Blocks+: a non-redundant database of protein alignment blocks derived from multiple compilations

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Protein sequence classification and function prediction are vital as biological databases expand.
• The original Blocks Database, based on Prosite, offered valuable sequence classification but had incomplete coverage.
• Expanding the Blocks Database with data from other protein family resources is necessary for broader application.

Purpose of the Study:

• To create a unified and comprehensive protein family database by integrating data from multiple sources.
• To enhance sequence classification accuracy and functional prediction capabilities.
• To develop a non-redundant, hierarchical compilation of protein families.

Main Methods:

• Constructed Blocks+ by integrating data from five protein family databases (Prosite, Prints, Pfam-A, ProDom, Domo).
• Utilized the PROTOMAT/BLOSUM scoring model and a single search algorithm for consistent classification.
• Employed the LAMA blocks-versus-blocks searching program to identify and manage overlapping protein families for non-redundancy.

Main Results:

• Blocks+ comprises 8909 blocks representing 1995 protein families, effectively doubling the coverage of the original Blocks Database.
• The unified database successfully integrates blocks from various sources, including those not present in Prosite.
• Demonstrated the ability to retain related yet distinct families and minimize redundancy, using the SNF2 family as a case study.

Conclusions:

• Blocks+ provides a significantly expanded and unified resource for protein family classification and functional prediction.
• The integration strategy effectively addresses the challenge of non-redundancy in compiling diverse protein family data.
• This enhanced database facilitates more accurate and comprehensive analysis of protein sequences and functions.