Defining clusters from a hierarchical cluster tree

摘要: Summary: Hierarchical clustering is a widely used method for detecting clusters in genomic data. Clusters are defined by cutting branches off the dendrogram. A common but inflexible uses constant height cutoff value; this exhibits suboptimal performance on complicated dendrograms. We present Dynamic Tree Cut R package that implements novel dynamic branch methods dendrogram depending their shape. Compared to method, our techniques offer following advantages: (1) they capable of identifying nested clusters; (2) flexible—cluster shape parameters can be tuned suit application at hand; (3) suitable automation; and (4) optionally combine advantages hierarchical partitioning around medoids, giving better detection outliers. illustrate use these applying them protein–protein interaction network data simulated gene expression set. Availability: The implemented an available http://www.genetics.ucla.edu/labs/horvath/CoexpressionNetwork/BranchCutting Contact: stevitihit@yahoo.com Supplementary information: Supplementary Bioinformatics online.