BreedPatH III: Zuchtwert-Mustererkennung in Hybridkulturarten

Abstract

According to breeding theory, breeding progress can be accelerated most rapidly by reduction of the time needed to complete breeding cycles. In this context, the methods developed in BreedPatH I and II represented a significant advance for practical breeding on winter oilseed rape, with highly valuable transfer potential for adaptation of other inbreeding crops. In BreedPatH Phase III we transferred methods and knowledge from Phases I and II to faba bean (Vicia faba L.). Faba bean represents an excellent model for translation of results, because best-performing varieties are based on synthetic hybrids derived from inbred lines with high per se performance. However, development of faba bean synthetics is an extremely time-consuming process which at the outset of this project was bred without genomic tools or data-driven breeding methods. Using the methods and concepts developed in BreedPatH Phases I and II, we used genomic data for simulation approaches based on genome-wide haplotypes to design heterotic pools and early selection strategies for high-performing faba bean synthetics. We anticipate that this strategy can reduce the breeding cycle of faba bean from 14 to 10 years. An effective hybrid breeding system for faba bean does not yet exist, however significant heterosis is observed and can be exploited via synthetic cultivars. Exploitation of heterotic pools represents a completely new approach for faba bean breeding that lays a foundation for systematic exploitation of heterosis and improvement of yield and yield stability in an important cool-season legume crop. By adapting methods that were developed in the first two funding periods for oilseed rape hybrid breeding, BreedPatH III aimed to reduce the time required for faba bean breeding cycles to 10 years by developing heterotic pools and genome-based prediction models to select lines that have high breeding values and good combining abilities for synthetic early testing at F5-level. In particular, prediction models were applied for design of crossing schemes leading to a rapid combination of beneficial alleles, specifically to separate diverse faba bean breeding materials into two distinct genetic pools. By incorporating observation trials with novel data from detection of genome-wide haplotype effects, prediction models and crossing strategies for breeding cycles were upgraded across breeding cycles in order to maximise both per se performance of inbred lines and general combining ability in a synthetic breeding context. In addition to saving considerable time in the breeding process, recurrent genomic selection procedures can also lead to a reduction in the number of field tests required and, with the same amount of testing, to greater selection efficiency and thus greater breeding progress during synthetic cultivar development. The scientific and technical objectives of BreedBatH III were as follows:

Separation of a non-diversified base population into distinct heterotic pools. Generation of two new, high-quality genome assemblies for contrasting pool representatives using long-read sequencing as a template for genome-based breeding. Application of cost-effective genotyping for genome-wide SNPs and structural variants. Development and implementation of a crossing strategy design based on simulation and prediction models. Prediction of per se performance and general combining ability (GA) for faba bean based on genome-wide haplotype patterns. Using the chain-crossing scheme and algorithms developed in BreedPatH I and genome-wide SNP genotype representing the diverse breeding pool of partner NPZi, crosses were designed for separation of the breeding pool into two distinct pools based on 60 founder lines each. A second cycle of crosses implemented genomic selection to generate families with high predicted yield performance. High resolution genome data were used to identify genome-wide haplotype effects and structure variants. The results provide novel plant materials and data for accelerated breeding using novel breeding methodologies in an important cool-season legume crop. Genome sequence data corresponding to important founder lines was generated collected and curated by partner FzJ. Based on these datasets along with phenotype data generated by partners NPZi, partner JLU established haplotype maps as a basis for simulation-based crossing schemes. Crosses and new plant populations were generated and tested by partner NPZi based on the results of the simulations performed by partner JLU. The project provides new plant materials, extensive genotype datasets and innovative new data-driven crossing designs as a basis to apply novel breeding methods for systematic exploitation of heterosis in faba bean and other important legume crops.