Hierarchical clustering
Hierarchical (or agglomerative) clustering methods are a class of bottom-up approaches involving the gradual pairing of samples into larger groups, to form a hierarchy. It is generally desirable to partition a large set of experimental observations into smaller subsets of similarly-behaving samples (variables). The upper limit for the number of clusters is actually the number of samples, this can be influenced by the selection of an appropriate cut off value. The hierarchical cluster tree is also known as a dendrogram. In phylogenetics, evolutionary trees are a form of clustering in which the branch (edge) lengths may carry evolutionary meaning, such as the genetic distance of a sample with respect to others in the same tree.
Applications
- Xie and co-workers used hierarchical clustering to group tumors based on genome-wide copy number data ref.
- 867 prokaryotic genomes were compared on the basis of the pairwise correlations of their genomic hexanucleotide signatures ref.
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.