Detecting the Dependence of Diversification on Multiple Traits from Phylogenetic Trees and Trait Data

Area of Science:

• Evolutionary Biology
• Computational Biology
• Phylogenetics

Background:

• Species diversification is influenced by various factors, including lineage traits.
• State-dependent speciation and extinction (SSE) models analyze trait-dependent diversification.
• Existing methods like MuSSE and HiSSE have limitations, including false positives and restricted trait state handling.

Purpose of the Study:

• Introduce SecSSE, an R package for inferring state-dependent diversification across multiple observed and hidden traits.
• Address limitations of previous SSE models, such as handling multi-state traits and simultaneous trait actions.
• Improve accuracy and reduce false positives in diversification analyses.

Main Methods:

• Developed the SecSSE R package integrating features of HiSSE and MuSSE.
• Implemented functionality for simultaneous inference across multiple observed traits (including multi-state traits) and a hidden trait.
• Corrected likelihood calculations for non-extinction conditioning, an issue in prior SSE models.

Main Results:

• Applied SecSSE to seven previous studies, revealing that conclusions from MuSSE were premature in five cases.
• Simulations demonstrated that SecSSE maintains statistical power while avoiding the high Type I error rate of MuSSE.
• SecSSE successfully detects trait-dependent diversification in most simulated scenarios.

Conclusions:

• SecSSE offers a more robust framework for analyzing trait-dependent diversification, especially with complex trait data.
• The package enhances the reliability of diversification studies by mitigating false positives and improving accuracy.
• SecSSE provides a valuable tool for evolutionary biologists studying the drivers of species diversification.