Empirical comparison of web-based antimicrobial peptide prediction tools

Area of Science:

• Biochemistry
• Bioinformatics
• Computational Biology

Background:

• Antimicrobial peptides (AMPs) are crucial innate immune molecules with broad-spectrum antimicrobial activity.
• Rising microbial resistance necessitates the development of novel antimicrobial agents.
• Numerous online tools exist for predicting AMPs, but their comparative performance is unclear.

Purpose of the Study:

• To systematically evaluate and compare the performance of ten publicly available antimicrobial peptide (AMP) prediction tools.
• To identify the most effective tools for predicting different classes of AMPs, including general antimicrobial, antibacterial, and bacteriocin peptides.
• To provide a benchmark for future development and selection of AMP prediction methods.

Main Methods:

• Compilation of two benchmark datasets comprising antimicrobial, antibacterial, and bacteriocin peptides, alongside non-AMPs.
• Systematic evaluation of ten publicly available AMP prediction tools using these datasets.
• Performance assessment based on metrics such as the area under the receiver operating characteristic (ROC) curve.

Main Results:

• CAMPR3(RF) demonstrated statistically significant superior performance among general AMP prediction tools compared to five other methods.
• For antibacterial peptide prediction, the original AntiBP method outperformed its successor, AntiBP2, on one benchmark dataset.
• Both BAGEL3 and BACTIBASE showed strong performance for bacteriocin prediction, with BAGEL3 outperforming BACTIBASE on the larger dataset.

Conclusions:

• CAMPR3(RF) is a highly effective tool for general antimicrobial peptide prediction.
• The performance of prediction tools can vary depending on the specific class of antimicrobial peptide.
• BAGEL3 represents an advancement in bacteriocin prediction compared to its predecessor.