Notes on the Statistical Power of the Binary State Speciation and Extinction (BiSSE) Model

Area of Science:

• Evolutionary Biology
• Phylogenetics
• Macroevolution

Background:

• The Binary State Speciation and Extinction (BiSSE) method is widely used for studying diversification and character evolution.
• Previous research suggested BiSSE requires large phylogenetic trees (>300 taxa) for reliable model comparisons.

Purpose of the Study:

• To reevaluate the power of the BiSSE method using simulations with varying phylogenetic tree sizes (30-300 taxa).
• To investigate the impact of root state assumptions on BiSSE's power to detect rate asymmetries.

Main Methods:

• Simulated phylogenetic trees with small and large sample sizes.
• Applied various asymmetry models and root state assumptions.
• Assessed the power of the BiSSE method to detect differences in speciation, extinction, and transition rates.

Main Results:

• BiSSE demonstrates higher power to detect speciation rate asymmetry than previously estimated, even with small sample sizes.
• The discrepancy in earlier studies was attributed to biased assumptions about the root state of the tree.
• Earlier simulations overestimated Type-II statistical errors for speciation rates but not for extinction or transition rates.

Conclusions:

• The BiSSE method's power is not inherently limited by small sample sizes.
• Correctly accounting for root state assumptions is crucial for accurate BiSSE analyses.
• The findings necessitate a revision of previous conclusions regarding BiSSE's sample size requirements.