Increased genomic prediction accuracy in wheat breeding using a large Australian panel

Evaluating genomic selection and speed breeding for Fusarium head blight resistance in wheat using stochastic simulations

Genomic selection-based breeding programs offer significant advantages over conventional phenotypic selection, particularly in accelerating genetic gains in plant breeding, as demonstrated by simulations focused on combating Fusarium head blight (FHB) in wheat. FHB resistance, a crucial trait, is challenging to breed for due to its quantitative inheritance and environmental influence, leading to slow progress using conventional breeding methods. Stochastic simulations in our study compared various breeding schemes, incorporating genomic selection (GS) and combining it with speed breeding, against conventional phenotypic selection. Two datasets were simulated, reflecting real-life genotypic data (MASBASIS) and a simulated wheat breeding program (EXAMPLE). Initially a 20-year burn-in phase using a conventional phenotypic selection method followed by a 20-year advancement phase with three GS-based breeding programs (GSF2F8, GSF8, and SpeedBreeding + GS) were evaluated alongside over a conventional phenotypic selection method. Results consistently showed significant increases in genetic gain with GS-based programs compared to phenotypic selection, irrespective of the selection strategies employed. Among the GS schemes, SpeedBreeding + GS consistently outperformed others, generating the highest genetic gains. This combination effectively minimized generation intervals within the breeding cycle, enhancing efficiency. This study underscores the advantages of genomic selection in accelerating breeding gains for wheat, particularly in combating FHB. By leveraging genomic information and innovative techniques like speed breeding, breeders can efficiently select for desired traits, significantly reducing testing time and costs associated with conventional phenotypic methods.

Supplementary Information

The online version contains supplementary material available at 10.1007/s11032-024-01527-z.

Wheat improvement through advances in single nucleotide polymorphism (SNP) detection and genotyping with a special emphasis on rust resistance

KEY MESSAGE

Single nucleotide polymorphism (SNP) markers in wheat and their prospects in breeding with special reference to rust resistance. Single nucleotide polymorphism (SNP)-based markers are increasingly gaining momentum for screening and utilizing vital agronomic traits in wheat. To date, more than 260 million SNPs have been detected in modern cultivars and landraces of wheat. This rapid SNP discovery was made possible through the release of near-complete reference and pan-genome assemblies of wheat and its wild relatives, coupled with whole genome sequencing (WGS) of thousands of wheat accessions. Further, genotyping customized SNP sites were facilitated by a series of arrays (9 to 820Ks), a cost effective substitute WGS. Lately, germplasm-specific SNP arrays have been introduced to characterize novel traits and detect closely linked SNPs for marker-assisted breeding. Subsequently, the kompetitive allele-specific PCR (KASP) assay was introduced for rapid and large-scale screening of specific SNP markers. Moreover, with the advances and reduction in sequencing costs, ample opportunities arise for generating SNPs artificially through mutations and in combination with next-generation sequencing and comparative genomic analyses. In this review, we provide historical developments and prospects of SNP markers in wheat breeding with special reference to rust resistance where over 50 genetic loci have been characterized through SNP markers. Rust resistance is one of the most essential traits for wheat breeding as new strains of the Puccinia fungus, responsible for rust diseases, evolve frequently and globally.

Genetic analysis of late-maturity α-amylase in twelve wheat populations

MAIN CONCLUSION

Genetic loci, particularly those with an effect in the independent panel, could be utilised to further reduce LMA expression when used with favourable combinations of genes known to affect LMA. Late maturity α-amylase (LMA) is a grain quality defect involving elevated α-amylase within the aleurone of wheat (Triticum aestivum L.) grains. The genes known to affect expression are the reduced height genes Rht-B1 (chromosome 4B) and Rht-D1 (chromosome 4D), and an ent-copalyl diphosphate synthase gene (LMA-1) on chromosome 7B. Other minor effect loci have been reported, but these are poorly characterised and further genetic understanding is needed. In this study, twelve F4-derived populations were created through single seed descent, genotyped and evaluated for LMA. LMA-1 haplotype C and the Rht-D1b allele substantially reduced LMA expression. The alternative dwarfing genes Rht13 and Rht18 had no significant effect on LMA expression. Additional quantitative trait loci (QTL) were mapped at 16 positions in the wheat genome. Effects on LMA expression were detected for four of these QTL in a large independent panel of Australian wheat lines. The QTL detected in mapping populations and confirmed in the large independent panel provide further opportunity for selection against LMA, especially if combined with Rht-D1b and/or favourable haplotypes of LMA-1.

Root system architecture associated zinc variability in wheat (Triticum aestivum L.)

Root system architecture (RSA) plays a fundamental role in nutrient uptake, including zinc (Zn). Wheat grains are inheritably low in Zn. As Zn is an essential nutrient for plants, improving its uptake will not only improve their growth and yield but also the nutritional quality of staple grains. A rhizobox study followed by a pot study was conducted to evaluate Zn variability with respect to RSA and its impact on grain Zn concentration. The grain Zn content of one hundred wheat varieties was determined and grown in rhizoboxes with differential Zn (no Zn and 0.05 mg L-1 ZnSO4). Seedlings were harvested 12 days after sowing, and root images were taken and analyzed by SmartRoot software. Using principal component analysis, twelve varieties were screened out based on vigorous and weaker RSA with high and low grain Zn content. The screened varieties were grown in pots with (11 mg ZnSO4 kg-1 soil) and without Zn application to the soil. Zinc translocation, localization, and agronomic parameters were recorded after harvesting at maturity. In the rhizobox experiment, 4% and 8% varieties showed higher grain Zn content with vigorous and weaker RSA, respectively, while 45% and 43% varieties had lower grain Zn content with vigorous and weaker RSA. However, the pot experiment revealed that varieties with vigorous root system led to higher grain yield, though the grain Zn concentration were variable, while all varieties with weaker root system had lower yield as well as grain Zn concentration. Zincol-16 revealed the highest Zn concentration (28.07 mg kg-1) and grain weight (47.9 g). Comparatively higher level of Zn was localized in the aleurone layer than in the embryonic region and endosperm. It is concluded that genetic variability exists among wheat varieties for RSA and grain Zn content, with a significant correlation. Therefore, RSA attributes are promising targets for the Zn biofortification breeding program. However, Zn localization in endosperm needs to be further investigated to achieve the goal of reducing Zn malnutrition.

Improved genomic prediction using machine learning with Variational Bayesian sparsity

BACKGROUND

Genomic prediction has become a powerful modelling tool for assessing line performance in plant and livestock breeding programmes. Among the genomic prediction modelling approaches, linear based models have proven to provide accurate predictions even when the number of genetic markers exceeds the number of data samples. However, breeding programmes are now compiling data from large numbers of lines and test environments for analyses, rendering these approaches computationally prohibitive. Machine learning (ML) now offers a solution to this problem through the construction of fully connected deep learning architectures and high parallelisation of the predictive task. However, the fully connected nature of these architectures immediately generates an over-parameterisation of the network that needs addressing for efficient and accurate predictions.

RESULTS

In this research we explore the use of an ML architecture governed by variational Bayesian sparsity in its initial layers that we have called VBS-ML. The use of VBS-ML provides a mechanism for feature selection of important markers linked to the trait, immediately reducing the network over-parameterisation. Selected markers then propagate to the remaining fully connected feed-forward components of the ML network to form the final genomic prediction. We illustrated the approach with four large Australian wheat breeding data sets that range from 2665 lines to 10375 lines genotyped across a large set of markers. For all data sets, the use of the VBS-ML architecture improved genomic prediction accuracy over legacy linear based modelling approaches.

CONCLUSIONS

An ML architecture governed under a variational Bayesian paradigm was shown to improve genomic prediction accuracy over legacy modelling approaches. This VBS-ML approach can be used to dramatically decrease the parameter burden on the network and provide a computationally feasible approach for improving genomic prediction conducted with large breeding population numbers and genetic markers.

An international wheat diversity panel reveals novel sources of genetic resistance to tan spot in Australia

KEY MESSAGE

Novel sources of genetic resistance to tan spot in Australia have been discovered using one-step GWAS and genomic prediction models that accounts for additive and non-additive genetic variation. Tan spot is a foliar disease in wheat caused by the fungal pathogen Pyrenophora tritici-repentis (Ptr) and has been reported to generate up to 50% yield losses under favourable disease conditions. Although farming management practices are available to reduce disease, the most economically sustainable approach is establishing genetic resistance through plant breeding. To further understand the genetic basis for disease resistance, we conducted a phenotypic and genetic analysis study using an international diversity panel of 192 wheat lines from the Maize and Wheat Improvement Centre (CIMMYT), the International Centre for Agriculture in the Dry Areas (ICARDA) and Australian (AUS) wheat research programmes. The panel was evaluated using Australian Ptr isolates in 12 experiments conducted in three Australian locations over two years, with assessment for tan spot symptoms at various plant development stages. Phenotypic modelling indicated high heritability for nearly all tan spot traits with ICARDA lines displaying the greatest average resistance. We then conducted a one-step whole-genome analysis of each trait using a high-density SNP array, revealing a large number of highly significant QTL exhibiting a distinct lack of repeatability across the traits. To better summarise the genetic resistance of the lines, a one-step genomic prediction of each tan spot trait was conducted by combining the additive and non-additive predicted genetic effects of the lines. This revealed multiple CIMMYT lines with broad genetic resistance across the developmental stages of the plant which can be utilised in Australian wheat breeding programmes to improve tan spot disease resistance.

Genomic prediction in plants: opportunities for ensemble machine learning based approaches

Background: Many studies have demonstrated the utility of machine learning (ML) methods for genomic prediction (GP) of various plant traits, but a clear rationale for choosing ML over conventionally used, often simpler parametric methods, is still lacking. Predictive performance of GP models might depend on a plethora of factors including sample size, number of markers, population structure and genetic architecture. Methods: Here, we investigate which problem and dataset characteristics are related to good performance of ML methods for genomic prediction. We compare the predictive performance of two frequently used ensemble ML methods (Random Forest and Extreme Gradient Boosting) with parametric methods including genomic best linear unbiased prediction (GBLUP), reproducing kernel Hilbert space regression (RKHS), BayesA and BayesB. To explore problem characteristics, we use simulated and real plant traits under different genetic complexity levels determined by the number of Quantitative Trait Loci (QTLs), heritability (h2 and h2e), population structure and linkage disequilibrium between causal nucleotides and other SNPs. Results: Decision tree based ensemble ML methods are a better choice for nonlinear phenotypes and are comparable to Bayesian methods for linear phenotypes in the case of large effect Quantitative Trait Nucleotides (QTNs). Furthermore, we find that ML methods are susceptible to confounding due to population structure but less sensitive to low linkage disequilibrium than linear parametric methods. Conclusions: Overall, this provides insights into the role of ML in GP as well as guidelines for practitioners.

Genomic prediction in plants: opportunities for ensemble machine learning based approaches

Background: Many studies have demonstrated the utility of machine learning (ML) methods for genomic prediction (GP) of various plant traits, but a clear rationale for choosing ML over conventionally used, often simpler parametric methods, is still lacking. Predictive performance of GP models might depend on a plethora of factors including sample size, number of markers, population structure and genetic architecture. Methods: Here, we investigate which problem and dataset characteristics are related to good performance of ML methods for genomic prediction. We compare the predictive performance of two frequently used ensemble ML methods (Random Forest and Extreme Gradient Boosting) with parametric methods including genomic best linear unbiased prediction (GBLUP), reproducing kernel Hilbert space regression (RKHS), BayesA and BayesB. To explore problem characteristics, we use simulated and real plant traits under different genetic complexity levels determined by the number of Quantitative Trait Loci (QTLs), heritability (h2 and h2e), population structure and linkage disequilibrium between causal nucleotides and other SNPs. Results: Decision tree based ensemble ML methods are a better choice for nonlinear phenotypes and are comparable to Bayesian methods for linear phenotypes in the case of large effect Quantitative Trait Nucleotides (QTNs). Furthermore, we find that ML methods are susceptible to confounding due to population structure but less sensitive to low linkage disequilibrium than linear parametric methods. Conclusions: Overall, this provides insights into the role of ML in GP as well as guidelines for practitioners.

Breeding Drought-Tolerant Pearl Millet Using Conventional and Genomic Approaches: Achievements and Prospects

Pearl millet [Pennisetum glaucum (L.) R. Br.] is a C4 crop cultivated for its grain and stover in crop-livestock-based rain-fed farming systems of tropics and subtropics in the Indian subcontinent and sub-Saharan Africa. The intensity of drought is predicted to further exacerbate because of looming climate change, necessitating greater focus on pearl millet breeding for drought tolerance. The nature of drought in different target populations of pearl millet-growing environments (TPEs) is highly variable in its timing, intensity, and duration. Pearl millet response to drought in various growth stages has been studied comprehensively. Dissection of drought tolerance physiology and phenology has helped in understanding the yield formation process under drought conditions. The overall understanding of TPEs and differential sensitivity of various growth stages to water stress helped to identify target traits for manipulation through breeding for drought tolerance. Recent advancement in high-throughput phenotyping platforms has made it more realistic to screen large populations/germplasm for drought-adaptive traits. The role of adapted germplasm has been emphasized for drought breeding, as the measured performance under drought stress is largely an outcome of adaptation to stress environments. Hybridization of adapted landraces with selected elite genetic material has been stated to amalgamate adaptation and productivity. Substantial progress has been made in the development of genomic resources that have been used to explore genetic diversity, linkage mapping (QTLs), marker-trait association (MTA), and genomic selection (GS) in pearl millet. High-throughput genotyping (HTPG) platforms are now available at a low cost, offering enormous opportunities to apply markers assisted selection (MAS) in conventional breeding programs targeting drought tolerance. Next-generation sequencing (NGS) technology, micro-environmental modeling, and pearl millet whole genome re-sequence information covering circa 1,000 wild and cultivated accessions have helped to greater understand germplasm, genomes, candidate genes, and markers. Their application in molecular breeding would lead to the development of high-yielding and drought-tolerant pearl millet cultivars. This review examines how the strategic use of genetic resources, modern genomics, molecular biology, and shuttle breeding can further enhance the development and delivery of drought-tolerant cultivars.

A multi-environment framework to evaluate the adaptation of wheat (Triticum aestivum) to heat stress

Assessing adaptation to abiotic stresses such as high temperature conditions across multiple environments presents opportunities for breeders to target selection for broad adaptation and specific adaptation. Adaptation of wheat to heat stress is an important component of adaptation in variable climates such as the cereal producing areas of Australia. However, in variable climates stress conditions may not be present in every season or are present to varying degrees, at different times during the season. Such conditions complicate plant breeders' ability to select for adaptation to abiotic stress. This study presents a framework for the assessment of the genetic basis of adaptation to heat stress conditions with improved relevance to breeders' selection objectives. The framework was applied here with the evaluation of 1225 doubled haploid lines from five populations across six environments (three environments selected for contrasting temperature stress conditions during anthesis and grain fill periods, over two consecutive seasons), using regionally best practice planting times to evaluate the role of heat stress conditions in genotype adaptation. Temperature co-variates were determined for each genotype, in each environment, for the anthesis and grain fill periods. Genome-wide QTL analysis identified performance QTL for stable effects across all environments, and QTL that illustrated responsiveness to heat stress conditions across the sampled environments. A total of 199 QTL were identified, including 60 performance QTL, and 139 responsiveness QTL. Of the identified QTL, 99 occurred independent of the 21 anthesis date QTL identified. Assessing adaptation to heat stress conditions as the combination of performance and responsiveness offers breeders opportunities to select for grain yield stability across a range of environments, as well as genotypes with higher relative yield in stress conditions.

Genomic prediction of preliminary yield trials in chickpea: Effect of functional annotation of SNPs and environment

Achieving yield potential in chickpea (Cicer arietinum L.) is limited by many constraints that include biotic and abiotic stresses. Combining next-generation sequencing technology with advanced statistical modeling has the potential to increase genetic gain efficiently. Whole genome resequencing data was obtained from 315 advanced chickpea breeding lines from the Australian chickpea breeding program resulting in more than 298,000 single nucleotide polymorphisms (SNPs) discovered. Analysis of population structure revealed a distinct group of breeding lines with many alleles that are absent from recently released Australian cultivars. Genome-wide association studies (GWAS) using these Australian breeding lines identified 20 SNPs significantly associated with grain yield in multiple field environments. A reduced level of nucleotide diversity and extended linkage disequilibrium suggested that some regions in these chickpea genomes may have been through selective breeding for yield or other traits. A large introgression segment that introduced from C. echinospermum for phytophthora root rot resistance was identified on chromosome 6, yet it also has unintended consequences of reducing yield due to linkage drag. We further investigated the effect of genotype by environment interaction on genomic prediction of yield. We found that the training set had better prediction accuracy when phenotyped under conditions relevant to the targeted environments. We also investigated the effect of SNP functional annotation on prediction accuracy using different subsets of SNPs based on their genomic locations: regulatory regions, exome, and alternative splice sites. Compared with the whole SNP dataset, a subset of SNPs did not significantly decrease prediction accuracy for grain yield despite consisting of a smaller number of SNPs.

Increased Predictive Accuracy of Multi-Environment Genomic Prediction Model for Yield and Related Traits in Spring Wheat (Triticum aestivum L.)

Genomic selection (GS) has the potential to improve the selection gain for complex traits in crop breeding programs from resource-poor countries. The GS model performance in multi-environment (ME) trials was assessed for 141 advanced breeding lines under four field environments via cross-predictions. We compared prediction accuracy (PA) of two GS models with or without accounting for the environmental variation on four quantitative traits of significant importance, i.e., grain yield (GRYLD), thousand-grain weight, days to heading, and days to maturity, under North and Central Indian conditions. For each trait, we generated PA using the following two different ME cross-validation (CV) schemes representing actual breeding scenarios: (1) predicting untested lines in tested environments through the ME model (ME_CV1) and (2) predicting tested lines in untested environments through the ME model (ME_CV2). The ME predictions were compared with the baseline single-environment (SE) GS model (SE_CV1) representing a breeding scenario, where relationships and interactions are not leveraged across environments. Our results suggested that the ME models provide a clear advantage over SE models in terms of robust trait predictions. Both ME models provided 2-3 times higher prediction accuracies for all four traits across the four tested environments, highlighting the importance of accounting environmental variance in GS models. While the improvement in PA from SE to ME models was significant, the CV1 and CV2 schemes did not show any clear differences within ME, indicating the ME model was able to predict the untested environments and lines equally well. Overall, our results provide an important insight into the impact of environmental variation on GS in smaller breeding programs where these programs can potentially increase the rate of genetic gain by leveraging the ME wheat breeding trials.

Genetic analysis of wheat (Triticum aestivum) adaptation to heat stress

Adaptation to abiotic stresses such as high-temperature conditions should be considered as its independent components of total performance and responsiveness. Understanding and identifying improved adaptation to abiotic stresses such as heat stress has been the focus of a number of studies in recent decades. However, confusing and potentially misleading terminology has made progress difficult and hard to apply within breeding programs selecting for improved adaption to heat stress conditions. This study proposes that adaption to heat stress (and other abiotic stresses) be considered as the combination of total performance and responsiveness to heat stress. In this study, 1413 doubled haploid lines from seven populations were screened through a controlled environment assay, subjecting plants to three consecutive eight hour days of an air temperature of 36 °C and a wind speed of 40 km h^-1, 10 days after the end of anthesis. QTL mapping identified a total of 96 QTL for grain yield determining traits and anthesis date with nine correlating to responsiveness, 72 for total performance and 15 for anthesis date. Responsiveness QTL were found both collocated with other performance QTL as well as independently. A sound understanding of genomic regions associated with total performance and responsiveness will be important for breeders. Genomic regions of total performance, those that show higher performance that is stable under both stressed and non-stressed conditions, potentially offer significant opportunities to breeders. We propose this as a definition and selection target that has not previously been defined for heat stress adaptation.

Machine learning approaches reveal genomic regions associated with sugarcane brown rust resistance

Sugarcane is an economically important crop, but its genomic complexity has hindered advances in molecular approaches for genetic breeding. New cultivars are released based on the identification of interesting traits, and for sugarcane, brown rust resistance is a desirable characteristic due to the large economic impact of the disease. Although marker-assisted selection for rust resistance has been successful, the genes involved are still unknown, and the associated regions vary among cultivars, thus restricting methodological generalization. We used genotyping by sequencing of full-sib progeny to relate genomic regions with brown rust phenotypes. We established a pipeline to identify reliable SNPs in complex polyploid data, which were used for phenotypic prediction via machine learning. We identified 14,540 SNPs, which led to a mean prediction accuracy of 50% when using different models. We also tested feature selection algorithms to increase predictive accuracy, resulting in a reduced dataset with more explanatory power for rust phenotypes. As a result of this approach, we achieved an accuracy of up to 95% with a dataset of 131 SNPs related to brown rust QTL regions and auxiliary genes. Therefore, our novel strategy has the potential to assist studies of the genomic organization of brown rust resistance in sugarcane.

Genomic Prediction for Grain Yield and Yield-Related Traits in Chinese Winter Wheat

Genomic selection (GS) is a strategy to predict the genetic merits of individuals using genome-wide markers. However, GS prediction accuracy is affected by many factors, including missing rate and minor allele frequency (MAF) of genotypic data, GS models, trait features, etc. In this study, we used one wheat population to investigate prediction accuracies of various GS models on yield and yield-related traits from various quality control (QC) scenarios, missing genotype imputation, and genome-wide association studies (GWAS)-derived markers. Missing rate and MAF of single nucleotide polymorphism (SNP) markers were two major factors in QC. Five missing rate levels (0%, 20%, 40%, 60%, and 80%) and three MAF levels (0%, 5%, and 10%) were considered and the five-fold cross validation was used to estimate the prediction accuracy. The results indicated that a moderate missing rate level (20% to 40%) and MAF (5%) threshold provided better prediction accuracy. Under this QC scenario, prediction accuracies were further calculated for imputed and GWAS-derived markers. It was observed that the accuracies of the six traits were related to their heritability and genetic architecture, as well as the GS prediction model. Moore–Penrose generalized inverse (GenInv), ridge regression (RidgeReg), and random forest (RForest) resulted in higher prediction accuracies than other GS models across traits. Imputation of missing genotypic data had marginal effect on prediction accuracy, while GWAS-derived markers improved the prediction accuracy in most cases. These results demonstrate that QC on missing rate and MAF had positive impact on the predictability of GS models. We failed to identify one single combination of QC scenarios that could outperform the others for all traits and GS models. However, the balance between marker number and marker quality is important for the deployment of GS in wheat breeding. GWAS is able to select markers which are mostly related to traits, and therefore can be used to improve the prediction accuracy of GS.

Genomic prediction of yield and root development in wheat under changing water availability

BACKGROUND

Deeper roots help plants take up available resources in deep soil ensuring better growth and higher yields under conditions of drought. A large-scale semi-field root phenotyping facility was developed to allow a water availability gradient and detect potential interaction of genotype by water availability gradient. Genotyped winter wheat lines were grown as rows in four beds of this facility, where indirect genetic effects from neighbors could be important to trait variation. The objective was to explore the possibility of genomic prediction for grain-related traits and deep root traits collected via images taken in a minirhizotron tube under each row of winter wheat measured.

RESULTS

The analysis comprised four grain-related traits: grain yield, thousand-kernel weight, protein concentration, and total nitrogen content measured on each half row that were harvested separately. Two root traits, total root length between 1.2 and 2 m depth and root length in four intervals on each tube were also analyzed. Two sets of models with or without the effects of neighbors from both sides of each row were applied. No interaction between genotypes and changing water availability were detected for any trait. Estimated genomic heritabilities ranged from 0.263 to 0.680 for grain-related traits and from 0.030 to 0.055 for root traits. The coefficients of genetic variation were similar for grain-related and root traits. The prediction accuracy of breeding values ranged from 0.440 to 0.598 for grain-related traits and from 0.264 to 0.334 for root traits. Including neighbor effects in the model generally increased the estimated genomic heritabilities and accuracy of predicted breeding values for grain yield and nitrogen content.

CONCLUSIONS

Similar relative amounts of additive genetic variance were found for both yield traits and root traits but no interaction between genotypes and water availability were detected. It is possible to obtain accurate genomic prediction of breeding values for grain-related traits and reasonably accurate predicted breeding values for deep root traits using records from the semi-field facility. Including neighbor effects increased the estimated additive genetic variance of grain-related traits and accuracy of predicting breeding values. High prediction accuracy can be obtained although heritability is low.

Genomic selection in multi-environment plant breeding trials using a factor analytic linear mixed model

Genomic selection (GS) is a statistical and breeding methodology designed to improve genetic gain. It has proven to be successful in animal breeding; however, key points of difference have not been fully considered in the transfer of GS from animal to plant breeding. In plant breeding, individuals (varieties) are typically evaluated across a number of locations in multiple years (environments) in formally designed comparative experiments, called multi-environment trials (METs). The design structure of individual trials can be complex and needs to be modelled appropriately. Another key feature of MET data sets is the presence of variety by environment interaction (VEI), that is the differential response of varieties to a change in environment. In this paper, a single-step factor analytic linear mixed model is developed for plant breeding MET data sets that incorporates molecular marker data, appropriately accommodates non-genetic sources of variation within trials and models VEI. A recently developed set of selection tools, which are natural derivatives of factor analytic models, are used to facilitate GS for a motivating data set from an Australian plant breeding company. The power and versatility of these tools is demonstrated for the variety by environment and marker by environment effects.

A hydroponics based high throughput screening system for Phytophthora root rot resistance in chickpea (Cicer arietinum L.)

BACKGROUND

Phytophthora root rot (PRR) caused by P. medicaginis is a major soil borne disease in chickpea growing regions of Australia. Sources of resistance have been identified in both cultivated and wild Cicer species. However, the molecular basis underlying PRR resistance is not known. Current phenotyping methods rely on mycelium slurry or oospore inoculum. Sensitive and reliable methods are desirable to study variation for PRR resistance in chickpea and allow for a controlled inoculation process to better capture early defence responses following PRR infection.

RESULTS

In this study, a procedure for P. medicaginis zoospore production was standardized and used as the inoculum to develop a hydroponics based in planta infection method to screen chickpea genotypes with established levels of PRR resistance. The efficiency of the system was both qualitatively validated based on observation of characteristic PRR symptom development, and quantitatively validated based on the amount of pathogen DNA in roots. This system was scaled up to screen two biparental mapping populations previously developed for PRR studies. For each of the screenings, plant survival time was measured after inoculation and used to derive Kaplan-Meier estimates of plant survival (KME-survival). KME-survival and canker length were then selected as phenotypic traits associated with PRR resistance. Genetic analysis of these traits was conducted which identified quantitative trait loci (QTL). Additionally, these hydroponic traits and a set of previously published plant survival traits obtained from multiple PRR field experiments were combined in a model-based correlation analysis. The results suggest that the underlying genetic basis for plant survival during PRR infection within hydroponics and field disease environments is linked. The QTL QRBprrkms03 and QRBprrck03 on chromosome 4 identified for the traits KME-survival and canker length, respectively, correspond to the same region reported for PRR resistance in a field disease experiment.

CONCLUSION

A hydroponics based screening system will facilitate reliable and rapid screening in both small- and large-scale experiments to study PRR disease in chickpea. It can be applied in chickpea breeding programs to screen for PRR resistance and classify the virulence of new and existing P. medicaginis isolates.

The effects of training population design on genomic prediction accuracy in wheat

Genomic selection offers several routes for increasing the genetic gain or efficiency of plant breeding programmes. In various species of livestock, there is empirical evidence of increased rates of genetic gain from the use of genomic selection to target different aspects of the breeder's equation. Accurate predictions of genomic breeding value are central to this, and the design of training sets is in turn central to achieving sufficient levels of accuracy. In summary, small numbers of close relatives and very large numbers of distant relatives are expected to enable predictions with higher accuracy. To quantify the effect of some of the properties of training sets on the accuracy of genomic selection in crops, we performed an extensive field-based winter wheat trial. In summary, this trial involved the construction of 44 F_2:4 bi- and tri-parental populations, from which 2992 lines were grown on four field locations and yield was measured. For each line, genotype data were generated for 25 K segregating SNP markers. The overall heritability of yield was estimated to 0.65, and estimates within individual families ranged between 0.10 and 0.85. Genomic prediction accuracies of yield BLUEs were 0.125-0.127 using two different cross-validation approaches and generally increased with training set size. Using related crosses in training and validation sets generally resulted in higher prediction accuracies than using unrelated crosses. The results of this study emphasise the importance of the training panel design in relation to the genetic material to which the resulting prediction model is to be applied.

From markers to genome-based breeding in wheat

Recent technological advances in wheat genomics provide new opportunities to uncover genetic variation in traits of breeding interest and enable genome-based breeding to deliver wheat cultivars for the projected food requirements for 2050. There has been tremendous progress in development of whole-genome sequencing resources in wheat and its progenitor species during the last 5 years. High-throughput genotyping is now possible in wheat not only for routine gene introgression but also for high-density genome-wide genotyping. This is a major transition phase to enable genome-based breeding to achieve progressive genetic gains to parallel to projected wheat production demands. These advances have intrigued wheat researchers to practice less pursued analytical approaches which were not practiced due to the short history of genome sequence availability. Such approaches have been successful in gene discovery and breeding applications in other crops and animals for which genome sequences have been available for much longer. These strategies include, (i) environmental genome-wide association studies in wheat genetic resources stored in genbanks to identify genes for local adaptation by using agroclimatic traits as phenotypes, (ii) haplotype-based analyses to improve the statistical power and resolution of genomic selection and gene mapping experiments, (iii) new breeding strategies for genome-based prediction of heterosis patterns in wheat, and (iv) ultimate use of genomics information to develop more efficient and robust genome-wide genotyping platforms to precisely predict higher yield potential and stability with greater precision. Genome-based breeding has potential to achieve the ultimate objective of ensuring sustainable wheat production through developing high yielding, climate-resilient wheat cultivars with high nutritional quality.

Optimising Genomic Selection in Wheat: Effect of Marker Density, Population Size and Population Structure on Prediction Accuracy

Genomic selection applied to plant breeding enables earlier estimates of a line’s performance and significant reductions in generation interval. Several factors affecting prediction accuracy should be well understood if breeders are to harness genomic selection to its full potential. We used a panel of 10,375 bread wheat (Triticum aestivum) lines genotyped with 18,101 SNP markers to investigate the effect and interaction of training set size, population structure and marker density on genomic prediction accuracy. Through assessing the effect of training set size we showed the rate at which prediction accuracy increases is slower beyond approximately 2,000 lines. The structure of the panel was assessed via principal component analysis and K-means clustering, and its effect on prediction accuracy was examined through a novel cross-validation analysis according to the K-means clusters and breeding cohorts. Here we showed that accuracy can be improved by increasing the diversity within the training set, particularly when relatedness between training and validation sets is low. The breeding cohort analysis revealed that traits with higher selection pressure (lower allelic diversity) can be more accurately predicted by including several previous cohorts in the training set. The effect of marker density and its interaction with population structure was assessed for marker subsets containing between 100 and 17,181 markers. This analysis showed that response to increased marker density is largest when using a diverse training set to predict between poorly related material. These findings represent a significant resource for plant breeders and contribute to the collective knowledge on the optimal structure of calibration panels for genomic prediction.

Genome-Wide Association Studies and Comparison of Models and Cross-Validation Strategies for Genomic Prediction of Quality Traits in Advanced Winter Wheat Breeding Lines

The aim of the this study was to identify SNP markers associated with five important wheat quality traits (grain protein content, Zeleny sedimentation, test weight, thousand-kernel weight, and falling number), and to investigate the predictive abilities of GBLUP and Bayesian Power Lasso models for genomic prediction of these traits. In total, 635 winter wheat lines from two breeding cycles in the Danish plant breeding company Nordic Seed A/S were phenotyped for the quality traits and genotyped for 10,802 SNPs. GWAS were performed using single marker regression and Bayesian Power Lasso models. SNPs with large effects on Zeleny sedimentation were found on chromosome 1B, 1D, and 5D. However, GWAS failed to identify single SNPs with significant effects on the other traits, indicating that these traits were controlled by many QTL with small effects. The predictive abilities of the models for genomic prediction were studied using different cross-validation strategies. Leave-One-Out cross-validations resulted in correlations between observed phenotypes corrected for fixed effects and genomic estimated breeding values of 0.50 for grain protein content, 0.66 for thousand-kernel weight, 0.70 for falling number, 0.71 for test weight, and 0.79 for Zeleny sedimentation. Alternative cross-validations showed that the genetic relationship between lines in training and validation sets had a bigger impact on predictive abilities than the number of lines included in the training set. Using Bayesian Power Lasso instead of GBLUP models, gave similar or slightly higher predictive abilities. Genomic prediction based on all SNPs was more effective than prediction based on few associated SNPs.