PremPLI

PremPLI predicts the impact of single amino acid mutations on protein-ligand interactions by quantitatively estimating changes in binding affinity using a structure-based machine learning framework.

Key Features:

• Quantitative Prediction: Provides numerical estimates of how single amino acid mutations alter binding affinity between proteins and ligands.
• Structure-Based Approach: Requires a 3D structure of the protein-ligand complex as input to ground predictions in molecular architecture.
• Machine Learning Integration: Utilizes structure-based machine learning algorithms to predict mutation effects, compared with physics-based approaches such as first-principle statistical mechanics and mixed physics- and knowledge-based potentials.
• Resource Efficiency: Achieves high predictive performance with lower computational resource requirements and improved computational speed relative to many existing methods.

Scientific Applications:

• Drug Design: Predicts mutation-induced changes in ligand binding to inform rational design of ligand-binding proteins and small-molecule therapeutics.
• Understanding Disease Mechanisms: Assesses the impact of mutations on protein-ligand interactions to identify and characterize disease-associated variants.
• Resistance Mutation Identification: Identifies potential resistance mutations that may reduce drug efficacy.
• Large-Scale Mutational Scanning: Supports systematic evaluation of numerous single amino acid substitutions on protein-ligand binding affinity.

Methodology:

Employs a structure-based machine learning framework and requires 3D protein-ligand complex structures; validated against benchmark datasets with predictive performance reported as comparable or superior to existing methods and with improved computational speed and reduced resource requirements.