Benchmarking predictive methods for small-angle X-ray scattering from atomic coordinates of proteins using maximum...

Area of Science:

• Biophysics
• Structural Biology
• Computational Biology

Background:

• Small-angle scattering (SAS) is crucial for determining protein structure in solution.
• Accurate modeling of protein hydration is essential for interpreting SAS data.
• Previous consensus profiles were generated from limited datasets.

Purpose of the Study:

• To generate improved consensus SAS profiles using a larger dataset and a maximum likelihood protocol.
• To compare the accuracy of different hydration models in fitting experimental SAS data.
• To evaluate the reliability of implicit hydration models versus all-atom molecular dynamics (MD) simulations.

Main Methods:

• A round-robin exercise generated extensive SAS data for five proteins.
• Maximum likelihood analysis was applied to a larger dataset for improved consensus profile generation.
• Fits of consensus profiles were investigated using atomic coordinates with implicit hydration models and all-atom MD simulations.

Main Results:

• Implicit hydration models, with adjustable parameters, showed good fits but could mask structural details.
• All-atom MD simulations, without adjustable parameters, yielded comparable or better fits.
• MD simulations accurately accounted for solvent composition (ions) and thermal fluctuations, improving radius of gyration predictions.

Conclusions:

• Adjustable parameters in implicit hydration models can obscure real structural differences.
• All-atom MD simulations are more robust against false positives and provide a more accurate representation of hydration.
• While slower, MD simulations offer superior accuracy for SAS profile interpretation, especially when accounting for solvent composition.