Genomic Prediction Approaches Research Articles

Application of deep learning with bivariate models for genomic prediction of sow lifetime productivity-related traits.

Pig breeders cannot obtain phenotypic information at the time of selection for sow lifetime productivity (SLP). They would benefit from obtaining genetic information of candidate sows. Genomic data interpreted using deep learning (DL) techniques could contribute to the genetic improvement of SLP to maximize farm profitability because DL models capture nonlinear genetic effects such as dominance and epistasis more efficiently than conventional genomic prediction methods based on linear models. This study aimed to investigate the usefulness of DL for the genomic prediction of two SLP-related traits; lifetime number of litters (LNL) and lifetime pig production (LPP). Two bivariate DL models, convolutional neural network (CNN) and local convolutional neural network (LCNN), were compared with conventional bivariate linear models (i.e., genomic best linear unbiased prediction, Bayesian ridge regression, Bayes A, and Bayes B). Phenotype and pedigree data were collected from 40,011 sows that had husbandry records. Among these, 3,652 pigs were genotyped using the PorcineSNP60K BeadChip. The best predictive correlation for LNL was obtained with CNN (0.28), followed by LCNN (0.26) and conventional linear models (approximately 0.21). For LPP, the best predictive correlation was also obtained with CNN (0.29), followed by LCNN (0.27) and conventional linear models (approximately 0.25). A similar trend was observed with the mean squared error of prediction for the SLP traits. This study provides an example of a CNN that can outperform against the linear model-based genomic prediction approaches when the nonlinear interaction components are important because LNL and LPP exhibited strong epistatic interaction components. Additionally, our results suggest that applying bivariate DL models could also contribute to the prediction accuracy by utilizing the genetic correlation between LNL and LPP.

Genomic Prediction from Multi-Environment Trials of Wheat Breeding.

Genomic prediction relates a set of markers to variability in observed phenotypes of cultivars and allows for the prediction of phenotypes or breeding values of genotypes on unobserved individuals. Most genomic prediction approaches predict breeding values based solely on additive effects. However, the economic value of wheat lines is not only influenced by their additive component but also encompasses a non-additive part (e.g., additive × additive epistasis interaction). In this study, genomic prediction models were implemented in three target populations of environments (TPE) in South Asia. Four models that incorporate genotype × environment interaction (G × E) and genotype × genotype (GG) were tested: Factor Analytic (FA), FA with genomic relationship matrix (FA + G), FA with epistatic relationship matrix (FA + GG), and FA with both genomic and epistatic relationship matrices (FA + G + GG). Results show that the FA + G and FA + G + GG models displayed the best and a similar performance across all tests, leading us to infer that the FA + G model effectively captures certain epistatic effects. The wheat lines tested in sites in different TPE were predicted with different precisions depending on the cross-validation employed. In general, the best prediction accuracy was obtained when some lines were observed in some sites of particular TPEs and the worse genomic prediction was observed when wheat lines were never observed in any site of one TPE.

Genetic gains underpinning a little-known strawberry Green Revolution

The annual production of strawberry has increased by one million tonnes in the US and 8.4 million tonnes worldwide since 1960. Here we show that the US expansion was driven by genetic gains from Green Revolution breeding and production advances that increased yields by 2,755%. Using a California population with a century-long breeding history and phenotypes of hybrids observed in coastal California environments, we estimate that breeding has increased fruit yields by 2,974-6,636%, counts by 1,454-3,940%, weights by 228-504%, and firmness by 239-769%. Using genomic prediction approaches, we pinpoint the origin of the Green Revolution to the early 1950s and uncover significant increases in additive genetic variation caused by transgressive segregation and phenotypic diversification. Lastly, we show that the most consequential Green Revolution breeding breakthrough was the introduction of photoperiod-insensitive, PERPETUAL FLOWERING hybrids in the 1970s that doubled yields and drove the dramatic expansion of strawberry production in California.

Genomic prediction reveals unexplored variation in grain protein and lysine content across a vast winter wheat genebank collection.

Globally, wheat (Triticum aestivum L.) is a major source of proteins in human nutrition despite its unbalanced amino acid composition. The low lysine content in the protein fraction of wheat can lead to protein-energy-malnutrition prominently in developing countries. A promising strategy to overcome this problem is to breed varieties which combine high protein content with high lysine content. Nevertheless, this requires the incorporation of yet undefined donor genotypes into pre-breeding programs. Genebank collections are suspected to harbor the needed genetic diversity. In the 1970s, a large-scale screening of protein traits was conducted for the wheat genebank collection in Gatersleben; however, this data has been poorly mined so far. In the present study, a large historical dataset on protein content and lysine content of 4,971 accessions was curated, strictly corrected for outliers as well as for unreplicated data and consolidated as the corresponding adjusted entry means. Four genomic prediction approaches were compared based on the ability to accurately predict the traits of interest. High-quality phenotypic data of 558 accessions was leveraged by engaging the best performing prediction model, namely EG-BLUP. Finally, this publication incorporates predicted phenotypes of 7,651 accessions of the winter wheat collection. Five accessions were proposed as donor genotypes due to the combination of outstanding high protein content as well as lysine content. Further investigation of the passport data suggested an association of the adjusted lysine content with the elevation of the collecting site. This publicly available information can facilitate future pre-breeding activities.

Accurate prediction of quantitative traits with failed SNP calls in canola and maize.

In modern plant breeding, genomic selection is becoming the gold standard to select superior genotypes in large breeding populations that are only partially phenotyped. Many breeding programs commonly rely on single-nucleotide polymorphism (SNP) markers to capture genome-wide data for selection candidates. For this purpose, SNP arrays with moderate to high marker density represent a robust and cost-effective tool to generate reproducible, easy-to-handle, high-throughput genotype data from large-scale breeding populations. However, SNP arrays are prone to technical errors that lead to failed allele calls. To overcome this problem, failed calls are often imputed, based on the assumption that failed SNP calls are purely technical. However, this ignores the biological causes for failed calls-for example: deletions-and there is increasing evidence that gene presence-absence and other kinds of genome structural variants can play a role in phenotypic expression. Because deletions are frequently not in linkage disequilibrium with their flanking SNPs, permutation of missing SNP calls can potentially obscure valuable marker-trait associations. In this study, we analyze published datasets for canola and maize using four parametric and two machine learning models and demonstrate that failed allele calls in genomic prediction are highly predictive for important agronomic traits. We present two statistical pipelines, based on population structure and linkage disequilibrium, that enable the filtering of failed SNP calls that are likely caused by biological reasons. For the population and trait examined, prediction accuracy based on these filtered failed allele calls was competitive to standard SNP-based prediction, underlying the potential value of missing data in genomic prediction approaches. The combination of SNPs with all failed allele calls or the filtered allele calls did not outperform predictions with only SNP-based prediction due to redundancy in genomic relationship estimates.

A Genome-Wide Association Study Reveals Region Associated with Seed Protein Content in Cowpea.

Cowpea (Vigna unguiculata L. Walp., 2n = 2x = 22) is a protein-rich crop that complements staple cereals for humans and serves as fodder for livestock. It is widely grown in Africa and other developing countries as the primary source of protein in the diet; therefore, it is necessary to identify the protein-related loci to improve cowpea breeding. In the current study, we conducted a genome-wide association study (GWAS) on 161 cowpea accessions (151 USDA germplasm plus 10 Arkansas breeding lines) with a wide range of seed protein contents (21.8~28.9%) with 110,155 high-quality whole-genome single-nucleotide polymorphisms (SNPs) to identify markers associated with protein content, then performed genomic prediction (GP) for future breeding. A total of seven significant SNP markers were identified using five GWAS models (single-marker regression (SMR), the general linear model (GLM), Mixed Linear Model (MLM), Fixed and Random Model Circulating Probability Unification (FarmCPU), and Bayesian-information and Linkage-disequilibrium Iteratively Nested Keyway (BLINK), which are located at the same locus on chromosome 8 for seed protein content. This locus was associated with the gene Vigun08g039200, which was annotated as the protein of the thioredoxin superfamily, playing a critical function for protein content increase and nutritional quality improvement. In this study, a genomic prediction (GP) approach was employed to assess the accuracy of predicting seed protein content in cowpea. The GP was conducted using cross-prediction with five models, namely ridge regression best linear unbiased prediction (rrBLUP), Bayesian ridge regression (BRR), Bayesian A (BA), Bayesian B (BB), and Bayesian least absolute shrinkage and selection operator (BL), applied to seven random whole genome marker sets with different densities (10 k, 5 k, 2 k, 1 k, 500, 200, and 7), as well as significant markers identified through GWAS. The accuracies of the GP varied between 42.9% and 52.1% across the seven SNPs considered, depending on the model used. These findings not only have the potential to expedite the breeding cycle through early prediction of individual performance prior to phenotyping, but also offer practical implications for cowpea breeding programs striving to enhance seed protein content and nutritional quality.

Genomic prediction for drought tolerance using multienvironment data in a common bean (Phaseolus vulgaris) breeding program

AbstractThis work evaluated the efficiency of different genomic prediction (GP) methods in a diverse Mesoamerican panel of 339 common bean accessions, genotyped with 3398 SNP markers. Field experiments were carried out for three consecutive years, with adequate water supply (non‐stress—NS) and water restriction imposition (water‐stress—WS), analyzing seed weight (SW) and grain yield (GY). Two methods to predict the accuracies (rĝg) were adopted (GBLUP and Bayes) and also considered the environmental variation (GBLUP‐based reaction norm model). Similar accuracies were observed for both methods. For GY, the highest rĝg were detected under NS (rĝg = 0.49) in 2016 (rĝg = 0.49) and in the joint analysis for the WS condition (rĝg = 0.33), both for models using local landraces. For SW under NS, the rĝg was higher for the elite lines (rĝg = 0.72), whereas for WS, the rĝg dropped considerably, ranging from 0.45 to 0.61 for the joint analysis, considering the landraces and all samples, respectively. For GY and SW, under NS, the rĝg using both models increased with increasing number of SNPs, until reaching a plateau of 800 and 300 SNPs, respectively. Increasing the training population (TP) size resulted in greater accuracy. Taking in account the Genotype × Environment, the multienvironment model performed better especially for more complex traits (GY/NS: rĝg = 0.32). The GP approach has great potential to help commercial bean breeding programs improving the performance of target quantitative traits.

The challenge of demonstrating contemporary natural selection on polygenic quantitative traits in the wild.

In a From the Cover article in this issue of Molecular Ecology, Ashraf et al.(2022) apply genomic prediction methods, devised by breeders to inform artificial selection, to understand the genetic component of variation in highly polygenic quantitative traits in Soay sheep (Figure1). These methods have allowed them to investigate the effects of contemporary natural selection on genetic variation underlying these traits in the wild (Hunter et al.,2022). Genomic prediction approaches promise to enhance our understanding of the evolution of highly polygenic quantitative traits in the wild and may allow us to document concrete examples of their natural selection in real time in systems that would otherwise be intractable.

A divide-and-conquer approach for genomic prediction in rubber tree using machine learning

Rubber tree (Hevea brasiliensis) is the main feedstock for commercial rubber; however, its long vegetative cycle has hindered the development of more productive varieties via breeding programs. With the availability of H. brasiliensis genomic data, several linkage maps with associated quantitative trait loci have been constructed and suggested as a tool for marker-assisted selection. Nonetheless, novel genomic strategies are still needed, and genomic selection (GS) may facilitate rubber tree breeding programs aimed at reducing the required cycles for performance assessment. Even though such a methodology has already been shown to be a promising tool for rubber tree breeding, increased model predictive capabilities and practical application are still needed. Here, we developed a novel machine learning-based approach for predicting rubber tree stem circumference based on molecular markers. Through a divide-and-conquer strategy, we propose a neural network prediction system with two stages: (1) subpopulation prediction and (2) phenotype estimation. This approach yielded higher accuracies than traditional statistical models in a single-environment scenario. By delivering large accuracy improvements, our methodology represents a powerful tool for use in Hevea GS strategies. Therefore, the incorporation of machine learning techniques into rubber tree GS represents an opportunity to build more robust models and optimize Hevea breeding programs.

Choosing the right tool: Leveraging of plant genetic resources in wheat (Triticum aestivum L.) benefits from selection of a suitable genomic prediction model.

Genomic prediction of genebank accessions benefits from the consideration of additive-by-additive epistasis and subpopulation-specific marker effects. Wheat (Triticum aestivum L.) and other species of the Triticum genus are well represented in genebank collections worldwide. The substantial genetic diversity harbored by more than 850,000 accessions can be explored for their potential use in modern plant breeding. Characterization of these large number of accessions is constrained by the required resources, and this fact limits their use so far. This limitation might be overcome by engaging genomic prediction. The present study compared ten different genomic prediction approaches to the prediction of four traits, namely flowering time, plant height, thousand grain weight, and yellow rust resistance, in a diverse set of 7745 accession samples from Germany's Federal ex situ genebank at the Leibniz Institute of Plant Genetics and Crop Plant Research in Gatersleben. Approaches were evaluated based on prediction ability and robustness to the confounding influence of strong population structure. The authors propose the wide application of extended genomic best linear unbiased prediction due to the observed benefit of incorporating additive-by-additive epistasis. General and subpopulation-specific additive ridge regression best linear unbiased prediction, which accounts for subpopulation-specific marker-effects, was shown to be a good option if contrasting clusters are encountered in the analyzed collection. The presented findings reaffirm that the trait's genetic architecture as well as the composition and relatedness of the training set and test set are major driving factors for the accuracy of genomic prediction.

An Integrative Genomic Prediction Approach for Predicting Buffalo Milk Traits by Incorporating Related Cattle QTLs.

Background: The 90K Axiom Buffalo SNP Array is expected to improve and speed up various genomic analyses for the buffalo (Bubalus bubalis). Genomic prediction is an effective approach in animal breeding to improve selection and reduce costs. As buffalo genome research is lagging behind that of the cow and production records are also limited, genomic prediction performance will be relatively poor. To improve the genomic prediction in buffalo, we introduced a new approach (pGBLUP) for genomic prediction of six buffalo milk traits by incorporating QTL information from the cattle milk traits in order to help improve the prediction performance for buffalo. Results: In simulations, the pGBLUP could outperform BayesR and the GBLUP if the prior biological information (i.e., the known causal loci) was appropriate; otherwise, it performed slightly worse than BayesR and equal to or better than the GBLUP. In real data, the heritability of the buffalo genomic region corresponding to the cattle milk trait QTLs was enriched (fold of enrichment > 1) in four buffalo milk traits (FY270, MY270, PY270, and PM) when the EBV was used as the response variable. The DEBV as the response variable yielded more reliable genomic predictions than the traditional EBV, as has been shown by previous research. The performance of the three approaches (GBLUP, BayesR, and pGBLUP) did not vary greatly in this study, probably due to the limited sample size, incomplete prior biological information, and less artificial selection in buffalo. Conclusions: To our knowledge, this study is the first to apply genomic prediction to buffalo by incorporating prior biological information. The genomic prediction of buffalo traits can be further improved with a larger sample size, higher-density SNP chips, and more precise prior biological information.

Selection of parental lines for plant breeding via genomic prediction.

A set of superior parental lines is imperative for the development of high-performing inbred lines in any biparental crossing program for crops. The main objectives of this study are to (a) develop a genomic prediction approach to identify superior parental lines for multi-trait selection, and (b) generate a software package for users to execute the proposed approach before conducting field experiments. According to different breeding goals of the target traits, a novel selection index integrating information from genomic-estimated breeding values (GEBVs) of candidate accessions was proposed to evaluate the composite performance of simulated progeny populations. Two rice (Oryza sativa L.) genome datasets were analyzed to illustrate the potential applications of the proposed approach. One dataset applied to the parental selection for producing inbred lines with satisfactory performance in primary and secondary traits simultaneously. The other one applied to demonstrate the application of producing inbred lines with high adaptability to different environments. Overall, the results showed that incorporating GEBV and genomic diversity into a selection strategy based on the proposed selection index could assist in selecting superior parents to meet the desired breeding goals and increasing long-term genetic gain. An R package, called IPLGP, was generated to facilitate the widespread application of the approach.

A joint learning approach for genomic prediction in polyploid grasses

Poaceae, among the most abundant plant families, includes many economically important polyploid species, such as forage grasses and sugarcane (Saccharum spp.). These species have elevated genomic complexities and limited genetic resources, hindering the application of marker-assisted selection strategies. Currently, the most promising approach for increasing genetic gains in plant breeding is genomic selection. However, due to the polyploidy nature of these polyploid species, more accurate models for incorporating genomic selection into breeding schemes are needed. This study aims to develop a machine learning method by using a joint learning approach to predict complex traits from genotypic data. Biparental populations of sugarcane and two species of forage grasses (Urochloa decumbens, Megathyrsus maximus) were genotyped, and several quantitative traits were measured. High-quality markers were used to predict several traits in different cross-validation scenarios. By combining classification and regression strategies, we developed a predictive system with promising results. Compared with traditional genomic prediction methods, the proposed strategy achieved accuracy improvements exceeding 50%. Our results suggest that the developed methodology could be implemented in breeding programs, helping reduce breeding cycles and increase genetic gains.

Using genomic prediction to detect microevolutionary change of a quantitative trait.

Detecting microevolutionary responses to natural selection by observing temporal changes in individual breeding values is challenging. The collection of suitable datasets can take many years and disentangling the contributions of the environment and genetics to phenotypic change is not trivial. Furthermore, pedigree-based methods of obtaining individual breeding values have known biases. Here, we apply a genomic prediction approach to estimate breeding values of adult weight in a 35-year dataset of Soay sheep (Ovis aries). Comparisons are made with a traditional pedigree-based approach. During the study period, adult body weight decreased, but the underlying genetic component of body weight increased, at a rate that is unlikely to be attributable to genetic drift. Thus cryptic microevolution of greater adult body weight has probably occurred. Genomic and pedigree-based approaches gave largely consistent results. Thus, using genomic prediction to study microevolution in wild populations can remove the requirement for pedigree data, potentially opening up new study systems for similar research.

The SoyaGen Project: Putting Genomics to Work for Soybean Breeders.

The SoyaGen project was a collaborative endeavor involving Canadian soybean researchers and breeders from academia and the private sector as well as international collaborators. Its aims were to develop genomics-derived solutions to real-world challenges faced by breeders. Based on the needs expressed by the stakeholders, the research efforts were focused on maximizing realized yield through optimization of maturity and improved disease resistance. The main deliverables related to molecular breeding in soybean will be reviewed here. These include: (1) SNP datasets capturing the genetic diversity within cultivated soybean (both within a worldwide collection of > 1,000 soybean accessions and a subset of 102 short-season accessions (MG0 and earlier) directly relevant to this group); (2) SNP markers for selecting favorable alleles at key maturity genes as well as loci associated with increased resistance to key pathogens and pests (Phytophthora sojae, Heterodera glycines, Sclerotinia sclerotiorum); (3) diagnostic tools to facilitate the identification and mapping of specific pathotypes of P. sojae; and (4) a genomic prediction approach to identify the most promising combinations of parents. As a result of this fruitful collaboration, breeders have gained new tools and approaches to implement molecular, genomics-informed breeding strategies. We believe these tools and approaches are broadly applicable to soybean breeding efforts around the world.

Extend mixed models to multilayer neural networks for genomic prediction including intermediate omics data.

With the growing amount and diversity of intermediate omics data complementary to genomics (e.g. DNA methylation, gene expression, and protein abundance), there is a need to develop methods to incorporate intermediate omics data into conventional genomic evaluation. The omics data help decode the multiple layers of regulation from genotypes to phenotypes, thus forms a connected multilayer network naturally. We developed a new method named NN-MM to model the multiple layers of regulation from genotypes to intermediate omics features, then to phenotypes, by extending conventional linear mixed models ("MM") to multilayer artificial neural networks ("NN"). NN-MM incorporates intermediate omics features by adding middle layers between genotypes and phenotypes. Linear mixed models (e.g. pedigree-based BLUP, GBLUP, Bayesian Alphabet, single-step GBLUP, or single-step Bayesian Alphabet) can be used to sample marker effects or genetic values on intermediate omics features, and activation functions in neural networks are used to capture the nonlinear relationships between intermediate omics features and phenotypes. NN-MM had significantly better prediction performance than the recently proposed single-step approach for genomic prediction with intermediate omics data. Compared to the single-step approach, NN-MM can handle various patterns of missing omics measures and allows nonlinear relationships between intermediate omics features and phenotypes. NN-MM has been implemented in an open-source package called "JWAS".

Genetic and genomic characterization followed by single-step genomic evaluation of withers height in German Warmblood horses

Reliability of genomic predictions is influenced by the size and genetic composition of the reference population. For German Warmblood horses, compilation of a reference population has been enabled through the cooperation of five German breeding associations. In this study, preliminary data from this joint reference population were used to genetically and genomically characterize withers height and to apply single-step methodology for estimating genomic breeding values for withers height. Using data on 2113 mares and their genomic information considering about 62,000 single nucleotide polymorphisms (SNPs), analysis of the genomic relationship revealed substructures reflecting breed origin and different breeding goals of the contributing breeding associations. A genome-wide association study confirmed a known quantitative trait locus (QTL) for withers height on equine chromosome (ECA) 3 close to LCORL and identified a further significant peak on ECA 1. Using a single-step approach with a combined relationship matrix, the estimated heritability for withers height was 0.31 (SE = 0.08) and the corresponding genomic breeding values ranged from − 2.94 to 2.96 cm. A mean reliability of 0.38 was realized for these breeding values. The analyses of withers height showed that compiling a reference population across breeds is a suitable strategy for German Warmblood horses. The single-step method is an appealing approach for practical genomic prediction in horses, because not many genotypes are available yet and animals without genotypes can by this way directly contribute to the estimation system.

Crop Improvement for Circular Bioeconomy Systems

HighlightsWe describe and demonstrate a multidimensional framework to integrate environmental and genomic predictors to enable crop improvement for a circular bioeconomy.A model training procedure based on multiple phenotypes is shown to improve predictive skill.The decision set comprised of model outputs can inform selection for both productivity and circularity metrics.Abstract. Contemporary agricultural systems are poised to transition from linear to circular, adopting concepts of recycling, repurposing, and regeneration. This transition will require changing crop improvement objectives to consider the entire system, and thus provide solutions to improve complex systems for higher productivity, resource use efficiency, and environmental quality. The methods and approaches that underpinned the doubling of yields during the last century may no longer be fully adequate to target crop improvement for circular agricultural systems. Here we propose a multidimensional framework for prediction with outcomes useful to assess both crop performance traits and environmental sustainability of the designed agricultural systems. The study focuses on maize harvestable grain yield and total carbon production, water use, and use efficiency for yield and carbon. The framework builds on the crop growth model whole genome prediction system, which is enabled by advanced phenomics and the integration of symbolic and sub-symbolic artificial intelligence. We demonstrate the approach and prediction accuracy advantages over a standard statistical genomic prediction approach used to breed maize hybrids for yield, flowering time, and kernel set using a dataset comprised of 7004 hybrids, 103 breeding populations, and 62 environments resulting from six years of experimentation in maize drought breeding in the U.S. We propose this framework to motivate a dialogue for how to enable circularity in agriculture through prediction-based systems design. Keywords: Circular bioeconomy, Circular economy, Crop improvement, Crop models, Drought, Gene editing, Genomic prediction, Maize, Plant breeding.

Towards a Cost-Effective Implementation of Genomic Prediction Based on Low Coverage Whole Genome Sequencing in Dezhou Donkey.

Low-coverage whole genome sequencing is a low-cost genotyping technology. Combined with genotype imputation approaches, it is likely to become a critical component of cost-effective genomic selection programs in agricultural livestock. Here, we used the low-coverage sequence data of 617 Dezhou donkeys to investigate the performance of genotype imputation for low-coverage whole genome sequence data and genomic prediction based on the imputed genotype data. The specific aims were as follows: 1) to measure the accuracy of genotype imputation under different sequencing depths, sample sizes, minor allele frequency (MAF), and imputation pipelines and 2) to assess the accuracy of genomic prediction under different marker densities derived from the imputed sequence data, different strategies for constructing the genomic relationship matrixes, and single-vs. multi-trait models. We found that a high imputation accuracy (>0.95) can be achieved for sequence data with a sequencing depth as low as 1x and the number of sequenced individuals ≥400. For genomic prediction, the best performance was obtained by using a marker density of 410K and a G matrix constructed using expected marker dosages. Multi-trait genomic best linear unbiased prediction (GBLUP) performed better than single-trait GBLUP. Our study demonstrates that low-coverage whole genome sequencing would be a cost-effective approach for genomic prediction in Dezhou donkey.

A novel genomic prediction method combining randomized Haseman-Elston regression with a modified algorithm for Proven and Young for large genomic data

Computational efficiency has become a key issue in genomic prediction (GP) owing to the massive historical datasets accumulated. We developed hereby a new super-fast GP approach (SHEAPY) combining randomized Haseman-Elston regression (RHE-reg) with a modified Algorithm for Proven and Young (APY) in an additive-effect model, using the former to estimate heritability and then the latter to invert a large genomic relationship matrix for best linear prediction. In simulation results with varied sizes of training population, GBLUP, HEAPY|A and SHEAPY showed similar predictive performance when the size of a core population was half that of a large training population and the heritability was a fixed value, and the computational speed of SHEAPY was faster than that of GBLUP and HEAPY|A. In simulation results with varied heritability, SHEAPY showed better predictive ability than GBLUP in all cases and than HEAPY|A in most cases when the size of a core population was 4/5 that of a small training population and the training population size was a fixed value. As a proof of concept, SHEAPY was applied to the analysis of two real datasets. In an Arabidopsis thaliana F2 population, the predictive performance of SHEAPY was similar to or better than that of GBLUP and HEAPY|A in most cases when the size of a core population (200) was 2/3 of that of a small training population (300). In a sorghum multiparental population, SHEAPY showed higher predictive accuracy than HEAPY|A for all of three traits, and than GBLUP for two traits. SHEAPY may become the GP method of choice for large-scale genomic data.