SNAP

SNAP performs non-sequential pair-wise structural alignment of proteins to detect structural similarity and infer evolutionary relationships when sequence similarity is low.

Key Features:

• Non-Sequential Alignment: Identifies and aligns structurally similar regions that are not contiguous in sequence, enabling detection of rearranged or permuted structural correspondences.
• Iterative Process: Employs a two-step iterative cycle of superposition and alignment that repeats until convergence to improve alignment quality.
• Greedy Algorithm: Uses a novel greedy algorithm to construct both sequential and non-sequential alignments efficiently across protein pairs.
• Performance Evaluation: Validated against manually curated reference alignments in the SISY and RIPC datasets, producing longer alignments with lower root-mean-square deviation (rmsd) compared to several methods.
• Fold Classification: Achieved high sensitivity and selectivity in fold classification on a dataset of 4,410 protein pairs from the CATH database.

Scientific Applications:

• Evolutionary Studies: Reveals structural similarities between proteins with low sequence identity to support inference of evolutionary relationships.
• Protein Structure Analysis: Identifies structurally conserved and functionally relevant regions that are not apparent from sequence alignments alone.
• Fold Classification and Comparison: Facilitates accurate classification and comparative analysis of protein folds and domains using structural alignment metrics.

Methodology:

Starts from an initial alignment and iteratively refines it by alternating superposition and alignment steps, employing a greedy algorithm to construct sequential and non-sequential alignments and repeating until convergence.