Divergence Time Estimation (Moore)

There will be a 1.5 hour lecture to provide an overview of various methods for inferring divergence times, followed by a 1.5 hour tutorial for estimating divergence times using BEAST.
Please download the required software and example data set prior to the tutorial.

Required software
BEAST v.1.4.8
Tracer v.1.4.1
FigTree v.1.2.1
TextWrangler v.2.3
TreeEdit.zip (This archive contains the TreeEdit application, for use with the Mac OS only.)

Example data set
DTE Input Files.zip (This archive contains all of the required input files for the tutorial.)

Suggested readings
Drummond_et_al_06.pdf (Description of the relaxed-clock models implemented in BEAST)
Ho_07.pdf (Description of prior probability distributions for calibrating divergence times)

Presentation
DTE_lecture.pdf (PDF of the lecture)

The following tutorial describes a series of steps required to estimate divergence times from an alignment of nucleotide sequence data using BEAST. This tutorial uses the example data set (provided above) for the sycamore genus, Platanus, and is based on the MAC OS version of the BEAST v.1.4.8 bundle, but other versions are apt to be similar. We assume that the sequences were sampled from species rather than from individuals within a population, although the latter inference scenario is quite similar. Moreover, we assume that a fossil representative of this group is available, which we will use to calibrate divergence times. Finally, we assume that these analyses are imbedded in a more comprehensive analysis pipeline, which we briefly outline in step one of the tutorial.

1. Overview of the Analysis Pipeline
What preliminary analyses are required prior to estimating divergence times.
More...

2. Formatting XML Input File with BEAUti
How to convert a NEXUS formatted data file to the XML format using BEAUti.
More...

3. Editing XML Input File
How to modify an XML file using a text editor. Specifically, we illustrate how to define a subset of species, monitor the monophyly of this clade, enforce the monophyly of this clade, monitor the age of this clade, and assign a calibration prior to the age of this clade.
More...

4. Specifying a Starting Tree
How to generate and specify an initial tree for the MCMC. Specifically, we illustrate how to create a phylogeny that satisfies the topological and temporal constraints that were defined and imposed in the previous step.
More...

5. Running a BEAST Analysis
How to initiate the MCMC, and some tips about common error messages.
More... (In development, March 19, 2009)

6. Diagnosing MCMC with Tracer
How to assess the performance of MCMC sampling. Specifically, we describe how to assess convergence of the chain to the stationary distribution, how to assess mixing of the chain over the target distribution, and how to assess the adequacy of sampling intensity from the MCMC run.
More... (In development, March 19, 2009)

7. Combining MCMC Samples with LogCombiner
How to pool parameter samples from independent MCMC analyses.
More... (In development, March 19, 2009)

8. Summarizing MCMC Samples with TreeAnnotator
How to generate a summary chronogram from an MCMC sample of trees.
More... (In development, March 19, 2009)

9. Viewing and Editing a Chronogram with FigTree
How to view the estimated chronogram and modify it for graphical output.
More... (In development, March 19, 2009)

10. Model Comparison/Hypothesis Evaluation using Bayes Factors
How to assess support for alternative models by comparing their corresponding marginal likelihoods.
More... (In development, March 20, 2009)