MAESTRO

MAESTRO predicts changes in protein stability (ΔΔG) caused by point mutations using structure-based analysis of wild-type three-dimensional protein structures (PDB), applicable to monomeric, multimeric proteins and PDB-defined biological assemblies.

Key Features:

• Structure-based analysis: Requires the three-dimensional structure of the wild-type protein (PDB) for stability prediction.
• Multi-Agent Machine Learning System: Employs a multi-agent machine learning approach to integrate predictive models for stability changes.
• Predicted Free Energy Change (ΔΔG) and Confidence Estimation: Outputs predicted ΔΔG values for mutations alongside an associated confidence estimate.
• High-throughput scanning for multi-point mutations: Performs high-throughput scans with user-controlled mutation sites and types for multi-point scenarios, including scans of n-point combinations up to n = 5.
• Mutation sensitivity profiles: Generates mutation sensitivity profiles to summarize the impact of substitutions across sites.
• Stabilizing disulfide bonds prediction mode: Includes a specific mode to predict potential stabilizing disulfide bonds.
• Versatility across protein types: Handles single-point and multi-point mutations and evaluates effects across monomeric, multimeric and biological assembly structures.

Scientific Applications:

• Protein engineering and design: Guides modification of protein properties by predicting stability effects of candidate mutations.
• Mutation impact and disease research: Aids investigation of how point mutations can alter protein stability and contribute to dysfunction or disease.
• Disulfide bond engineering: Supports evaluation of potential disulfide bonds to improve protein folding and stability.

Methodology:

Structure-based analysis of wild-type 3D protein structures (PDB); a multi-agent machine learning system for prediction; high-throughput scanning of defined mutation sites and types (including multi-point up to n = 5); a dedicated mode for predicting stabilizing disulfide bonds; outputs of ΔΔG and prediction confidence estimates.