Does Hamiltonian Replica Exchange via Lambda-Hopping Enhance the Sampling in Alchemical Free Energy Calculations?

Abstract

In the context of computational drug design, we examine the effectiveness of the enhanced sampling techniques in state-of-the-art free energy calculations based on alchemical molecular dynamics simulations. In a paradigmatic molecule with competition between conformationally restrained E and Z isomers whose probability ratio is strongly affected by the coupling with the environment, we compare the so-called λ-hopping technique to the Hamiltonian replica exchange methods assessing their convergence behavior as a function of the enhanced sampling protocols (number of replicas, scaling factors, simulation times). We found that the pure λ-hopping, commonly used in solvation and binding free energy calculations via alchemical free energy perturbation techniques, is ineffective in enhancing the sampling of the isomeric states, exhibiting a pathological dependence on the initial conditions. Correct sampling can be restored in λ-hopping simulation by the addition of a “hot-zone” scaling factor to the λ-stratification (FEP+ approach), provided that the additive hot-zone scaling factors are tuned and optimized using preliminary ordinary replica-exchange simulation of the end-states.

Comparison of Equilibrium and Nonequilibrium Approaches for Relative Binding Free Energy Predictions

“…An essential corollary is that we cannot rely on approaches based on one-off simulations despite the fact that many authors continue to report them. Enhanced sampling methods, such as REST2 (replica exchange with solute scaling) as implemented in the free energy perturbation approach FEP+ by Schrödinger Inc., have been shown to be unreliable. − While the oft-vaunted FEP+ method continues to be reported in the context of one-off simulations and, indeed, one still reads many papers in which one or a small number of “repeats” are performed, a growing number of authors now recognize that ensembles are, in fact, essential. , …”

Dealing with Induced Fit, Conformational Selection, and Secondary Poses in Molecular Dynamics Simulations for Reliable Free Energy Predictions

“…However, in many cases, this time span can be insufficient for obtaining statistically converged simulations since we are dealing with a single-molecule system and not with a thermodynamic ensemble. Unconverged runs on single complexes typically plague the final outcome of the prediction using FEP or TI-based alchemical approaches for binding free energy calculations. , The sampling problem is patently exposed when results are found to critically depend on similar starting conditions so that in general replicates of costly alchemical free energy calculations must be often carried to deliver a reliable prediction with a credible confidence interval. − …”

“…Unconverged runs on single complexes typically plague the final outcome of the prediction using FEP or TI-based alchemical approaches for binding free energy calculations. 34 , 35 The sampling problem is patently exposed when results are found to critically depend on similar starting conditions so that in general replicates of costly alchemical free energy calculations must be often carried to deliver a reliable prediction with a credible confidence interval. 36 − 40 …”

NE‐RDFE: A protocol and toolkit for computing relative dissociation free energies with GROMACS between dissimilar molecules using bidirectional nonequilibrium dual topology schemes

“…FEP+, while using simultaneously the alchemical and the hotzone scaling factors in the generalized ensemble, is nonetheless implemented as a one-dimensional HREM with no scaling at the end-states. 16 As such, the methodology may suffer from limitations such a sluggish convergence 17,18 or poor enhanced sampling capabilities in challenging systems. 16,19,20 Also, transmutations are usually limited to the so-called R-group perturbations, where a single substituent or atom is transformed into another substituent or atom.…”

“…16 As such, the methodology may suffer from limitations such a sluggish convergence 17,18 or poor enhanced sampling capabilities in challenging systems. 16,19,20 Also, transmutations are usually limited to the so-called R-group perturbations, where a single substituent or atom is transformed into another substituent or atom. FEP-based RDFE generally do not involve scaffold-hopping or ring breaking 21,22 and transmutations between congeneric molecules differing by more than one substituent are generally carried out through the so-called perturbation graphs, using uninteresting connecting intermediate ligands.…”