De novo protein design

01Overview and value chain#

Markers EC: EU ATMP Regulation (EC) No 1394/2007 + FDA/NMPA biologics pathways for designed proteins | OECD: Biotechnology and health | Regulator: FDA (USA), EMA (EU), NMPA (China)

De novo protein design inverts the relationship between biology and engineering that held for the whole molecular age: instead of finding a protein that does something and improving it, the designer specifies the function — bind this target, catalyze that reaction, self-assemble into this cage — and computes a sequence that folds into a structure performing it. The discipline stands on a physical claim proven repeatedly since its founding demonstrations: that the mapping from amino-acid sequence to folded structure is computable, so a backbone never seen in nature can be specified, produced by solid-phase or recombinant synthesis, and verified against its design model by X-ray crystallography or cryo-EM. The Institute for Protein Design at the University of Washington — the field’s center of gravity under David Baker — runs this loop as an institution: a $33 million annual research budget, a $7 million grant from the Washington Research Foundation in 2026, core laboratories spanning protein production, peptide synthesis, electron and light microscopy and X-ray crystallography, more than 250 members, 50-plus scientific publications per year and over 100 patents issued. The applied output is already a product class — designed vaccines, binders and self-assembling nanomaterials — and the discipline’s industrialization runs through synthetic-biology manufacturing on the other side of the Pacific: Tianjin University’s School of Synthetic Biology and Biomanufacturing, home to the State Key Laboratory of Synthetic Biology, ranks first globally in synthetic-biology research papers in Scopus and trains roughly 2,000 postgraduates.

Key directions of de novo protein design:

1. Computational backbone and interface design (Design-Build-Test): specified geometry computed from physical principles, synthesized, and verified against the design model by crystallography or cryo-EM — the core loop the Institute for Protein Design runs across 50-plus publications a year on a $33 million annual budget.
2. Designed vaccines and therapeutics (Function by Specification): immunogens, binders and therapeutic candidates designed rather than discovered, moving through the ATMP and biologics frameworks that the FDA, EMA and NMPA operate.
3. Self-assembling nanomaterials (Protein Architectures): designed cages, lattices and arrays whose assembly is programmed by the sequence itself — the materials wing of the discipline.
4. Genome-scale synthesis as the manufacturing arm (Synthetic Biology Institutes): the design-build-test loop at chromosome and genome scale — Tianjin University’s yeast genome synthesis work anchors the synthesis-and-assembly layer that turns designs into buildable biology.

Sectoral value chain#

[function specification] ──> [computational design] ──> [gene synthesis, expression]
                                                                      │
                                                        (structure-function verification)
                                                                      │
                                                                      ▼
[product, material, therapy] <── [scale-up biomanufacturing] <── [designed protein validated]

Value chain levels#

Cross-cutting technologies of the sector:

• Structure prediction networks (Computed Folds): the models that made backbone computation tractable, now the substrate design iterates on.
• Gene synthesis and DNA fabrication (Build Layer): the industrial synthesis that turns sequences into expressible constructs at design-loop speed.
• Cryo-EM and high-throughput characterization (Test Layer): the verification instruments that close the design loop.

02US#

The US hosts the field’s institutional center of gravity and its venture layer; the design loop’s published output and patent estate are American in majority.

The Baker-lab institute, patent estate, designed therapeutics pipeline#

• Institute for Protein Design (University of Washington): housed within the School of Medicine, directed by David Baker, with 250+ members, a $33 million annual research budget and a $7 million Washington Research Foundation grant in 2026, core R&D labs for protein production, peptide synthesis, electron and light microscopy and X-ray crystallography, 50-plus publications per year and 100-plus patents issued.
• Design of vaccine and therapeutic candidates: the institute’s research focus spans de novo protein design, computational biomolecular design, vaccine and therapeutic development and self-assembling systems — the pipeline venture companies license from.
• FDA pathway for designed biologics: designed protein therapeutics file as biologics under FDA review — the regulatory classification that turns a computed sequence into a licensable product.

03CN#

China industrializes the synthesis-and-build arm of the discipline at university-institute scale, with Tianjin University the anchor institution by publication volume and training throughput.

Tianjin University’s synthetic biology complex, genome synthesis, talent throughput#

• School of Synthetic Biology and Biomanufacturing (Tianjin University): home to the State Key Laboratory of Synthetic Biology, the Frontier Science Center for Synthetic Biology and the Key Laboratory of Systems Bioengineering — a stack of national facilities dedicated to the build side of the design loop.
• Genome-scale synthesis: research spanning yeast genome synthesis, DNA information storage and artificial cell construction — the chromosome-scale end of the build layer that makes designed genomes constructible.
• Publication and talent throughput: first globally in synthetic-biology research papers in Scopus, with 20 senior national-level leading talents, 37 national-level young talents and roughly 2,000 postgraduates trained as of 2022 — the workforce the design discipline scales through.

04EU#

Europe contributes the ATMP regulatory framework that designed-protein therapeutics file under and a growing computational-design academic layer.

ATMP framework for designed proteins, European design groups, regulator readiness#

• EMA’s ATMP framework (Regulation EC No 1394/2007): designed-protein therapeutics and vaccines file under the advanced-therapy regime the EU pioneered — the classification that gives the field’s outputs a European regulatory route.
• Academic design groups across member states: European structural-biology and computational-design groups feed the field’s method base, with cryo-EM facilities among the strongest verification infrastructures anywhere.
• Regulator readiness as a field asset: a mature ATMP pathway means a designed therapeutic has a defined route to European patients — the institutional half of translation that pure software fields lack.

05Leading companies and research institutes#

06Tech stack and innovations#

The stack is a loop — specify, compute, build, verify — and its innovations are the instruments that shortened the loop from years to weeks.

1. Computational structure and interface design (Specify-to-Sequence):
   • backbone geometry and binding interfaces are computed from physical principles; the output is an amino-acid sequence that has never existed.
   • case: the Institute for Protein Design’s design-build-test loop — 50-plus publications per year, 100-plus patents issued, self-assembling systems among its research focuses.
2. Gene synthesis and genome-scale construction (Build at Scale):
   • designed sequences become expressible constructs through commercial and institute gene synthesis; at chromosome scale, whole genomes are synthesized and assembled.
   • case: Tianjin University’s yeast genome synthesis program inside the State Key Laboratory of Synthetic Biology.
3. High-throughput structural verification (Close the Loop):
   • cryo-EM and X-ray crystallography verify designed structures against their models — the step that keeps design honest.
   • case: the institute’s core labs span protein production, peptide synthesis, electron and light microscopy and X-ray crystallography in one facility stack.

07Value chains and production pipelines#

Industrial pipeline of a designed protein campaign (design-build-test-verify regime)#

┌───────────────────────────┐      ┌───────────────────────────┐
│ 1. Function               │ ───> │ 2. Computational          │
│     specification         │      │ design                    │
└───────────────────────────┘      └───────────────────────────┘
                                                 │
                                                 ▼
┌───────────────────────────┐      ┌───────────────────────────┐
│ 4. Structural             │ <─── │ 3. Gene synthesis and     │
│     verification          │      │ expression                │
└───────────────────────────┘      └───────────────────────────┘
              │
              ▼
┌───────────────────────────┐      ┌───────────────────────────┐
│ 5. Iterative              │ ───> │ 6. Scale-up               │
│     optimization          │      │ biomanufacturing          │
└───────────────────────────┘      └───────────────────────────┘

Stage 1: Function specification

The need — bind a target, catalyze a reaction, assemble into a cage — is written as a design brief with quantitative acceptance criteria.

Stage 2: Computational design

Backbone geometry, interfaces and sequences are computed; the Institute for Protein Design’s pipeline is the reference implementation of this stage as an institutional process.

Stage 3: Gene synthesis and expression

Designed sequences are synthesized to order and expressed recombinantly — Tianjin University’s national facility stack is the industrial form of this stage at genome scale.

Stage 4: Structural verification

Expressed proteins are verified against design models by cryo-EM or X-ray crystallography — the institute’s core labs keep verification inside the design loop rather than after it.

Stage 5: Iterative optimization

Verification data feeds the next design generation — the cycle that turns single successful designs into designable families.

Stage 6: Scale-up biomanufacturing

Verified designs move to recombinant production and, for therapeutics, through the FDA, EMA ATMP or NMPA frameworks that turn a computed sequence into a licensed product.