Desenvolvimento de marcadores ambientômicos para arroz de terras altas (Oryza sativa L.) em território brasileiro

• Ano de defesa: 2023
• Autor(a) principal: Bahia, Marco Antônio Marcelino
• Orientador(a): Resende, Rafael Tassinari
• Banca de defesa: Resende, Rafael Tassinari, Melo, Patrícia Guimarães Santos, Zaidan, Úrsula Ramos
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Goiás
• Programa de Pós-Graduação: Programa de Pós-graduação em Genética e Melhoramento de Plantas (EA)
• Departamento: Escola de Agronomia - EA (RMG)
• País: Brasil
• Palavras-chave em Português: Ambientômica; Melhoramento genético; Bioinformática; Rizicultura; Monte Carlo
• Palavras-chave em Inglês: Enviromics; Plant breeding; Bioinformatics; Rice cultivation; Monte Carlo
• Área do conhecimento CNPq: CIENCIAS EXATAS E DA TERRA
• Link de acesso: http://repositorio.bc.ufg.br/tede/handle/tede/12956

Rice (Oryza sativa L.) is one of the staple foods in the Brazilian diet, and therefore, its cultivation and productive independence are strategically essential for ensuring the population's food security. Within rice farming, selecting the appropriate genotype for planting is the factor that most strongly impacts the outcome of the endeavor. In order to support this decision-making process, enviromics has been applied with the objective of selecting genotypes with higher productive potential for specific areas of interest. The aim of this study was to generate and analyze the contribution of enviromic markers to the total upland rice production data in Brazilian territory. The experimental data were provided by Embrapa Rice and Beans and involved the evaluation of 2,119 rice genotypes in 187 municipalities or localities across the country, spanning the period from 1982 to 2018. For the generation of enviromic markers, data from the SoilGrids, WorldClim, and NASA POWER platforms were used, resulting in a total of 393 environmental covariates collected. The generation of enviromic markers was performed using the Monte Carlo method, with 10,000 iterations and always considering the presence of the 187 municipalities where the Embrapa experiments were conducted. The Random Forest package and the IncMSE and IncNodePurity methods were used to evaluate the importance of each covariate for the model applied throughout the Brazilian territory. The results showed that the coefficient of variation for seasonal precipitation was the most important covariate for both models.