A Robust ANOVA Approach to Estimating a Phylogeny from Multiple Genes.

Area of Science:

• Bioinformatics
• Computational Biology
• Evolutionary Genetics

Background:

• Phylogenetic tree estimation is crucial for understanding evolutionary relationships.
• Individual gene trees can conflict due to evolutionary processes like lateral gene transfer.
• Existing methods may be computationally intensive or less robust to conflicting gene histories.

Purpose of the Study:

• To develop a robust algorithm for phylogenetic tree estimation from multi-gene sequence data.
• To address the challenge of incongruent evolutionary histories among individual genes.
• To provide a computationally efficient and statistically sound method for phylogenetic inference.

Main Methods:

• A novel algorithm for tree estimation based on pairwise distances computed gene-by-gene.
• Robust analysis of variance (ANOVA) to combine distances across genes, generating summary distances.
• Tree construction using standard distance-based methods (e.g., BIONJ) with ANOVA weights.

Main Results:

• The robust ANOVA effectively combines gene-specific distances, yielding a summary distance matrix.
• The method allows identification of outlying genes and taxa through ANOVA weights.
• The distance-based approach demonstrates significantly faster computation compared to maximum likelihood on concatenated genes.

Conclusions:

• The proposed method provides a robust and efficient approach to phylogenetic tree estimation from multi-gene data.
• It effectively handles gene tree incongruence and facilitates outlier detection.
• The method is computationally advantageous and amenable to standard bootstrap analysis for statistical validation.