Enhancing the quality of duck egg protein during the pickling process using machine learning and voltage-pulsed electric field

Abstract

This study investigated the effects of voltage-pulsed electric field (PEF) treatment on the quality and structure of duck eggs during pickling. Furthermore, a model was developed to predict the salt content utilizing the predictive abilities of decision trees, support vector regression, artificial neural network (ANN), and genetic algorithm (GA)-ANN models. The GA-ANN model demonstrated a superior predictive accuracy for salt content. The optimal PEF treatment conditions were as follows: pickling solution concentration of 22 %, subjected to 30 voltage-pulses, and PEF intensity of 2.5 kV/cm, with a prediction error of 0.66 % compared to the actual values. After PEF treatment, the pickling process was completed within 20 days, resulting in increased salt and oil content and decreased moisture content in the eggs. Additionally, the textural properties of egg whites and yolks were enhanced. The secondary structure of egg white proteins shifted from β-sheet to random coil, with an increase in α-helix conformation. Compared to the untreated group, the treated egg whites exhibited increased free sulfhydryl groups and surface hydrophobicity. Furthermore, scanning electron microscopy revealed the presence of numerous pores on the surfaces of the treated samples. Taken together, the PEF treatment effectively shortened the pickling time and improved the quality of duck eggs.