Consistency and identifiability of the polymorphism-aware phylogenetic models

Area of Science:

• Phylogenetics and Evolutionary Biology
• Population Genetics
• Bioinformatics

Background:

• Polymorphism-aware phylogenetic models (PoMo) extend standard DNA substitution models by incorporating polymorphic states.
• This approach accounts for ancestral and current intra-population variation and population-level processes like genetic drift and selection.
• While valuable, PoMo lacked formal proofs of statistical consistency and identifiability.

Discussion:

• This study provides the first proof of PoMo's identifiability, a prerequisite for statistical consistency.
• The maximum a posteriori (MAP) tree estimator within the PoMo framework is demonstrated to be a consistent estimator of the species tree.
• Theoretical findings are validated using simulated genome-wide data reflecting natural population diversity and incomplete lineage sorting.

Key Insights:

• PoMo is statistically identifiable, confirming its theoretical soundness for phylogenetic analysis.
• The MAP tree estimator in PoMo is statistically consistent, ensuring reliable species tree reconstruction.
• The implemented Bayesian PoMo framework successfully recovers true species trees with typical genome-wide data.

Outlook:

• Further applications of statistically consistent PoMo in diverse phylogenetic studies.
• Potential for refining population genetics inference using polymorphism-aware models.
• Advancements in computational methods for large-scale phylogenetic analyses incorporating intra-population variation.