Integrated Boltzmann Sampling: A Few-State Approach for Efficient Multistate Free Energy Calculations

Area of Science:

• Computational chemistry
• Molecular modeling
• Biophysics

Background:

• Free energy calculations are crucial for computational chemistry tasks like drug discovery.
• Current methods are computationally expensive due to extensive sampling requirements.

Purpose of the Study:

• Introduce Integrated Boltzmann Sampling (IBS), a novel few-state framework.
• Reduce the computational cost and time for free energy calculations.

Main Methods:

• IBS integrates multistate thermodynamic sampling into artificial ensembles.
• Trajectories are reweighted to recover thermodynamic information from fewer states.
• Reduces sampling cost from K·S to (1 - ϵ + Kϵ)·S.

Main Results:

• IBS achieved accuracy comparable to replica-based methods on SAMPL6 and FX receptor benchmarks.
• Reduced computational wall time by approximately 50-60%.

Conclusions:

• Chemically accurate free energy predictions do not necessitate exhaustive replica sampling.
• IBS provides an efficient alternative for applications requiring accurate free energy differences.