Growth curves in meat-type and laying quail: a Bayesian perspective

Abstract

The aim of this study was to assess the goodness of fit for nonlinear models, using the best model to describe body growth curves, comparing the parameters obtained for gender and one meat-type (Coturnix coturnix coturnix) and two laying (Coturnix coturnix japonica) quail strains, as well as nesting via MCMC (Markov chain Monte Carlo processes) methods under a Bayesian approach. A total of 1,350 one-day-old mixed quail were used: 400 of meat-type, 450 of yellow laying, and 500 of red laying strains distributed in a completely randomized design with three treatments (each treatment corresponded to one strain) and five replications. The experimental period consisted of 1 to 42 days of age. At 21 days of age, quail sexing was performed by means of sexual dimorphism, being individually identified at one day of age with numbered rings, allowing determining growth curves by gender. Birds were reared in a conventional system, fed ad libitum with diets formulated to meet nutritional requirements. Body weight was determined weekly and assessed using nonlinear models: Logistic, Brody, Von Bertalanffy, and Gompertz, whose parameters were estimated under a Bayesian approach via MCMC algorithm by means of BRugs package from the software R. DIC (Deviance Information Criterion) criterion was used to select the best nonlinear model, i.e. the lower the DIC value is, the better the model goodness of fit to the data. Gompertz model was better adjusted to the data regardless the gender or strain. Meat-type quail had the highest asymptotic weights and the highest age at which growth rate was maximum, followed by red and yellow strains. All nestings presented significant differences (p < 0.05) between gender for contrasted parameters. Meat-type, yellow, and red females presented values significantly (p < 0.05) higher for asymptotic weight (370, 203, and 215 g, respectively) when compared to males (274, 131, and 143 g, respectively), which were earlier in body growth. Gompertz model was better adjusted to body weight data of quail regardless the gender or strain and the Bayesian approach allowed obtaining accurate estimations. Meat-type strain presented the highest body asymptotic weight, followed by red and yellow laying strains. Females presented higher asymptotic weight than that found for males of their respective strains but were later in growth.

Keywords

• Deviance Information Criterion
• Equação de Gompertz
• Modelos não lineares
• Peso corporal.