Optically Programming DNA Computing in Microflow Reactors

John S. McCaskill

GMD - National Research Center for Information Technology, Schloß Birlinghoven, D-53754 St. Augustin, Germany

e-mail: john.mccaskill@gmd.de

DNA Computing became an experimental subject in 1994 when L. Adleman demonstrated a simple in vitro procedure for finding a solution to a "hard" combinatorial problem. Since then, DNA Computing has been a rapid growth field, especially in spawning new theoretical models for universal computation. Natural and artificial variants of DNA expand the technical range of DNA Computing which is attractive for computation because of its massive parallelism (1015/ml) and low energy dissipation. Practically though, the time scale of molecular recognition, side reactions, DNA-DNA recognition errors and manual manipulations currently limit large scale computations. Progress in resolving some of these difficulties will be discussed, including our own approach in microflow reactors. In particular, a strategy for purely optical programming of DNA Computing will be presented, using the example of the Maximal Clique Problem, along with first results in this direction. While more sophisticated algorithms than brute search can and must be implemented to attack larger combinatorial problems, pure computation is not the only or most interesting goal of DNA Computing.