Similarity search for local protein structures at atomic resolution by exploiting a database management system

Area of Science:

• Structural biology
• Bioinformatics
• Computational chemistry

Background:

• Identifying local structural similarities in proteins is crucial for understanding protein function and evolution.
• Analyzing vast protein structural databases requires efficient computational methods.

Purpose of the Study:

• To develop and present a novel method for searching local structural similarities in proteins at atomic resolution.
• To demonstrate the efficiency and applicability of this method for large-scale structural data analysis.

Main Methods:

• Geometric indexing of protein structural data within a relational database.
• Iterative application of the Hungarian algorithm for atomic-resolution structural alignments.
• Statistical significance estimation using an empirical model based on gamma distribution.

Main Results:

• The geometric indexing method efficiently handles large structural datasets (over 160,000 candidates) within reasonable CPU time on standard hardware.
• Structurally similar ligand binding sites can be rapidly enumerated.
• Significant local structural similarities are detected between homologous and non-homologous proteins.

Conclusions:

• Geometric indexing provides an effective approach for discovering local structural similarities in protein databases.
• The method facilitates the identification of functionally relevant structural motifs across diverse protein families.
• This technique enhances the capabilities of structural bioinformatics for large-scale data mining.