STAMP

STAMP aligns protein sequences using three-dimensional (3D) structural comparisons to generate structure-based multiple sequence alignments and assess residue-level alignment quality.

Key Features:

• Structure-Based Alignment: Starts from an initial sequence-based multiple alignment to inform an early superposition of protein structures for detailed structural comparison.
• Algorithmic Approach: Compares all pairs of superposed proteins to compute a similarity tree that guides generation of multiple sequence alignments by tracing paths from branches to the root.
• Output at Branchpoints and Root: Produces structure-based sequence alignments and coordinate transformations at each branchpoint and a comprehensive multiple alignment at the root.
• Quality Assessment: Introduces two indices—Sc, which measures global structural similarity between pairs or groups of proteins, and Pij', which provides a normalized per-residue alignment confidence measure.
• Reproducibility and Reliability: Characterizes alignments with Sc and Pij' values to systematically assess and document alignment quality for studies of residue conservation within structural motifs.

Scientific Applications:

• Fold-specific alignment: Aligns proteins with well-characterized folds such as dehydrogenase fold domains, globins, and serine proteinases using structural information.
• Alignment validation against published structures: Produces alignments consistent with published structural accounts for benchmarking and comparative structural analysis.
• Residue conservation and motif analysis: Supports studies of residue conservation and functional motifs by providing structure-informed alignment and per-residue confidence (Pij').

Methodology:

Generates an initial sequence-based multiple alignment, superposes structures based on that alignment, applies a structure comparison algorithm to derive a similarity tree, and produces alignments at each branchpoint and the root with Sc and Pij' quality indices.