PSVII-17 Correlation between loin depth and eye area in pigs: Development of an efficient prediction model using ultrasound measurements.

Abstract This study analyzed the correlation between loin depth and eye area to develop an efficient prediction model for standardizing measurements of key pig carcass traits, including backfat thickness, loin eye area, and intramuscular fat. Using 87 pigs (16 Duroc, 12 Landrace, and 59 Yorkshire), ultrasound measurements were taken with A-mode and B-mode devices to measure loin depth and eye area. Breed-specific ratios between loin depth and width were calculated from actual carcass measurements. Results showed distinct ratio patterns across breeds: for Duroc, loin width = loin depth / 1.55; for Yorkshire, loin width = loin depth / 1.73; and for Landrace, loin width = loin depth / 1.66. A prediction model was developed using the elliptical area formula based on these breed-specific characteristics. The prediction equation was formulated as: Predicted loin eye area = (loin depth/2) × (predicted loin width based on breed ratio/2) × π. This model uses ultrasound-measured loin depth and applies breed-specific ratios to estimate loin width, which is then incorporated into the elliptical area formula to calculate loin eye area. Statistical analysis revealed strong correlations between predicted and actual loin eye areas across all breeds, with coefficient of determination (R²) values consistently exceeding 0.94. Specifically, Landrace showed the highest R² value at 0.961, while Duroc and Yorkshire showed 0.947 and 0.952, respectively. These findings confirm that estimating the loin eye area using only loin depth measurements is highly reliable across all three breeds when appropriate breed-specific ratios are applied. This prediction model simplifies the pig carcass evaluation process while maintaining high accuracy, potentially contributing to improved efficiency in swine breeding programs. The model allows for accurate assessment of production traits with reduced time and effort during ultrasound measurement of live animals, offering practical value for field applications. Future research should focus on validating the model across more diverse environments and conditions to enhance its generalizability.