A solid-phase method for solving the Hamiltonian path problem (HPP) is described. The method employs only fast and simple DNA operations amenable to full automation. Single-stranded DNA molecules representing paths with no city visited twice are synthesized city-by-city from the start city on the surface of a solid support. The solution can thus be found in the computation time proportional to the number of cities. As well as the stepwise path synthesis, a pruning technique developed for the removal of looping paths helps the reduction of DNA molecules necessary for the computation; thus deenitely increasing the size of problems solvable on a DNA-based computer. Experiments using Adleman's seven-city instance of the HPP showed that the path extension cycle was very accurate and took only about 45 min. Our solid-phase method has originally been developed for solving the HPP, but it could also be applied to other problems requiring a massive parallelism in computation.
[1]
D M Crothers,et al.
Prediction of RNA secondary structure.
,
1971,
Proceedings of the National Academy of Sciences of the United States of America.
[2]
H. Blöcker,et al.
Predicting DNA duplex stability from the base sequence.
,
1986,
Proceedings of the National Academy of Sciences of the United States of America.
[3]
L M Adleman,et al.
Molecular computation of solutions to combinatorial problems.
,
1994,
Science.
[4]
R J Lipton,et al.
DNA solution of hard computational problems.
,
1995,
Science.
[5]
Anne Condon,et al.
Power of surface-based DNA computation
,
1997
.