Enhanced Sampling in the Well-Tempered Ensemble

We introduce the well-tempered ensemble (WTE) which is the biased ensemble sampled by well-tempered metadynamics when the energy is used as collective variable. WTE can be designed so as to have approximately the same average energy as the canonical ensemble but much larger fluctuations. These two properties lead to an extremely fast exploration of phase space. An even greater efficiency is obtained when WTE is combined with parallel tempering. Unbiased Boltzmann averages are computed on the fly by a recently developed reweighting method [M. Bonomi et al. J. Comput. Chem. 30, 1615 (2009)]. We apply WTE and its parallel tempering variant to the 2d Ising model and to a Go-model of HIV protease, demonstrating in these two representative cases that convergence is accelerated by orders of magnitude.

💡 Research Summary

The paper introduces the Well‑Tempered Ensemble (WTE), a biased statistical ensemble that naturally arises when the energy of a system is used as the collective variable (CV) in well‑tempered metadynamics (WT‑MetaD). In WT‑MetaD a time‑dependent bias potential V(s,t) is deposited on the CV s, with a tempering factor γ > 1 that controls how quickly the bias “flattens” the free‑energy surface. When the CV is the potential energy E, the resulting stationary distribution has the same average energy as the canonical (Boltzmann) ensemble but its variance is amplified by a factor γ. Consequently, the system explores both low‑energy and high‑energy regions much more frequently, while preserving the correct average thermodynamic quantities. This property makes WTE similar in spirit to multicanonical or umbrella‑sampling approaches, but it requires no a‑priori knowledge of the density of states because the bias is built adaptively during the simulation.

A key insight of the work is that the enlarged energy fluctuations dramatically accelerate the crossing of free‑energy barriers. The authors further combine WTE with Parallel Tempering (PT), a replica‑exchange technique in which several copies of the system are simulated at different temperatures and periodically attempt swaps. In a PT‑WTE scheme each replica is run under the WTE bias, and the acceptance probability for a swap between replicas i and j becomes exp