DNA computing and information storage is a viable, rapidly emerging and quickly commercialised technology. Our capacity to manipulate, synthetise, and digitally interface with DNA as a biochemical medium is setting the stage for a radical reappraisal of biocomputing in general, leading to the idea of bioprogramming.

Perhaps you have heard about nanorobots or other microscopic machines. All of them can be described as applied biocomputational apparatuses of sorts - synthetic creatures controlled by biochemical signals. However, what if we could directly program organisms and cells using state-of-the-art digital interfaces? How much of the world could become programmable? The possibility of this vision of a programmable world lies in DNA.

From the computational perspective, DNA offers both information-storing and information-processing solutions, given that it is not just a static information medium, but a medium that actively computes upon itself by replicating and rewriting its own code. Moreover, enzymes that regulate DNA copying can check and repair the errors in data duplication, which turns DNA into a self-repairing information archive. DNA thus represents a durable and flexible computational substrate that enables scale-free, in-memory, parallel processing that bypasses the inherent limitations of von Neumann computer architecture.

However, one can move beyond the DNA itself and assemble whole computational circuits from the building blocks of DNA, RNA, enzymes, and proteins. These are analogous to electronic semiconductor integrated circuits, which consist of logical gates representing different operations with information. However, instead of electrons conducted through the network of logical gates carved on the surface of the chip, biological circuits rely on a large portfolio of bioanalogs: conducting flux of RNA polymerase alongside DNA strands, controlling the production of proteins in ribosomes, or exploiting regulatory networks in cells and bacteria.

Moreover, while the logical gates in integrated semiconductor circuits are made of transistors, their biological counterparts use transcriptors, which are based on input-output biochemical regulators, such as promoter and repressor proteins.^[1] For example, a biological version of NOT gate uses a combination of biochemical promoters and repressors to control the flux of RNA polymerase (RNAP). If the input promoter is switched ON (= value 1), then the gate produces a repressor on its output, which halts the RNAP flux (= 0). Conversely, if the promoter is switched OFF (= 0), then the output is a promoter that sends RNAP flux further (= 1).

Compared to electronic circuits, the advantage of biological circuits is little to no information loss. Moreover, they have superior spatial efficiency. Think about the Apollo Guidance Computer - the first digital computer made of silicon integrated circuits. This was fairly simple hardware composed of 5600 NOR gates, fit into a machine half a meter long and weighing over 30 kg. Compared to this hardware, its biocomputational version could be easily contained in a single cell.^[2] However, they have also major limitations - while the semiconductor circuits have distinct pathways that conduct electrons through the maze of gates, biological computation happens in the organic “soup” inside the cell, where molecules interfere with each other in an open environment. Moreover, different promoters and repressors produce distinct chemicals, so there is no universal chemical gate structure for all operations - different gates need different promoters and repressors, and they have to be insulated from each other.

Despite these design hurdles, biological circuits unlock a whole new field of computation - programming life at its root, by means of synthesising and editing DNA sequences inside cells and microorganisms. Hence, one of the possible futures of computation is its extension into bioeconomy through programming microorganisms, making them produce specific materials, recycling waste, or controlling the growth of organs.

Instead of programming computers that govern environments in which organisms exist, it will be suddenly possible to program the organisms right away. Here, biocomputing finds itself on the interface with synthetic biology, using hardware-independent coding languages such as Verilog to rewrite or update the properties and behaviour of organic matter.

This leads some scholars (such as Christopher Voigt from MIT) to quite an audacious vision of the programmable world, where modified organisms run human-installed algorithms that change the way they function or express their genome. In some instances, the programmable world is already here. Many of us are intimately familiar with mRNA Covid-19 vaccines - the feats of synthetic biology, initially shipped around the globe in the form of digitised RNA sequences and synthesised on demand by different laboratories.

These vaccines are the results of the interfacing between good old-fashioned computation and biocomputing: unlike vaccines that had to be reproduced from material samples, mRNA vaccines are essentially cloud vaccines, downloadable from email attachments.^[3]