Inferring Phylogenetic Networks with Maximum Pseudolikelihood under Incomplete Lineage Sorting.

Area of Science:

• Evolutionary biology
• Computational biology
• Genetics

Background:

• Phylogenetic networks are crucial for depicting complex evolutionary histories beyond simple tree structures.
• Existing methods for inferring phylogenetic networks often lack scalability, hindering their application to large datasets.
• Many species exhibit non-tree-like inheritance patterns, including horizontal gene transfer and hybridization.

Purpose of the Study:

• To develop a computationally efficient and scalable statistical method for inferring phylogenetic networks.
• To account for both incomplete lineage sorting and horizontal gene transfer in evolutionary reconstructions.
• To provide a robust tool for analyzing complex evolutionary relationships in diverse taxa.

Main Methods:

• A pseudolikelihood framework was developed for phylogenetic network inference from multi-locus genetic data.
• The model incorporates the coalescent model to address incomplete lineage sorting.
• Reticulation nodes were used to represent horizontal gene inheritance, and quartet-level estimation enabled parallelization.

Main Results:

• The pseudolikelihood method demonstrated significant speed improvements over full likelihood approaches without sacrificing accuracy.
• The method effectively handles incomplete lineage sorting and horizontal gene transfer.
• Successful application to reconstruct evolutionary relationships in the genus Xiphophorus, known for hybridization.

Conclusions:

• The developed pseudolikelihood method offers a scalable and accurate solution for phylogenetic network inference.
• This approach enhances the study of complex evolutionary histories, particularly in species with reticulate evolution.
• The method provides a valuable tool for evolutionary biologists studying hybridization and gene flow.