微阵列数据中常用的分类分析方法 - 第二军医大学卫生统计学教研室.PDFVIP

HE Xian Min, Department of Health Statistics, Second Military Medical University, Shanghai, China Post Code: 200433 TEL: (021 Email: hxmine@ 微阵列数据中常用的分类分析方法 贺宪民,贺 佳,吴 骋,范思昌 （第二军医大学卫生统计学教研室, 上海 200433 ） XIANG Zhaoying (Microarray Core Facility, Weill Medical College of Cornell University, NEW YORK, USA) 摘要：微阵列技术的出现改变了生物医学研究的前景。微阵列技术产生的大量数据是限制其发展的一个主要瓶颈， 为提取其中所隐含的有价值的信息，在微阵列数据分析的复杂计算工具和方法方面都有很多尝试。本文对基因表达 模式识别中的分类方法进行了综述。 关键词：基因芯片；微阵列；分类 中图分类号： Classification methods used for mic roarray data analysis HE Xian-min, HE Jia, WU Cheng, FAN Si-chang (Department of health statistics, Second Military Medical University, Shanghai 200433, China) XIANG Zhaoying (Microarray Core Facility, Weill Medical College of Cornell University, NEW YORK, USA) Abstract: The advent of microarray -based technologies has changed the landscape of biomedical research. Microarrays produce huge quantities of data, which poses a major bottleneck. In order to obtain a meaningful knowledge out of this numerical data, abundant efforts have been made to develop sophisticated computational tools and methods for microarray data analysis. In this article, some specific classification methods that have been used to identify patterns of gene expression are reviewed. Key words: gene chip, microarray, classification 2 微阵列数据的产生及数据的标准化 1 前言 生物芯片反应包括四个基本步骤：芯片制备、 随着人类基因组计划的发展，编码人类全部染 样品制备、生物分子反应和芯片信号检测和分析。 色体的约3-4 万条基因被发现，人类基因组计划由此 生物芯片上探针与经荧