Maxflow

Maxflow aligns distantly related protein sequences by leveraging transitive alignments through intermediate sequences to estimate positional structural equivalence and improve alignment reliability across evolutionary distances.

Key Features:

• Transitive alignments through intermediates: Uses intermediate sequences as stepping stones to connect proteins that are not directly or closely related.
• Greedy algorithm with consistency score: Employs a greedy algorithm combined with a novel consistency score to estimate the relative likelihood of alternative transitive alignment paths.
• Structural-equivalence probability model: Models the probability that two positions within protein sequences are structurally equivalent rather than focusing solely on amino acid preferences as in profile models.
• Retention of information across sequence space: Maintains high information content across extensive distances in sequence space to support alignments of highly divergent proteins.
• Identification of sparse active-site signatures: Detects sparse and narrow active-site sequence signatures embedded within high-entropy sequence segments.
• Structure-based multiple alignment support: Applicable to structure-based multiple alignments of large and diverse enzyme superfamilies.
• Benchmark performance on urease superfamily: Demonstrated superior reliability and double coverage in benchmark tests using the urease superfamily compared to existing sequence alignment software.

Scientific Applications:

• Functional genomics: Improves transferability of functional characterizations from model proteins to members of entire protein superfamilies.
• Structural genomics: Enhances the reliability of structure-based multiple alignments for inferring structural equivalence across divergent proteins.
• Annotation of enzyme superfamilies: Facilitates identification of active-site signatures and functional residues within large, diverse enzyme superfamilies.
• Alignment of evolutionarily distant proteins: Addresses the challenge of reliably aligning proteins with significant evolutionary divergence.

Methodology:

Performs transitive alignments through intermediate sequences using a greedy algorithm and a novel consistency score, and models the probability that two sequence positions are structurally equivalent for structure-based multiple alignments.