摘要
提出一种基于柔性神经树的蛋白质结构预测方法,将近似熵和蛋白质序列的疏水特性作为伪氨基酸组成的特征。对数据集中的每一条蛋白质进行特征提取。对于一个蛋白质样本,用一个27-D伪氨基酸组成作为其特征,伪氨基酸组成特征作为输入数据,柔性神经树作为预测工具,分类方法采用M-ary方法,数据集选用640数据集。仿真结果表明,该方法具有较好的优化性能,提高了预测的准确率。
This paper proposes a method of protein structural prediction classes based on flexible neural tree. The approximate entropy and hydrophobicity pattern of a protein sequence are used to characterize the Pseudo-Amino Acid（PseAA） components. It extracts features of protein in data set. For a given protein sequence sample, a 27-D PseAA composition is gen^rated as its descriptor. PseAA composition features as input data, the flexible neural tree is adopted as the prediction engine. A classification method named M-ary classifier is introduced. The 640 protein sequence is used as the dataset. Experimental result shows the method has better optimization of performance and improves the predictive accuracy rate.
出处
《计算机工程》
CAS
CSCD
北大核心
2011年第1期159-160,163,共3页
Computer Engineering
基金
国家自然科学基金资助项目(60573065)
山东省自然科学基金资助项目(Y2007G33)
关键词
蛋白质结构分类
伪氨基酸组成
近似熵
疏水性
柔性神经树
protein structure classification
Pseudo-Amino Acid（PseAA） composition
approximate entropy
hydrophobicity
flexible neural tree
