We've moved from discovering life to designing it from basic chemicals.
In a Minnesota laboratory, researchers have assembled a living cell from non-living chemicals alone — not by editing what nature provided, but by composing something biology never wrote. Named SpudCell for its uneven shape, this minimal 36-gene creation marks the moment humanity moved from interpreting life to authoring it. The achievement is fragile, contested, and incomplete, yet it opens a threshold that science has long approached but never crossed: a biological system that is entirely transparent, entirely designed, and entirely ours to understand.
- SpudCell cannot feed itself, cannot divide cleanly, and loses genomic integrity in most of its offspring — yet it still competes, adapts, and demonstrates rudimentary evolution in a petri dish.
- The research was rejected by Cell on the grounds that it was engineering rather than biology, then released to journalists before peer review, igniting a dispute about whether the science or the story arrived first.
- A Stanford co-founder and a $10 million nonprofit are already positioning SpudCell as open-source infrastructure for biology, with AI-driven cell design projected within a decade.
- Security experts and AI companies are sounding alarms to Congress: the same tools that could accelerate medicine could lower the barrier to engineering a pathogen, placing synthetic biology alongside nuclear technology as a dual-use civilizational risk.
In a University of Minnesota laboratory, Kate Adamala's team assembled a cell from purified enzymes, synthetic lipids, and a protein-synthesis system called PURE — components that had never been alive, combined until they were. The result, SpudCell, named for its lumpy appearance under the microscope, is the first cell built entirely bottom-up from chemicals, with no evolutionary inheritance and no hidden biological logic. Every part is known. Every interaction is legible. This distinguishes it sharply from earlier milestones: when Craig Venter's team synthesized life in 2010, they transplanted a synthetic genome into an existing bacterial shell, and even their stripped-down version left 149 genes whose functions remained unexplained. SpudCell carries no such mystery.
The cell is radically minimal — 36 genes spread across seven DNA molecules, compared to E. coli's 4.6 million base pairs. It cannot build its own ribosomes or feed itself. Survival depends on nutrient-loaded liposomes that scientists deliver externally; when SpudCell's membrane fuses with one, it absorbs the energy needed to copy its own genetic material. Reproduction is equally crude: researchers force the cells through a filter to split them, a mechanical violence that leaves only 30 percent of fifth-generation offspring with a complete genome.
Yet within that fragility, something recognizable emerged. When the team created a MAX variant with a larger membrane better at capturing nutrients, it outcompeted the standard model — rising from 50 to 61 percent of the population over five generations, and nearly monopolizing resources when they grew scarce. Designed evolution, observed in real time.
The scientific community's reception has been uneven. Cell rejected the paper, with one reviewer arguing it was engineering rather than biology — a distinction Adamala's team would likely accept as a compliment. The manuscript now lives on the preprint server bioRxiv, and the decision to share it with journalists before formal peer review drew criticism from colleagues who felt the announcement outpaced the science.
What follows is both promising and sobering. Adamala and Stanford's Drew Endy have founded a nonprofit called Biotic, raising roughly $10 million to make SpudCell an open platform — a kind of Linux for biology. Researchers anticipate that within a decade, AI will be able to model entire synthetic cells virtually before a single molecule is assembled, compressing the timeline for drug development, anti-aging research, and clean-energy organisms. The same capability, critics note, could make pathogen engineering more accessible than ever, and major AI companies have already urged Congress to mandate security reviews for DNA synthesis orders.
SpudCell is not yet alive by any classical measure — too simple, too dependent, too easily broken. But it has made biology fully writable for the first time, and the question of what humanity chooses to write next may be the defining one of this century.
In a laboratory at the University of Minnesota, researchers have done something that reads like speculative fiction but is now documented fact: they built a living cell from scratch, assembling it piece by piece from chemicals that had never been alive. The cell, named SpudCell for its bumpy, potato-like appearance under the microscope, represents a fundamental shift in how humans approach biology—not by tinkering with what nature already made, but by designing life from the ground up.
Kate Adamala's team pursued what they call a "bottom-up" approach, which means they started with purified enzymes, synthetic lipids, and a cell-free protein-making system called PURE, then combined them in precise concentrations to create something that could eat, grow, and reproduce. Every component is known. Every interaction is transparent. There is no evolutionary mystery hiding inside the machinery, no "black box" of biological logic that accumulated over millions of years and no one fully understands. This is the opposite of how Craig Venter's team created synthetic life in 2010—they had taken a natural bacterium and swapped out its genome for a chemically synthesized one. Even when Venter's team later stripped that cell down to its bare minimum, 149 of its remaining genes still did things nobody could fully explain. SpudCell has no such hidden inheritance. It is, in the truest sense, designed rather than discovered.
But SpudCell is also radically minimal. It contains only 36 genes scattered across seven separate DNA molecules—compare that to the 4.6 million base pairs in E. coli or the 3 billion in humans. It cannot make its own ribosomes, the factories that build proteins. It cannot feed itself. To stay alive, it depends on tiny fat bubbles called liposomes, about 0.4 micrometers across, loaded with nutrients and energy molecules that scientists deliver to it. When SpudCell's outer membrane touches one of these delivery packages and fuses with it, the cell can absorb what it needs and activate the machinery to copy its own genetic material.
Reproduction is crude. SpudCell has no sophisticated cell division mechanism, so the researchers literally squeeze the cells through a filter membrane to force them apart. This violent method produces unstable offspring—after five generations of division, only about 30 percent of the surviving cells still carry a complete genome. Yet even in this fragile state, SpudCell demonstrates something profound: it competes. When the team created two versions—a standard model and a "MAX" version with a larger membrane better at capturing nutrient liposomes—the MAX cells outcompeted the standard ones. Starting at 50 percent of the population, MAX cells rose to 61 percent after five generations. When resources were scarce, MAX cells nearly monopolized them, reaching 70 percent of the population. Evolution, in miniature, in a petri dish, designed by humans.
The academic reception has been complicated. Adamala's team submitted their work to Cell, one of biology's most prestigious journals, and was rejected. A reviewer stated flatly that SpudCell was not really biology but engineering—as if that were a criticism rather than the entire point. The paper now sits on bioRxiv, a preprint server, awaiting formal peer review. The controversy deepened when Adamala emailed the manuscript to journalists before it had been formally reviewed, a move that struck many in the field as putting hype before science. Yet the underlying achievement remains: humanity has moved from modifying life to writing it.
What comes next is both exhilarating and unsettling. Adamala and Drew Endy of Stanford have founded a nonprofit called Biotic, which has raised roughly $10 million to open-source SpudCell as a foundational platform—imagining it as a kind of Linux for biology. Within five to ten years, researchers predict that artificial intelligence will be able to model entire cells virtually, allowing scientists to design organisms the way engineers design machines: simulate first, then assemble the DNA and proteins. This could accelerate drug development, anti-aging research, and the creation of organisms that produce clean energy. It could also, critics warn, make it far easier for someone with bad intentions to engineer a dangerous pathogen. Major AI companies have already begun calling on Congress to require strict security reviews before DNA synthesis companies fulfill orders, a recognition that the tools for creating life are becoming as dual-use as nuclear technology.
SpudCell itself is not yet a true living thing by any classical definition. It is too simple, too dependent, too fragile. But it has opened a door that will not close. For the first time, humanity possesses a biological system that is entirely knowable, entirely controllable, and entirely human-made. What we build with it next will define whether the twenty-first century becomes the century of biology.
Citações Notáveis
SpudCell is compared by some scholars to the Wright brothers' first flight in biology, marking a critical step for humanity in synthetic biology from 'modifying life' to 'designing life.'— Academic observers cited in the research
Drew Endy compares SpudCell to the Wright brothers' 1903 'Flyer I'—though the first flight lasted only 12 seconds, it inaugurated the entire aviation age.— Drew Endy, Stanford University and co-founder of Biotic
A Conversa do Hearth Outra perspectiva sobre a história
Why does it matter that SpudCell is built from scratch rather than modified from something that already existed?
Because it means we finally understand what we're working with. Natural cells are like old software that's been patched a thousand times—it works, but nobody knows why half of it is there. SpudCell is written in a language we chose.
But it sounds incredibly fragile. It needs to be fed constantly and can barely reproduce. How is that a breakthrough?
The fragility is almost beside the point. What matters is that we proved we could do it at all. The Wright brothers' first flight lasted twelve seconds. SpudCell is proof of concept. The question now is what we build once we know it's possible.
The competition experiment—where the MAX cells outcompeted the others—that's evolution happening in real time, isn't it?
Yes, but evolution we designed. We created the rules, the players, the resources. We watched Darwinian selection unfold in a system we fully understood. That's never happened before.
Why did Cell reject the paper? That seems like exactly the kind of thing they'd want.
One reviewer said it wasn't biology, it was engineering. Which is true. But that's the whole point. We've moved past discovering what life is and started designing what it could be. Some people in traditional biology see that as leaving the field.
And the biosecurity concern—is that real, or is it hype?
It's real. If designing cells becomes as routine as writing code, and AI can help you do it faster, then yes, the barrier to creating something dangerous drops significantly. That's why the AI companies are asking for regulation now, before it's too late.