Artificial intelligence-driven discovery of bioactive peptides: Computational approaches and future perspectives

Area of Science:

• Biochemistry
• Bioinformatics
• Computational Biology

Background:

• Bioactive peptides are amino acid chains with diverse biological functions, including antimicrobial, antioxidant, and anti-inflammatory activities.
• Traditional methods for discovering bioactive peptides involve complex, time-consuming separation processes like chromatography and ultrafiltration, requiring significant lab resources.
• These conventional techniques face limitations in terms of efficiency and speed for large-scale peptide discovery.

Purpose of the Study:

• To review and evaluate the application of artificial intelligence (AI) in the discovery of bioactive peptides.
• To highlight the advantages of AI-driven approaches over conventional methods in terms of speed and resource utilization.
• To explore future directions and enhancement strategies for AI in bioactive peptide research.

Main Methods:

• AI-driven computational pipelines encompassing data acquisition, feature engineering, machine learning model construction, training, validation, and prediction.
• Systematic evaluation of recent AI applications across antimicrobial, antioxidant, anti-inflammatory, and multifunctional peptides.
• Review of integrated enhancement strategies, including functional mechanism classification and database-independent modeling.

Main Results:

• AI approaches significantly reduce the time and laboratory resources required for bioactive peptide discovery compared to conventional multi-stage separation processes.
• AI pipelines systematically address critical phases from data curation to high-throughput prediction.
• The review critically assesses AI applications in four key bioactive peptide categories.

Conclusions:

• AI represents a transformative platform for accelerating the discovery and prediction of bioactive peptides.
• Future advancements include scenario-specific peptide customization and prediction of bioactivity in digested proteomes using AI.
• Integrated strategies can further enhance AI's utility in classifying peptides by functional mechanism and enabling database-independent modeling.