Candidate genes for longitudinal traits under sequential sampling in beef cattle.

Both the measurement age of a longitudinal trait and the common pre-sampling procedures used in beef cattle herds may affect the identification of a functional candidate gene (FCG) that is potentially associated with a trait. To identify the FCG that takes part in the genetic control of body weight at five different ages in a beef cattle population with and without sequential sampling, the animals were weighed at different measurement events, around 330, 385, 440, 495 and 550 days old. Genetic parameters were estimated for body weight at each age using a single trait (STM) and a random regression model (RRM). In addition, two different databases were used to estimate the genetic parameters: the first (DB100) was formed by all animals that were weighed in the five measurement events, and the second (DB70) has records of the same population, considering that 70% of the heaviest animals were selected after each measurement event. For DB100, genome-wide association studies (GWAS) were performed with 21,667 SNP markers to identify genomic windows that explained at least 1% of the genetic variance. Additionally, prioritization analyses were performed and FCGs were selected. We associated seven different FCGs with body weight at different ages. Among them, the gene DUSP10 was suggested as FCG in all five ages evaluated. Genetic parameters estimated for body weight using DB100 were similar when STM and RRM were applied. However, when DB70 was used as phenotypic data, there were differences between the two models. When the STM was applied, there were differences between the genetic parameters estimated for body weight when DB100 or DB70 were used as sources of phenotypes, but not for the estimates obtained with RRM. The importance of each gene for animal growth can change at different ages, and different genes may be more relevant to body weight at each different growth stage for beef cattle. Besides, sequential sampling can affect the GWAS results of a longitudinal trait. The age of the animal when a longitudinal trait is measured and pre-sampling can also contribute to inconsistencies in GWAS results for body weight in beef cattle, depending on the time when that data were collected, and consequently on the identification of FCG between studies, even when models that consider a covariance structure are used.

Utilização de funções lineares e não lineares para ajuste do crescimento de bovinos santa gertrudis, criados a pasto / Use of linear and nonlinear functions to fit the growth of santa gertrudis cattle raised under grazing

Objetivou-se com este trabalho ajustar a curva de crescimento de bovinos da raça Santa Gertrudis utilizando funções spline e não lineares derivadas do modelo Richards e verificar a qualidade do ajuste dos modelos testados, a partir de registros de pedigree provenientes do banco de dados da Associação Brasileira de Santa Gertrudis (ABSG). O melhor modelo para ajustes em fases determinadas foi obtido a partir do modelo linear com cinco segmentos, com pontos de junção aos 179, 264, 421 e 850 dias de idade. Os valores obtidos dos coeficientes de determinação (R2), Quadrado Médio do Resíduo (QMRes) e Erro de predição Médio (EPM) foram de 0,76 a 0,95; 6201,18 a 6840,5 e - 0,0005 a - 0,0493, este último indicando uma pequena superestimação do peso médio observado. Os animais atingiram, em média, 542,8 kg de peso adulto e 9,7 meses no ponto de inflexão. O uso de funções splines é uma alternativa viável ao ajuste de curvas de crescimento e para a correta determinação do desempenho do animal em idades iniciais e tardias. Os valores de peso assintótico e taxa de maturidade obtidos pelas funções não lineares fornecem uma estimativa prática sobre o desempenho final no desenvolvimento dos animais e espera-se que animais maiores sejam menos precoces. Além disso, pesagens mais frequentes deverão proporcionar melhores qualidades de ajuste.(AU)

This research aimed at to adjust the growth curve of Santa Gertrudis cattle using spline functions and nonlinear functions, derivated from Richards model, and to verify the quality of the fitted models, using pedigree records from the Brazilian Association of Santa Gertrudis (ABSG) datafiles. The best model to fit in certain phases was obtained using the linear model with five segments, with the junction points at 179, 264, 421 and 850 days of age. The values of R2, EMS and PEM ranged from 0.76 to 0.95; 6201.18 to 6840.5 and 0.000507 to 0.049348, the latter indicating the slightly overestimation of average weight observed. Spline functions provide a feasible alternative for the adjustment of growth curves, and accurately predict the performance of the animals in early and late ages. Asymptotic weight and mature rate obtained from nonlinear functions provide a practical estimate on final performance of animal growth and it is expected that larger animals will be worse in early finishing. In order to have better goodness of fit, more frequent records of weight should be provided.(AU)

Utilização de funções lineares e não lineares para ajuste do crescimento de bovinos santa gertrudis, criados a pasto / Use of linear and nonlinear functions to fit the growth of santa gertrudis cattle raised under grazing

Objetivou-se com este trabalho ajustar a curva de crescimento de bovinos da raça Santa Gertrudis utilizando funções spline e não lineares derivadas do modelo Richards e verificar a qualidade do ajuste dos modelos testados, a partir de registros de pedigree provenientes do banco de dados da Associação Brasileira de Santa Gertrudis (ABSG). O melhor modelo para ajustes em fases determinadas foi obtido a partir do modelo linear com cinco segmentos, com pontos de junção aos 179, 264, 421 e 850 dias de idade. Os valores obtidos dos coeficientes de determinação (R2), Quadrado Médio do Resíduo (QMRes) e Erro de predição Médio (EPM) foram de 0,76 a 0,95; 6201,18 a 6840,5 e - 0,0005 a - 0,0493, este último indicando uma pequena superestimação do peso médio observado. Os animais atingiram, em média, 542,8 kg de peso adulto e 9,7 meses no ponto de inflexão. O uso de funções splines é uma alternativa viável ao ajuste de curvas de crescimento e para a correta determinação do desempenho do animal em idades iniciais e tardias. Os valores de peso assintótico e taxa de maturidade obtidos pelas funções não lineares fornecem uma estimativa prática sobre o desempenho final no desenvolvimento dos animais e espera-se que animais maiores sejam menos precoces. Além disso, pesagens mais frequentes deverão proporcionar melhores qualidades de ajuste.

This research aimed at to adjust the growth curve of Santa Gertrudis cattle using spline functions and nonlinear functions, derivated from Richards model, and to verify the quality of the fitted models, using pedigree records from the Brazilian Association of Santa Gertrudis (ABSG) datafiles. The best model to fit in certain phases was obtained using the linear model with five segments, with the junction points at 179, 264, 421 and 850 days of age. The values of R2, EMS and PEM ranged from 0.76 to 0.95; 6201.18 to 6840.5 and – 0.000507 to – 0.049348, the latter indicating the slightly overestimation of average weight observed. Spline functions provide a feasible alternative for the adjustment of growth curves, and accurately predict the performance of the animals in early and late ages. Asymptotic weight and mature rate obtained from nonlinear functions provide a practical estimate on final performance of animal growth and it is expected that larger animals will be worse in early finishing. In order to have better goodness of fit, more frequent records of weight should be provided.