北京大学学报(自然科学版)

利用PLS-VIP方法筛选差异表达基因

韩放1,吴晶辰1,徐江峰1,邓明华1,2   

  1. 1.北京大学数学科学学院,北京100871;2.北京大学理论生物学中心,北京100871;,E-mail:dengmh@math.pku.edu.cn
  • 收稿日期:2007-08-06 出版日期:2009-01-20 发布日期:2009-01-20

Searching for Differentially Expressed Genes by PLS-VIP Method

HAN Fang1, WU Jingchen1, XU Jiangfeng1, DENG Minghua1,2   

  1. 1. School of Mathematical Science, Peking University, Beijing 100871; 2. Center for Theoretical Biology, Peking University, Beijing 100871; , E-mail: dengmh@math.pku.edu.cn?
  • Received:2007-08-06 Online:2009-01-20 Published:2009-01-20

摘要: 提出一种基于变量权重寻找差异表达基因的新方法。该方法的最终目的是从微阵列数据中抽取出核心变量(基因)。将该种方法抽取出的差异表达基因判别样本的能力和普通的PLS方法以及判别最小二乘方法进行比较,结果表明该方法的错误率明显低于其他两种传统方法。因此,PLS-VIP方法是一种较为合适的抽取差异表达基因并判别样本的方法。

关键词: 微阵列, 降维, PLS-VIP, 差异表达基因

Abstract: A new approach called PLS-VIP based on variables' importance of projection (VIP) for detecting differentially expressed genes is proposed. It is a new method to sum up the contribution of components based on Partial Least Square method (PLS). As it takes the correlation of different genes into consideration, it is more suitable than those classification methods based on reviewing distinct genes separately. The effect of genes extracted by ordinary PLS, discriminant PLS and proposed method to classify multi-class tumors is compared. Results show that the error rate of the new method is obviously lower than the other two methods in most cases.

Key words: reduction of dimension, PLS-VIP, microarray, differentially expressed genes

中图分类号: