes manquants sur puce haute densité (puce HD). Dans cette optique, différentes études ont été menées afin de choisir la stratégie de génotypages basse densité la mieux adaptée à une lignée de poule pondeuse. Différentes densités de génotypages, plusieurs scenarii de populations et deux logiciels ont été testés. Les puces basse densité ont été construites selon deux stratégies : choix des SNP selon un clustering basé sur un seuil de déséquilibre de liaison ou choix de SNP à intervalles réguliers. Trois critères d’efficacité des imputations ont été comparés : taux d’erreur génotypique, taux d’erreur allélique et corrélation entre génotypes imputés et réels. Le taux d’erreur génotypique s’est révélé un critère plus discriminant que les autres. Le logiciel FImpute apparaît plus intéressant que le logiciel Beagle car c’est un meilleur compromis entre efficacité de l’imputation et temps de calcul. L’étude des différentes puces basse densité montre que les imputations sont meilleures lorsque la densité des SNP augmente, lorsque le seuil de DL augmente et lorsque des SNP avec une faible MAF (Fréquence Allélique Mineure) sont pris en compte. Les études de populations montrent que les imputations sont meilleures lorsque la taille de la population de référence augmente. L’influence du degré d’apparentement entre population de référence et candidats ainsi que l’impact sur la précision des évaluations génomiques restent à approfondir., Genomic selection uses information on DNA polymorphisms, in particular SNP, in addition of performance measures, in order to choose breeders of the next generation among a set of selection candidates. Since 2013, a commercial high density genotyping chip (600000 markers) for chicken allowed the implementation of genomic selection in layer and broiler breeding. However, genotyping costs with this chip still remain high for a routine use on a large number of selection candidates. Consequently, one of the current stakes of genomic selection is the development at a lower cost of low density genotyping chips (LD chips). To do so, a set of SNP markers has to be selected to enable an imputation (prediction) of missing genotypes on a high density chip (HD chip). [br/] In this regard, various studies were conducted to choose the low density genotyping strategy the best suited to laying hens line. Different genotyping densities, several population scenarii and two software were tested. Low density genotyping chips were built according to two strategies: a choice of SNP depending on a clustering based on linkage disequilibrium threshold or a choice of SNP at regular intervals. Three criteria of imputation accuracy were compared: genotypic error rate, allelic error rate and correlation. The genotypic error rate appeared to be a criterion more discriminating than the others. FImpute software is preferred to Beagle software because FImpute is a better compromise between imputation accuracy and calculation time. The study of different low density genotyping chips shows that imputations improve with SNP density, when the LD threshold increases and when SNP with low MAF (Minor Allele Frequency) are considered. Population studies show that imputations are better when the size of the reference population increases. The influence of the kinship degree between reference population and target population, and the impact on accuracy of genomic evaluations still remain to be deepened.

Key message We propose a criterion to predict genomic selection efficiency for structured populations. This criterion is useful to define optimal calibration set and to estimate prediction reliability for multiparental populations. Abstract Genomic selection refers to the use of genotypic information for predicting the performance of selection candidates. It has been shown that prediction accuracy depends on various parameters including the composition of the calibration set (CS). Assessing the level of accuracy of a given prediction scenario is of highest importance because it can be used to optimize CS sampling before collecting phenotypes, and once the breeding values are predicted it informs the breeders about the reliability of these predictions. Different criteria were proposed to optimize CS sampling in highly diverse panels, which can be useful to screen collections of genotypes. But plant breeders often work on structured material such as biparental or multiparental populations, for which these criteria are less adapted. We derived from the generalized coefficient of determination (CD) theory different criteria to optimize CS sampling and to assess the reliability associated to predictions in structured populations. These criteria were evaluated on two nested association mapping (NAM) populations and two highly diverse panels of maize. They were efficient to sample optimized CS in most situations. They could also estimate at least partly the reliability associated to predictions between NAM families, but they could not estimate differences in the reliability associated to the predictions of NAM families using the highly diverse panels as calibration sets. We illustrated that the CD criteria could be adapted to various prediction scenarios including inter and intra-family predictions, resulting in higher prediction accuracies. Electronic supplementary material The online version of this article (doi:10.1007/s00122-017-2956-7) contains supplementary material, which is available to authorized users.

International audience; Genomic selection is promising for plant breeding, particularly for perennial crops. Multivariate analysis, which considers several traits jointly, takes advantage of the genetic correlations to increase accuracy. The aim of this study was to empirically evaluate the potential of a univariate and multivariate genomic mixed model (G-BLUP) compared to the traditional univariate pedigree-based BLUP (T-BLUP) when analyzing progeny tests of oil palm, the world’s major oil crop. The dataset comprised 478 crosses between two heterotic groups, A and B, with 140 and 131 parents, respectively, genotyped with 313 simple sequence repeat markers. The traits were bunch number and average bunch weight. We found that G-BLUP with a genomic matrix based on a similarity index gave a higher likelihood than T-BLUP. In addition, multivariate G-BLUP improved the accuracy of additive effects (breeding values or general combining abilities, GCAs), in particular for the less heritable trait, and of dominance effects (specific combining abilities, SCAs). The average increase in accuracy was 22.5 % for GCAs and 18.7 % for SCAs. Using 160 markers in group A and 90 in group B was enough to reach maximum GCA prediction accuracy. The contrasted history of the parental groups likely explained the higher benefit of G-BLUP over T-BLUP for group A than for group B. Finally, G-BLUP should be used instead of T-BLUP to analyze oil palm progeny tests, with a multivariate approach for correlated traits. G-BLUP will allow breeders to consider SCAs in addition to GCAs when selecting between the progeny-tested parents.