In the forthcoming funding period we plan to develop methods to estimate complex kinetic models directly from single-molecule experimental data. In its spirit, our approach is related to Hidden Markov modeling but goes much beyond that and resolves some of the fundamental limitations of Hidden Markov modeling for the context of biophysical systems. The result of this approach will be a model of the conformational substates present in the data, their interconversion rates, and the distribution of values in the molecular observable(s) each conformation is associated with.