Predicting hydration free energies using all-atom molecular dynamics simulations and multiple starting conformations

Area of Science:

• Computational chemistry
• Physical chemistry

Background:

• Predicting hydration free energies is crucial for understanding molecular behavior in solution.
• Molecular dynamics (MD) simulations offer a powerful computational approach for these predictions.

Purpose of the Study:

• To evaluate the accuracy of MD simulations in predicting hydration free energies for small organic molecules.
• To assess the performance of different solvent models and identify areas for force field improvement.

Main Methods:

• Molecular dynamics simulations in explicit solvent.
• Utilized TIP3P and TIP4P-Ew solvent models.
• Applied umbrella sampling to capture conformational changes in water.

Main Results:

• Achieved good agreement with experimental hydration free energies (RMS error of 2.82 kcal/mol).
• Identified systematic force field errors for hydroxyl-rich compounds.
• Observed minimal differences between TIP3P and TIP4P-Ew solvent models (RMS difference of 0.64 kcal/mol).
• Found that conformational changes of carboxylic acids in water were not adequately sampled by standard MD.

Conclusions:

• MD simulations provide reliable hydration free energy predictions for many organic molecules.
• Force field parameters require refinement for specific compound classes, particularly hydroxyl-rich molecules.
• Advanced sampling techniques like umbrella sampling are necessary to accurately model conformational changes in solution.