Robust regression and posterior predictive simulation increase power to detect early bursts of trait evolution

Area of Science:

• Evolutionary biology
• Phylogenetics
• Macroevolution

Background:

• The "Early Burst" model predicts rapid trait evolution early in a clade's history, followed by a slowdown.
• Evidence for Early Bursts in phylogenetic comparative data is often equivocal, potentially due to methodological limitations.

Purpose of the Study:

• To improve methods for detecting "Early Burst" patterns in phylogenetic comparative data.
• To address limitations of existing methods when burst strength is low or morphological convergence is present.

Main Methods:

• Developed posterior predictive simulation approaches to enhance detection power.
• Utilized robust regression to identify and down-weight convergent taxa.
• Applied these methods to body size evolution in cetaceans.

Main Results:

• Posterior predictive simulations outperformed maximum likelihood for detecting moderate-strength early bursts.
• Robust regression provided moderate improvements in method performance.
• Analysis of cetacean body size evolution favored the Early Burst model over Brownian motion or rate shift models.

Conclusions:

• Modified comparative methods offer increased power to detect "Early Burst" evolutionary patterns.
• Posterior predictive simulation and robust regression are valuable tools for macroevolutionary model assessment.
• Further research is needed to determine the prevalence of "Early Burst" evolution in nature.