基于多特征信息及Ma-Ada 多分类器融合的蛋白质结构类预测

doi:10.3969/j.issn.0258-8021. 2013. 05.010

摘要
图/表
参考文献
相关文章 (0)

全文: PDF (496 KB) HTML (1 KB)
输出: BibTeX | EndNote (RIS)

摘要蛋白质序列特征表示和机器学习算法是影响蛋白质结构类预测效果好坏的两个重要方面。本研究基于k字统计频率和k片段位置分布两种特征提取方法，将分别提取到的氨基酸序列信息和物理化学性质信息同蛋白质二级结构信息进行融合，建立17维和57维的特征信息集，并尝试在Adaboost.M1算法中引入MultiAgent多智能体融合的思想，提出了一种Ma-Ada多分类器融合算法。该算法作为蛋白质结构类的预测工具，充分挖掘了单分类器度量层信息以及各个单分类器之间的交互融合信息。实验结果表明，Ma-Ada算法在Z277、Z498、1189和D640四个蛋白质数据集的57维特征信息集上的分类率分别达到了91.3%、96.8%、85.3%和87.2%，在17维特征信息集上的分类率也分别达到了90.6%、95.8%、84.8%和88.3%。与其它蛋白质结构类预测方法的结果相比，本方法能够获得较好的分类率。

	服务

	把本文推荐给朋友
	加入我的书架
	加入引用管理器
	E-mail Alert
	RSS
	作者相关文章
	郑斌厉力华^*

关键词 ：蛋白质结构类预测, 特征信息集, Ma-Ada多分类器融合

Abstract：Protein sequence feature and machine learning algorithm are two important aspects to determine the results of protein structural class prediction. In this study, we established 17-D and 57-D feature information sets through fusing the sequence information, physical and chemical information with the secondary structure information based on the kword statistical frequency and the kfragment distribution feature extraction method. By introducing MultiAgent's idea into Adaboost.M1 algorithm, a novel method for protein structural class prediction, called MaAda multiclassifier fusion algorithm, was proposed, which fully utilized the information of the single classifier metric layer and the fusion of information among individual classifiers. Four protein datasets including Z277, Z498, 1189, D640 were used to validate the performance of the Ma-Ada algorithm. Classification accuracies are 91.3 %, 96.8 %, 85.3% and 87.2 % with 57-D features, and 90.6 % , 95.8 %, 84.8 % and 88.3 % with 17 D features on datasets Z277, Z498, 1189 and D640, respectively. The experimental results show better.

Key words： protein structural class prediction feature information set Ma-Ada multiclassifier fusion

基金资助:国家自然科学基金（61271063）；国家重点基础研究发展计划（973计划）（2013CB329502）；国家杰出青年科学基金（60788101）

引用本文:

郑斌厉力华^*. 基于多特征信息及Ma-Ada 多分类器融合的蛋白质结构类预测[J]. 中国生物医学工程学报, 2013, 32(5): 580-587.
ZHENG Bin LI Li Hua^*. Protein Structural Class Prediction Based on Multi-Feature and Ma-Ada Multi-Classifier Fusion. journal1, 2013, 32(5): 580-587.

链接本文:

http://cjbme.csbme.org/CN/10.3969/j.issn.0258-8021. 2013. 05.010 或 http://cjbme.csbme.org/CN/Y2013/V32/I5/580

［1］Levitt M, Chothia C. Structural patterns in globular proteins［J］. Nature, 1976, 261(5561): 552-558.
［2］Nakashima H, Nishikawa K. Discrimination of intracellular and extracellular proteins using amino acid composition and residuepair frequencies［J］. Journal of Molecular Biology, 1994, 238(1): 54-61.
［3］Bu Weishu, Feng Zhiping, Zhang Ziding, et al. Prediction of protein (domain) structural classes based on aminoacid index［J］. European Journal of Biochemistry, 1999, 266(3): 1043-1049.
［4］Chou Kuochen. Prediction of protein cellular attributes using pseudoamino acid composition ［J］. Proteins: Structure, Function, and Bioinformatics, 2001, 43(3): 246-255.
［5］Ding CHQ, Dubchak I. Multiclass protein fold recognition using support vector machines and neural networks［J］. Bioinformatics, 2001, 17(4): 349-358.
［6］Liu Taigang, Zheng Xiaoqi, Wang Jun. Prediction of protein structural class using a complexitybased distance measure［J］. Amino Acids, 2010, 38(3): 721-728.
［7］Wu Li, Dai Qi, Han Bin, et al. Prediction of protein structural class using a combined representation of proteinsquence information and support vector machine［C］//Bioinformatics and Biomedicine Workshops (BIBMW). HongKong: IEEE, 2010:101-106.
［8］Cai YD, Feng KY, Lu WC, et al. Using logitboost classifier to predict protein structural classes［J］. Journal of Theoretical Biology［J］.2006, 238(1): 172-176.
［9］Feng KY, Cai YD, Chou KC. Boosting classifier for predicting protein domain structural class［J］. Biochemical and Biophysical Research Communications, 2005, 334(1): 213-217. 
［10］Dai Qi, Wu Li, Li Lihua. Improving protein structural class prediction using novel combined sequence information and predicted secondary structural features［J］. Journal of Computational Chemistry, 2011,32(16): 3393-3398.
［11］Freund Y, Schapire RE. Experiment with a new boosting algorithm［C］//Machine LearningInternational Workshop. Morgan: Kaufmann Publishers, 1996:148-156. 
［12］彭芳青，厉力华，徐伟栋，等. 基于MultiAgent的乳腺钼靶图像肿块分类方法［J］. 传感技术学报, 2010, 23(2): 153-157.
［13］赵浣萍，徐伟栋，厉力华，等. 一种基于改进型MultiAgent多分类器融合的乳腺钼靶肿块分类算法［J］. 仪器仪表学报, 2011, 32(9): 2034-2040.
［14］Liu Tian, Jia Cangzhi. A highaccuracy protein structural class prediction algorithm using predicted secondary structural information［J］. Journal of Theoretical Biology, 2010, 267(3): 272-275.
［15］武勃，黄畅，艾海舟，等. 基于连续Adaboost算法的多视角人脸检测［J］. 计算机研究与发展, 2005, 42(9): 1612-1621.
［16］江林升，朱学芳. 一种基于新特征的车牌检测方法［J］. 计算机工程与应用, 2011, 47(20): 188-190.
［17］寇忠宝，张长水. 基于MultiAgent 的分类器融合［J］. 计算机学报, 2003, 26(2): 174-179.
［18］Kurgan LA, Homaeian L. Prediction of structural classes for protein sequences and domains—Impact of prediction algorithms，sequence representation and homology, and test procedures on accuracy［J］. Pattern Recognition, 2006,39(12): 2323-2343.
［19］Chen Ke, Kurgan LA, Ruan Jishou. Prediction of protein structural class using novel evolutionary collocationbased sequence representation［J］. Journal of computional chemistry, 2008, 29(10):1596-1604.
［20］Zhou Guoping. An intriguing controversy over protein structural class prediction［J］. Journal of Protein Chemistry, 1998, 17(8): 729-738.
［21］Vapnik VN. An overview of statistical learning theory［J］. IEEE Transactions on Neural Networks, 1999, 10(5): 988-999.
［12］Sun XD, Huang RB. Prediction of protein structural classes using support vector machines［J］. Amino Acids, 2006, 30(4): 469-475.
［23］Cai YD, Liu XJ, Xu X, et al. Support vector machines for predicting protein structural class［J］. BMC Bioinformatics, 2001, 2(1): 1-5. 
［24］Cao Youfang, Liu Shi, Zhang Lida, et al. Prediction of protein structural class with rough sets［J］. BMC bioinformatics. BMC Bioinformatics, 2006, 7(1): 7-20.
［25］Yang Jianyi, Peng Zhenling, Chen Xin. Prediction of protein structural classes for lowhomology sequences based on predicted secondary structure［J］. BMC Bioinformatics, 2010, 11(Suppl 1):S9.