Stand at the edge of a tide pool and you hear waves. Underneath the sound you can hear, there is another one you cannot: thousands of species — bacteria, plankton, fungi, things without common names — living, dividing, and trading chemistry inside that small pocket of water. A coral reef is not quiet. We are just deaf to most of it.
What if you could hear it?
That question is the seed of the World Genome Symphony — a project we are building to turn environmental DNA into music and light. Not a soundtrack laid over a chart. The actual genetic data of an ecosystem, thousands of species in a single sample, used to generate a score in real time. Reading all of them at once is itself recent magic: metagenomic sequencing untangles the mixed-together DNA in one scoop of water and sorts it back into thousands of separate organisms — something that only became possible when computers grew powerful enough to do the untangling. Biodiversity you can listen to. Ecosystem health you can feel in the room.
An instrument, not a visualization
Here is the part I find genuinely thrilling. The design is not "play a recording while a graph animates." It is an instrument. A musician stands on stage and plays alongside the data — a duet between a human performer and the living chemistry of a reef or a patch of soil, the organisms supplying a part no composer wrote.
To make that work, a lot has to happen in well under a tenth of a second: sequence the sample, classify what is in it, measure its diversity, turn that into a score, synthesize the audio, and render the visuals — fast enough that a performer can answer it as it happens. The design targets under 100 milliseconds end to end, running across federated machines on Lattice Protocol so the heavy computation can sit wherever the GPUs are while the sound stays in sync at the venue.
None of this is built yet. The Symphony is in its design phase — the architecture is drawn, the audio synthesis is not written, the debut is a plan and not a memory. I would rather tell you that plainly than oversell a prototype. The aim is a first performance at Muscle Beach, in Venice Beach, in the summer of 2027.
Three ways in
The design has three modes, because biodiversity should not only be a show you watch.
In Show Mode, musicians perform live duets with ocean microbiome data in front of an audience. In Gallery Mode, you walk up to an installation and a rainforest, a kelp forest, a prairie, or a city park plays its own genetic score while you stand inside it. In Educational Mode, students run the whole loop themselves: extract DNA from a sample, sequence it on a pocket-sized device, classify the organisms, design how the data becomes sound, and listen to a piece of music no one has ever heard — because no one has ever sampled exactly that handful of dirt before.
That last mode is the one I care about most. A student does not receive the Symphony. She makes one.
Why bother making biology audible
Science is already beautiful. The complexity in a gram of soil needs no decoration. But beauty you cannot perceive does not move anyone, and most of biology's beauty is locked behind microscopes and databases that only specialists ever open.
Sound is a way in that asks for no degree. You do not have to understand metagenomics to feel a reef's diversity thin out as it dies, or thicken as it recovers. The ocean is a body, and the water is its blood; the Symphony is an attempt to let a room full of people feel that pulse directly, without translation.
We are still building the instrument. But the idea underneath it is simple, and I think it is right: every organism in that sample is a real, living thing, and behind every data point is something alive. Make the data sing, and you have not invented the music. You have finally turned up the volume on a song that has been playing under your feet, and under the waves, the whole time.