基于计算机辅助水解技术的芝麻寡肽体内转运性质预测

Prediction on transport properties in vivo of oligopeptides peptide in Zhima based on computer-assisted proteolysis

目的基于计算机辅助技术预测芝麻寡肽的体内转运性质。方法收集芝麻中主要蛋白序列,以计算机辅助技术进行蛋白模拟消化水解,将水解获得的寡肽构建虚拟结构数据库。构建寡肽转运蛋白2(Pep T2)底物的支持向量机(SVM)模型后,通过数据处理和参数寻优方法挑选最优模型,预测芝麻Pep T2的底物成分。结果芝麻寡肽数据库共包括346条寡肽,利用筛选获得45个描述符,创建了Pep T2底物的SVM模型,模型的准确度、灵敏度和特异性均在85%以上。经受试者工作特征(ROC)分析,其曲线下面积(AUC)>0.9。利用Pep T2底物模型预测芝麻寡肽数据库,发现36%的芝麻寡肽具有通过Pep T2转运的潜在活性。结论利用计算机辅助水解技术构建了中药芝麻寡肽数据库,以Pep T2底物SVM模型预测芝麻寡肽的体内转运性质,模型准确、稳定、灵敏、特异性高,为开展寡肽类中药成分的性质研究提供了新的方法。

Objective To predict transport properties in vivo of oligopeptides in Zhima（ Sesame,Sesamum indicum） based on computer-assisted proteolysis. Methods Main protein sequences in Zhima were collected,protein simulated digestive hydrolysis was conducted with computer-assisted proteolysis,and obtained oligopeptides were taken to establish virtual structure database. A support vector machine（ SVM） model of Pep T2 substrate was established. Through data treating and parameter optimization to select the bast modef for predicting Pep T2 substrate components in Zima. Results There were 346 oligopeptides in the database of Zhima oligopeptides and 45 descriptors obtained after screening. SVM model of Pep T2 substrate was established and its accuracy,sensitivity and specificity were all higher than85 %. The analysis of receiver operating characteristic curve（ ROC） showed that area under curve（ AUC） was greater than 0. 9. The prediction of the database of Zhima oligopeptides by using SVM model of Pep T2 substrate showed that 36 % Zhima oligopeptides had latent activities of transporting by Pep T2.Conclusion A database of Zhima oligopeptides is established by using computer-assisted proteolysis. The SVM model of Pep T2 substrate is used to predict transport properties of Zhima oligopeptides,and it is accuracy,stable and sensitive with high specificity. The study provides a new method for the researches on absorption properties of oligopeptides components in Chinese materia medica.