Deep-AVPpred

Deep-AVPpred identifies potential antiviral peptides (AVPs) in protein sequences to support discovery of candidate antiviral therapeutics.

Key Features:

• Deep learning classifier: Employs deep learning to classify sequences for antiviral peptide activity.
• Transfer learning: Applies transfer learning within its algorithmic framework to improve identification accuracy.
• Input data: Operates on protein sequences and includes analysis of human interferons-α family proteins.
• Predictive performance: Reports approximately 94% precision on validation and 93% precision on test datasets.
• Candidate generation: Proposes peptide candidates suitable for chemical synthesis and experimental validation.

Scientific Applications:

• AVP discovery: Enables identification of novel antiviral peptides from protein sequences.
• Candidate prioritization: Prioritizes peptides for chemical synthesis and experimental antiviral testing.
• Therapeutic development: Supports antiviral drug discovery efforts for human and veterinary medicine.
• Interferon analysis: Facilitates analysis of human interferons-α family proteins to propose AVP candidates.

Methodology:

The method uses a deep learning classifier applied to protein sequences and incorporates transfer learning to improve AVP identification.