Multiple Sequence Alignment Averaging Improves Phylogeny Reconstruction.

Area of Science:

• Bioinformatics
• Computational Biology
• Evolutionary Biology

Background:

• Phylogenetic tree inference traditionally relies on a single multiple sequence alignment (MSA), which can be inaccurate.
• Alignment errors in MSAs can reduce the accuracy of phylogenetic tree reconstruction.
• Filtering unreliable alignment regions or weighting columns has limitations due to signal loss or complexity.

Purpose of the Study:

• To develop and evaluate a novel method for improving phylogenetic tree inference accuracy.
• To address the limitations of single MSA-based approaches by incorporating uncertainty.
• To provide a criterion for assessing the utility of the proposed method.

Main Methods:

• Generation of a large set of alternative multiple sequence alignments (MSAs).
• Concatenation of alternative MSAs into a single SuperMSA.
• Phylogenetic tree reconstruction using the SuperMSA.

Main Results:

• The SuperMSA approach resulted in more accurate phylogenetic trees on average compared to single unfiltered MSAs or weighted MSAs.
• Simulations demonstrated the superiority of the SuperMSA method across various scenarios.
• Analysis identified specific regions within the MSA space where the SuperMSA approach is most beneficial.

Conclusions:

• The SuperMSA methodology offers a robust alternative for phylogenetic tree inference, particularly for diverged sequences.
• This approach effectively captures broader phylogenetic signals often missed by single MSAs.
• A simple criterion is provided to guide the decision-making process for employing the SuperMSA method.