Improving the Accuracy of Protein Thermostability Predictions for Single Point Mutations.

Area of Science:

• Computational Biology
• Protein Engineering
• Biophysics

Background:

• Accurate prediction of protein thermostability changes from single point mutations is vital for understanding diseases and industrial protein engineering.
• Free Energy Perturbation (FEP) is a long-standing method for predicting protein stability changes but has limitations.
• Existing FEP methods struggle with proline mutations and charge-changing mutations, hindering industrial applications.

Purpose of the Study:

• To extend the FEP+ protocol for accurate modeling of proline and charge-changing mutations.
• To evaluate the impact of the unfolded model on protein stability calculations.
• To improve the accuracy of *in silico* prediction of protein thermostability changes.

Main Methods:

• Extended the FEP+ protocol to include proline and charge-changing mutations.
• Evaluated the influence of the unfolded model using native sequence and conformation peptides.
• Applied the improved FEP+ protocol to a dataset of 87 mutations across five proteins.

Main Results:

• Achieved a mean unsigned error of 0.86 kcal/mol and a root mean square error of 1.11 kcal/mol for 87 protein mutations.
• The enhanced FEP+ protocol demonstrated significantly improved accuracy compared to previous FEP studies.
• Prediction accuracy is comparable to state-of-the-art small-molecule relative binding affinity calculations.

Conclusions:

• The improved FEP+ protocol accurately predicts protein stability changes for proline and charge mutations.
• This advancement enhances the reliability of *in silico* methods for protein engineering and disease research.
• The enhanced FEP+ method shows potential for driving protein engineering discovery projects.