AI Model Predicts Gene Activity in Human Cells, Transforming Biological Research

Scientists trained AI to predict gene activity, a potentially powerful tool

Researchers hope the AI tool will aid in the development of cell-specific gene therapies to treat diseases such as cancer. Scientists led by a team at Columbia University have trained a model to predict how the genes inside a cell will drive its behavior, which could be a powerful tool with the potential to broaden our understanding of cancer and genetic diseases, and even to pave the way for cell-specific gene therapies to treat them. The researchers trained the new artificial intelligence tool, an algorithm dubbed General Expression Transformer, or GET, using an approach similar to that used by creators of the language program ChatGPT. While ChatGPT learned the grammar of language, GET has learned underlying rules governing genes: how they are turned on or off like a light switch, or dialed up or down like a volume control. This complex process, known as gene expression, determines which proteins we make and whether we make them in the correct amounts, crucial work given that proteins play a role in virtually every action in the body -- fighting disease, moving, breathing, even thinking. Although GET is at a much earlier stage of development, it could play a similar role to AlphaFold2, the AI system that predicts the three-dimensional structure of proteins. The transformative technology was recognized with the 2024 Nobel Prize in chemistry and has now been updated to AlphaFold3. The regulation of genes and the structure of proteins both are fundamental to life, and problems in either can trigger disease. "Biology is being transformed into something that is a predictive science," said Raul Rabadan, one of the authors of a paper reporting the work Wednesday in the journal Nature and director of the Program for Mathematical Genomics at Columbia. "We're seeing a revolution in biology." Mark Gerstein, a professor of biomedical informatics at Yale School of Medicine, who was not involved in the new study, said that for 15 to 20 years experts have been systematically trying to make predictions about gene regulation, building on a trove of carefully made datasets. The data examined all genes in specific types of human cells -- for example, retinal cells or neurons -- measuring, among other things, gene expression and the binding of key proteins called transcription factors. "This is a field poised to have this type of advancement by AI," Gerstein said. While other scientific groups have trained models using abnormal cells such as those found in different cancers, Xi Fu, a graduate student in Rabadan's lab, decided to train GET using information from cells in normal human tissue. The training used data from more than 1.3 million cells, spanning 213 different types found in the human body. Rabadan's team found that they could omit one cell type from the data -- for example, astrocytes, which are found in the central nervous system -- and the model could make accurate predictions about astrocytes based on what it had learned from all of the other cells. Mike Pazin, a program director at the National Human Genome Research Institute, part of the National Institutes of Health, said learning about one cell type and then making predictions about another is an especially daunting challenge. "In some ways, it's like if I handed somebody a bunch of books in English, and said: 'Okay, now this is in Russian. What does it say?' And they say: 'Ah. I understand grammar, syntax and words. I'm going to make predictions about this even though it is written in a different language.' I would be like, 'Wow, is that possible?'" The work described in the Nature paper "directly tackles one of biology's major challenges: understanding how the same genome can drive such diverse behaviors in different cell types," said Jian Ma, professor of computational biology and director of the Center for AI-Driven Biomedical Research in the School of Computer Science at Carnegie Mellon University. Ma, who was not involved in the new study, commented by email. All 30 trillion or so cells in the human body contain the same complete set of DNA. "Yet each cell type -- be it a neuron, a muscle cell, or a skin cell -- expresses a unique set of genes," Ma said. Humans have about 20,0000 genes, some of which may be turned on in a retinal cell, but off in a skin cell. While "much of this regulatory grammar remains poorly understood," Ma said, "The GET model takes an important step toward decoding this language." Understanding the language of gene regulation holds the potential for great benefit to human health, said Yang E. Li, assistant professor in the departments of neurosurgery and genetics at Washington University School of Medicine in St. Louis. "We want to learn the grammar and prioritize the key players in different cell types," Li said, "because many human diseases are caused by a disruption of that grammar." Scientists hope the model will help in other ways, such as in development of gene therapies to correct a mutation -- an error in the genetic code -- that harms a specific kind of cell. Such therapies have to be precisely designed so that they fix the cells harmed by a disease without disrupting other cell types. "We can design gene therapies that deliver a gene that is only expressed in one cell type, and not in another," Rabadan said. Being able to predict which genes are turned on, off, up or down in different cells could help determine the cell of origin for a disease. A model that makes accurate predictions about gene regulation also holds the promise of lessening one of the more grueling tasks in science: deciding which of a massive number of possible experiments are the ones most likely to answer the researcher's question. A cancer, for example, may contain more than 1,000 mutations in the genome that have developed after conception. The effects of most of these mutations are unknown, Rabadan said. That leaves scientists with the enormous task of determining where to start. "The number of potential genetic combinations is more than the number of atoms in the universe," Rabadan said. "What are the ones that are relevant?"

New AI method predicts gene activity in human cells

Columbia University Irving Medical CenterJan 8 2025 Using a new artificial intelligence method, researchers at Columbia University Vagelos College of Physicians and Surgeons can accurately predict the activity of genes within any human cell, essentially revealing the cell's inner mechanisms. The system, described in the current issue of Nature, could transform the way scientists work to understand everything from cancer to genetic diseases. Predictive generalizable computational models allow to uncover biological processes in a fast and accurate way. These methods can effectively conduct large-scale computational experiments, boosting and guiding traditional experimental approaches." Raul Rabadan, professor of systems biology and senior author of the new paper Traditional research methods in biology are good at revealing how cells perform their jobs or react to disturbances. But they cannot make predictions about how cells work or how cells will react to change, like a cancer-causing mutation. "Having the ability to accurately predict a cell's activities would transform our understanding of fundamental biological processes," Rabadan says. "It would turn biology from a science that describes seemingly random processes into one that can predict the underlying systems that govern cell behavior." In recent years, the accumulation of massive amounts of data from cells and more powerful AI models are starting to transform biology into a more predictive science. The 2024 Nobel Prize in Chemistry was awarded to researchers for their groundbreaking work in using AI to predict protein structures. But the use of AI methods to predict the activities of genes and proteins inside cells has proven more difficult. New AI method predicts gene expression in any cell In the new study, Rabadan and his colleagues tried to use AI to predict which genes are active within specific cells. Such information about gene expression can tell researchers the identity of the cell and how the cell performs its functions. "Previous models have been trained on data in particular cell types, usually cancer cell lines or something else that has little resemblance to normal cells," Rabadan says. Xi Fu, a graduate student in Rabadan's lab, decided to take a different approach, training a machine learning model on gene expression data from millions of cells obtained from normal human tissues. The inputs consisted of genome sequences and data showing which parts of the genome are accessible and expressed. The overall approach resembles the way ChatGPT and other popular "foundation" models work. These systems use a set of training data to identify underlying rules, the grammar of language, and then apply those inferred rules to new situations. "Here it's exactly the same thing: we learn the grammar in many different cellular states, and then we go into a particular condition-it can be a diseased or it can be a normal cell type-and we can try to see how well we predict patterns from this information," says Rabadan. Fu and Rabadan soon enlisted a team of collaborators, including co-first authors Alejandro Buendia, now a Stanford PhD student formerly in the Rabadan lab, and Shentong Mo of Carnegie Mellon, to train and test the new model. After training on data from more than 1.3 million human cells, the system became accurate enough to predict gene expression in cell types it had never seen, yielding results that agreed closely with experimental data. New AI methods reveal drivers of a pediatric cancer Next, the investigators showed the power of their AI system when they asked it to uncover still hidden biology of diseased cells, in this case, an inherited form of pediatric leukemia. "These kids inherit a gene that is mutated, and it was unclear exactly what it is these mutations are doing," says Rabadan, who also co-directs the cancer genomics and epigenomics research program at Columbia's Herbert Irving Comprehensive Cancer Center. With AI, the researchers predicted that the mutations disrupt the interaction between two different transcription factors that determine the fate of leukemic cells. Laboratory experiments confirmed AI's prediction. Understanding the effect of these mutations uncovers specific mechanisms that drive this disease. AI could reveal "dark matter" in genome The new computational methods should also allow researchers to start exploring the role of genome's "dark matter"-a term borrowed from cosmology that refers to the vast majority of the genome, which does not encode known genes-in cancer and other diseases. "The vast majority of mutations found in cancer patients are in so-called dark regions of the genome. These mutations do not affect the function of a protein and have remained mostly unexplored. says Rabadan. "The idea is that using these models, we can look at mutations and illuminate that part of the genome." Already, Rabadan is working with researchers at Columbia and other universities, exploring different cancers, from brain to blood cancers, learning the grammar of regulation in normal cells, and how cells change in the process of cancer development. The work also opens new avenues for understanding many diseases beyond cancer and potentially identifying targets for new treatments. By presenting novel mutations to the computer model, researchers can now gain deep insights and predictions about exactly how those mutations affect a cell. Coming on the heels of other recent advances in artificial intelligence for biology, Rabadan sees the work as part of a major trend: "It's really a new era in biology that is extremely exciting; transforming biology into a predictive science." Columbia University Irving Medical Center Journal reference: Fu, X., et al. (2025). A foundation model of transcription across human cell types. Nature. doi.org/10.1038/s41586-024-08391-z.

Computational biologists develop AI that predicts inner workings of cells

Using a new artificial intelligence method, researchers at Columbia University Vagelos College of Physicians and Surgeons can accurately predict the activity of genes within any human cell, essentially revealing the cell's inner mechanisms. The system, described in Nature, could transform the way scientists work to understand everything from cancer to genetic diseases. "Predictive generalizable computational models allow to uncover biological processes in a fast and accurate way. These methods can effectively conduct large-scale computational experiments, boosting and guiding traditional experimental approaches," says Raul Rabadan, professor of systems biology and senior author of the new paper. Traditional research methods in biology are good at revealing how cells perform their jobs or react to disturbances. But they cannot make predictions about how cells work or how cells will react to change, like a cancer-causing mutation. "Having the ability to accurately predict a cell's activities would transform our understanding of fundamental biological processes," Rabadan says. "It would turn biology from a science that describes seemingly random processes into one that can predict the underlying systems that govern cell behavior." In recent years, the accumulation of massive amounts of data from cells and more powerful AI models are starting to transform biology into a more predictive science. The 2024 Nobel Prize in Chemistry was awarded to researchers for their groundbreaking work in using AI to predict protein structures. But the use of AI methods to predict the activities of genes and proteins inside cells has proven more difficult. New AI method predicts gene expression in any cell In the new study, Rabadan and his colleagues tried to use AI to predict which genes are active within specific cells. Such information about gene expression can tell researchers the identity of the cell and how the cell performs its functions. "Previous models have been trained on data in particular cell types, usually cancer cell lines or something else that has little resemblance to normal cells," Rabadan says. Xi Fu, a graduate student in Rabadan's lab, decided to take a different approach, training a machine learning model on gene expression data from millions of cells obtained from normal human tissues. The inputs consisted of genome sequences and data showing which parts of the genome are accessible and expressed. The overall approach resembles the way ChatGPT and other popular "foundation" models work. These systems use a set of training data to identify underlying rules, the grammar of language, and then apply those inferred rules to new situations. "Here it's exactly the same thing: we learn the grammar in many different cellular states, and then we go into a particular condition -- it can be a diseased [cell type] or it can be a normal cell type -- and we can try to see how well we predict patterns from this information," says Rabadan. Fu and Rabadan soon enlisted a team of collaborators, including co-first authors Alejandro Buendia, now a Stanford Ph.D. student formerly in the Rabadan lab, and Shentong Mo of Carnegie Mellon, to train and test the new model. After training on data from more than 1.3 million human cells, the system became accurate enough to predict gene expression in cell types it had never seen, yielding results that agreed closely with experimental data. New AI methods reveal drivers of a pediatric cancer Next, the investigators showed the power of their AI system when they asked it to uncover still-hidden biology of diseased cells, in this case, an inherited form of pediatric leukemia. "These kids inherit a gene that is mutated, and it was unclear exactly what it is these mutations are doing," says Rabadan, who also co-directs the cancer genomics and epigenomics research program at Columbia's Herbert Irving Comprehensive Cancer Center. With AI, the researchers predicted that the mutations disrupt the interaction between two different transcription factors that determine the fate of leukemic cells. Laboratory experiments confirmed AI's prediction. Understanding the effect of these mutations uncovers specific mechanisms that drive this disease. AI could reveal 'dark matter' in genome The new computational methods should also allow researchers to start exploring the role of genome's "dark matter" -- a term borrowed from cosmology that refers to the vast majority of the genome, which does not encode known genes -- in cancer and other diseases. "The vast majority of mutations found in cancer patients are in so-called dark regions of the genome. These mutations do not affect the function of a protein and have remained mostly unexplored," says Rabadan. "The idea is that using these models, we can look at mutations and illuminate that part of the genome." Rabadan is working with researchers at Columbia and other universities, exploring different cancers, from brain to blood cancers, learning the grammar of regulation in normal cells, and how cells change in the process of cancer development. The work also opens new avenues for understanding many diseases beyond cancer and potentially identifying targets for new treatments. By presenting novel mutations to the computer model, researchers can now gain deep insights and predictions about exactly how those mutations affect a cell. Coming on the heels of other recent advances in artificial intelligence for biology, Rabadan sees the work as part of a major trend: "It's really a new era in biology that is extremely exciting; transforming biology into a predictive science."

New AI predicts inner workings of cells

In the same way that ChatGPT understands human language, a new AI model developed by Columbia computational biologists captures the language of cells to accurately predict their activities. Using a new artificial intelligence method, researchers at Columbia University Vagelos College of Physicians and Surgeons can accurately predict the activity of genes within any human cell, essentially revealing the cell's inner mechanisms. The system, described in the current issue of Nature, could transform the way scientists work to understand everything from cancer to genetic diseases. "Predictive generalizable computational models allow to uncover biological processes in a fast and accurate way. These methods can effectively conduct large-scale computational experiments, boosting and guiding traditional experimental approaches," says Raul Rabadan, professor of systems biology and senior author of the new paper. Traditional research methods in biology are good at revealing how cells perform their jobs or react to disturbances. But they cannot make predictions about how cells work or how cells will react to change, like a cancer-causing mutation. "Having the ability to accurately predict a cell's activities would transform our understanding of fundamental biological processes," Rabadan says. "It would turn biology from a science that describes seemingly random processes into one that can predict the underlying systems that govern cell behavior." In recent years, the accumulation of massive amounts of data from cells and more powerful AI models are starting to transform biology into a more predictive science. The 2024 Nobel Prize in Chemistry was awarded to researchers for their groundbreaking work in using AI to predict protein structures. But the use of AI methods to predict the activities of genes and proteins inside cells has proven more difficult. New AI method predicts gene expression in any cell In the new study, Rabadan and his colleagues tried to use AI to predict which genes are active within specific cells. Such information about gene expression can tell researchers the identity of the cell and how the cell performs its functions. "Previous models have been trained on data in particular cell types, usually cancer cell lines or something else that has little resemblance to normal cells," Rabadan says. Xi Fu, a graduate student in Rabadan's lab, decided to take a different approach, training a machine learning model on gene expression data from millions of cells obtained from normal human tissues. The inputs consisted of genome sequences and data showing which parts of the genome are accessible and expressed. The overall approach resembles the way ChatGPT and other popular "foundation" models work. These systems use a set of training data to identify underlying rules, the grammar of language, and then apply those inferred rules to new situations. "Here it's exactly the same thing: we learn the grammar in many different cellular states, and then we go into a particular condition -- it can be a diseased or it can be a normal cell type -- and we can try to see how well we predict patterns from this information," says Rabadan. Fu and Rabadan soon enlisted a team of collaborators, including co-first authors Alejandro Buendia, now a Stanford PhD student formerly in the Rabadan lab, and Shentong Mo of Carnegie Mellon, to train and test the new model. After training on data from more than 1.3 million human cells, the system became accurate enough to predict gene expression in cell types it had never seen, yielding results that agreed closely with experimental data. New AI methods reveal drivers of a pediatric cancer Next, the investigators showed the power of their AI system when they asked it to uncover still hidden biology of diseased cells, in this case, an inherited form of pediatric leukemia. "These kids inherit a gene that is mutated, and it was unclear exactly what it is these mutations are doing," says Rabadan, who also co-directs the cancer genomics and epigenomics research program at Columbia's Herbert Irving Comprehensive Cancer Center. With AI, the researchers predicted that the mutations disrupt the interaction between two different transcription factors that determine the fate of leukemic cells. Laboratory experiments confirmed AI's prediction. Understanding the effect of these mutations uncovers specific mechanisms that drive this disease. AI could reveal "dark matter" in genome The new computational methods should also allow researchers to start exploring the role of genome's "dark matter" -- a term borrowed from cosmology that refers to the vast majority of the genome, which does not encode known genes -- in cancer and other diseases. "The vast majority of mutations found in cancer patients are in so-called dark regions of the genome. These mutations do not affect the function of a protein and have remained mostly unexplored. says Rabadan. "The idea is that using these models, we can look at mutations and illuminate that part of the genome." Already, Rabadan is working with researchers at Columbia and other universities, exploring different cancers, from brain to blood cancers, learning the grammar of regulation in normal cells, and how cells change in the process of cancer development. The work also opens new avenues for understanding many diseases beyond cancer and potentially identifying targets for new treatments. By presenting novel mutations to the computer model, researchers can now gain deep insights and predictions about exactly how those mutations affect a cell. Coming on the heels of other recent advances in artificial intelligence for biology, Rabadan sees the work as part of a major trend: "It's really a new era in biology that is extremely exciting; transforming biology into a predictive science."

AI Reveals Gene Activity in Human Cells - Neuroscience News

Summary: Researchers have developed an AI model that accurately predicts gene activity in any human cell, providing insights into cellular functions and disease mechanisms. Trained on data from over 1.3 million cells, the model can predict gene expression in unseen cell types with high accuracy. It has already uncovered mechanisms driving a pediatric leukemia and may help explore the genome's "dark matter," where most cancer mutations occur. Using a new artificial intelligence method, researchers at Columbia University Vagelos College of Physicians and Surgeons can accurately predict the activity of genes within any human cell, essentially revealing the cell's inner mechanisms. The system, described in the current issue of Nature, could transform the way scientists work to understand everything from cancer to genetic diseases. "Predictive generalizable computational models allow to uncover biological processes in a fast and accurate way. These methods can effectively conduct large-scale computational experiments, boosting and guiding traditional experimental approaches," says Raul Rabadan, professor of systems biology and senior author of the new paper. Traditional research methods in biology are good at revealing how cells perform their jobs or react to disturbances. But they cannot make predictions about how cells work or how cells will react to change, like a cancer-causing mutation. "Having the ability to accurately predict a cell's activities would transform our understanding of fundamental biological processes," Rabadan says. "It would turn biology from a science that describes seemingly random processes into one that can predict the underlying systems that govern cell behavior." In recent years, the accumulation of massive amounts of data from cells and more powerful AI models are starting to transform biology into a more predictive science. The 2024 Nobel Prize in Chemistry was awarded to researchers for their groundbreaking work in using AI to predict protein structures. But the use of AI methods to predict the activities of genes and proteins inside cells has proven more difficult. New AI method predicts gene expression in any cell In the new study, Rabadan and his colleagues tried to use AI to predict which genes are active within specific cells. Such information about gene expression can tell researchers the identity of the cell and how the cell performs its functions. "Previous models have been trained on data in particular cell types, usually cancer cell lines or something else that has little resemblance to normal cells," Rabadan says. Xi Fu, a graduate student in Rabadan's lab, decided to take a different approach, training a machine learning model on gene expression data from millions of cells obtained from normal human tissues. The inputs consisted of genome sequences and data showing which parts of the genome are accessible and expressed. The overall approach resembles the way ChatGPT and other popular "foundation" models work. These systems use a set of training data to identify underlying rules, the grammar of language, and then apply those inferred rules to new situations. "Here it's exactly the same thing: we learn the grammar in many different cellular states, and then we go into a particular condition -- it can be a diseased or it can be a normal cell type -- and we can try to see how well we predict patterns from this information," says Rabadan. Fu and Rabadan soon enlisted a team of collaborators, including co-first authors Alejandro Buendia, now a Stanford PhD student formerly in the Rabadan lab, and Shentong Mo of Carnegie Mellon, to train and test the new model. After training on data from more than 1.3 million human cells, the system became accurate enough to predict gene expression in cell types it had never seen, yielding results that agreed closely with experimental data. New AI methods reveal drivers of a pediatric cancer Next, the investigators showed the power of their AI system when they asked it to uncover still hidden biology of diseased cells, in this case, an inherited form of pediatric leukemia. "These kids inherit a gene that is mutated, and it was unclear exactly what it is these mutations are doing," says Rabadan, who also co-directs the cancer genomics and epigenomics research program at Columbia's Herbert Irving Comprehensive Cancer Center. With AI, the researchers predicted that the mutations disrupt the interaction between two different transcription factors that determine the fate of leukemic cells. Laboratory experiments confirmed AI's prediction. Understanding the effect of these mutations uncovers specific mechanisms that drive this disease. AI could reveal "dark matter" in genome The new computational methods should also allow researchers to start exploring the role of genome's "dark matter" -- a term borrowed from cosmology that refers to the vast majority of the genome, which does not encode known genes -- in cancer and other diseases. "The vast majority of mutations found in cancer patients are in so-called dark regions of the genome. These mutations do not affect the function of a protein and have remained mostly unexplored. says Rabadan. "The idea is that using these models, we can look at mutations and illuminate that part of the genome." Already, Rabadan is working with researchers at Columbia and other universities, exploring different cancers, from brain to blood cancers, learning the grammar of regulation in normal cells, and how cells change in the process of cancer development. The work also opens new avenues for understanding many diseases beyond cancer and potentially identifying targets for new treatments. By presenting novel mutations to the computer model, researchers can now gain deep insights and predictions about exactly how those mutations affect a cell. Coming on the heels of other recent advances in artificial intelligence for biology, Rabadan sees the work as part of a major trend: "It's really a new era in biology that is extremely exciting; transforming biology into a predictive science." Additional information The paper, titled "A foundational model of transcription across human cell types," was published Jan. 8 in Nature. Authors (all from Columbia except where noted): Xi Fu, Shentong Mo (Mohamed bin Zayed University of Artificial Intelligence, Abu Dhabi, UAE, and Carnegie Mellon University, Pittsburgh, PA), Alejandro Buendia, Anouchka P. Laurent, Anqi Shao, Maria del Mar Alvarez-Torres, Tianji Yu, Jimin Tan (New York University Grossman School of Medicine, New York, NY), Jiayu Su, Romella Sagatelian, Adolfo A. Ferrando (Columbia and Regeneron, Tarrytown, NY), Alberto Ciccia, Yanyan Lan (Tsinghua University, Beijing, China), David M. Owens Teresa Palomero, Eric P. Xing (Mohamed bin Zayed University of Artificial Intelligence and Carnegie Mellon University), and Raul Rabadan. A foundation model of transcription across human cell types Transcriptional regulation, which involves a complex interplay between regulatory sequences and proteins, directs all biological processes. Computational models of transcription lack generalizability to accurately extrapolate to unseen cell types and conditions. Here we introduce GET (general expression transformer), an interpretable foundation model designed to uncover regulatory grammars across 213 human fetal and adult cell types. Relying exclusively on chromatin accessibility data and sequence information, GET achieves experimental-level accuracy in predicting gene expression even in previously unseen cell types. GET also shows remarkable adaptability across new sequencing platforms and assays, enabling regulatory inference across a broad range of cell types and conditions, and uncovers universal and cell-type-specific transcription factor interaction networks. We evaluated its performance in prediction of regulatory activity, inference of regulatory elements and regulators, and identification of physical interactions between transcription factors and found that it outperforms current models in predicting lentivirus-based massively parallel reporter assay readout. In fetal erythroblasts, we identified distal (greater than 1 Mbp) regulatory regions that were missed by previous models, and, in B cells, we identified a lymphocyte-specific transcription factor-transcription factor interaction that explains the functional significance of a leukaemia risk predisposing germline mutation. In sum, we provide a generalizable and accurate model for transcription together with catalogues of gene regulation and transcription factor interactions, all with cell type specificity.