Search

International audience; Crop species rely on genetic and phenotypic diversity to adapt to various environments. Genotypes respond differently, resulting in genotype-by-environment interactions (GxE). Recent advances in phenomics and modeling allow for the prediction of GxE for traits related to sustainability and productivity. However, a phenotype-centered framework that effectively navigates the interplay between genetics, physiology, and the environment is needed. Here we present maize data from phenotyping platforms and field experiments and explore strategies to integrate concepts and techniques from ecophysiology, ecology, and math to: (i) define and describe phenotypic space across spatial, temporal, and biological scales; (ii) identify unique phenotypic combinations and constraints; and (iii) identify territories of the phenotypic space representing favorable adaptations for crop improvement. Our findings demonstrate the value of phenotype-focused perspectives for adapting crops to climate change.

The quality of yield prediction is linked to that of leaf area. We first analysed the consequences of flowering time and environmental conditions on the area of individual leaves in 127 genotypes presenting contrasting flowering times in fields of Europe, Mexico, and Kenya. Flowering time was the strongest determinant of leaf area. Combined with a detailed field experiment, this experiment showed a large effect of flowering time on the final leaf number and on the distribution of leaf growth rate and growth duration along leaf ranks, in terms of both length and width. Equations with a limited number of genetic parameters predicted the beginning, end, and maximum growth rate (length and width) for each leaf rank. The genotype-specific environmental effects were analysed with datasets in phenotyping platforms that assessed the effects (i) of the amount of intercepted light on leaf width, and (ii) of temperature, evaporative demand, and soil water potential on leaf elongation rate. The resulting model was successfully tested for 31 hybrids in 15 European and Mexican fields. It potentially allows prediction of the vertical distribution of leaf area of a large number of genotypes in contrasting field conditions, based on phenomics and on sensor networks., Flowering time affects all processes governing whole-plant leaf area, namely rates, durations, and sensitivity to environmental conditions. Genotype dependency was assessed via large phenomic datasets in field and controlled conditions.

International audience; There is potential sources of alleles and genes currently locked into wheat-related species that could be introduced into wheat breeding programs for current and future hot and dry climates. However, neither the intra- nor the inter-specific diversity of the responses of leaf growth and transpiration to temperature and evaporative demand have been investigated in a large diversity of wheat-related species. By analysing 12 groups of wheat-related sub-species, we questioned the n-dimensional structure of the genetic diversity for traits linked to plant vegetative structures and development, leaf expansion and transpiration together with their responses to “non-stressing” range of temperature and evaporative demand. In addition to provide new insight on how genome type, ploidy level, phylogeny and breeding pressure together structure this genetic diversity, this study provides new mathematical formalisms and the associated parameters of trait responses in the large genetic diversity of wheat-related species. This potentially allow crop models predicting the impact of this diversity on yield, and indicate potential sources of varietal improvement for modern wheat germplasms, through interspecific crosses.

Impact de la variabilité intra-et inter-spécifique des réponses à la température et au déficit hydrique dans la diversité des scénarios E x M

International audience

Crop models were developed by ecophysiologists to describe plant development. They allow in particular to report difference existing between several genotypes in several environments, due to genotype by environment interaction. It is first necessary to calibrate these models to use them for prediction purpose. We consider the crop model APSIM and present a joint bayesian model with mixed effects. We infer models parametervalues from data collected in the field and in phenotyping platform. Prior distribution are chosen in order to integrate expert knowledge. We implement an hybrid Gibbs algorithm to simulate the posterior distribution. Results obtained from simulated and real data highlight clearly the advantage of using phenotyping platform data in addition to field data., Les modèles de culture élaborés par des écophysiologistes décrivent les processus de développement d"une plante. Ils permettent en particulier de rendre compte des différences de comportement de plusieurs variétés dans différents environnements, dues aux interactions génotype-environnement. Pour les utiliser à des fins prédictives, il est nécessaire de calibrer auparavant leurs paramètres. Nous considérons le modèle de culture APSIM et proposons un modèle joint bayésien à effets mixtes dans lequel nous inférons la valeur des paramètres inconnus à partir de données issues d'expérience de plein champ et mesurées en plateforme de phénotypage. Nous choisissons des lois a priori informatives pour intégrer les connaissances d'expert et implémentons un algorithme de type Gibbs hybride pour simuler la loi a posteriori. Les résultats obtenus sur données simulées et réelles mettent en évidence le gain obtenu sur la précision des estimations en utilisant les données issues de plateforme de phénotypage en sus des données de champ.

Plants are subjected every day to rapid variation of evaporative demand and soil water availability, resulting in rapid changes in stomatal conductance, expansive growth and metabolism over minutes. Because yield involves several months, the connection between physiological mechanisms and response of yield to drought scenarios faces a massive problem of time scales. Furthermore, yield results from optimization between traits and alleles that lead to either minimize the risk of crop failure or to increase crop production. Evolution has tended to favour conservative processes (short crop cycle, low transpiration and leaf area, large root systems) which are favourable under severe stresses, whereas yield in milder water deficits is associated with the opposite traits. Hence, one aims at identifying which traits and alleles are favourable in which drought scenarios, rather than at a generic ‘drought tolerance’. We deal with these methodological difficulties by combining phenomics, modelling, genetic analysis and genomic prediction. A first strategy explores the genetic variability of key processes, which are translated into parameters of a crop model. This requires detailed analyses in phenotyping platforms with a capacity of thousands of plants, with the relevant time scales. These parameters are analysed by GWAS and simulated via genomic prediction. The model can then simulate yield in hundreds of fields for hundreds of genotypes, from genetic parameters of each genotype and environmental conditions in each field. A second strategy directly explores the responses of yield to environmental conditions in contrasting environmental scenarios, e.g. in 40 fields. This results in a mixed model whose parameters are analysed genetically and can be estimated by genomic prediction, thereby allowing one to predict yields in new genotypes and fields. As a whole, the combination of field and platform data allows identification of combination

Crop improvement for drought is based on the selection of alleles that increase yield in dry or hot conditions. The Genotype by Environment Interactions (GxE) is typically high in these environments, with alleles conferring either positive or negative effects, depending on drought scenarios (Tardieu, 2012). Rather than trying to over-simplify GxE, for instance, in managed drought experiments, we propose an integrative approach using genome wide association studies (GWAS), phenotyping and modelling. It aims at predicting in which drought scenarios a combination of trait/allele could confer advantages (Parent and Tardieu, 2015). Indeed, (i) we phenotype the intra- and inter- specific variability of development and growth responses to temperature, evaporative demand and water deficit with phenotyping platforms. (ii) We develop ecophysiological models with parameters which can be directly extracted from measurements in platforms and in the field. (iii) We carry out GWAS at different scales, from -omic to plant scale in platform, and to yield components in network of field experiments to identify QTLs linked to conditional allelic effects depending on environmental conditions, and to values of model parameters. (iv) We use either direct measurements or the allelic compositions at target QTLs to determine the phenotypic profiles (set of parameter values) of real or virtual genotypes. (v) We simulate genotypic performance and the contribution of genomic regions under current and future stress situations over Europe via modelling. Results are compared to the observed genetic variability in networks of field experiments and are used as feedbacks for improving our phenotyping routines and ecophysiological models.

The plant science community has to design new genotypes that are able to cope with the diversity of environmental conditions linked to climate change. A major issue is to combine genomic selection with knowledge originating from phenotyping. We have adopted for that a multi-scale multi-environment approach. In the field, it consisted of clustering time courses of environmental variables over Europe into six scenarios of temperature and water deficit experienced by plants; of performing field experiments in 29 contrasting environmental conditions across Europe with a panel of 250 maize hybrids; assigning individual experiments to previously defined scenarios, and finally analysing the genetic variation of plant performance for each environmental scenario via genome wide association studies (GWAS). Large variations of QTL effects depending on environmental scenarios resulted in a pattern associated with each QTL. In a phenotyping platform (Phenoarch), we have estimated intercepted light and radiation use efficiency of each hybrid of the same panel via a functional-structural model using 3D reconstructions of each plant, and the sensitivity of growth to water deficit of each hybrid via a joint analysis of several experiments with contrasting light, evaporative demand and soil water potential. As a whole, the combination of field and platform approaches results in a dataset that allows one to identify genomic regions associated with tolerance in specific scenarios of heat and drought, and with traits associated to these genomic regions. Finally, models allow identifying geographical regions in which a given combination of alleles is likely to have comparative advantages.

International audience; Leaf expansion depends on both carbon and water availabilities. In cereals, most of experimental effort has focused on leaf elongation, with essentially hydraulic effects. We have tested if evaporative demand and light could have distinct effects on leaf elongation and widening, and if short term effects could translate into final leaf dimensions. For that, we have monitored leaf widening and elongation in a field experiment with temporary shading, and in a platform experiment with 15-min temporal resolution and contrasting evaporative demands. Leaf widening showed a strong (positive) sensitivity to whole-plant intercepted light and no response to evaporative demand. Leaf elongation was (negatively) sensitive to evaporative demand, without effect of intercepted light per se. We have successfully tested resulting equations to predict leaf length and width in an external dataset of 15 field and 6 platform experiments. These effects also applied to a panel of 251 maize hybrids. Leaf length and width presented quantitative trait loci (QTLs) whose allelic effects largely differed between both dimensions but were consistent in the field and the platform, with high QTLxEnvironment interaction. It is therefore worthwhile to identify the genetic and environmental controls of leaf width and leaf length for prediction of plant leaf area.

Growth under water deficit is controlled by short-term mechanisms but, because of numerous feedbacks, the combination of these mechanisms over time often results in outputs that cannot be deduced from the simple inspection of individual mechanisms. It can be analysed with dynamic models in which causal relationships between variables are considered at each time-step, allowing calculation of outputs that are routed back to inputs for the next time-step and that can change the system itself. We first review physiological mechanisms involved in seven feedbacks of transpiration on plant growth, involving changes in tissue hydraulic conductance, stomatal conductance, plant architecture and underlying factors such as hormones or aquaporins. The combination of these mechanisms over time can result in non-straightforward conclusions as shown by examples of simulation outputs: ‘over production of abscisic acid (ABA) can cause a lower concentration of ABA in the xylem sap ’, ‘decreasing root hydraulic conductance when evaporative demand is maximum can improve plant performance’ and ‘rapid root growth can decrease yield’. Systems of equations simulating feedbacks over numerous time-steps result in logical and reproducible emergent properties that can be viewed as ‘meta-mechanisms’ at plant level, which have similar roles as mechanisms at cell level.

Yield is subject to strong Genotype by Environment interactions (G x E) in the field, especially under abiotic constraints such as soil water deficit (Drought, D) and high temperature (Heat, H). Since environmental conditions show strong fluctuations during the whole crop cycle, geneticists usually do not consider the environmental measures as quantitative variables, but rather as factors in multi-environment analyses. Based on 11 experiments in a field platform with contrasted temperature and soil water deficit, we determined the periods of sensitivity to drought and heat constraints in wheat (Triticum aestivum L.) and determined the average sensitivities for major yield components. G x E interactions were separated into their underlying components, constitutive (G), G x D, G x H and G x H x D, and were analysed for two genotypes, highlighting contrasted responses to heat and drought constraints. We then tested the constitutive and responsive behaviours of two strong Quantitative Trait Loci (QTL) previously associated with yield components. This analysis confirmed the constitutive effect of chromosome 1B QTL, and explained the G x E interaction of chromosome 3B QTL by a benefit of one allele when temperature rises. In addition to the method itself which can be applied to other datasets and populations, this study will support the cloning of a major yield QTL on chromosome 3B which is highly dependent on environmental conditions, and for which the climatic interaction is now quantified.

In Australian dryland agriculture, grain crop yields are approximately 50% of their potential and are highly unpredictable. During the 1990’s, the rate of productivity increase in Australian broad acre cropping improved by 3.4% annually but has since slowed and declined by -1.4% in drought years. A way to improve the drought tolerance of crops varieties is to discover new genes and alleles that allow plants to continue to grow and yield grain under water limited conditions. Although many quantitative trait loci (QTL) have been identified in wheat, few have been deployed in breeding programmes. We cumulated QTL over 10 years on three wheat populations for yield, agronomical, physiological and morphological traits in various locations in Australia, India and Mexico. Genomic resources now enable us make progress in fine mapping and positional cloning of QTL in wheat. Target QTL that increases yield and yield components in hot and dry climates were fine mapped to genes level using the cv Chinese Spring reference sequence and whole genome shotgun sequences of Australian parental lines. We also investigated QTL function under controlled conditions to measure growth rate, transpiration, stomatal traits, and semi-controlled evironments using deep soil bins, rainout shelter and irrigation. By measuring accurately the environmental variables and using ecophysiological models, we can dissect the response to the environment into elementary and simpler traits and identify the conditions where a QTL is specifically expressed. Such detailed information on QTL x environment interaction, physiological mechanism and fine mapping is crucial for breeding application.

International audience

National audience; Genetic progress will largely depend on the adaptation of new genotypes to the diversity of environments, market requirements and types of agriculture. Most experiments for cultivar evaluation use networks of field experiments to assess the suitability of hundreds of genotypes to environmental conditions. In most cases, most of them run for a few years, thereby taking into account a small fraction of the climatic variability. The use of crop models for predicting plant performance under different climatic conditions (annual variation and/or different locations) is a feasible alternative , assuming that models have an efficient capacity for simulating genotype by environment interactions (i.e. : response of plant to specific environmental conditions using genotypic parameters). If the designed model is convenient, it can be used to analyse the sensitivity of a given trait to changes in specific environment variables, assisting breeding decisions. Our first aim is to adapt the current formalisms of the APSIM model to integrate genotypic variability on key development processes affected by environmental stress, with specific parameters for each genotype that can be measured in a phenotyping platform. For that purpose, a new leaf growth module has been implemented in the model, incorporating genotypic variability of phyllochron, number of leaves and sensitivity of leaf growth to water deficit and vapour pressure deficit. All these parameters are measured in the platform PhenoArch, thereby generating a specific vector of parameters for each genotype. The objective will be to simulate the ranking of different existing genotypes in a grid of locations around Europe and analyse specific traits advantages to define the best suited ideotype for each environmental conditions, under current and future climatic conditions (incorporating climate change scenarios).

Maintenance of expansive growth under water deficit has been selected as a key target trait of DROPS because of its early response in drying conditions, its large genetic variability, its partially common control with reproductive growth and its consequences on light interception and transpiration. Development of methods to measure shoot growth in Phenotyping platforms (PhenoArch and Phenodyn, M3P, Montpellier, France; The Plant Accelerator, Adelaide, Australia) allowed identification of a large genetic variability in maize, wheat and durum wheat. Three-D characterization of individual plant architecture and of the spatial distribution of light in phenotyping platforms allowed calculation of light interception and of Radiation Use Efficiency (RUE) of hundreds of genotypes. Measurements of leaf elongation rate with a temporal definition of a few minutes allowed comparison of time courses of growth and of physiological variables that are candidate for the control of expansive growth. Overall, these methods resulted in progress in physiological understanding and in dozens of QTLs identified for shoot growth in bread wheat and maize, peduncle in durum wheat, RUE, transpiration and phenology in all three species. The importance of hydraulics in the control of leaf expansion rate (LER) has been demonstrated. Changes in LER occurred over a few minutes upon changes in evaporative demand and soil water status. LER also varied with circadian-driven changes in plant hydraulic conductance in continuous light. Rapid changes of LER with water deficit and evaporative demand are due to stomatal movements and changes in hydraulic conductivity related to aquaporin Plasma Membrane Intrinsic Proteins (PIPs). Both PIP expression and plant hydraulic conductance rapidly changed with environmental conditions and with the circadian clock. In addition to its value, this result was essential to establish a protocol for comparing transcript amounts of PIPs in genotypes of the maize panel (eQTLs). A large genetic variability has been observed for the progression of shoot development (phyllochron), for whole-plant leaf growth and for their responses to water deficit. Accurate QTLs have been identified for each of these variables, which are compared with the positions of QTLs of PIP transcript amounts, and of accumulation of abscisic acid and of metabolites in the leaf. This will results in new insights on the control of growth under water deficit. The position of QTLs of growth in response to drought has been compared to the positions of QTLs of yield in the field in the three species. In addition, the effects of introgressed genomic regions have been analyzed on the growth of several organs (leaf, peduncle, roots) and on yield of durum wheat and maize. The analysis of impact on yield of growth QTLs is based on colocation of QTLs between platforms and fields, and on a model for scaling up the effects of growth in the field. We have designed for that a model of shoot development that takes into account the responses of individual leaves to water deficit and evaporative demand, with parameters that are measured in a phenotyping platform for all genotypes of the panels. The model is incorporated into the crop model APSIM-maize, for estimation of the effect of QTLs of leaf growth on crop yield in hundreds of climatic scenarios, current or future.

Leaf elongation rate (LER) is the first trait affected by water deficit or high evaporative demand, with typical time constants of 30 min for change in LER upon rapid changes in soil water content or air vapour pressure deficit (VPD). The same applies to other organs such as maize silks. Phenodyn (https://www6.montpellier.inra.fr/lepse/M3P/plateforme-PHENODYN) measures organ elongation rate and transpiration rate of hundreds of plants with a temporal resolution of 3 min (or more if required) in order to follow the changes in LER and transpiration in fluctuating conditions of soil water content, evaporative demand and temperature. Phenodyn imposes known soil water potentials to each plant via independent automatic irrigation. Climatic conditions are either imposed in the growth chamber or left to naturally fluctuate in the greenhouse. Elongation rate is measured with 500 rotational displacement transducers with a precision of 0.2 mm. Transpiration and soil water content are measured with scales; changes in weight are attributed to changes in soil water content after correction for the increase in plant biomass. A set of sensors measures meristem temperature, incident light, air temperature and VPD every minute. Phenodyn is associated to an information system for real time monitoring of experiments, post-analysis of large datasets (around 700.000 data points are generated in each experiment) and identification of genotypic parameters such as rates or time constants. It has been used (i) for analyzing the response of LER to soil water potential and to VPD in mapping populations, diversity panel for association genetics or insertion lines, (ii) for establishing response curves to temperature in different species and genotypes, (iii) for following jointly changes in transpiration and in elongation rates of leaves or silks together with hydraulic variables. It has been used in maize, rice, wheat, sorghum, millet, apple tree and vine. Phenodyn is part of the M3P facility (https://www6.montpellier.inra.fr/lepse/M3P). It is accessible to public or private scientists via the website of the national project Phenome-FPPN (https://www.phenome-fppn.fr/).