hlp

hlp predicts peptide half-life in an intestine-like environment and assists design of single-point mutant peptides with optimized physicochemical properties to improve proteolytic stability.

Key Features:

• Half-Life Prediction: SVM-based models predict peptide half-life using datasets of 10mer (HL10) and 16mer (HL16) peptides and features including amino acid composition, dipeptide composition, and tripeptide composition.
• Model Performance: For HL10, correlations (R/R²) are 0.57/0.32, 0.68/0.46, and 0.69/0.47 for amino acid, dipeptide, and tripeptide compositions respectively; for HL16, correlations are 0.91/0.82, 0.90/0.39, and 0.90/0.31 respectively, with selected feature-based models achieving R = 0.70 for HL10 and R = 0.98 for HL16.
• Physicochemical Properties Calculation: Computes peptide properties including charge, polarity, hydrophobicity, volume, and pK values.
• Mutant Peptide Design: Generates all possible single-point mutants for a given peptide sequence to evaluate effects on half-life and physicochemical properties.

Scientific Applications:

• Peptide therapeutic optimization: Guides selection and modification of peptide sequences to enhance stability against intestinal proteases for therapeutic development.
• Oral bioavailability design: Assists design of peptides with improved resistance to proteolytic degradation in intestine-like conditions to support oral administration strategies.

Methodology:

Models were trained using Support Vector Machine algorithms on experimentally determined half-life data from peptides exposed to crude intestinal protease preparations, with features comprising amino acid, dipeptide, and tripeptide compositions; the system also computes physicochemical properties and enumerates all single-point mutants.