Seq2Enz

Seq2Enz predicts whether a protein sequence functions as an enzyme and assigns it to a specific enzyme class using mask BLAST augmented with novel structural-chemical properties (NCL) derived from amino acids.

Key Features:

• Enzyme detection and classification: Identifies whether a protein sequence is enzymatic and assigns it to a specific enzyme class.
• Mask BLAST integration: Leverages an advanced application of mask BLAST for sequence similarity-based inference.
• Novel structural-chemical properties (NCL): Incorporates NCL features derived from the fundamental building blocks of proteins (amino acids) into the prediction framework.
• Evaluation dataset: Evaluated on 267,276 reviewed enzyme sequences from UniProt/SwissProt plus 7,062 recent depositions not used in ECPred training.
• Benchmarking: Compared against conventional BLAST methods and the state-of-the-art ECPred tool.
• Hierarchical performance: Demonstrated consistent improvement across all enzyme classes at four hierarchical levels of classification.

Scientific Applications:

• Enzyme function prediction: Assigns enzyme classes to protein sequences for functional inference.
• Enzyme discovery: Supports identification of enzymatic sequences within large protein datasets.
• Functional annotation: Enhances annotation workflows by providing enzyme class predictions for UniProt/SwissProt-derived sequences.
• Comparative benchmarking: Serves as a basis for method comparison against BLAST and ECPred using curated UniProt/SwissProt datasets.

Methodology:

Seq2Enz applies an advanced application of mask BLAST integrated with novel structural-chemical properties (NCL) derived from the fundamental building blocks of proteins (amino acids).