生物大数据中的聚类方法分析.pdf

1.海丈军￥tl (自然科学版) Vol. 22 No. 1 第 22 卷第1 期 2016年2月 JOURNAL OF SHANGHAI UNIVERSITY (NATURAL SCIENCE) Feb.2016 DOI; 10.3969/j. issn.1007-286 1.2015.04.018 生物大数据中的聚类方法分析 路东方，许俊富，项超娟，谢江 (上海大学计算机工程与科学学院，上海 200444) 摘要: 随着人类基因组计划的实施和完成，生物实验技术快速发展7 生物数据呈现爆发式增 长并不断积累，生命科学迎来了大数据时代.在后基因组时代，单一的统计模式逐渐被智能化 与综合分析相结合的方式所取代，聚类分析便是核心的数据挖掘方式.描述了生物信息学领域 中的大数据现状，总结基因表达谱分析和生物网络分析中常用的聚类方法，并对小自1胚胎成纤 维细胞的时间序列数据进行实验对比.实验结果表明，不同的聚类方法生成了不同的实验结 果，面临高噪声的生物大数据，选择或结合合适的聚类方法进行综合分析将有助于获得更可靠 的分析结果. 关键词:生物大数据;数据分析;聚类方法 中圄分类号: TP39 文献标志码:A 文章编号: 1007-2861(2016)01-0045-13 Survey of clustering methods for big data in biology LU Dongfang , XU Junfu , XIANG Chaojuan, XIE Jiang (School of Computer Engineering and Science , Shanghai University, Shanghai 200444 , China) Abstract: With the implementation of the Human Genome Project and the rapid devel- opment of biological experiment technology, biological data sharply grow and continuous accumulate. Age of big data in biology is coming. In the post genomic era , single statistical models 缸e gradually replaced with combination of intelligent and comprehensive analyses. Clustering is the core of data mining. This paper describes the stat←oιthe-art technology of big data in bioinformatics , and summarizes several popular clustering methods on gene expression profiling and biological networks. Furthermore , some experiments are made to compare different clustering methods on the time seri臼 data of mouse embryonic fibrob- lasts , showing that different clustering methods have different results. To achieve more reliable conclusions for highly noisy biological data , it is nece