基于特征选择和支持向量机的HIV-1型蛋白酶剪切位点预测  

HIV-1 protease cleavage site prediction based on feature selection and support vector machine

作　　者：袁哲明[1,2] 张弘杨[1,2] 陈渊[1] 

机构地区：[1]湖南农业大学湖南省作物种质创新与资源利用重点实验室,长沙410128 [2]湖南农业大学植物病虫害生物学与防控湖南省重点实验室,长沙410128

出　　处：《吉林大学学报（工学版）》2017年第2期639-646,共8页Journal of Jilin University:Engineering and Technology Edition

基　　金：高等学校博士学科点专项科研基金项目(20124320110002);长沙市科技计划项目(K1406018-21)

摘　　要：为了提高HIV-1型蛋白酶剪切位点的预测准确性,提出一种基于特征选择和支持向量机的剪切位点预测模型。首先,通过对5830个样本的HIV-1型蛋白酶剪切位点数据集进行分析,根据最小冗余最大相关理念,采用可自动终止法选择剪切位点的特征向量;然后,将特征向量输入到支持向量机进行学习和训练,建立HIV-1型蛋白酶剪切位点的分类模型;最后,采用Matlab 2014的仿真工具箱进行仿真测试。实验结果表明:本文模型在特征最少的条件下,剪切位点预测精度优于参比模型及文献报道,且所选择的特征向量具有较好的可解释性及生物学意义。In order to improve the prediction accuracy of the HIV-1 protease cleavage site, a shear prediction model based on feature selection and support vector machine is proposed. First, by analysis of the cleavage site dataset of 5830 samples, and using absorption minimum redundancy maximum relevance concept, the automatic termination method is employed to select the cleavage site feature vectors. Then, the feature vector is input to a support vector machine for learning and training to build the classification model of splice sites. Finally, simulation is carried out using MATLAB 2004 simulation toolbox. Results show that the proposed model has better prediction accuracy than that of the reference models and literature report. The selected features have good interpretability and biological significance.

关 键 词：生物物理学 剪切位点预测 特征选择 最小冗余最大相关 支持向量机 

分 类 号：Q6[生物学—生物物理学]