PREvaIL, an integrative approach for inferring catalytic residues using sequence, structural, and network features in...

Area of Science:

• Biochemistry and Structural Biology
• Computational Biology and Bioinformatics
• Enzyme Engineering

Background:

• Identifying enzyme catalytic residues is crucial for understanding protein function and designing new enzymes.
• Experimental methods for enzyme characterization are resource-intensive, highlighting the need for computational approaches.
• Computational tools can bridge the gap between protein sequence, structure, and function.

Purpose of the Study:

• To introduce PREvaIL, a novel computational method for predicting enzyme catalytic residues.
• To leverage multi-level features (sequence, structure, residue-contact network) for improved prediction accuracy.
• To provide a valuable tool for enzyme functional characterization and protein design.

Main Methods:

• Development of the PREvaIL method using a random forest machine-learning framework.
• Extraction of informative features from sequence, 3D structure, and residue-contact network data.
• Rigorous benchmarking using 10-fold cross-validation and independent tests on eight datasets.

Main Results:

• PREvaIL demonstrated competitive predictive performance, with AUC values ranging from 0.896 to 0.973.
• The method effectively integrated diverse features from multiple biological levels.
• Performance comparisons showed PREvaIL outperforming or matching seven other modern prediction methods.

Conclusions:

• PREvaIL offers a significant advancement in computational catalytic residue prediction.
• The method enhances understanding of enzyme sequence-structure-function relationships.
• PREvaIL can accelerate the characterization of enzymes with unknown functions.