DNA has been used in vitro as a computational substrate due to programmable base-pairing interactions. This allows the construction of logic gates, self-assembled shapes, motors, walkers and other nanoscale devices comprised of DNA strands. These strands must be carefully designed in order to correctly perform their function, and this requires good models of the DNA system. The thermodynamics of multiple interacting DNA strands is a well-studied model which can make equilibrium predictions on these systems, but many devices operate far from equilibrium. The prediction of large non- equilibrium systems requires the simulation of a kinetics model, as these are both stochastic and computationally difficult to solve in an exact form. We developed the Multistrand kinetics simulator, which extends the previous work by expanding the kinetics model to handle multiple nucleic acid strands, as well as using optimized algorithms in order to speed up the simulations. We examine the extra parameters found in the expanded kinetics model, and evaluate whether we can predict known experiments with a good choice of parameters.

Joseph Schaeffer is a Graduate Student in Computer Science in Erik Winfree's group at Caltech. He received a B.S. in Engineering and Applied Science from Caltech in 2002. His current research is in the optimization of algorithms for the simulation of DNA folding kinetics.