Sources of variation influencing the use of real-time ultrasound to predict intramuscular fat in live beef cattle

A total of 123 steers of six European breeds (Angus, Simmental, Charolais, Limousin, Blonde d'Aquitaine, Piedmontese) were used (i) to evaluate the precision of the ultrasound-predicted intramuscular fat (USIMF) and its sources of variation using the current Pie QUIP technology and (ii) to develop improved models for predicting USIMF. Steers were slaughtered when they reached the target value of 3.5% USIMF. Hide samples were obtained 3 d before slaughter by shot-biopsy. After slaughter, a sample of the longissimus muscle was used to determine actual chemical intramuscular fat (EEIMF), collagen content and solubility. Among the variables available during a chute-side scanning session, hide thickness and ultrasound subcutaneous fat thickness at the 12th and 13th ribs were shown to be significantly correlated with EEIMF. These two variables were selected as possible independent variables to evaluate the construction of new models. The model with the best fit included USIMF, hide thickness and liveweight and had a standard error of prediction of 0.96%, which is similar to other published technologies. Breed group and collagen-related traits did not influence USIMF estimation. Finally, the revised Pie QUIP technology should be considered as one technology of choice to predict EEIMF content in live animals.