Conclusion  

This document constitutes a well thought-out proposal for thesis research in several important respects. First, it presents a research plan that is quite feasible, because it is tightly focused on a very concrete goal, namely that of getting CMM working to the point of being able to execute simple machines, and also because CMM itself is designed to be a very realistic approach to computation with DNA, since it does not require any new lab techniques beyond ones that are already in standard use in molecular biology labs such as the one in which we are working, and since the amount of manual steps required to execute our DNA machines is kept to a minimum. In addition, much work has already been accomplished, namely researching the relevant literature (appendix D), designing particular DNA machines (appendix B), creating a CMM simulator to test them (appendix C), and putting together details of some of the experimental protocols involved in their execution (appendix A).

Second, the research proposed here would make a significant contribution to science, because it is of inherent scientific interest to show that a system of biomolecules capable of universal computation can actually be made to work. It is of interest to computer scientists because it shows that general-purpose molecular computation is possible, and it illustrates some important theoretical issues involving logical reversibility and the thermodynamics of computation. It is of interest to biologists, because it shows that life might potentially, at some point during evolution or development, perform arbitrarily complex computations in this way. Also, apart from the pure scientific interest, there may be eventual practical applications, perhaps involving the use of nondeterministic algorithms running under CMM to solve large search problems with a high degree of parallelism, or perhaps involving the use of CMM in living cell lines to cause their genes to change in desired ways over time. Although those particlar applications are both somewhat speculative, an achievement as fundamental as a universal bio-molecular computer seems likely to yield some sort of important future consequences, perhaps of a nature which currently remains unforeseen. Additionally, a subsidiary contribution of this work will be that our experiments will help to advance the state of knowledge regarding binding of mismatched DNA strands, a subject which is useful in other areas of biology and molecular medicine, e.g.,for detecting mutations [Lerman-et-al-86], inhibiting genes [Wagner-94], and finding unknown genes similar to a known one.

Finally, the project is well-conceived as a program of personal research training for a young scientist, because it involves many important aspects of science: careful empirical lab work, understanding of principles of physics and computation (e.g.,sec. 2.9, sec. 2.4.3-2.6), extensive literature research, and not least important, experience with being at the forefront of a fast-growing new field.

In summary, this document has presented a highly concrete and focused plan for research in a subject area of considerable scientific and perhaps (eventually) technological interest. Moreover, significant amounts of conceptual development and planning work have already been accomplished in the last five months and presented in at least roughly polished form already in this document. The pace and thoroughness of the work accomplished already indicates that there should be no difficulty in completing work of an amount and quality sufficient for a Ph.D. in the time suggested.