Molecular computing is the science of using individual molecules to build computer programs. Instead of running software on a traditional computer, some scientists are now trying to replace the silicon chip with test tubes, liquids, and even living cells due to a concern about the limits of miniaturization—a real and pressing problem that threatens the future advancement of computers.
To get an idea of how small computers have already become, think of a standard processor chip and imagine this: if the chip circuit were magnified such that the individual components were the size of office buildings, and the interconnections between them were the size of streets and avenues, then the entire circuit at this scale would stretch from London to San Francisco. Very impressive, but can it continue? The answer is "no."
When it comes to creating smaller and smaller computers, existing technology—white-suited technicians in clean-rooms making silicon chips—will eventually hit a very solid wall imposed by the realities of physics. Once circuits are miniaturized down to the atomic level, components begin to interfere with one another, and the whole chip becomes useless. It is for this reason that some people are investigating alternatives to silicon to build the computers of the future.
Scientists in the field of nanocomputing are investigating several different possibilities, including the use of biological molecules. It seems that deoxyribonucleic acid (DNA), the very stuff of life, may hold the key. A tiny, almost invisible drop of water can contain trillions of molecules of DNA. Nature has information storage down to a fine art: a human body contains countless copies of the genetic sequence that makes a person who he or she is, and yet one single copy of that sequence would occupy a large encyclopedia if it were printed out.
Moreover, when scientists manipulate solutions of DNA, they operate on trillions of strands simultaneously. This massive parallel processing, combined with the incredible degree of miniaturization offered by DNA, leads scientists to believe it could form one of the main components of twenty-first-century computers.
The personal computer or laptop stores information in the form of bits, each of which may take the value one or zero. A computer software program is nothing more than a string of ones and zeroes, which is interpreted by the computer processor. DNA molecules are similar in that they are simply strings of, not ones and zeroes, but As, Gs, Cs, and Ts (this is how the 1997 film Gattaca, starring Ethan Hawke and Jude Law, got its name). The human genetic sequence can be thought of as "software," which is then interpreted by human "hardware"—the various processes that guide development from a single cell to a fully functioning human being.
The key to using DNA to compute is that it can be manufactured in the laboratory. A request for a particular sequence (say, AGTTCA) can be given to a technician, and, after a short wait, a machine produces countless copies of the same short DNA sequence, ready for use. So how can DNA be used to solve a computing problem?
The problem of "coloring" has a long history. Given a map of the mainland United States, each state can be colored one of four colors such that no two states sharing a border are colored the same. However, what happens if there are only three colors, say, red, green, and blue? Will it still be possible to not color two adjacent states the same? This problem is easy to describe, but fiendishly difficult to solve once the map gets only moderately large.
The first thing to do is to generate all possible ways to color the map, each way represented by a single long sequence of DNA. This is done by mixing together trillions of copies of smaller sequences each encoding, say, "color Michigan green," "color Wisconsin blue," or "color Michigan red" as a distinct sequence of bases. If the sequence encoding is right, these sequences stick together to form larger sequences, where each state is represented only once. Of course, most of these longer sequences encode undesirable colorings, for example, where both California and Nevada (which are, of course, neighbors) are colored green.
However, if enough DNA is used, a correct coloring is probably in there somewhere; it is the needle in a very large haystack. The next step is to remove from the test tube all of the undesirable colorings—the equivalent of the "hay." For each state border, any sequences that color two neighboring states the same is removed. This is done by adding extra DNA sequences and a dash of chemicals to the tube. This process is repeated for each border until all that is left is the "needle"—a sequence encoding a correct coloring of the map.
Of course, biological operations work on geological time scales compared to twenty-first-century supercomputers . The power of the DNA computer lies in its massive parallel processing capability; when a chemical is added to the test tube it acts on every strand simultaneously. Because the average tube holds trillions of strands, that is a lot of computing going on at once.
Some scientists are going one stage further and re-engineering the genetic programs of living cells. The machinery of life that controls the development of cells can now be re-programmed to give the cells simple, human-defined "decision-making" capabilities. By replacing specific sequences within their genetic code, it may soon be possible to engineer cells to act as microbial "robots" that seek out disease or deliver drugs at the point of infection. Of course, these various developments may take decades to bear fruit, and some may never get beyond the concept stage. What is clear, however, is that the fusion of computers and biology will provide some of the most exciting scientific breakthroughs of the twenty-first century.
see also Medical Systems; Molecular Biology; Nanocomputing.
Adleman, Leonard M. "Computing with DNA." Scientific American (August 1998): 54–61.
——. "Molecular Computation of Solutions to Combinatorial Problems." Science 266 (1994): 1021–1024.