Biocomputers

Nowadays, when every new step in improving semiconductor technologies is given with great difficulty, scientists are looking for alternative possibilities for the development of computing systems. The natural interest of a number of research groups (among them Oxford and Texas Universities, Massachusetts Institute of Technology, Berkeley Laboratories, Sandia and Rockefeller) aroused natural ways of storing and processing information in biological systems. The result of their research was (or, more precisely, still has to be) a hybrid of information and molecular technologies and biochemistry - a biocomputer. There are developments of several types of biocomputers, which are based on different biological processes. These are, first of all, DNA and cellular biocomputers that are under development.

DNA computers

In living cells, as is known, genetic information is encoded in a DNA molecule (deoxyribonucleic acid). DNA is a polymer composed of subunits called nucleotides. The nucleotide is a combination of sugar (deoxyribose), phosphate, and one of the four nitrogenous bases that make up DNA: adenine (A), thymine (T), guanine (G), and cytosine (C). The DNA molecule forms a helix consisting of two chains connected by hydrogen bonds. In this case, the base A of one chain can be connected by hydrogen bonds only with the base T of the other chain, and the base G - only with the base C. That is, having one of the DNA chains, you can always restore the structure of the second. Due to this fundamental property of DNA, called complementarity, genetic information can be accurately copied and passed from mother cells to daughter cells. Replication of the DNA molecule occurs due to the work of a special enzyme DNA polymerase. This enzyme slides along the DNA and synthesizes a new molecule based on it, in which all bases are replaced with the corresponding pairs. Moreover, the enzyme begins to work only if a short piece-seed (primer) is attached to the DNA. In cells, there is also a matrix ribonucleic acid (RNA) molecule related to the DNA molecule. It is synthesized by a special enzyme that uses one of the DNA chains as a sample, and is complementary to it. It is on the RNA molecule in the cell, as on the matrix, with the help of special enzymes and auxiliary factors that protein synthesis occurs. The RNA molecule is chemically more stable than DNA, therefore it is more convenient for experimenters to work with it. The sequence of nucleotides in the DNA / RNA chain determines the genetic code. The unit of the genetic code, the codon, is a sequence of three nucleotides.

Scientists decided to try, using the example of nature, to use DNA molecules for storing and processing data in biocomputers.

The first of these was Leonard Adleman from the University of Southern California (see: "Molecular Computation of Solutions to Combinatorial Problems. Science, 1994, No. 266, p. 1021), who was able to solve the Hamiltonian problem. Its essence was to find the route with given start and finish points between several cities (in this case, a family), each of which is allowed to visit only once. The “road network” is a unidirectional graph. This problem is solved by brute force, however, as the number of cities increases, its complexity increases exponentially .Ed each city A lmen identified a unique sequence of 20 nucleotides. Then the path between any two cities will consist of the second half of the coding sequence for the starting point and the first half of the coding sequence for the finish point (the DNA molecule, like the vector, has a direction) .To synthesize such sequences modern molecular apparatus allows very quickly. As a result, the DNA sequence with the solution will be 140 nucleotides (7x20).

It remains only to synthesize and isolate such a DNA molecule. For this, about 100 trillion DNA molecules containing all possible 20-nucleotide sequences encoding cities and the paths between them are placed in a test tube. Further, due to the mutual attraction of nucleotides A-T and GC, individual DNA chains interlock with each other randomly, and the special enzyme ligase stitches the resulting short molecules into larger formations. At the same time, DNA molecules are synthesized, reproducing all possible routes between cities. It is only necessary to select from them those that correspond to the desired solution.

Adlmen solved this problem by biochemical methods, by successively removing first the chains that did not start from the first city — the starting point — and did not end at the finish point, then those that contained more than seven cities or did not contain at least one. It is easy to understand that any DNA molecule remaining after such a selection is a solution to the problem. (For more, see: V. Borkus, "DNA is the basis of computers". PC Week / RE, No. 29-30 / 99, p. 29).

Following the work of Adlman, others followed. Lloyd Smith of the University of Wisconsin solved with the help of DNA the task of delivering four varieties of pizza to four addresses, which meant 16 answer choices. Scientists from Princeton University solved the combinatorial chess problem: with the help of RNA, they found the right move of a chess horse on a nine-cell board (512 variants in total).

Richard Lipton from Princeton was the first to show how to encode binary numbers using DNA and solve the problem of satisfying a logical expression. Its essence is that, having a logical expression that includes n logical variables, you need to find all combinations of variable values ​​that make the expression true. The problem can be solved only by searching 2n combinations. All these combinations are easy to encode using DNA, and then act according to the method of Adleman. Lipton also proposed a way to break the DES (American cryptographic) cipher, interpreted as a kind of logical expression. The first model of a biocomputer, however, in the form of a mechanism made of plastic, in 1999 created Ihud Shapiro from the Weizmann Institute of Natural Sciences. She imitated the work of the "molecular machine" in a living cell that collects protein molecules according to DNA information, using RNA as an intermediary between DNA and protein.

And in 2001, Shapiro managed to implement the model in a real biocomputer (see Programmable and autonomous computing machine made of biomoleciles, Nature, 2001, No. 44, p. 430), which consisted of DNA molecules, RNA and special enzymes. Enzyme molecules served as hardware, and DNA molecules - software. In this case, about one trillion elementary computational modules were placed in one tube. As a result, the computation speed could reach a billion operations per second, and the accuracy - 99.8%.

So far, the Shapiro bio-computer can be used only to solve the simplest tasks, giving out only two types of answers: "true" or "false." In the experiments performed in one cycle, all DNA molecules simultaneously solved a single problem. However, they can potentially work at the same time on different tasks, while traditional PCs are, in fact, single-task.

At the end of February 2002, it was reported that Olympus Optical is claiming primacy in creating a commercial version of a DNA computer designed for genetic analysis. The machine was created in collaboration with the Associate Professor of Tokyo University Akira Toyama.

A computer built by Olympus Optical has molecular and electronic components. The first carries out chemical reactions between DNA molecules, provides search and selection of the result of calculations. The second one processes information and analyzes the results obtained.

Gene analysis is usually done manually and takes a lot of time: this creates numerous DNA fragments and monitors the course of chemical reactions. "When DNA computing is used for genetic analysis, tasks that were previously performed for three days can be solved in six hours," said Olympus Optical employee Satoshi Ikuta.

The company hopes to put a DNA-based genetic analysis technology on a commercial basis. It will find application in medicine and pharmacy. Scientists plan to introduce molecular nanodevices in the human body to monitor its health and the synthesis of necessary drugs.

The capabilities of biocomputers are also interested in the military. The American Defense Research Agency DARPA is running a project called Bio-Comp (Biological Computations). His goal - the creation of powerful computing systems based on DNA. Along the way, researchers hope to learn how to manage the interaction of proteins and genes. To this end, it is planned to create a powerful simulator Bio-SPICE, capable of visualizing biomolecular processes by means of computer graphics. Bio-SPICE is planned to be developed on the principles of open source (open source). The program is designed for five years.

Cell computers

Another interesting direction is the creation of cellular computers. Bacteria would be ideally suited for this purpose if it were possible to incorporate into their genome a certain logical scheme that could be activated in the presence of a certain substance. Such computers are very cheap to manufacture. They do not need such a sterile atmosphere, as in the manufacture of semiconductors. And once you program a cell, you can easily and quickly grow thousands of cells with the same program.

In 2001, American scientists created transgenic microorganisms (i.e. microorganisms with artificially modified genes) whose cells can perform logical operations AND and OR.

Laboratory specialists in Oak Ridge, Tennessee, used the ability of genes to synthesize a particular protein under the influence of a certain group of chemical stimuli. Scientists changed the genetic code of the bacteria Pseudomonas putida in such a way that their cells acquired the ability to perform simple logical operations. For example, during the operation And in the cell are fed two substances (in fact - input operands), under the influence of which the gene produces a specific protein. Now scientists are trying to create more complex logical elements on the basis of these cells, and are also thinking about the possibility of creating a cell that performs several logical operations in parallel.

The potential of biocomputers is very high. Compared with conventional computing devices, they have a number of unique features. First, they use not binary, but ternary code (as the information in them is encoded by triples of nucleotides). Secondly, since the calculations are performed by simultaneously entering into the reaction trillions of DNA molecules, they can perform up to 1014 operations per second (although the extraction of the results of calculations involves several stages of very thorough biochemical analysis and is much slower). Thirdly, DNA-based computing devices store data with a density that is trillions of times higher than optical discs. Finally, DNA computers have extremely low power consumption.

However, in the development of biocomputers, scientists are faced with a number of serious problems. The first is associated with reading the result - modern sequencing methods (determining the coding sequence) are not perfect: it is impossible to sequentially sequence at least several thousand bases at a time. In addition, it is a very expensive, complex and time-consuming operation.

The second problem is calculation errors. For biologists, an accuracy of 1% in the synthesis and sequencing of bases is considered very good. For IT, it is unacceptable: problem solutions can be lost when molecules simply stick to the walls of blood vessels; there are no guarantees that point mutations will not occur in DNA, etc. And yet - DNA breaks down over time, and the results of calculations disappear before our eyes! But cellular computers are slow and easy to confuse. With all these problems, scientists are actively fighting. How successful - time will tell.

Biocomputers are not designed for the broad masses of users. But scientists are hoping that they will find their place in medicine and pharmacy. The head of the Israeli research group, Professor Ehud Shapiro, is confident that in the future DNA nanomachines will be able to interact with human cells, monitor potential pathogenic changes and synthesize drugs to combat them.

Finally, with the help of cellular computers it will be possible to combine information and biotechnologies. For example, they will be able to manage a chemical plant, regulate biological processes inside the human body, produce hormones and medicinal substances, and deliver the required dose of drugs to a specific organ.