UniAlign

UniAlign integrates evolutionary information with 3D geometric similarity to perform protein pairwise structure alignment that identifies functionally relevant residue correspondences between homologous proteins.

Key Features:

• Incorporation of Evolutionary Information: Integrates sequence similarity, sequence profiles, and residue conservation into the structure alignment process.
• Per-Residue Scoring System (UniScore): Defines UniScore as a per-residue weighted sum of sequence similarity, profile information, and residue conservation to evaluate potential alignments.
• Iterative Optimization Procedure: Employs an iterative optimization routine to search for the alignment that maximizes the overall UniScore.
• Benchmark Performance: Demonstrates consistent outperformance of commonly used structure alignment methods on CDD, HOMSTRAD, and BAliBASE benchmark datasets.
• Family-Specific Model Development: Supports development of family-specific models to improve alignment quality within particular protein families.

Scientific Applications:

• Functional Annotation: Improves identification of functional sites on proteins by incorporating evolutionary conservation into alignments.
• Evolutionary Studies: Enhances analysis of evolutionary relationships between distantly related homologous proteins by focusing on conserved functional and structural features.
• Drug Design: Enables more precise identification of functionally relevant residues for structure-based drug design through evolution-aware alignments.

Methodology:

Collects sequence similarity data, sequence profiles, and residue conservation metrics; integrates these features into the UniScore per-residue weighted sum; and applies an iterative optimization procedure to find the alignment that maximizes overall UniScore.