Important progress in antimicrobial peptide prediction research in the past five years

Antimicrobial peptides (AMPs), as key targets for novel anti-infection therapy, have made their efficient and precise identification technology a critical breakthrough in biomedical research. Currently, the high cost of experimental validation severely restricts the development process of antimicrobial peptides, while computational biology methods demonstrate significant advantages due to their cost-effectiveness and efficiency. This review comprehensively examines antimicrobial peptide database resources, with a focus on analyzing the characteristics and application value of mainstream databases such as APD3, DRAMP, and DBAASP, and systematically summarizes the latest advances in feature encoding techniques in recent years. At the methodological level, we discuss in detail the innovative prediction methods that have emerged in the past five years (2020-2024): from traditional structural feature analysis and machine learning algorithms to cutting-edge deep learning and generative model technologies, providing an in-depth comparison of the predictive performance of various methods on different datasets. Finally, based on current research bottlenecks, we prospectively propose future development trends in antimicrobial peptide prediction research, providing direction for scientific research efforts in related fields.