When does molecular dynamics improve RNA models? Insights from CASP15 and practical guidelines.

Area of Science:

• Biomolecular modeling
• Computational biology
• Structural bioinformatics

Background:

• Molecular dynamics (MD) simulations are increasingly used for refining biomolecular models.
• The utility of MD for RNA structure prediction requires systematic evaluation.

Purpose of the Study:

• To benchmark the impact of MD simulations on RNA models within the CASP15 framework.
• To determine optimal MD simulation parameters for RNA structure refinement.

Main Methods:

• Systematic benchmarking of MD simulation effects on 61 RNA models from CASP15.
• Utilized Amber software with the RNA-specific χOL3 force field.
• Evaluated models subjected to short (10-50 ns) and long (>50 ns) simulations.

Main Results:

• Short MD simulations (10-50 ns) provided modest improvements for high-quality starting models, enhancing stability of stacking and non-canonical base pairs.
• Poorly predicted models showed little to no benefit and often deteriorated after MD simulations.
• Longer simulations (>50 ns) generally led to structural drift and reduced model fidelity.

Conclusions:

• MD simulations are most effective for fine-tuning reliable RNA models and assessing their stability.
• MD is not a universal corrective method for RNA structure prediction errors.
• Guidelines are provided for selecting input models and optimal simulation lengths for RNA refinement.