Macrel

Macrel identifies and classifies antimicrobial peptides (AMPs) in genomes and metagenomes to predict AMP candidates and assess antimicrobial and hemolytic activities from sequence data.

Key Features:

• End-to-end pipeline: Processes raw genomic and metagenomic data to produce high-quality AMP candidate identifications.
• 22 peptide feature set: Extracts a specific set of 22 peptide features engineered for AMP prediction.
• Activity classifiers: Machine-learning classifiers predict both antimicrobial and hemolytic activities of peptide sequences.
• Precision and validation: Classifiers were evaluated for precision using realistic simulations and real-world data.
• Implementation: Implemented in Python 3.

Scientific Applications:

• Novel AMP discovery: Mining genomes and metagenomes to identify candidate antimicrobial peptides for therapeutic development.
• Metagenomic screening: Large-scale screening of metagenomic contigs for AMP-encoding sequences.
• Safety assessment: Concurrent prediction of hemolytic activity to inform potential toxicity of AMP candidates.

Methodology:

Feature engineering to compute 22 peptide features and machine-learning classifiers implemented in Python 3 are applied within a pipeline from raw genomes/metagenomes to AMP candidate identification, with classifier validation using realistic simulations and real-world data.