Effect of short-term depuration on the flavor of crucian carp (Carassius auratus) and exploration of prediction model for fishy odor

Abstract

This study investigated the changes of flavor in crucian carp under short-term depuration (0, 2, 4, 6, 8, 10 and 12 days), and established prediction Model 1 (depuration time - fishy odor in meat) and Model 2 (fishy odor in mucus - fishy odor in meat) using back propagation-artificial neural network (BP-ANN) and optimized by genetic algorithm (GA). It showed that the earthy odor 2-MIB and GSM were significantly reduced by 28.87 % and 46.62 % after 6 days of depuration. Hexanal, nonanal and 1-hexanol were decreased significantly on the 6th day (P < 0.05) compared to the control group, indicating 6 days of depuration is effective to improve the odor of crucian carp. The prediction model based on depuration time and mucus odor showed that the BP-ANN could predict with the relative error ranged from 0.2 to 2.73, while GA could optimize the accuracy of the BP-ANN with the relative error 1.62 and 0.42 in Model 1 and 2, which showed that the GA-BP-ANN could better predict fishy odor of crucian carp during depuration, and also confirmed the feasibility of fishy odor prediction using fish mucus.