BIASES ASSOCIATED WITH GENOTYPE AND SEX IN PREDICTION OF FAT-FREE LEAN MASS AND CARCASS VALUE IN HOGS

Carcass and live measurements of 165 market hogs that represented seve n genotypes were used to investigate genotype and sex biases associate d with the prediction of fat-free lean mass (FFLM) and carcass value. Carcass value was determined as the sum of the product of weight of in dividual cuts and their average unit prices adjusted for slaughter and processing costs. Independent variables used in the prediction equati ons included carcass measurements, such as optical probe, midline rule r, ribbed carcass measurements, and electromagnetic scanning (EMSCAN), and live animal ultrasonic scanning. The effect of including subpopul ation mean values of independent variables in the prediction equations for FFLM and carcass value was also investigated. Genotype and sex bi ases were found in equations in which midline backfat, ribbed carcass, EMSCAN, and live ultrasonic scanning were used as single technology s ets of measurements. The prediction equations generally undervalued ge notypes with above-average carcass value. Biases were reduced when mea surements of combined technologies and mean adjusted variables were us ed. The FFLM and carcass value of gilts were underestimated, and they were overestimated of barrows. Equations that combined OP and EMSCAN t echnologies were the most accurate and least biased for both FFLM and carcass value. Equations that included carcass weight and midline last -rib backfat thickness measurements were the least accurate and most b iased. Genotype and sex biases must be considered when predicting FFLM and carcass value.