A new distance measurement for clustering time-course gene expression data

The purpose of this paper is two-fold. First, a new distance measurement is proposed for temporal microarray gene expression data based on the angles of line segments in the curve of each individual gene expression profile. The hierarchical agglomerative clustering methods are used to incorporate this distance definition. Second, the assessment of the quality of clusterings obtained from the methods are provided by the use of the Davies-Bouldin validity index (DBI). We conclude that the DBI may not be an appropriate indicator for the quality assessment of clusters for time-course gene expression data. We provide an alternative DBI based on the normalized Pearson correlation for this purpose.