DNA computers: invisible monitors of health?

Each year, zettabytes of data are produced globally, which is comparable to the number of stars in the observable universe. This number is increasing at a rapid pace, and in 2025, global yearly data production is set to hit 175 zettabytes. This is the digital cosmos we live in.

A surprising solution to handle this data deluge might be a substance operating at a much smaller scale – DNA. The promise? All the world’s data could hypothetically be stored in a cup of DNA. We are nowhere near achieving this, but researchers have managed to store an episode of a Netflix show on DNA, and more recently, the world’s first ‘DNA book’ was published.

DNA for storage, which we wrote about in a briefing last year, isn’t the only use of DNA as computer hardware. Since the 1990s, scientists have been exploring the potential of DNA to perform computational tasks, and hence, the development of a DNA computer.

What is a DNA computer?

In simple terms, a DNA computer uses DNA as the hardware and biochemical reactions as the software. The basic units of a DNA computer are the bases As, Ts, Gs and Cs, much like how traditional computers use 1s and 0s. ‘Operations’ are performed through hybridisation (base-pair binding) and ligation (joining) of DNA strands. These computers are powered not by electric supply but by the power of biochemical reactions.

Definitions of DNA computing vary across disciplines. In biology, the definition of DNA computing might overlap with efforts in biotechnology, where DNA is being used to perform simple logical tasks such as detecting the presence of a compound. Whereas, in computer science, DNA computing often refers to the use of DNA to complete computational tasks that go beyond simple logical calculations. Here, we use a very broad definition of DNA computing to encompass the breadth of healthcare applications that we found.

How does a DNA computer work?

Let’s take an example. Say you want to find the shortest way to visit cities A to Z, visiting each city only once and ending your trip back in City A. This problem is the classic travelling salesman problem that a DNA computer can help solve. To solve this, first, DNA with random but unique sequences representing each of the cities is designed. Then, information about any roads connecting the different cities (e.g., City A and City D) is incorporated. This is done using linker sequences – unique base pairs added to the end of the DNA strand representing City A, such that it is complementary to the linker sequence of the DNA strand representing City D. One way to make the final DNA sequence is using DNA printers that can create customised DNA for you. 

Multiple copies of all the DNA strands representing all the cities are then put in a test tube with a suitable reagent and shaken. The movement causes different DNA strands to bind to each other due to sheer proximity and following biochemical rules. These reactions are the computation. 

To identify the solution, one could separate DNA using gel electrophoresis (a method that separates DNA by length) and pull out the smallest linked strands. These isolated strands are then sequenced to determine the solution.

Why do we need this? 

We are approaching the physical limits of how small silicon chips (a key component of computer hardware) can be made – something which has helped laptops become slimmer and more powerful over time. We are also facing a tidal wave of data and traditional data centres are massive, consume a lot of electricity and generate a lot of heat. In contrast, DNA is compact, energy-efficient and offers solutions for data processing and storage at the nanoscale.

What makes DNA computers special?

• One gram of DNA can theoretically store 215 petabytes of data (equivalent to the Library of Congress’ digital collections 10 x over). This capacity is likely to further increase as new innovations push the storage limit that can be achieved with DNA 
• DNA computing reactions use very little energy 
• DNA computers can check for trillions of possible solutions simultaneously and can be used for parallel processing. So instead of solving a 1000-piece puzzle one-piece-a-time, a DNA computer can simultaneously look for solutions for each piece, reducing the time taken to solve the puzzle. 

Some healthcare-related use cases in development 

There are several healthcare-related uses of DNA computers under development, each with a different mode of operation. For example, researchers have combined DNA computing with nanopore sequencing to detect cancer-associated microRNAs (miRNA) in the blood of patients with bile cancer. Using a nanopore sequencer, researchers are able to distinguish a DNA plus miRNA strand from a DNA-only strand with high sensitivity. So, this method can be used to detect very low concentrations of miRNA biomarkers in patient blood, allowing early detection of bile duct cancer.

Another example is the use of DNA computing to distinguish between bacterial and viral acute respiratory infections (ARIs), a process that is usually time-consuming and costly. RNAs produced in the patient’s blood in response to an ARI are detected by combining the blood sample with DNA strands specifically designed to bind to those RNAs. The binding releases and activates a fluorophore, a light-emitting molecule, which is attached to the DNA. The intensity of the light from the fluorophore indicates whether the ARI is bacterial or viral. This method has an 87% accuracy rate and can be performed in under four hours.

What’s holding DNA computing back?

The interest in DNA computing comes off the back of a period of relatively slow progress following the invention of the first DNA computer by Leonard Adleman in 1994. The field struggled due to a lack of tools available to modify DNA which also limited scale-up. Recent transformations have made it significantly easier to edit, copy, read and write DNA, however a number of key barriers still remain:

• Prone to errors: Biochemical operations are not always reliable as errors in synthesis or hybridisation can affect results.
• Still too slow: In a majority of cases, DNA reactions still cannot surpass standard electronic computers.
• Not ready for scale-up: Building and reading massive DNA systems is still costly and complex. You also need sequencers to read the output and convert this to digital information which limits where and when they can be used.
• Lack of standards: Unlike digital computing, a universal programming language for DNA remains to be developed.

Ethics and governance challenges

Barriers to implementing DNA computing mean the ethics and governance of this technology have received little attention so far but there are several issues to consider as the field develops. The question of purpose is crucial: what uses could – and should – DNA computing be put to? Given the potential to combine programming and biology, it is speculated that DNA computing could open the door to new ways of intervening with life, including through augmentation. Whilst this may still be some way off, proactive governance will be essential to ensure the technology is pursued in line with societal expectations. 

DNA computing also poses safety challenges. Understanding how and when errors arise, and ensuring error mitigation steps are in place is necessary before implementing DNA computers in healthcare. This becomes more pertinent if and when DNA computers are deployed in vivo, as they risk physical harm through injury or illness. This is further complicated by limitations when it comes to directly observing or monitoring operations occurring at a molecular level. Therefore, rigorous clinical utility assessments, implications for patient safety, and the information patients need to know about possible outcomes will be important components of future regulation. 

The extent to which DNA computing can be equitably accessed remains to be seen. In theory, computations in DNA computers are powered by biochemical reactions and do not require electricity like traditional computers. However, in reality the role of surrounding infrastructure needs to be factored in. A reliance on specialist knowledge and equipment risks restricting their implementation to areas where there is already a concentration of research expertise and investment. Therefore, methods to support decentralisation will be necessary if DNA computing is to reduce, rather than exacerbate, existing inequalities in access to health (and other) technologies.

Invisible doctors of the future?

The future of DNA computers holds great promise across several fields including healthcare. It has been twenty years since scientists developed a DNA computer that could detect a disease-related molecule and produce a therapeutic compound in response. Although there hasn’t been much progress since, this application of DNA computers as biological control systems is particularly exciting for future consideration.

Imagine you are genetically prone to cancer in a future where tiny DNA computers live inside your body – computers smaller than a cell, floating through your bloodstream. They continuously monitor for early warning signs. If they detect unusual gene expression, they flip a chemical switch to shut down the risky cell, alert your doctor, or even begin localised treatment. It would be like having millions of invisible, always-on physicians, working at the molecular level to keep you healthy. Like all computers, when DNA computers work, they make everything simpler – there are no symptoms or trips to the hospital – but these computers also carry risks – they can be turned on and off, they can be corrupted by viruses, and they could even be hijacked to do things you don’t want them to.

Whether and when such a scenario might transpire is unclear, but ongoing developments in DNA computing are attracting commercial interest. Compound Annual Growth Rate (CAGR) is a commonly used measure of average yearly growth in investment over a given period. The DNA computing market is estimated to grow at a CAGR of 35.85% over the next 5 years, which is similar to other reportedly high-growth industries such as AI in drug discovery. Unsurprisingly, several companies are invested in developing DNA Computers, with the UK-based company DNA Ascendancy having patented a proof-of-concept model for a practical DNA computer. It remains to be seen how this technology evolves, what new applications it opens up for healthcare and how the ethics and governance challenges that arise can best be approached.