A viral-protein library the size of a small city genome

On 10 July 2026, a research team at the Centre for Infectious Disease Research in Berlin published a catalogue of more than 100,000 individual proteins drawn from the genomes of about 1,500 viral species, an order-of-magnitude jump over the most ambitious previous effort and the first made fully open to outside laboratories. The library is the work of nearly 100 scientists across Europe and North America and is being distributed through a non-profit mechanism rather than a commercial supplier, a deliberate choice the authors link to pandemic preparedness.

The bet is simple. A virus, reduced to its working parts, becomes a drug target. If researchers can map every protein that an emerging pathogen uses to enter a human cell, evade an immune response, or copy itself, vaccine designers and small-molecule chemists can move before the first human clusters become an epidemic. For most of the past two decades that map was hand-drawn, one paper at a time, with each lab producing a few hundred proteins at most.

The mechanics of a working library

Producing proteins at this scale required standardising what had been bespoke protocols. The team behind the new release spent years reducing each step to a written recipe: insert a gene into a bacterial expression vector, grow it in a cell culture, purify the protein by affinity column, verify it by mass spectrometry. Each of the roughly 100,000 proteins passed quality control, with the project's most striking claim being that more than 70% of those attempted ended up soluble and characterisable, a yield that the authors call high for any structural-biology resource and a necessary precondition for the work to be useful.

The library is, in effect, a phonebook for the viral kingdom. A researcher hunting for an antibody target against, say, a newly sequenced bat-borne coronavirus can look up the closest known relatives, request physical samples, and run binding assays within weeks. The same logic applies to drug discovery: a structure of a viral enzyme in complex with a candidate inhibitor can be solved by mail rather than built from scratch.

Why an open catalogue, not a commercial one

The decision to release the proteins through a non-profit rather than a contract-research-organisation catalogue is itself the story. Several of the largest reagent companies in the United States and Germany already sell viral-protein arrays, including panels focused on SARS-CoV-2 variants and influenza subtypes. The new resource complements that commercial market by providing thousands of proteins that no pharmaceutical catalogue covers, particularly from under-studied viral families that infect livestock, bats, and arthropods. The framing the authors prefer is that pandemic preparedness is a public-goods problem: the marginal cost of releasing one more protein is small, the marginal value to a future responder is large, and the commercialisation premium is therefore wasteful.

This sits inside a slow but durable shift in biomedical infrastructure. Initiatives like the National Institutes of Health's Structural Biology program in the United States, the European Bioinformatics Institute in Hinxton, and China's National Protein Science Center have all invested in shared data resources over the past five years. The new library does not displace those resources, but it makes them more usable: each protein is paired with a structure prediction and a sequence file, so wet-lab researchers do not have to recompute either.

What the library cannot yet do

The catalogue is broad rather than deep. Almost all the proteins were expressed in a single bacterial host, which means those requiring mammalian chaperones, post-translational modification, or membrane insertion may be misfolded or absent. The team acknowledges the gap in the published manuscript, noting that the 1,500 viral species covered represent roughly one-tenth of all known viral diversity.

A more pointed critique is the library's silence on viral dynamics. A static protein does not capture the conformational changes a virus undergoes during cell entry, fusion, and replication. Antibody and drug screens using these reagents may therefore identify targets that look promising in a test tube but fail in a cell. Two senior virologists consulted by SciXchange called for parallel investment in live-virus assays, citing the gap between soluble protein and infectious biology.

What to watch next

Three dates matter in the near term. A consortium in Geneva plans to use the library to screen the entire World Health Organization Essential Medicines list for off-target antiviral activity, with results expected by the first quarter of 2027. A United States National Institute of Allergy and Infectious Diseases programme will fold the resource into its pre-clinical antibody platform in the coming months. And a Chinese Academy of Sciences group in Shanghai has separately signalled it will use a comparable in-house catalogue to cross-check the European release.

The bigger structural question is whether open reagent libraries like this one will continue to attract funding. Pandemic preparedness budgets in the United States and Europe have been politically contested since 2024; the cost of a single resource of this scale is reported in the mid-single-digit millions of euros. If this release performs as advertised, the case for replication is strong. If it sits underused, the larger bet on open biomaterials will take longer to pay off than its defenders hope.

Monexus framed this release as a public-infrastructure story first and a science story second — a deliberate choice given that the methodology itself, structural biology at scale, is increasingly a question of who funds and who runs shared resources.