DNA²DNA Computations: A Potential "Killer App"?

Ever since Adleman's seminal paper [1] there has been a flood of ideas on how one could use DNA to compute. Lipton was the first to show that DNA could be used to solve more than just a variation of the famous travelling salesman problem [12]. Since then there have been many other papers on using DNA to solve various computational problems. [3,5,4,6,7,15] At the top level all these papers are. similar: they all at tempt to use DNA computation to solve some large search problem. Since a liter of water can hold 10 2-" bases of DNA, there is the possibility that one can outperform electronic machines. However, this is currently problematic. There are several reasons for this. First, electronic machines are very fast; moreover, they are getting faster every day. Second, there are many models of how to do DNA computations. Yet, it is unclear if any of these models wilt be practical. The problem is mainly that DNA technology is not perfect. DNA operations are not error free. Finally, there is the lack of a killer app. A killer app is an application that fits the DNA model; cannot be solved by the current or even future electronic machines; and is important. The latter is critical: to be a killer app the problem must be one for which people are willing to "pay money" for solutions. To date there are no viable candidates for the killer app. We propose a new way to use DNA computations. This way allows us to use DNA computations to solve important and potentially killer applications. The potential applications include: (1) DNA sequencing; (2) DNA fingerprinting; (3) DNA mutation detection or population screening; (~) Other fundamental operations on DNA. The key new idea is to use DNA computat ion to operate on unknown pieces of DNA. This is a fundamental change in the way that we use DNA computation. We call these DNA~-DNA computations: DNA to DNA computations. This idea was first proposed in [8] and called "analog" DNA computations there. The key idea is the following. Suppose that one has a test tube that contains multiple copies of some unknown strand X of DNA, By unknown we mean that we do not known