A Survey of Methods for Constructing Rooted Phylogenetic Networks

Area of Science:

• Phylogenetics and Evolutionary Biology
• Computational Biology
• Bioinformatics

Background:

• Rooted phylogenetic networks are crucial for depicting evolutionary reticulations and conflicting data.
• Several methods exist for constructing these networks, with decomposition-based and incompatible graph-based approaches being prominent.
• Efficient construction of accurate phylogenetic networks remains an active research area.

Purpose of the Study:

• To compare the efficiency and effectiveness of rooted phylogenetic network construction methods.
• To evaluate methods based on network decomposition and incompatible graphs, specifically CASS, LNETWORK, and BIMLR.
• To identify which method yields simpler and more accurate phylogenetic networks.

Main Methods:

• Comparative analysis of phylogenetic network construction algorithms.
• Utilizing both practical and artificial datasets for evaluation.
• Assessing methods based on computational running time and the quality (simplicity) of the resulting networks.

Main Results:

• Methods utilizing network decomposition and incompatible graphs (CASS, LNETWORK, BIMLR) are generally more efficient.
• LNETWORK demonstrated a superior ability to construct significantly simpler rooted phylogenetic networks compared to CASS and BIMLR.
• Performance was evaluated across diverse datasets to ensure robustness.

Conclusions:

• LNETWORK is a highly effective method for constructing simplified rooted phylogenetic networks.
• The choice of method significantly impacts the complexity and interpretability of evolutionary networks.
• Further research can explore optimizations for other network construction algorithms.