COMPARISON OF THREE NONLINEAR MODELS TO DESCRIBE THE GROWTH CURVE OF HOLSTEIN-FRIESIAN BULLS RAISED UNDER EGYPTIAN CONDITIONS

Abstract

The current study aimed to estimate the growth curve parameters through three non-linear models (Logistic, Gompertz and Richards) to determine which model best fits the data. Live weight records of 102 HolsteinFriesian bulls collected between 2017-2019 from a Holstein-Friesian herd that belongs to the Association of Livestock Development (ELLahhamy farm), located thirty kilometers west of Fayoum Governorate. In this work, the parameters of the studied models, asymptotic weight (A), constant of integration (b) and the maturation rate (K) ranged from 626.15 kg to 879.82 kg, 2.708 to 11.08, and 0.0035 to 0.008, respectively. According to the studied parameters of growth functions, Gompertz reached the highest numerical estimated value for (A) and the Logistic function had the lowest value. Parameter (K) estimate by the Gompertz model was similar to that obtained by the Richards model; both values were lower than those attained through the Logistic model (0.008). The inflectionpoint traits, time at Point of inflection (IPT) and weight at point of inflection (IPW) estimates ranged from 300.64 kg to 314.59 kg and 323.92 days to 336.14 days, respectively. The Richards model has the highest estimates of IPW and IPT comparedto the other models, also it had the best adjustment according to model goodness of fitcriteria, by having the lowest values for Akaike information criterion (AIC), Schwarz Bayesian information criter ion (BIC), Mean square error (MSE) and highest coefficient of determination (R2, ,14489.18, 14510.6, 317.37 and 0.9983) followed by the Gompertz, and logistic functions.