AlphaFold Database adds predicted protein pairs for 2,800 viruses to aid pandemic research

A free database now shows how proteins from thousands of viruses may pair up. It could give vaccine and drug researchers a head start, but every prediction still needs checking in a laboratory.

By Utopia Data & AI Team6 min read

A modern two-storey building with a wide timber-lined canopy and pale panelled walls, fronted by two round trees in autumn yellow and a blurred strip of grass
Licensed photo. The Thornton Building at EMBL's European Bioinformatics Institute on the Wellcome Genome Campus near Cambridge, England, October 2024. EMBL-EBI runs the AlphaFold Database with Google DeepMind. The photo shows no part of the viral prediction work. Photo: Jeff Dowling/EMBL-EBI/Wikimedia Commons, CC BY-SA 4.0[7][2]

In brief

  • On 24 September 2026, EMBL-EBI, Google DeepMind, NVIDIA and six other partners opened predicted protein complex structures for more than 2,800 viruses in the AlphaFold Database.[2][1]
  • The database says the effort covered 2,812 viral proteomes from 23 human-relevant families and kept 8,028 high-confidence protein pairs.[3]
  • The partners say the predictions cannot show how mutations change a virus. Malaysia, with 72,049 dengue cases recorded this year, is not named in the release.[2][5]

What was released

Scientists preparing for the next pandemic have gained a large, free reference library. The AlphaFold Database, run by the European Molecular Biology Laboratory's European Bioinformatics Institute (EMBL-EBI), now holds AI-predicted 3D structures showing how proteins from more than 2,800 viruses may fit together.[2]

The release was announced on 24 September, a day before a United Nations General Assembly high-level meeting on pandemic prevention, preparedness and response in New York. The partners say the aim is to have structural knowledge ready before an outbreak, rather than building it after one starts.[2][1]

Close-up of a striped Aedes aegypti mosquito standing on human skin, its thin mouthpart pierced into the skin and turning red with blood
Licensed photo. A female Aedes aegypti mosquito, the main carrier of dengue, feeding on a person, photographed for the US Centers for Disease Control and Prevention in 2006. The release does not say which dengue proteins it includes. Photo: James Gathany/CDC/Wikimedia Commons, public domain[8][5]

What the database now holds

The database's new Pandemic Preparedness Portal says the team predicted structures across 2,812 viral proteomes from 23 virus families relevant to human health. It kept 5,279 high-confidence pairs of different proteins and 2,749 pairs of identical ones. Another 4,681 identical pairs came from a separate Viral AlphaFold Database project.[3]

Nature reports that the researchers worked from 41,774 viral protein sequences and predicted nearly 1.7 million pairs in total. Only the most reliable made it into the main database, although all predictions were made public. The viruses range from common-cold families to emerging threats such as mpox, measles and hepatitis B.[4][2]

How the predictions were made

The structures were inferred with AlphaFold2, Google DeepMind's protein-folding model, and sped up with NVIDIA's BioNeMo Inference Runtime so the team could run thousands of viral proteomes. NVIDIA has also released the GPU workflow it used as open source, so labs can predict structures for their own targets.[1]

The partners are EMBL-EBI, Google DeepMind, NVIDIA, Seoul National University, Sungkyunkwan University, the University of Glasgow, the Swiss Institute of Bioinformatics and the Coalition for Epidemic Preparedness Innovations. NVIDIA says about 30% of the protein interactions added are new to science, with no match in the main experimental structure archive.[2][1]

Hundreds of large bats flying across an orange and blue dusk sky above a calm bay, with a small blue boat and a forested island below
Licensed photo. Large flying foxes leaving their roost at dusk near Rinca in Komodo, Indonesia, May 2015. Fruit bats of this family are the natural host of Nipah virus, according to WHO. These bats are not linked to any outbreak. Photo: Charles J. Sharp/Wikimedia Commons, CC BY-SA 4.0[9][6]

What the predictions cannot tell scientists

EMBL-EBI is explicit about the limits. The predictions do not show how genetic changes affect a virus or how it interacts with its host. They cannot show what makes a virus more deadly or transmissible, and they cannot be used to engineer viruses that infect humans. Real behaviour still has to be tested in a laboratory.[2]

Nature adds two gaps. The models leave out the sugar molecules that help many viral proteins hide from the immune system, and many complexes larger than pairs are missing, including the three-part coronavirus spike. Flaviviruses such as dengue and Zika have lacked good entries because their proteins are cut from one long chain, a problem this effort tried to address.[4]

Why Malaysia has a stake

Malaysia lives with viral disease every year. The Ministry of Health's iDengue portal recorded 72,049 dengue cases and 66 deaths between 4 January and 27 September 2026, with Selangor alone accounting for 30,523 cases. The new portal also has a separate view for neglected tropical diseases.[5][3]

The World Health Organization lists Malaysia among the countries that have reported Nipah virus, whose case fatality rate it estimates at 40% to 75%. Fruit bats are its natural host, and there is still no treatment or vaccine. EMBL-EBI's interim director says open data lowers barriers for scientists in low-resource settings who face outbreaks first-hand.[6][2]

Why Malaysia should care

Malaysia has recorded more than 72,000 dengue cases this year and has seen Nipah outbreaks before. Free viral structure data lowers one cost for local researchers, but no Malaysian body is part of this release.

University virologists

Free structural leads for local viruses.[3]

Practical move: Search the portal for current projects.

Vaccine and drug startups

Cheaper first look at targets.[2]

Practical move: Budget lab work to confirm predictions.

Health agencies

Better outbreak research groundwork.[6][5]

Practical move: Map gaps for Nipah and dengue.

What Malaysians can do now

  1. Researchers can check whether viruses they already study appear in the Pandemic Preparedness Portal and compare the predicted pairs with their own data.
  2. Treat each structure as a hypothesis, use its confidence label, and plan experiments before relying on it for vaccine or drug work.
  3. Labs with GPUs can test NVIDIA's open prediction pipeline on proteins the release does not yet cover.

What we still do not know

What we still do not know

  • Which dengue and Nipah proteins passed the high-confidence bar.
  • How many predictions will hold up in laboratory tests.
  • Whether larger complexes such as spike trimers will be added.

Sources

  1. How Open Science Can Help Researchers Prepare for the Next PandemicNVIDIA, 24 September 2026
  2. AlphaFold Database adds viral protein complexes to support pandemic preparednessEMBL, 24 September 2026
  3. AlphaFold Protein Structure DatabaseEMBL-EBI
  4. AlphaFold 'goes viral': database adds protein complexes of common virusesNature, 24 September 2026
  5. iDengueMinistry of Health Malaysia and Malaysian Space Agency
  6. Nipah virusWorld Health Organization, 29 January 2026
  7. File:EMBL-EBI Thornton Building Exterior.jpgWikimedia Commons, 25 October 2024
  8. File:Aedes aegypti CDC9177.tifWikimedia Commons
  9. File:Large flying foxes (Pteropus vampyrus) in flight from Pulau Kalong Rinca.jpgWikimedia Commons, 20 May 2015

All three images come from Wikimedia Commons, with licence, author and date read live from the Commons API on 28 September 2026: two Creative Commons photographs and one US government public-domain photograph. None shows a viral protein, the prediction work or any Malaysian laboratory, and each caption states what the image actually shows. No image was generated, composited or retouched; source-preserving framing was the only change made.

Last fact check: 28 September 2026. Factual correction? devteam.utopia@gmail.com.

More AI News

Back to AI NewsAll 201 stories