You might not believe it, but none of the pulsating little cells in this video are real. Instead, they're 3D cell avatars created by a visual effects specialist turned digital artist who is fascinated by artificial life.
For the past decade, Andy Lomas has dedicated a fair amount of after-work hours into perfecting his simulations of biological cells.
"[Hybrid Forms] is all about exploring the relationship between growth and form, in particular how really complex intricate structures can emerge from growth processes deep down at the cellular level," Lomas told me over email. "What biologists call morphogenesis."
Hailing from a mathematics background, Lomas told me he tries to "use the simplest possible system to create the most interesting results." In his case, creating intricately evolving simulated cells that look like everything from "pollen grains, reptile skin and brains."
Lomas' current project, Hybrid Forms, builds on a decades-worth of similar but slightly less complex projects such as Aggregation, Flow, and more recently Cellular Forms, which won the Lumen Prize exhibition's top digital art award in 2014. For his works, Lomas doesn't use any conventional techniques from computational biology such as introducing signals to explicitly change cell behaviour over time. His computer-generated cells do that all by themselves. Just watch as black and white simulated sim-cells indistinguishable from the real thing morph to musician Max Cooper's hypnotic music.
When Lomas created Cellular Forms, all the structures in the cells he simulated had the same properties, which were fixed at the beginning of the the sim, and unable to vary over time. For Hybrid Forms, however, Lomas wanted to add an additional layer of complexity by "using more than one cell type," which he said results in regions with "divergent properties."
In other words, cell types that spar with one another in the sim to create ever-more complex forms. To give you a rough idea of how complex, some of the final structures have over a hundred million cells, according to Lomas.
"The different cell types effectively compete with each other for space during development," explained Lomas. "When each cell is born it gets a percentage of the properties of its parent, as well as a percentage of properties of the immediately adjacent cells. The range of forms appears to be very rich, with strong echoes of protozoa and other simple biological organisms. There's definitely a lot more life-like properties in the generated forms."
Lomas' delicate black and white simulations involve reams of code. To date, he estimates that he's written between 30,000 to 50,000 lines of it to feed into his simulation engines. His cell sims are also so lifelike that when he presented his project at this year's European Conference on Artificial Life, he told me that many researchers in the room were pretty shocked they weren't looking at something biological on the screen.
Next up, Lomas wants to explore cell death as well as growth, and how multiple types of cells affect each other through "expressions of growth promoting and inhibiting chemicals." He also wants to visualize his cell sims through augmented and virtual reality, and maybe even as physical sculptures.
"One thing I'm very interested when presenting work is the idea that there is no one true way of properly seeing and understanding the data," said Lomas. "Different presentational methods tend to highlight different aspects, leading to modified understandings and interpretations."