Evaluation of Non- linear Growth Curves Models for Native Slow-growing Khazak Chickens

Native poultry is a valuable genetic source with high resistance against diseases providing an important subject for breeding programs. The non-linear mathematical modeling of the growth pattern may partly explain the relationship between requirements and body weight to precise feeding that plays a vital role in the animal enterprises. A study was conducted to compare five non-linear models including Gompertz, Richards, Lopez, Logistic, and Von Bertalanffy describing the growth curve of Khazak native chickens. A total of 120 Khazak chickens (male and female) were individually weighed from 0 to 29 weeks under the same condition. The models were fitted on body weight data set and then evaluated by goodness-of-fit criteriaincluding root mean square error (RMSE), Bayesian information criterion (BIC), Akaike information criterion (AIC), and adjusted coefficient of determination (R2Adj). Based on goodness-of-fit criteria, Lopez model was the most suitable one for describing the growth curves in female and male chickens. The effect of sex was significantly important on curve parameters in all models (P< 0.05). The highest and lowest initial weight (W0) parameter was estimated by Logistic and Richards models, respectively, however, the other parameters of the growth curves were higher in Lopez model compared to others. Male chickens had higher values for age (ti), and weight (wi) at inflection point than females. Using an appropriate model to describe the growth curve in native Khazak chickens could increase the accuracy of selection for rapid growth at early stages of age.