摘要
金属离子是一种重要的蛋白质结合配体,蛋白质与离子配体结合才能实现其功能,预测蛋白质—金属离子配体结合残基对分子药物设计研究具有重要意义.本文以5种金属离子配体为研究对象,基于蛋白质序列,使用深度学习算法对蛋白质—金属离子配体结合残基进行预测,五交叉检验下得到了较好的预测结果.通过对深度学习算法超参数进行优化,预测性能进一步提升.在最优超参数下,深度学习算法独立检验的预测结果好于随机森林算法.
As metal ion is an important protein binding ligand and protein can only perform its function after combining with an ionic ligand,the prediction of protein-metal ion ligand binding residue is of great significance for molecular drug design and research.In this paper,five metal ion ligands are taken as the research objects and a deep learning algorithm is used to predict the protein-metal ion ligand binding residues based on protein sequences.Quite good prediction results are obtained by five-fold cross-validation.The prediction properties are further improved by the optimization of the hyper-parameters of the deep learning algorithm.The prediction results of the independent test of the deep learning algorithm are better than those of the random forest algorithm.
作者
孙锴
胡秀珍
王子洋
SUN Kai;HU Xiu-zhen;WANG Zi-yang(College of Scuence,Inner Mongolia University of Technology,Hohhot 010051,China)
出处
《内蒙古工业大学学报（自然科学版）》
2020年第6期410-417,共8页
Journal of Inner Mongolia University of Technology：Natural Science Edition
基金
国家自然科学基金项目(61961032)
内蒙古自然科学基金项目(2019BS03025)。
