Genomic Prediction Accounting for Genotype by Environment Interaction Offers an Effective Framework for Breeding Simultaneously for Adaptation to an Abiotic Stress and Performance Under Normal Cropping Conditions in Rice

Abstract

Developing rice varieties adapted to alternate wetting and drying water management is crucial for the sustainability of irrigated rice cropping systems. Here we report the first study exploring the feasibility of breeding rice for adaptation to alternate wetting and drying using genomic prediction methods that account for genotype by environment interactions. Two breeding populations (a reference panel of 284 accessions and a progeny population of 97 advanced lines) were evaluated under alternate wetting and drying and continuous flooding management systems. The predictive ability of genomic prediction for response variables (index of relative performance and the slope of the joint regression) and for multi-environment genomic prediction models were compared. For the three traits considered (days to flowering, panicle weight and nitrogen-balance index), significant genotype by environment interactions were observed in both populations. In cross validation, predictive ability for the index was on average lower (0.31) than that of the slope of the joint regression (0.64) whatever the trait considered. Similar results were found for progeny validation. Both cross-validation and progeny validation experiments showed that the performance of multi-environment models predicting unobserved phenotypes of untested entrees was similar to the performance of single environment models with differences in predictive ability ranging from -6–4% depending on the trait and on the statistical model concerned. The predictive ability of multi-environment models predicting unobserved phenotypes of entrees evaluated under both water management systems outperformed single environment models by an average of 30%. Practical implications for breeding rice for adaptation to alternate wetting and drying system are discussed.

Selection of trait-specific markers and multi-environment models improve genomic predictive ability in rice

“…While in the case of wheat, data from a multi-local trial across a natural continuum of environments were grouped in four subsets, our data were produced in a managed-environment trial with clear-cut treatments to assess the effect of drought stress on rice development end yield. Overall, these results reinforce the conclusion drawn by [42] that, multi-environment genomic prediction models that account for G×E interactions as evaluated from multi-local trials, are also of interest for breeding for tolerance to abiotic stresses, including drought.…”

“…The RKHS multi-environment model enabled gains in PA of up to 32%. These gains are much lower than those of up to 68% reported in [39] in wheat, similar to the ones reported by [42], and in accordance with [41], higher than the GBLUP multi-environment model. The differences in the amplitude of gains of PA observed in our study and the ones reported in [39,41] are probably due to several factors, including the size of the population (599 wheat lines against 204 accessions in our case), the number of environment (large number of environments that were grouped in four target sets of environments versus three in our case), and the distinctive features of those environments.…”

“…These gains were significantly lower than the 30% gain reported in [38] comparing multi-environment and single environment GBLUP models for GY in wheat. This was also the case for the gains in PA reported in [42] using rice data from two managed environments, continued flooding and alternate watering and drying.…”

“…the genomic best linear unbiased prediction (GBLUP) and the reproducing kernel Hilbert spaces regressions, RKHS, [51]. The implementation procedures of the two models is detailed in [42].…”

“…Application of these multi-environment models to data from multilocal trials of CIMMYT' maize and wheat breeding programs confirmed their superiority over the single environment models. Multi-environment prediction models were also reported to be better than single environment models when used with rice phenotypic data from managed environments representing two water management systems: continuous flooding and alternate watering and drying [42]. However, the differences between the PA of different categories multienvironment models were not significant.…”