\n\nTheodoris, Christina V., et al. \"Transfer learning enables predictions in network biology.\" Nature (2023): 1-9.
\n
\n\nKamimoto, Kenji, et al. \"Dissecting cell identity via network inference and in silico gene perturbation.\" Nature 614.7949 (2023): 742-751.
\n
\n", "messages": [ { "id": "2435629", "postDate": "09/13/2023 05:38:28", "content": "Roohani, Yusuf, Kexin Huang, and Jure Leskovec. \"Predicting transcriptional outcomes of novel multigene perturbations with GEARS.\" Nature Biotechnology (2023): 1-9.
\n
\n\nTheodoris, Christina V., et al. \"Transfer learning enables predictions in network biology.\" Nature (2023): 1-9.
\n
\n\nKamimoto, Kenji, et al. \"Dissecting cell identity via network inference and in silico gene perturbation.\" Nature 614.7949 (2023): 742-751.
\n
\n", "rawMarkdown": "1. [Geneformer](https://huggingface.co/ctheodoris/Geneformer)\nTransfer learning enables predictions in network biology\nMapping gene networks requires large amounts of transcriptomic data to learn the connections between genes, which impedes discoveries in settings with limited data, including rare diseases and diseases affecting clinically inaccessible tissues. Recently, transfer learning has revolutionized fields such as natural language understanding1,2 and computer vision3 by leveraging deep learning models pretrained on large-scale general datasets that can then be fine-tuned towards a vast array of downstream tasks with limited task-specific data. Here, we developed a context-aware, attention-based deep learning model, Geneformer, pretrained on a large-scale corpus of about 30 million single-cell transcriptomes to enable context-specific predictions in settings with limited data in network biology. During pretraining, Geneformer gained a fundamental understanding of network dynamics, encoding network hierarchy in the attention weights of the model in a completely self-supervised manner. Fine-tuning towards a diverse panel of downstream tasks relevant to chromatin and network dynamics using limited task-specific data demonstrated that Geneformer consistently boosted predictive accuracy. Applied to disease modelling with limited patient data, Geneformer identified candidate therapeutic targets for cardiomyopathy. Overall, Geneformer represents a pretrained deep learning model from which fine-tuning towards a broad range of downstream applications can be pursued to accelerate discovery of key network regulators and candidate therapeutic targets.\n\n>Theodoris, Christina V., et al. \"Transfer learning enables predictions in network biology.\" Nature (2023): 1-9.\n\n2. [CellOracle](https://morris-lab.github.io/CellOracle.documentation/tutorials/index.html)\nDissecting cell identity via network inference and in silico gene perturbation\nCell identity is governed by the complex regulation of gene expression, represented as gene-regulatory networks1. Here we use gene-regulatory networks inferred from single-cell multi-omics data to perform in silico transcription factor perturbations, simulating the consequent changes in cell identity using only unperturbed wild-type data. We apply this machine-learning-based approach, CellOracle, to well-established paradigms—mouse and human haematopoiesis, and zebrafish embryogenesis—and we correctly model reported changes in phenotype that occur as a result of transcription factor perturbation. Through systematic in silico transcription factor perturbation in the developing zebrafish, we simulate and experimentally validate a previously unreported phenotype that results from the loss of noto, an established notochord regulator. Furthermore, we identify an axial mesoderm regulator, lhx1a. Together, these results show that CellOracle can be used to analyse the regulation of cell identity by transcription factors, and can provide mechanistic insights into development and differentiation.\n\n>Kamimoto, Kenji, et al. \"Dissecting cell identity via network inference and in silico gene perturbation.\" Nature 614.7949 (2023): 742-751.\n\n3. [GEARS](https://github.com/snap-stanford/GEARS)\nPredicting transcriptional outcomes of novel multigene perturbations with GEARS\nUnderstanding cellular responses to genetic perturbation is central to numerous biomedical applications, from identifying genetic interactions involved in cancer to developing methods for regenerative medicine. However, the combinatorial explosion in the number of possible multigene perturbations severely limits experimental interrogation. Here, we present graph-enhanced gene activation and repression simulator (GEARS), a method that integrates deep learning with a knowledge graph of gene–gene relationships to predict transcriptional responses to both single and multigene perturbations using single-cell RNA-sequencing data from perturbational screens. GEARS is able to predict outcomes of perturbing combinations consisting of genes that were never experimentally perturbed. GEARS exhibited 40% higher precision than existing approaches in predicting four distinct genetic interaction subtypes in a combinatorial perturbation screen and identified the strongest interactions twice as well as prior approaches. Overall, GEARS can predict phenotypically distinct effects of multigene perturbations and thus guide the design of perturbational experiments.\n\n>Roohani, Yusuf, Kexin Huang, and Jure Leskovec. \"Predicting transcriptional outcomes of novel multigene perturbations with GEARS.\" Nature Biotechnology (2023): 1-9.", "votes": null }, { "id": "2485808", "postDate": "10/17/2023 13:21:06", "content": "Roohani, Yusuf, Kexin Huang, and Jure Leskovec. \"Predicting transcriptional outcomes of novel multigene perturbations with GEARS.\" Nature Biotechnology (2023): 1-9.
\n
Good work! Useful to learn
", "rawMarkdown": "Good work! Useful to learn", "votes": null } ], "comments": [ { "id": 2485808, "author_name": "hanyangwangnb", "author_url": "", "post_date": "10/17/2023 13:21:06", "content": "Good work! Useful to learn
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "2435629": "1. [Geneformer](https://huggingface.co/ctheodoris/Geneformer)\nTransfer learning enables predictions in network biology\nMapping gene networks requires large amounts of transcriptomic data to learn the connections between genes, which impedes discoveries in settings with limited data, including rare diseases and diseases affecting clinically inaccessible tissues. Recently, transfer learning has revolutionized fields such as natural language understanding1,2 and computer vision3 by leveraging deep learning models pretrained on large-scale general datasets that can then be fine-tuned towards a vast array of downstream tasks with limited task-specific data. Here, we developed a context-aware, attention-based deep learning model, Geneformer, pretrained on a large-scale corpus of about 30 million single-cell transcriptomes to enable context-specific predictions in settings with limited data in network biology. During pretraining, Geneformer gained a fundamental understanding of network dynamics, encoding network hierarchy in the attention weights of the model in a completely self-supervised manner. Fine-tuning towards a diverse panel of downstream tasks relevant to chromatin and network dynamics using limited task-specific data demonstrated that Geneformer consistently boosted predictive accuracy. Applied to disease modelling with limited patient data, Geneformer identified candidate therapeutic targets for cardiomyopathy. Overall, Geneformer represents a pretrained deep learning model from which fine-tuning towards a broad range of downstream applications can be pursued to accelerate discovery of key network regulators and candidate therapeutic targets.\n\n>Theodoris, Christina V., et al. \"Transfer learning enables predictions in network biology.\" Nature (2023): 1-9.\n\n2. [CellOracle](https://morris-lab.github.io/CellOracle.documentation/tutorials/index.html)\nDissecting cell identity via network inference and in silico gene perturbation\nCell identity is governed by the complex regulation of gene expression, represented as gene-regulatory networks1. Here we use gene-regulatory networks inferred from single-cell multi-omics data to perform in silico transcription factor perturbations, simulating the consequent changes in cell identity using only unperturbed wild-type data. We apply this machine-learning-based approach, CellOracle, to well-established paradigms—mouse and human haematopoiesis, and zebrafish embryogenesis—and we correctly model reported changes in phenotype that occur as a result of transcription factor perturbation. Through systematic in silico transcription factor perturbation in the developing zebrafish, we simulate and experimentally validate a previously unreported phenotype that results from the loss of noto, an established notochord regulator. Furthermore, we identify an axial mesoderm regulator, lhx1a. Together, these results show that CellOracle can be used to analyse the regulation of cell identity by transcription factors, and can provide mechanistic insights into development and differentiation.\n\n>Kamimoto, Kenji, et al. \"Dissecting cell identity via network inference and in silico gene perturbation.\" Nature 614.7949 (2023): 742-751.\n\n3. [GEARS](https://github.com/snap-stanford/GEARS)\nPredicting transcriptional outcomes of novel multigene perturbations with GEARS\nUnderstanding cellular responses to genetic perturbation is central to numerous biomedical applications, from identifying genetic interactions involved in cancer to developing methods for regenerative medicine. However, the combinatorial explosion in the number of possible multigene perturbations severely limits experimental interrogation. Here, we present graph-enhanced gene activation and repression simulator (GEARS), a method that integrates deep learning with a knowledge graph of gene–gene relationships to predict transcriptional responses to both single and multigene perturbations using single-cell RNA-sequencing data from perturbational screens. GEARS is able to predict outcomes of perturbing combinations consisting of genes that were never experimentally perturbed. GEARS exhibited 40% higher precision than existing approaches in predicting four distinct genetic interaction subtypes in a combinatorial perturbation screen and identified the strongest interactions twice as well as prior approaches. Overall, GEARS can predict phenotypically distinct effects of multigene perturbations and thus guide the design of perturbational experiments.\n\n>Roohani, Yusuf, Kexin Huang, and Jure Leskovec. \"Predicting transcriptional outcomes of novel multigene perturbations with GEARS.\" Nature Biotechnology (2023): 1-9.", "2485808": "Good work! Useful to learn" }, "source": "meta" }