Loading
What changed between revision 1 and revision 4, compared over the article's markdown source with its frontmatter — so a change to status, evidence or reversibility shows up as the edit it is.
Older · revision 1
Import· initial corpus import, not an edit
Initial import of content/virtual-cell-models.md — the filesystem corpus, unchanged. Not an edit.
19.5 kB
Revision 1
Revision 4
lines 8–14 → 8–147 unchanged lines not shown
categories: ["foundations", "genetics"]categories: ["foundations", "genetics"]tags: ["machine learning", "systems biology", "simulation", "computational modelling", "single cell", "foundation models"]tags: ["machine learning", "systems biology", "simulation", "computational modelling", "single cell", "foundation models"]summary: "Computational models that simulate a cell's molecular behaviour, from complete cell cycles of minimal bacteria to learned predictions of how human cells respond to perturbation."summary: "Computational models that simulate a cell's molecular behaviour, from complete cell cycles of minimal bacteria to learned predictions of how human cells respond to perturbation."updated: "2026-08-22"updated: "2026-10-27"humanEvidence: "Complete whole-cell models exist only for minimal bacteria; learned models of human cell lines predict part of the transcriptional response to gene silencing, and none is validated as a model of a whole human cell."humanEvidence: "Complete whole-cell models exist only for minimal bacteria; learned models of human cell lines predict part of the transcriptional response to gene silencing, and none is validated as a model of a whole human cell."access: "Most models and datasets are public, including Recon3D, Geneformer, scGPT and Arc's State; commercial world models are proprietary and released as previews to selected academic and industry users."access: "Most models and datasets are public, including Recon3D, Geneformer, scGPT and Arc's State; commercial world models are proprietary and released as previews to selected academic and industry users."reversibility: "context"reversibility: "context"lines 49–54 → 49–5834 unchanged lines not shown
them, assembled from the published literature on an organism with about 525 genes. A human cell hasthem, assembled from the published literature on an organism with about 525 genes. A human cell hasroughly twenty thousand.roughly twenty thousand. ```figure{"key": "mycoplasma-3d-cell-model", "caption": "A spatial model of the same bacterium the 2012 dynamical model simulated: a molecular census placed in space at one instant. No model of any cell yet combines this structure with dynamics."}``` Learned models take the opposite route and assume nothing about mechanism. Geneformer converts eachLearned models take the opposite route and assume nothing about mechanism. Geneformer converts eachcell's expression profile into a sequence of gene tokens ranked by abundance and trains acell's expression profile into a sequence of gene tokens ranked by abundance and trains atransformer over roughly 30 million single-cell transcriptomes;[^theodoris2023] scGPT and itstransformer over roughly 30 million single-cell transcriptomes;[^theodoris2023] scGPT and itslines 75–81 → 79–8620 unchanged lines not shown
{ "year": "2023–2024", "title": "Single-cell foundation models", "text": "Geneformer and scGPT are trained on tens of millions of transcriptomes and released, moving the field's centre of gravity from equations to learned representations." }, { "year": "2023–2024", "title": "Single-cell foundation models", "text": "Geneformer and scGPT are trained on tens of millions of transcriptomes and released, moving the field's centre of gravity from equations to learned representations." }, { "year": "2024", "title": "The AI virtual cell proposed", "text": "A 42-author perspective in Cell sets out what an AI Virtual Cell would have to be: universal representations across scales, queried through virtual instruments, judged against experiment." }, { "year": "2024", "title": "The AI virtual cell proposed", "text": "A 42-author perspective in Cell sets out what an AI Virtual Cell would have to be: universal representations across scales, queried through virtual instruments, judged against experiment." }, { "year": "2025", "title": "Benchmarks and data at scale", "text": "The Arc Institute releases State and launches the Virtual Cell Challenge; the Chan Zuckerberg Initiative starts the Billion Cells Project with 10x Genomics and Ultima Genomics." }, { "year": "2025", "title": "Benchmarks and data at scale", "text": "The Arc Institute releases State and launches the Virtual Cell Challenge; the Chan Zuckerberg Initiative starts the Billion Cells Project with 10x Genomics and Ultima Genomics." }, { "year": "2026", "title": "A bacterial cell cycle in four dimensions", "text": "A spatial and kinetic model simulates the roughly 100-minute cycle of JCVI-syn3A end to end, including growth and division." } { "year": "2026", "title": "A bacterial cell cycle in four dimensions", "text": "A spatial and kinetic model simulates the roughly 100-minute cycle of JCVI-syn3A end to end, including growth and division." }, { "year": "2026", "title": "The Virtual Biology Initiative", "text": "Biohub, the Chan Zuckerberg Initiative's science organisation, commits $500 million in April to open data generation for predictive cell models; in October it, the US Department of Energy and the National Institutes of Health announce an expansion the partners value at $1.8 billion." }]]`````` lines 88–103 → 93–1226 unchanged lines not shown
existing demonstration that the approach works, and it took thirty years and an organism with fewerexisting demonstration that the approach works, and it took thirty years and an organism with fewerthan five hundred genes.than five hundred genes. The learned side is where the money and the human cells are. Arc Institute's State was trained onThe learned side is where the money and the human cells are. Arc Institute's [State](https://aiwiki.ai/wiki/arc_state_model) was trained onobservational data from 167 million cells and perturbational data from over 100 million more,observational data from 167 million cells and perturbational data from over 100 million more,spanning seventy human cell contexts.[^arc2025] Reporting in *Nature* put its performance at roughly a thirdspanning seventy human cell contexts.[^arc2025] Reporting in *Nature* put its performance at roughly a thirdof the genes most strongly affected by a given perturbation in a test set, against 7% forof the genes most strongly affected by a given perturbation in a test set, against 7% forconventional methods.[^eisenstein2026] Xaira Therapeutics has released a perturbation atlas of 25.6conventional methods.[^eisenstein2026] Xaira Therapeutics has released a perturbation atlas of 25.6million cells and a model trained on it; Emma Lundberg's SubCell learns from microscope imagesmillion cells and a model trained on it; Emma Lundberg's SubCell learns from microscope imagesrather than transcriptomes, which is one answer to the objection that expression is not the cell.rather than transcriptomes, which is one answer to the objection that expression is not the cell.The Chan Zuckerberg Initiative's Billion Cells Project, launched in 2025 with 10x Genomics andThe [Chan Zuckerberg Initiative](https://aiwiki.ai/wiki/chan_zuckerberg_initiative)'s Billion Cells Project, launched in 2025 with 10x Genomics andUltima Genomics, exists to supply training data on a scale nobody has had.Ultima Genomics, exists to supply training data on a scale nobody has had. The largest bet on that premise came in 2026. [[biohub]], the Chan Zuckerberg Initiative's scienceorganisation, launched the [Virtual Biology Initiative](https://aiwiki.ai/wiki/virtual_biology_initiative) in April with $500 million over fiveyears, $400 million of it for data generation and new measurement technology and $100 million foroutside groups, and said the data it generates would be open and freely available.[^biohub2026a] InOctober Biohub, the US Department of Energy and the National Institutes of Health announced anexpansion that the partners valued at $1.8 billion in funding, data, computation and measurementtechnology, and described as the largest coordinated commitment to AI-ready biological data todate. The figure combines Biohub's original commitment, more than $500 million of planned DOEspending over five years, routed through the federal Genesis Mission for AI in science, NIH datasetsbuilt through more than $500 million of earlier federal investment, which add data rather than newmoney, and $300 million from Google DeepMind, [[isomorphic-labs|Isomorphic Labs]] and Meta, with NVIDIA supplyingcomputing infrastructure.[^biohub2026b] It is a commitment to measure cells, not a model: neitherannouncement names a system to be built or a test it would have to pass. Commercial systems claim more. In August 2026 GenBio AI, founded in 2024 with headquarters in PaloCommercial systems claim more. In August 2026 GenBio AI, founded in 2024 with headquarters in PaloAlto and laboratories in Paris and Abu Dhabi, previewed AIDO Cell, which the company describes as aAlto and laboratories in Paris and Abu Dhabi, previewed AIDO Cell, which the company describes as aworld model of the human cell spanning DNA, RNA, protein, regulatory networks and whole-cellworld model of the human cell spanning DNA, RNA, protein, regulatory networks and whole-celllines 196–208 → 215–24692 unchanged lines not shown
applies weakly here, since a model of how a cell responds to a perturbation is not a design tool forapplies weakly here, since a model of how a cell responds to a perturbation is not a design tool forpathogens, though the distinction narrows as the models grow.pathogens, though the distinction narrows as the models grow. The 2026 funding brings these questions to the data rather than the models. Its stated rationale isthe proponents' side of the argument set out below, that data is the binding constraint, put byBiohub's head of science as a need for "orders of magnitude more data than exists today".[^biohub2026a]That is a premise, not a finding. Neither announcement names a licence beyond calling the result anopen resource, gives a date for releasing data, describes how consent and access will be governedfor the human donors whose cells and records are pooled, or mentions misuse.[^biohub2026a][^biohub2026b]What the corporate funders, one of them Meta, receive was explained to Reuters rather than in theannouncements: data paid for by commercial funders is held for an embargo period in which they canwork on it before it becomes a public resource, while the government-funded work carries no suchrestriction. Biohub's head of science also told Reuters he expected a first dataset in about a yearand accurate predictive models within five years.[^reuters2026]The gaps are the questions that attach to any pooled biomedical resource: who controls access, whethera donor can withdraw, whether data held in federal repositories stays available as policy changes,and whether models that predict how a person's cells respond could inform[[genetic-discrimination|decisions by insurers or employers]]. The embargo also sharpens theconcentration described above: the companies that pay for the largest datasets are the first totrain on them. ## Outlook## Outlook The test the field has set itself is clear enough to fail. A virtual cell would predict aThe test the field has set itself is clear enough to fail. A virtual cell would predict aperturbation's effect in a cell type it was not trained on, and the prediction would hold whenperturbation's effect in a cell type it was not trained on, and the prediction would hold whensomeone ran the experiment. Nothing published meets that standard, and the honest reading of thesomeone ran the experiment. Nothing published meets that standard, and the honest reading of the2025 benchmarks is that scale alone has not yet produced it. Proponents argue that data is the2025 benchmarks is that scale alone has not yet produced it. Proponents argue that data is thebinding constraint and that billion-cell corpora will change the picture; sceptics point out thatbinding constraint and that billion-cell corpora will change the picture, the premise the 2026 Virtual BiologyInitiative commitments rest on; sceptics point out thatthe same argument was made at each previous order of magnitude, and that a model trained only onthe same argument was made at each previous order of magnitude, and that a model trained only ontranscriptomes may be missing not data but variables.transcriptomes may be missing not data but variables. lines 240–244 → 278–28531 unchanged lines not shown
[^roohani2025]: `paper` Roohani, Y. H. et al. "Virtual Cell Challenge: Toward a Turing test for the virtual cell." *Cell*, 2025. {Describes the benchmark and its dataset of around 300 CRISPR-interference perturbations in H1 human embryonic stem cells.}[^roohani2025]: `paper` Roohani, Y. H. et al. "Virtual Cell Challenge: Toward a Turing test for the virtual cell." *Cell*, 2025. {Describes the benchmark and its dataset of around 300 CRISPR-interference perturbations in H1 human embryonic stem cells.}[^eisenstein2026]: `news` Eisenstein, M. "Can biology move into the Matrix?" *Nature*, vol. 654, 4 June 2026, pp. 286–288. {A reported feature; the State and challenge figures in it are attributed to the researchers involved rather than independently verified.}[^eisenstein2026]: `news` Eisenstein, M. "Can biology move into the Matrix?" *Nature*, vol. 654, 4 June 2026, pp. 286–288. {A reported feature; the State and challenge figures in it are attributed to the researchers involved rather than independently verified.}[^genbio2026]: `statement` GenBio AI. "AIDO Cell: A General-Purpose Simulator for Cell Biology." Preview announcement, 18 August 2026. {A company describing its own unreleased system; no peer-reviewed description or prospective experimental validation accompanied it.}[^genbio2026]: `statement` GenBio AI. "AIDO Cell: A General-Purpose Simulator for Cell Biology." Preview announcement, 18 August 2026. {A company describing its own unreleased system; no peer-reviewed description or prospective experimental validation accompanied it.}[^biohub2026a]: `statement` Biohub. "Biohub Launches the Virtual Biology Initiative to Galvanize a Global Effort to Create the Open Data Foundation for AI-Accelerated Biology." Press release, 29 April 2026. {The funder announcing its own programme; it promises open data but names no licence, release date or validation target, and does not address donor consent or misuse.}[^biohub2026b]: `statement` Biohub. "International, cross-sector collaboration commits nearly $2 billion to build foundational data for AI models to predict and treat disease." Press release, 7 October 2026. {The partners' own valuation of a total that combines new money, planned agency spending, earlier federal investment and in-kind data and computing; it does not say what the corporate funders receive.}[^reuters2026]: `news` Hu, K. "US government, Google join Zuckerberg-backed Biohub in $1.8 billion push for AI biology data." Reuters, 7 October 2026. {Reports the embargo arrangement and the timelines as Biohub's head of science described them; the length of the embargo is not given.}[^arc2025]: `statement` Arc Institute. "Arc Institute's first virtual cell model: State," June 2025; and "Virtual Cell Challenge 2025 Wrap-Up: Winners and Reflections," December 2025. {The institute's own accounts: the State training figures and the challenge prize results are reported by the party that built one and ran the other.}[^arc2025]: `statement` Arc Institute. "Arc Institute's first virtual cell model: State," June 2025; and "Virtual Cell Challenge 2025 Wrap-Up: Winners and Reflections," December 2025. {The institute's own accounts: the State training figures and the challenge prize results are reported by the party that built one and ran the other.} removed, struck through added, underlinedLine numbers count the serialised markdown of each revision, frontmatter included.
239 lines are the same in both revisions, and 190 of them are not shown. Neither revision here is necessarily the current article — read Virtual cell models for the text this wiki stands behind today.