Member of NVIDIA Inception
Reactions nature was never optimized to solve.
DeNovoLabs designs enzymes that don't exist in nature. We build the protein and the metal cofactor together, starting with the forever chemicals nothing alive can break.
Member of NVIDIA Inception
PFAS are a class of synthetic chemicals prized for their durability. That same durability makes them persistent, mobile, and remarkably difficult to destroy once they enter the environment.
PFAS are defined by carbon–fluorine bonds — among the strongest single bonds in organic chemistry. The very feature that makes them useful is what makes them nearly indestructible.
Because they resist natural breakdown, PFAS accumulate in water, soil, and living tissue over time, spreading far from where they were first released.
Most current approaches capture and concentrate PFAS rather than destroy them — shifting the burden downstream instead of eliminating it.
Among the strongest single bonds in organic chemistry — the structural reason PFAS resist conventional breakdown.
Biology can't help here either. Carbon–fluorine bonds barely exist in nature, so evolution never built an enzyme to break one. The tool doesn't exist yet. That is the gap we design into.
Designing a protein to do something nature never asked it to do was, until recently, out of reach. Several independent advances now make it tractable.
Generative models now design proteins toward a target function, compressing what used to take years of trial and error into computation.
Candidate designs can be ranked and triaged computationally — so experimental effort goes to the most promising before a single sample is made.
Gene synthesis and automated wet-labs make each design–build–test cycle faster and far less expensive than it once was.
Tightening global limits on PFAS are turning genuine destruction — not just capture — from a preference into a requirement.
The same approach extends beyond PFAS to other reactions with no natural enzyme: plastic breakdown, carbon conversion, and pharmaceutical chemistry. PFAS is the first proof, not the ceiling.
Enzymes already run the most precise chemistry on Earth. We believe they can be designed deliberately — directed at problems evolution never had reason to solve.
We design reaction-first. We start from the exact bond transformation that has to happen, then design two things together: a protein that grips and orients the target molecule, and a metal cofactor that drives chemistry a plain protein can't. That co-design lets us reach reactions natural enzymes never evolved to do.
We specify a target reaction and design candidate enzymes in silico, exploring sequence and structure space far faster than the lab alone allows.
Promising candidates move to the wet lab to be expressed and characterized — grounding computational predictions in measured results.
Experimental data feeds back into design. Each cycle sharpens our models and moves candidates closer to the performance a real-world application needs.
PFAS are urgent, global, and chemically unforgiving — exactly the kind of problem our platform is built for. Proving the approach here is the clearest test of what programmable enzymes can do.
With computational proof-of-concept complete, our PFAS program is now in wet-lab validation. We are not claiming a working PFAS-degrading enzyme today — we are testing our designs experimentally to build the evidence toward one.
Established the design–build–test approach and selected PFAS — with PFOA as the flagship validation target — as the first proving ground.
Designed and computationally evaluated candidate enzymes against the flagship target, building the case for experimental work.
Expressing and characterizing leading candidates with validation partners to test predictions against measured activity.
Use experimental results to refine designs cycle over cycle — the aim being enzymes effective under realistic conditions.
A small, focused founding team — and a clear picture of the specialists and partners we are building toward.
Sets the company's direction and strategy, leading DeNovoLabs as it builds its enzyme-design platform and partnerships.
Structural biologist with Cambridge research experience, leading the scientific approach behind our enzyme-design work.
Runs operations and execution, turning the research program into a company that ships while leading hiring, partnership operations, and go-to-market.
Leads finance and operations, structuring the funding and partnerships that carry the platform from the lab toward real-world impact.
Mechanism, kinetics, and assay design for candidate enzymes.
Producing and purifying designed proteins for testing.
Rigorous detection and measurement of reaction outcomes.
Real-world context, samples, and routes to deployment.
We're talking with investors, grant bodies, and wet-lab validation partners who want to help turn programmable enzymes into a tool for the world's hardest chemistry.