Enhanced Sampling with Sub-optimal Collective Variables: Reconciling Accuracy and Convergence Speed

Abstract

We introduce an enhanced sampling algorithm to obtain converged free energy landscapes of molecular rare events, even when the collective variable (CV) used for biasing is not optimal. Our approach is a combination of the On-the-fly probability enhanced sampling (OPES) and its exploratory variant, OPES Explore (OPESe). We demonstrate the successful application of this combined algorithm on the two-dimensional Wolfe-Quapp potential, ligand-receptor binding in trypsin-benzamidine complex, and folding-unfolding of chignolin. Apart from computing accurate free energy profiles, we can discover additional metastable configurations not distinguished by the sub-optimal CV space. Moreover, we can control the trade-off between accuracy and convergence speed by varying the ratio of the barrier parameters in OPES and OPESe components. The improved efficiency and accuracy of free energy calculation, and the possibility of using generic and intuitive collective variables, make our proposed algorithm particularly promising for the simulation of complex molecular systems.