Beastiary

Beastiary visualizes and analyzes Markov chain Monte Carlo (MCMC) trace files from Bayesian phylogenetic software to summarize outputs and assess model convergence and parameter estimates.

Key Features:

• Real-time inspection of log and trace files: Performs real-time inspection of MCMC log and trace files to monitor ongoing analyses and provide immediate insight into trace behavior.
• Remote server compatibility: Operates on remote servers, including high-performance computing (HPC) environments, to handle computationally intensive Bayesian phylogenetic runs.
• Compatibility with Bayesian phylogenetic software: Supports outputs from BEAST, BEAST2, RevBayes, and MrBayes for direct integration with common Bayesian workflows.
• Summarization and visualization of MCMC outputs: Summarizes and visualizes MCMC outputs to present parameter estimates and trace dynamics in compact visual formats.
• Assessment of convergence and parameter estimation: Provides computational summaries used to assess model convergence and the accuracy/stability of parameter estimates.

Scientific Applications:

• Validation of Bayesian phylogenetic inference: Evaluates convergence and parameter stability to validate inferences from Bayesian phylogenetic analyses.
• Assessment of phylogenetic tree reliability: Assesses the reliability of phylogenetic trees and evolutionary hypotheses derived from Bayesian analyses by examining MCMC behavior.
• Monitoring long-running MCMC analyses on HPC: Monitors long-running MCMC analyses on remote servers or HPC to detect convergence issues and parameter instability during runtime.

Methodology:

Performs real-time monitoring and remote analysis of MCMC trace and log files and summarizes those outputs into visual formats to identify lack of convergence or parameter instability.