Fundamental Redesign of the TIGER2hs Kernel to Address Severe Parameter Sensitivity

Area of Science:

• Computational chemistry
• Molecular dynamics simulations
• Biophysics

Background:

• Replica exchange molecular dynamics (REMD) is popular for enhancing conformational sampling in molecular systems.
• Applications include protein folding, protein-protein interactions, and binding free energy calculations.
• REMD methods yield accurate results but are computationally expensive.

Purpose of the Study:

• To introduce TIGER2hs, an improved version of the temperature intervals with global exchange of replicas (TIGER2) algorithm.
• To address sampling imbalance and parameter sensitivity issues in previous TIGER2 algorithms.
• To enable accurate enhanced sampling with reduced computational resources.

Main Methods:

• Developed TIGER2hs, combining replica-based enhanced sampling in explicit solvent with hybrid solvent energy evaluation.
• Replaced bulk water with an implicit solvent model during exchange attempts.
• Implemented a pairwise exchange kernel (PE) and controlled average temperature change per exchange (⟨ΔT/χ⟩) below 30 K.

Main Results:

• TIGER2hs requires significantly fewer replicas compared to REMD for similar sampling efficiency.
• The pairwise exchange kernel eliminated sensitivity to replica number and maximum temperature.
• Adjusting ⟨ΔT/χ⟩ below 30 K ensures controlled temperature mixing, mimicking REMD.
• Results demonstrated excellent agreement with REMD for three different peptides.

Conclusions:

• TIGER2hs offers a robust and computationally efficient alternative to REMD for enhanced conformational sampling.
• The method provides accurate results with flexible resource investment.
• The pairwise exchange kernel and controlled temperature change enhance the reliability of TIGER2 simulations.