Using MOEA with Redistribution and Consensus Branches to Infer Phylogenies

Area of Science:

• Computational Biology
• Bioinformatics
• Evolutionary Biology

Background:

• Phylogenetic inference is crucial for understanding evolutionary relationships but is computationally challenging (NP-hard).
• Metaheuristics, including Maximum Parsimony and Maximum Likelihood, are increasingly used for phylogenetic reconstruction.
• Existing multi-objective metaheuristics combining these methods are often slow due to the computational demands of individual objectives.

Purpose of the Study:

• To develop a novel, faster multi-objective optimization algorithm for phylogenetic reconstruction.
• To accelerate the process of rebuilding phylogenies while maintaining solution quality.

Main Methods:

• Proposed a new multi-objective optimization algorithm, MOEA-RC.
• Incorporated the structural information of elite individuals from current populations to enhance processing speed.
• Compared MOEA-RC against established algorithms (MOEA/D, NAGA-II) on real-world datasets.

Main Results:

• MOEA-RC demonstrated superior performance compared to MOEA/D and NAGA-II.
• The algorithm achieved better solutions within a limited number of iterations.
• The effectiveness was validated on three distinct real-world datasets.

Conclusions:

• MOEA-RC offers a significant advancement in accelerating phylogenetic reconstruction.
• The algorithm's efficiency stems from its novel approach to utilizing population structures.
• This method provides a faster and effective alternative for complex phylogenetic analyses.