Use este identificador para citar ou linkar para este item: http://repositorio.ufla.br/jspui/handle/1/38358
Título: Statistical modeling and selection efficiency on Urochloa ruziziensis breeding
Título(s) alternativo(s): Modelagem estatística e eficiência da seleção no melhoramento de Urochloa ruziziensis
Autores: Gonçalves, Flávia Maria Avelar
Souza Sobrinho, Fausto de
Gonçalves, Flávia Maria Avelar
Nunes, José Airton Rodrigues
Pessoa Filho, Marco Aurelio Caldas de Pinho
Souza Sobrinho, Fausto de
Souza, João Cândido de
Palavras-chave: Brachiaria ruziziensis
Modelos mistos
Seleção indireta
BLUP unitrait
BLUP multi-trait
Mixed models
Indirect selection
Data do documento: 19-Dez-2019
Editor: Universidade Federal de Lavras
Citação: DIAS, J. A. Statistical modeling and selection efficiency on Urochloa ruziziensis breeding. 2019. 56 p. Tese (Doutorado em Genética e Melhoramento de Plantas)–Universidade Federal de Lavras, Lavras, 2019.
Resumo: The Urochloa ruziziensis (R. Germ. & C.M. Evrard) Crins (sin. Brachiaria ruziziensis) is diploid and presents sexual reproduction, has great value for pasture diversification, particularly for milk production and it has great pasture potential due to its high nutritional quality, good animal acceptance and good leaf/stem ratio. Evaluations and selection of genotypes in breeding programs are usually based on measurements made on several cuttings over time in the same experimental plot, being a time consuming process that requiring methodologies capable of dealing with the complexity of the data generated. Therefore, statistical methods for analysis of this type of data should consider the spatial variation and temporal correlation between repeated measures, as well as the possibility of heterogeneity of variance and appropriately model genetic effects over time. In addition, the identification of traits closely related to yield may enable indirect selection for forage production, making evaluations more agile and accurate, maximizing gains and minimizing time and cost in launching new cultivars. Thus, the present study aimed to evaluate different covariance structures for the residual and genetic matrices, verify their implications for clone selection and verify the efficiency of visual selection for green biomass production in U. ruziziensis using the plant vigor on unitrait and multitrait analysis verifying which approach maximizes the predictive accuracy of genetic values. A total of 254 U. ruziziensis clones were evaluated in nine cuts in an incomplete block de sign with three replicates and plot of one plant together with the cultivars Marandu ( U. brizantha) and Basilisk (U. decumbens) on the experimental field of Embrapa Gado de Leite (Coronel Pacheco - MG). To investigate the statistical modeling in clone selection, the covariance matrices for genetic and residual effects were modeled for green biomass production and the choice of the best model was by BIC. To verify the efficiency of visual selection, unitrait and multitrait analyzes of green biomass production and vigor were performed, considering different selection strategies. It was found that the covariance structures change according to the data under study, and it is necessary to verify which structures provide the best quality of fit to the data, due to the impact on genetic parameter estimates, experimental precision and on the selection and continuity of the breeding programs. In addition, it was possible to verify that the visual selection for forage production through plant vigor can be a useful tool in U. ruziziensis breeding programs, especially when using the multitrait approach.
URI: http://repositorio.ufla.br/jspui/handle/1/38358
Aparece nas coleções:Genética e Melhoramento de Plantas - Doutorado (Teses)

Arquivos associados a este item:
Arquivo Descrição TamanhoFormato 
TESE_Statistical modeling and selection efficiency on Urochloa ruziziensis breeding.pdf1,01 MBAdobe PDFVisualizar/Abrir


Os itens no repositório estão protegidos por copyright, com todos os direitos reservados, salvo quando é indicado o contrário.