A simulation comparison of phylogeny algorithms under equal and unequal evolutionary rates

Area of Science:

• Computational Biology
• Phylogenetics
• Bioinformatics

Background:

• Phylogenetic trees are crucial for understanding evolutionary relationships.
• Accurate estimation of these trees is vital for biological research.
• Various computational methods exist for phylogenetic tree inference.

Purpose of the Study:

• To compare the accuracy and bias of five common phylogenetic tree estimation methods.
• To evaluate method performance under varying evolutionary parameters (substitution rates, sequence length).
• To identify the most robust and reliable phylogenetic inference techniques.

Main Methods:

• Simulated phylogenetic datasets were generated across diverse conditions.
• Five methods were evaluated: parsimony, compatibility, maximum likelihood, Fitch-Margoliash, and neighbor joining.
• Tree accuracy was assessed using distance metrics, with and without branch length sensitivity.

Main Results:

• Maximum likelihood demonstrated superior overall accuracy.
• Distance-matrix methods (Fitch-Margoliash, neighbor joining) performed well when negative branch lengths were constrained.
• Parsimony and compatibility methods showed limitations, especially with rate variation across branches.
• All methods exhibited inaccuracy and bias when evolutionary rates varied across sites.

Conclusions:

• Maximum likelihood is the recommended method for general phylogenetic tree estimation.
• Neighbor joining and Fitch-Margoliash can be effective for shorter sequences.
• Understanding method limitations under rate variation is critical for reliable evolutionary inference.