OpenWorm brings simulated life one step closer with ‘real’ digital muscles

The important thing about Caenorhabditis elegans, also called a roundworm or nematode, is how simple it is. A mouse may be small, but it’s still a mammal, and composed of many billions of cells. Even a fruit fly is a relatively complex organism, and as scientists manipulate its genes they see all kinds of emergent properties arising from the incredible number and variety of cells affected. But c. elegans is both tiny and enormous, simple and complete, a multicellular organism with all the features of a real animal — but just barely. The world’s most important worm has things like a mouth and digestive tract, reproductive organs, and neurons. At just 1mm long, though, all those features come packed into an organism with just 959 total cells — and hey, that’s a low enough number of cells that we might actually be able to figure out what each of them does! For years, a project called OpenWorm has been trying to do just that, and this week it reached a major milestone on that path: muscles.

On a nematode, muscles run in four bands along the length of the body and allow it to move various segments of the body back and forth. In the video below, the OpenWorm team used its hard-coded abstraction for muscular contraction to drive the worm forward through a medium of simulated water particles. Though it takes place over just a fraction of a second, the simulation is so complex, it took a full three days to render. Each muscle segment receives an independent contractile signal, just like the real things. It’s not literally simulating signaling and contraction, modeling the rush of calcium ions or the ratcheting of myosin, but boiling biological processes down to their practical effects and hard-coding those into the model. The sum total of the team’s work is a nematode that swims naturally though water, closely mirroring the movements of a real nematode.

This is important for a number of reasons, not the least of which is that it is cool. Beyond that, though, OpenWorm represents a novel, non-arbitrary way to quantify just how well we really do understand physiology. It’s ultimately a metaphor, as anything will be until we can simulate biology down to the quantum level, but even so, medicine is waiting eagerly for the ability to compile and run a biological entity. Though a nematode is certainly not human, when OpenWorm gains the ability to simulate a nematode’s tiny, 302-neuron brain in its entirety, the project will have made the first step toward true, usefully modern brain science. As mentioned, the only truly important difference between a nematode and mammalian brain is scale and complexity — and humans are nothing if not good at improving things we’ve already done.

This is what a real nematode’s movement looks like.

The simulation is complex enough, which several different existing and custom simulation models running at once; when the worm moves, the Sybernetic Engine determines the effect on surrounding water particles and a bio-simulation engine called Geppetto models the worm. All muscle cells were modeled one-to-one on the worm’s body, meaning that every contractile unit on the real animal is accounted for in this program. That sort of uncompromising, ultra-literal simulation is what gave rise to OpenWorm’s ultimate goal: to simulate the full nematode brain well enough that we can carry out preliminary neurological experiments on a computer, rather than in a lab.

Though it’s in the earliest stages, supporters already talk about the brain simulation effort as an attempt at immortality; if we know how brain’s connections function in terms of data, then we could simulate that data and download human consciousness into a machine — you know, maybe. These are the sorts of things people naturally consider, though, when faced with the prospect of recreating life’s basic processes. It’s fun to let yourself get caught up in the excitement.

Now read: HIV structure cracked using GPU-based simulations