Practice

Introduction

There are two widely used approaches to getting a time-stamped phylogeny or 'time tree'

• 'Clockify' a maximum likelihood tree
  • Rate smoothing
  • Least squares dating
• Fit a model that explicitly considers a strict or molecular clock

Rate smoothing with chronos

• Rate smoothing involves constraining a tree to have clock-like branch lengths, but allowing the evolutionary rate to vary in a possibly complex manner
• Given a fixed, rooted phylogeny, and a set of temporal constraints on sampling time, the function chronos will return a time tree
  • We will use a slightly modified version of the code

Rooting the tree

• chronos needs a rooted tree
• We can estimate the root of the tree by maximising the association between the root-to-tip distance and the sampling times
• The function rtt in the library ape allows one to do this

Least squares dating

• Least squares dating, implemented in a program LSD also takes a fixed tree, and also estimates the root
• We can use R to generate the data files and run lsd

Bayesian approaches

• We can also fit phylogenies with time constraints using the popular Bayesian phylogenetic programs BEAST (version 1 or 2) and MrBayes
• R can help to generate the input files for these programs