Inferring time-dependent migration and coalescence patterns from genetic sequence and predictor data in structured...

Area of Science:

• Evolutionary biology
• Computational biology
• Population genetics

Background:

• Phylodynamic methods infer population dynamics from genetic data.
• Current methods often assume constant rates or unstructured populations.
• Inferring time-varying parameters in structured populations is computationally challenging.

Purpose of the Study:

• To develop a method for inferring predictors of time-varying migration rates and effective population sizes in structured populations.
• To overcome limitations of existing phylodynamic models in handling complex population structures and temporal variations.

Main Methods:

• Utilizes a generalized linear model (GLM) approach with the marginal approximation of the structured coalescent.
• Applies the method to infer parameters and predictors from phylogenetic trees.
• Implements the framework within the BEAST2 package MASCOT for joint inference.

Main Results:

• Successfully infers model parameters and their predictors from simulated phylogenetic trees.
• Demonstrates superior performance compared to discrete trait GLM models.
• Identifies weekly cases as a key predictor for effective population size and geographic distance for migration in an Ebola virus dataset.

Conclusions:

• The developed method reliably infers predictors of population dynamics in structured populations.
• Provides a powerful tool for understanding factors influencing migration and population size changes.
• Enables joint inference of population dynamics, phylogenetic tree, and evolutionary parameters.