Modeling the relationships between quality and biochemical composition of fatty liver in mule ducks

The fatty liver of mule ducks, i.e. French 'foie gras', is the most valuable product in duck production systems. Its quality is measured by the technological yield, which is the opposite of the fat loss during cooking. The purpose of this study was to determine whether biochemical measures of fatty liver could be used to accurately predict the technological yield (TY). Ninety one male mule ducks were bred, overfed and slaughtered under commercial conditions. Fatty liver weight (FLW) and biochemical variables, such as dry matter content (DM) and lipid (LIP) and protein content (PROT) were collected. To evaluate evidence for nonlinear fat loss during cooking, we compared regression models describing linear and nonlinear relations between biochemical measures and TY. We detected significantly greater than linear relation between DM and TY. Our results indicate that LIP and PROT follow a different pattern (linear) than DM and showed that LIP and PROT are nonexclusive contributing factor to TY. Other components such as carbohydrates other than those measured in this study could contribute to DM. Stepwise regression for TY was performed. The traditional model with the FLW was tested. The results showed that the weight of the liver is of limited value in the determination of fat loss during cooking (R2 = 0.14). The most accurate TY prediction equation included DM (in linear and quadratic terms), FLW and PROT (R2 = 0.43). Biochemical measures in the fatty liver were more accurate predictors of TY than liver weight. The model is useful in commercial conditions because DM, PROT an FLW are non invasive measures.