Bioinformatics & Molecular Design Research Center (BMDRC), Seoul, Korea.
BMC Genomics. 2009 Dec 3;10 Suppl 3(Suppl 3):S15. doi: 10.1186/1471-2164-10-S3-S15.
Gene expression similarity measuring methods were developed and applied to search rapidly growing public microarray databases. However, current expression similarity measuring methods need to be improved to accurately measure similarity between gene expression profiles from different platforms or different experiments.
We devised new gene expression similarity measuring method based on pathway information. In short, newly devised method measure similarity between gene expression profiles after converting them into pathway based expression profiles. To evaluate pathway based gene expression similarity measuring method, we conducted cell type classification test. Pathway based similarity measuring method shows higher classification accuracy. Especially, pathway based methods outperform at most 50% and 10% over conventional gene expression similarity method when search databases are limited to cross-platform profiles and cross-experiment profiles.
The pathway based gene expression similarity measuring method outperforms commonly used similarity measuring methods. Considering the fact that public microarray database is consist of gene expression profiles of various experiments with various type of platform, pathway based gene expression similarity measuring method could be successfully applied for searching large public microarray databases.
基因表达相似性测量方法被开发并应用于快速搜索不断增长的公共微阵列数据库。然而，目前的表达相似性测量方法需要改进，以准确测量来自不同平台或不同实验的基因表达谱之间的相似性。
我们设计了一种新的基于通路信息的基因表达相似性测量方法。简而言之，新设计的方法在将基因表达谱转换为基于通路的表达谱后测量它们之间的相似性。为了评估基于通路的基因表达相似性测量方法，我们进行了细胞类型分类测试。基于通路的相似性测量方法显示出更高的分类准确性。特别是，当搜索数据库仅限于跨平台的谱和跨实验的谱时，基于通路的方法的表现优于传统基因表达相似性方法，最高可达 50%和 10%。
基于通路的基因表达相似性测量方法优于常用的相似性测量方法。考虑到公共微阵列数据库由具有各种类型平台的各种实验的基因表达谱组成，基于通路的基因表达相似性测量方法可以成功地应用于搜索大型公共微阵列数据库。
