Do model-based phylogenetic analyses perform better than parsimony? A test with empirical data.

Area of Science:

• Evolutionary biology
• Phylogenetics
• Bioinformatics

Background:

• Model-based phylogenetic inference methods are often preferred over parsimony (MP) due to concerns about potential inaccuracies.
• The frequency of topological incongruence among different phylogenetic inference methods using empirical data remains largely unknown.

Discussion:

• A study analyzed 1000 recent articles to assess congruence between phylogenetic hypotheses generated by various methods.
• Out of 504 articles using multiple methods, only two showed significant incongruence, challenging the perceived need for complex approaches.
• This suggests that parsimony does not commonly produce erroneous topologies due to long-branch attraction in real-world datasets.

Key Insights:

• Phylogenetic tree inference methods, including parsimony, maximum likelihood, and Bayesian inference, show high congruence on empirical DNA sequence data.
• The perceived necessity of sophisticated model-based phylogenetic methods may be overstated, as simpler methods like parsimony appear robust.
• The study advocates for analytical approaches that are less encumbered by complex assumptions.

Outlook:

• Future research could explore specific conditions where incongruence might arise, if any.
• Further investigation into the performance of different phylogenetic methods across diverse datasets is warranted.
• Simplifying phylogenetic analyses could enhance the clarity and explanatory power of evolutionary evidence.