Clustering of Gene Expression Data: Performance and Similarity Analysis

Recent advances of the DNA microarray technology allow monitoring gene expression profiles of thousands of genes simultaneously. However, the analysis and handling of such fast growing data is becoming the major bottleneck in the utilization of the technology. Clustering analysis is one of the most effective methods for analyzing such gene expression data. In this paper we first experimentally study three major clustering algorithms: hierarchical clustering, self-organizing map (SOM), and self organizing tree algorithm (SOTA), using yeast saccharomyces cerevisiae gene expression data, and compare their performance. Then, we present a data mining tool, cluster diff, which allows the similarity analysis of clusters generated by different algorithms. A case study is conducted based on clusters generated by SOTA and SOM