酪蛋白-胰酶水解历程分子量变化模拟与三维表征

食源性活性多肽的生物活性功能具有很强的分子量依赖性,不同分子量的多肽有不同的功能。本文运用人工神经网络模拟了酪蛋白在胰酶作用下酶解全过程的分子量分布变化,并基于此模型进行了酶解历程三维表征。使用BP神经网络模拟,隐含层为2层,每层含30个节点时拟合效果最好,回归系数R²可达0.9922。将预测数据用于复杂酶解历程的三维表征,可通过该三维表征图迅速判断各集总分子量多肽区的最佳制备水解度。

Simulation and 3D plot of molecular weight distribution of released peptides from pancreatic hydrolysis of casein

Biological activities of released peptides from pancreatic hydrolysis of bovine casein depend on the molecular weight.Peptides with different molecular weights have different functions.In this paper, typical neural network was applied to simulate molecular weight distribution and get 3D plot of released peptides from pancreatic hydrolysis of casein.During the simulation, when hidden layer was 2 and hidden nodes were 30, R² was 0.9922.The result was good when back propagation (BP) artificial neural network (ANN) was used to predict molecular weight distribution of released peptides.The prediction result was further used to get 3D plot of molecular weight distribution of casein pancreatic hydrolysates.The best degree of hydrolysis could be chosen by this 3D plot, when the peptides with given molecular weights were required.