Estimation of Liveweight from Body Measurements through Best Fitted Regression Model in Dezhou Donkey Breed

Abstract

Dezhou donkey is an excellent local hide, meat and milk breed in Shandong Province. To accelerate the speed of breeding, reproduction and conversation, correlation and multiple regression analysis between body weight (BW) and body size of Dezhou donkey rearing under intensive farms was made by SAS 9.4 software (Statistical Analysis for Windows). A total of 162 donkeys of both gender of age 2 ~ 10 years old were used to investigate the relationships between BW and body dimensional traits (cm) including height of withers, body length; thoracic depth, thoracic girth (TG), thoracic width (TW), circumference of cannon bone (CB), height of rump, rump length (RL) and rump width (RW). The results showed that BW and body measurements have positive and great correlations with R² value ranged from 0.58 to 0.88 (P <0.01). The R² values from the single-parameter equations showed that the TG was highly related to BW (0.72, P < 0.01). The stepwise regression equations were applied to obtain the best prediction equations, and the results indicated that the prediction accuracy for BW was improved with the inclusion of more body measurement variables. The "best fit" models were: BW (kg) = 1.88 × TG + 1.27 × BL + 2.55 × TW + 4.61 × CB + 2.18 × RW + 1.78 × RL - 422.8 (R² = 0.906, P < .01). The predicted BW from the present equations showed the nearest value to the real BW (R² > 0.94, P < .01). In addition, the equations derived to predict the BW of donkeys in Britain, Morocco were less satisfactory for use with the present Dezhou donkey breed because they overestimated or underestimated the BW due to the different donkey breed.

Keywords: Body measurements; Body weight; Dezhou donkey; Prediction equation.