Integrating a growth degree-days based reaction norm methodology and multi-trait modeling for genomic prediction in wheat

Spectral-genomic chain-model approach enhances the wheat yield component prediction under the Mediterranean climate

In light of the changing climate that jeopardizes future food security, genomic selection is emerging as a valuable tool for breeders to enhance genetic gains and introduce high-yielding varieties. However, predicting grain yield is challenging due to the genetic and physiological complexities involved and the effect of genetic-by-environment interactions on prediction accuracy. We utilized a chained model approach to address these challenges, breaking down the complex prediction task into simpler steps. A diversity panel with a narrow phenological range was phenotyped across three Mediterranean environments for various morpho-physiological and yield-related traits. The results indicated that a multi-environment model outperformed a single-environment model in prediction accuracy for most traits. However, prediction accuracy for grain yield was not improved. Thus, in an attempt to ameliorate the grain yield prediction accuracy, we integrated a spectral estimation of spike number, being a major wheat yield component, with genomic data. A machine learning approach was used for spike number estimation from canopy hyperspectral reflectance captured by an unmanned aerial vehicle. The spectral-based estimated spike number was utilized as a secondary trait in a multi-trait genomic selection, significantly improving grain yield prediction accuracy. Moreover, the ability to predict the spike number based on data from previous seasons implies that it could be applied to new trials at various scales, even in small plot sizes. Overall, we demonstrate here that incorporating a novel spectral-genomic chain-model workflow, which utilizes spectral-based phenotypes as a secondary trait, improves the predictive accuracy of wheat grain yield.

Confidence intervals for validation statistics with data truncation in genomic prediction

BACKGROUND

Validation by data truncation is a common practice in genetic evaluations because of the interest in predicting the genetic merit of a set of young selection candidates. Two of the most used validation methods in genetic evaluations use a single data partition: predictivity or predictive ability (correlation between pre-adjusted phenotypes and estimated breeding values (EBV) divided by the square root of the heritability) and the linear regression (LR) method (comparison of "early" and "late" EBV). Both methods compare predictions with the whole dataset and a partial dataset that is obtained by removing the information related to a set of validation individuals. EBV obtained with the partial dataset are compared against adjusted phenotypes for the predictivity or EBV obtained with the whole dataset in the LR method. Confidence intervals for predictivity and the LR method can be obtained by replicating the validation for different samples (or folds), or bootstrapping. Analytical confidence intervals would be beneficial to avoid running several validations and to test the quality of the bootstrap intervals. However, analytical confidence intervals are unavailable for predictivity and the LR method.

RESULTS

We derived standard errors and Wald confidence intervals for the predictivity and statistics included in the LR method (bias, dispersion, ratio of accuracies, and reliability). The confidence intervals for the bias, dispersion, and reliability depend on the relationships and prediction error variances and covariances across the individuals in the validation set. We developed approximations for large datasets that only need the reliabilities of the individuals in the validation set. The confidence intervals for the ratio of accuracies and predictivity were obtained through the Fisher transformation. We show the adequacy of both the analytical and approximated analytical confidence intervals and compare them versus bootstrap confidence intervals using two simulated examples. The analytical confidence intervals were closer to the simulated ones for both examples. Bootstrap confidence intervals tend to be narrower than the simulated ones. The approximated analytical confidence intervals were similar to those obtained by bootstrapping.

CONCLUSIONS

Estimating the sampling variation of predictivity and the statistics in the LR method without replication or bootstrap is possible for any dataset with the formulas presented in this study.

Genomic prediction for root and yield traits of barley under a water availability gradient: a case study comparing different spatial adjustments

BACKGROUND

In drought periods, water use efficiency depends on the capacity of roots to extract water from deep soil. A semi-field phenotyping facility (RadiMax) was used to investigate above-ground and root traits in spring barley when grown under a water availability gradient. Above-ground traits included grain yield, grain protein concentration, grain nitrogen removal, and thousand kernel weight. Root traits were obtained through digital images measuring the root length at different depths. Two nearest-neighbor adjustments (M1 and M2) to model spatial variation were used for genetic parameter estimation and genomic prediction (GP). M1 and M2 used (co)variance structures and differed in the distance function to calculate between-neighbor correlations. M2 was the most developed adjustment, as accounted by the Euclidean distance between neighbors.

RESULTS

The estimated heritabilities ([Formula: see text]) ranged from low to medium for root and above-ground traits. The genetic coefficient of variation ([Formula: see text]) ranged from 3.2 to 7.0% for above-ground and 4.7 to 10.4% for root traits, indicating good breeding potential for the measured traits. The highest [Formula: see text] observed for root traits revealed that significant genetic change in root development can be achieved through selection. We studied the genotype-by-water availability interaction, but no relevant interaction effects were detected. GP was assessed using leave-one-line-out (LOO) cross-validation. The predictive ability (PA) estimated as the correlation between phenotypes corrected by fixed effects and genomic estimated breeding values ranged from 0.33 to 0.49 for above-ground and 0.15 to 0.27 for root traits, and no substantial variance inflation in predicted genetic effects was observed. Significant differences in PA were observed in favor of M2.

CONCLUSIONS

The significant [Formula: see text] and the accurate prediction of breeding values for above-ground and root traits revealed that developing genetically superior barley lines with improved root systems is possible. In addition, we found significant spatial variation in the experiment, highlighting the relevance of correctly accounting for spatial effects in statistical models. In this sense, the proposed nearest-neighbor adjustments are flexible approaches in terms of assumptions that can be useful for semi-field or field experiments.

Frost Damage Index: The Antipode of Growing Degree Days

Abiotic stresses such as heat and frost limit plant growth and productivity. Image-based field phenotyping methods allow quantifying not only plant growth but also plant senescence. Winter crops show senescence caused by cold spells, visible as declines in leaf area. We accurately quantified such declines by monitoring changes in canopy cover based on time-resolved high-resolution imagery in the field. Thirty-six winter wheat genotypes were measured in multiple years. A concept termed "frost damage index" (FDI) was developed that, in analogy to growing degree days, summarizes frost events in a cumulative way. The measured sensitivity of genotypes to the FDI correlated with visual scorings commonly used in breeding to assess winter hardiness. The FDI concept could be adapted to other factors such as drought or heat stress. While commonly not considered in plant growth modeling, integrating such degradation processes may be key to improving the prediction of plant performance for future climate scenarios.

Partitioning the forms of genotype-by-environment interaction in the reaction norm analysis of stability

KEY MESSAGE

The reaction norm analysis of stability can be enhanced by partitioning the contribution of different types of G × E to the variation in slope. The slope of regression in a reaction norm model, where the performance of a genotype is regressed over an environmental covariable, is often used as a measure of stability of genotype performance. This method could be developed further by partitioning variation in the slope of regression into the two sources of genotype-by-environment interaction (G × E) which cause it: scale-type G × E (heterogeneity of variance) and rank-type G × E (heterogeneity of correlation). Because the two types of G × E have very different properties, separating their effect would enable a clearer understanding of stability. The aim of this paper was to demonstrate two methods which seek to achieve this in reaction norm models. Reaction norm models were fit to yield data from a multi-environment trial in Barley (Hordeum vulgare), with the adjusted mean yield from each environment used as the environmental covariable. Stability estimated from factor-analytic models, which can disentangle the two types of G × E and estimate stability based on rank-type G × E, was used for comparison. Adjusting the reaction norm slope to account for scale-type G × E using a genetic regression more than tripled the correlation with factor-analytic estimates of stability (0.24-0.26 to 0.80-0.85), indicating that it removed variation in the reaction norm slope that originated from scale-type G × E. A standardisation procedure had a more modest increase (055-0.59) but could be useful when curvilinear reaction norms are required. Analyses which use reaction norms to explore the stability of genotypes could gain additional insight into the mechanisms of stability by applying the methods outlined in this study.

Genomic prediction for grain yield and micro-environmental sensitivity in winter wheat

Individuals within a common environment experience variations due to unique and non-identifiable micro-environmental factors. Genetic sensitivity to micro-environmental variation (i.e. micro-environmental sensitivity) can be identified in residuals, and genotypes with lower micro-environmental sensitivity can show greater resilience towards environmental perturbations. Micro-environmental sensitivity has been studied in animals; however, research on this topic is limited in plants and lacking in wheat. In this article, we aimed to (i) quantify the influence of genetic variation on residual dispersion and the genetic correlation between genetic effects on (expressed) phenotypes and residual dispersion for wheat grain yield using a double hierarchical generalized linear model (DHGLM); and (ii) evaluate the predictive performance of the proposed DHGLM for prediction of additive genetic effects on (expressed) phenotypes and its residual dispersion. Analyses were based on 2,456 advanced breeding lines tested in replicated trials within and across different environments in Denmark and genotyped with a 15K SNP-Illumina-BeadChip. We found that micro-environmental sensitivity for grain yield is heritable, and there is potential for its reduction. The genetic correlation between additive effects on (expressed) phenotypes and dispersion was investigated, and we observed an intermediate correlation. From these results, we concluded that breeding for reduced micro-environmental sensitivity is possible and can be included within breeding objectives without compromising selection for increased yield. The predictive ability and variance inflation for predictions of the DHGLM and a linear mixed model allowing heteroscedasticity of residual variance in different environments (LMM-HET) were evaluated using leave-one-line-out cross-validation. The LMM-HET and DHGLM showed good and similar performance for predicting additive effects on (expressed) phenotypes. In addition, the accuracy of predicting genetic effects on residual dispersion was sufficient to allow genetic selection for resilience. Such findings suggests that DHGLM may be a good choice to increase grain yield and reduce its micro-environmental sensitivity.