Skip to Main Content
Clustering gene expression data is an important task in bioinformatics research and biomedical applications. In this paper, we present an effective clustering algorithm for gene expression data. The clustering algorithm is based on the analysis of data's density distribution. We propose an intersecting partition of gene expression data into the supports of data points. Density clusters are maximally connected regions at certain density levels, and thus, can be organized in a hierarchical structure. For interactive visual exploration, we use a 2D radial layout of the hierarchical density cluster tree with linked as well as embedded views of parallel coordinates and heat maps. Our system supports the understanding of the distribution of density clusters and the patterns of the density clusters. Experimental results for common gene expression data sets shows the effectiveness and scalability of the algorithm.