Protein functional class prediction using global encoding of amino acid sequence

作     者：Li, Xi Liao, Bo Shu, Yu Zeng, Qingguang Luo, Jiawei 

作者机构：Hunan Univ Sch Comp & Commun Changsha 410082 Hunan Peoples R China Sch Langshan Shaoyang 422000 Hunan Peoples R China 

出 版 物：《JOURNAL OF THEORETICAL BIOLOGY》 (理论生物学杂志)

年 卷 期：2009年第261卷第2期

页      面：290-293页

核心收录：

学科分类：0710[理学-生物学] 07[理学] 09[农学] 

基　　金：National Nature Science Foundation of China [10571019, 60873184] National Nature Science Foundation of Hunan province [07JJ5080, 06JJ2090] 

主　　题：Protein functional class prediction Global encoding Nearest neighbor algorithm Physiochemical property 

摘      要：A key goal of the post-genomic era is to determine protein functions. In this paper, we proposed a global encoding method of protein sequence (GE) to descript global information of amino acid sequence, and then assign protein functional class using machine learning methods nearest neighbor algorithm (NNA). We predicted the function of 1818 Saccharomyces cerevisiae proteins which was used in Vazquez s global optimization method (GOM) except eight proteins which cannot get from the data base now or whose sequence length is too short. Using our approach, the computed accuracy is better than Vazquez s global optimization method (GOM) in some cases. The experiment results show that our new method is efficient to predict functional class of unknown proteins. (C) 2009 Elsevier Ltd. All rights reserved.