Decoding Historical Breeding Strategies from Expected Progeny Distributions of Past Crosses: Retrospective Genomic Analysis of Japanese Citrus Breeding

Cross breeding is a fundamental plant breeding practice, but the selection of parental combinations as well as superior individuals among progeny has largely relied on breeders’ tacit and undocumented experience, making it difficult to quantitatively characterize historical breeding processes. Therefore, retrospective analyses of accumulated breeding records may be useful for elucidating how past breeding decisions shaped modern cultivars. Recent advances in genomic predictions enable the reconstruction of expected progeny distributions using genomic information without generating actual progeny, making it possible to retrospectively analyze breeding outcomes, even when sibling records are unavailable. In this study, we used historical records from Japanese citrus breeding programs to develop a framework for retrospective genomic analyses based on expected progeny distributions and estimated individual genomic selection intensity for released cultivars. We incorporated additive and dominance effects to evaluate their influence on progeny distributions and retrospective inferences of selection patterns. Our results demonstrated that individual genomic selection intensity captured directional selection patterns associated with high-sugar and low-acidity cultivars, and that accounting for dominance effects decreased potential biases in retrospective analyses. Detailed analyses further revealed how ‘Kiyomi’, a key maternal parent in Japanese citrus breeding programs, contributed to long-term cultivar development despite its relatively unfavorable trait values. Although this study focused on an outcrossing fruit tree species, the proposed framework is potentially applicable to diverse breeding systems because expected progeny distributions can be predicted for breeding schemes involving repeated selfing and fixation. These results demonstrate the potential utility of retrospective genomic analyses for quantitatively reconstructing historical breeding processes and interpreting breeding strategies using accumulated breeding records.