Machine Learning Prediction of Antimicrobial Peptides.

Area of Science:

• Bioinformatics and Computational Biology
• Drug Discovery and Development
• Microbiology and Infectious Diseases

Background:

• Antibiotic resistance is a critical global health threat, necessitating the development of new antimicrobial agents.
• Naturally occurring antimicrobial peptides (AMPs) serve as promising templates for novel antibiotic development.
• Predicting AMPs from genomic data is crucial for accelerating the discovery of new antimicrobials.

Purpose of the Study:

• To provide an overview of machine-learning (ML) based prediction methods for antimicrobial peptides (AMPs).
• To explore the evolution of AMP prediction from single-label to multi-label functional predictions.
• To discuss the integration of peptide characteristics like posttranslational modification and 3D structure in ML predictions.

Main Methods:

• Review of existing ML-based AMP prediction tools (e.g., AntiBP, CAMP, iAMPpred, iAMP-2L, MLAMP, AMAP).
• Analysis of single-label predictions for various antimicrobial activities (antifungal, antiviral, antibiofilm, etc.).
• Examination of multi-label predictions encompassing a broad spectrum of biological activities.
• Consideration of peptide features such as posttranslational modification, 3D structure, and microbial species specificity.

Main Results:

• Machine learning models can predict antimicrobial activity with high accuracy, expanding to diverse functions beyond antibacterial effects.
• Multi-label prediction frameworks enable the identification of AMPs with multiple biological roles, including anticancer and antiviral activities.
• Comparison of ML-identified key amino acids with natural AMP residue patterns provides insights into AMP structure-activity relationships.

Conclusions:

• Machine learning is a powerful tool for accelerating the discovery of novel antimicrobial peptides (AMPs).
• The predictive capabilities extend to a wide range of AMP functions, aiding in the development of broad-spectrum antimicrobials.
• Future directions include developing comprehensive prediction pipelines that integrate diverse functional predictions to combat antimicrobial resistance.