摘要
金属离子作为一种重要的配体与蛋白质结合在细胞代谢、物质运输和信号传导等过程中发挥着重要作用。因此,理论上准确识别蛋白质与金属离子配体结合残基有十分重要的意义。选取10种金属离子配体作为研究对象,以氨基酸、氨基酸的物化特征和预测的结构信息作为基础特征参数,添加能量和倾向性因子两种新的预测参数,使用支持向量机(SVM)算法对蛋白质-金属离子配体结合残基进行预测,五交叉检验和独立检验都得到了较好的预测结果,独立检验下Fe^(2+)、Fe^(3+)、Co^(2+)和K^(+)得到了好于前人的预测结果。
As an important ligand binding to proteins,metal ions play an important role in the process of cell metabolism,material transport and signal transduction.Therefore,it is of great significance to accurately identify the binding residues of proteins and metal ion ligands in theory.In this paper,10 kinds of metal ion ligands are selected as the research objects,and amino acids,amino acid physicochemical characteristics and predicted structural information are used as basic characteristic parameters,and two new prediction parameters,energy and propensity factor,are added,to predict protein-metal ion ligand binding residues with the support vector machine algorithm.Both the 5-fold cross-validation and the independent test have obtained good prediction results.The independent tests indicate that the prediction results of Fe^(2+),Fe^(3+),Co^(2+) and K^(+) are better than the previous ones.
作者
杨彩芸
胡秀珍
尤肖肖
王子洋
郝四喜
胡慧敏
YANG Caiyun;HU Xiuzhen;YOU Xiaoxiao;WANG Ziyang;HAO Sixi;HU Huimin(School of Sciences,Inner Mongolia University of Technology,Hohhot,Inner Mongolia 010051,China)
出处
《内蒙古工业大学学报（自然科学版）》
2022年第2期105-114,共10页
Journal of Inner Mongolia University of Technology：Natural Science Edition
基金
国家自然科学基金项目(61961032)。
