酶蛋白质β-发夹模体配体结合位点的统计分析及预测

摘要

从酶蛋白质loop分类数据库ArchDB_EC中整理得到了1397个含配体结合位点β-发夹,对β-发夹中的配体结合位点进行了统计分析.在此基础上,提取序列片段的氨基酸组分、位点氨基酸组分和位点亲疏水特征组分为参数集,采用mRMR方法对该参数集进行优化筛选,将优化后的最优参数子集输入支持向量机,对酶蛋白质β-发夹中几种常见配体结合位点类型进行了初步预测.%A dataset containing 1397 β-hairpins with ligand binding sites is obtained from ArchDB_EC database. Through statistical analysis, a method for predicting ligand binding sites of β- hairpin motif in enzymes is proposed. Based on amino acid composition, amino acid composition of position and Hydropathy characteristics for amino acids of position as parameters, after the selection and optimization of the parameters with the principle of mRMR, by using Support Vector Machine, we simply predict high -frequency types of ligand binding sites in β- hairpin motif of Enzymes.