CSM-peptides

CSM-peptides predicts therapeutic peptide activity using machine learning to identify candidate peptides across eight classes: anti-angiogenic, anti-bacterial, anti-cancer, anti-inflammatory, anti-viral, cell-penetrating, quorum sensing, and surface-binding.

Key Features:

• Machine Learning Methodology: Employs a novel machine learning methodology developed to identify therapeutic peptides.
• Supported Peptide Classes: Identifies eight distinct therapeutic peptide types: anti-angiogenic, anti-bacterial, anti-cancer, anti-inflammatory, anti-viral, cell-penetrating, quorum sensing, and surface-binding.
• Performance: Achieves an Area Under the Curve (AUC) of up to 0.92 in independent blind tests and maintains predictive performance across cross-validation.
• Data-Driven Development: Trained using extensive collections of experimentally characterized therapeutic peptides.

Scientific Applications:

• High-Throughput Screening: Screens large peptide libraries to identify novel candidate therapeutic peptides.
• Therapeutic Development Support: Supports the development of new peptide-based treatments across various medical fields by identifying candidate peptides.

Methodology:

Machine learning models were trained on experimentally characterized therapeutic peptide datasets and validated using cross-validation and independent blind tests.