A Bayesian approach to interpret time-resolved experiments using molecular simulations

Time-resolved experiments can provide unique insights into dynamical processes such as protein folding, ligand binding, and many other molecular processes. These experiments are, however, difficult to interpret at the molecular level because they generally report on signals that are averaged over multiple configurational states, and because they often report on processes that are well beyond what can be studied using most simulation methods. Here we describe an approach to use molecular simulations to model and interpret time-resolved experiments. The method, which we term trBME (time-resolved Bayesian/Maximum Entropy), is based on combining a model for the dynamics of the system, with 3D structural models of the protein that can be compared to the experiments. We illustrate the utility of the model using synthetic time-resolved small-angle X-ray scattering data and show how it can be used to extract detailed information from the experimental data on the process of protein unfolding.