Abordagem multivariada na seleção de progênies de soja superiores e portadoras do gene RR
Carregando...
Arquivos
Data
2013-07-05
Autores
Dallastra, Anderson [UNESP]
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Estadual Paulista (Unesp)
Resumo
O melhoramento genético de plantas é considerado um processo complexo que gera múltiplas informações e, em muitos casos, de difícil compreensão. O objetivo deste trabalho foi selecionar progênies com caracteres superiores provenientes de cruzamentos bi-parentais de soja com fonte de resistência ao glifosato (RR), além de identificar cruzamentos e genitores mais eficientes, por meio de abordagens multivariadas. Além disso, objetivou-se ainda testar a eficiência dos métodos no processo seletivo para múltiplos caracteres de interesse. O experimento foi conduzido no delineamento experimental do tipo famílias com testemunhas intercalares, no ano agrícola 2010/2011 e 2011/2012 em Jaboticabal-SP sendo que, nas populações F3 foram selecionadas seis plantas fenotipicamente superiores e avaliadas para os caracteres: número de dias para o florescimento (NDF), número de dias para a maturidade (NDM), altura de inserção da primeira vagem (AIV), altura de planta na maturidade (APM), acamamento (Ac), valor agronômico (VA), número de ramos (NR), número de vagens por planta (NV), peso de cem sementes (PCS), número de sementes por planta (NS) e produção de grãos (PG). Os dados foram analisados utilizando-se o software Statistica 7.0. Os resultados obtidos possibilitaram a seleção de 77 progênies superiores através da Análise de Componentes Principais. A análise de Agrupamentos, pelo do método de K-means, agregou as progênies em seis grupos de acordo com os caracteres de maior importância em cada um e, através do método de Ward, identificou por meio do dendrograma a estrutura de similaridade e divergência entre as progênies selecionadas. Por fim, comparou-se os métodos de agrupamentos e verificou-se que houve concordância entre ambos quanto aos resultados obtidos
The plant breeding is considered to be a complex process that generates multiple information sources and in many cases, difficult to understand. The aim of this work was to select progenies with superior characters from bi-parental crosses with soy source of resistance to glyphosate (RR), and identify intersections and parents more efficient through multivariate approaches. Furthermore, aimed to further test the efficiency of the methods in the selection process for multiple traits of interest. The experiment was conducted in the experimental design of type families to witness progress in the agricultural year 2010/2011 and 2011/2012 in Jaboticabal being that in F3 populations selected six plants were phenotypically superior and evaluated for the traits: number of days to flowering (NDF), number of days to maturity (NDM), height of the first pod (AIV), plant height at maturity (APM), lodging (Ac), agronomic value (VA), number of branches (NR), number of pods per plant (NV), one hundred seed weight (PCS), number of seeds per plant (NS) and grain yield (GY). Data were analyzed using the software Statistica 7.0. The results allowed the selection of 77 superior progenies by Principal Component Analysis. Cluster analysis by the K-means method, all progenies added into six groups according to the characters in each of utmost importance, and by the method of Ward identified by the dendrogram structure similarity and divergence between the progenies. Finally, we compared the clustering methods and found that there was an agreement between them as to the results obtained
The plant breeding is considered to be a complex process that generates multiple information sources and in many cases, difficult to understand. The aim of this work was to select progenies with superior characters from bi-parental crosses with soy source of resistance to glyphosate (RR), and identify intersections and parents more efficient through multivariate approaches. Furthermore, aimed to further test the efficiency of the methods in the selection process for multiple traits of interest. The experiment was conducted in the experimental design of type families to witness progress in the agricultural year 2010/2011 and 2011/2012 in Jaboticabal being that in F3 populations selected six plants were phenotypically superior and evaluated for the traits: number of days to flowering (NDF), number of days to maturity (NDM), height of the first pod (AIV), plant height at maturity (APM), lodging (Ac), agronomic value (VA), number of branches (NR), number of pods per plant (NV), one hundred seed weight (PCS), number of seeds per plant (NS) and grain yield (GY). Data were analyzed using the software Statistica 7.0. The results allowed the selection of 77 superior progenies by Principal Component Analysis. Cluster analysis by the K-means method, all progenies added into six groups according to the characters in each of utmost importance, and by the method of Ward identified by the dendrogram structure similarity and divergence between the progenies. Finally, we compared the clustering methods and found that there was an agreement between them as to the results obtained
Descrição
Palavras-chave
Soja, Plantas - Resistencia, Resistência a herbicidas, Herbicidas, Genes, Análise multivariada, Herbicide resistance
Como citar
DALLASTRA, Anderson. Abordagem multivariada na seleção de progênies de soja superiores e portadoras do gene RR. 2013. iii, 63 p. Dissertação (mestrado) - Universidade Estadual Paulista, Faculdade de Ciências Agrárias e Veterinárias de Jaboticabal, 2013.