Several factors that affect estimates of scanned loin muscle area at the position of half body length (SEM1/2) by using scanned loin muscle area (SEMLAST) or depth at the position of the last rib were investigated. Records on 566 pigs (Berkshire, Landrace, Large White and Duroc) in 19 herds from different research stations were analyzed. Loin muscle area at body weights of about 90, 105, and 120kg were measured by ultrasound. SEM1/2 were analyzed using the fixed procedure of SAS. A covariate of SEMLAST scanned loin muscle depth at the position of 3 or 6cm side from the last rib using B-mode (BDP3, BDP6) or A-mode (ADP3, ADP6) ultrasonic device and fixed effects of breed, sex, place, weight category, and year-month were included in the model. Models with different regressions within breed or weight category were also used (intra class regression model). Covariates of SEMLAST were statistically significant for all fixed models. Analyses of each breed and regression coefficients of SEMLAST within weight category were similar and not statistically significant. Covariates of loin depth were statistically significant for all fixed models. Coefficients of determination of BDP3 and BDP6 were larger than ADP3 and ADP6, respectively. Coefficients of determination were larger in models with covariate of BDP6 than in BDP3. Covariates of BDP6 within breed or weight category were significantly different. These results show that it is possible to estimate SEM1/2 from SEMLAST by using the model for each breed without the intra class regression model. Additionally, the results suggest that the relationships between loin muscle area and loin depth were different in breeds or stages of growth.

View full abstract