BEAST 2

BEAST 2 implements Bayesian evolutionary analysis using Markov chain Monte Carlo to estimate rooted, time-measured phylogenies and support testing of evolutionary hypotheses.

Key Features:

• Bayesian MCMC inference: Uses Markov chain Monte Carlo to sample rooted, time-measured phylogenies and weight tree topologies by posterior probability.
• Molecular clock models: Supports both strict and relaxed molecular clock models for time calibration.
• Tree priors: Implements birth-death-sampling tree priors.
• Phylodynamics: Provides models for phylodynamic analysis and population dynamics inference.
• Model averaging and partitioning: Supports model averaging for substitution models and site partitioning.
• Package management: Includes a package management system that enables third-party modules to extend analytical functionality.
• Checkpointing and state I/O: Can read and write the entire MCMC chain state to/from disk to enable checkpointing and sharing across instances for multi-processor and high-end computing environments.
• XML format and templates: Uses a consistent XML file format with an XML template-based mechanism for adding functionality.

Scientific Applications:

• Time-calibrated phylogeny reconstruction: Estimating rooted, time-measured phylogenies and divergence times.
• Evolutionary hypothesis testing: Testing evolutionary hypotheses within a Bayesian posterior-sampling framework.
• Phylodynamics: Inferring population dynamics and epidemiological processes.
• Birth-death sampling analyses: Modeling sampling and diversification processes using birth-death-sampling tree priors.
• Model selection and partitioning: Comparing and averaging substitution models and site-partitioning schemes.

Methodology:

Performs Markov chain Monte Carlo (MCMC) sampling of rooted, time-measured phylogenies under strict and relaxed molecular clock models, implements birth-death-sampling tree priors, phylodynamic models, and model averaging for substitution models and site partitioning, supports package-based extensions and read/write of the full MCMC state to disk, and uses a consistent XML file format with templates.