NAPS

NAPS predicts amino acid residues that mediate DNA and RNA binding in nucleic acid-binding proteins.

Key Features:

• Residue prediction: Predicts specific amino acid residues that mediate binding to DNA and RNA.
• Separate DNA/RNA models: Provides distinct models for DNA-binding and RNA-binding residue prediction.
• Input features: Uses sequence-based attributes as input features for prediction.
• Core algorithm: Implements the C4.5 decision tree algorithm.
• Ensemble and learning techniques: Enhances C4.5 with bootstrap aggregation (bagging) and cost-sensitive learning.
• Performance metrics: DNA-binding model accuracy 79.1% and RNA-binding model accuracy 73.2%.

Scientific Applications:

• Interaction characterization: Characterizing mechanisms of nucleic acid-binding proteins interacting with DNA and RNA.
• Functional annotation: Annotating nucleic acid-binding proteins by identifying candidate binding residues.
• Mutagenesis guidance: Guiding site-directed mutagenesis experiments by prioritizing putative binding residues.

Methodology:

Uses sequence-based attributes with a C4.5 decision tree algorithm enhanced by bootstrap aggregation (bagging) and cost-sensitive learning.