Modern Phylogenomics: Building Phylogenetic Trees Using the Multispecies Coalescent Model

Area of Science:

• Phylogenetics and Evolutionary Biology
• Computational Biology
• Genomics

Background:

• The multispecies coalescent (MSC) model is a key framework for inferring species phylogenies from multilocus DNA sequence data.
• Traditional MSC approaches and early implementations faced limitations, particularly in 'anomaly zones' where gene tree frequencies conflict with the true species tree.
• Concatenation and simple vote-based methods can be inconsistent, converging on incorrect species trees with increased data.

Purpose of the Study:

• To review and evaluate different implementations of the multispecies coalescent model for phylogenetic tree construction.
• To highlight the strengths and weaknesses of various MSC methods, including older and newer approaches.
• To demonstrate the utility of statistical tests, such as likelihood ratio tests, for evaluating competing species tree hypotheses.

Main Methods:

• Discussion of early MSC methods (parsimony, democratic vote, concatenation) and their inconsistencies.
• Examination of second-generation MSC models (Bayesian, likelihood-based) and their scalability issues with large datasets.
• Focus on two-step methods (MP-EST, ASTRAL) for estimating species trees from gene trees, emphasizing their consistency in anomaly zones.
• Illustration of likelihood ratio tests using MP-EST for statistical hypothesis testing of species trees.

Main Results:

• Early MSC methods are inconsistent in certain phylogenetic scenarios, notably the anomaly zone.
• Second-generation Bayesian methods are consistent but computationally intensive for large phylogenomic datasets.
• Two-step methods like MP-EST and ASTRAL are consistent in anomaly zones and handle large datasets effectively.
• MP-EST facilitates statistical tests of species trees, offering an alternative to multilocus bootstrapping.

Conclusions:

• Multispecies coalescent (MSC) models are powerful tools for inferring species phylogenies from multilocus data.
• Modern two-step methods (MP-EST, ASTRAL) provide consistent and scalable solutions for phylogenomic analysis.
• Statistical hypothesis testing using likelihood models enhances the reliability of phylogenetic inference.