Accuracy of Genomic Prediction in Switchgrass (Panicum virgatum L.) Improved by Accounting for Linkage Disequilibrium

Physiological Responses of C4 Perennial Bioenergy Grasses to Climate Change: Causes, Consequences, and Constraints

C4 perennial bioenergy grasses are an economically and ecologically important group whose responses to climate change will be important to the future bioeconomy. These grasses are highly productive and frequently possess large geographic ranges and broad environmental tolerances, which may contribute to the evolution of ecotypes that differ in physiological acclimation capacity and the evolution of distinct functional strategies. C4 perennial bioenergy grasses are predicted to thrive under climate change-C4 photosynthesis likely evolved to enhance photosynthetic efficiency under stressful conditions of low [CO2], high temperature, and drought-although few studies have examined how these species will respond to combined stresses or to extremes of temperature and precipitation. Important targets for C4 perennial bioenergy production in a changing world, such as sustainability and resilience, can benefit from combining knowledge of C4 physiology with recent advances in crop improvement, especially genomic selection.

Forecasting trait responses in novel environments to aid seed provenancing under climate change

Revegetation projects face the major challenge of sourcing optimal plant material. This is often done with limited information about plant performance and increasingly requires factoring resilience to climate change. Functional traits can be used as quantitative indices of plant performance and guide seed provenancing, but trait values expected under novel conditions are often unknown. To support climate-resilient provenancing efforts, we develop a trait prediction model that integrates the effect of genetic variation with fine-scale temperature variation. We train our model on multiple field plantings of Arabidopsis thaliana and predict two relevant fitness traits-days-to-bolting and fecundity-across the species' European range. Prediction accuracy was high for days-to-bolting and moderate for fecundity, with the majority of trait variation explained by temperature differences between plantings. Projection under future climate predicted a decline in fecundity, although this response was heterogeneous across the range. In response, we identified novel genotypes that could be introduced to genetically offset the fitness decay. Our study highlights the value of predictive models to aid seed provenancing and improve the success of revegetation projects.

Genomic Prediction of Complex Traits in Forage Plants Species: Perennial Grasses Case

The majority of forage grass species are obligate outbreeders. Their breeding classically consists of an initial selection on spaced plants for highly heritable traits such as disease resistances and heading date, followed by familial selection on swards for forage yield and quality traits. The high level of diversity and heterozygosity, and associated decay of linkage disequilibrium (LD) over very short genomic distances, has hampered the implementation of genomic selection (GS) in these species. However, next generation sequencing technologies in combination with the development of genomic resources have recently facilitated implementation of GS in forage grass species such as perennial ryegrass (Lolium perenne L.), switchgrass (Panicum virgatum L.), and timothy (Phleum pratense L.). Experimental work and simulations have shown that GS can increase significantly the genetic gain per unit of time for traits with different levels of heritability. The main reasons are (1) the possibility to select single plants based on their genomic estimated breeding values (GEBV) for traits measured at sward level, (2) a reduction in the duration of selection cycles, and less importantly (3) an increase in the selection intensity associated with an increase in the genetic variance used for selection. Nevertheless, several factors should be taken into account for the successful implementation of GS in forage grasses. For example, it has been shown that the level of relatedness between the training and the selection population is particularly critical when working with highly structured meta-populations consisting of several genetic groups. A sufficient number of markers should be used to estimate properly the kinship between individuals and to reflect the variability of major QTLs. It is also important that the prediction models are trained for relevant environments when dealing with traits with high genotype × environment interaction (G × E). Finally, in these outbreeding species, measures to reduce inbreeding should be used to counterbalance the high selection intensity that can be achieved in GS.

Subsampling and DNA pooling can increase gains through genomic selection in switchgrass

Genomic selection (GS) can accelerate breeding cycles in perennial crops such as the bioenergy grass switchgrass (Panicum virgatum L.). The sequencing costs of GS can be reduced by pooling DNA samples in the training population (TP), only sequencing TP phenotypic outliers, or pooling candidate population (CP) samples. These strategies were simulated for two traits (spring vigor and anthesis date) in three breeding populations. Sequencing only the outlier 50% of the TP phenotype distribution resulted in a penalty of <5% of the predictive ability, measured using cross-validation. Predictive ability also decreased when sequencing progressively fewer TP DNA pools, but TPs constructed from only two phenotypically contrasting DNA samples retained a mean of >80% predictive ability relative to individual TP sequencing. Novel group testing methods allowed greater than one CP individual to be screened per sequenced DNA sample but resulted in a predictive ability penalty. To determine the impact of reduced sequencing, genetic gain was calculated for seven GS scenarios with variable sequencing budgets. Reduced TP sequencing and most CP pooling methods were superior to individual sequence-based GS when sequencing resources were restricted (2,000 DNA samples per 5-yr cycle). Only one scenario was superior to individual sequencing when sequencing budgets were large (8,000 DNA samples per 5-yr cycle). This study highlights multiple routes for reduced sequencing costs in GS.

Genetic loci associated with winter survivorship in diverse lowland switchgrass populations

High winter mortality limits biomass yield of lowland switchgrass (Panicum virgatum L.) planted in the northern latitudes of North America. Breeding of cold tolerant switchgrass cultivars requires many years due to its perennial growth habit and the unpredictable winter selection pressure that is required to identify winter-hardy individuals. Identification of causal genetic variants for winter survivorship would accelerate the improvement of switchgrass biomass production. The objective of this study was to identify allelic variation associated with winter survivorship in lowland switchgrass populations using bulk segregant analysis (BSA). Twenty-nine lowland switchgrass populations were evaluated for winter survival at two locations in southern Wisconsin and 21 populations with differential winter survivorship were used for BSA. A maximum of 10% of the individuals (8-20) were bulked to create survivor and nonsurvivor DNA pools from each population and location. The DNA pools were evaluated using exome capture sequencing, and allele frequencies were used to conduct statistical tests. The BSA tests revealed nine quatitative trait loci (QTL) from tetraploid populations and seven QTL from octoploid populations. Many QTL were population-specific, but some were identified in multiple populations that originated across a broad geographic landscape. Four QTL (at positions 88 Mb on chromosome 2N, 115 Mb on chromosome 5K, and 1 and 100 Mb on chromosome 9N) were potentially the most useful QTL. Markers associated with winter survivorship in this study can be used to accelerate breeding cycles of lowland switchgrass populations and should lead to improvements in adaptation within USDA hardiness zones 4 and 5.

Pairwise linkage disequilibrium estimation for polyploids

Many tasks in statistical genetics involve pairwise estimation of linkage disequilibrium (LD). The study of LD in diploids is mature. However, in polyploids, the field lacks a comprehensive characterization of LD. Polyploids also exhibit greater levels of genotype uncertainty than diploids, yet no methods currently exist to estimate LD in polyploids in the presence of such genotype uncertainty. Furthermore, most LD estimation methods do not quantify the level of uncertainty in their LD estimates. Our study contains three major contributions. (i) We characterize haplotypic and composite measures of LD in polyploids. These composite measures of LD turn out to be functions of common statistical measures of association. (ii) We derive procedures to estimate haplotypic and composite LD in polyploids in the presence of genotype uncertainty. We do this by estimating LD directly from genotype likelihoods, which may be obtained from many genotyping platforms. (iii) We derive standard errors of all LD estimators that we discuss. We validate our methods on both real and simulated data. Our methods are implemented in the R package ldsep, available on the Comprehensive R Archive Network https://cran.r-project.org/package=ldsep.

Advancing crested wheatgrass [Agropyron cristatum (L.) Gaertn.] breeding through genotyping-by-sequencing and genomic selection

Crested wheatgrass [Agropyron cristatum (L.) Gaertn.] provides high quality, highly palatable forage for early season grazing. Genetic improvement of crested wheatgrass has been challenged by its complex genome, outcrossing nature, long breeding cycle, and lack of informative molecular markers. Genomic selection (GS) has potential for improving traits of perennial forage species, and genotyping-by-sequencing (GBS) has enabled the development of genome-wide markers in non-model polyploid plants. An attempt was made to explore the utility of GBS and GS in crested wheatgrass breeding. Sequencing and phenotyping 325 genotypes representing 10 diverse breeding lines were performed. Bioinformatics analysis identified 827, 3,616, 14,090 and 46,136 single nucleotide polymorphism markers at 20%, 30%, 40% and 50% missing marker levels, respectively. Four GS models (BayesA, BayesB, BayesCπ, and rrBLUP) were examined for the accuracy of predicting nine agro-morphological and three nutritive value traits. Moderate accuracy (0.20 to 0.32) was obtained for the prediction of heading days, leaf width, plant height, clump diameter, tillers per plant and early spring vigor for genotypes evaluated at Saskatoon, Canada. Similar accuracy (0.29 to 0.35) was obtained for predicting fall regrowth and plant height for genotypes evaluated at Swift Current, Canada. The Bayesian models displayed similar or higher accuracy than rrBLUP. These findings show the feasibility of GS application for a non-model species to advance plant breeding.

High-Throughput Genome-Wide Genotyping To Optimize the Use of Natural Genetic Resources in the Grassland Species Perennial Ryegrass (Lolium perenne L.)

The natural genetic diversity of agricultural species is an essential genetic resource for breeding programs aiming to improve their ecosystem and production services. A large natural ecotype diversity is usually available for most grassland species. This could be used to recombine natural climatic adaptations and agronomic value to create improved populations of grassland species adapted to future regional climates. However describing natural genetic resources can be long and costly. Molecular markers may provide useful information to help this task. This opportunity was investigated for Lolium perenne L., using a set of 385 accessions from the natural diversity of this species collected right across Europe and provided by genebanks of several countries. For each of these populations, genotyping provided the allele frequencies of 189,781 SNP markers. GWAS were implemented for over 30 agronomic and/or putatively adaptive traits recorded in three climatically contrasted locations (France, Belgium, Germany). Significant associations were detected for hundreds of markers despite a strong confounding effect of the genetic background; most of them pertained to phenology traits. It is likely that genetic variability in these traits has had an important contribution to environmental adaptation and ecotype differentiation. Genomic prediction models calibrated using natural diversity were found to be highly effective to describe natural populations for almost all traits as well as commercial synthetic populations for some important traits such as disease resistance, spring growth or phenological traits. These results will certainly be valuable information to help the use of natural genetic resources of other species.

Benchmarking Parametric and Machine Learning Models for Genomic Prediction of Complex Traits

The usefulness of genomic prediction in crop and livestock breeding programs has prompted efforts to develop new and improved genomic prediction algorithms, such as artificial neural networks and gradient tree boosting. However, the performance of these algorithms has not been compared in a systematic manner using a wide range of datasets and models. Using data of 18 traits across six plant species with different marker densities and training population sizes, we compared the performance of six linear and six non-linear algorithms. First, we found that hyperparameter selection was necessary for all non-linear algorithms and that feature selection prior to model training was critical for artificial neural networks when the markers greatly outnumbered the number of training lines. Across all species and trait combinations, no one algorithm performed best, however predictions based on a combination of results from multiple algorithms (i.e., ensemble predictions) performed consistently well. While linear and non-linear algorithms performed best for a similar number of traits, the performance of non-linear algorithms vary more between traits. Although artificial neural networks did not perform best for any trait, we identified strategies (i.e., feature selection, seeded starting weights) that boosted their performance to near the level of other algorithms. Our results highlight the importance of algorithm selection for the prediction of trait values.

Genomic Prediction for Winter Survival of Lowland Switchgrass in the Northern USA

The lowland ecotype of switchgrass has generated considerable interest because of its higher biomass yield and late flowering characteristics compared to the upland ecotype. However, lowland ecotypes planted in northern latitudes exhibit very low winter survival. Implementation of genomic selection could potentially enhance switchgrass breeding for winter survival by reducing generation time while eliminating the dependence on weather. The objectives of this study were to assess the potential of genomic selection for winter survival in lowland switchgrass by combining multiple populations in the training set and applying the selected model in two independent testing datasets for validation. Marker data were generated using exome capture sequencing. Validation was conducted using (1) indirect indicators of winter adaptation based on geographic and climatic variables of accessions from different source locations and (2) winter survival estimates of the phenotype. The prediction accuracies were significantly higher when the training dataset comprising all populations was used in fivefold cross validation but its application was not useful in the independent validation dataset. Nevertheless, modeling for population heterogeneity improved the prediction accuracy to some extent but the genetic relationship between the training and validation populations was found to be more influential. The predicted winter survival of lowland switchgrass indicated latitudinal and longitudinal variability, with the northeast USA the region for most cold tolerant lowland populations. Our results suggested that GS could provide valuable opportunities for improving winter survival and accelerate the lowland switchgrass breeding programs toward the development of cold tolerant cultivars suitable for northern latitudes.

Extensions of BLUP Models for Genomic Prediction in Heterogeneous Populations: Application in a Diverse Switchgrass Sample

Genomic prediction is a useful tool to accelerate genetic gain in selection using DNA marker information. However, this technology typically relies on standard prediction procedures, such as genomic BLUP, that are not designed to accommodate population heterogeneity resulting from differences in marker effects across populations. In this study, we assayed different prediction procedures to capture marker-by-population interactions in genomic prediction models. Prediction procedures included genomic BLUP and two kernel-based extensions of genomic BLUP which explicitly accounted for population heterogeneity. To model population heterogeneity, dissemblance between populations was either depicted by a unique coefficient (as previously reported), or a more flexible function of genetic distance between populations (proposed herein). Models under investigation were applied in a diverse switchgrass sample under two validation schemes: whole-sample calibration, where all individuals except selection candidates are included in the calibration set, and cross-population calibration, where the target population is entirely excluded from the calibration set. First, we showed that using fixed effects, from principal components or putative population groups, appeared detrimental to prediction accuracy, especially in cross-population calibration. Then we showed that modeling population heterogeneity by our proposed procedure resulted in highly significant improvements in model fit. In such cases, gains in accuracy were often positive. These results suggest that population heterogeneity may be parsimoniously captured by kernel methods. However, in cases where improvement in model fit by our proposed procedure is null-to-moderate, ignoring heterogeneity should probably be preferred due to the robustness and simplicity of the standard genomic BLUP model.

Breeding progress and preparedness for mass-scale deployment of perennial lignocellulosic biomass crops switchgrass, miscanthus, willow and poplar

Genetic improvement through breeding is one of the key approaches to increasing biomass supply. This paper documents the breeding progress to date for four perennial biomass crops (PBCs) that have high output-input energy ratios: namely Panicum virgatum (switchgrass), species of the genera Miscanthus (miscanthus), Salix (willow) and Populus (poplar). For each crop, we report on the size of germplasm collections, the efforts to date to phenotype and genotype, the diversity available for breeding and on the scale of breeding work as indicated by number of attempted crosses. We also report on the development of faster and more precise breeding using molecular breeding techniques. Poplar is the model tree for genetic studies and is furthest ahead in terms of biological knowledge and genetic resources. Linkage maps, transgenesis and genome editing methods are now being used in commercially focused poplar breeding. These are in development in switchgrass, miscanthus and willow generating large genetic and phenotypic data sets requiring concomitant efforts in informatics to create summaries that can be accessed and used by practical breeders. Cultivars of switchgrass and miscanthus can be seed-based synthetic populations, semihybrids or clones. Willow and poplar cultivars are commercially deployed as clones. At local and regional level, the most advanced cultivars in each crop are at technology readiness levels which could be scaled to planting rates of thousands of hectares per year in about 5 years with existing commercial developers. Investment in further development of better cultivars is subject to current market failure and the long breeding cycles. We conclude that sustained public investment in breeding plays a key role in delivering future mass-scale deployment of PBCs.

Genomic Prediction of Complex Phenotypes Using Genic Similarity Based Relatedness Matrix

In the last years, a series of methods for genomic prediction (GP) have been established, and the advantages of GP over pedigree best linear unbiased prediction (BLUP) have been reported. However, the majority of previously proposed GP models are purely based on mathematical considerations while seldom take the abundant biological knowledge into account. Prediction ability of those models largely depends on the consistency between the statistical assumptions and the underlying genetic architectures of traits of interest. In this study, gene annotation information was incorporated into GP models by constructing haplotypes with SNPs mapped to genic regions. Haplotype allele similarity between pairs of individuals was measured through different approaches at single gene level and then converted into whole genome level, which was then treated as a special kernel and used in kernel based GP models. Results shown that the gene annotation guided methods gave higher or at least comparable predictive ability in some traits, especially in the Arabidopsis dataset and the rice breeding population. Compared to SNP models and haplotype models without gene annotation, the gene annotation based models improved the predictive ability by 0.56~26.67% in the Arabidopsis and 1.62~16.53% in the rice breeding population, respectively. However, incorporating gene annotation slightly improved the predictive ability for several traits but did not show any extra gain for the rest traits in a chicken population. In conclusion, integrating gene annotation into GP models could be beneficial for some traits, species, and populations compared to SNP models and haplotype models without gene annotation. However, more studies are yet to be conducted to implicitly investigate the characteristics of these gene annotation guided models.

Genomic prediction accuracy for switchgrass traits related to bioenergy within differentiated populations

BACKGROUND

Switchgrass breeders need to improve the rates of genetic gain in many bioenergy-related traits in order to create improved cultivars that are higher yielding and have optimal biomass composition. One way to achieve this is through genomic selection. However, the heritability of traits needs to be determined as well as the accuracy of prediction in order to determine if efficient selection is possible.

RESULTS

Using five distinct switchgrass populations comprised of three lowland, one upland and one hybrid accession, the accuracy of genomic predictions under different cross-validation strategies and prediction methods was investigated. Individual genotypes were collected using GBS while kin-BLUP, partial least squares, sparse partial least squares, and BayesB methods were employed to predict yield, morphological, and NIRS-based compositional data collected in 2012-2013 from a replicated Nebraska field trial. Population structure was assessed by F statistics which ranged from 0.3952 between lowland and upland accessions to 0.0131 among the lowland accessions. Prediction accuracy ranged from 0.57-0.52 for cell wall soluble glucose and fructose respectively, to insignificant for traits with low repeatability. Ratios of heritability across to within-population ranged from 15 to 0.6.

CONCLUSIONS

Accuracy was significantly affected by both cross-validation strategy and trait. Accounting for population structure with a cross-validation strategy constrained by accession resulted in accuracies that were 69% lower than apparent accuracies using unconstrained cross-validation. Less accurate genomic selection is anticipated when most of the phenotypic variation exists between populations such as with spring regreening and yield phenotypes.

Predictive ability of genomic selection models in a multi-population perennial ryegrass training set using genotyping-by-sequencing

KEY MESSAGE

Genomic prediction models for multi-year dry matter yield, via genotyping-by-sequencing in a composite training set, demonstrate potential for genetic gain improvement through within-half sibling family selection. Perennial ryegrass (Lolium perenne L.) is a key source of nutrition for ruminant livestock in temperate environments worldwide. Higher seasonal and annual yield of herbage dry matter (DMY) is a principal breeding objective but the historical realised rate of genetic gain for DMY is modest. Genomic selection was investigated as a tool to enhance the rate of genetic gain. Genotyping-by-sequencing (GBS) was undertaken in a multi-population (MP) training set of five populations, phenotyped as half-sibling (HS) families in five environments over 2 years for mean herbage accumulation (HA), a measure of DMY potential. GBS using the ApeKI enzyme yielded 1.02 million single-nucleotide polymorphism (SNP) markers from a training set of n = 517. MP-based genomic prediction models for HA were effective in all five populations, cross-validation-predictive ability (PA) ranging from 0.07 to 0.43, by trait and target population, and 0.40-0.52 for days-to-heading. Best linear unbiased predictor (BLUP)-based prediction methods, including GBLUP with either a standard or a recently developed (KGD) relatedness estimation, were marginally superior or equal to ridge regression and random forest computational approaches. PA was principally an outcome of SNP modelling genetic relationships between training and validation sets, which may limit application for long-term genomic selection, due to PA decay. However, simulation using data from the training experiment indicated a twofold increase in genetic gain for HA, when applying a prediction model with moderate PA in a single selection cycle, by combining among-HS family selection, based on phenotype, with within-HS family selection using genomic prediction.

Extensive Genetic Diversity is Present within North American Switchgrass Germplasm

Switchgrass ( is a perennial native North American grass present in two ecotypes: upland, found primarily in the northern range of switchgrass habitats, and lowland, found largely in the southern reaches of switchgrass habitats. Previous studies focused on a diversity panel of primarily northern switchgrass, so to expand our knowledge of genetic diversity in a broader set of North American switchgrass, exome capture sequence data were generated for 632 additional, primarily lowland individuals. In total, over 37 million single nucleotide polymorphisms (SNPs) were identified and a set of 1.9 million high-confidence SNPs were obtained from 1169 individuals from 140 populations (67 upland, 65 lowland, 8 admixed) were used in downstream analyses of genetic diversity and population structure. Seven separate population groups were identified with moderate genetic differentiation [mean fixation index (Fst) estimate of 0.06] between the lowland and the upland populations. Ecotype-specific and population-specific SNPs were identified for use in germplasm evaluations. Relative to rice ( L.), maize ( L.), soybean [ (L.) Merr.], and Gaertn., analyses of nucleotide diversity revealed a high degree of genetic diversity (0.0135) across all individuals, consistent with the outcrossing mode of reproduction and the polyploidy of switchgrass. This study supports the hypothesis that repeated glaciation events, ploidy barriers, and restricted gene flow caused by flowering time differences have resulted in distinct gene pools across ecotypes and geographic regions. These data provide a resource to associate alleles with traits of interest for forage, restoration, and biofuel feedstock efforts in switchgrass.

Rice diversity panel provides accurate genomic predictions for complex traits in the progenies of biparental crosses involving members of the panel

KEY MESSAGE

Rice breeding programs based on pedigree schemes can use a genomic model trained with data from their working collection to predict performances of progenies produced through rapid generation advancement. So far, most potential applications of genomic prediction in plant improvement have been explored using cross validation approaches. This is the first empirical study to evaluate the accuracy of genomic prediction of the performances of progenies in a typical rice breeding program. Using a cross validation approach, we first analyzed the effects of marker selection and statistical methods on the accuracy of prediction of three different heritability traits in a reference population (RP) of 284 inbred accessions. Next, we investigated the size and the degree of relatedness with the progeny population (PP) of sub-sets of the RP that maximize the accuracy of prediction of phenotype across generations, i.e., for 97 F5-F7 lines derived from biparental crosses between 31 accessions of the RP. The extent of linkage disequilibrium was high (r 2 = 0.2 at 0.80 Mb in RP and at 1.1 Mb in PP). Consequently, average marker density above one per 22 kb did not improve the accuracy of predictions in the RP. The accuracy of progeny prediction varied greatly depending on the composition of the training set, the trait, LD and minor allele frequency. The highest accuracy achieved for each trait exceeded 0.50 and was only slightly below the accuracy achieved by cross validation in the RP. Our results thus show that relatively high accuracy (0.41-0.54) can be achieved using only a rather small share of the RP, most related to the PP, as the training set. The practical implications of these results for rice breeding programs are discussed.

Incorporating Gene Annotation into Genomic Prediction of Complex Phenotypes

Today, genomic prediction (GP) is an established technology in plant and animal breeding programs. Current standard methods are purely based on statistical considerations but do not make use of the abundant biological knowledge, which is easily available from public databases. Major questions that have to be answered before biological prior information can be used routinely in GP approaches are which types of information can be used, and at which points they can be incorporated into prediction methods. In this study, we propose a novel strategy to incorporate gene annotation into GP of complex phenotypes by defining haploblocks according to gene positions. Haplotype effects are then modeled as categorical or as numerical allele dosage variables. The underlying concept of this approach is to build the statistical model on variables representing the biologically functional units. We evaluate the new methods with data from a heterogeneous stock mouse population, the Drosophila Genetic Reference Panel (DGRP), and a rice breeding population from the Rice Diversity Panel. Our results show that using gene annotation to define haploblocks often leads to a comparable, but for some traits to a higher, predictive ability compared to SNP-based models or to haplotype models that do not use gene annotation information. Modeling gene interaction effects can further improve predictive ability. We also illustrate that the additional use of markers that have not been mapped to any gene in a second separate relatedness matrix does in many cases not lead to a relevant additional increase in predictive ability when the first matrix is based on haploblocks defined with gene annotation data, suggesting that intergenic markers only provide redundant information on the considered data sets. Therefore, gene annotation information seems to be appropriate to perceive the importance of DNA segments. Finally, we discuss the effects of gene annotation quality, marker density, and linkage disequilibrium on the performance of the new methods. To our knowledge, this is the first work that incorporates epistatic interaction or gene annotation into haplotype-based prediction approaches.