A Tale of Too Many Trees: A Conundrum for Phylogenetic Regression

Area of Science:

• Evolutionary Biology
• Phylogenetics
• Comparative Genomics

Background:

• Phylogenetic comparative methods are crucial for testing evolutionary hypotheses.
• Accurate phylogenetic trees are required inputs for these methods.
• Phylogenetic conflict (gene tree vs. species tree) is common and poses challenges.

Purpose of the Study:

• To investigate the consequences of tree-trait mismatch on phylogenetic regression.
• To assess the impact of gene tree-species tree conflict on evolutionary analyses.

Main Methods:

• Conducted simulation experiments with varying tree sizes, trait complexities, and phylogeny estimation methods.
• Analyzed a cross-species gene expression dataset to evaluate tree choice in a real-world scenario.
• Explored the utility of robust estimators as a potential solution.

Main Results:

• Tree-trait mismatch led to excessively high false positive rates across diverse simulation settings.
• Observed directionality of error, where assuming a species tree for gene tree-evolved traits was sometimes worse.
• A robust estimator showed promise in mitigating some tree mismatch issues.

Conclusions:

• Inaccurate phylogenetic tree assumptions significantly compromise the reliability of evolutionary hypothesis testing.
• Careful consideration of tree choice and phylogenetic modeling is essential for robust comparative studies.
• Further research into robust methods is needed to address tree-trait mismatch challenges.