Modelos de dimensão finita e infinita para avaliação da produção de ovos em aves de postura
Carregando...
Data
2009-02-20
Autores
Venturini, Guilherme Costa [UNESP]
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Estadual Paulista (Unesp)
Resumo
Objetivou-se estimar parâmetros genéticos e fenotípicos da produção de ovos total e parcial em aves de postura por meio de modelos de dimensão finita e infinita, visando à seleção das aves para a característica produção de ovos. Utilizou-se dados de fêmeas, nascidas a partir de três incubações, provenientes de uma geração de uma linhagem de poedeiras, desenvolvida pela EMBRAPA Suínos e Aves em Concórdia, SC. A produção de ovos foi avaliada individualmente por meio do número de ovos, a partir de 20 até 70 semanas de idade. As produções foram divididas em períodos, sendo que cada período consistiu em três semanas de produção. As características estudadas foram os registros parciais nos períodos e o registro total da produção de ovos (PTOTAL). Para análise em modelo animal uni e bicaracterística, utilizou-se o programa computacional MTDFREML. Maior herdabilidade foi estimada para o período de 20 a 22 semanas de idade (P1) e PTOTAL (0,33 ± 0,07 e 0,23 ± 0,06). Correlações genéticas entre os períodos parciais e PTOTAL variaram de 0,42 ± 0,15 (PTOTAL com P2, de 20 a 22 semanas de idade) até 1,00 (PTOTAL com P9 e P11, de 44 a 46 e 50 a 52 semanas de idade, respectivamente). A correlação de Spearman entre as ordens dos 43 galos quanto aos valores genéticos para produção de P9, P10 (de 47 a 49 semanas de idade) e de P11 com PTOTAL foi significativa (p<0,001) para P9 e P11, igual a 1,00 e para P10 (p<0,0028), igual a 0,99. Concluiu-se que o registro parcial de ovos para os períodos P9, P10 e P11 poderiam ser considerados no processo de seleção visando à produção total de ovos. Porém, devido os mesmos apresentarem baixas estimativas de herdabilidade (0,06; 0,12 e 0,10, respectivamente para P9, P10 e P11) a seleção seria mais eficiente considerando o PTOTAL. Portanto, seria recomendada a seleção baseada na produção total de ovos, pois a mesma...
The aim was to estimate genetic and phenotypic parameters for total and partial records of egg production of laying hens using finite and infinite-dimension models, with a view to selecting birds for egg production. Data on females born from three incubations, from a generation of an egg-laying strain developed by EMBRAPA Suínos e Aves, Concórdia, SC, were used. Egg production was evaluated individually as the number of eggs produced between the ages of 20 and 70 weeks. The production was divided into three-week periods. The traits studied were the partial records for each period (P1 to P17) and the total recorded egg production (PTOTAL). For analysis in one and two-trait animal models, the MTDFREML software was used. Higher heritabilities were estimated for the periods of 20 to 22 weeks of age (P1) (0.33 ± 0.07) and PTOTAL (0.23 ± 0.06). The genetic correlations between the partial periods and PTOTAL ranged from 0.42 ± 0.15 (P1) to 1.00 (for P9 and P11, representing 44-46 and 50-52 weeks of age, respectively). Spearman’s correlation between the ranks of the 43 sires, regarding the genetic values for production in P9, P10 (47-49 weeks of age) and P11 in relation to PTOTAL, was significant and equal to 1.00 (P<0.001) for P9 and P11 and equal to 0.99 (P<0.0028) for P10. Although highly correlated with PTOTAL, these periods had low heritability estimates in two-trait analysis (0.06, 0.12 and 0.10, respectively for P9, P10 and P11). Thus, selection based on these partial periods would be inefficient and basing it on total egg production would be recommended, since this presented higher heritability estimates. To estimate the covariance functions using random regression models, the DXMRR option in the DFREML statistical software was used. The most appropriate model regarding goodness-of-fit of the egg production data in these laying hens was the one composed of third-order... (Complete abstract click electronic access below)
The aim was to estimate genetic and phenotypic parameters for total and partial records of egg production of laying hens using finite and infinite-dimension models, with a view to selecting birds for egg production. Data on females born from three incubations, from a generation of an egg-laying strain developed by EMBRAPA Suínos e Aves, Concórdia, SC, were used. Egg production was evaluated individually as the number of eggs produced between the ages of 20 and 70 weeks. The production was divided into three-week periods. The traits studied were the partial records for each period (P1 to P17) and the total recorded egg production (PTOTAL). For analysis in one and two-trait animal models, the MTDFREML software was used. Higher heritabilities were estimated for the periods of 20 to 22 weeks of age (P1) (0.33 ± 0.07) and PTOTAL (0.23 ± 0.06). The genetic correlations between the partial periods and PTOTAL ranged from 0.42 ± 0.15 (P1) to 1.00 (for P9 and P11, representing 44-46 and 50-52 weeks of age, respectively). Spearman’s correlation between the ranks of the 43 sires, regarding the genetic values for production in P9, P10 (47-49 weeks of age) and P11 in relation to PTOTAL, was significant and equal to 1.00 (P<0.001) for P9 and P11 and equal to 0.99 (P<0.0028) for P10. Although highly correlated with PTOTAL, these periods had low heritability estimates in two-trait analysis (0.06, 0.12 and 0.10, respectively for P9, P10 and P11). Thus, selection based on these partial periods would be inefficient and basing it on total egg production would be recommended, since this presented higher heritability estimates. To estimate the covariance functions using random regression models, the DXMRR option in the DFREML statistical software was used. The most appropriate model regarding goodness-of-fit of the egg production data in these laying hens was the one composed of third-order... (Complete abstract click electronic access below)
Descrição
Palavras-chave
Ave poedeira, Ovos - Produção, Regressão aleatória, Curva de produção, Avaliação genética, random regression, Production curve, Layers, Genetic evaluation
Como citar
VENTURINI, Guilherme Costa. Modelos de dimensão finita e infinita para avaliação da produção de ovos em aves de postura. 2009. xi, 45 f. Dissertação (mestrado) - Universidade Estadual Paulista, Faculdade de Ciências Agrárias e Veterinárias, 2009.