Tutorials

From BEAST Software

1 Tutorials in running BEAST for various types of analyses
2 Self-contained practicals to download
3 Introductory tutorials by example
4 Advanced tutorials
5 Model testing and model selection
6 Demographic model tutorials
7 Bayesian skyline plot tutorials
8 Extended Bayesian skyline plot (EBSP) tutorial and script
9 Trait analysis tutorials
- 9.1 Continuous morphological characters
10 Phylogeography tutorials
- 10.1 Discrete phylogeographic diffusion models
- 10.2 Continuous phylogeographic diffusion models
11 Hierarchical phylogenetic model (HPM) tutorials
12 *BEAST tutorial and example

Tutorials in running BEAST for various types of analyses

This document gives a brief guide to the practicalities of performing an MCMC analysis with the BEAST software package, by going through a number of examples in detail.

Self-contained practicals to download

These archives contain complete practicals to using BEAST, Tracer and FigTree complete with the required data sets.

Continuous_Phylogeography_1.7.zip (BEAST v1.7.x): Practical on continuous diffusion phylogeography of rabies virus in racoons

Divergence Dating (Primates) v1.2a.zip (BEAST v1.6.x): Estimating a date of divergence using fossil calibration for primates.

Virus_Practical_1_6_1.zip (BEAST v1.6.x): A three-part virus practical that covers the same material as the two above but has an additional part on Bayesian Skyline Plots.

Influenza BEAST Practical.zip (BEAST v1.6.x): A two-part practical focused on estimating rates of human influenza virus A using dated sequences and constructing Bayesian skyline plots of fluctuating populations (A/H1N1pdm and A/H3N2).

Divergence Dating (Papillomaviruses) v1.0.zip: Estimating a date of divergence using a host co-divergence for feline papillomaviruses (similar to the primate practical but virus-orientated).

Estimating Rates in Viruses (RSVA) v1.0.zip: Estimating the rate of evolution from serially-sampled sequences (dated tips) using an RSVA (human respiratory syncytial virus subgroup A) data set.

Introductory tutorials by example

Using sequences sampled at different points in time to estimate rates in dengue virus: Describes the use of the BEAUti GUI application to create a BEAST XML file for analysing data where sequences have known dates of sampling.

Analysing BEAST output using Tracer: Describes the use of the Tracer GUI application to analyse the output of BEAST.

Summarizing BEAST trees using TreeAnnotator and FigTree: Describes the use of the TreeAnnotator utility and FigTree to summarize the trees produced by BEAST.; This tutorial has yet to be written.

Estimating the divergence time of a monophyletic group using a known mutation rate: This tutorial describes the use of BEAUti and BEAST to analyse some primate sequences and estimate the date of divergence of humans and chimps when the mutation rate is known.

How to increase the ESS of a parameter: A short list of techniques for increasing the effective sample size (ESS) for a BEAST analysis.

Advanced tutorials

How to use statistics: Describes how to create and log various statistics in order to estimate posterior distributions of particular hypotheses of interest.

Placing an upper limit on the root height: Describes using the BEAST XML format to place an upper limit on the age of the root of the tree.

Calibrating the date of an MRCA using a bounded range or parametric probability distribution: Describes how to specify a uniform bounded or parametrically distributed prior on the age of the MRCA of a subset of taxa.

Providing a user-specified starting tree: Describes editing the BEAST XML input file to provide an initial user-specified tree for the MCMC to start with.

Starting with a UPGMA tree: Describes editing the BEAST XML input file to create a UPGMA tree for the MCMC to start with.

Constraining a group of taxa to be monophyletic: Describes editing the BEAST XML input file to constrain a particular group of taxa to be kept monophyletic throughout the analysis. This can be used to keep an outgroup as the outgroup.

Creating partitions of sites with different substitution parameters: Describes editing the BEAST XML to add an additional data partition with a different substitution model and rate.

Imposing a prior distribution on a parameter: Describes how to place a probability distribution as a prior on the age of the root of the tree or the substitution rate.

Analyzing multi-locus data: Describes how to incorporate multiple unlinked loci into a single analysis.

Creating your own general data type: Describes how to create a general data type for a custom type of discrete characters. (Updated on 22/01/2010)

Using a nucleotide substitution model other than HKY and GTR: Describes how to set up time-reversible nucleotide models other than HKY and GTR.

Running BEAST without data to sample from the Prior: Describes how to alter a BEAST XML file so that BEAST only samples from the Prior distribution.

Reconstructing ancestral states/sequences: Describes how to reconstruct an ancestral state or sequence at a particular node

Model testing and model selection

Model selection in BEAST: Describes how to perform model selection in BEAST using the harmonic mean estimator (HME), a posterior simulation-based analogue of Akaike's information criteration (AICM), path sampling (PS) and stepping-stone sampling (SS). This tutorial replaces the previous model comparison tutorial.

Demographic model tutorials

Setting up a two-epoch model: Describes how to set up a demographic model involving two epochs and a transition time.

Bayesian skyline plot tutorials

Creating an upper limit on population size
BSP.pdf (BEAST v1.6.x): Describes how to set up an upper limit on population size in a Bayesian Skyline Plot.

Extended Bayesian skyline plot (EBSP) tutorial and script

EBSP.zip (BEAST 1.6.x): This short practical explains how to set up an Extended Bayesian Skyline Plot (EBSP) analysis in BEAST, and how to generate some EBSP plots.

Trait analysis tutorials

Continuous morphological characters

Continuous traits and the comparative method: This tutorial describes how to setup an analysis of continuous traits to look at the correlation between them in a phylogenetically corrected way (i.e., the comparative method). This examples uses some mitochondrial sequence data to sample trees whilst jointly estimating the coevolution of 5 continuous morphological traits.

Phylogeography tutorials

Discrete phylogeographic diffusion models

These tutorials describe how to set up the analyses described in Lemey, Rambaut, Drummond & Suchard (2009) PLoS Comput Biol.

Setting up a standard discrete phylogeographic analysis: Describes how to edit a BEAST XML file to set up a discrete phylogeographic analysis using a standard continuous-time Markov chain.

A Bayesian stochastic search variable selection (BSSVS) extension of the discrete phylogeographic model: Describes how to modify a discrete phylogeographic model XML file to set up BSSVS.

Summarizing a discrete phylogeographic inference using an MCC tree and visualization in FigTree: Describes how to annotate an MCC tree using the modal location states.

Visualizing a location-annotated MCC tree in Google Earth: Describes how to convert a location-annotated MCC to KML for visualization in Google Earth.

Performing a Bayes factor test to identify well-supported rates: Describes how to identify rates that are frequently invoked to explain the diffusion process and how to visualize these in Google Earth.

Continuous phylogeographic diffusion models

These tutorials describe how to set up the analyses described in Lemey, Rambaut, Welch & Suchard (2010) MBE.

Setting up a standard continuous phylogeographic analysis: Describes how to edit a BEAST XML file to set up a phylogeographic analysis in continuous space using a standard random walk.

Setting up a continuous phylogeographic analysis using relaxed random walks: Describes how to modify an XML file that specifies a homogenous Brownian diffusion phylogeographic model to set up a relaxed random walk.

Hierarchical phylogenetic model (HPM) tutorials

These tutorials (coming shortly) will describe how to set up HPMs in BEAST as described in Suchard, Kitchen, Sinsheimer & Weiss (2003) Syst Biol

Setting up a hierarchical phylogenetic model in BEAST: Describes how to edit a BEAST XML file to set up a hierarchical phylogenetic model across several evolutionary parameters.

*BEAST tutorial and example

STARBEAST.pdf (BEAST 1.7.x): The objective of this tutorial is to estimate the species tree that is most probably given the multi-individual multi-locus sequence data. The species tree has 9 taxa, whereas each gene tree has 26 taxa. *BEAST will co-estimate three gene trees embedded in a shared species tree (see Heled and Drummond, 2010 for details).

Return to contents