Given an arbitrary set of cities, what is the shortest route that links all the cities, passing through each one only once? With a small number of cities, it is an easy problem to solve; limit it to five cities and you can doodle out the answer on a piece of paper. But as the number of cities grows, the problem becomes increasingly difficult to answer, and by the time you get 30 cities, there are more than a billion possibilities. At 100 cities, it would take the fastest supercomputer more than a billion years to dig up the answer. The difficulty lies in the fact that a traditional computer has to crunch through all possibilities one at a time. However, with a DNA computer, these doubts can be solved easily. The deoxyribonucleic acid plows through all of the combinations more or less simultaneously. The trick lies in creating the right set of DNA strands to start the process, and then weeding out the wrong answers.

Leonard Adleman invented the DNA computer in 1994. According to legend, the University of Southern California computer scientist and mathematician was inspired by reading Molecular Biology of the Gene, a textbook written by James Watson, a co-discoverer of the structure of DNA. Adleman tested out the viability of the DNA computer by solving the Hamilton path question for seven cities. His test-tube computer created all of the possible Hamilton paths in just seconds, but it took him a few weeks to filter out the correct answer. DNA has the ability to store and manipulate information. So, genes are controlled by small DNA sequence elements that run up and down the gene, like beads on a string, making copies that can be introduced into different contexts to arrive at an identical output.

The initial reaction to Adleman's publication was mixed. Although everyone agreed that his ideas were revolutionary, many scientists questioned whether Adleman's small demonstration experiment could be scaled up to solve real-world problems, and whether any practical problems existed that needed DNA computation to solve. These attacks were premature, and numerous scientists were soon proposing solutions to both the problem of the quantity of DNA needed and of what to do with DNA computers. Using Adleman's initial technique, a huge quantity of DNA would be required to solve practical problems because there is no easy way to custom-make the DNA strands. The DNA strands must be produced in all possible combinations, and the required strands must be laboriously extracted from the brew. Scientists have already developed a number of promising theories to make DNA-based computation workable. One of the most interesting proposals, the "sticker model," dispenses with the problem of having to make large quantities of custom DNA. This method would involve the use of short pieces of DNA, the stickers, which would be attached to longer "memory" strands of DNA. This method mimics conventional digital computers in that a memory strand with a sticker attached would correspond to a 1, and a strand without a sticker would correspond to a 0.