采用广义回归神经网络建立酪蛋白乳化性与疏水性关系

为了研究琥珀酰化修饰后酪蛋白乳化性与疏水性关系,该文以琥珀酰化牦牛乳酪蛋白为研究对象,分析了不同酰化程度酪蛋白乳化性及疏水性变化趋势,采用广义回归神经网络建立了牦牛乳酰化酪蛋白乳化性与疏水性关系模型。结果显示,琥珀酰化牦牛乳酪蛋白乳化性和疏水性均与酰化程度、pH值有关,pH值为5以上,随着酰化程度的增加,酪蛋白乳化活性增大;等电点附近,酪蛋白乳化活性较差,等电点之后乳化活性迅速增大。pH值介于2-6时,所有酪蛋白乳化稳定性较强,pH值介于6-11之间时,酪蛋白乳化稳定性差异较小,pH值为12时乳化稳定性有所增加。酪蛋白内荧光与1-苯胺基萘-8-磺酸(1-aniline napthalene-8-sulfonic acid,ANS)外源荧光最大荧光强度和最大发射波长随酰化程度及pH值变化表现出较为复杂的关系。通过广义回归神经网络(generalized-regression-neu-network,GRNN)建立了牦牛乳酪蛋白疏水性参数、pH值、酰化程度与乳化性关系,网络模型对乳化性的预测相对误差小于10%,预测结果良好。研究结果为酪蛋白乳化性研究提供了参考依据。

Establishing relationship between hydrophobicity and emulsification of caseins using generalized-regression-neural-network

Abstract: Yaks (Bos grunniens) from the Qinghai-Tibetan plateau are the sole source of milk of local inhabitants. In the past years, yak milk has attracted more and more attention, due to its increasing demand. Yak milk is widely used to produce butter, soft and hard cheeses, yogurt, milk powder, Qula, and casein-containing products. Caseins, the major protein in milk, are commonly used in the food, chemical, cosmetic, and pharmaceutical industries. Yak caseins are widely used to produce high quality food ingredients and soaps, glues, leather polishing reagents, and clothing, among others. However, compared to cow caseins, yak caseins have poor solubility, which affects the other functional properties. The poor solubility of yak caseins is mostly attributed to its conformation. Succinylation is the most frequently used modification method, which changes the conformation and increases the net charge of protein, which improves its functional properties consequently. It was indicated that the reaction conditions were mild and easy to control, the modified level was higher than other chemical modification, and its modification effect was outstanding.In this study, yak milk caseins were selected as materials and succinylation was used to modify the yak casein, and the effect of succinylation on casein emulsification and hydrophobility was studied. The GRNN was also used to build the model of relationship between pH, succinylated level, the main parameters of caseins hydrophobility, emulsifying properties, and it was proved to have good predictability.The results showed that the emulsifying properties of protein were affected by pH, spacial conformation, and hydrophobility mainly, and they were non-linearly related. The emulsification and hydrophobility of yak caseins were influenced by the degree of succinylation and pH. Above a pH of 5, with the increase of degree of succinylation, emulsifying activity of caseins increased. Near the isoelectric point, emulsifying activity was poorer, and then increased rapidly with the increasing of pH. The emulsion stability of modified casein was stronger with pH between 2-6, which changed slowly with pH between 6-11, and then increased greatly with pH at 12, and the emulsion stability of 10 min increased obviously. The values of native and ANS fluorescence intensity and maximum emission wavelength of native and ANS fluorescence of modified casein showed a more complex relationship with succinylated degree and the pH. GRNN had a good performance in deal with complex relationships of non-linear, which was used to establish the relationship model between succinlated level, pH, vital parameters of hydrophobility, and emulsifying properties in order to predict the emulsifying activity and emulsion stability with 10 min and 30 min. The error of predicted values was within 10%, so the results of prediction were credible. The model would save time and the cost of test about studying and predicting yak caseins emulsifying properties. The results could provide references for the study on emulsification of yak caseins and succinylation caseins.