PSP-GNM

PSP-GNM predicts changes in protein stability caused by point mutations by modeling protein dynamics with a coarse-grained Gaussian Network Model and estimating unfolding ΔΔG.

Key Features:

• Gaussian Network Model: Utilizes a coarse-grained Gaussian Network Model to simulate protein dynamics with residue interactions weighted by the Miyazawa-Jernigan statistical potential.
• ΔΔG calculation: Computes unfolding Gibbs free energy change (ΔΔG) by simulating partial unfolding of both wildtype and mutant structures and calculating differences in energies and entropies.
• Benchmarking performance: Evaluated on datasets S350 (350 forward mutations), S669 (669 forward and reverse mutations), and S611 (611 forward and reverse mutations) with Pearson correlation up to 0.61 against experimental ΔΔG.
• Experimental condition sensitivity: Shows improved agreement with experimental ΔΔG at temperatures near 25 °C and at neutral pH.
• Antisymmetry assessment: Tested on Ssym+ (352 forward and reverse mutations) and exhibited near-perfect antisymmetry with a Pearson correlation of -0.97 between reverse and forward ΔΔG.
• Implementation: Implemented in Python.

Scientific Applications:

• Mutation Impact Analysis: Quantifies how specific amino acid substitutions alter protein stability for studies of missense mutation effects.
• Protein Engineering: Predicts stability changes of engineered mutations to inform design of proteins with desired stability properties.
• Structural Biology: Provides computational insight into protein folding and stability landscapes to aid interpretation of experimental data.

Methodology:

Applies a coarse-grained Gaussian Network Model with interactions weighted by the Miyazawa-Jernigan statistical potential, simulates partial unfolding of wildtype and mutant structures, and computes ΔΔG from differences in energies and entropies.