KPP

KPP constructs phylogenetic profiles from sequence contigs to generate evolutionary information features for predicting pathogen–host interactions (PHIs).

Key Features:

• Phylogenetic profile construction: Compresses sequences into contigs to build phylogenetic profiles that capture evolutionary signals.
• Evolutionary information features: Generates evolutionary information features derived from the constructed phylogenetic profiles.
• Feature extraction models: Implements five different feature-extraction models to represent evolutionary patterns relevant to PHI prediction.
• Integration with structure-based data: Designed to combine phylogenetic-profile features with structure-based information extracted directly from sequences to improve predictive performance.
• Machine-learning compatibility: Produces features intended for use in ML-based PHI prediction workflows, distinct from purely structure-based ML approaches.

Scientific Applications:

• PHI prediction: Improves prediction accuracy for pathogen–host interactions by supplying evolutionary features for classifiers.
• Molecular mechanism exploration: Supports studies of the molecular mechanisms underlying PHIs through evolutionary pattern analysis.
• Discovery of novel relationships: Aids identification of previously unknown biological relationships between pathogens and hosts by integrating evolutionary and structure-based signals.

Methodology:

Sequences are compressed into contigs to construct phylogenetic profiles; evolutionary information features are generated and represented using five distinct feature-extraction models, which can be combined with structure-based information for ML-based PHI prediction.