Uso de informação genômica para estimação de parâmetros genéticos para características de crescimento e carcaça em bovinos Nelore
Carregando...
Arquivos
Data
2019-02-20
Autores
Costa, Rebeka Magalhães da
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Estadual Paulista (Unesp)
Resumo
A seleção tradicional para características quantitativas de importância econômica é realizada geralmente com base nos valores genéticos preditos a partir de registros fenotípicos de um indivíduo e de seus parentes. Com a disponibilidade da informação genômica, a predição do valor genético para características complexas tem sido amplamente aperfeiçoada. A utilização da predição genômica poderá levar a um ganho genético mais rápido do que o alcançado com métodos tradicionais de seleção, com base apenas em dados de pedigree e fenotípicos. O objetivo deste trabalho foi avaliar a influência da inclusão da informação genômica nas estimativas dos parâmetros genéticos e tendências genéticas para ganho de peso diário do nascimento aos 120 dias de idade (GP1), dos 120 aos 210 dias de idade (GP2), dos 210 aos 365 dias de idade (GP3), dos 365 aos 450 dias de idade (GP4), área de olho de lombo (AOL), espessura de gordura subcutânea (EGS) e espessura de gordura subcutânea na garupa (EGP8) em bovinos da raça Nelore. Além disso, avaliou-se a associação genética entre as características estudadas e, por meio das análises de agrupamento não-hierárquicas, verificou-se quais os grupos de animais mais indicados para atender aos objetivos de seleção, visando contribuir para o processo de seleção do programa de melhoramento da raça Nelore. As estimativas de parâmetros genéticos foram obtidas com base em registros de pedigree de 192.483 animais, nascidos entre 1934 e 2016, registros fenotípicos de 80.114 animais e genótipos de 8.652 animais, os quais foram cedidos pela Associação Nacional de Criadores e Pesquisadores (ANCP). Os componentes de (co)variância foram obtidos por meio de metodologia bayesiana via amostragem de Gibbs. Os valores genéticos (EBVs) foram estimados por meio do modelo animal, via equações de modelos mistos (o qual utiliza apenas a matriz A) e os valores genéticos genômicos (GEBVs) por meio da metodologia “single-step genomic BLUP” (ssGBLUP) (a qual utiliza a matriz H). As tendências genéticas foram estimadas por meio de regressão linear considerando-se a variação dos EBVs e dos GEBVs médios anuais em função do ano de nascimento. As análises multivariadas de agrupamento foram utilizadas para agrupar os animais com base nos EBVs e GEBVs para as características estudadas e explorar o padrão genético em cada agrupamento obtido. Essas análises foram divididas em dois cenários, sendo um cenário considerando somente as fêmeas e outro somente os machos. Utilizando somente a matriz A, construída com base apenas em registros de pedigree, as médias das estimativas de herdabilidade direta (ℎ2) foram 0,14±0,02, 0,18±0,01, 0,14±0,02, 0,12±0,02, 0,37±0,03, 0,17±0,02, 0,28±0,02, respectivamente para GP1, GP2, GP3, GP4, AOL, EGS e EGP8. Com a inclusão de informação genômica, utilizando a matriz H, as estimativas de ℎ2 para GP1, GP2, GP3, GP4, AOL, EGS e EGP8 foram 0,17±0,02, 0,19±0,01, 0,15±0,02, 0,11±0,01, 0,32±0,02, 0,18±0,02 e 0,25±0,02, respectivamente. As correlações genéticas médias obtidas pelos dois métodos foram muito semelhantes. As maiores médias de correlações genéticas obtidas (matriz A/matriz H) foram entre AOL e as características GP1 (0,34±0,10/0,40±0,08), GP2 (0,52±0,06/0,48±0,06) e GP3 (0,37±0,09/0,39±0,07) e entre EGS e EGP8 (0,69±0,05/0,66±0,04). Com a inclusão de informações genômicas, as acurácias dos valores genéticos apresentaram pequeno aumento e maiores mudanças nas acurácias foram observadas para os touros com poucos filhos. Para as tendências genéticas foram obtidos coeficientes angulares significativos (p<0,05) para a maioria das características estudadas, exceto aquela obtida com o uso do GEBV para EGP8. Evidenciou-se o progresso genético para as características estudadas, tanto para os GEBVs como para os EBVs. A quantidade de animais por grupo nos dois cenários indica grande diferença entre grupos com EBV e com GEBV. A inclusão de informação genômica resultou em pequeno aumento nas estimativas de acurácia dos valores genéticos para as características estudadas, em relação aqueles obtidos com base apenas no fenótipo e no pedigree. Touros mais jovens, com poucos filhos, apresentaram maiores acurácias dos GEBVs se comparadas as dos EBVs para GP1, GP4, AOL e EGS, portanto, a seleção desses touros pode favorecer a redução do intervalo de gerações. As características de crescimento, em geral, apresentaram associação genética favorável com as características de carcaça, portanto essa associação poderá ser considerada no programa de melhoramento genético de bovinos de corte da raça Nelore, de modo a auxiliar na seleção para tais características. As análises de agrupamento não-hierárquicas podem ajudar a visualizar as associações genéticas existentes entre as características estudadas, definir quais os animais mais indicados para atender aos objetivos de seleção e observar as diferenças nos valores genéticos dos animais, preditos com e sem o uso de informação genômica.
Traditional selection for economically important quantitative traits is generally performed on the basis of breeding values predicted from phenotypic records of an individual and their relatives. With the availability of genomic information, the prediction of breeding values for complex traits has been vastly improved. The use of genomic prediction may lead to a faster genetic gain compared to that achieved by traditional pedigree-based selection methods. The aim of this study was to evaluate the influence of the inclusion of genomic information on the genetic parameters estimates and the genetic trends for daily weight gain from birth to 120 days of age (GP1), from 120 to 210 days of age (GP2), from 210 to 365 days of age (GP3), from 365 to 450 days of age (GP4), ribeye area (AOL), subcutaneous backfat thickness (EGS) and rump fat (EGP8) in Nelore cattle. Further, the genetic associations between the studied traits were evaluated and, by using non-hierarchical clustering analyses it was verified which groups of animals were most suitable for meeting the selection objectives, in order to contribute to the selection process of the breeding program of the Nelore breed. The genetic parameters estimates were obtained based on the pedigree records of 192,483 animals born between 1934 and 2016, phenotypic records of 80,114 animals and genotypes of 8,652 animals, which were provided by the “Associação Nacional de Criadores e Pesquisadores” (ANCP). The (co) variance components were obtained by Bayesian methodology via Gibbs sampling. The expected breeding values (EBVs) were estimated through the animal model, using the mixed model equations (which uses only matrix A), whereas the genomic EBVs (GEBVs) were obtained through the single-step genomic BLUP methodology (ssGBLUP) (which uses the matrix H). The genetic trends were estimated using a linear regression considering the mean variation of the EBVs and GEBVs according to the year of birth. The multivariate clustering analyses were used for grouping the animals according to the EBVs and GEBVs for the traits studied and to explore the genetic pattern in each grouping obtained. These analyses were divided into two scenarios, one scenario considering only the females and the other only the males. Using the A matrix, built based only on pedigree records, the average estimates for direct heritability (ℎ2) were 0.14±0.02, 0.18±0.01, 0.14±0, 02, 0.12±0.02, 0.37±0.03, 0.17±0.02, 0.28±0.02 for GP1, GP2, GP3, GP4, AOL, EGS and EGP8, respectively. With the inclusion of the genomic information using the H matrix, the average estimates of ℎ2 for GP1, GP2, GP3, GP4, AOL, EGS and EGP8 were 0.17±0.02, 0.19±0.01, 0.15±0.02, 0.11±0.01, 0.32±0.02, 0.18±0.02 and 0.25±0.02, respectively. The average genetic correlations obtained by the two methods were very similar. The highest genetic correlations (matrix A/matrix H) were between AOL and GP1 (0.34±0.10/0.40±0.08), GP2 (0.52±0.06/0.48±0.06) and GP3 (0.37±0.09/0.39±0.07) and between EGS and EGP8 (0.69±0.05/0.66±0.04). Including the genomic information, slight increases in the breeding values accuracy were obtained, with the greater changes in accuracy observed for sires with few offspring. For the genetic trends, significant angular coefficients (p<0.05) were obtained for most of the studied traits, excepting for EGP8 using the GEBV. It was shown that genetic progress for the traits studied was achieved with both GEBVs and EBVs. The number of animals per group in the two scenarios indicates a large difference between groups with EBV and GEBV. The inclusion of the genomic information resulted in a small increase in the breeding values accuracy estimates for the studied traits, in comparison to those obtained with the phenotype and pedigree only based models. Younger sires, with few offspring, presented greater accuracies of the GEBVs when compared to the EBVs for GP1, GP4, AOL, and EGS, so that the selection of these sires may favor the reduction of the generation interval. In general, the growth traits presented a favorable genetic association with the carcass traits, thus, this association could be considered in the Nelore beef cattle breeding program, in order to assist in the selection for such traits. The non-hierarchical clustering analyzes can help to depict the genetic associations between the traits studied, to define the most suitable animals to meet the selection objectives and to observe the differences in the breeding values of the animals predicted with and without using the genomic information.
Traditional selection for economically important quantitative traits is generally performed on the basis of breeding values predicted from phenotypic records of an individual and their relatives. With the availability of genomic information, the prediction of breeding values for complex traits has been vastly improved. The use of genomic prediction may lead to a faster genetic gain compared to that achieved by traditional pedigree-based selection methods. The aim of this study was to evaluate the influence of the inclusion of genomic information on the genetic parameters estimates and the genetic trends for daily weight gain from birth to 120 days of age (GP1), from 120 to 210 days of age (GP2), from 210 to 365 days of age (GP3), from 365 to 450 days of age (GP4), ribeye area (AOL), subcutaneous backfat thickness (EGS) and rump fat (EGP8) in Nelore cattle. Further, the genetic associations between the studied traits were evaluated and, by using non-hierarchical clustering analyses it was verified which groups of animals were most suitable for meeting the selection objectives, in order to contribute to the selection process of the breeding program of the Nelore breed. The genetic parameters estimates were obtained based on the pedigree records of 192,483 animals born between 1934 and 2016, phenotypic records of 80,114 animals and genotypes of 8,652 animals, which were provided by the “Associação Nacional de Criadores e Pesquisadores” (ANCP). The (co) variance components were obtained by Bayesian methodology via Gibbs sampling. The expected breeding values (EBVs) were estimated through the animal model, using the mixed model equations (which uses only matrix A), whereas the genomic EBVs (GEBVs) were obtained through the single-step genomic BLUP methodology (ssGBLUP) (which uses the matrix H). The genetic trends were estimated using a linear regression considering the mean variation of the EBVs and GEBVs according to the year of birth. The multivariate clustering analyses were used for grouping the animals according to the EBVs and GEBVs for the traits studied and to explore the genetic pattern in each grouping obtained. These analyses were divided into two scenarios, one scenario considering only the females and the other only the males. Using the A matrix, built based only on pedigree records, the average estimates for direct heritability (ℎ2) were 0.14±0.02, 0.18±0.01, 0.14±0, 02, 0.12±0.02, 0.37±0.03, 0.17±0.02, 0.28±0.02 for GP1, GP2, GP3, GP4, AOL, EGS and EGP8, respectively. With the inclusion of the genomic information using the H matrix, the average estimates of ℎ2 for GP1, GP2, GP3, GP4, AOL, EGS and EGP8 were 0.17±0.02, 0.19±0.01, 0.15±0.02, 0.11±0.01, 0.32±0.02, 0.18±0.02 and 0.25±0.02, respectively. The average genetic correlations obtained by the two methods were very similar. The highest genetic correlations (matrix A/matrix H) were between AOL and GP1 (0.34±0.10/0.40±0.08), GP2 (0.52±0.06/0.48±0.06) and GP3 (0.37±0.09/0.39±0.07) and between EGS and EGP8 (0.69±0.05/0.66±0.04). Including the genomic information, slight increases in the breeding values accuracy were obtained, with the greater changes in accuracy observed for sires with few offspring. For the genetic trends, significant angular coefficients (p<0.05) were obtained for most of the studied traits, excepting for EGP8 using the GEBV. It was shown that genetic progress for the traits studied was achieved with both GEBVs and EBVs. The number of animals per group in the two scenarios indicates a large difference between groups with EBV and GEBV. The inclusion of the genomic information resulted in a small increase in the breeding values accuracy estimates for the studied traits, in comparison to those obtained with the phenotype and pedigree only based models. Younger sires, with few offspring, presented greater accuracies of the GEBVs when compared to the EBVs for GP1, GP4, AOL, and EGS, so that the selection of these sires may favor the reduction of the generation interval. In general, the growth traits presented a favorable genetic association with the carcass traits, thus, this association could be considered in the Nelore beef cattle breeding program, in order to assist in the selection for such traits. The non-hierarchical clustering analyzes can help to depict the genetic associations between the traits studied, to define the most suitable animals to meet the selection objectives and to observe the differences in the breeding values of the animals predicted with and without using the genomic information.
Descrição
Palavras-chave
Análise de agrupamento, Correlação genética, Herdabilidade, Inferência bayesiana, Tendência genética, ssGBLUP, Bayesian inference, Cluster analyses, Genetic correlation, Genetic trend, Heritability