RAxML-NG: a fast, scalable and user-friendly tool for maximum likelihood phylogenetic inference

Area of Science:

• Computational biology
• Bioinformatics
• Evolutionary biology

Background:

• Phylogenetic trees are crucial for biological research, agriculture, biotechnology, and medicine.
• Determining the optimal phylogenetic tree using maximum likelihood (ML) is computationally challenging (NP-hard).
• There is a need for efficient and scalable software to analyze increasingly large biological datasets.

Purpose of the Study:

• To introduce RAxML-NG, a re-implementation of the RAxML/ExaML greedy tree search algorithm.
• To enhance accuracy, flexibility, speed, scalability, and usability in phylogenetic inference.
• To provide new features for tree space exploration and statistical support.

Main Methods:

• A from-scratch re-implementation of the greedy tree search algorithm used in RAxML/ExaML.
• Development of RAxML-NG with a focus on performance and scalability for large datasets.
• Incorporation of novel features like terrace detection and transfer bootstrap support.

Main Results:

• RAxML-NG demonstrates improved accuracy, speed, and scalability compared to its predecessors.
• For datasets with many taxa, RAxML-NG often identifies higher-scoring trees than IQTree.
• New features, including terrace detection and transfer bootstrap support, are introduced.

Conclusions:

• RAxML-NG provides a powerful and efficient tool for maximum likelihood-based phylogenetic inference.
• The software is suitable for analyzing large and complex biological datasets.
• RAxML-NG offers enhanced capabilities for exploring phylogenetic tree space and assessing tree reliability.