Monteverde, Eliana2019-10-152019-12-052019-05-30Monteverde_cornellgrad_0058F_11296http://dissertations.umi.com/cornellgrad:11296bibid: 11050412https://hdl.handle.net/1813/67430Uruguay is the top rice exporter in Latin America, and among the top eight high-quality rice exporters in the world. As Uruguayan farmers are reaching the yield potential of current varieties, new varieties with higher yield potential and milling quality must be developed. With the recent developments in genomics, rapid gains can be achieved through the integration of conventional breeding methods with genomic selection (GS). However, the best strategy for the efficient implementation of these techniques in specific breeding programs must be carefully analyzed. This work addresses some aspects of the implementation of GS in the Uruguayan breeding program and provides a model for other breeding programs of similar size and complexity. First, the impact on prediction accuracies of modeling genotype by environment interactions (G×E) was tested, and we found that modeling covariance structures that accommodate correlations between environments was beneficial for predicting yield and milling quality in both indica and tropical japonica rice. Different approaches for including weather information to assist genomic predictions were compared, and the impact of certain weather components on yield and milling quality were assessed. Modeling environmental effects by using weather variables provided an advantage in terms of prediction accuracy when predicting untested environments. Results from both genomic prediction and QTL×E analyses provided clues about the main weather variables affecting milling yield in rice grown in subtropical regions. We also tested the use of genomic prediction for selection of parents in a tropical japonica rice breeding program. Starting from a population of 19 families, we evaluated several strategies for parental selection based on cross and progeny simulations to improve grain yield and milling quality traits. We also performed a field evaluation of the progeny from some of these crosses to compare genomic predictions to empirical data. Finally, a genome-wide association study was performed in order to find genomic regions associated with anther culture response in tropical japonica germplasm. The analysis identified 21 significant regions of the rice genome involved in callus induction and plant regeneration. Some of the same regions were reported in previous studies for anther culture response in rice. Future validations of the strategies outlined in this research will provide the foundation for future decision-making about the role that GS may play in the Uruguayan rice breeding program.en-USAttribution-NonCommercial-NoDerivatives 4.0 InternationalAgricultureGenomic predictionricePlant Breedingdissertation or thesishttps://doi.org/10.7298/xawd-aj36