The University of Copenhagen has opened the Novo Nordisk Foundation Center for Protein Design with a grant of 700 million Danish kroner, about 109 million US dollars, led by Dek Woolfson. Opened in August 2025, the center designs proteins from first principles across biology, chemistry, drug…
The University of Copenhagen has opened the Novo Nordisk Foundation Center for Protein Design , backed by a 700 million Danish kroner grant from the Novo Nordisk Foundation . That is approximately 109 million US dollars, and it amounts to more than 100 million US dollars in private philanthropic funding committed to a public university. The center opened in August 2025, nearly a year ago, and is led by Professor Dek Woolfson .
The size of the commitment is itself a finding. A private philanthropic foundation has decided that the ability to build proteins from scratch, rather than borrow them from nature, warrants support at nine figures. The center is a bet that designed proteins will reshape medicine, materials, and environmental science, and one of the clearest signals that protein design has moved from promising idea to global priority.
The physical center is only just coming together. Center staff moved into the new laboratories and offices about one month before Woolfson described the timetable. His phrasing is exact: "We haven't had the official inauguration of the CPD yet and moved into our new labs and offices just a month ago." Funding and leadership were in place before the laboratories were ready, and the official inauguration, which had not taken place at the time of writing, will mark the moment the center's work fully begins.
The center's operation is spread across four departments of the University of Copenhagen. Its main operation is shared between the Department of Biology and the Department of Drug Design and Pharmacology , with further activity in the departments of Chemistry and Computer Science. It draws together biologists, chemists, drug designers, and computer scientists under a single program with a single stated aim: to design proteins from first principles, engineering new sequences that fold into structures and functions nature never evolved.
The geography is part of the design. The relevant departments sit within walking distance of one another in Innovation District Copenhagen , a proximity intended to turn separate disciplines into a single working group. That arrangement is more than convenience. Protein design is a loop, not a single step. A computer scientist proposes a sequence, a chemist synthesizes it, a biologist folds and characterizes it, and a pharmacologist asks whether the molecule does anything useful. When the people doing those steps are in different buildings, each handoff costs time and information. When they can reach one another's labs in minutes, a failed fold in one building becomes the subject of an afternoon conversation in another.
Each department contributes a distinct piece of the pipeline. Biology supplies the experimental end: expression systems, folding assays, and biophysical characterization of whatever the computers propose. Drug design and pharmacology gives the program an orientation toward therapeutic molecules. Chemistry provides synthesis for the molecules that biology cannot express. Computer science supplies the search machinery that generates candidate sequences. The center is, in effect, a vertical organization inside a horizontal university.
The center is therefore as much a building project as a research program, and its first year has been measured in moves rather than in publications. Proximity is a bet that separate disciplines will merge into a working group. The street plan only gets people talking; the science has to keep them talking.
De novo protein design differs from conventional protein engineering in a fundamental way. Most engineering takes a natural protein and modifies it: a few mutations here, a domain swap there, with the hope of preserving the fold while changing a function. De novo design abandons the natural starting point. The researcher specifies a target structure, often a fold that has never appeared in nature, and then searches for an amino acid sequence predicted to fold into it. Designing from first principles means exactly this: no natural template, only the physical rules of folding and an intended outcome.
The search is the hard part. A protein sequence is a string of choices among twenty amino acids. For a chain of 50 residues, the number of possible sequences is 20^50, on the order of 10^65. Nature has sampled only a vanishing fraction of that space, and most of what it has sampled is tied to the functions life happens to need. A designer needs more than a sequence that can fold; the sequence must fold specifically into the target structure and not into any of the countless alternative states that a floppy chain could adopt. Hydrophobic side chains must pack into a buried core. Backbone amides must find hydrogen-bonding partners. Charged residues must sit where water can reach them. The whole molecule must hold its shape in solution at body temperature.
The 2024 Nobel Prize in Chemistry recognized the two computational advances that made this search tractable: the prediction of protein structure from sequence and the design of proteins from scratch on a computer. The two are inverse problems. Prediction asks what a given sequence folds into. Design asks what sequence folds into a given structure. Design is the harder direction, because it requires the model not merely to recognize a fold but to find one of the rare sequences that adopts it as a stable and unique state.
The current generation of design tools couples two approaches. Physics-based energy functions estimate the free energy of a sequence in a proposed fold, penalizing buried charges, exposed hydrophobics, and strained backbone geometry. Learned models trained on the known structures of natural proteins propose sequences consistent with a desired architecture. Neither approach is exact. Energy functions are approximations, and learned models inherit the biases of their training data. Every sequence that leaves a computer carries the errors of both.
The chemistry of folding sets the failure modes. A design with a cavity in its core will collapse. A design that buries a charged residue will misfold to escape it. A design that satisfies every local requirement but ignores the global energy landscape will adopt a different structure entirely, and the gap between the prediction and the reality may be invisible until the molecule is made.
For peptide research, the relevance is direct. De novo design can generate new protein and peptide sequences, scaffolds, and functions not found in nature: short peptides locked into defined conformations, miniature proteins assembled around a designed core, or scaffolds that present a chosen binding surface. These are the kinds of molecules that feed peptide-based discovery, either as therapeutic leads themselves or as frameworks that carry other sequences. A designed protein must also do more than fold in a simulation. It has to fold in water, at a useful temperature and concentration, and often in serum, where proteases are waiting. Stability is a design input, not an afterthought.
Woolfson's appointment connects the new center to one of the field's longest-running programs. He pioneered de novo protein design, led the BrisSynBio synthetic biology center, holds a chair in chemistry and biochemistry at the University of Bristol , and is a Fellow of the Royal Society. He keeps the Bristol chair and holds a joint appointment between Bristol and Copenhagen, an arrangement that gives the new center access to a research lineage that predates the current surge of institutional money.
What has changed is the scale of support. Protein design has been practiced for years; what is new is a private foundation backing a university-wide program at this scale and putting a drug design department at its center. That choice tells the peptide community how the funders see the field: not as a computational curiosity, but as a route to molecules that matter clinically.
The grant is also notable for where it landed. Philanthropic capital on this scale often funds named buildings and endowed chairs at private institutions. Here it funds an ongoing research program at a public university, with the agenda set by the science rather than by commercial timelines. The trade-off is accountability of a particular kind: a nine-figure gift invites a public accounting of what the money produced, and the accounting will be written in published molecules, not press releases.
The joint appointment cuts both ways. Copenhagen gains a leader who has spent his career on this problem; Bristol retains a professor who keeps an independent program running. The arrangement is unusual in academic science, where senior posts tend to be exclusive, and it keeps the center connected to Bristol's synthetic biology community as it builds its own. The timing matters as well. The Nobel Prize for 2024 recognized protein design on computers, and it raised the public stakes for any institution claiming to practice the field. The Copenhagen center will be read as one of the field's institutional answers to that recognition.
No designed sequence is accepted on the strength of a computer model. The field's standard of proof is experimental: the gene is synthesized, the protein is expressed and purified, and its behavior in solution is measured against the prediction.
For a first-pass design, confirmation usually begins with a circular dichroism spectrum to ask whether the molecule adopts any regular secondary structure at all. Size-exclusion chromatography asks whether it forms the intended monomer, dimer, or larger assembly, or whether it has aggregated into something the model never proposed. Thermal denaturation asks how much heat the fold tolerates. The decisive test is structural: an X-ray crystal structure, an NMR structure, or a cryo-electron microscopy reconstruction showing the atoms where the model placed them. For a designed protein, that means the experimental density must match the computational model residue by residue.
The process is slow because it is honest. A sequence that buries a charged side chain, leaves a cavity in its core, or finds a lower-energy alternative fold will come back from the lab as an aggregate or a clear solution that folds to nothing in particular. The designer revises, and the revised design re-enters the queue. Computational tools can propose large numbers of candidates in a single run, but every candidate that matters is eventually tested in a tube, and the tube is the rate limit. Failure is cheap in computation and expensive in the lab, which is why design groups publish validated structures sparingly and only after years of iteration.
This is the context in which the center's first-year silence should be read. The absence of reported research results from Copenhagen is not evidence of failure; it is the normal shape of a field in which a single design-test-revise cycle can take months. The milestones that would indicate the center is delivering are specific. A designed sequence whose experimentally determined structure matches the model. A designed binder that recognizes a chosen target in a biochemical or cellular assay. A designed enzyme with an activity nature never evolved. A designed material that assembles exactly as programmed. None has been reported from the center yet, and Woolfson's own timeline makes clear why: the center moved into its laboratories only recently.
The center's structure points directly at therapeutic applications. Drug design has a seat at the table: the Department of Drug Design and Pharmacology shares the main operation with the Department of Biology. For peptide researchers,…
Related reading: First Woman Wins Emil Fischer Medal for Synthetic Peptide Research, Cyclized Cecropin A Peptide Reaches Phase IIa for CRAB Pneumonia, FDA Panel Backs Six Unapproved Peptides for 503A Compounding List, GLP-1 multi-target therapies and manufacturing near 2025 milestones.