Reconciling transition path time and rate measurements in reactions with large entropic barriers

Recent experiments and simulation studies showed that protein/DNA folding barriers inferred from folding rates or from potentials of mean force are often much higher than the barriers estimated from the distributions of transition path times. Here a toy model is used to explain a possible origin of this effect: It is shown that when the transition in question involves an entropic barrier, the one-dimensional Langevin model commonly used to interpret experimental data, while adequately predicting the transition rate, fails to describe the properties of the subset of the trajectories that form the transition path ensemble; the latter may still be describable in terms of a one-dimensional model, but with a different potential, just as observed experimentally.