A Noninvasive Approach to Predict Body and Carcass Weights in Goats and Sheep using in vivo Body Measurements

This study aimed to develop and validate predictive models for estimating body weight (BW) and hot car-cass weight (HCW) in small ruminants using in vivo morphometric measurements. A total of 400 animals (250 sheep and 150 goats) were used for BW prediction, and among them, 200 sheep and 64 goats were slaughtered to develop HCW models. The in vivo measurements included chest circumference (CC), body length (BL), and withers height (WH). Model performance was evaluated through K-fold cross-validation, considering the coefficient of determination (R²), root mean square error of cross-validation (RMSECV), and prediction bias (BIAS). For BW, the combined-species simple model achieved R²= 0.90 and RMSECV= 3.34 kg, while the multiple model yielded R²= 0.91 and RMSECV= 4.44 kg. Sheep-specific models showed R²= 0.82 and RMSECV= 3.47 kg for the simple model, and R²= 0.85 and RMSECV= 4.33 kg for the multiple model. Goat models reached R²= 0.89 and RMSECV= 2.95 kg (simple), and R²= 0.90 and RMSECV= 4.29 kg (multiple). For HCW, the combined simple model (R²=0.79; RMSECV=1.89 kg) and the sheep-specific simple model (R²=0.75; RMSECV=1.90 kg) showed good predictive ability. The simple model for goats presented moderate predictive power (R²=0.73; RMSECV=1.75 kg), whereas the multiple model was not significant. In conclusion, BW and HCW can be accurately estimated using simple linear regression models, which may be applied either separately or jointly across species.