ArticlePublisher preview available

Realized genetic gain with reciprocal recurrent selection in a Eucalyptus breeding program

Authors:
Saulo Chaves at Escola Superior de Agricultura "Luiz de Queiroz"
Saulo Chaves
LAS Dias at Federal University of Viçosa
LAS Dias
Rodrigo S. Alves
Rodrigo S. Alves
Rodrigo S. Alves
  • This person is not on ResearchGate, or hasn't claimed this research yet.
Filipe Manoel Ferreira at São Paulo State University
Filipe Manoel Ferreira
Show all 11 authorsHide
Read publisher preview
Download citation
To read the full-text of this research, you can request a copy directly from the authors.

Abstract and Figures

Key message:Eucalyptusbreeding can benefit from strategies that capture dominance effects, as shown by the improvement in mean annual increment of wood volume across cycles of RRS. Abstract: There is no empirical validation of reciprocal recurrent selection (RRS) in Eucalyptus breeding. Our study helps to fill this gap by quantifying the realized response to selection achieved after two cycles of RRS involving Eucalyptus urophylla and E. grandis. We also investigated the selection effects on the genetic parameters of the breeding populations. We evaluated 25 trials of the first cycle (C1) of RRS and 12 trials of the second cycle (C2) of RRS. These trials were established in two different regions, separated according to altitude. Fitting linear mixed models enabled the estimation of variance components and the prediction of mean components (general and specific hybridizing abilities). The realized response to selection was calculated as the difference between the mean of the predicted genotypic values of the C1 and C2. The RRS effectively improved the mean annual increment of wood volume by 28.5% in the high-altitude region and 12.3% in the low-altitude region from the C1 to C2. The genetic variability also increased as a result of the new genotypes that arose through recombination. These findings provide insights for decision-making and reinforce that Eucalyptus breeding can benefit from strategies that capture dominance effects.
Reciprocal recurrent selection (RRS) scheme evaluated in this study. The legends on the bottom and top indicate the cycle and the time (in years) of each phase in the breeding pipeline, respectively. The program started with the evaluation of pure-species progenies in pure-progeny trials, where the best trees were selected for intercrossing. The resulting hybrids were evaluated in hybrid-progeny trials, which were used to select the best pure-species parents based on their GHA. The selected parents of each species were recombined to form the population of the next cycle. Then, the process repeats for each cycle of RRS
Reciprocal recurrent selection (RRS) scheme evaluated in this study. The legends on the bottom and top indicate the cycle and the time (in years) of each phase in the breeding pipeline, respectively. The program started with the evaluation of pure-species progenies in pure-progeny trials, where the best trees were selected for intercrossing. The resulting hybrids were evaluated in hybrid-progeny trials, which were used to select the best pure-species parents based on their GHA. The selected parents of each species were recombined to form the population of the next cycle. Then, the process repeats for each cycle of RRS
… 
Heatmaps depicting the distribution of Eucalyptus grandis and Eucalyptus urophylla parents across trials of the first RRS cycle (a and b, represent the high-altitude region and c and d illustrate the low-altitude region, respectively), and of the second RRS cycle (e and f illustrate the high-altitude region and g and h represent the low-altitude region). Intense-colored cells indicate the presence of the corresponding parent in the x-axis in the trial of the y-axis, and light-colored cells, the absence. The numbers in the x- and y-axes represent the number of trials and parents in that particular region and cycle, respectively. The percentage of filled cells, in the top right of each heatmap, represents the degree of connectivity between trials
Heatmaps depicting the distribution of Eucalyptus grandis and Eucalyptus urophylla parents across trials of the first RRS cycle (a and b, represent the high-altitude region and c and d illustrate the low-altitude region, respectively), and of the second RRS cycle (e and f illustrate the high-altitude region and g and h represent the low-altitude region). Intense-colored cells indicate the presence of the corresponding parent in the x-axis in the trial of the y-axis, and light-colored cells, the absence. The numbers in the x- and y-axes represent the number of trials and parents in that particular region and cycle, respectively. The percentage of filled cells, in the top right of each heatmap, represents the degree of connectivity between trials
… 
Distribution of the ratios of genetic effects’ variance estimates to total variance: general hybridizing ability (GHA) of Eucalyptus grandis, GHA of E. urophylla, specific hybridizing ability (SHA), and total genotypic variance among hybrid families in individual trials of the first and second reciprocal recurrent selection cycles in both high- and low-altitude regions for the mean annual increment of wood volume
Distribution of the ratios of genetic effects’ variance estimates to total variance: general hybridizing ability (GHA) of Eucalyptus grandis, GHA of E. urophylla, specific hybridizing ability (SHA), and total genotypic variance among hybrid families in individual trials of the first and second reciprocal recurrent selection cycles in both high- and low-altitude regions for the mean annual increment of wood volume
… 
Mean of the ratios of genetic effects’ variances estimates to total variance: general hybridizing ability (GHA) of Eucalyptus grandis, GHA of E. urophylla, specific hybridizing ability (SHA), and total genotypic variance among hybrid families across trials in the same reciprocal recurrent selection cycle, in the high- and low-altitude regions for the mean annual increment of wood volume
Mean of the ratios of genetic effects’ variances estimates to total variance: general hybridizing ability (GHA) of Eucalyptus grandis, GHA of E. urophylla, specific hybridizing ability (SHA), and total genotypic variance among hybrid families across trials in the same reciprocal recurrent selection cycle, in the high- and low-altitude regions for the mean annual increment of wood volume
… 
Distribution of the genotypic values of the mean annual increment of wood volume of the “urograndis” hybrids in each cycle in the high-altitude and in the low-altitude regions. The caption in the upper right of each plot has the overall mean of the genotypic values in each cycle. Below each plot, the difference between cycles (Δ\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\Delta $$\end{document}), in percentage, is depicted. The vertical lines in the density plots represent the mean genotypic value of each cycle
+1
Distribution of the genotypic values of the mean annual increment of wood volume of the “urograndis” hybrids in each cycle in the high-altitude and in the low-altitude regions. The caption in the upper right of each plot has the overall mean of the genotypic values in each cycle. Below each plot, the difference between cycles (Δ\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\Delta $$\end{document}), in percentage, is depicted. The vertical lines in the density plots represent the mean genotypic value of each cycle
… 
Figures - available from: Tree Genetics & Genomes
This content is subject to copyright. Terms and conditions apply.
A preview of this full-text is provided by Springer Nature.
Learn more
Content available from Tree Genetics & Genomes
This content is subject to copyright. Terms and conditions apply.
https://doi.org/10.1007/s11295-024-01678-2
ORIGINAL ARTICLE
Realized genetic gain with reciprocal recurrent selection
in a Eucalyptus breeding program
Saulo F. S. Chaves1·Luiz A. S. Dias1·Rodrigo S. Alves2·Filipe M. Ferreira3·Maurício S. Araújo4·
Marcos D. V. Resende5,6 ·Elizabete K. Takahashi2·João E. Souza2·Fernando P. Leite2·Samuel B. Fernandes7·
Kaio Olimpio G. Dias8
Received: 30 November 2023 / Revised: 21 August 2024 / Accepted: 13 October 2024
© The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2024
Abstract
Key message: Eucalyptus breeding can benefit from strategies that capture dominance effects, as shown by the improve-
ment in mean annual increment of wood volume across cycles of RRS.
Abstract: There is no empirical validation of reciprocal recurrent selection (RRS) in Eucalyptus breeding. Our study helps to
fill this gap by quantifying the realized response to selection achieved after two cycles of RRS involving Eucalyptus urophylla
and E. grandis. We also investigated the selection effects on the genetic parameters of the breeding populations. We evaluated
25 trials of the first cycle (C1) of RRS and 12 trials of the second cycle (C2) of RRS. These trials were established in two
different regions, separated according to altitude. Fitting linear mixed models enabled the estimation of variance components
and the prediction of mean components (general and specific hybridizing abilities). The realized response to selection was cal-
culated as the difference between the mean of the predicted genotypic values of the C1 and C2. The RRS effectively improved
the mean annual increment of wood volume by 28.5% in the high-altitude region and 12.3% in the low-altitude region from
the C1 to C2. The genetic variability also increased as a result of the new genotypes that arose through recombination. These
findings provide insights for decision-making and reinforce that Eucalyptus breeding can benefit from strategies that capture
dominance effects.
Keywords Tree breeding ·Heterosis ·Response to selection ·Genetic parameters ·Eucalyptus grandis ·Eucalyptus urophylla
Introduction
Eucalyptus L’Hér breeding programs aim to release clones
for forest-based industries while increasing the frequency
of favorable alleles in the breeding population (Rezende
et al. 2014). Recurrent selection strategies allow breed-
ers to simultaneously select high-performance candidates
(testing/selection step) while keeping the decay of genetic
variability between cycles at low levels (recombination step)
(Kerr et al. 2004; Souza Jr. 2015). For successfully reaching
these objectives, two factors play a major role: the genetic
architecture of the traits under selection and the effective
population size required to initiate (founders) and maintain
(selected genotypes from one cycle to another) the breeding
program in the medium and long terms (Isik and McKeand
2019).
Communicated by C. Kulheim.
Extended author information available on the last page of the article
Recurrent selection in Eucalyptus breeding programs can
be conducted either within a single population or by employ-
ing two populations (Assis and Resende 2011). The latter
strategy is known as reciprocal recurrent selection (RRS)
(Comstock et al. 1949). RRS is particularly suitable for
hybrid breeding, where the focus is not only on increasing the
additive value of breeding populations but also on exploring
the dominance effects via heterosis (Resende and Barbosa
2005; Covarrubias-Pazaran et al. 2023; Labroo et al. 2023).
The concept of heterosis was first defined by Shull (1908)
as the increased vigour of the heterozygote compared to
their homozygous parents. This concept was later outlined
as “baseline heterosis” in the population level (Lamkey and
Edwards 1999). However, in Eucalyptus breeding, heterosis
is achieved by crossing trees - which rarely are inbred - from
two different species, deviating from the classical concept
outlined by Shull (1908). Instead, Eucalyptus breeding RRS
explores the mid-parent panmictic heterosis, defined as the
difference between the value of the interspecific hybrid popu-
lation and the mean value of the two panmictic pure-species
0123456789().: V,-vol 123
Tree Genetics & Genomes (2024) 20:47
/ Published online: 5 November 2024
Content courtesy of Springer Nature, terms of use apply. Rights reserved.
ResearchGate has not been able to resolve any citations for this publication.
Models to estimate genetic gain of soybean seed yield from annual multi-environment field trials
Article
Full-text available
  • Nov 2023
  • THEOR APPL GENET
Simulations demonstrated that estimates of realized genetic gain from linear mixed models using regional trials are biased to some degree. Thus, we recommend multiple selected models to obtain a range of reasonable estimates. Genetic improvements of discrete characteristics are obvious and easy to demonstrate, while quantitative traits require reliable and accurate methods to disentangle the confounding genetic and non-genetic components. Stochastic simulations of soybean [Glycine max (L.) Merr.] breeding programs were performed to evaluate linear mixed models to estimate the realized genetic gain (RGG) from annual multi-environment trials (MET). True breeding values were simulated under an infinitesimal model to represent the genetic contributions to soybean seed yield under various MET conditions. Estimators were evaluated using objective criteria of bias and linearity. Covariance modeling and direct versus indirect estimation-based models resulted in a substantial range of estimated values, all of which were biased to some degree. Although no models produced unbiased estimates, the three best-performing models resulted in an average bias of ± 7.41\pm\, 7.41 kg/ha−1^{-1}/yr−1^{-1} (± 0.11\pm\, 0.11 bu/ac−1^{-1}/yr−1^{-1}). Rather than relying on a single model to estimate RGG, we recommend the application of several models with minimal and directional bias. Further, based on the parameters used in the simulations, we do not think it is appropriate to use any single model to compare breeding programs or quantify the efficiency of proposed new breeding strategies. Lastly, for public soybean programs breeding for maturity groups II and III in North America, the estimated RGG values ranged from 18.16 to 39.68 kg/ha−1^{-1}/yr−1^{-1} (0.27–0.59 bu/ac−1^{-1}/yr−1^{-1}) from 1989 to 2019. These results provide strong evidence that public breeders have significantly improved soybean germplasm for seed yield in the primary production areas of North America.
View
Realized genomic selection across generations in a reciprocal recurrent selection breeding program of Eucalyptus hybrids
Article
Full-text available
  • Oct 2023
Introduction Genomic selection (GS) experiments in forest trees have largely reported estimates of predictive abilities from cross-validation among individuals in the same breeding generation. In such conditions, no effects of recombination, selection, drift, and environmental changes are accounted for. Here, we assessed the effectively realized predictive ability (RPA) for volume growth at harvest age by GS across generations in an operational reciprocal recurrent selection (RRS) program of hybrid Eucalyptus. Methods Genomic best linear unbiased prediction with additive (GBLUP_G), additive plus dominance (GBLUP_G+D), and additive single-step (HBLUP) models were trained with different combinations of growth data of hybrids and pure species individuals (N = 17,462) of the G1 generation, 1,944 of which were genotyped with ~16,000 SNPs from SNP arrays. The hybrid G2 progeny trial (HPT267) was the GS target, with 1,400 selection candidates, 197 of which were genotyped still at the seedling stage, and genomically predicted for their breeding and genotypic values at the operational harvest age (6 years). Seedlings were then grown to harvest and measured, and their pedigree-based breeding and genotypic values were compared to their originally predicted genomic counterparts. Results Genomic RPAs ≥0.80 were obtained as the genetic relatedness between G1 and G2 increased, especially when the direct parents of selection candidates were used in training. GBLUP_G+D reached RPAs ≥0.70 only when hybrid or pure species data of G1 were included in training. HBLUP was only marginally better than GBLUP. Correlations ≥0.80 were obtained between pedigree and genomic individual ranks. Rank coincidence of the top 2.5% selections was the highest for GBLUP_G (45% to 60%) compared to GBLUP_G+D. To advance the pure species RRS populations, GS models were best when trained on pure species than hybrid data, and HBLUP yielded ~20% higher predictive abilities than GBLUP, but was not better than ABLUP for ungenotyped trees. Discussion We demonstrate that genomic data effectively enable accurate ranking of eucalypt hybrid seedlings for their yet-to-be observed volume growth at harvest age. Our results support a two-stage GS approach involving family selection by average genomic breeding value, followed by within-top-families individual GS, significantly increasing selection intensity, optimizing genotyping costs, and accelerating RRS breeding.
View
Genomic prediction in hybrid breeding: II. Reciprocal recurrent genomic selection with full-sib and half-sib families
Article
Full-text available
  • Aug 2023
  • THEOR APPL GENET
Key message Genomic prediction of GCA effects based on model training with full-sib rather than half-sib families yields higher short- and long-term selection gain in reciprocal recurrent genomic selection for hybrid breeding, if SCA effects are important. Abstract Reciprocal recurrent genomic selection (RRGS) is a powerful tool for ensuring sustainable selection progress in hybrid breeding. For training the statistical model, one can use half-sib (HS) or full-sib (FS) families produced by inter-population crosses of candidates from the two parent populations. Our objective was to compare HS-RRGS and FS-RRGS for the cumulative selection gain (ΣΔGΣΔG\Sigma \Delta G), the genetic, GCA and SCA variances (σG2σG2\sigma_{G}^{2},σgca2σgca2\sigma_{gca}^{2}, σsca2σsca2\sigma_{sca}^{2}) of the hybrid population, and prediction accuracy (rgcargcar_{gca}) for GCA effects across cycles. Using SNP data from maize and wheat, we simulated RRGS programs over 10 cycles, each consisting of four sub-cycles with genomic selection of Ne=20Ne=20N_{e} = 20 out of 950 candidates in each parent population. Scenarios differed for heritability h2(h2)\left( {h^{2} } \right) and the proportion τ=100×σsca2:σG2τ=100×σsca2:σG2\tau = 100 \times \sigma_{sca}^{2} :\sigma_{G}^{2} of traits, training set (TS) size (NTSNTSN_{TS}), and maize vs. wheat. Curves of ΣΔGΣΔG\Sigma \Delta G over selection cycles showed no crossing of both methods. If ττ\tau was high, ΣΔGΣΔG\Sigma \Delta G was generally higher for FS-RRGS than HS-RRGS due to higher rgcargcar_{gca}. In contrast, HS-RRGS was superior or on par with FS-RRGS, if ττ\tau or h2h2h^{2} and NTSNTSN_{TS} were low. ΣΔGΣΔG\Sigma \Delta G showed a steeper increase and higher selection limit for scenarios with low ττ\tau, high h2h2h^{2} and large NTSNTSN_{TS}. σgca2σgca2\sigma_{gca}^{2} and even more so σsca2σsca2\sigma_{sca}^{2} decreased rapidly over cycles for both methods due to the high selection intensity and the role of the Bulmer effect for reducing σgca2σgca2\sigma_{gca}^{2}. Since the TS for FS-RRGS can additionally be used for hybrid prediction, we recommend this method for achieving simultaneously the two major goals in hybrid breeding: population improvement and cultivar development.
View
Clonal diploid and autopolyploid breeding strategies to harness heterosis: insights from stochastic simulation
Article
Full-text available
  • Jun 2023
  • THEOR APPL GENET
Key message Reciprocal recurrent selection sometimes increases genetic gain per unit cost in clonal diploids with heterosis due to dominance, but it typically does not benefit autopolyploids. Abstract Breeding can change the dominance as well as additive genetic value of populations, thus utilizing heterosis. A common hybrid breeding strategy is reciprocal recurrent selection (RRS), in which parents of hybrids are typically recycled within pools based on general combining ability. However, the relative performances of RRS and other breeding strategies have not been thoroughly compared. RRS can have relatively increased costs and longer cycle lengths, but these are sometimes outweighed by its ability to harness heterosis due to dominance. Here, we used stochastic simulation to compare genetic gain per unit cost of RRS, terminal crossing, recurrent selection on breeding value, and recurrent selection on cross performance considering different amounts of population heterosis due to dominance, relative cycle lengths, time horizons, estimation methods, selection intensities, and ploidy levels. In diploids with phenotypic selection at high intensity, whether RRS was the optimal breeding strategy depended on the initial population heterosis. However, in diploids with rapid-cycling genomic selection at high intensity, RRS was the optimal breeding strategy after 50 years over almost all amounts of initial population heterosis under the study assumptions. Diploid RRS required more population heterosis to outperform other strategies as its relative cycle length increased and as selection intensity and time horizon decreased. The optimal strategy depended on selection intensity, a proxy for inbreeding rate. Use of diploid fully inbred parents vs. outbred parents with RRS typically did not affect genetic gain. In autopolyploids, RRS typically did not outperform one-pool strategies regardless of the initial population heterosis.
View
Reciprocal recurrent selection based on genetic complementation: An efficient way to build heterosis in diploids due to directional dominance
Article
Full-text available
Depending on the trait architecture and reproduction system, selection strategies in plant breeding focus on the accumulation of additive, dominance effects, or both. Innovation in the exploitation of dominance‐effect‐based heterosis has been limited since the proposal of general combining ability (GCA)‐based approaches. We propose the use of a new surrogate of genetic complementation between genetic pools to increase accumulation of dominance effects and heterosis. We simulated breeding programs to show how reciprocal recurrent selection (RRS) by genetic complementation would build the dominance‐based heterosis cheaper than GCA‐based approaches and used real phenotypic data from hybrid maize (Zea mays) to demonstrate the underlying concepts. We found RRS by genetic complementation to be an attractive and viable strategy to exploit dominance, build de novo heterotic pools, and enhance the current GCA‐based approaches. If demonstrated in practice, we hypothesized that this approach would lower the cost of hybrid breeding drastically and contribute to food security.
View
Quantitative Genetics of a Hybrid Population of Eucalyptus nitens × Eucalyptus globulus: Estimation of Genetic Parameters and Implications for Breeding Strategies
Article
Full-text available
  • Feb 2023
In Chile, interspecific hybrids between Eucalyptus nitens × Eucalyptus globulus (GloNi) were developed by Arauco Forestry to capture specific traits from each parental species: growth rate and cold resistance from E. nitens (NIT) and wood properties from E. globulus (GLO). Field tests of E. nitens × E. globulus were distributed in two geographic zones: Arauco (12 tests) and Valdivia (15 tests), where growth and wood properties measurements were recorded at different ages. The hybrid population is composed of clones from 28 full-sib families, being the result of crossing 12 E. nitens females and 8 E. globulus males. Progeny from each of these families were vegetatively propagated and tested on each growth zone, with a total of 1214 clones developed. The quantitative genetic parameter estimates reveal high genetic variation in hybrid volume gain and wood properties, which make possible large genetic gain in all traits analyzed. Additionally, E. nitens has a considerable impact on the volume gain of the hybrid, making it important to test more parents in future interspecific crosses in both hybrid zones. In contrast, E. globulus demonstrated zero impact in volume. In wood traits, E. globulus in Arauco zone demonstrates a large effect on the genetic variability of these traits; meanwhile, in the Valdivia zone, E. nitens and E. globulus parents contributed roughly similar amounts of genetic variation. The high General Hybridizing Ability (GHA) and General Combining Ability (GCA) relationship between hybrid progeny and pure species progeny performance indicates that parents could be selected for interspecific crosses based on pure species test results for volume and wood properties.
View
Application of linear mixed models for multiple harvest/site trial analyses in perennial plant breeding
Article
Full-text available
  • Dec 2022
  • TREE GENET GENOMES
The optimization of perennial plant breeding necessarily involves the evaluation of multi-harvest and/or multi-site trials. In these situations, modeling covariance structures can elevate accuracy. This study aimed to evaluate different covariance structures for multi-harvest and multi-site trial analyses, using two datasets (D1 and D2). In D1, 25 hybrids of Theobroma grandiflorum were evaluated in a complete randomized block design, during twelve consecutive harvest years. In D2, 215 clones of Eucalyptus spp. were evaluated in a complete randomized block design, in four sites. For both datasets, the covariance structures of the random effects were modeled, and their adequacy was tested by the Akaike and Bayesian information criteria. From the selected model, the variance components and genetic parameters were estimated. We also compared the expected genetic gains and the rankings of genotypes based on the genotypic values provided by the basic and the selected models. For D1, the third-order factor analytic model was the most suitable for genetic effects, while for D2, the unstructured model showed the best fit for such effects. The models provided a better insight into the variances dynamics over the harvest years/sites. The genetic gains were 3.52 percentage points higher in D1 and did not change in D2. Despite similar results, the standard model, modeled with covariance structures that assume homogeneity of co-variances, was not the most statistically appropriate model for D2 according to the information criteria. Therefore, the modeling of covariance structures can and should be used in the genetic evaluation of perennial plants.
View
Multibreed genomic evaluation for production traits of dairy cattle in the United States using single-step genomic best linear unbiased predictor
Article
Full-text available
  • Mar 2022
  • J DAIRY SCI
Official multibreed genomic evaluations for dairy cattle in the United States are based on multibreed BLUP evaluation followed by single-breed estimation of SNP effects. Single-step genomic BLUP (ssGBLUP) allows the straight computation of genomic (G)EBV in a multibreed context. This work aimed to develop ssGBLUP multibreed genomic predictions for US dairy cattle using the algorithm for proven and young (APY) to compute the inverse of the genomic relationship matrix. Only purebred Ayrshire (AY), Brown Swiss (BS), Guernsey (GU), Holstein (HO), and Jersey (JE) animals were considered. A 3-trait model with milk (MY), fat (FY), and protein (PY) yields was applied using about 45 million phenotypes recorded from January 2000 to June 2020. The whole data set included about 29.5 million animals, of which almost 4 million were genotyped. All the effects in the model were breed specific, and breed was also considered as fixed unknown parent groups. Evaluations were done for (1) each single breed separately (single); (2) HO and JE together (HO_JE); (3) AY, BS, and GU together (AY_BS_GU); (4) all the 5 breeds together (5_BREEDS). Initially, 15k core animals were used in APY for AY_BS_GU and 5_BREEDS, but larger core sets with more animals from the least represented breeds were also tested. The HO_JE evaluation had a fixed set of 30k core animals, with an equal representation of the 2 breeds, whereas HO and JE single-breed analysis involved 15k core animals. Validation for cows was based on correlations between adjusted phenotypes and (G)EBV, whereas for bulls on the regression of daughter yield deviations on (G)EBV. Because breed was correctly considered in the model, BLUP results for single and multibreed analyses were the same. Under ssGBLUP, predictability and reliability for AY, BS, and GU were on average 7% and 2% lower in 5_BREEDS compared with single-breed evaluations, respectively. However, validation parameters for these 3 breeds became better than in the single-breed evaluations when 45k animals were included in the core set for 5_BREEDS. Evaluations for Holsteins were more stable across scenarios because of the greatest number of genotyped animals and amount of data. Combining AY, BS, and GU into one evaluation resulted in predictions similar to the ones from single breed, especially when using about 30k core animals in APY. The results showed that single-step large-scale multibreed evaluations are computationally feasible, but fine tuning is needed to avoid a reduction in reliability when numerically dominant breeds are combined. Having evaluations for AY, BS, and GU separated from HO and JE may reduce inflation of GEBV for the first 3 breeds.
View
Computing strategies for multi-population genomic evaluation
Article
Full-text available
  • Feb 2022
  • GENET SEL EVOL
Background Multiple breed evaluation using genomic prediction includes the use of data from multiple populations, or from parental breeds and crosses, and is expected to lead to better genomic predictions. Increased complexity comes from the need to fit non-additive effects such as dominance and/or genotype-by-environment interactions. In these models, marker effects (and breeding values) are modelled as correlated between breeds, which leads to multiple trait formulations that are based either on markers [single nucleotide polymorphism best linear unbiased prediction (SNP-BLUP)] or on individuals [genomic(G)BLUP)]. As an alternative, we propose the use of generalized least squares (GLS) followed by backsolving of marker effects using selection index (SI) theory. Results All investigated options have advantages and inconveniences. The SNP-BLUP yields marker effects directly, which are useful for indirect prediction and for planned matings, but is very large in number of equations and is structured in dense and sparse blocks that do not allow for simple solving. GBLUP uses a multiple trait formulation and is very general, but results in many equations that are not used, which increase memory needs, and is also structured in dense and sparse blocks. An alternative formulation of GBLUP is more compact but requires tailored programming. The alternative of solving by GLS + SI is the least consuming, both in number of operations and in memory, and it uses only single dense blocks. However, it requires dedicated programming. Computational complexity problems are exacerbated when more than additive effects are fitted, e.g. dominance effects or genotype x environment interactions. Conclusions As multi-breed predictions become more frequent and non-additive effects are more often included, standard equations for genomic prediction based on Henderson’s mixed model equations become less practical and may need to be replaced by more efficient (although less general) approaches such as the GLS + SI approach proposed here.
View
View