Bayesian Inference of Species Trees using Diffusion Models.

Area of Science:

• Population Genetics
• Computational Biology
• Evolutionary Biology

Background:

• Inferring species trees and demographic history is crucial for understanding evolutionary processes.
• Existing methods can be computationally intensive, limiting the analysis of large datasets.

Purpose of the Study:

• To develop a computationally efficient Bayesian methodology for inferring species trees and demographics.
• To enable the analysis of large population genetics datasets using diffusion models.

Main Methods:

• Utilized diffusion models of allele frequency dynamics for likelihood calculations.
• Developed novel numerical algorithms for efficient computation.
• Implemented the Snapper method within the BEAST2 package.

Main Results:

• Snapper allows for the analysis of datasets with hundreds or thousands of individuals.
• Simulation experiments confirmed the method's accuracy, computational efficiency, and low numerical error.
• Demonstrated the scalability of Snapper with a large SNP dataset from freshwater turtles.

Conclusions:

• Snapper provides a powerful and efficient tool for Bayesian inference of species trees and demographics.
• The diffusion-based approach facilitates large-scale population genetic analyses.
• The method is suitable for analyzing complex SNP data across numerous populations.