MSARC: Multiple sequence alignment by residue clustering.

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Progressive methods are common for multiple sequence alignment but can be biased by guide-trees, particularly for distant sequences.
• This bias can affect the accuracy of alignments in phylogenetic analyses.

Purpose of the Study:

• To introduce MSARC, a novel algorithm for multiple sequence alignment.
• To evaluate MSARC's performance compared to existing progressive methods, focusing on its ability to mitigate guide-tree bias.

Main Methods:

• MSARC utilizes a graph-clustering approach to align sets of sequences.
• The algorithm operates without the need for guide-trees, a departure from traditional progressive methods.

Main Results:

• MSARC demonstrates alignment quality comparable to leading progressive methods on standard datasets like BAliBASE.
• The algorithm shows superior performance on sequence sets with evolutionary distances poorly represented by phylogenetic trees, areas prone to guide-tree bias.

Conclusions:

• MSARC offers an effective alternative for multiple sequence alignment, particularly for complex or divergent sequence sets.
• The graph-clustering approach successfully addresses the limitations of guide-tree dependent methods.