Predicting fat, lean and the weights of primal cuts for growing pigs of different genotypes and sexes using computed tomography.

Abstract

The aim of the present study was to find single equations to predict the amounts of fat, lean, and the weights of the primal cuts (ham, loin, belly, and shoulder) as well as ham composition of pigs from 30 to 120 kg BW of different genotypes (GEN; Exp. 1) and sexual conditions (SEX; Exp. 2). Two types of regression equations, taking into account different work situations, were developed: 1) research applications, using computed tomography (CT) parameters, and 2) potential on-farm applications, which could be obtained using easily accessible equipment. Two data sets were used: Exp. 1 included 90 gilts from 3 different GEN: 30 Duroc × (Landrace × Large White), 30 Pietrain × (Landrace × Large White), and 30 Landrace × Large White, and Exp. 2 included 92 Pietrain × (Landrace × Duroc) pigs of different SEX: 24 each of females, entire males, castrated males, and 20 immunocastrated males. Pigs were fully CT scanned in vivo at 30, 70, 100, and 120 kg BW. A subsample of pigs of each GEN ( = 5) or SEX ( = 4) were slaughtered at 30, 70, and 100 kg BW, and all remaining pigs were slaughtered after weighing and scanning at 120 kg BW. For all the slaughtered pigs, the 4 main cuts were fully (GEN) or partially dissected (SEX). CT images were analyzed and used to predict the lean and fat contents as well as the weights of the primal cuts and the composition of the ham. Total amounts of fat and lean for both populations were predicted with high levels of accuracy ( = 0.994 and 0.993, respectively) and proportions of random error for GEN and SEX effects (0.998 and 0.946 for the fat and 0.997 and 0.836 for the lean predictions, respectively). Moreover, the composition of ham (fat, lean, and bone) was very well predicted with high proportions (> 80%) of random error for GEN and SEX effect using CT and potential on-farm predictors.