Algebraic Invariants for Inferring 4-Leaf Semi-Directed Phylogenetic Networks

Area of Science:

• Evolutionary biology
• Computational phylogenetics
• Bioinformatics

Background:

• Phylogenomics aims to reconstruct species evolutionary history using sequence data.
• Phylogenetic networks model complex evolutionary events but are challenging to infer.
• Semi-directed phylogenetic networks are of growing interest, often built by combining smaller networks.

Purpose of the Study:

• To investigate an algebraic method for inferring semi-directed phylogenetic networks.
• To analyze leaf-pattern probabilities for phylogenetic network inference.
• To assess the method's accuracy with varying sequence lengths.

Main Methods:

• Utilized an algebraic approach analyzing leaf-pattern probabilities.
• Applied the method to infer 4-leaf semi-directed phylogenetic networks from nucleotide sequences.
• Tested the method on simulated data and a real dataset from Xiphophorus species.

Main Results:

• Accurate identification of undirected phylogenetic networks with sequences >= 10kbp.
• Accurate identification of semi-directed phylogenetic networks requires sequences approaching 10Mbp.
• Successfully identified tree-like evolution and underlying trees, and applied to Xiphophorus data.

Conclusions:

• The algebraic method is effective for inferring phylogenetic networks, particularly undirected ones.
• Inferring semi-directed networks is more data-intensive, requiring significantly longer sequences.
• The approach is valuable for both simulated and real-world phylogenetic analyses.