Selecting informative genes from microarray dataset using fuzzy relational clustering

Abstract

Selecting informative genes from microarray experiments is one of the most important data analysis steps for deciphering biological information imbedded in such experiments. This paper presents a novel approach for selecting informative genes in two steps. First, fuzzy relational clustering is used to cluster co-expressed genes and select genes that express differently in distinct sample conditions. Second, Support Vector Machine Recursive Feature Elimination (SVM-RFE) method is applied to rank genes. The proposed method is tested on cancer datasets for cancer classification. The results show that the proposed feature selection method selects better subset of genes than the original SVM-RFE does and improves the classification accuracy. © 2008 Springer-Verlag.