STEP3

STEP3 leverages evolutionary features in phage genome sequences to improve genome annotation and predict phage virion components relevant to phage therapy and antimicrobial resistance.

Key Features:

• Evolutionary Feature Recognition: Identifies and utilizes "evolutionary features" within phage genomes to distinguish specific and universal characteristics across diverse phages and improve annotation quality for genomes from infrequently sampled environmental sources.
• Ensemble Framework Integration: Integrates evolutionary features into an ensemble framework to achieve stable and robust prediction performance and to outperform existing prediction tools when benchmarked against a wide array of phages from diverse origins.
• Machine Learning Application: Employs machine learning techniques to exploit diversity in evolutionary features and enhance prediction accuracy for phage virion components, including protein components that dictate shelf-life and therapeutic applicability.
• Validation through High-Resolution Mass Spectrometry: Predictive capabilities have been validated using high-resolution mass spectrometry analysis on two novel phages isolated from a Southern Hemisphere watercourse.

Scientific Applications:

• Phage therapy candidate selection: Improves selection of effective therapeutic phages by predicting relevant virion components from genome sequences.
• Genome annotation improvement: Addresses poor annotation quality for phage genomes, particularly those derived from undersampled or infrequently sampled environmental sources.
• Evaluation of phage component parts: Recognizes and evaluates protein components that influence shelf-life and applicability for therapeutic use.

Methodology:

Performs comprehensive analysis of phage genomes to extract evolutionary features, integrates those features into an ensemble framework, and applies machine learning algorithms to refine prediction accuracy.