DNA computing

by Francesc Rosselló


A molecule of the desoxyribonucleic acid (DNA), a polymer made of nucleotides of four kinds, can be conceived as a chain composed of four letters: A (adenine), T (thymine), C (cytosine) and G (guanine). A human DNA particle consists of nucleotides.

In the last 20 years the methods for DNA chain manipulation have developed considerably. With the use of enzymes a chain can be cut at any point, glued together (at matching letters) multiplied etc. Moreover, there are means for selecting those that contain some specified sequence of nucleotides. Those methods found a surprising application in 1994, when Leonard Adleman of the University of South Carolina solved a simple case of an old problem, known as the problem of Hamiltonian paths in a graph.

Adleman considered a graph with 7 vertices and 14 edges. First of all, he produced seven DNA chains of 20 letters each to represent the vertices. The edges were represented by 20-letter chains, too, chosen so that the first 10 letters of an edge were the same as the last 10 letters of the initial vertex of the edge, whereas the last 10 letters of the edge were also the first 10 letters of its terminal vertex. Thus chains corresponding to edges sharing a common vertex could be glued together, as well as the chains corresponding to vertices with those of the incident edges - just like Lego blocks. Using DNA molecule manipulation methods mentioned above, Adleman first selected all the paths in the graph, then multiplied those that contained all the vertices, and finally selected out of these the paths which contain each vertex exactly once. In fact, the "computation" extended over an entire week (a traditional computer would do the job in a fraction of a second), nevertheless a DNA-based computer proved to be possible, opening a new field of intensive research.

The first results are overwhelming. It is estimated that DNA molecules can hold data with a density of 1 bit/nm3, while our computers have density 1 bit/1012 nm3. Moreover, DNA-based computers can be 1,200,000 times faster than today's supercomputers, and use 1010 less energy!

Since Adleman's first experiment the technology for DNA has developed to such an extent, that oceans of DNA molecules and tons of enzymes are now replaced by a few grams of the first and vestigial quantities of the second.

Work is going on. Even if test-tubes full of DNA are not expected to take the place of electronic computers in this century, the results of Adleman and others open wide the doors to the world of DNA computers for specific tasks requiring a huge amount of parallel computations.