Practice

• The phangorn library in R contains functions to perform maximum likelihood based analyses of sequence data
  • While not the fastest program, it is very flexible
  • Can fit a wide variety of substitution models
• For large-scale phylogenetic inference, there are specialised (external) software packages available
  • RAxML/ExaML
  • PhyML
  • IQTREE
• As with alignment, we can use R to format the data, and call these external programs
• There is also an interface to RAxML in the ips package

phangorn

• Takes data in a special format, phyDat
  • Sequences read in using ape are in DNAbin format
  • There is a convenience function called as.phyDat that can convert between them
• Can compare between different models using modelTest
• Can fit models (and trees) using pml

RAxML

• RAxML is one of the fastest, most accurate phylogeny reconstruction software packages available
• It does not implement a wide variety of models, instead, it assumes a fairly complex one, which will probably be the best model if the dataset is variable enough

PhyML

• PhyML is slower than RAxML, but offers a wider variety of models
• Also offers an interactive (text) interface

IQTREE

• IQTREE is a fairly new phylogenetic software package
• It does implement a variety of models, and includes model selection and a fast bootstrap-like procedure