davidcain / upgma

Efficient UPGMA implementation

UPGMA clustering

This is an efficient implementation of a hierarchical clustering method, UPGMA. UPGMA is used (most commonly) to create phylogenetic trees for use in biological research.

This implementation allows creation of a tree with any arbitrary distance function and member objects. While Python is generally not ideal for a highly optimized data structure, the flexibility allowed by Python and the gains to time complexity motivated me to implement UPGMA clustering myself.

This method has superior time complexity to other implementations (see Gronau I, Moran S (2007) Optimal implementations of UPGMA and other common clustering algorithms, from which this implementation was adapted). Note that one proposed optimization - storing distances in a heap - was not implemented.