Enhanced Sampling\nwith Suboptimal Collective Variables:\nReconciling Accuracy and Convergence Speed

We introduce an enhanced sampling algorithm to obtain\nconverged\nfree energy landscapes of molecular rare events, even when the collective\nvariable (CV) used for biasing is not optimal. Our approach samples\na time-dependent target distribution by combining the on-the-fly probability\nenhanced sampling and its exploratory variant, OPES Explore. This\npromotes more transitions between the relevant metastable states and\naccelerates the convergence speed of the free energy estimate. We\ndemonstrate the successful application of this combined algorithm\non the two-dimensional Wolfe–Quapp potential, millisecond time-scale\nligand–receptor binding in the trypsin–benzamidine complex,\nand folding–unfolding transitions in chignolin mini-protein.\nOur proposed algorithm can compute accurate free energies at an affordable\ncomputational cost and is robust in terms of the choice of CVs, making\nit particularly promising for the simulation of complex biomolecular\nsystems.