Protein Structure Prediction from NMR Hydrogen-Deuterium Exchange Data

Area of Science:

• Structural biology
• Computational biology
• Biophysics

Background:

• Amide hydrogen-deuterium exchange (HDX) provides insights into protein flexibility and binding sites.
• HDX-NMR offers higher throughput than traditional structure determination methods like X-ray crystallography or cryo-EM.
• HDX data inherently contains structural information, making it suitable for computational integration.

Purpose of the Study:

• To develop and validate a computational methodology for incorporating HDX-NMR data into *ab initio* protein structure prediction.
• To assess the impact of HDX-NMR data on the accuracy of predicted protein models using the Rosetta software framework.
• To establish a confidence metric for identifying near-native protein structures predicted with HDX-NMR data.

Main Methods:

• Developed a computational approach to integrate HDX-NMR data into Rosetta for *ab initio* protein structure prediction.
• Scored 38 proteins with available HDX-NMR data, comparing predictions with and without HDX data incorporation.
• Validated the method on two additional proteins and developed a confidence metric for model selection.

Main Results:

• Incorporating HDX-NMR data improved the root-mean-square deviation (rmsd) of predicted models by an average of 1.42 Å.
• Significant rmsd improvements (average 3.63 Å) were observed in models where HDX data altered selection, with some improving over 11 Å.
• The developed confidence metric successfully identified near-native models, even without a known native structure.

Conclusions:

• Integrating HDX-NMR data into *ab initio* protein structure prediction using Rosetta substantially enhances model accuracy.
• The method provides a powerful and computationally efficient tool for predicting protein structures and identifying high-quality models.
• HDX-NMR data integration offers a valuable complement to experimental structure determination, improving prediction confidence and accuracy.