Abstract
An efficient method is proposed for building Markov models with
discrete states able to accurately describe the slow relaxation of a
complex system with two stable conformations. First, the reaction
pathway described by a set of collective variables between the two
stable states is determined using the string method with swarms of
trajectories. Then, short trajectories are initiated at different
points along this pathway to build the state-to-state transition
probability matrix. It is shown, using a model system, how this
strategy makes it possible to use trajectories that are significantly
shorter than the slowest relaxation time to efficiently build a
reliable and accurate Markov model. Extensions of the method to
multiple pathways, as well as some common pitfalls arising from poorly
relaxed paths or an inappropriate choice of collective variables, are
illustrated and discussed. (C) 2008 American Institute of Physics.
Users
Please
log in to take part in the discussion (add own reviews or comments).