Evaluation of nonlinear models to define growth curve in Lohi sheep

Abstract

In Pakistan, very few studies have been conducted to define the growth of small ruminants. The objective of present study was to determine the best nonlinear model to describe growth in Lohi sheep. Four nonlinear models (Brody, Gompertz, logistic and von Bertalanffy) with three parameters were used. Data on Lohi sheep was obtained from Small Ruminant Training and Research Centre, Pattoki, Pakistan. The records comprised 10526 age-weight observations on 554 animals (358 females, 196 males; 420 singles and 134 twins) from birth to 18 months of age. Nonlinear models were fitted using easyreg R package in R program to estimate growth curve parameters. The best model was selected on goodness of fit criteria like adjusted coefficient of determination (R2adj), root mean square error (RMSE) and Akaike’s (AIC) and Bayesian information criteria (BIC). Where marginal differences in fit statistic criteria for all models were observed. Brody was the best fitted model to estimate growth curves for all categories of the Lohi sheep. Males had higher mature weight than females. Females had higher rate of maturity and attained mature weight at an earlier age as compared to males. Similarly, differences were also observed in the growth performance of single and twin lambs. Age of dam, year of birth and season of birth also affected growth of animals. The results suggested that Brody model can be used to predict body weight at any age in Lohi sheep.