Automating the DNA computer: solving n-Variable 3-SAT problems
- 79 Downloads
In the decade since the first molecular computation was performed, it has been shown that DNA molecules can perform sophisticated, massively parallel computations avoiding the Von Neumann bottleneck. However, progress in the field has been slow. The largest problem solved to date is an instance of the 20-variable 3-CNF SAT problem. Performing the computation took more than two man-weeks to complete because every aspect of the computation was performed by hand. Molecular computations are extremely labor intensive and error prone–automation is necessary for further progress.
The next step, (the second generation DNA computer—that of taking the laborious, laboratory bench protocols performed by hand, and automating them), has been achieved with the construction of an automated DNA computer dubbed EDNAC. It employs the same paradigm that was used to solve the labor-intensive instance of the 20-variable 3-CNF SAT problem. Using a combinatorial DNA library and complementary probes, EDNAC solves instances of the n-variable 3-CNF SAT problem. A 10 variable instance of the 3-CNF SAT problem was essayed. The computation took 28 h to perform. EDNAC correctly computed nine of the 10 variables, with a tenth variable remaining ambiguous. This result is comparable to current results in the molecular computation community. This research tested the critical properties, such as complexity, robustness, reliability, and repeatability necessary for the successful automation of a molecular computer.
KeywordsDNA Computation Molecular Computation Natural Computing SAT Computation
I would like to thank friends, and colleagues at the USC Laboratory for Molecular Science for their support, in particular Len Adleman, Rebecca Anderson, Ravi Braich, Nickolas Chelyapov, Ramon Del Gadillo, Victor Jordan, Howard Lukefahr, Jim Merritt, Dustin Reishus, Paul Rothemund, Bilal Shaw, Areio Soltani, and Don Wiggins.
- Braich R, Johnson C, Rothemund PWK, Hwang D, Chelyapov N, Adleman L (2000) Satisfiability problem on a gel based DNA computer DNA computing—DNA 6 2000, vol. 2054. Springer-Verlag, New YorkGoogle Scholar