Overview of Phylogenetic Methods

Common elements of most methods

Trees as a representation of evolutionary history

Dichotomously branching trees, perhaps with zero-length branches leading to polytomies

In mathematical terms, an acyclic graph with interior vertices of degree three (or higher)

Characters, character states, character-state changes

Branch length

Taxa (also called OTUs), in our case often represented by biological sequences such as DNA

Parsimony

Parsimony methods favor the tree that requires the smallest number of character-state changes

Parsimony as an optimality criterion

In standard parsimony, branch length is not important per se; it is the overall length of the tree that is optimized.

Tree searching (branch swapping) to find the best tree

Can be prone to long-branch attraction if there are superimposed substitutions

Distance (Clustering)

Distance methods cluster taxa according to a distance measure

Distances can be calculated (or measured) in many ways

Some distance measures take into account superimposed substitutions

Branch lengths are intrinsically meaningful (as is the "diameter" of the tree

There are many different distance methods depending upon how distances are calculated and on how tree is found

Likelihood

Uses a model of character-state substitution to relate data to hypothesis

Particularly valuable for DNA, RNA, and Amino-Acid data

Models of point mutation have been described

There are also well-established models of site-to-site rate variation

Codon-specific models

Tree topology includes both the branching order and branch lengths

Likelihood methods can also be used to test other hypotheses

Maximum Likelihood

Uses tree-search methods similar to those used in Parsimony analysis

If hypothesis is tree topology, best tree is the tree with the highest likelihood

This is the tree that would be most likely to give rise to the data given the model

Bayesian

Uses a Metropolis-Hastings Coupled Markov Chain Monte Carlo (MCMCMC) algorithm to estimate the probability of different trees given the data

Confidence assessment

Baldauf SL (2003) The deep roots of eukaryotes. Science 300:1703-1706

Stewart CB (1993) The Powers and Pitfalls of Parsimony. Nature 361:603-607