Fast structural search for classification of gut bacterial mucin O-glycan degrading enzymes.

DEFT achieves over 90% accuracy in classifying gut bacterial enzymes at all EC levels, outperforming existing methods and enabling rapid genome-wide annotations.

• Why it matters: Accurate enzyme classification is crucial for understanding microbial functions and interactions, but current methods struggle with fine-grained distinctions, especially at the catalytic domain level.
• What they did: The approach combines PLM-based annotations for the top two EC levels with structure-based predictions for the remaining levels, using benchmarking datasets to validate its performance.
• The result: This method enables efficient, high-throughput enzyme annotation across entire genomes, as demonstrated by experimentally validating glycoside hydrolase profiles in gut bacteria.