Field guide

What is biocomputing?

Biocomputing means using biological material to compute or to store information. I'd split it into three branches, which are DNA computing, DNA data storage and computing with living neurons, and you'll find they're at very different stages.

Peer-reviewed sourcesUpdated 12 November 2026

Biocomputing in three results

One from each branch. You'll find the papers in the sources below.

0first DNA computation
0DNA storage density in a test
0what a human brain runs on
Part 01

Three kinds of biocomputing

KindWhat does the workA landmark result
DNA computingStrands of DNA and the enzymes that cut and join themAdleman, 1994: a small Hamiltonian path problem solved in a test tube
DNA data storageSynthetic DNA used as a hard driveErlich and Zielinski, 2017: 215 petabytes per gram in a test
Living neurons, or wetwareHuman brain cells grown on electrodesDishBrain, 2022: cultured neurons that got better at Pong

I'd split it this way because the three have very little in common apart from the biology. DNA computing and DNA storage use molecules that aren't alive, while the third kind uses living cells that have to be fed and kept warm (and that die, which you'll read about in the guide on how long brain organoids live).

Part 02

DNA computing: Adleman's test tube

If you want a start date, I'd use 1994. Leonard Adleman encoded a small graph in DNA molecules and solved an instance of the directed Hamiltonian path problem (you're looking for a route that visits every point exactly once), with the operations done by standard lab protocols and enzymes. His paper in Science called it a demonstration that computing at the molecular level was feasible, and I think that's the fair way to read it. It worked, but as far as I know nobody has turned it into a general purpose computer since.

Part 03

DNA data storage: where the numbers are big

I'd say this is the branch that's furthest along. In 2012 George Church's group at Harvard wrote a 5.27 megabit book into DNA and read it back with ordinary sequencing. Then in 2017 Yaniv Erlich and Dina Zielinski stored a whole computer operating system, a film and some other files (2.14 million bytes in all) and got every byte back. They also pushed the density as far as they could and still recovered the data perfectly at 215 petabytes per gram of DNA.

You shouldn't read that as a product you can buy, though. Church's paper says new synthesis and sequencing technology is making DNA an increasingly feasible place to store data, and I'd take "increasingly" at its word (it's research for now).

Part 04

Living neurons: the part people call wetware

This is the branch the rest of this site covers. In 2022 the DishBrain team put neurons on a chip (a multielectrode array) and they got better at a simulated Pong game within five minutes. In 2023 a group writing in Frontiers in Science named the field organoid intelligence, and research groups can now get access to living organoids through FinalSpark's Neuroplatform or buy a Cortical Labs CL1. The appeal is energy, since your brain runs on about 20 watts (the brain versus GPU guide has the figures), but I wouldn't claim any of these systems beats silicon at a real task yet (I've compared them with neuromorphic chips).

Part 05

Questions

What is biocomputing?

I'd define it as computing or storing data with biological material. So far that means DNA computing, DNA data storage and living neurons on electrodes.

What was the first DNA computer?

Leonard Adleman's 1994 experiment is the usual starting point. He encoded a small graph in DNA and solved a directed Hamiltonian path problem with standard lab enzymes.

How much data can DNA hold?

In a 2017 test, Erlich and Zielinski recovered data perfectly at 215 petabytes per gram of DNA. That's a lab figure, though, and I wouldn't treat it as a product spec.