From Silicon to Cells: How Biology Could Redefine Computing

By Yuseong Hong, Jungmin Heo

Living cells are now beginning to do what we would usually expect from computers—solving problems and storing information. What if computers are not built from wires and silicon but from life? Instead of electrons in circuits, molecules could handle information, remember signals, and make decisions—using far less energy than modern computers.

Generally, silicon-based computers can do most of the computational tasks. But some inherent problems limit the performance and sustainability of such computers. The following are the main reasons for the necessity of biocomputers.

Firstly, the shrinking size of transistors now threatens the accuracy of silicon semiconductor computers. As of now, the scale of an individual transistor is around two to three nanometers, which is equal to a few dozen atoms. This causes quantum tunneling, in which electrons pass through the barriers, leading to signal loss, errors, and heat generation. 

And then the tremendous electricity demands of these computers threaten sustainable development. According to Deloitte, a network of professional service firms, the power consumption of global data centers is expected to reach 681 terawatt-hours by 2026, and AI data centers alone will consume around 90 terawatt-hours. Note that the entire country of Finland spends 90 terawatt-hours a year. This figure tends to increase proportionately to the development of the AI industry. 

In short, silicon computers meet natural limits, such as errors and energy loss. At the same time, data centers running AI and cloud services consume electricity in enormous amounts. And recently, research has increased to create an entirely new computer system. 

To get over these issues, scientists are researching biocomputing, a field that programs cells to act like what we know as computers. The structure of DNA allows dense storage of information, proteins carry out difficult tasks, and cells adapt to environments. Through these advancements, researchers aim to create computers that are smaller, more efficient, and capable of performing tasks that silicon-based computers cannot.

Quantum computers are on the way to solving problems that many supercomputers have difficulty in solving. But limitations of their functionality and technical barriers prevent them from becoming general computers. What is better than ordinary computers but also more realistic than quantum computers?

One of the possible answers to the problem is closer than expected; in fact, the key may actually be what constitutes your body. Biomaterials and cells are emerging as a new way to process information, using biocomputing. Small in volume but highly compressed materials and complex networks of cells are a way to compute and store information much more energy efficiently than their silicon counterparts.

The history of biocomputing is no longer than the inception of computers. After Watson and Crick's discovery of the double helix, some scientists thought that genetic code could encode information similar to electronic bits. 

And in 1994, Adleman, who is also known for co-inventing the RSA, first used DNA as a computation source. In Molecular Computation of Solutions to Combinatorial, he demonstrated that encoded DNA can solve computational problems. This experiment became the starting point of the applications of biotic materials as computing resources.

An introduction to types of biocomputers:

• DNA-Based Biocomputers: As mentioned above, DNA molecules can encode information by setting up the sequence of bases. Hybridization (joining two single-stranded DNA to form a complete double helix) and other enzymatic reactions are the mechanism of computation. DNA can store information, solve combinatorial problems, and simulate complex biological situations.

• RNA-Based Biocomputers: RNA molecules are known to serve regulatory and computational functions, and they are used for therapeutic applications and biosensors.

• Protein-Based Biocomputers: The complex structure and biochemical reactions of proteins enable biosensing, molecular diagnostics and assembly, and so on.

• Cellular Biocomputers: Genetically engineered bacterial or mammalian cells behave as logic gates and can form complex biology circuits, performing logical functions, memory storage, biological information processing, etc.

• Organoid Intelligence and Multi-Electrode Arrays (MEAs): Miniaturized and simple versions of organs are called organoids. Brain organoids are composed of human neurons and can simulate neural networks for computation, and they are expected to contribute to speed and energy efficiency and modeling personalized medicine or neurological diseases. MEAs can act as an interface for communication between systems of cellular networks by sending stimuli and recording responses.

Imagine a little factory in a cell, where there are circuits made out of molecules instead of wires. In cellular computing, researchers create gene circuits, which are DNA sequences that tell the cell to give a response to a certain signal. Signals come in different forms; they can be chemical, light, or some change in the environment. If an intended change takes place, the circuit will stimulate certain genes, causing the cell to make an output.

One of the results from this idea is DOMINO from the Massachusetts Institute of Technology. This allows cells to record information at fine resolutions—information not only about the existence of the signal but also about the time, order of signals, and intensity. That is, a cell can remember a series of events, not just whether something happened.

To make up these circuits, scientists use parts like promoters, activator and repressor proteins, and sometimes enzymes that can rearrange DNA. These tools let the cell use logic similar to computers. One challenge is that active living systems are noisy and unpredictable: genes can vary, cells may mutate, or the environment can fluctuate. To make circuits reliable, researchers try to use parts that do not interfere with each other and build feedback mechanisms so that small changes do not ruin the system.

Another application of this technology is storage. Some circuits are only temporary; they only respond when the signal appears, then shut down. Conversely, another type of circuit memorizes a specific signal. Memory can be permanent or semi-permanent. In DOMINO, for instance, the memory is coded at the level of DNA bases, so the cell’s DNA literally records the history of observed signals.

Once you have cells that can sense, compute, and remember information, many new possibilities open. In medicine, researchers are exploring living diagnostics, cells engineered to detect things and give signals. For example, bacteria have been adjusted so that when they detect a tumor, they activate a gene that generates a substance visible in urine, letting doctors find out without complex procedures.

Another medical prospect is smart therapies. In this idea, engineered cells do not only detect; they act. If an unusual signal is present, they trigger immune responses. Some circuits are being designed to respond to several simultaneous signals. For example, there is a gene circuit in stem cells where the cell is engineered to produce an anti-inflammatory biologic when two rules are true: inflammation being high and according to the circadian rhythm. In this way, unwanted treatments do not occur, improving control.

These circuits are capable of taking the efficiency of biomanufacturing to the next level. Imagine bacteria that swap between diverse metabolic techniques depending on present nutrients or shut down reactions if conditions are not optimal, saving energy. This kind of control is being researched in synthetic biology broadly.

Some additional ideas are quite intriguing as well. One is building adaptive circuits that change their behavior with repeated signals, refining their responses. Another is integrating living computing with electronics, so you can use living cells to process biological events. And safety and ethical issues are prominent. 'Wild' cells can cause unintended or uncontrollable mutation, which requires kill switches to prevent them from happening. Some papers on this topic emphasize biosafety as in need of development.

A novel system using cells, cellular computers can overcome some of the problems their silicon counterparts couldn't. However, life is yet a subject that needs further research, and there are some limitations found in the process of development. Living systems have technical difficulties such as noise and are, on multiple occasions, unpredictable. Thus, it requires reliable control and improvement. 

Manipulating life necessitates caution: to prevent biohazards from small accidents to epidemics and to encourage humane technology, safety precautions and ethical guidelines related to biocomputing must be implemented. And beyond the basics, there is a need for technical safety mechanisms, such as the previously mentioned kill switches. 

Cellular biocomputers integrated into machines or server computers pretty soon. Currently in the alpha stage of development, this biocomputing system opens many possibilities, further benefiting mankind soon enough.