三元分子体系氨基酸+蛋白质变性剂+水的密度预测

Density prediction of ternary systems of α-amino acids-denaturant-water via neural networks

以298.15K时α-氨基酸+水和蛋白质变性剂+水二元实验数据为基础,利用改进的逆传播算法训练神经网络,构造混合物组成与密度关系的模型,用于对α-氨基酸(甘氨酸、DL-α-丙氨酸、DL-α-氨基正丁酸、DL-α-缬氨酸和DL-α-亮氨酸)-蛋白质变性剂(乙酸钠,丁酸钠,己酸钠和辛酸钠)-水三元混合物体系的密度进行预测,所得预测结果与实验值进行了对比。结果表明,神经网络作为三元混合物体系密度的预测是令人满意的。

Muhilayer artifial neutral networks with improved reversed-propagation of error were employed to predict the densities of system of α-amino acids (glycine, DL-alanine, DL-α-amino-n-butyric acid, DL-α-valine and DL-α-leucine) - denaturant (sodium acetate, sodium butyrate, sodium caproate and sodium caprylate) - water based on the binary experimental data of amino acids-water and denaturant-water systems at 298.15 K. Comparing with the experimental values the prediction is satisfactory.