Análise multivariada para caracterização e divergência de genótipos e correlação entre caracteres em milho
Carregando...
Arquivos
Data
2018-03-22
Autores
Dutra, Sophia Mangussi Franchi [UNESP]
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Estadual Paulista (Unesp)
Resumo
Os objetivos principais com a pesquisa foram: alocar genótipos de milho em grupos divergentes pela abordagem multivariada e correlacionar caracteres agronômicos com caracteres relacionados à qualidade de sementes de milho. Para alocação em grupos heteróticos, foram utilizados dois métodos multivariados, método K-means e método Tocher. A partir da avaliação de caracteres agronômicos (altura de planta, posição relativa da espiga, acamamento, quebramento, prolificidade e produtividade de grãos), foram alocadas 229 linhagens parcialmente endogâmicas (S3), avaliadas em três safras. Já para correlação entre caracteres foram utilizados os métodos Redes Bayesiana e correlação linear simples de Pearson. Os caracteres avaliados, de 31 genótipos de milho, foram caracteres agronômicos (altura de planta, posição relativa da espiga, acamamento, quebramento, prolificidade e produtividade de grãos), avaliados em três safras, e caracteres de sementes (primeira contagem de germinação, germinação final, primeira contagem de emergência, emergência final e índice de velocidade de germinação). As análises estatísticas foram realizadas utilizando-se o programa computacional SAS, Selegen e RStudio. Na análise de variância conjunta foi constatado efeito significativo para safra em todas caracteres analisadas, caracterizando a influência das safras e indicando diferença entre os ambientes de avaliação. Foi possível alocar os genótipos de milho em grupos heteróticos a partir dos métodos de agrupamento. A correlação linear de Pearson variou de -0,511 (entre quebramento e produtividade) a 0,987 (entre primeira contagem de emergência e emergência final). A partir do Directed Acyclic Graph é possível observar a correlação entre caracteres agronômicos e caracteres relacionados à qualidade de sementes de milho. Assim, os genótipos de milho foram alocados quanto à divergência a partir das análises multivariadas de agrupamento K-médias e Tocher. Foi possível identificar sete grupos divergêntes pelo método Kmédias e 21 pelo método Tocher. E também, observa-se pouca correlação existente entre os caracteres agronômicos e os caracteres relacionados à qualidade de sementes de milho.
The main objectives of the research were: to allocate corn genotypes in divergent groups by the multivariate approach and to correlate agronomic traits with traits related to maize seed quality. For allocation in heterotic groups, two multivariate methods, K-means method and Tocher method were used. Based on the evaluation of agronomic characters (plant height, relative position of the spike, lodging, breaking, prolificacy and grain yield), 229 partially inbred lines (S3) were evaluated, evaluated in three harvests. Already for the correlation between characters were used the methods Bayesian Networks and simple linear correlation of Pearson. The evaluated traits of 31 corn genotypes were agronomic characteristics (plant height, relative position of the spike, lodging, breaking, prolificacy and grain yield) evaluated in three harvests, and seed characteristics (first germination count, germination endpoint, first emergency count, final emergency and germination speed index). Statistical analyzes were performed using the SAS software, Selegen and RStudio. In the analysis of joint variance it was verified a significant effect for harvest in all characters analyzed, characterizing the influence of the crops and indicating difference between the evaluation environments. It was possible to allocate the maize genotypes to heterotic groups from the clustering methods. Pearson's linear correlation ranged from -0.511 (between breakage and productivity) to 0.987 (between first emergency count and final emergency). From the Directed Acyclic Graph it is possible to observe the correlation between agronomic characters and characters related to the quality of maize seeds. Thus, maize genotypes were allocated for divergence from the multivariate K-medium and Tocher cluster analyzes. It was possible to identify seven divergent groups by the K-means method and 21 by the Tocher method. Also, there is little correlation between agronomic characters and traits related to maize seed quality.
The main objectives of the research were: to allocate corn genotypes in divergent groups by the multivariate approach and to correlate agronomic traits with traits related to maize seed quality. For allocation in heterotic groups, two multivariate methods, K-means method and Tocher method were used. Based on the evaluation of agronomic characters (plant height, relative position of the spike, lodging, breaking, prolificacy and grain yield), 229 partially inbred lines (S3) were evaluated, evaluated in three harvests. Already for the correlation between characters were used the methods Bayesian Networks and simple linear correlation of Pearson. The evaluated traits of 31 corn genotypes were agronomic characteristics (plant height, relative position of the spike, lodging, breaking, prolificacy and grain yield) evaluated in three harvests, and seed characteristics (first germination count, germination endpoint, first emergency count, final emergency and germination speed index). Statistical analyzes were performed using the SAS software, Selegen and RStudio. In the analysis of joint variance it was verified a significant effect for harvest in all characters analyzed, characterizing the influence of the crops and indicating difference between the evaluation environments. It was possible to allocate the maize genotypes to heterotic groups from the clustering methods. Pearson's linear correlation ranged from -0.511 (between breakage and productivity) to 0.987 (between first emergency count and final emergency). From the Directed Acyclic Graph it is possible to observe the correlation between agronomic characters and characters related to the quality of maize seeds. Thus, maize genotypes were allocated for divergence from the multivariate K-medium and Tocher cluster analyzes. It was possible to identify seven divergent groups by the K-means method and 21 by the Tocher method. Also, there is little correlation between agronomic characters and traits related to maize seed quality.
Descrição
Palavras-chave
agrupamento, correlação, melhoramento de plantas, correlation, clustering, plant breeding