Pub Date : 2024-11-15DOI: 10.1038/s42256-024-00920-9
Zaixi Zhang, Wan Xiang Shen, Qi Liu, Marinka Zitnik
Designing protein-binding proteins is critical for drug discovery. However, artificial-intelligence-based design of such proteins is challenging due to the complexity of protein–ligand interactions, the flexibility of ligand molecules and amino acid side chains, and sequence–structure dependencies. We introduce PocketGen, a deep generative model that produces residue sequence and atomic structure of the protein regions in which ligand interactions occur. PocketGen promotes consistency between protein sequence and structure by using a graph transformer for structural encoding and a sequence refinement module based on a protein language model. The graph transformer captures interactions at multiple scales, including atom, residue and ligand levels. For sequence refinement, PocketGen integrates a structural adapter into the protein language model, ensuring that structure-based predictions align with sequence-based predictions. PocketGen can generate high-fidelity protein pockets with enhanced binding affinity and structural validity. It operates ten times faster than physics-based methods and achieves a 97% success rate, defined as the percentage of generated pockets with higher binding affinity than reference pockets. Additionally, it attains an amino acid recovery rate exceeding 63%. A generative model that leverages a graph transformer and protein language model to generate residue sequences and full-atom structures of protein pockets is introduced, which outperforms state-of-the-art approaches.
{"title":"Efficient generation of protein pockets with PocketGen","authors":"Zaixi Zhang, Wan Xiang Shen, Qi Liu, Marinka Zitnik","doi":"10.1038/s42256-024-00920-9","DOIUrl":"10.1038/s42256-024-00920-9","url":null,"abstract":"Designing protein-binding proteins is critical for drug discovery. However, artificial-intelligence-based design of such proteins is challenging due to the complexity of protein–ligand interactions, the flexibility of ligand molecules and amino acid side chains, and sequence–structure dependencies. We introduce PocketGen, a deep generative model that produces residue sequence and atomic structure of the protein regions in which ligand interactions occur. PocketGen promotes consistency between protein sequence and structure by using a graph transformer for structural encoding and a sequence refinement module based on a protein language model. The graph transformer captures interactions at multiple scales, including atom, residue and ligand levels. For sequence refinement, PocketGen integrates a structural adapter into the protein language model, ensuring that structure-based predictions align with sequence-based predictions. PocketGen can generate high-fidelity protein pockets with enhanced binding affinity and structural validity. It operates ten times faster than physics-based methods and achieves a 97% success rate, defined as the percentage of generated pockets with higher binding affinity than reference pockets. Additionally, it attains an amino acid recovery rate exceeding 63%. A generative model that leverages a graph transformer and protein language model to generate residue sequences and full-atom structures of protein pockets is introduced, which outperforms state-of-the-art approaches.","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"6 11","pages":"1382-1395"},"PeriodicalIF":18.8,"publicationDate":"2024-11-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.nature.com/articles/s42256-024-00920-9.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142637798","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-14DOI: 10.1038/s42256-024-00914-7
Yunqi Cai, Jiangnan Li, Dong Wang
Important progress has been made in micromagnetics, driven by its wide-ranging applications in magnetic storage design. Numerical simulation, a cornerstone of micromagnetics research, relies on first-principles rules to compute the dynamic evolution of micromagnetic systems using the renowned Landau–Lifshitz–Gilbert equation, named after Landau, Lifshitz and Gilbert. However, these simulations are often hindered by their slow speeds. Although fast Fourier transformation calculations reduce the computational complexity to O(Nlog(N)), it remains impractical for large-scale simulations. Here we introduce NeuralMAG, a deep learning approach to micromagnetic simulation. Our approach follows the Landau–Lifshitz–Gilbert iterative framework but accelerates computation of demagnetizing fields by employing a U-shaped neural network. This neural network architecture comprises an encoder that extracts aggregated spins at various scales and learns the local interaction at each scale, followed by a decoder that accumulates the local interactions at different scales to approximate the global convolution. This divide-and-accumulate scheme achieves a time complexity of O(N), notably enhancing the speed and feasibility of large-scale simulations. Unlike existing neural methods, NeuralMAG concentrates on the core computation—rather than an end-to-end approximation for a specific task—making it inherently generalizable. To validate the new approach, we trained a single model and evaluated it on two micromagnetics tasks with various sample sizes, shapes and material settings. Many physical systems involve long-range interactions, which present a considerable obstacle to large-scale simulations. Cai, Li and Wang introduce NeuralMAG, a deep learning approach to reduce complexity and accelerate micromagnetic simulations.
{"title":"Fast and generalizable micromagnetic simulation with deep neural nets","authors":"Yunqi Cai, Jiangnan Li, Dong Wang","doi":"10.1038/s42256-024-00914-7","DOIUrl":"10.1038/s42256-024-00914-7","url":null,"abstract":"Important progress has been made in micromagnetics, driven by its wide-ranging applications in magnetic storage design. Numerical simulation, a cornerstone of micromagnetics research, relies on first-principles rules to compute the dynamic evolution of micromagnetic systems using the renowned Landau–Lifshitz–Gilbert equation, named after Landau, Lifshitz and Gilbert. However, these simulations are often hindered by their slow speeds. Although fast Fourier transformation calculations reduce the computational complexity to O(Nlog(N)), it remains impractical for large-scale simulations. Here we introduce NeuralMAG, a deep learning approach to micromagnetic simulation. Our approach follows the Landau–Lifshitz–Gilbert iterative framework but accelerates computation of demagnetizing fields by employing a U-shaped neural network. This neural network architecture comprises an encoder that extracts aggregated spins at various scales and learns the local interaction at each scale, followed by a decoder that accumulates the local interactions at different scales to approximate the global convolution. This divide-and-accumulate scheme achieves a time complexity of O(N), notably enhancing the speed and feasibility of large-scale simulations. Unlike existing neural methods, NeuralMAG concentrates on the core computation—rather than an end-to-end approximation for a specific task—making it inherently generalizable. To validate the new approach, we trained a single model and evaluated it on two micromagnetics tasks with various sample sizes, shapes and material settings. Many physical systems involve long-range interactions, which present a considerable obstacle to large-scale simulations. Cai, Li and Wang introduce NeuralMAG, a deep learning approach to reduce complexity and accelerate micromagnetic simulations.","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"6 11","pages":"1330-1343"},"PeriodicalIF":18.8,"publicationDate":"2024-11-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142610316","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-13DOI: 10.1038/s42256-024-00922-7
Sebastian Porsdam Mann, Anuraag A. Vazirani, Mateo Aboy, Brian D. Earp, Timo Minssen, I. Glenn Cohen, Julian Savulescu
In this Comment, we propose a cumulative set of three essential criteria for the ethical use of LLMs in academic writing, and present a statement that researchers can quote when submitting LLM-assisted manuscripts in order to testify to their adherence to them.
{"title":"Guidelines for ethical use and acknowledgement of large language models in academic writing","authors":"Sebastian Porsdam Mann, Anuraag A. Vazirani, Mateo Aboy, Brian D. Earp, Timo Minssen, I. Glenn Cohen, Julian Savulescu","doi":"10.1038/s42256-024-00922-7","DOIUrl":"10.1038/s42256-024-00922-7","url":null,"abstract":"In this Comment, we propose a cumulative set of three essential criteria for the ethical use of LLMs in academic writing, and present a statement that researchers can quote when submitting LLM-assisted manuscripts in order to testify to their adherence to them.","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"6 11","pages":"1272-1274"},"PeriodicalIF":18.8,"publicationDate":"2024-11-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142601024","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-11DOI: 10.1038/s42256-024-00924-5
Thomas Grote, Timo Freiesleben, Philipp Berens
A new class of AI models, called foundation models, has entered healthcare. Foundation models violate several basic principles of the standard machine learning paradigm for assessing reliability, making it necessary to rethink what guarantees are required to establish warranted trust in them.
{"title":"Foundation models in healthcare require rethinking reliability","authors":"Thomas Grote, Timo Freiesleben, Philipp Berens","doi":"10.1038/s42256-024-00924-5","DOIUrl":"10.1038/s42256-024-00924-5","url":null,"abstract":"A new class of AI models, called foundation models, has entered healthcare. Foundation models violate several basic principles of the standard machine learning paradigm for assessing reliability, making it necessary to rethink what guarantees are required to establish warranted trust in them.","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"6 12","pages":"1421-1423"},"PeriodicalIF":18.8,"publicationDate":"2024-11-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142598345","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-08DOI: 10.1038/s42256-024-00923-6
Ruijiang Li, Jiang Lu, Ziyi Liu, Duoyun Yi, Mengxuan Wan, Yixin Zhang, Peng Zan, Song He, Xiaochen Bo
Variational graph encoders effectively combine graph convolutional networks with variational autoencoders, and have been widely employed for biomedical graph-structured data. Lam and colleagues developed a framework based on the variational graph encoder, NYAN, to facilitate the prediction of molecular properties in computer-assisted drug design. In NYAN, the low-dimensional latent variables derived from the variational graph autoencoder are leveraged as a universal molecular representation, yielding remarkable performance and versatility throughout the drug discovery process. In this study we assess the reusability of NYAN and investigate its applicability within the context of specific chemical toxicity prediction. The prediction accuracy—based on NYAN latent representations and other popular molecular feature representations—is benchmarked across a broad spectrum of toxicity datasets, and the adaptation of NYAN latent representation to other surrogate models is also explored. NYAN, equipped with common surrogate models, shows competitive or better performance in toxicity prediction compared with other state-of-the-art molecular property prediction methods. We also devise a multi-task learning strategy with feature enhancement and consensus inference by leveraging the low dimensionality and feature diversity of NYAN latent space, further boosting the multi-endpoint acute toxicity estimation. The analysis delves into the adaptability of the generic graph variational model, showcasing its aptitude for tailored tasks within the realm of drug discovery. Ruijiang Li et al. assess the reproducibility of a variational graph encoder-based framework and examines its reusability for chemical toxicity prediction. It explores how a generalist model can function as a specialist model with adaptation.
{"title":"Reusability report: exploring the utility of variational graph encoders for predicting molecular toxicity in drug design","authors":"Ruijiang Li, Jiang Lu, Ziyi Liu, Duoyun Yi, Mengxuan Wan, Yixin Zhang, Peng Zan, Song He, Xiaochen Bo","doi":"10.1038/s42256-024-00923-6","DOIUrl":"10.1038/s42256-024-00923-6","url":null,"abstract":"Variational graph encoders effectively combine graph convolutional networks with variational autoencoders, and have been widely employed for biomedical graph-structured data. Lam and colleagues developed a framework based on the variational graph encoder, NYAN, to facilitate the prediction of molecular properties in computer-assisted drug design. In NYAN, the low-dimensional latent variables derived from the variational graph autoencoder are leveraged as a universal molecular representation, yielding remarkable performance and versatility throughout the drug discovery process. In this study we assess the reusability of NYAN and investigate its applicability within the context of specific chemical toxicity prediction. The prediction accuracy—based on NYAN latent representations and other popular molecular feature representations—is benchmarked across a broad spectrum of toxicity datasets, and the adaptation of NYAN latent representation to other surrogate models is also explored. NYAN, equipped with common surrogate models, shows competitive or better performance in toxicity prediction compared with other state-of-the-art molecular property prediction methods. We also devise a multi-task learning strategy with feature enhancement and consensus inference by leveraging the low dimensionality and feature diversity of NYAN latent space, further boosting the multi-endpoint acute toxicity estimation. The analysis delves into the adaptability of the generic graph variational model, showcasing its aptitude for tailored tasks within the realm of drug discovery. Ruijiang Li et al. assess the reproducibility of a variational graph encoder-based framework and examines its reusability for chemical toxicity prediction. It explores how a generalist model can function as a specialist model with adaptation.","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"6 12","pages":"1457-1466"},"PeriodicalIF":18.8,"publicationDate":"2024-11-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.nature.com/articles/s42256-024-00923-6.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142596493","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-01DOI: 10.1038/s42256-024-00917-4
Samuel Schmidgall, Ji Woong Kim, Alan Kuntz, Ahmed Ezzat Ghazi, Axel Krieger
The dominant paradigm for end-to-end robot learning focuses on optimizing task-specific objectives that solve a single robotic problem such as picking up an object or reaching a target position. However, recent work on high-capacity models in robotics has shown promise towards being trained on large collections of diverse and task-agnostic datasets of video demonstrations. These models have shown impressive levels of generalization to unseen circumstances, especially as the amount of data and the model complexity scale. Surgical robot systems that learn from data have struggled to advance as quickly as other fields of robot learning for a few reasons: there is a lack of existing large-scale open-source data to train models; it is challenging to model the soft-body deformations that these robots work with during surgery because simulation cannot match the physical and visual complexity of biological tissue; and surgical robots risk harming patients when tested in clinical trials and require more extensive safety measures. This Perspective aims to provide a path towards increasing robot autonomy in robot-assisted surgery through the development of a multi-modal, multi-task, vision–language–action model for surgical robots. Ultimately, we argue that surgical robots are uniquely positioned to benefit from general-purpose models and provide four guiding actions towards increased autonomy in robot-assisted surgery. Schmidgall et al. describe a pathway for building general-purpose machine learning models for robot-assisted surgery, including mechanisms for avoiding risk and handing over control to surgeons, and improving safety and outcomes beyond demonstration data.
{"title":"General-purpose foundation models for increased autonomy in robot-assisted surgery","authors":"Samuel Schmidgall, Ji Woong Kim, Alan Kuntz, Ahmed Ezzat Ghazi, Axel Krieger","doi":"10.1038/s42256-024-00917-4","DOIUrl":"10.1038/s42256-024-00917-4","url":null,"abstract":"The dominant paradigm for end-to-end robot learning focuses on optimizing task-specific objectives that solve a single robotic problem such as picking up an object or reaching a target position. However, recent work on high-capacity models in robotics has shown promise towards being trained on large collections of diverse and task-agnostic datasets of video demonstrations. These models have shown impressive levels of generalization to unseen circumstances, especially as the amount of data and the model complexity scale. Surgical robot systems that learn from data have struggled to advance as quickly as other fields of robot learning for a few reasons: there is a lack of existing large-scale open-source data to train models; it is challenging to model the soft-body deformations that these robots work with during surgery because simulation cannot match the physical and visual complexity of biological tissue; and surgical robots risk harming patients when tested in clinical trials and require more extensive safety measures. This Perspective aims to provide a path towards increasing robot autonomy in robot-assisted surgery through the development of a multi-modal, multi-task, vision–language–action model for surgical robots. Ultimately, we argue that surgical robots are uniquely positioned to benefit from general-purpose models and provide four guiding actions towards increased autonomy in robot-assisted surgery. Schmidgall et al. describe a pathway for building general-purpose machine learning models for robot-assisted surgery, including mechanisms for avoiding risk and handing over control to surgeons, and improving safety and outcomes beyond demonstration data.","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"6 11","pages":"1275-1283"},"PeriodicalIF":18.8,"publicationDate":"2024-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142562164","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-10-30DOI: 10.1038/s42256-024-00912-9
Sergios Gatidis, Marcel Früh, Matthias P. Fabritius, Sijing Gu, Konstantin Nikolaou, Christian La Fougère, Jin Ye, Junjun He, Yige Peng, Lei Bi, Jun Ma, Bo Wang, Jia Zhang, Yukun Huang, Lars Heiliger, Zdravko Marinov, Rainer Stiefelhagen, Jan Egger, Jens Kleesiek, Ludovic Sibille, Lei Xiang, Simone Bendazzoli, Mehdi Astaraki, Michael Ingrisch, Clemens C. Cyran, Thomas Küstner
Automated detection of tumour lesions on positron emission tomography–computed tomography (PET/CT) image data is a clinically relevant but highly challenging task. Progress in this field has been hampered in the past owing to the lack of publicly available annotated data and limited availability of platforms for inter-institutional collaboration. Here we describe the results of the autoPET challenge, a biomedical image analysis challenge aimed to motivate research in the field of automated PET/CT image analysis. The challenge task was the automated segmentation of metabolically active tumour lesions on whole-body 18F-fluorodeoxyglucose PET/CT. Challenge participants had access to a large publicly available annotated PET/CT dataset for algorithm training. All algorithms submitted to the final challenge phase were based on deep learning methods, mostly using three-dimensional U-Net architectures. Submitted algorithms were evaluated on a private test set composed of 150 PET/CT studies from two institutions. An ensemble model of the highest-ranking algorithms achieved favourable performance compared with individual algorithms. Algorithm performance was dependent on the quality and quantity of data and on algorithm design choices, such as tailored post-processing of predicted segmentations. Future iterations of this challenge will focus on generalization and clinical translation. Automating the image analysis process for oncologic whole-body positron emission tomography–computed tomography data is a key area of interest. Gatidis et al. describe the autoPET 2022 challenge, an international competition focused on the segmentation of metabolically active tumour lesions, aiming to advance techniques in the field.
{"title":"Results from the autoPET challenge on fully automated lesion segmentation in oncologic PET/CT imaging","authors":"Sergios Gatidis, Marcel Früh, Matthias P. Fabritius, Sijing Gu, Konstantin Nikolaou, Christian La Fougère, Jin Ye, Junjun He, Yige Peng, Lei Bi, Jun Ma, Bo Wang, Jia Zhang, Yukun Huang, Lars Heiliger, Zdravko Marinov, Rainer Stiefelhagen, Jan Egger, Jens Kleesiek, Ludovic Sibille, Lei Xiang, Simone Bendazzoli, Mehdi Astaraki, Michael Ingrisch, Clemens C. Cyran, Thomas Küstner","doi":"10.1038/s42256-024-00912-9","DOIUrl":"10.1038/s42256-024-00912-9","url":null,"abstract":"Automated detection of tumour lesions on positron emission tomography–computed tomography (PET/CT) image data is a clinically relevant but highly challenging task. Progress in this field has been hampered in the past owing to the lack of publicly available annotated data and limited availability of platforms for inter-institutional collaboration. Here we describe the results of the autoPET challenge, a biomedical image analysis challenge aimed to motivate research in the field of automated PET/CT image analysis. The challenge task was the automated segmentation of metabolically active tumour lesions on whole-body 18F-fluorodeoxyglucose PET/CT. Challenge participants had access to a large publicly available annotated PET/CT dataset for algorithm training. All algorithms submitted to the final challenge phase were based on deep learning methods, mostly using three-dimensional U-Net architectures. Submitted algorithms were evaluated on a private test set composed of 150 PET/CT studies from two institutions. An ensemble model of the highest-ranking algorithms achieved favourable performance compared with individual algorithms. Algorithm performance was dependent on the quality and quantity of data and on algorithm design choices, such as tailored post-processing of predicted segmentations. Future iterations of this challenge will focus on generalization and clinical translation. Automating the image analysis process for oncologic whole-body positron emission tomography–computed tomography data is a key area of interest. Gatidis et al. describe the autoPET 2022 challenge, an international competition focused on the segmentation of metabolically active tumour lesions, aiming to advance techniques in the field.","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"6 11","pages":"1396-1405"},"PeriodicalIF":18.8,"publicationDate":"2024-10-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142541712","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-10-23DOI: 10.1038/s42256-024-00915-6
Jialin He, Lei Xiong, Shaohui Shi, Chengyu Li, Kexuan Chen, Qianchen Fang, Jiuhong Nan, Ke Ding, Yuanhui Mao, Carles A. Boix, Xinyang Hu, Manolis Kellis, Jingyun Li, Xushen Xiong
Gene expression involves transcription and translation. Despite large datasets and increasingly powerful methods devoted to calculating genetic variants’ effects on transcription, discrepancy between messenger RNA and protein levels hinders the systematic interpretation of the regulatory effects of disease-associated variants. Accurate models of the sequence determinants of translation are needed to close this gap and to interpret disease-associated variants that act on translation. Here we present Translatomer, a multimodal transformer framework that predicts cell-type-specific translation from messenger RNA expression and gene sequence. We train the Translatomer on 33 tissues and cell lines, and show that the inclusion of sequence improves the prediction of ribosome profiling signal, indicating that the Translatomer captures sequence-dependent translational regulatory information. The Translatomer achieves accuracies of 0.72 to 0.80 for the de novo prediction of cell-type-specific ribosome profiling. We develop an in silico mutagenesis tool to estimate mutational effects on translation and demonstrate that variants associated with translation regulation are evolutionarily constrained, both in the human population and across species. In particular, we identify cell-type-specific translational regulatory mechanisms independent of the expression quantitative trait loci for 3,041 non-coding and synonymous variants associated with complex diseases, including Alzheimer’s disease, schizophrenia and congenital heart disease. The Translatomer accurately models the genetic underpinnings of translation, bridging the gap between messenger RNA and protein levels as well as providing valuable mechanistic insights for uninterpreted disease variants. A transformer-based approach called Translatomer is presented, which models cell-type-specific translation from messenger RNA expression and gene sequence, bridging the gap between messenger RNA and protein levels as well as providing a mechanistic insight into the genetic regulation of translation.
{"title":"Deep learning prediction of ribosome profiling with Translatomer reveals translational regulation and interprets disease variants","authors":"Jialin He, Lei Xiong, Shaohui Shi, Chengyu Li, Kexuan Chen, Qianchen Fang, Jiuhong Nan, Ke Ding, Yuanhui Mao, Carles A. Boix, Xinyang Hu, Manolis Kellis, Jingyun Li, Xushen Xiong","doi":"10.1038/s42256-024-00915-6","DOIUrl":"10.1038/s42256-024-00915-6","url":null,"abstract":"Gene expression involves transcription and translation. Despite large datasets and increasingly powerful methods devoted to calculating genetic variants’ effects on transcription, discrepancy between messenger RNA and protein levels hinders the systematic interpretation of the regulatory effects of disease-associated variants. Accurate models of the sequence determinants of translation are needed to close this gap and to interpret disease-associated variants that act on translation. Here we present Translatomer, a multimodal transformer framework that predicts cell-type-specific translation from messenger RNA expression and gene sequence. We train the Translatomer on 33 tissues and cell lines, and show that the inclusion of sequence improves the prediction of ribosome profiling signal, indicating that the Translatomer captures sequence-dependent translational regulatory information. The Translatomer achieves accuracies of 0.72 to 0.80 for the de novo prediction of cell-type-specific ribosome profiling. We develop an in silico mutagenesis tool to estimate mutational effects on translation and demonstrate that variants associated with translation regulation are evolutionarily constrained, both in the human population and across species. In particular, we identify cell-type-specific translational regulatory mechanisms independent of the expression quantitative trait loci for 3,041 non-coding and synonymous variants associated with complex diseases, including Alzheimer’s disease, schizophrenia and congenital heart disease. The Translatomer accurately models the genetic underpinnings of translation, bridging the gap between messenger RNA and protein levels as well as providing valuable mechanistic insights for uninterpreted disease variants. A transformer-based approach called Translatomer is presented, which models cell-type-specific translation from messenger RNA expression and gene sequence, bridging the gap between messenger RNA and protein levels as well as providing a mechanistic insight into the genetic regulation of translation.","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"6 11","pages":"1314-1329"},"PeriodicalIF":18.8,"publicationDate":"2024-10-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142488349","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-10-22DOI: 10.1038/s42256-024-00913-8
Yumeng Zhang, Zhikang Wang, Yunzhe Jiang, Dene R. Littler, Mark Gerstein, Anthony W. Purcell, Jamie Rossjohn, Hong-Yu Ou, Jiangning Song
Understanding the mechanisms of T cell antigen recognition that underpin adaptive immune responses is critical for developing vaccines, immunotherapies and treatments against autoimmune diseases. Despite extensive research efforts, accurate prediction of T cell receptor (TCR)–antigen binding pairs remains a great challenge due to the vast diversity and cross-reactivity of TCRs. Here we propose a deep-learning-based framework termed epitope-anchored contrastive transfer learning (EPACT) tailored to paired human CD8+ TCRs. Harnessing the pretrained representations and co-embeddings of peptide–major histocompatibility complex (pMHC) and TCR, EPACT demonstrated generalizability in predicting binding specificity for unseen epitopes and distinct TCR repertoires. Contrastive learning enabled highly precise predictions for immunodominant epitopes and interpretable analysis of epitope-specific T cells. We applied EPACT to SARS-CoV-2-responsive T cells, and the predicted binding strength aligned well with the surge in spike-specific immune responses after vaccination. We further fine-tuned EPACT on structural data to decipher the residue-level interactions involved in TCR–antigen recognition. EPACT was capable of quantifying interchain distance matrices and identifying contact residues, corroborating the presence of TCR cross-reactivity across multiple tumour-associated antigens. Together, EPACT can serve as a useful artificial intelligence approach with important potential in practical applications and contribute towards the development of TCR-based immunotherapies. Accurate prediction of T cell receptor (TCR)–antigen recognition remains a challenge. Zhang et al. propose a contrastive transfer learning model to predict TCR–pMHC binding that enables interpretable analyses of epitope-specific T cells and can decipher residue-level interactions.
了解支撑适应性免疫反应的 T 细胞抗原识别机制对于开发疫苗、免疫疗法和治疗自身免疫性疾病至关重要。尽管开展了大量研究工作，但由于 TCR 的多样性和交叉反应性，准确预测 T 细胞受体（TCR）与抗原的结合对仍然是一项巨大的挑战。在这里，我们提出了一种基于深度学习的框架，称为表位锚定对比转移学习（EPACT），专门针对成对的人类 CD8+ TCR。利用肽-主要组织相容性复合体（pMHC）和TCR的预训练表征和共嵌入，EPACT在预测未知表位和不同TCR复合物的结合特异性方面展示了通用性。对比学习可以对免疫优势表位进行高度精确的预测，并对表位特异性 T 细胞进行可解释的分析。我们将 EPACT 应用于 SARS-CoV-2 反应性 T 细胞，预测的结合强度与接种疫苗后尖峰特异性免疫反应的激增非常吻合。我们根据结构数据进一步微调了 EPACT，以破译 TCR 与抗原识别中涉及的残基级相互作用。EPACT 能够量化链间距离矩阵并识别接触残基，从而证实多种肿瘤相关抗原之间存在 TCR 交叉反应。总之，EPACT可以作为一种有用的人工智能方法，在实际应用中具有重要潜力，并有助于开发基于TCR的免疫疗法。
{"title":"Epitope-anchored contrastive transfer learning for paired CD8+ T cell receptor–antigen recognition","authors":"Yumeng Zhang, Zhikang Wang, Yunzhe Jiang, Dene R. Littler, Mark Gerstein, Anthony W. Purcell, Jamie Rossjohn, Hong-Yu Ou, Jiangning Song","doi":"10.1038/s42256-024-00913-8","DOIUrl":"10.1038/s42256-024-00913-8","url":null,"abstract":"Understanding the mechanisms of T cell antigen recognition that underpin adaptive immune responses is critical for developing vaccines, immunotherapies and treatments against autoimmune diseases. Despite extensive research efforts, accurate prediction of T cell receptor (TCR)–antigen binding pairs remains a great challenge due to the vast diversity and cross-reactivity of TCRs. Here we propose a deep-learning-based framework termed epitope-anchored contrastive transfer learning (EPACT) tailored to paired human CD8+ TCRs. Harnessing the pretrained representations and co-embeddings of peptide–major histocompatibility complex (pMHC) and TCR, EPACT demonstrated generalizability in predicting binding specificity for unseen epitopes and distinct TCR repertoires. Contrastive learning enabled highly precise predictions for immunodominant epitopes and interpretable analysis of epitope-specific T cells. We applied EPACT to SARS-CoV-2-responsive T cells, and the predicted binding strength aligned well with the surge in spike-specific immune responses after vaccination. We further fine-tuned EPACT on structural data to decipher the residue-level interactions involved in TCR–antigen recognition. EPACT was capable of quantifying interchain distance matrices and identifying contact residues, corroborating the presence of TCR cross-reactivity across multiple tumour-associated antigens. Together, EPACT can serve as a useful artificial intelligence approach with important potential in practical applications and contribute towards the development of TCR-based immunotherapies. Accurate prediction of T cell receptor (TCR)–antigen recognition remains a challenge. Zhang et al. propose a contrastive transfer learning model to predict TCR–pMHC binding that enables interpretable analyses of epitope-specific T cells and can decipher residue-level interactions.","PeriodicalId":48533,"journal":{"name":"Nature Machine Intelligence","volume":"6 11","pages":"1344-1358"},"PeriodicalIF":18.8,"publicationDate":"2024-10-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142486862","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：