应用聚类分析和GO数据库分析间变性星形细胞瘤的微阵列数据

Applying cluster analysis and Gene Ontology^(TM) Consortium(GO) to the cDNA array data of anaplastic astrocytomas.

目的 探讨从巨量的微阵列基因数据中挖掘肿瘤相关分子机理及功能信息 ;方法 以人肿瘤cDNA微阵列1 2分析并获得间变性星形细胞瘤和正常脑组织的差异表达基因数据 ,进行基于CityBlock距离和平均距离法的聚类分析 ,应用超几何分布的概率模型计算聚类分析所得的各基因类与GO数据库注释的各基因功能类之间的随机关联概率 ,给各基因类标以显著关联的GO数据库的功能类标签。结果 从数以千计的基因数据中获得了 12 1个差异显示基因并分成 6类 ,这 6类基因的生物学功能基本与该肿瘤的生物学特点符合。GO数据库中 6个基因通路功能类分别与该分类相关度最大。结论 应用聚类分析方法和GO数据库对微阵列研究获得的基因信息进行进一步分析 ,有利于提取巨量基因数据中的有效信息 ,可能提供进一步研究的有价值线索。

Objective To explore the tumor related molecular mechanism and functional message from the gigantic cDNA array gene data. Methods The data relative to the differential gene expression between the anaplastic astrocytomas and normal brain tissue were acquired through the comparing analysis with Atlas Human Cancer Array 1.2, on which the cluster analysis based on both City Block Distance and Average Distance was performed. The random associated ratios between the different gene clusters and the gene function clusters annotated in Gene OntologyTM Consortium(GO) were calculated by Super Geometry Distribution Probability. The notable associated functional cluster of GO was tagged on the different gene clusters. Results 121 genes were calculated from thousands of genes and were then clustered as 6 kinds, whose biological functions were in coincidence with the clinical biological features of the anaplastic astrocytomas. 6 functional clusters pathways in GO showed the highest associated ratios with these 6 kinds genes respectively. Conclusion It is helpful to extract the effective message from the gigantic gene data when further bioimformatic analysis is performed on the gene information obtained in cDNA array research, from which it is possible to get some valuable clew for further study.