Soybean maturity groups in Eastern Africa: method and genotype classification

• Ano de defesa: 2024
• Autor(a) principal: Brandão, Leonardo Martins
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Biblioteca Digitais de Teses e Dissertações da USP
• Palavras-chave em Português: Glycine max; Ciclo de desenvolvimento de plantas; Data-driven modeling; Melhoramento de plantas; Modelagem orientada por dados; Plant breeding; Plant growth cycle
• Link de acesso: https://www.teses.usp.br/teses/disponiveis/11/11137/tde-13032025-170228/

Soybean [Glycine max (L.) Merr.] is a light-sensitive species and its cycle is very much affected by day length. Thus, the usage of relative maturity groups (RMGs) is an important tool to provide information about positioning soybean cultivars. In many Eastern African countries, soybean is still being introduced, and there have been no prior efforts to establish RMGs for the region. Therefore, this dissertation aimed to assign RMGs to soybean genotypes cultivated in different environments in Sub-Saharan Eastern African countries; to recommend local soybean genotypes as RMG standards for future trials; to establish a comprehensive methodology for filtering and selecting suitable data, and for calculating RMGs; and finally, to provide criteria for verifying the accuracy of these selections and calculations. The study used the Soybean Innovation Lab\'s Pan African Trials dataset, which includes data on 315 genotypes collected between 2015 and 2022. After selecting the most suitable environment and the cultivars to be used as RMG standards, a linear regression model was designed, using the number of days to maturity to calculate RMGs. In order to expand to different environments, the rest of the dataset was submitted to a series of criteria to elect the environments and cultivars suitable for calculations. The model was then adjusted to each selected genotype, in each specific suitable environment, and their RMGs were calculated. In total, 46 genotypes, including 43 developed by breeding programs from Eastern African countries, had their RMG groups calculated based on their numbers of days to maturity. Afterwards, five genotypes originated from African countries were indicated to be used as RMG standards for future trials in different environments in Africa. This dissertation provided comprehensive methods to assign RMGs to genotypes based on their cycles that could be applied to various African regions. The proposed steps also include strategies to overcome the difficulties of processing challenging soybean growth cycle datasets. Moreover, the methods provide tools to verify the accuracy of the calculations, ensuring more reliable results for soybean breeding programs.