Genomic prediction models based on a large-scale recombinant population allow quick breeding of high-yield rice

Abstract Genomics-based breeding is a promising approach to generate crops with high yield and quality in a short timeframe. However, this approach is currently not suitable for simultaneously controlling multiple traits, necessitating development of a more efficient methodology. Here, we present a quick breeding strategy employing interpretable genomic prediction models generated using a large-scale recombinant population, which enables optimizing multiple traits in cultivars. To validate our strategy, we developed a nested association mapping population in rice ( Oryza sativa ) and generated its associated genomic models, which allowed rapid improvement of high-yielding rice cultivars. Our genomic breeding strategy provides a general framework to quickly breed high-yielding cultivars capable of coping with the challenges caused by a rapidly changing environment.