从时间序列基因芯片数据中挖掘跨事务关联规则  

Mining inter-transaction association rules from time series microarray data

作　　者：彭斌[1] 李辉智[2] 易东[1] 

机构地区：[1]第三军医大学军事预防医学院卫生统计学教研室 [2]西南政法大学刑侦学院,重庆400031

出　　处：《第三军医大学学报》2008年第5期396-398,共3页Journal of Third Military Medical University

基　　金：国家自然科学基金(60371034)~~

摘　　要：目的探讨从时间序列基因芯片数据中发掘基因之间的作用关系。方法采用跨事务关联规则挖掘技术及Gene Ontology标注库对时间序列基因芯片数据进行分析处理。结果根据6个不同发育时间段的基因芯片表达数据,采用2倍法从10080个探针中筛选出119个差异表达基因参与规则挖掘,产生的规则达1300多条,筛选保留J-Measure值最大的10条关联规则并构建基因关联网络图。结论跨事务关联规则能够描述时序基因之间的作用关系,根据关联规则可以对基因的表达状态进行预测。Objective To deduce the interactions between genes from time series microarray data. Methods We used inter-transaction association rules mining technique and GO （Gene Ontology） annotation to analyze the microarray data. Results Using 2-fold-change method, 119 differential expression genes were identified from total 10 080 genes or ESTs, whose expression levels varied significantly on 6 periods of fetus cerebellar development. As a result, about 1 300 inter-transaction association rules were extracted and 10 top rules were kept for their maximum J-measure values. A genes association network graph was deduced based on the 10 top rules. Conclusion Inter-transaction association rules are able to deduce the interactions between genes from time series microarray data and the gene expression status can be predicted based on the association rules.

关 键 词：跨事务关联规则 时间序列基因芯片数据 数据挖掘 

分 类 号：R195.1[医药卫生—卫生统计学] R318.04[医药卫生—卫生事业管理]