{"id":116505,"date":"2025-08-01T17:45:00","date_gmt":"2025-08-01T21:45:00","guid":{"rendered":"https:\/\/www.freethink.com\/?post_type=ftm_article&p=116505"},"modified":"2025-08-04T09:06:51","modified_gmt":"2025-08-04T13:06:51","slug":"virtual-cells","status":"publish","type":"ftm_article","link":"https:\/\/www.freethink.com\/artificial-intelligence\/virtual-cells","title":{"rendered":"AI’s next frontier: Modeling life itself"},"content":{"rendered":"\n

This article is an installment of Future Explored<\/a>, a monthly guide to world-changing technology. You can get stories like this one straight to your inbox every month by subscribing above.<\/em><\/p>\n\n\n\n

It\u2019s 2040. Drug discovery is booming, thanks to virtual cells. These AI-powered models of living cells have become indispensable tools in biomedical research, helping scientists test treatments in silico before they ever reach a lab \u2014 saving time, money, and lives.<\/p>\n\n\n\n

Virtual Cells<\/strong><\/h2>\n\n\n\n

Artificial intelligence is quickly becoming biology\u2019s most powerful microscope.<\/p>\n\n\n\n

Top research centers are using the tech to develop virtual cells, AI-based simulations of the core building blocks of all lifeforms \u2014 and it\u2019s hard to overstate the impact their models could have on the world of health.<\/p>\n\n\n\n

\u201cThe vision that we can really understand everything about a cell \u2014 from its molecular structure to its function to how cells interact and operate in living organisms to how they respond and react to any intervention \u2014 will go a long way to helping us cure, prevent, and manage disease,\u201d said Patricia Brennan, VP of Science Technology and General Manager for Science at the Chan Zuckerberg Initiative (CZI).<\/p>\n\n\n\n

To find out how we got here \u2014 and where we\u2019re going \u2014 this month\u2019s Future Explored is taking a close look at virtual cells: what they are, who\u2019s making them, and how they could shape the future of medicine.<\/p>\n\n\n\n

Where we\u2019ve been<\/h3>\n\n\n\n
\"1665<\/figure>\n\n\n\n

Where we\u2019re going (maybe)<\/h3>\n\n\n\n

The first virtual cell was built over a decade ago \u2014 so why is interest in them surging now? The answer lies in fundamental differences in how today\u2019s models work compared to ones like what the Stanford team created in 2012.<\/p>\n\n\n\n

Their virtual cell simulated all of the molecular processes of Mycoplasma genitalium<\/a> (Mgen) by breaking them into 28 modules, each governed by its own set of mathematical equations and biological rules \u2014 a simplified example might be, \u201cIf Gene A is expressed, Gene B will be activated.\u201d<\/p>\n\n\n\n

The team developed those instructions by painstakingly digging through more than 900 papers, books, and databases, essentially distilling all of the scientific community\u2019s knowledge of how Mgen functions into 28 algorithms. Once complete, their model could simulate the entire Mgen lifecycle, from birth to division, in about 10 hours \u2014 roughly the same amount of time an actual Mgen cell takes to divide. <\/p>\n\n\n\n

Simply being able to observe this process was valuable, but the real utility of the virtual cell was that researchers could experiment on it. They could knock out a gene with keystrokes instead of CRISPR and then run the simulation to predict how its loss would affect Mgen. If they saw something interesting, they could then spend time on a lab experiment. <\/p>\n\n\n\n

\u201cIf you use a model to guide your experiments, you\u2019re going to discover things faster,\u201d said study leader Markus Covert<\/a> in 2012. \u201cWe\u2019ve shown that time and time again.\u201d<\/p>\n\n\n\n

Most of the time, the virtual cell\u2019s predictions would match the results of real-world experiments. When they didn\u2019t, the discrepancies usually involved genes that were poorly understood in scientific literature. That made sense: The model was limited by its programming. If scientists wanted to improve it, they\u2019d need to update their algorithms.<\/p>\n\n\n\n

\n

\u201cIf you use a model to guide your experiments, you\u2019re going to discover things faster.”<\/p>\nMarkus Covert<\/cite><\/blockquote>\n\n\n\n

In the years following the Stanford breakthrough, other groups made their own virtual cells using the same method: distill the known literature into rule-based algorithms. <\/p>\n\n\n\n

Today\u2019s virtual cells, however, are built on artificial intelligence, usually a specific kind of model called a transformer. Google researchers<\/a> first proposed this AI architecture in 2017, and it\u2019s the basis for many of today\u2019s best generative AIs, including ChatGPT.<\/p>\n\n\n\n

Transformer-based AIs learn to spot relationships between tokens (small units of data) by training on huge datasets. Once trained, they can then generate new content by predicting the most likely next token in a sequence. <\/p>\n\n\n\n

For ChatGPT, for example, tokens are words or parts of words. The huge training dataset was the internet, and once trained, ChatGPT was able to generate text by predicting the most likely token to come next in its response, over and over again, based on the ones that came before it.<\/p>\n\n\n\n

One of the most remarkable things about transformer-based AIs is that they can output content that isn\u2019t included in their training data. An AI image generator, for example, can output a photorealistic picture of a cat made of spaghetti even if it wasn\u2019t explicitly shown what that should look like.<\/p>\n\n\n\n

“The goal is for virtual cell models to serve as digital twins or computational stand-ins for experimental systems.”<\/p>Marinka Zitnik<\/cite><\/blockquote><\/figure>\n\n\n\n

Researchers are now building virtual cells on the transformer architecture \u2014 and the results are remarkable.<\/p>\n\n\n\n

CZI\u2019s TranscriptFormer model<\/a>, for example, was trained on datasets containing images, RNA sequences, and other biological data from 112 million cells. These were drawn from 12 different species across 1.5 billion years of evolution. A researcher can now prompt it with data from a cell they\u2019re studying, and the AI can predict its cell type, infection status, and more \u2014 even if the cell comes from a species that wasn\u2019t included in the model\u2019s training data.<\/p>\n\n\n\n

\u201cWe just trained it on natural variability, but this natural variability follows the tree of life: It has a lot of structure. There’s a lot of interesting stuff happening there,\u201d said Theofanis Karaletsos, Senior Director of AI at CZI. \u201cThe model actually becomes extremely rich and extremely performant on doing all kinds of tasks.\u201d<\/p>\n\n\n\n

CZI is now offering researchers early access to a one-stop platform<\/a> that includes several virtual cell models, including TranscriptFormer, and the datasets used to train them. These models have specific use cases \u2014 CZI\u2019s GREmLN model<\/a>, for example, predicts how genes work together \u2014 but the vision is to develop models that can simulate more complex cellular behavior.<\/p>\n\n\n\n

\u201cLooking forward, the goal is for virtual cell models to serve as digital twins or computational stand-ins for experimental systems,\u201d said Marinka Zitnik, a CZI collaborator and assistant professor at Harvard. \u201cFor example, a validated virtual cell could simulate the outcome of a drug or genetic intervention in silico, potentially reducing the need for animal experiments or guiding the design of laboratory studies.\u201d<\/p>\n\n\n\n

\n

“You ask the model, \u2018What perturbations do I need to make to move this cell from this diseased state to this healthy state?\u2019\u201d<\/p>\nDavid Burke<\/em><\/cite><\/blockquote>\n\n\n\n

Biomedical research nonprofit Arc Institute is building virtual cells, too. It recently opened access to its first model, STATE<\/a>, which was trained on observational data from nearly 170 million cells and perturbational data from over 100 million cells. Perturbational data captures how a cell reacts when its normal function is disrupted by a drug, genetic edit, or some other external stimulus.<\/p>\n\n\n\n

Researchers input a cell\u2019s transcriptome \u2014 the full collection of genes being actively expressed at a given moment \u2014 and a proposed perturbation, and STATE predicts how the cell\u2019s gene expression patterns are likely to change. This can give scientists a way to test the potential impact of disease treatments without having to actually perform experiments. <\/p>\n\n\n\n

By running this process in reverse, STATE can even point researchers toward promising interventions they hadn\u2019t considered.<\/p>\n\n\n\n

\u201cYou take a cell that’s in a diseased state \u2014 like maybe it’s got an Alzheimer’s disease transcriptomic profile \u2014 and one in a healthy state, and then you ask the model, \u2018What perturbations do I need to make to move this cell from this diseased state to this healthy state?\u2019\u201d said David Burke, Arc Institute\u2019s CTO. <\/p>\n\n\n\n

STATE\u2019s predictions range in accuracy from 40% to 60%, depending on the type of perturbation, according to Burke. He thinks 75% would be good enough that biologists could start relying on the AI\u2019s predictions without having to run experiments in the wet lab.<\/p>\n\n\n\n

\u201cThat might seem a bit low,\u201d said Burke, \u201cbut when you look at all the different datasets from different labs, the concordance between them is only about 75% because single-cell sequencing and perturbation screens are very noisy, so that’s our target.\u201d<\/p>\n\n\n\n

“We will need a lot more data.”<\/p>Patricia Brennan<\/cite><\/blockquote><\/figure>\n\n\n\n

A transformer model is only as good as the quality and quantity of its training data, so if we want to improve today\u2019s virtual cells, we\u2019re going to need to improve our datasets.<\/p>\n\n\n\n

\u201cWhile the scale of the data sets has been growing over the last number of years, we will need a lot more data,\u201d said CZI\u2019s Brennan, who noted that much of the data we already have wasn\u2019t necessarily collected with the training of virtual cells in mind, which complicates its use as training material. <\/p>\n\n\n\n

To help close the data gap as quickly as possible, CZI launched the open-source Billion Cells Project (BCP)<\/a> in February. The goal of the initiative is to quickly and cost-effectively generate a public dataset containing one billion cells through partnerships with scientists and the developers of cutting-edge cell analysis technologies.<\/p>\n\n\n\n

\u201cTraditional data generation pipelines can take three to four years, whereas BCP is compressing that timeline to months,\u201d said Bailey Marshall, Senior Program Associate, Single-Cell Biology at CZI.<\/p>\n\n\n\n

\u201cOne of the project\u2019s most important innovations is in interoperability,\u201d she added. \u201cBy aligning from the outset on standardized kits, protocols, and technologies across the BCP\u2026data from diverse tissues, species, and modalities can be easily integrated. This allows researchers and AI developers to train models that are consistent, reproducible, and broadly applicable.\u201d<\/p>\n\n\n\n

\n

\u201cHow far can we go? That’s an open scientific question.”<\/p>\nDavid Burke<\/cite><\/blockquote>\n\n\n\n

So, in the short term, the scientific community knows it needs more data to make its virtual cells more robust, but the extent to which it can improve them is still unknown.<\/p>\n\n\n\n

With enough high-quality data, is it possible to make a single model that accurately predicts everything that will happen in a cell when it\u2019s perturbed in every way possible? How about a model that takes into account how cells work in context with one another? Can we create virtual models of entire tissues, organs, or even bodies? <\/p>\n\n\n\n

If so, it would mark a turning point in biology \u2014 moving from a science that observes life to one that can simulate and predict it. The consequences for medicine, longevity, and our understanding of health would be transformative.<\/p>\n\n\n\n

\u201cHow far can we go? That’s an open scientific question,\u201d said Burke.<\/p>\n\n\n\n

We\u2019d love to hear from you! If you have a comment about this article or if you have a tip for a future Freethink story, please email us at tips@freethink.com<\/a>.<\/em><\/p>\n","protected":false},"excerpt":{"rendered":"

Biologists are skipping the petri dish and using AI-powered virtual cells to experiment in silico.<\/p>\n","protected":false},"author":25,"featured_media":116506,"template":"","ftm_taxonomy_fields":[46,57],"ftm_taxonomy_challenges":[],"ftm_taxonomy_statuses":[36],"ftm_taxonomy_hidden_tags":[1939],"class_list":["post-116505","ftm_article","type-ftm_article","status-publish","has-post-thumbnail","hentry","ftm_taxonomy_fields-ai","ftm_taxonomy_fields-biology","ftm_taxonomy_statuses-featured"],"acf":[],"apple_news_notices":[],"yoast_head":"\nAI's next frontier: Modeling life itself<\/title>\n<meta name=\"description\" content=\"Biologists are skipping the petri dish and using AI-powered virtual cells to experiment in silico.\" \/>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/www.freethink.com\/artificial-intelligence\/virtual-cells\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"AI's next frontier: Modeling life itself\" \/>\n<meta property=\"og:description\" content=\"Biologists are skipping the petri dish and using AI-powered virtual cells to experiment in silico.\" \/>\n<meta property=\"og:url\" content=\"https:\/\/www.freethink.com\/artificial-intelligence\/virtual-cells\" \/>\n<meta property=\"og:site_name\" content=\"Freethink\" \/>\n<meta property=\"article:modified_time\" content=\"2025-08-04T13:06:51+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/www.freethink.com\/wp-content\/uploads\/2025\/08\/virtual-cells-thumb.jpg?resize=1200,630\" \/>\n\t<meta property=\"og:image:width\" content=\"1200\" \/>\n\t<meta property=\"og:image:height\" content=\"630\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/jpeg\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:description\" content=\"Biologists are skipping the petri dish and using AI-powered virtual cells to experiment in silico.\" \/>\n<meta name=\"twitter:label1\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data1\" content=\"8 minutes\" \/>\n<!-- \/ Yoast SEO Premium plugin. -->","yoast_head_json":{"title":"AI's next frontier: Modeling life itself","description":"Biologists are skipping the petri dish and using AI-powered virtual cells to experiment in silico.","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/www.freethink.com\/artificial-intelligence\/virtual-cells","og_locale":"en_US","og_type":"article","og_title":"AI's next frontier: Modeling life itself","og_description":"Biologists are skipping the petri dish and using AI-powered virtual cells to experiment in silico.","og_url":"https:\/\/www.freethink.com\/artificial-intelligence\/virtual-cells","og_site_name":"Freethink","article_modified_time":"2025-08-04T13:06:51+00:00","og_image":[{"width":1200,"height":630,"url":"https:\/\/www.freethink.com\/wp-content\/uploads\/2025\/08\/virtual-cells-thumb.jpg?resize=1200,630","type":"image\/jpeg"}],"twitter_card":"summary_large_image","twitter_description":"Biologists are skipping the petri dish and using AI-powered virtual cells to experiment in silico.","twitter_misc":{"Est. reading time":"8 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/www.freethink.com\/artificial-intelligence\/virtual-cells#article","isPartOf":{"@id":"https:\/\/www.freethink.com\/artificial-intelligence\/virtual-cells"},"author":{"name":"kristinhouser","@id":"https:\/\/www.freethink.com\/#\/schema\/person\/e45bf79276f6c14454ee4e1dfa7aca8c"},"headline":"AI’s next frontier: Modeling life itself","datePublished":"2025-08-01T21:45:00+00:00","dateModified":"2025-08-04T13:06:51+00:00","mainEntityOfPage":{"@id":"https:\/\/www.freethink.com\/artificial-intelligence\/virtual-cells"},"wordCount":1775,"publisher":{"@id":"https:\/\/www.freethink.com\/#organization"},"image":{"@id":"https:\/\/www.freethink.com\/artificial-intelligence\/virtual-cells#primaryimage"},"thumbnailUrl":"https:\/\/www.freethink.com\/wp-content\/uploads\/2025\/08\/virtual-cells-thumb.jpg?quality=75","inLanguage":"en-US"},{"@type":"WebPage","@id":"https:\/\/www.freethink.com\/artificial-intelligence\/virtual-cells","url":"https:\/\/www.freethink.com\/artificial-intelligence\/virtual-cells","name":"AI's next frontier: Modeling life itself","isPartOf":{"@id":"https:\/\/www.freethink.com\/#website"},"primaryImageOfPage":{"@id":"https:\/\/www.freethink.com\/artificial-intelligence\/virtual-cells#primaryimage"},"image":{"@id":"https:\/\/www.freethink.com\/artificial-intelligence\/virtual-cells#primaryimage"},"thumbnailUrl":"https:\/\/www.freethink.com\/wp-content\/uploads\/2025\/08\/virtual-cells-thumb.jpg?quality=75","datePublished":"2025-08-01T21:45:00+00:00","dateModified":"2025-08-04T13:06:51+00:00","description":"Biologists are skipping the petri dish and using AI-powered virtual cells to experiment in silico.","breadcrumb":{"@id":"https:\/\/www.freethink.com\/artificial-intelligence\/virtual-cells#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/www.freethink.com\/artificial-intelligence\/virtual-cells"]}]},{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/www.freethink.com\/artificial-intelligence\/virtual-cells#primaryimage","url":"https:\/\/www.freethink.com\/wp-content\/uploads\/2025\/08\/virtual-cells-thumb.jpg?quality=75","contentUrl":"https:\/\/www.freethink.com\/wp-content\/uploads\/2025\/08\/virtual-cells-thumb.jpg?quality=75","width":1200,"height":675,"caption":"Chan Zuckerberg Initiative"},{"@type":"BreadcrumbList","@id":"https:\/\/www.freethink.com\/artificial-intelligence\/virtual-cells#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Articles","item":"https:\/\/www.freethink.com\/articles"},{"@type":"ListItem","position":2,"name":"AI’s next frontier: Modeling life itself"}]},{"@type":"WebSite","@id":"https:\/\/www.freethink.com\/#website","url":"https:\/\/www.freethink.com\/","name":"Freethink","description":"Move the world","publisher":{"@id":"https:\/\/www.freethink.com\/#organization"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/www.freethink.com\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"en-US"},{"@type":"Organization","@id":"https:\/\/www.freethink.com\/#organization","name":"Freethink Media","url":"https:\/\/www.freethink.com\/","logo":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/www.freethink.com\/#\/schema\/logo\/image\/","url":"https:\/\/www.freethink.com\/wp-content\/uploads\/2021\/06\/logo.svg","contentUrl":"https:\/\/www.freethink.com\/wp-content\/uploads\/2021\/06\/logo.svg","width":651,"height":124,"caption":"Freethink Media"},"image":{"@id":"https:\/\/www.freethink.com\/#\/schema\/logo\/image\/"}},{"@type":"Person","@id":"https:\/\/www.freethink.com\/#\/schema\/person\/e45bf79276f6c14454ee4e1dfa7aca8c","name":"kristinhouser","image":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/www.freethink.com\/#\/schema\/person\/image\/","url":"https:\/\/secure.gravatar.com\/avatar\/ff88759e0ed195de655c7703310050f17b921ae4fc276d7eb5930cddafa694f9?s=96&d=mm&r=g","contentUrl":"https:\/\/secure.gravatar.com\/avatar\/ff88759e0ed195de655c7703310050f17b921ae4fc276d7eb5930cddafa694f9?s=96&d=mm&r=g","caption":"kristinhouser"},"url":"https:\/\/www.freethink.com\/author\/kristinhouser"}]}},"jetpack_sharing_enabled":true,"amp_enabled":true,"_links":{"self":[{"href":"https:\/\/www.freethink.com\/wp-json\/wp\/v2\/ftm_article\/116505","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.freethink.com\/wp-json\/wp\/v2\/ftm_article"}],"about":[{"href":"https:\/\/www.freethink.com\/wp-json\/wp\/v2\/types\/ftm_article"}],"author":[{"embeddable":true,"href":"https:\/\/www.freethink.com\/wp-json\/wp\/v2\/users\/25"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.freethink.com\/wp-json\/wp\/v2\/media\/116506"}],"wp:attachment":[{"href":"https:\/\/www.freethink.com\/wp-json\/wp\/v2\/media?parent=116505"}],"wp:term":[{"taxonomy":"ftm_taxonomy_fields","embeddable":true,"href":"https:\/\/www.freethink.com\/wp-json\/wp\/v2\/ftm_taxonomy_fields?post=116505"},{"taxonomy":"ftm_taxonomy_challenges","embeddable":true,"href":"https:\/\/www.freethink.com\/wp-json\/wp\/v2\/ftm_taxonomy_challenges?post=116505"},{"taxonomy":"ftm_taxonomy_statuses","embeddable":true,"href":"https:\/\/www.freethink.com\/wp-json\/wp\/v2\/ftm_taxonomy_statuses?post=116505"},{"taxonomy":"ftm_taxonomy_hidden_tags","embeddable":true,"href":"https:\/\/www.freethink.com\/wp-json\/wp\/v2\/ftm_taxonomy_hidden_tags?post=116505"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}