Abstract
The amount of digital memory is increasing day by day, while the power potential of typical silicon and other semiconductor transistors is severely limited. The researchers propose as an alternative to use natural biomolecules contained in the organisms of all living things, namely, DNA. DNA, with its unique chemical composition and unusual structure, can replace the classic binary system, becoming the basis for new processors. The possibility of using such DNA computers for high-tech calculations is discussed in detail in this paper. The reader is offered a short discussion of DNA structure, DNA computing logic determination, and analysis of current applications of such computers for solving ordinary arithmetic tasks.
Keywords: DNA, algorithms, computers, DNA computing, molecular data
Introduction
The amount of digital information stored and generated by humankind is increasing immeasurably. This suggests that there is an urgent need for a large-scale storage facility that can accommodate not only the actual amount but also all data that will be generated in the future. The innovative solution to this problem lies in the tiniest molecule that exists in our body cells – DNA. Information about all signs of a human being and human life is embedded in the four-letter alphabet of DNA. For this reason, IT researchers are faced with a significant question concerning the development of bioinformatics and the creation of biocomputers based on DNA computing.
Biocomputers are a hybrid of information technology and biochemistry. Researchers from different fields of science are trying to use real biological processes to create artificial computational schemes. There are several fundamentally different types of biological computers based on different biological processes: artificial neural circuits, evolutionary programming, gene algorithms, DNA computers, and cellular computers. The first two have been researched since the early 40s, but so far, these researches have resulted in nothing working. The last three, based on genetic engineering methods, have much higher promise, but work in these areas has begun relatively recently.
DNA computing is a section of the field of molecular computing, a new interdisciplinary research direction on the border of molecular biology and computer sciences. The main idea of DNA computing is to build a new paradigm of computation, new models, new algorithms based on knowledge about the structure and functions of the DNA molecule and operations that are performed in living cells over DNA molecules using different enzymes. The chief hopes that lie in the field of DNA computing in the practical sense are new methods for the synthesis of substances and objects at the molecular level. On the other hand. For specialists in the field of computer science and theory of computation, the paradigm of DNA-computations is engaging in new opening possibilities: new models of computations, new algorithms, and also possibility of solving the tasks not solved within the framework of the classical model of computations. For specialists in molecular biology, the area of DNA-computations can give new ideas, which were developed earlier in computer sciences, new instrumental means, new approaches in modeling of living matter on the molecular level. This research paper consistently describes the structure of DNA, gives the reader an idea of why biocomputers use this particular molecule, then demonstrates the basics of biocomputing and provides information on problems and results already achieved in this area. This paper will take a position that supports the potential of computers because, already today, there is a tremendous amount of research that confirms the incredible capabilities of such machines.
Unique Features of DNA
DNA is a unit of genetic information and a repository of genes that is characteristic of all living organisms. Almost in every cell of the human organism, it is possible to find such molecules; their content is the highest in cell nuclei. The molecule is a double elongated spiral divided into many sections and consisting of nucleotides (Cherry & Qi, 2018). Nucleotides in biology are chemical complex compounds consisting of five-carbon carbohydrate, phosphate groups, and a nitrogen base: adenine (A), guanine (G), cytosine (C), and thymine (T) (Ullah, Aslam, & Nazir, 2019). In general, the entire DNA strand can be considered as an irregular alternation of the four mentioned letters. In other words, all information about the body is enclosed in the four-letter alphabet of DNA.
As it is known, modern computers work with binary logic, which implies only two states: logical zeros and one. Using such binary code, including a sequence of zeros and units, any information can be encoded (Cherry & Qi, 2018). As mentioned above, there are four basic bases in DNA molecules that are connected in a chain. That is, one of the paired chains of DNA can have, for example, this type of chain: ATTTACGGCC – here, the quadruple logic, not the binary one, is used. Moreover, just as in binary logic, any information can be encoded as a sequence of zeros and units. In DNA molecules, any information can be encoded by combining basic bases.
The excellent research value of the molecule is not only its chemical composition but also its structure. A valuable property of DNA molecules is that they can take the form of a regular double helix with a diameter of only 2 nanometers (Cherry & Qi, 2018). Such a spiral consists of two chains-sequences of base bases, and the content of the first chain strictly corresponds to the content of the second. This correspondence is achieved due to the presence of hydrogen bonds between the two chains – G and C in pairs or A and T – directed towards each other. Describing this property of the double helix, molecular biologists say that DNA chains are complementary due to the formation of pairs G-C and A-T (Ullah et al., 2019). For example, if the sequence S is written as ATTACGTCG, then the complementary sequence S’ will take the form of TAATGCAGC.
Different enzymes are used for different manipulations of DNA molecules. Just as modern microprocessors have a set of basic operations such as addition, shift, logical operations AND, OR, and NOT NOR, DNA molecules can perform basic operations such as cutting, copying, and inserting under the influence of enzymes (Woods et al., 2019). Moreover, all operations on DNA molecules can be performed in parallel and independently of other operations, for example, the addition of the DNA chain is carried out when the original molecule of enzymes — polymerase.
Literary Review
While writing this research work, we used a whole lot of scientific sources, which differently translated the possibility of using DNA molecules for computational processes. Theoretical analysis of literature allows us to highlight a promising direction of development: cryptography on DNA (Mondal & Mandal, 2017). Chinese mathematicians and programmers have demonstrated that DNA threads obtained by biotechnological methods can be used for data processing in the process of chemical reactions (Zhong et al., 2020). The development proposed by researchers is a molecular solution poured into a test tube and trained to recognize nine different images. In fact, researchers from China proposed methods to evaluate the effectiveness of “learning” DNA threads and their ability to recognize images. Such studies have already been published in the article Cherry and Qian two years earlier (2018), but by that time, scientists have proposed another option of using acid molecules, namely – for handwriting expertise. Cherry and Qian (2019) drew numbers from 1 to 9 using DNA and created a biotechnical neural network. Scientists introduced one of the DNA sequences into the neural network, and the solution began to glow with specific colors that corresponded to this or that sample.
It is worth noting that using DNA molecules to organize calculations was not a new concept. In each case, careful development was needed in order to perform one particular algorithm that would generate the DNA structure (Woods et al., 2019). The difference in this case is that researchers have developed a system in which the same basic DNA sequences can be ordered to create completely different algorithms – and thus obtain completely different results. The researchers emphasize that the first computer based on DNA was created in 1994, the American scientist Leonard Adleman (Ullah et al., 2019). Scientific novelty Adleman consisted in manipulations of DNA molecules, which caused a controlled interaction between the programmer and the molecule. The test tube mixed DNA molecules in which the original data were encoded and specially selected enzymes. To solve the problem using DNA calculations, Adleman encoded the name of four cities in the United States as a single chain of DNA. The researcher tried to solve the hypothetical problem of a traveler visiting each of the cities only once to get from Atlanta to Detroit. This problem is solved by a direct search, but with increasing numbers of cities, its complexity increases exponentially. Adleman identified each city with 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 start point and the first half of the coding sequence for the finish point because the DNA molecule, like a vector, has direction. Modern molecular equipment allows synthesizing such sequences very quickly (Woods et al., 2019). As a result, the DNA sequence with the solution was 140 nucleotides.
DNA Molecules as a Computer
Devices that convert information from one form to another according to specific rules are called automata. The universal Turing machine, which gave impetus to the development of modern computers, is one of such hypothetical devices (Ullah et al., 2019). The principle of the Turing machine and its variations, including finite automata, is to read some data tape: such an organization is a direct analogy with information encoding biopolymers, which, in turn, served as a basis for the creation of several DNA-computers.
DNA-molecules with their unique structure form and possibility to realize parallel calculations allow to look at the problem of computer calculations in another way. Traditional processors execute programs sequentially (Woods et al., 2019). Despite the existence of multi-processor systems, multi-core processors, and various technologies aimed at increasing the level of parallelism, all computers built based on von Neumann architecture are devices with sequential command execution mode. All modern processors implement the following algorithm for processing commands and data: the selection of commands and data from memory and execution of instructions over selected data (Ullah et al., 2019). This cycle is repeated many times and with high speed. DNA-calculations have different, parallel architecture on their basis, and in some cases, it is thanks to this that they can easily calculate those tasks for which computers based on the von Neumann architecture would take years to solve.
The hardware base of such an automatic machine consists of limiting nucleotides and a ligase enzyme, while the software and the input are encoded with a double-helix DNA molecule. Under such conditions, programming is reduced to the selection of the software molecules needed in a particular case. Mixing the solutions containing the above components, the automaton processes the input molecule with a restriction cascade, hybridization and ligation cycles, producing a defined output molecule capable of encoding the final state of the automaton and thus determining the result of the calculation.
DNA Computer Logic
To understand the principle of the work of DNA technology, it is essential to note that bits are binary units of information in electronic computers. They represent the physical discrete state of the hardware needed, such as the presence or lack of electric current (Woods et al., 2019). These bits, or to be more precise the electrical signals, pass through circuits composed of logical elements that perform an operation with one or more input bits and generate one bit as an output. By combining these simple building blocks over and over, computers can launch surprisingly sophisticated programs (Takahashi, Nguyen, Strauss, & Ceze, 2019). The idea behind DNA calculations is to replace electrical signals with chemical bonds, and silicon with nucleic acids to create biomolecular software (Woods et al., 2019). In a DNA computer, unlike traditional types of devices, data input and output are demonstrated by changing the concentration of molecules in a test tube containing acid molecules and enzymes.
Challenges and Achievements of DNA Computers
As with any other processor, the DNA processor is characterized by the structure and command set, the structure of the processor is the structure of a DNA molecule, and the command set is a list of biochemical operations with molecules. The principle of computer DNA memory is based on the serial connection of four nucleotides (Stefano, Wang, & Kream, 2018). The three nucleotides, connected in any sequence, form an elementary memory cell, the codon, whose aggregate then forms a DNA chain (Ullah et al., 20199). The main difficulty in the development of DNA computers is related to the selective single-codon reactions within the DNA chain (Ullah et al., 20199). Another critical problem is the natural phenomenon of DNA self-assembly, resulting in information loss (Woods et al., 2019). It is overcome by the introduction of particular inhibitors in the cell, substances that prevent the chemical reaction of the self-assembly.
It is important to note that all existing DNA-systems have a problem: they are unique proprietary developments that do not have any flexibility. If it is compared with silicon technology, each group of researchers from scratch has to develop a new computer architecture, for which they need to write new software. Nevertheless, this could change with the first programmable DNA computer developed at the University of California at Davis (Woods et al., 2019). The first DNA-programmable computer is described in a scientific article, the authors of which have shown that with a simple trigger, the same basic set of DNA molecules can implement many different algorithms. Although the study is a simple laboratory experiment, in the future programmable molecular algorithms can be used, for example, to program DNA robots that have already successfully delivered drugs to cancer cells.
In addition, researchers from Washington University and Microsoft recently built the world’s first DNA auger. This design is the first to be able to record and read the information in a DNA warehouse without human involvement (Takahashi et al., 2019). During the first launch, the device prototype successfully translated the word “HELLO” into an unusual format. The device encoded the bits in DNA sequence (A, C, T, G) and synthesized the molecule itself, saving it as a liquid. The stored DNA was then decoded back into bits after enzyme treatment. It took the device 21 hours to encode and then decrypt a 5-byte recording, but the researchers said they had already found a way to reduce this time by 10-12 hours (Takahashi et al., 2019). In nucleotide form, the liquid version of the word HELLO had a mass of only four µg of approximately 1 mg of synthesized DNA (Takahashi et al., 2019). It was estimated that this method would fit all the information stored in a large data center into a drive slightly larger than a few playing cubes. High reliability of such disks is also noted: as Microsoft points out, some DNA samples were stored for tens of thousands of years — for example, in mammoth tusks.
Moreover, French scientists from the Sadron Institute successfully encoded and then read the word “Sequence,” which was presented in the ASCII-code using a sequence of synthetic polymers. Thus, they proved that DNA molecules could store information and take up 100 times less space than conventional hard drives (Al Ouahabi, Amalian, Charles, & Lutz, 2017). It is interesting to note that until recently, biocomputers have not been able to perform complex mathematical operations. The situation changed with the release of research papers by Zhou, Geng, Wang, and Guo (2019), which showed that modern DNA biocomputers are able to calculate the square root of numbers up to 900. During the experiment, the researchers encoded the ten links in DNA number from 1 to 900 and attached to them fluorescent marks. Since each DNA chain completes itself paired on the principle of complementarity or correspondence, the scientists used this process, controlling the process of completion so that the color of the fluorescence corresponds to a particular square root.
Benefits of Using DNA Computing
The DNA molecules may be the material that will soon be able to replace the classic silicon transistors with their binary logic completely. The most important advantage, which determines the considerable potential for the use of natural molecules, is that a small mass of DNA can adequately accommodate the total amount of information available to date (Stefano et al., 2018). Another advantage of DNA processor over conventional silicon processors is that they can perform all calculations in parallel rather than in series, ensuring that the most complicated mathematical calculations can be performed in minutes (Woods et al., 2019). Traditional computers would take months and years to perform such calculations.
Conclusion
One of the alternatives to modern semiconductor technology in the future could be so-called biological computers or biocomputers. A typical computer operates electrical signals, binary bits of information, while in DNA computers, electrical signals are replaced by chemical bonds, and nucleic acids replace silicon. It is assumed that computers based on such molecules can create massive and parallel computing architectures that are difficult to implement on powerful silicon computers.
Each DNA molecule corresponds to another DNA, which meets the conditions of the principle of complementarity. The second circuit has an opposite orientation to the original molecule. The attraction of adenine to thymine and cytosine to guanine results in a double helix, providing the possibility of doubling DNA in cell division. The task of doubling is solved by using a particular enzyme protein. Essentially, such a protein is an implementation of the Turing machine, consisting of two tapes and a programmable control panel. The remote read data from one ribbon process it according to some algorithm and writes it to another ribbon. The unit also sequentially reads the raw data from one DNA ribbon and forms a calculation ribbon based on that data.
Whole generations of researchers have demonstrated to the world community the possibility of using molecules of natural acids as computer components, such as hardcore and processor. Many scientists are convinced that the issue of large-scale and mass introduction of such technologies, allowing to perform system processes in a short time, is only a matter of the current decade. However, today DNA-calculations are nothing more than promising technologies at the level of laboratory research, and they can be in this state for many more years. In fact, at the current stage of development, it is necessary to answer the global technological question related to the determination of the universality of DNA computing functions.
References
Al Ouahabi, A., Amalian, J. A., Charles, L., & Lutz, J. F. (2017). Mass spectrometry sequencing of long digital polymers facilitated by programmed inter-byte fragmentation. Nature Communications, 8(1), 1-8.
Cherry, K. M., & Qian, L. (2018). Scaling up molecular pattern recognition with DNA-based winner-take-all neural networks. Nature, 559(7714), 370-376.
Mondal, B., & Mandal, T. (2017). A lightweight secure image encryption scheme based on chaos & DNA computing. Journal of King Saud University-Computer and Information Sciences, 29(4), 499-504.
Stefano, G. B., Wang, F., & Kream, R. M. (2018). DNA MemoChip: Long-Term and high capacity information storage and select retrieval. Medical Science Monitor: International Medical Journal of Experimental and Clinical Research, 24(1), 1185-1187.
Takahashi, C. N., Nguyen, B. H., Strauss, K., & Ceze, L. (2019). Demonstration of end-to-end automation of DNA data storage. Scientific Reports, 9(1), 1-5.
Ullah, M. S., Aslam, W., & Nazir, H. (2019). DNA computing: A review of promises and potential. Journal of Information Communication Technologies and Robotic Applications, 10(1), 93-101.
Woods, D., Doty, D., Myhrvold, C., Hui, J., Zhou, F., Yin, P., & Winfree, E. (2019). Diverse and robust molecular algorithms using reprogrammable DNA self-assembly. Nature, 567(7748), 366-372.
Zhong, G., Li, T., Jiao, W., Wang, L. N., Dong, J., & Liu, C. L. (2020). DNA computing inspired deep networks design. Neurocomputing, 382(1), 140-147.
Zhou, C., Geng, H., Wang, P., & Guo, C. (2019). Programmable DNA nanoindicator‐based platform for large‐scale square root logic biocomputing. Small, 15(49), 1-9.