Développer de nouvelles variétés de blé pour une agriculture durable

BreedWheat (BW) was conceived to strengthen the competitiveness of the French breeding sector as well as to address the societal demand for sustainability, quality, and safety. BW aimed at (i) sequencing the bread wheat genome in collaboration within the International Wheat Genome Sequencing Consortium (IWGSC) and developing molecular markers, (ii) expanding and facilitating the use of original genetic resources to increase allelic variability in the elite wheat gene pool, (iii) deciphering the genetic and ecophysiological bases of abiotic and biotic stress tolerance for yield and quality, (iv) evaluating and integrating novel and innovative breeding methods and (v) providing new bioinformatics tools and databases to efficiently store and visualize the data and disseminate the results of the project to researchers, breeders, and growers.

The objectives were reached and, in some cases, largely exceeded thanks to methodological progresses and the ability of work package leaders and partners to take advantage of these evolutions. BW contributed significantly to the sequencing and the analysis of the first reference sequence of the hexaploid wheat genome. This achievement is a major breakthrough for the wheat community. The sequence as well as other resources developed by the IWGSC have been made available through the information system of INRAE URGI. Based on these resources, two genotyping arrays have been developed: (i) the TaBW420K array comprising ~423,385 SNPs (Single Nucleotide Polymorphism) and (ii) the TaBW35K array comprising 34,746 SNPs. These arrays have been used in the frame of BW, as well as in other projects since they have been made available to the wheat community. More than 13,000 cultivated and wild wheat accessions have been genotyped. These data were used in other work packages, as well as in collaboration with other projects to perform genome-wide association studies, to implement genomic and phenomic selection, to characterize the worldwide diversity and its history, and to build a catalogue of structural variations of the genome sequence.

The aim of the BW was also to broaden the genetic basis of French elite varieties through the introduction of novel sources of genetic diversity that contain favourable alleles for resistance to abiotic and biotic stresses. For that, it was decided (i) to select a panel of 450 accessions representing the world genetic diversity for winter wheat and (ii) to create pre-breeding populations crossing elite European varieties and genitors bringing new tolerances to biotic (Fusarium Head Blight (FHB), Septoria tritici Blotch (STB)) and abiotic (high temperature, drought) stresses. To select the panel, 4,600 accessions of the INRAE bread wheat collection were genotyped and characterized for simple traits (plant height, precocity) in the field. The 450 accessions were then chosen to maximize the possibility to detect chromosomal regions in genome-wide association studies. To create the pre-breeding populations, genitors were either selected in the worldwide diversity or in the panel of 450 accessions. With these selections, BW partners created 36 populations for a total of almost 5,000 pre-breeding lines that contain novel sources of diversity for tolerance to stresses.

The genetic and ecophysiological tolerance to biotic (FHB, STB) and abiotic (drought, nitrogen and sulphur deficiency, high temperature) stresses were investigated. A set of ecophysiological models was developed to simulate (i) the Carbon-Nitrogen metabolism of a wheat plant within a 3D description of the architecture to investigate Genotype × Environment interactions; (ii) the development of Septoria epidemics in a wheat canopy and to investigate specifically the impact of plant architecture and nitrogen and (iii) the signals from wheat canopies perceived by phenotyping sensors and to develop/improve methods for high throughput phenotyping. Candidate genes and pathways were identified for the response of grain weight to high temperature and the response of grain protein composition to nitrogen and sulphur supplies. This was done using different approaches (network inference, multivariate analyses) to analyse transcriptomics and proteomics data. Finally, genetic approaches were undertaken to identify genomic regions of interest (Quantitative Trait Loci, QTL) combining genotypic data and phenotypic data coming from large field trial networks. Two different panels, the first one composed of 220 European elite varieties and the second one corresponding the 450 accessions worldwide diversity panel) were experimented and phenotyped in large trial networks. We used for that information coming from the ecophysiological studies and new phenotyping tools developed by the PHENOME project. Marker x traits associations were identified for tolerance to FHB, STB, drought and N deficiency. Seedling stage and field evaluations of accessions combining five QTL validated the efficiency of these chromosomal regions against STB. At the end, BW was able to transfer to breeders lists of tolerant varieties for N and drought stresses and markers to follow genetic areas of interest for biotic and abiotic stress tolerance.

A R-package named BWGS, which gathers and makes easy-to use several methods for data cleaning, sampling and imputation, as well as 15 methods for genomic prediction of breeding value was developed. This package was disseminated through national and International training sessions. An experimental comparison of lines selected by either genomic selection (GS) or phenotypic selection (PS) after one or two cycles of GS was applied to a real breeding programme. After having designed the characteristics of lines that better fit given pedo-climatic conditions and farming systems (these characteristics define ideotypes), 25 varieties showing contrasted resistances/tolerance and characteristics close to the desired ones were identified and tested in trials carried out for three years in eight sites. These trials were dedicated to studying the tolerance of the varieties to water and nitrogen stresses and their adaptation to low input cropping practices. Results from these trials allowed to assess the ideotypes proposed and make recommendations for variety evaluation in innovative cropping systems. Finally, to address both technical and socio-economic aspect of the implementation of novel breeding strategies, a research cost function for different breeding strategies (genomic selection, conventional breeding) was defined. This function relates the level of research investment to the level of innovation. Then a virtual model of plant breeding was developed, including cost parameters. This model aims at comparing the economic efficiency of genomic selection in simulated but realistic wheat breeding programmes. The simulation pipeline is expected to become a decision tool to help breeders adjusting breeding schemes, according to their short or long-term objectives. A second study was conducted to better characterize the link between the innovation embedded in new seed variety and their sales. The sale of one variety is define as the product between the market size of the seed related to wheat and the market share of the variety. Two complementary empirical analysis were applied to the French context.

The objective to develop a centralized repository suited to the data generated in BW has been reached with (i) the development of the BW Information System (BIS) which relies on GnpIS that was updated to consider the user needs, (ii) the integration of the project data in the information system and (iii) the definition of data access rights following the consortium agreement. BIS has been specifically developed to manage and link different data types: genomics data, germplasm description, genotyping data, genetic maps, ontologies, phenotyping data, and results of association studies. BIS has been successfully implemented with the project data and a dedicated webpage has been set-up to easily access them.

The initial objective to “provide advanced tools, knowledge, material, and databases all of which should enhance the competitiveness of the French wheat breeding sector” has then been globally reached.