BrEPS: a flexible and automatic protocol to compute enzyme-specific sequence profiles for functional annotation

Area of Science:

• Bioinformatics
• Computational Biology
• Enzymology

Background:

• Accurate enzyme function prediction is crucial for metabolic network simulation.
• Current methods rely on sequence databases and operon analysis.
• Novel approaches are needed to enhance functional annotation accuracy.

Purpose of the Study:

• To develop an automatic method, BrEPS, for creating highly specific sequence patterns.
• To improve the functional annotation of enzymes based on genomic sequences.
• To support the reconstruction of metabolic networks.

Main Methods:

• Enzyme sequences from UniProtKB were clustered and aligned using BLAST and ClustalW.
• Conserved columns in alignments were used to construct sequence patterns.
• Pattern specificity was computed and refined iteratively; BrEPS was compared to PRIAM.

Main Results:

• The BrEPS protocol generated highly specific sequence patterns for enzyme annotation.
• Performance was evaluated on Swiss-Prot, showing comparable results to the PRIAM method.
• True positive annotations were determined for five microorganisms using BRENDA and AMENDA data.

Conclusions:

• BrEPS provides a valuable, automatic, and unsupervised method for enzyme functional annotation.
• The generated sequence patterns support metabolic network reconstruction.
• The method is efficient, particularly after pattern evaluation.