Robust Free Energy Perturbation Protocols for Creating Molecules in Solution.

Area of Science:

• Computational chemistry
• Biochemistry
• Statistical mechanics

Background:

• Accurate free energy estimation is vital for condensed-phase problems in chemistry and biochemistry.
• Free energy perturbation (FEP) and thermodynamic integration (TI) are common methods used with molecular dynamics (MD) and Monte Carlo (MC) simulations.
• Calculating absolute free energies of solvation or binding often involves creating or annihilating solutes in simulations.

Purpose of the Study:

• To identify optimal protocols for calculating the free energy of creation/annihilation of organic molecules in aqueous solution.
• To investigate the efficiency and convergence of FEP calculations for molecular creation versus annihilation.
• To determine the minimum number of simulation windows required for accurate free energy estimates.

Main Methods:

• Decoupling of electrostatic and Lennard-Jones interactions was performed.
• Free energy perturbation (FEP) calculations were employed for molecular creation and annihilation.
• Simulations were conducted using varying numbers of "λ-windows" to assess convergence.

Main Results:

• FEP calculations for molecular creation are significantly more efficient than annihilation.
• As few as four windows can be adequate for creation calculations of small to medium-sized organic molecules.
• For a larger druglike molecule (MIF180), 10 creation windows yielded accurate hydration free energy.
• Convergence is hindered by sampling in the annihilation direction, negating the need for postprocessing methods like Bennett acceptance ratio (BAR).

Conclusions:

• Optimized protocols using molecular creation simulations can efficiently yield accurate free energies of hydration.
• The number of simulation windows required is dependent on molecular size and complexity.
• Avoiding annihilation sampling simplifies protocols and improves convergence for free energy calculations.