Antimicrobial peptide identification using multi-scale convolutional network

Area of Science:

• Computational biology
• Bioinformatics
• Drug discovery

Background:

• Antibiotic resistance is a growing global health threat.
• Antimicrobial peptides (AMPs) are a promising alternative to traditional antibiotics.
• Machine learning and deep learning are increasingly used for AMP identification.

Purpose of the Study:

• To develop and evaluate a deep learning model for identifying antimicrobial peptide sequences.
• To improve AMP identification by incorporating additional information into a fusion model.
• To assess the model's performance against state-of-the-art methods.

Main Methods:

• Designed a deep learning model utilizing an embedding layer and a multi-scale convolutional network.
• The multi-scale convolutional network employs varying filter lengths to capture diverse features.
• Developed a fusion model by integrating additional information to enhance predictive capabilities.

Main Results:

• The proposed deep learning model demonstrated superior performance on AMP datasets and the APD3 benchmark.
• The fusion model achieved higher accuracy on an anti-inflammatory peptide (AIP) dataset compared to existing models.
• The multi-scale convolutional network effectively utilizes latent features for improved identification.

Conclusions:

• The multi-scale convolutional network represents a novel advancement in deep neural network models for peptide analysis.
• The developed deep learning and fusion models offer improved capabilities for discovering novel AMPs and AIPs.
• The study provides accessible source code and data for further research in AMP discovery.