Structure and ensemble refinement against SAXS data: Combining MD simulations with Bayesian inference or with the...

Area of Science:

• Biophysics
• Structural Biology
• Computational Biology

Background:

• Small-angle X-ray scattering (SAXS) is crucial for studying conformational changes in proteins and soft-matter complexes in solution.
• SAXS data interpretation faces challenges due to low spatial resolution and limited information content, increasing the risk of overinterpretation.

Purpose of the Study:

• To integrate SAXS data with all-atom molecular dynamics (MD) simulations for deriving accurate atomic structures and heterogeneous ensembles.
• To reduce the risk of overinterpreting SAXS data by incorporating physicochemical information from MD simulations.

Main Methods:

• Utilized SAXS data integration with all-atom MD simulations.
• Implemented SAXS-driven parallel-replica ensemble refinement and Bayesian structure refinement.
• Employed GROMACS-SWAXS, a modified GROMACS software version, for SAXS-driven MD simulations.

Main Results:

• Developed a method to derive atomic structures and heterogeneous ensembles compatible with SAXS data.
• Demonstrated that MD simulations add crucial physicochemical context, mitigating SAXS data overinterpretation.
• Provided practical guidelines for conducting and interpreting SAXS-driven MD simulations.

Conclusions:

• SAXS-driven MD simulations offer a powerful approach to obtain atomistic insights into protein and complex dynamics.
• The integration enhances the reliability of structural interpretations from SAXS experiments.
• Freely available methods (GROMACS-SWAXS) facilitate broader application in structural biology research.