Phylogenetic Inference and Visualization: Community Insights
In the rapidly evolving field of viral genomics, selecting the appropriate computational framework for phylogenetic analysis is as critical as the data itself. Following a recent Journal Club presentation, members of the RdRp Summit community engaged in a technical exchange that provided a snapshot of the current software landscape used by experts to reconstruct and visualize evolutionary histories.
This post archives those recommendations, ranging from industry standards to niche libraries for specialized genomic architectures.
1. Phylogenetic Tree Construction
The discussion highlighted a clear distinction between rapid, maximum-likelihood (ML) heuristics for routine surveillance and rigorous Bayesian frameworks for complex evolutionary modeling.
- Routine & Rapid Inference: IQ-TREE 2 remains the community favorite for its speed and robust model selection. For real-time epidemiology, it is frequently paired with TreeTime (Augur) to generate time-resolved phylogenies.
- Deep Evolutionary Relationships: When resolving deep nodes, PhyloBayes was noted for its ability to handle complex site-heterogeneous models.
- Phylodynamics & Reassortment: BEAST 1 and BEAST 2 are the go-to frameworks for phylodynamic inference. Specifically, BEAST 2 was highlighted for its specialized packages like CoalRe, which is essential for analyzing reassortment networks in multi-segmented viruses.
| Tool | Category | Primary Use / Strengths | Link |
|---|---|---|---|
| IQ-TREE 2 | ML Inference | High-performance, automatic model selection | Website |
| PhyML | ML Inference | Classical, well-validated ML approach | Website |
| FastTree | ML Inference | Optimized for extremely large datasets | Website |
| MrBayes | Bayesian | Standard Bayesian MCMC inference | Website |
| PhyloBayes | Bayesian | Specialized for deep-level phylogenetics | Website |
| BEAST / BEAST 2 | Bayesian | Phylodynamics, molecular dating, reassortment | Website |
| TreeTime | Post-processing | ML-based temporal dating and Augur pipelines | GitHub |
2. Visualization and Post-Processing
Visualization remains a matter of both aesthetic preference and functional requirement (e.g., handling non-treelike evolution).
- Aesthetics: baltic (Python) was highly recommended for producing publication-quality figures, though it carries a steeper learning curve.
- Interactivity: For exploring Nextstrain-style JSON trees, Auspice is the standard. For visualizing complex reassortment networks, IcyTree remains a unique and vital web-based tool.
- Programmatic Integration: ggtree (R) and ete3 (Python) continue to be the workhorses for researchers integrating tree plotting into automated pipelines.
| Tool | Type | Key Feature | Link |
|---|---|---|---|
| baltic | Python | Highly customizable, publication-ready plots | GitHub |
| ggtree | R | Seamless integration with the R/Tidyverse ecosystem | Website |
| FigTree | Desktop GUI | Classic, user-friendly tree viewing/editing | Website |
| iTOL | Web-based | Interactive, easy annotation of large trees | Website |
| Auspice | Web-based | Interactive viewer for Nextstrain datasets | Website |
| IcyTree | Web-based | Specialized for reassortment networks | Website |
| ete3 | Python | Comprehensive toolkit for tree manipulation | Website |
| phylotree.hyphy.org | Web-based | Lightweight, interactive JS viewer | Website |
| phyTreeViz | Python | Minimalist plotting library | GitHub |
| TreeSwift | Processing | Ultra-fast tree traversal and manipulation | GitHub |
| DendroPy | Processing | Professional-grade library for phylogenetic data | Website |
| toytree | Python | Interactive plotting in Jupyter notebooks | GitHub |
Join the Community
This summary represents only a fraction of the daily knowledge exchange within our community. If you are a virologist, bioinformatician, or student working with RNA viruses, consider joining our Slack to participate in these discussions in real-time.