Efficient use of non-equilibrium measurement to estimate free energy differences for molecular systems

Abstract

A promising method for calculating free energy differences Delta F is to generate non-equilibrium data via ``fast-growth'' simulations or experiments -- and then use Jarzynski's equality. However, a difficulty with using Jarzynski's equality is that Delta F estimates converge very slowly and unreliably due to the nonlinear nature of the calculation -- thus requiring large, costly data sets. Here, we present new analyses of non-equilibrium data from various simulated molecular systems exploiting statistical properties of Jarzynski's equality. Using a fully automated procedure, with no user-input parameters, our results suggest that good estimates of Delta F can be obtained using 6-15 fold less data than was previously possible. Systematizing and extending previous work [1], the new results exploit the systematic behavior of bias due to finite sample size. A key innovation is better use of the more statistically reliable information available from the raw data.