Measuring guide-tree dependency of inferred gaps in progressive aligners

Area of Science:

• Bioinformatics
• Computational Biology
• Evolutionary Biology

Background:

• Multiple sequence alignments are crucial for evolutionary studies.
• Progressive alignment methods use guide-trees and introduce gaps to maximize residue pairing.
• The evolutionary meaning of inferred gaps remains unclear.

Purpose of the Study:

• To develop a method for separating phylogenetic signals from guide-tree bias in sequence alignment gaps.
• To assess the informativeness of inferred gap patterns for reconstructing phylogenies.

Main Methods:

• Developed a novel framework to disentangle phylogenetic signal from guide-tree information.
• Analyzed patterns of inferred gaps in multiple sequence alignments.

Main Results:

• Most inferred gaps are incorrectly inserted and primarily reflect the guide-tree structure.
• Gap patterns in current alignments are largely uninformative for phylogeny reconstruction.
• The program PRANK is particularly sensitive to guide-tree bias.

Conclusions:

• The reliance on guide-trees in progressive alignment can lead to non-informative gap patterns.
• Current alignment methods may overestimate the phylogenetic signal from gaps.
• Further research is needed to develop alignment strategies that better capture evolutionary information.