BioMatics 1.0: A Wasserstein Distance Approach for Next-Generation Multiple Sequence Alignment

Area of Science:

• Bioinformatics
• Computational Biology
• Structural Biology

Background:

• Accurate multiple sequence alignment (MSA) is crucial for understanding protein evolution, structure, and function.
• Existing MSA methods often struggle with accuracy in diverse protein families and conserved regions.

Purpose of the Study:

• Introduce BioMatics 1.0, a novel MSA algorithm.
• Enhance the detection of structural and evolutionary patterns through refined alignment.
• Improve residue-level alignment precision for downstream applications.

Main Methods:

• Utilizes optimal transport principles, specifically the Wasserstein first-order distance, for profile-to-profile alignment.
• Employs Earth Mover's Distance on per-position amino acid frequency vectors, guided by BLOSUM62 similarity.
• Incorporates entropy-adaptive gap penalties to adjust alignment in variable regions.

Main Results:

• BioMatics 1.0 demonstrates superior performance in column score (CS) accuracy compared to widely used tools.
• Achieves competitive or comparable sum-of-pairs score (SPS) results.
• Outperforms existing methods on benchmark datasets including conserved domains, structural motifs, and heterogeneous families.

Conclusions:

• BioMatics 1.0 offers a significant advancement in MSA methodology.
• Its residue-level precision enhances downstream phylogenetic reconstruction and structure-informed modeling.
• The optimal transport approach provides a robust framework for future MSA algorithm development.