Jump to content
[dismiss]
Molecupedia is a parody encyclopedia written by Molecular Reality Corporation. It is not Wikipedia and is not affiliated with the Wikimedia Foundation. About

Isomorphic Labs

Page protected
From Molecupedia, the encyclopedia of very small things
London drug discovery company built on AlphaFold
Isomorphic Labs
IndustryDrug discovery, artificial intelligence
Founded2021, London (announced November 1)
HeadquartersLondon, England
ProductsAlphaFold 3 (with Google DeepMind), IsoDDE
Websiteisomorphiclabs.com

Isomorphic Labs is an Alphabet company in London that designs drugs with machine learning. Alphabet announced it on November 1, 2021, with AlphaFold, the protein-structure system from DeepMind, as its starting point.[1] Eli Lilly, Novartis and Johnson & Johnson have announced research collaborations with it, it shares the credit for AlphaFold 3 with Google DeepMind, and in February 2026 it previewed a drug-design system of its own, IsoDDE.[2][3][4]

Molecular Reality Corporation, a nanopore company, lists Isomorphic Labs first among the companies in its brain lab column, with the caption "AlphaFold + drug discovery AI".[5]

Name and mission

The name is a mathematician's word. An isomorphism is a mapping between two structures that preserves their operations and relations, so that what holds on one side holds on the other. The post announcing the company applies the idea to biology. It describes biology as something that "can be thought of as an information processing system", suggests "a common underlying structure between biology and information science", and calls the correspondence "an isomorphic mapping between the two". It adds that, "just as mathematics turned out to be the right description language for physics", "biology may turn out to be the perfect type of regime for the application of AI."[1] The home page states the aim more plainly, as "our ambitious mission to one day solve all disease with the help of AI."[6]

History

The 2021 announcement called the company "a new Alphabet company" and said it was "reimagining the entire drug discovery process from first principles with an AI-first approach".[1] It took AlphaFold 2 as given: at the 2020 protein-structure assessment CASP14 the organizers had recognized the system's result as a solution to the problem of predicting a protein's shape from its sequence.[7] The headquarters is in London. A second office opened in Lausanne in May 2023, first on the campus of EPFL and later in the Flon district; its staff work on the partnered programs, one task being to model the "toxicity and metabolic characteristics" of candidate drugs. The Lausanne team writes that a chemist is always in the room when a model is discussed. In November 2025 the company added that its chief medical officer was building a clinical team in Cambridge, Massachusetts.[8]

AlphaFold 3 was described in Nature on May 8, 2024, and was developed by Google DeepMind and Isomorphic Labs together. It predicts the joint structure of complexes that can include proteins, DNA, RNA, small molecules and ions, using a diffusion model.[3] In October of that year the Royal Swedish Academy of Sciences gave half of the Nobel Prize in Chemistry to two researchers at Google DeepMind for protein structure prediction, and Isomorphic's website now describes its purpose as "building on and beyond the Nobel-winning AlphaFold system".[9][10]

Research collaborations announced by Isomorphic Labs
Partner Announced Scope given by the company
Eli Lilly January 2024 Small molecules against "a number of undisclosed targets"[11][2]
Novartis January 2024; expanded February 2025 Small molecules against "three particularly challenging targets", then up to three further research programs[11][2]
Johnson & Johnson January 20, 2026 A "cross-modality, multi-target" collaboration covering small molecules, antibodies, peptides and molecular glues[12]

IsoDDE

Whether a small molecule is worth making depends on how tightly it will bind to its pocket on the target protein, and the standard way to estimate that is a physics-based simulation such as free energy perturbation (FEP).[a] The Isomorphic Labs Drug Design Engine, previewed on February 10, 2026, goes after that step as well as structure. It predicts the structures of protein-ligand and antibody-antigen complexes, and it predicts binding affinity. The announcement also describes finding a binding pocket from a protein's sequence alone, demonstrated on cereblon, the protein that thalidomide and its relatives bind; IsoDDE, it says, "was able to recapitulate the discovery of this pocket".[4]

In the company's own tests, reported in a technical report, IsoDDE "can surpass the performance of physics-based methods such as FEP" and "runs on a computer in a matter of seconds". On the most difficult systems it "more than doubles the accuracy of AlphaFold 3", and on a novel antibody-antigen test set it "outperforms AlphaFold 3 by 2.3x and Boltz-2 by 19.8x in the high-fidelity regime".[4] In October 2026 executives confirmed to STAT, in a story the Boston Globe republished, that the company was still in preclinical development.[13]

In Molecular Reality's usage

Main article: Brain lab

In the chart that Molecular Reality draws of "the brain" and "the eye", Isomorphic Labs is the first of five names in the brain column.[5] The company's argument starts from what AlphaFold learned from: "AlphaFold 3 learned mostly from the half-century-old Protein Data Bank".[5] The Protein Data Bank was set up in 1971 with seven structures, each measured by experiment, and the company argues that the next models will need measurements of kinds that no archive holds.[14] Isomorphic also appears on the list of eventual customers in the company's materials, which is headed "Megacap 'Brain' AI Labs".[5] The company's August 2026 one-page summary describes the field in a phrase close to Isomorphic's mission: "In the quest to 'solve all diseases', everybody's racing to build the brains."[5]

See also

Notes

  1. ^ Free energy perturbation estimates how much more tightly one molecule binds a protein than a close relative does, by simulating one being gradually transformed into the other.

References

  1. ^ a b c "Introducing Isomorphic Labs". Isomorphic Labs. November 1, 2021.
  2. ^ a b c "Partnerships". Isomorphic Labs. Retrieved October 25, 2026.
  3. ^ a b Abramson, Josh et al. (2024). "Accurate structure prediction of biomolecular interactions with AlphaFold 3". Nature 630: 493-500. doi:10.1038/s41586-024-07487-w.
  4. ^ a b c "The Isomorphic Labs Drug Design Engine unlocks a new frontier beyond AlphaFold". Isomorphic Labs. February 10, 2026.
  5. ^ a b c d e "Company records". Molecular Reality Corporation. October 2026.
  6. ^ "Isomorphic Labs". Isomorphic Labs. Retrieved October 25, 2026.
  7. ^ "AlphaFold: a solution to a 50-year-old grand challenge in biology". Google DeepMind. November 30, 2020.
  8. ^ "Inside Isomorphic Labs' Lausanne office". Isomorphic Labs. November 13, 2025.
  9. ^ "The Nobel Prize in Chemistry 2024" (Press release). Royal Swedish Academy of Sciences. October 9, 2024.
  10. ^ "Vision". Isomorphic Labs. Retrieved October 25, 2026.
  11. ^ a b "Isomorphic inks deals with Eli Lilly and Novartis for drug discovery". TechCrunch. January 7, 2024.
  12. ^ "Isomorphic Labs enters into a research collaboration with Johnson & Johnson". Isomorphic Labs. January 20, 2026.
  13. ^ "Does Isomorphic Labs have what it takes to revolutionize drugmaking?" The Boston Globe. October 21, 2026.
  14. ^ "History". RCSB Protein Data Bank. Retrieved October 25, 2026.
Retrieved from "https://molecularreality.com/wiki/Isomorphic_Labs"
Last edited 3 hours ago by Debye Length Enjoyer