BEAST, Bayesian Evolutionary Analysis Sampling Trees, is a cross-platform program for Bayesian analysis of molecular sequences using MCMC. The program is orientated towards (strict and relaxed) molecular clock analyses. It can be used as a method of constructing phylogenies, but it is also intended for testing evolutionary hypotheses without conditioning on a single tree topology. BEAST uses MCMC to average over tree space, so that each tree is weighted proportional to its posterior probability. It uses an XML input format that allows the user to design and run a large range of models. We also include a graphical application that for generating this format under a wide variety of models.

BEAST documentation, tutorials, and help is available at documentation site.

Information about downloading and installing BEAST.

Compiled binaries (for Mac, Windows and Linux) are available from the BEAST GitHub repository.

Source code is available from GitHub.

Version History

2018-08-01 - v1.10.1

BEAST is no longer distributed from this website
2016-06-17 - v1.8.4

Bug fixes and new features
New features:
  • New structured list of citations printed to screen before running.
  • Option ('-citation_file') to write citation list to file.
  • Option in BEAUti Priors panel to set parameters to 'Fixed Value'.
Bugs fixed:
  • Issue 808: Set autoOptimize to false in the randomWalkOperator on Pagel's lambda
  • Issue 806: SRD06 in BEAUTi selecting incorrect options.
  • Issue 799: Relative rate parameters for partitions were not being created. All partitions within a clock model have a relative rate if their substitution models are unlinked.
  • Issue 798: Calculating pairwise distances was slow for big data sets - removed this (but initial values no longer suggested based on data).
  • Issue 797: Removed 'meanRate' from Priors tab in BEAUti.
  • Issue 794: Running with empty command line causes error.
  • Issue 792: Check to see that the same likelihood isn't included multiple times into the density.
2016-02-13 - v1.8.3

Bug fixes and new features
New features:
  • Generalized Stepping Stone sampling for Marginal Likelihood Estimation.
  • Continuous quantile version of uncorrelated relaxed clock as option in BEAUti [from Li & Drummond 2012, MBE].
  • Option in BEAUti to log complete histories in Markov Jump Counting.
  • Jukes-Cantor added as an option for BEAUti nucleotide substition model.
  • New tree transition kernel (SubTreeLeap operator) implementd in BEAST.
  • Defaulting to randomizing rate categories for DiscretizedBranchRates.
  • BEAUti 'guess dates' options are now persistent from run to run (and shared with Path-O-Gen/Tempest).
  • Transmission tree model of Hall et al (2016) implemented.
  • Fast, general multidimensional scaling (MDS) implemented.
  • Clock panel in BEAUti simplified.
  • Parameter linking available in Priors panel in BEAUti.
  • Command line is logged into log file headers for reference.
Bugs fixed:
  • Issue 779: Ancestral sequence reconstruction broken with elements.
  • Issue 772: BEAUTi should include BranchRatesModels in prior density.
  • Issue 768: BEAUti creates a subtree slide operator with size zero when data is invariable.
  • Issue 758: Full evaluation error.
  • Issue 304: Label individual transition rate dimensions for CTM models in log files.
2015-03-05 - v1.8.2

Bug fixes
Bugs fixed:
  • Issue 747: Robust counting with multiple partitions generates unique reference IDs in BEAUti and unique total counts for Tracer.
  • Issue 756: FEE with random local clock + codon position partitions + multiple threads.
  • Issue 754: BEAST v1.8.1 is not terminating leaving 'zombie' processes
  • Issue 749: Multiple partitions with unlinked substitution models can turn on ancestral state reconstruction.
  • Issue 747: Duplicate references in robust counting with multiple partitions.
  • Issue 745: Removing an HPM from some parameters doesn't remove the model in the XML
2014-09-20 - v1.8.1

Bug fixes and new features
New features:
  • Local clock (fixed) model as option in 'Clocks' panel. Creates local rate for each taxon set with enforced monophyly.
  • Multilocus models with discrete traits can be generated by BEAUti.
  • Selection filters in Taxon Set panel allows quick selection by name.
  • Improved robust counting under null distribution for measuring selection.
  • Can link parameters in BEAUti to sample together.
  • TreeAnnotator can specify burnin as number of states or trees.
Bugs fixed:
  • Issue 747: Robust counting with multiple partitions generates unique reference IDs in BEAUti and unique total counts for Tracer.
  • Issue 743: Partitions in BEAUti should have the same number of taxa as the tree they are using.
  • Issue 741: Fix negative branch length issues when sampling taxa ages from precision, or set different (default) operator.
  • Issue 740: Add jitter option to continuous diffusion options in the BEAUti sites panel.
  • Issue 734: BEAUti duplicates leafHeight parameters in multiple partition XMLs.
  • Issue 729: Setting initial values for infinite uniform improper priors in BEAUTi.
  • Issue 728: Having taxon sets defined without heights when using a user starting tree causes parsing issues.
  • Issue 725: BEAUti duplicates parameter id of Markov jumps for multiple traits.
  • Issue 722: Complex charsets in nexus files don't parse.
  • Issue 714: sets block not recognized in 1.8.
  • Issue 709: MCMC tab in Beautie v1.8.0 problem with multiple unlinked loci.
Andrew Rambaut, 2007