SAMPL7 blind predictions using nonequilibrium alchemical approaches.

Area of Science:

• Computational Chemistry
• Molecular Modeling
• Supramolecular Chemistry

Background:

• Accurate prediction of binding free energies is crucial for molecular design.
• Host-guest chemistry, particularly with cavitands and cyclodextrins, is a key area in supramolecular chemistry.
• The SAMPL (Statistical Assessment of Modeling for Predicting Ligand Affinities) challenge benchmarks computational methods.

Purpose of the Study:

• To compute the standard binding free energy for host-guest systems involving Isaac's TrimerTrip and Gilson's Cyclodextrin derivatives.
• To evaluate the performance of a non-equilibrium (NE) alchemical technique within the SAMPL7 challenge.
• To assess the accuracy and limitations of the GAFF2 non-polarizable force field for these systems.

Main Methods:

• Employed a non-equilibrium (NE) alchemical technique combining enhanced sampling molecular dynamics simulations.
• Utilized driven fast out-of-equilibrium alchemical trajectories to recover free energy via Jarzynski and Crooks NE theorems.
• Parametrized systems using the GAFF2 non-polarizable force field.

Main Results:

• Calculated standard binding free energies for two series of host-guest systems.
• Achieved acceptable performance, comparable in accuracy to previous SAMPL6 submissions.
• Identified specific instances where the GAFF2 non-polarizable force field exhibited deficiencies.

Conclusions:

• The adopted computational approach, utilizing NE alchemical techniques, is reliable for predicting binding free energies.
• The study confirmed the robustness of the computational strategy across different host-guest systems.
• The findings underscore the need for force field refinement, particularly for complex host-guest interactions.