ArchDB14

ArchDB14 catalogs protein loop structures and classifies them to support analysis of loop conformations in protein structure, folding dynamics, and function.

Key Features:

• Extensive loop dataset: Contains 149,134 protein loops extracted from known structures.
• Hierarchical classification: Organizes loops into 5,739 distinct classes and 9,608 subclasses.
• Graph-based clustering algorithm: Employs a graph-based clustering algorithm to identify structural similarities among loop conformations.

Scientific Applications:

• Homology modeling: Provides loop classifications that can be used to select loop templates for modeling unknown protein regions.
• Protein structure prediction: Supplies reference loop conformations to inform computational modeling of protein structures.
• Protein design and engineering: Informs design of enzymes and catalytic loops by characterizing loop variability and structural motifs.
• Function prediction: Aids prediction of protein functions by relating loop conformations to dynamic regions involved in binding or conformational change.

Methodology:

Loops are extracted from known protein structures and classified into classes and subclasses using a graph-based clustering algorithm based on structural similarity.