The Antimicrobial Peptide Pipeline: A Bacteria-Centric AMP Predictor.

Abstract

Introduction: Antimicrobial peptides (AMPs), unlike antibiotics, are encoded in genomes. AMPs are exported from the cell after expression and translation. In the case of bacteria, the exported peptides target other microbes to give the producing bacterium a competitive edge. While AMPs are sought after for their similar antimicrobial activity to traditional antibiotics, it is difficult to predict which combinations of amino acids will confer antimicrobial activity. Many computer algorithms have been designed to predict whether a sequence of amino acids will exhibit antimicrobial activity, but the vast majority of validated AMPs in databases are still of eukaryotic origin. This defies common sense since the vast majority of life on Earth is prokaryotic.

Methods: The antimicrobial peptide pipeline, presented here, is a bacteria-centric AMP predictor that predicts AMPs by taking design inspiration from the sequence properties of bacterial genomes with the intention to improve the detection of naturally occurring bacterial AMPs. The pipeline integrates multiple concepts of comparative biology to search for candidate AMPs at the primary, secondary, and tertiary peptide structure levels.

Results: Results showed that the antimicrobial peptide pipeline identifies known AMPs that are missed by state-of-the-art AMP predictors and that the pipeline yields more AMP candidates from real bacterial genomes than from fake genomes, with the rate of AMP detection being significantly higher in the genomes of six nosocomial pathogens than in the fake genomes.

Conclusion: This bacteria-centric AMP pipeline enhances the detection of bacterial AMPs by incorporating sequence properties unique to bacterial genomes. It complements existing tools, addressing gaps in AMP detection and providing a promising avenue for discovering novel antimicrobial peptides.