Unifying vertical and nonvertical evolution: a stochastic ARG-based framework.

Area of Science:

• Evolutionary biology
• Phylogenetics
• Computational biology

Background:

• Traditional methods using gene tree collections for nonvertical evolution are indirect and hard to generalize.
• Phylogenetic networks model nonvertical evolution but lack statistical significance for specific events.

Purpose of the Study:

• To introduce the Stochastic Model for Reassortment and Transfer Events (SMARTIE) to address limitations in analyzing nonvertical evolution.
• To provide a framework for statistically significant inference of nonvertical evolutionary events using ancestral recombination graphs (ARGs).

Main Methods:

• Developed SMARTIE, a model based on ancestral recombination graphs (ARGs) for formal probabilistic inference.
• Implemented a reversible jump Markov chain Monte Carlo sampler within the BEAST software for posterior distribution approximation.
• Utilized a parallel computing approach for efficient inference on large datasets.

Main Results:

• SMARTIE allows for the inference of a single most probable ARG, offering more direct interpretation than population dynamic statistics.
• The model enables novel probability distributions on ARGs derived from phylogenetic data.
• Demonstrated SMARTIE's applicability through analyses of pathogenic Leptospirochete and Saccharomyces data.

Conclusions:

• SMARTIE offers a statistically robust and interpretable method for studying nonvertical evolutionary processes.
• The model advances phylogenetic inference by integrating ancestral recombination graphs and advanced computational techniques.