A Graph-Based Approach for Clustering Analysis of Gene Expression Data by Using Topological Features

This paper proposed a graph-based clustering approach for gene expression data. The new method is based on regulatory network graph obtained from gene expression data. Clustering is performed based on the topological features of the graph which characterizes the regulatory relationships between genes, which is different from the conventional methods that simply group genes with similar gene expression patterns. The performance of the proposed method is assessed by real gene expression data clustering. The results clearly show that the proposed method can give higher accuracies in clustering recognition than the traditional approaches which are based on similarity between gene expression patterns.