HOMSTRAD: recent developments of the Homologous Protein Structure Alignment Database

Area of Science:

• Structural bioinformatics
• Computational biology
• Protein databases

Background:

• HOMSTRAD is a curated collection of protein families grouped by sequence and structural similarity.
• Sequence alignments are annotated with structural features using JOY, aiding in the identification of conserved residues.
• The database historically utilized flat files and indexes, with tools like COMPARER and MNYFIT for family generation.

Purpose of the Study:

• To introduce HOMSTRAD BETA, a new relational database version of HOMSTRAD.
• To enhance data accessibility, flexibility, and update efficiency for the HOMSTRAD resource.
• To provide improved tools for comparative modeling and identifying distant protein sequence/structure similarities.

Main Methods:

• Development of a relational database using MySQL for HOMSTRAD.
• Integration of JOY for detailed structural feature annotation of sequence alignments.
• Implementation of the FUGUE search engine for structure-aided sequence similarity searches.
• Inclusion of structural superposition data for each protein family.

Main Results:

• HOMSTRAD BETA offers a more flexible and accessible data structure compared to the previous version.
• New web features, including a custom alignment facility, have been added.
• The enhanced database facilitates efficient identification of structurally conserved residues.
• The FUGUE search engine enables sensitive detection of sequence/structure matches.

Conclusions:

• HOMSTRAD and its BETA version represent a valuable resource for structural bioinformatics.
• The relational database structure supports future development and improved data management.
• The integrated tools and annotations enhance the utility of HOMSTRAD for protein modeling and similarity analysis.