ProP

ProP predicts propeptide cleavage sites in eukaryotic protein sequences, identifying arginine and lysine residues recognized by subtilisin/kexin-like proprotein convertases (PCs) to pinpoint processing sites that activate secretory protein precursors.

Key Features:

• Cleavage specificity: Predicts propeptide cleavage at basic amino acids arginine and lysine recognized by subtilisin/kexin-like proprotein convertases, including furin.
• Artificial Neural Networks: Employs neural network models to analyze sequence context for cleavage-site prediction.
• Furin-specific network: A neural network trained on curated experimental data from the literature for furin cleavage-site prediction.
• General PC-specific network: A neural network trained using sequence data from the Swiss-Prot protein database to predict cleavage by subtilisin/kexin-like PCs.
• Predictive performance: Reported sensitivity of 95% for the furin-specific network and 62% for the general PC-specific network on independent sequences.

Scientific Applications:

• Post-translational processing studies: Identification of proteolytic processing sites in secretory proteins to study conversion of inactive precursors into active polypeptides.
• Disease mechanism investigation: Analysis of cleavage events implicated in Alzheimer's disease, cancer, anthrax, and Ebola fever to inform understanding of pathogenic processes.
• Therapeutic target analysis: Mapping PC cleavage sites to support identification of potential intervention points related to furin and other proprotein convertases.

Methodology:

ProP uses artificial neural networks trained on curated experimental data for the furin-specific model and on Swiss-Prot sequence data for the general PC-specific model to analyze protein sequences.