This special topic highlights recent developments in enhanced sampling methods for molecular-level simulations of chemical and biological systems. These methods are designed to enable more efficient exploration of phase space and extend the time scales that can be explored by simulations.
Enhanced sampling methods nowadays are routinely deployed to perform molecular simulations in solid- and liquid-state physics, biophysics, materials science, and chemistry. These approaches are improving the ability of computer simulations to be predictive and useful because they have the potential to provide quantitative information on processes that would just not happen in a brute force molecular dynamics or Monte Carlo run. Enhanced methods are evolving constantly to meet new challenges and are the object of intense theoretical investigation worldwide.
The present special topic provides a broad sample of recent developments in enhanced sampling techniques, covering a wide variety of methods and application areas. Some articles compare different methods applied to the same systems and/or provide a critical analysis of existing strategies. Other articles introduce novel approaches or more powerful variants of old algorithms. A truly impressive range of systems is studied—complex fluids, peptides, protein-ligand complexes, and water are just a few examples. These systems are studied using a variety of levels of description, ranging from ab initio to coarse-grained.
Some of the earliest strategies, replica exchange and simulated tempering, are still the object of intense investigation. The classical idea of exchanging among temperatures has been generalized in numerous ways. The special topic section includes studies which further generalize and optimize the approach, developing exchange schemes for subsystems, thermodynamic states, and different thermodynamic ensembles.
Equilibrium sampling is pursued by a range of additional strategies. Very modern approaches incorporating methodology from machine learning and Bayesian statistics are presented, in addition to variants on long-established methods such as the Wang-Landau and multicanonical approaches. Coarse-grained models are also exploited. In some cases, existing approaches are combined to achieve higher efficiency.
Related to the primary focus of the special topic, a number of articles describe methods for calculating non-equilibrium properties (e.g., kinetic rates), exploiting the forward flux approach, metadynamics, and the weighted ensemble method.
We hope this collection of articles will provide readers with an up-to-date snapshot of some of the most important methods used to address sampling in a broad spectrum of molecular systems.