Identification of prokaryotic and eukaryotic signal peptides and prediction of their cleavage sites

Key Points

• To develop a neural network-based computational method for identifying signal peptides and accurately predicting their cleavage sites in both prokaryotic and eukaryotic protein sequences.
• Trained artificial neural networks on separate, curated datasets of prokaryotic and eukaryotic protein sequences.
• Designed the prediction model to operate on genome-wide datasets and differentiate between cleaved signal peptides and uncleaved N-terminal signal-anchor sequences.
• Achieved significantly superior prediction accuracy for signal peptides and cleavage sites compared to previous prediction schemes.
• Successfully discriminated between cleaved signal peptides and uncleaved N-terminal signal-anchor sequences, though with lower precision relative to cleavage site localization.