Flexible Alternatives to Modelswidely Used for Describing Growth in Japanese Quail

Abstract

Mathematical functions for modeling the growth of animals are separated into two groups called fixed and flexible according to their inflection behaviors. The aims of this study were to compare fixed and flexible growth functions and to determine the best fit model for the growth data of Japanese quail. The long term growth patterns of Japanese quail were monitored during 110 days from day-old to slaughtering, and were analyzed. The fixed and flexible models used in the study were Gompertz, Logistic, Von Bertalanffy and Levakovich, Janoschek, Morgan Mercer Flodin, respectively. The differences in weight between growth profiles of female and male birds were tested via profile analysis also known as Repeated MANOVA. The results of Profile Analysis revealed that females' live weights were heavier (P<0.01). In order to evaluate growth functions, the R-2 (coefficient of determination), adjusted R-2 (adjusted coefficient of determination), MSE (mean square error), AIC (Akaike's information criterion) and BIC (Bayesian information criterion) goodness of fit criteria were used. In the study, it was determined that the best fit model from female and male quail growth data on the basis of these goodness of fit criteria is Janoschek model which has a flexible structure. The Janoschek model is not only important because it has a higher number of parameters with biological meaning than the other functions (the mature weight and initial weight parameters), but also because it was not previously used in the modeling of the quail growth.