Single-molecule kinetic data often appear as stochastic time trajectories. These trajectories show the different states visited by the single molecule as a function of time. Many Markov model adaptations are widely used in interpreting this data though such models have limitations: (i) the underlying kinetic state topology is assumed in advance, (ii) transition times between microscopic states are assumed to be distributed exponentially, and (iii) they may waste data by forcing the data to fit a particular model. Here we show that Markov models can be formulated as a Non-Markov Memory-Kernel (NMMK) model in which the kinetics is captured by memory kernels, rather than by explicit states, from which a Markov model can be reconstructed. We extract the memory kernel with error bars, and thus the Markov model, from single molecule data using the method of image reconstruction which is grounded in the principle of Maximum Entropy. We show how this method predicts models without prespecifying a topology or assuming Markovianty using both simulated and real data.

Using the full single molecule data set to extract a kinetic model with minimal bias

243rd National Meeting (2012)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.