6: Multiple Sequence Alignment (MSA) & Progressive Methods

Chapter 6: Multiple Sequence Alignment (MSA) & Progressive Methods

While pairwise alignment compares two sequences, biological families often comprise dozens, hundreds, or thousands of homologous proteins or genes. A Multiple Sequence Alignment (MSA) arranges three or more biological sequences simultaneously, revealing evolutionary conservation, active catalytic residues, and phylogenetic lineages.

1. The Progressive Alignment Paradigm (Clustal)

Developed by Des Higgins and Paul Sharp, progressive alignment solves the MSA problem in three sequential steps:

• Step 1: All-Against-All Pairwise Alignment: Performs \(rac{N(N-1)}{2}\) pairwise alignments to compute an evolutionary distance matrix between all pairs of sequences.
• Step 2: Guide Tree Construction: Constructs a phylogenetic guide tree using hierarchical clustering (e.g. Neighbor-Joining or UPGMA) based on the distance matrix.
• Step 3: Progressive Profile Alignment: Aligns the most closely related pair of sequences first. As it traverses up the guide tree, sequences and existing alignments are converted into profiles (consensus frequency matrices), and profile-profile alignments are calculated until all sequences are merged.

2. Modern Iterative Tools: MUSCLE & MAFFT

# 1. Run MUSCLE with iterative profile refinement
muscle -align unaligned_proteins.fasta -output aligned_muscle.fasta

# 2. Run MAFFT using high-accuracy L-INS-i mode (ideal for <200 sequences)
mafft --localpair --maxiterate 1000 input_sequences.fasta > mafft_linsi.fasta

# 3. Run MAFFT using ultra-fast FFT-NS-2 mode (ideal for >10,000 viral genomes)
mafft --retree 2 --maxiterate 2 input_genomes.fasta > mafft_rapid.fasta