Accelerating the 3D reference interaction site model theory of molecular solvation with treecode summation and...

Area of Science:

• Computational chemistry
• Molecular modeling
• Biophysics

Background:

• The 3D reference interaction site model (3D-RISM) is crucial for molecular solvation studies.
• Calculating solvent interactions for large molecules like proteins is computationally intensive.
• Current 3D-RISM methods face performance limitations due to long-range interaction calculations.

Purpose of the Study:

• To develop and implement computationally efficient methods for 3D-RISM calculations.
• To accelerate the computation of potential energy and long-range asymptotics in non-periodic 3D-RISM.
• To enable the study of solvation in large, complex molecular systems.

Main Methods:

• Implemented treecode summation for long-range interactions.
• Applied analytically corrected cut-offs for short-range interactions.
• Integrated new methods into the AmberTools molecular modeling suite for non-periodic 3D-RISM.

Main Results:

• Achieved a 4-fold reduction in computation time for large proteins like tubulin.
• Demonstrated near-linear scaling for parallel calculations with the new methods.
• Successfully applied the accelerated 3D-RISM to a large microtubule system (910 tubulin dimers).

Conclusions:

• The developed methods significantly enhance the computational efficiency of 3D-RISM.
• These advancements facilitate the analysis of solvation thermodynamics and density distributions in large biomolecular systems.
• The approach is effective for studying complex biological structures, such as microtubules.