gblocks

gblocks removes poorly aligned positions and divergent regions from multiple sequence alignments to improve their suitability for phylogenetic and comparative analyses.

Key Features:

• Alignment Cleaning: Systematically identifies and removes poorly aligned positions and divergent regions from multiple sequence alignments.
• Block Selection Criteria: Retains blocks based on the number of contiguous conserved positions, absence of gaps, and conservation of flanking positions.
• Adaptability to Alignment Lengths: Provides adjustable stringency levels, with relaxed selection for shorter alignments and stringent criteria for longer ones.
• Phylogenetic Signal Enhancement: Reduces noise from non-homologous or saturated positions to increase the phylogenetic signal in the retained alignment.
• Compatibility with Various Phylogenetic Methods: Has been tested with alignment methods such as ClustalW, Mafft, Probcons and phylogenetic reconstruction techniques including maximum likelihood, neighbor-joining, and parsimony.

Scientific Applications:

• Phylogenetic reconstruction: Produces refined alignments that improve tree reconstruction accuracy for methods such as maximum likelihood, neighbor-joining, and parsimony.
• Alignment refinement for poorly conserved sequences: Removes non-homologous or saturated positions from low-conservation alignments to reduce bias in comparative analyses.
• Mitochondrial protein analyses in diverse eukaryotes: Applied to mitochondrial protein datasets to reduce sequence divergence, yielding more uniform amino acid compositions, smaller pairwise distances, and clearer topologies with stronger support for evolutionary groupings.

Methodology:

Automated identification and removal of poorly aligned positions and divergent regions using block selection criteria based on contiguous conserved positions, absence of gaps, conservation of flanking positions, and adjustable stringency levels.